article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
the magnetic moment is an important characteristic of elementary particles , but the magnetic moments of many particles ( e.g. , charm and beauty baryons ) are , as yet , not measured .
this is because for particles with a short lifetime @xmath0 ( @xmath1s for @xmath2 and @xmath3s for @xmath4 ) , the decay length @xmath5 cm if the energy acquired through the production reaction equals 1 tev @xcite .
for this reason , the anomalous magnetic moments of short - lived particles can not be measured with conventional methods .
the existence of the spin rotation phenomenon for high - energy particles moving in bent crystals in the channeling regime was first established in @xcite .
the spin rotation angle is determined by the anomalous magnetic moment @xmath6 and , according to @xcite , can attain values as high as many radians .
the unique feature of the effect is that for lorentz - factors @xmath7 , the value of the limit spin rotation angle per unit length of 1 cm in a crystal with a limit radius of curvature @xmath8 at which the particle still moves in the channeling regime without being expelled from the channel by the centrifugal force is independent of particle energy .
the idea advanced in @xcite was experimentally verified and confirmed for @xmath9 hyperons in fermilab @xcite .
a detailed analysis of the experiment @xcite given in @xcite suggests the feasibility of using the spin rotation effect in bent crystals to measure the anomalous magnetic moment of short - lived charm baryons @xmath2 and @xmath4 . in @xcite , a thorough consideration
was given to charm baryons @xmath2 produced by a beam of protons whose momentum in a tungsten target was @xmath10gev / s .
the characteristic momentum of the produced @xmath2 was @xmath11 gev / s ( characteristic lorentz factor @xmath12 ) . at the same time , it was stated in @xcite that since the beauty quark has a negative charge , the beauty baryons have a negative charge , too , and can not be channeled .
as a result , the spin rotation phenomena for them does not exists .
positively - charged antiparticles could be channeled , but the production rates at lhc energies would be too small . it is noteworthy here that with growing particle energy the dechanneling length is increased not only for positively - charged particles , but also for their negatively charged counterparts .
what is more , successful experiments recently demonstrated the possibility to deflect negatively - charged particles at gev energies @xcite .
for this reason , in light of the projected increase in energy at lhc and fcc , it seems pertinent to return to the issue of using the spin rotation effect in bent crystals for measuring the magnetic moment of not only positively- but also negatively - charged particles .
the effect of spin depolarization of high - energy particles scattered by crystal axes ( planes ) provides another possibility of measuring the anomalous magnetic moment of short - lived , high - energy particles @xcite .
what is more , an important advantage of the spin depolarization effect is that it applies to both positively- and negatively - charged particles .
thus , this effect enables magnetic moment measurements of not only positively - charged baryons , but also beauty baryons , whose negative charge and rapid dechanneling make such measurements quite a challenge even in the range of high energies .
this paper considers the effects of particle spin rotation and spin depolarization in crystals at lhc and fcc energies .
it is shown that these effects can be used for measuring anomalous magnetic moments of short - lived particles in this range of energies .
it is also demonstrated that the effect of spin depolarization of high - energy particles scattered by crystal axes ( planes ) can be used effectively for measuring the anomalous magnetic moments of negatively - charged beauty baryons , too .
because the depolarization effect is possible at wider angles as compared with the angles of capture into channeling regime ( determined by the lindhard angle ) , the number of usable events in this case appears to be greater than in the experiments with bent crystals , thus giving us hope for measuring the anomalous magnetic moment of beauty baryons . before proceeding to the consideration of the possibility to use the phenomena of spin rotation and depolarization of high - energy particles moving in bent or straight crystals for measuring the anomalous magnetic moment of charm and beauty baryons ,
let us recall briefly the theory underlying these phenomena , as it is stated in @xcite .
in 1976 , e. tsyganov @xcite demonstrated that high - energy particles may move in the channeling regime in a bent crystal , thus moving along the curved path and deviating from the initial direction . in 1979 ,
the predicted effect was experimentally verified in jinr ( dubna , russia ) @xcite : the proton beam with the kinetic energy of 8.4gev was rotated through the angle of 26mrad by means of the silicon crystal of length 2.0 cm . at present , the experiments on channeling in bent crystals are performed at the world s largest accelerators . in these experiments a great variety of crystal optical elements are widely used to manipulate the beams of high - energy particles .
the next step forward was made in 1979 by v. baryshevsky , who demonstrated @xcite that the spin of a particle moving in a bent crystal rotates with respect to the momentum direction figure ( [ fig16.1 ] ) .
= 6 cm indeed , in a bent crystal a particle moves along a curved path under the action of the electric field @xmath13 induced by the crystallographic plane . in the instantaneous rest frame of the particle , due to relativistic effects , this field produces a magnetic field @xmath14 , which acts on the particle magnetic moment , thus causing spin rotation in this field ( figure [ fig16.1 ] ) . owing to an extremely large magnitude of the field @xmath15 cgse and , as a result , a large magnitude of the field @xmath16 gauss , the spin rotation angle in the crystal of several centimeters in length can reach several radians .
now , let us proceed to a more detailed consideration of the effect of particle spin rotation in a bent crystal . from the bargmann - michel - telegdi ( bmt )
equation @xcite follows that as the energy of particles increases , their spin precession frequency in the external magnetic field changes , being determined in the ultrarelativistic case by the anomalous magnetic moment . as a result , for example
, in a magnetic field @xmath17gs , the electron ( proton ) spin precession frequency @xmath18s@xmath19 ( @xmath20 is the anomalous part of the magnetic moment ) , and the spin rotation angle per one centimeter path length @xmath21 is just @xmath22 as mentioned above , a relativistic particle in a bent crystal `` senses '' a much greater magnetic field . as the particle energy is increased (
particle wavelength is decreased ) , the quasi - classical character of particle motion enables one to apply the laws of motion determined by classical mechanics .
if a crystal is nonmagnetic , the bmt equation for the spin polarization vector @xmath23 can be written in the form @xcite @xmath24\,,\ ] ] where @xmath13 is the electric field at the point of particle location and @xmath25 ( where @xmath26 is the particle velocity ) .
let us recall here that vector @xmath27 characterizes the polarization of a particle in its `` instantaneous '' rest frame .
intracrystalline fields @xmath13 are large , reaching the values of @xmath28 cgse and even greater .
therefore from ( [ chan_25.1 ] ) follows that for constant intracrystalline fields , the spin precession frequency could reach @xmath29s@xmath19 and the angle @xmath30 could be of the order of @xmath31rad / cm .
however , when a particle moves through a crystal in arbitrary direction , the field @xmath13 , in a similar manner as in the amorphous medium , takes on random values at the particle location point . as a consequence ,
such a field causes spin depolarization . under channeling conditions
, the situation is basically different .
if the crystal bending radius is @xmath32 , then , for example , a beam of high energy protons will change its direction following the crystal bend @xcite , i.e. , the particle will move along a curved path .
the stated motion is due to a constant mean electric field acting on a particle in a bent crystal
. equation ( [ chan_25.1 ] ) for a particle moving in a crystal , for example , in a planar channel bent to a radius of curvature @xmath32 around the @xmath33-axis , has a form ( @xmath34 , @xmath35 the trajectory lies in the @xmath36 , @xmath37 plane ) @xmath38 the position vector @xmath39 of a particle in such a channel rotates about the @xmath33-axis with the frequency @xmath40 .
its magnitude oscillates about the particle equilibrium position @xmath41 in the channel with the frequency @xmath42 , the amplitude @xmath43 , and the initial phase @xmath44 . in the explicit form @xmath45 , @xmath46 , and @xmath47 .
we point out that , due to the presence of centrifugal forces in a bent crystal , the equilibrium point @xmath41 does not coincide with the position @xmath32 of the minimum of the electrostatic potential @xmath48 of the channel . for example , for a particle moving in a harmonic well @xmath49 where the constant @xmath50 is determined by the crystal properties .
integration of ( [ chan_25.2 ] ) in the polar coordinate system gives ( @xmath51 , @xmath52 ) : @xmath53 for a harmonic well we can write ( [ chan_25.3 ] ) accurate up to the terms of the order @xmath54 and @xmath55 in the form @xmath56+\arctan\frac{\zeta_{x}(0)}{\zeta_{z}(0)}\right\}\,,\ ] ] where @xmath57 and @xmath58 is the electric field at the location point of the particle center of equilibrium in the bent crystal ; @xmath59 the coefficient @xmath60 in ( [ chan_25.4 ] ) is small ( for si , the coefficient @xmath61ev/@xmath62 , @xmath63s@xmath19 for protons with the energy @xmath64gev , as a result , @xmath65 ) . neglecting the term containing @xmath60
, we obtain that the spin rotates with frequency @xmath66 ( with growing energy @xmath67 , the coefficient @xmath68 increases , and the spin rotation turns into oscillations at frequency @xmath66 and the frequencies multiple of @xmath42 ) . due to a large magnitude of the field
@xmath69 curving the particle trajectory ( @xmath70 cgse ) , the frequency @xmath71s@xmath19 , and the rotation angle @xmath72rad / cm . if the radius of curvature @xmath73 ( a straight channel ) , then only spin oscillations due to the term containing @xmath60 remain . in this case , a significant spin rotation occurs only at high energies : for @xmath74tev protons , for example , @xmath75 .
thus , according to ( [ chan_25.1 ] ) , the spin precession frequency of a particle moving in a bent crystal is @xmath76 as stated above , this expression for the frequency follows immediately if we consider what field acts on the spin in the particle s instantaneous rest frame . as a result of relativistic transformations ,
the electric field that is transverse relative to the particle velocity in the particle rest frame generates a magnetic field @xmath77 that is orthogonal to @xmath78 and has a magnitude equal to @xmath79 .
the spin precession frequency associated with the anomalous magnetic moment in the particle rest frame is @xmath80 the spin precession frequency in the laboratory frame is @xmath81 it is noteworthy that with zero particle s anomalous magnetic moment @xmath6 , the particle spin precession frequency would equal the orbital rotation frequency of the momentum , and the particle spin direction would follow that of the momentum . if @xmath82 , the angle between the polarization and momentum direction vectors will change .
the angle @xmath83 of particle spin rotation relative to the particle momentum direction is @xmath84 , where the time @xmath85 , with @xmath86 being the path length traveled by the channeled particle in the bent crystal . from this follows that the rotation angle per unit path length @xmath87 let us also remember that the magnetic moment of the particle with spin @xmath88 is related to the gyromagnetic ( land ) factor @xmath89 ( see , e.g. , @xcite ) as @xmath90 where @xmath91 is the particle mass .
when @xmath92 we can write @xmath93 where @xmath94 is the bohr magneton .
consequently , we have @xmath95 here @xmath96 is the force responsible for rotating the particle momentum that should be equal to the centrifugal force @xmath97 ; @xmath98 , where @xmath99 is the radius of curvature of the channel . from this
we have @xmath100 we shall take into account that @xmath101 equals the value of the momentum s angle of rotation @xmath102 .
hence , @xmath103 equation ( [ e16 ] ) was derived by v. lyuboshitz using the bmt equation @xcite .
it follows from ( [ e2 ] ) that the spin rotation angle per unit path length ( the angle of rotation per 1 cm ) is @xmath104 on the other hand , the equality @xmath105 yields @xmath106 . the quantity @xmath107 , where @xmath108 is the particle s potential energy in the channel .
as a consequence , we have @xmath109 the minimum value of the radius of curvature ( the critical radius ) @xmath8 required for the motion in a bent channel to be still possible for ceratin @xmath110 is achieved at maximum @xmath111 . consequently
, the maximum rotation angle per unit * length * is @xmath112 as is seen , the maximum spin rotation angle @xmath113 is energy - independent .
however , at @xmath114 , the capturing acceptance ( the fraction of particles captured into channeling regime ) vanishes @xcite . for certain @xmath115 , the spin rotation angle can be written in the form @xmath116 as a result , we can see that for the given ratio @xmath117 , the angle @xmath118 is energy - independent , too .
the dechanneling length increases as the energy is increased @xcite .
this enables us to increase the rotation angle by increasing the length of the bent crystal channel @xmath119 where @xmath120 is the dechanneling length . for protons with energies higher than @xmath121 gev , the length @xmath122 cm .
according to @xcite , channeling of protons in 15-cm - long si crystals is currently available .
as reported in @xcite , for si , @xmath123 gev cm@xmath19 , and hence the rotation angle per 1 cm path length in a si crystal is @xmath125 according to the estimates @xcite , for a particle with mass @xmath126 gev , the @xmath89-factor lies in the range @xmath127 . from this
we get the minimum value of the ratio @xmath128 .
taking @xmath129 ( in which case the dechanneling length is maximum @xcite ) , we obtain that for @xmath2 , the rotation angle per unit length in a bent silicon crystal equals @xmath130 for @xmath131 , we have @xmath132rad ; for @xmath133 cm , we have @xmath134rad .
thus , if the ratio @xmath135 remains constant , the value of the spin rotation angle appears to be constant and rather large in a wide range of energies .
the decrease in the lindhard angle @xmath136 with increasing energy seemingly impairs the observation conditions .
however , as the energy is increased , the angular width @xmath137 in which the produced particles move is @xmath138 , and hence the fraction of particles captured into the channel is increased with energy as @xmath139 . according to @xcite , as the energy is increased , the probability of @xmath2 production also increases . for this reason , the increase in energy leads to an increasing number of usable events , i.e. , the number of deflected @xmath2 entering the detector .
for example , if the energy of @xmath2 is increased from 300 gev to 3 tev , the number of produced @xmath2 increases tenfold if the linear growth of the production cross section for @xmath2 , suggested in @xcite , continues as the energy of the incident protons rises .
moreover , the amount of @xmath2 captured into channeling regime is increased threefold .
as a result , the time required to observe the effect is reduced .
when a particle travels through a crystal , the field @xmath140 at the point of particle location takes on random values .
this is due to fluctuations of the position of the nucleus ( electron ) in the crystal ( in particular , under thermal vibrations ) and to fluctuations of the electric field that come from particle scattering by various atomic chains .
the fluctuations of the field @xmath13 lead to the stochastic spin precession frequency in the electric field . owing to the presence of fluctuations of the spin rotation frequency , the spin rotation angle of any particle that has traveled through a crystal takes on different values . as a result , after passing through a target ,
every particle has a different direction of the polarization vector @xmath141 , and the distribution of the particle spin direction in the beam may even become isotropic , i.e. , the beam polarization vector goes to zero .
the stochasticity of @xmath66 leads to spin depolarization of particles moving in the crystal @xcite . according to @xcite , depolarization of particles in the case of scattering by crystal planes ( axes )
is determined by the intensity of multiple scattering by crystal planes ( axes ) .
it was shown that depolarization has practically no effect on the possibility of measuring the magnetic moment of particles moving in bent crystals in channeling regime .
a much stronger depolarization is observed for the particles that are not involved in the channeling regime .
these particles may undergo collisions with atomic chains as a whole .
this leads to a sharp increase in the root - mean - square angle of multiple scattering and , as a consequence , to a large depolarization .
for example , according to @xcite , the longitudinal polarization @xmath142 of particles undergoing scattering by the axes decreases as @xmath143 @xmath144 where @xmath145 is the mean square angle of multiple scattering of the particle by the crystal axes per 1 centimeter path length in the crystal . according to @xcite , the mean square angle of particle scattering by the crystal axes for a path length @xmath21 traveled in the crystal
can be written as @xmath146 so we have @xmath147 where @xmath148 , @xmath149 is the atomic number , @xmath150 is the number of atoms per @xmath151 , @xmath152 is the numerical coefficient equal to 1.16 for the axial potential in the lindhard model , and @xmath153 is the angle between the direction of the particle momentum and the crystallographic axis ; @xmath154{z}\,,\ ] ] where @xmath155 is the screening radius , @xmath156 is the distance between the axes , and the @xmath37-axis is parallel to the crystallographic axis ; @xmath157 is the ratio between the charge of a particle and that of an electron .
for example , for particles with the parameters of the proton ( i.e. , @xmath158 , @xmath159mev ) , we have for w - type crystals the magnitude @xmath160 and , according to the value of @xmath153 , it may reach @xmath161 , i.e. , for crystal thickness of the order of 1 cm , the depolarization may amount to a few tens of percent .
equation ( [ nim90_1.40 ] ) is valid for @xmath153 , which are larger than the lindhard angle @xmath162 ( @xmath163 is the averaged axial potential ) .
it follows from ( [ nim90_1.39 ] ) and ( [ ins_nim90_1.40 ] ) that @xmath164 when the particle energy corresponds to @xmath7 , from ( [ nim90_1.39 ] ) and ( [ ins_nim90_1.40 ] ) we may obtain the following expression for @xmath165 @xcite : @xmath166 let us note that for electrons and positrons , as reported in @xcite , the experiments on spin rotation in bent crystals provide a unique possibility for studying the effects of quantum electrodynamics of a strong field , namely , the dependence of the anomalous magnetic moment , i.e. , @xmath167 , on the crystal electric field intensity and the particle energy .
the same possibility exists in studying the depolarization of @xmath168 .
thus , with experimentally determined values of @xmath169 and @xmath170 , we can find the value of ( @xmath171 ) , which gives us another possibility to measure the anomalous magnetic moment of particles .
it is noteworthy that strong fluctuating fields act on neutral particles moving at a small angle to the crystal axes ( planes ) , too , leading to the depolarization of their spin , which fact can be used to measure the magnetic moment of neutral partticles .
we should pay attention to the fact that theoretical considerations and experimental results both reveal that @xmath172 produced through collisions between nonpolarized nucleons and nonpolarized nucleons ( nuclei ) are polarized , and their polarization vector is oriented orthogonally to the reaction plane , i.e. , the plane where the momenta of the incident nucleon and the produced baryon lie .
a similar situation would occur in production of other baryons charm and beauty .
this is why the reaction amplitude @xmath173 includes the term of the form @xmath174\ ] ] where @xmath175 is the spin of the produced particle , @xmath176 is the momentum of the incident particle , and @xmath177 is the momentum of the produced baryon .
consequently , the polarization component @xmath178 of the produced baryon that is parallel to the momentum @xmath177 equals zero .
depolarization occurs to the polarization vector component @xmath179 lying in the plane transverse relative to vector @xmath177 .
let the produced baryon be incident on the crystal axis at a small angle @xmath180 ( according to @xcite , @xmath181{z}$ ] ) . the baryon s velocity component parallel to the axes @xmath182 , where @xmath183 is the particle velocity .
the component @xmath184 of the particle s initial velocity is orthogonal to the axis , and @xmath185 .
let us choose the @xmath36-axis along vector @xmath186 .
multiple scattering leads to appearing of the velocity s component @xmath187 that is orthogonal to the @xmath36-axis , and the changes in @xmath187 are much greater than those in @xmath188 in the direction of the @xmath36-axis . as a consequence , multiple scattering by the axes is strongly asymmetric in the plane perpendicular to the direction of the baryon momentum @xcite .
this means that the depolarization is also asymmetric in this plane , i.e. , @xmath189 .
for this reason , the maximum depolarization of the produced baryons can be obtained if the crystal axes are oriented in such a way that the initial vector of baryon polarization is oriented in the direction of the @xmath33-axis . according to @xcite , the component @xmath190 of the polarization vector relates to @xmath21 as @xmath191.\ ] ]
let us recall that in view of the analysis @xcite , the contributions to multiple scattering by the axes come not only from the averaged potential , but also from particle scattering by the fluctuations of the potential of the axes that are due to thermal oscillations of atoms . as a result , @xmath192 @xcite , where @xmath193 is the mean square angle of multiple scattering due to thermal oscillations and @xmath194 is the mean square angle of multiple scattering due to scattering by averaged over thermal oscillations potential of the axes .
let us also note that @xmath195 , that is , @xmath196 is independent of the particle energy .
it also follows from ( [ new2 ] ) that @xmath197 in our further estimations we shall consider only the contribution coming from multiple scattering by the averaged over thermal oscillations potential of the axes @xmath198 and assume that @xmath199 . as a result
, we have the following expression for the damping rate @xmath200 where @xmath201 is defined by ( [ nim90_1.40 ] ) .
thus , the depolarization rate in this case is similar to that in ( [ nim90_1.39 ] ) .
now let @xmath2 be incident on the crystal .
for our estimate we use the value of the @xmath89-factor given in @xcite : @xmath202 .
the minimum value is estimated as @xmath203 .
so we have the following estimate for @xmath201 in tungsten ( w ) : @xmath204 ( the beam is assumed to be incident on the axes at an angle of about several lindhard angles ) ; then for the length @xmath205 cm , for example , @xmath206 .
hence , the polarization in this case decreases by a factor of @xmath207 , i.e. , quite appreciably .
more precise estimates require thorough simulations , but even the estimates given here indicate that this phenomenon can hopefully be used for negatively - charged beauty baryons .
now , let us note that when moving in the crystal , such particles as baryons ( antibaryons and nuclei ) take part not only in electromagnetic , but also in strong interactions . as a result ,
the spatially periodic effective potential energy of interaction between a particle and a crystal is determined by two interactions : coulomb and strong @xcite .
the effective periodic potential @xmath208 can be introduced to describe the passage of a coherent wave in the crystal : @xmath209 where @xmath210 is the reciprocal lattice vector of the crystal , @xmath211 here @xmath212 is the volume of the crystal elementary cell , @xmath213 is the coordinate of the atom ( nucleus ) of type @xmath214 in the crystal elementary cell , and the squared @xmath215 equals the thermal - factor ( i.e. , the debye - waller factor ) , well - known for x - ray scattering @xcite : @xmath216 where @xmath217 is the mass of the incident particle , @xmath110 is its lorentz - factor , @xmath218 is the amplitude of elastic coherent scattering of the particle by the atom , @xmath219 is the wave vector of the incident wave , and @xmath220 is the wave vector of the scattered wave . it will be recalled that in the case of a crystal , the imaginary part of the amplitude @xmath221 does not contain the contribution from the total cross section of elastic coherent scattering . the imaginary part of @xmath221 in a crystal is only determined by the total cross sections of inelastic processes .
this occurs because in a crystal , unlike in amorphous matter , the wave elastically scattered at a nonzero angle , due to rescattering by periodically located centers , is involved in formation of a coherent wave propagating through the crystal .
elastic coherent scattering of a particle by an atom is caused by both the coulomb interaction of the particle with the atom s electrons and nucleus and its nuclear interaction with the nucleus .
therefore , the scattering amplitude can be presented as a sum of two amplitudes : @xmath222 where @xmath223 is the amplitude of particle scattering caused by the coulomb interaction with the atom ( nucleus ) ( it contains contributions from the coulomb interaction of the particle with the atom along with the spin - orbit interaction with the coulomb field of the atom ( nucleus ) ) ; @xmath224 is the amplitude of elastic coherent scattering of the particle caused by the nuclear interaction ( this amplitude contains the terms independent of the incident particle spin along with the terms depending on spin of both the incident particle and nucleus , in particular , spin - orbit interaction ) .
therefore , @xmath208 and @xmath225 can also be expressed : @xmath226 let us suppose that a high energy particle moves in a crystal at a small angle to the crystallographic planes ( axes ) close to the lindhard angle @xmath227 ( in a relativistic case @xmath228 ) , where @xmath78 is the energy of the particle , @xmath163 is the height of the potential barrier created by the crystallographic plane ( axis ) .
this motion is determined by the plane ( axis ) potential @xmath229 , which could be derived from @xmath230 by averaging over the distribution of atoms ( nuclei ) in the crystal plane ( axis ) . as a consequence ,
the potential @xmath231 for a particle channeled in a plane ( or axis ) channel or moving over the barrier at a small angle , close to the lindhard angle , can be expressed as a sum @xcite : @xmath232 where @xmath233 is the potential energy of particle coulomb interaction with the crystallographic plane ( axis ) , @xmath234 is the energy of spin - orbit interaction with the coulomb field of the plane ( axis ) and spin - orbit nuclear interaction with the effective nuclear field of the plane ( axis ) , and @xmath235 is the effective potential energy of nuclear interaction of the incident particle with the crystallographic plane ( axis ) .
@xmath236 where @xmath237 is the density of nuclei at point @xmath238 of the crystallographic plane ( axis ) , the @xmath37-axis is directed along the crystal axis ( along the velocity component parallel to the crystallographic plane in the case of plane channeling ) , @xmath239 is the amplitude of elastic coherent forward scattering , which depends on the incident particle spin and nucleus polarization , @xmath240 is the energy of magnetic interaction of the particle with the magnetic field produced by the electrons ( nuclei ) .
thus in describing particle motion in crystals , the contribution of strong interactions to the formation of the effective potential acting on the particle from the crystallographic planes ( axes ) should be considered along with the coulomb interaction . in the case of high energies , the particle motion in the potential @xmath241
can be described in the quasi - classical approximation .
the spin - evolution equations for a particle moving in straight and bent crystals in the presence of the contribution from @xmath242 is discussed in @xcite .
let us assume now that the crystal nuclei are polarized .
the cross section of high - energy particle scattering by a nucleus ( the amplitude of zero - angle scattering ) in a crystal with polarized nuclei depends on the particle spin orientation with respect to the polarization of the target .
therefore , the absorption coefficient also depends on particle spin orientation .
let a particle move in a crystal at a small angle to the crystallographic plane ( axis ) close to the lindhard angle .
in this case the coefficient of particle absorption by nuclei can differ from the absorption coefficient for the particle moving in the crystal at a large angle to the crystallographic plane ( axis ) .
moreover , it can appear to be greater even for positively - charged particles .
let us discuss the possibility to use a polarized crystal to obtain a polarized beam of high - energy particles and to analyze the polarization state of high - energy particles @xcite .
let us consider a positively - charged , high - energy particle moving close to the top of the potential barrier .
therefore , it moves in the range with nuclei density higher than that of an amorphous medium . in this range ,
the growth is described by @xmath243 here @xmath244 is the lattice period , @xmath245 is the amplitude of the nucleus thermal oscillations . for the over - barrier motion in the vicinity of the crystallographic axis ,
the growth is described by the relation @xmath246 after passing through the polarized crystal of thickness @xmath86 , the particles moving in the region where the crystal nuclei are located acquire the polarization degree that can be described by formula @xmath247 where @xmath248 is the total scattering cross section for the particle whose spin is parallel to the polarization vector of nuclei and @xmath249 is the total scattering cross section for the particle whose spin is anti - parallel to the polarization vector of nuclei . to avoid misunderstanding
, it will be recalled that the exponential absorption law holds true if the characteristic scattering angle of a particle is much larger than the angle at which the observer from the point of target location may see the input window of the detector , registering the flux of particles transmitted through the target . in the opposite case
all of the scattered particles enter the detector , and by @xmath250 one should mean the total cross section of absorption and reactions that remove the particle from the incident beam ( causing appreciable change in its energy ) .
the density matrix techniques should be used as a tool in considering the general case .
the magnitude of the effect is determined by the parameter @xmath251 . with @xmath252 being the average density of nuclei in the crystal , @xmath253 for a plane and @xmath254 for an axis .
the length @xmath255 corresponding to @xmath256 is @xmath257 . if we suppose that @xmath258 and @xmath259@xmath260 , then for a plane @xmath261 cm and for an axis @xmath262 cm .
as is seen , in this case a beam of polarized particles can be obtained using rather small - sized polarized targets .
let positively - charged particles move parallel to a crystallographic plane . in this case , most of them appear close to the bottom of the potential well and far from the nuclei , therefore , the absorption coefficient for this fraction of particles is smaller comparing to the average crystal absorption
nevertheless , in this case there is also a fraction of particles that moves in the region where the nuclei are located .
a fraction of particles moving near the barrier can be sorted out by means of angular distribution or by using bent crystals . moving in a bent crystal ,
the fraction of channeled particles that due to coulomb repulsion is hardly involved in the nuclear interaction is deflected from the initial direction of motion .
thus the particles that have undergone the interaction with the nuclei will change the beam moving in the initial direction .
because the absorption for this fraction of particles increases , it becomes possible to achieve quite a significant polarization of the initially nonpolarized beam .
a bent crystal can be used to increase the interaction of a positively - charged particle with a nucleus , too . in this case
, the centrifugal forces push the channeled beam to the range with high density of nuclei . a detailed analysis of the polarization state of transient particles
requires a numerical simulation .
in contrast to positively charged particles , negatively charged particles ( for example , @xmath263 hyperons , antiprotons , and beauty hyperons ) , being channeled , move in the region with high density of nuclei , therefore , even a very thin polarized crystal ( @xmath264 cm ) can be an effective polarizer and polarization analyzer .
moreover , for a negatively - charged channeled particle , the angle of spin rotation in the nuclear pseudo - magnetic field of the polarized crystal @xcite can be larger than that for amorphous matter @xcite .
this is the reason for the thin polarized crystals to be used as nuclear - optical elements that can guide polarization of high - energy particle beams .
note that because of the presence of the amplitude similar to that in ( [ new1 ] ) , the spin - orbit interaction also leads to particle polarization , as well as to the left
right asymmetry in scattering of polarized particles ( both charged and neutral ) .
the density of nuclei near the axis is much higher than the density of an amorphous target .
this makes the induced polarization ( left right asymmetry ) for particles scattered by the crystal axes noticeably higher , as compared with amorphous matter , due to interference of the coulomb and nuclear interactions @xcite . in this context , intensively developing techniques for collimating and handling the beams of high - energy particles , using the volume reflection of charged particles from bent crystal planes deserve attention @xcite-@xcite . in particular , according to @xcite , multiple volume reflection from different planes of one bent crystal becomes possible when particles move at a small angle with respect to the crystal axis .
such multiple volume reflection enables a severalfold increase in the particle deflection angle inside one crystal .
this effect is experimentally revealed now , and it is noteworthy that it leads to an increasing probability of particle - nucleus interaction .
channeling of particles in either straight or bent crystals with polarized nuclei could be used to polarize high - energy particles or analyze their polarization .
the beam of nonpolarized particles extracted from the storage ring could be significantly polarized by applying the additional polarized bent ( straight ) crystal .
a few more words should be said about the possibility of quadrupole moment measurements for @xmath263-hyperons moving in a crystal @xcite .
theoretical estimates of the quadrupole moment of short - lived particles , for instance of an @xmath263-hyperon @xcite , based on the nonrelativistic quark model depend on the model parameters and can differ significantly .
experimental measurement of the quantity under consideration using the convenient method of passing through an inhomogeneous electromagnetic field is at present practically unrealizable because of the smallness of quadrupole interaction at attainable strengths of the electromagnetic field . despite the seeming simplicity ,
the use of the effect of spin oscillation of @xmath263-hyperons channeled in crystals is a challenging task , too .
this is , above all , due to the sharp increase of multiple scattering by nuclei in the case of negatively - charged particles ( @xmath263-hyperons ) .
however , it should be emphasized that the difficulties can be overcome to a great extent using the effect of spin oscillation occurring in a nonchanneled state of motion ( not too far above the critical angle ) relative to the chosen family of crystallographic planes for particles with spin @xmath265 .
the equations describing the spin dynamics of a particle possessing magnetic and quadrupole moments can be derived using the results obtained in @xcite . in consequence ,
the relativistic equation of spin motion in this case has the following form : @xmath266 where @xmath267 is a component of the vector @xmath268v^{2}\vec{\omega}_{0}$ ] and @xmath269 .
let us consider ( [ hyperon_1.8 ] ) under the condition that the particle moves in a nonchanneled state at an angle @xmath270 with respect to the chosen family of crystallographic planes , where @xmath271 is the lindhard angle .
suppose that the coordinate axis @xmath36 is perpendicular to the given crystallographic planes .
taking into account that @xmath272 , @xmath273 , we obtain @xmath274 @xmath275 @xmath276 equations ( [ hyperon_1.8])([hyperon_1.9c ] ) enable one to analyze whether the spin oscillation of a particle moving in a nonchanneled state can be used for measuring the electric quadrupole moment of the @xmath263-hyperon . in view of ( [ hyperon_1.9a])([hyperon_1.9c ] ) , when a particle with a quadrupole moment moves in a crystal , not only spin oscillation and rotation appear , but also the transitions between tensor @xmath277 and vector @xmath278 polarizations of the particle .
as follows from ( [ hyperon_1.8])-([hyperon_1.9c ] ) , under the initial conditions @xmath279 and @xmath280 , there is a maximal change of the @xmath263-hyperon spin projection @xmath281 due to the quadrupole interaction . according to the estimates made in @xcite , in such experiments for the @xmath263-hyperon beam with lorentz factor @xmath282 , intensity @xmath283 particle / s , and the beam divergence angle @xmath284mrad ,
it is possible to measure the quadrupole moment @xmath285 of the @xmath263-hyperon on the level @xmath286@xmath62 in a tungsten crystal of length @xmath287 cm .
it should be noted that the measurement scheme suggested here for @xmath263-hyperon quadrupole moment measurements requires neither a high quality crystal nor a monochromatic hyperon beam ! for the realization of this measuring procedure it is quite sufficient to have a crystal with a mosaic spread @xmath288mrad or a set of crystals arranged with an exactness of not more than @xmath289mrad relative to the chosen family of crystallographic planes .
this condition justifies the possibility of creating large crystal systems necessary for carrying out this experiment .
the analysis given in this paper shows that the phenomena of spin rotation and depolarization of high - energy particles in crystals at energies that will be available at lhc and fcc can be used to measure the anomalous magnetic moments of short - lived particles at such energies .
it has also been demonstrated that for negatively - charged particles ( e.g. , beauty baryons ) , the phenomenon of spin depolarization in crystals is a promising tool allowing the anomalous magnetic moment measurements .
moreover it has been noted that the spin depolarization effect occurs for neutral particles incident at small angles to crystal axes ( planes ) , and this opens the potential for magnetic moment measurements of such short - lived particles .
channeling of particles in either straight or bent crystals with polarized nuclei could be used for polarization or polarization analysis of high - energy particles , including neutral particles .
depolarization and spin rotation of that fraction of neutral particles which moves in the region of high concentration of nuclei can be defined by particle reactions in the crystal : if at a certain point the particle is scattered at a large angle , this means that it has undergone a collision , and hence it was in the vicinity of the nuclei . at high energies , particles move along a straight line .
particles moving between the planes do not undergo collisions , while those moving along the axis or a plane do .
baublis v. v. _ et al . _
first observation of spin precession of polarized @xmath290 hyperons channeled in bent crystals , lnpi research report ( 19901991 ) e761 collaboration ( st . pertersburg ) ( 1992 ) 2426 .
baryshevsky v. g. and and gurinovich a. a. , spin rotation and oscillations of high energy particles in a crystal and possibility to measure the quadrupole moments and tensor polarizabilities of elementary particles and nuclei , _ nuclear instrum .
methods b _ * 252 * , 1 , ( 2006 ) 136141 . baryshevsky v. g. and shechtman a. g. , spin oscillation and the possibility of quadrupole moment measurement for @xmath292hyperons moving in a crystal , _ nucl . instr .
methods b _ * 83 * , 1 - 2 ( 1993 ) 250254 . | we study the phenomena of spin rotation and depolarization of high - energy particles in crystals in the range of high energies that will be available at hadron collider ( lhc ) and future circular collider ( fcc ) .
it is shown that these phenomena can be used to measure the anomalous magnetic moments of short - lived particles in this range of energies .
we also demonstrate that the phenomenon of particle spin depolarization in crystals provides a unique possibility of measuring the anomalous magnetic moment of negatively - charged particles ( e.g. , beauty baryons ) , for which the channeling effect is hampered due to far more rapid dechanneling as compared to that for positively - charged particles .
channeling of particles in either straight or bent crystals with polarized nuclei could be used for polarization and the analysis thereof of high - energy particles . |
the dwarf - irregular galaxy ic 10 in cassiopeia presents one of the most active starbursts in the local group and a unique environment to study the immediate relics of the most massive stars : black holes and neutron stars in high mass x - ray binaries ( hmxbs ) .
motivated by the discovery of the black hole hmxb ic 10 x-1 @xcite and the very large amplitude transient ic 10 x-2 @xcite we conducted a 7-shot monitoring campaign with _
chandra _ during 2009 - 10 . combined with deep observations taken in 2003 and 2006 ,
we were able to span timescales ranging up to 7 years reaching a limiting sensitivity of @xmath2erg s@xmath3 .
this cadence and sensitivity , coupled with the angular resolution of _ chandra _ , enable light - curves to be constructed for individual point sources .
ic 10 hosts a young ( @xmath56 @xmath6 yr ) stellar population @xcite accompanied by the highest known space - density of wolf rayet ( wr ) stars @xcite , which are the evolved helium cores of stars with initial stellar masses in excess of 20 @xmath7 .
the star formation rate ( sfr ) is reported to be as high as 0.5 @xmath7@xmath8 @xcite , which when normalized by the mass of the galaxy ( 2 @xmath9 @xmath7 ; @xcite ) yields 2.5 @xmath10 @xmath7@xmath8 @xmath7@xmath3 which is among the highest specific sfrs in the local universe exceeding the values for the milky way ( mw ) , small magellanic cloud ( smc ) , and large magellanic cloud ( lmc ) .
depending on the indicators and methods used , there is an order of magnitude dispersion among sfr values for ic 10 @xcite .
the chaotic structure of ic 10 is probably part of the picture and indeed much of the galaxy resembles an ob association , but on a kiloparsec scale @xcite .
ic 10 is replete with a complex structure of @xmath11-bright and radio - bright bubbles , filaments , and hii regions , all signaling its turbulent evolution under the influence of the kinetic energy being injected by winds and supernovae .
the similarities between ic10 and the smc are striking ; both are gas - rich dwarf irregulars and both are thought to have undergone tidal disruption in the recent past @xcite leading to triggered star formation activity .
however , two critical differences exist between these galaxies that make ic 10 an exciting new laboratory for stellar astrophysics .
the first and most important contrast is the time since star formation began .
this is @xmath56 @xmath6 yr for ic 10 compared to 40 - 200 @xmath6 yr for the various distinct age - segregated populations identified in the smc @xcite , where the hmxbs are found in the 40 - 70 myr sub - population @xcite .
the second contrast is in metallicity which for ic 10 is @xmath12 @xcite , intermediate between the smc and the mw .
complete censuses of x - ray binary populations in local group galaxies are a powerful approach to discover the underlying physics of stellar evolution , for example the effect of metallicity .
the magellanic clouds have historically served this purpose , but gaining proper control of secular differences between galaxies requires new independent laboratories such as ic 10 to be added .
important large samples have been constructed by mining the _ chandra _ archive , ( for example @xcite ) and by concerted multi - year efforts to intensively observe nearby galaxies ( e.g. , @xcite ) .
an x - ray study of point sources in ic 10 was conducted by @xcite using single epoch data from _ xmm - newton _ and _
chandra_. among their important findings was a population of point sources significantly above the expected background density and spatially concentrated within the optical outline of the galaxy .
the combined x - ray spectrum of these point - sources was consistent with an absorbed power - law with parameters characteristic of hmxbs .
key physics goals to be addressed by ic 10 s x - ray binary ( xrb ) sample include : constructing the x - ray luminosity function ( xlf ) in order to compare it with the sfr history ; discovering what type of xrbs are formed at the earliest times ; and breaking the degeneracy between compact - object mass and accretion - mechanism that currently bedevils attempts to understand the nature of both ultraluminous x - ray ( ulx ) sources and the xlf - sfr relation .
the relative proximity of ic10 is crucial because the individual xrbs can be studied in detail once they are identified . in the larger samples of galaxies ,
the questions are addressed on a purely statistical basis . identifying optical counterparts to x - ray sources at the 660 kpc distance of ic 10
is a daunting prospect since it lies at low galactic latitude and it is viewed by us through the outer portion of the galactic plane .
since most xrbs are highly variable , a monitoring campaign offers an alternative path to discovering the nature of these individual sources .
this paper begins with a uniform reduction / analysis of the _ chandra _ monitoring and archival observations and visualizations of the ccd data ( [ sect : obs ] ) followed by the construction of the source catalog ( [ sect : catalog ] ) and the generation of x - ray light curves ( [ sect : lc ] ) .
we outline our spectral analysis using quantiles and ccd pulse - height spectra in [ sect : quantile ] , we briefly describe how optical counterparts were identified ( [ sect : optical ] ) , and then we proceed to [ sect : discussion ] , where we look at each variable source in turn in order to summarize the available information and to discuss its likely nature . finally , in [ sect : conclusions ] , we present an overview of the results and we lay out the direction of future progress .
a monitoring series of 7@xmath1315 ksec _ chandra _ observations , spaced at roughly 6-week intervals was obtained during 2009 - 2010 .
we used the advanced ccd imaging spectrometer ( acis ) , positioning the galaxy ic 10 on the back - illuminated chip s3 .
the observing cadence was designed to match the native variability timescales of hmxbs which tend to be transients with typical outbursts lasting for 1 - 3 weeks repeating on intervals of months .
the individual exposure times were calculated to reach a limiting luminosity of 10@xmath14 erg s@xmath3 ( assuming 5 net acis - s counts at the high resolution mirror assembly ( hrma ) aim - point for a detection threshold ) .
this sensitivity would reach the typical luminosity of hmxbs ( both persistent supergiant hmxbs and be - hmxbs in outburst ) .
a pair of very deep acis - s3 observations ( 2@xmath1345 ksec ) made in november 2006 provide a reference dataset for improved source positions and spectral information .
the original @xcite _ chandra _ ( acis - s3 , 1/2 subarray ) observation of 30 ksec made in 2003 was included in our analysis to provide additional time baseline for the variability study and to make use of their published source identifications .
the complete listing of 10 observation identifiers ( obsids ) comprising the dataset is summarized in table [ tab : dataset ] .
all of the data were processed in a uniform way following the standard _ ciao _
threads , and the prescription of @xcite for consistent selection of energy bands
b ( 0.3 - 8 kev ) , s ( 0.3 - 1.5 kev ) , and h ( 2.5 - 8 kev)and extraction regions .
llllllll 52710.7 & 12 mar 2003 & 03953(a ) & 28.9 & 00:20:25 & 59:16:55 & 339.27 & 31 + 54041.8 & 2 nov 2006 & 07082 & 40.1 & 00:20:04 & 59:16:45 & 223.70 & 48 + 54044.2 & 5 nov 2006 & 08458 & 40.5 & 00:20:04 & 59:16:45 & 223.70 & 41 + 55140.7 & 5 nov 2009 & 11080 & 14.6 & 00:20:17 & 59:17:56 & 226.53 & 19 + 55190.2 & 25 dec 2009 & 11081 & 8.1 & 00:20:19 & 59:18:02 & 286.15 & 24 + 55238.5 & 11 feb 2010 & 11082 & 14.7 & 00:20:23 & 59:17:10 & 320.56 & 24 + 55290.6 & 4 apr 2010 & 11083 & 14.7 & 00:20:34 & 59:19:01 & 10.32 & 25 + 55337.8 & 21 may 2010 & 11084 & 14.2 & 00:20:25 & 59:20:16 & 67.89 & 27 + 55397.5 & 20 jul 2010 & 11085 & 14.5 & 00:20:11 & 59:19:13 & 121.25 & 22 + 55444.6 & 5 sep 2010 & 11086 & 14.7 & 00:20:15 & 59:18:11 & 157.71 & 27 + ( b ) & 2 - 5 nov 2009 & 57082 & 80.6 & - & - & - & 63 + images for visual inspection were generated using both gaussian - smoothed and adaptively - smoothed b ( 0.3 - 8 kev ) , s ( 0.3 - 1.5 kev ) , and h ( 2.5 - 8 kev ) band data .
smoothing is necessary to create an image from sparse data .
the adaptive smoothing maps created with _
ciao smooth _ for the b - band images were applied to make the s & h band images , and all 3 bands used the exposure maps generated for 1.5 kev photons ( roughly the peak of acis sensitivity ) . to illustrate our search for transient sources we present all 10 b - band images in figure [ fig : cxo_images ] which shows a number of prominent transient events and
also illustrates the effects of shifting spacecraft roll - angle and aim - point .
the individual variable sources can be identified with the aid of figure [ fig : cxo_image ] which is a representative color image constructed using the visualization and analysis program _
saoimage ds9 _ from the deep merged dataset 57082 ( table [ tab : dataset ] ) .
for each _ chandra _ observation we ran _ ciao _ _ wavdetect _ on each of the three energy bands ( s , b , h ) .
_ wavdetect _ was run with the default significance threshold of @xmath15 at wavelet scales of 1 , 2 , 4 , 8 , and 16 pixels .
each detection is given its own source @xmath16 , and the detections in each band are cross - matched by position ( 95% uncertainty radius computed by the formula of @xcite ) to produce a list of unique sources within the observation , linked by an identifier called @xmath17 .
the list derived from the 90 ksec stacked @xmath18 57082 was bore - sighted against the optical catalog of @xcite .
source lists from each @xmath18 were then corrected to the same reference frame , using the coordinates of ic 10 x-1 ( by far the brightest source ) .
we then compiled the final point - source catalog from the entire set of observations by using the tool _
classifytable _ in the _ starbase _ relational database package .
_ classifytable _ is used to group together sources that lie within a certain distance ( 5 ) of each other , these matches are then filtered using the 95% uncertainty radii .
all detections belonging to a unique source are linked by the identifier _ srcnumber_. coordinates and error radius are inherited from the @xmath16 with the best " detection ( most counts , smallest error radius ) . the number of unique x - ray point sources detected in each observation is listed in the final column of table [ tab : dataset ] .
next , the unique source list for each observation was bore - sighted to the unique source list obtained from the deep merged dataset of obsids 07082 & 08485 ( which had identical pointing ) .
this reference data set had itself already been bore - sighted to the optical catalog of @xcite .
the bore - sighting procedure between x - ray observations used only the brightest source ( ic 10 x-1 ) due to its very small statistical positional uncertainty arising from its large count rate .
once bore - sighting was complete , we constructed a master catalog using a 5 arcsec matching radius around each source ( @xmath17 ) .
finally we applied a filter to eliminate false matches : requiring the matching @xmath17 coordinates to lie within the quadrature sum of their 95% error radii and an additional 0.75 " arcsec accounting for statistical and systematic errors .
the resulting master catalog contains 110 unique point sources , many of which have multiple detections .
all detections of the same source share the same _ srcnumber _ identifier and a _ chandra _ source name based on the usual convention ( cxou j hhmmss@xmath19ddmmss ) taking the coordinates from the @xmath16 with the smallest error circle .
the _ chandra _ observations all have slightly different fields of view as a consequence of the operational requirements of the spacecraft . a certain pitch and roll angle
are required to maintain the proper orientation with respect to the sun .
we worked with the mission planners to adjust the exact pointing for each observation once the pitch / roll combination was known .
this enabled us to keep most of the same point sources within the acis - s3 ccd during each observation .
the master catalog was processed with a _ unix _ script that called _ ciao _ to query the exposure map for the coordinates of every source in each observation to determine whether the source was actually observed with acis - s3 or not .
if the source was observed _ and not detected _ by _ wavdetect _ , then a new row was added to the master catalog with the modified julian date ( mjd ) of the observation , the value @xmath20 = null " and an estimated upper - limit for the broad - band count rate .
this upper limit was computed as the equivalent count - rate that would generate 5 net counts at the hrma aim - point according to pimms ( portable multi - mission simulatorbin / tools / w3pimms / w3pimms.pl ] ) for an absorbed power - law spectrum ( @xmath21 @xmath22 ) divided by the normalized exposure - map value at the source location .
the simulations of @xcite indicate this is a @xmath2375% detection probability for the off - axis distances encountered within the s3 chip , at low background rates .
if the source was found to lie outside the s3 chip , then no modifications were made to the table .
the master source catalog is maintained as a single table , combining all detections and upper limits , from which light curves are extracted as described in the next section .
positional matching to the catalog of @xcite was performed using our and their 95% error radii in quadrature ; the resulting ids are provided in table [ tab : variables ] . to facilitate observations at other wavelengths , and cross - correlation with other databases we provide our complete point - source catalog as supplemental data in the online journal .
light curves for all of the 110 unique sources were constructed by processing the master catalog with a _ unix _ script which computed the following parameters for each group of rows bearing the same @xmath24 identifier : number of occasions observed ( @xmath25 ) , number of detections in the broad - band ( @xmath26 ) , minimum count rate ( @xmath27 ) , maximum count rate ( @xmath28 ) , average count rate ( @xmath29 ) ; root - mean - square @xmath30 ; variability range ( @xmath31 = @xmath28 - @xmath27 ) ; variability ratio ( @xmath32 ) ; relative variability ( @xmath33 ) ; and total number of net broad - band counts ( @xmath34 ) obtained by summing the values from each detection . in computing the above parameters we included all b - band detections and b - band upper limits , except that @xmath35 ( and its associated error bar ) are required for a positive detection .
the complete set of parameters characterizing each source are provided in the online journal as supplemental data accompanying table [ tab : variables ] , which shows only the variable objects , whose selection is described below .
light curves were plotted for all x - ray sources in the master catalog ; selected examples are shown in figures [ fig : lcs_gt10 ] , [ fig : lcs_gt5 ] , and [ fig : lcs_lt5 ] .
the plots display detections in all 3 energy bands and upper limits for observations with no b - band detection .
all 110 light curves were visually examined alongside the _ acis _ images to verify the detections and to note any abnormalities such as sources falling very close to the edge of a ccd or having a close neighbor . in this way
, we flagged 8 sources as probably false " ( objects detected only once at the extreme edge of the ccd ) and 4 sources as being duplicates ( multiple detections of a real source being falsely identified as separate sources when detected in subsequent observations ) .
these eventualities are expected , and the number of cases was within the expected range .
_ wavdetect _ expects to generate one false source per ccd ( i.e. @xmath23 10 in total ) and faint off - axis sources are subject to larger positional errors .
we note that most of the sources were detected in at least two independent observations .
specifically , 43 sources were detected more than once in the b - band , 36 only once in b - band ( some of which were also detected in a separate observation in s or h only ) , and 25 were only ever detected in the s or h band .
the probability of a source with multiple detections being false is extremely small given the low surface density of sources .
the light - curve parameters described above ( and given in table [ tab : variables ] ) were used to identify variable and transient sources by sorting on the @xmath36 and @xmath37 parameters . assuming the usual statistical definitions , a light curve with @xmath38 is variable at the 99% confidence level .
there are 21 such objects , listed in table [ tab : variables ] ranked in descending order by @xmath38 .
our criterion for detecting a variable or transient source is at least one b - band detection with count rate at least @xmath39 above the lowest upper limit ; thus , only real sources are included in table [ tab : variables ] .
complete data for each variable source are provided in tabular form as `` data - behind - the figure '' accompanying this paper in the online journal .
lllllllllllll 46 & 11c & 10 & 2 & 1.61 & 0.1 & 11.75 & 11.65 & 117.5 & 18.2 & 370 & 3.35 & t , x + 90 & - & 9 & 1 & 0.45 & 0.1 & 1.83 & 1.73 & 18.3 & 4.7 & 26 & 1.64 & t , s + 28 & 24c & 8 & 3 & 0.73 & 0.18 & 2.86 & 2.67 & 15.6 & 9.2 & 130 & 1.13 & t , s + 50 & 17c , 46x & 8 & 6 & 0.58 & 0.1 & 1.24 & 1.14 & 12.4 & 5.7 & 88 & 2.07 & t , x ? + 13 & - & 7 & 2 & 0.39 & 0.1 & 0.96 & 0.86 & 9.6 & 3.1 & 19 & 1.69 & tc , s ? + 30 & 19c & 10 & 4 & 0.36 & 0.1 & 0.92 & 0.82 & 9.2 & 3.1 & 33 & 1.97 & tc , x ?
+ 41 & - & 10 & 5 & 0.63 & 0.2 & 1.68 & 1.48 & 8.4 & 7.4 & 154 & 2.53 & tc , x + 43 & 24x & 8 & 7 & 1.11 & 0.3 & 2.45 & 2.15 & 8.2 & 5.1 & 170 & 2.06 & v , x + 20 & 18c , 39x & 10 & 10 & 105.8 & 23.7 & 179.87 & 156.17 & 7.6 & 55.6 & 22632 & 1.75 & v , x + 53 & - & 8 & 1 & 0.47 & 0.1 & 0.66 & 0.56 & 6.6 & 3.7 & 27 & 1.38 & tc , u + 5 & 14c , 35x & 8 & 8 & 1.55 & 0.54 & 3.17 & 2.63 & 5.9 & 6.7 & 292 & 0.89 & v , u + 29 & 4c & 10 & 1 & 0.31 & 0.1 & 0.51 & 0.41 & 5.1 & 3.6 & 22 & 3.60 & tc , x + 23 & 17x & 9 & 7 & 0.98 & 0.33 & 1.62 & 1.29 & 4.9 & 4.9 & 203 & 1.86 & v , u + 52 & - & 9 & 2 & 0.53 & 0.3 & 1.21 & 0.91 & 4.0 & 4.7 & 64 & 1.24 & v , s + 14 & 3c , 25x & 8 & 7 & 0.76 & 0.39 & 1.68 & 1.29 & 4.3 & 3.3 & 117 & 3.64 & v , x + 42 & 12c , 33x & 10 & 9 & 0.73 & 0.4 & 1.44 & 1.04 & 3.6 & 3.3 & 144 & 2.70 & v , x / a ? + 33 & 16x & 9 & 7 & 0.87 & 0.4 & 1.38 & 0.98 & 3.4 & 3.0 & 132 & 2.55 & v , x / a ? + 1 & 3c , 36x & 5 & 4 & 1.24 & 0.68 & 1.89 & 1.21 & 2.8 & 3.0 & 70 & 1.76 & v , x + 12 & 15c , 36x & 9 & 9 & 4.96 & 2.84 & 7.93 & 5.1 & 2.8 & 8.1 & 1137 & 1.69 & v , u + 18 & 8c & 10 & 10 & 1.99 & 1.18 & 2.77 & 1.59 & 2.3 & 3.6 & 435 & 1.12 & v , u + 8 & 7c , 28x & 8 & 8 & 4.22 & 3.1 & 5.8 & 2.7 & 1.9 & 4.5 & 863 & 1.83 & v , x + given the large distance to ic 10 we are able to extract spectra for relatively few sources .
consequently the median energy for the combined events is reported in table [ tab : variables ] as an indication of spectral hardness .
we plotted a hardness intensity diagram for all 110 point sources in figure [ fig : hardint ] , using cumulative events as a proxy for intensity , and the median event energy for the hardness axis .
this parameter space serves to illustrate some key properties of the objects in our catalog .
figure [ fig : hardint ] shows that large amplitude variability is seen at all intensity levels , that most bright sources are moderately variable ( @xmath40 ) , and that below 100 counts our selection method ( section [ sect : lc ] ) is sensitive only to large amplitude variables .
thus we are not picking up random fluctuations in faint sources as variables .
figure [ fig : hardint ] also shows that the distribution of median event energy is bimodal , with a hard / soft split at 1.4 kev .
this was revealed by a kernel density estimator ( _ density _ in the _ r _ package , with bandwidth = 0.1 kev ) which we ran both with and without weighting by the errors .
weighting increased the strength of the bimodality , as seen by comparing the solid and dashed lines in the right - hand panel of figure 6 .
we performed a bootstrap resampling ( 10@xmath41 trials ) using the error - weighted values to generate 95% confidence intervals for the density plot .
the _ kmeans _ clustering test identifies a similar hard - soft split , which can be attributed to different extinction values for ic 10 vs foreground galactic objects .
energy quantiles @xcite were computed from the event lists for each unique source in the master catalog .
quantiles were computed for each observation and also for the combined event list , which provides the maximum number of counts and hence delivers better s / n values .
these values and the accumulated number of events in the source extraction regions is listed in the supplementary table in the online journal . in our analysis
we use two of the quantiles defined by @xcite : the slope quantile @xmath42 which is related to the steepness of the spectrum , and the curvature quantile @xmath43 which is constructed from the ratio of the quartiles @xmath44 .
full details are available in @xcite and in the source code .. the @xmath42 , @xmath43 parameter space shown in figure [ fig : quantiles ] was designed to explore in a single analysis , the spectral shape of low count sources that differ widely in hardness .
panel a shows spectral model grids generated for the back - illuminated _ acis - s3 _ detector for two classes of sources : power law ( photon index @xmath45 ) and thermal bremsstrahlung ( @xmath46 ) , both of which with absorption varying between @xmath47@xmath22 .
these models are intended to represent accretion - powered and stellar coronal sources , respectively .
the quantile values for the variable sources selected in table [ tab : variables ] are plotted in panel b , with labels to facilitate discussion .
error bars are in panel c , and those sources with optical counterparts ( table [ tab : optical ] ) are highlighted in panel d. pulse - height spectra were extracted for all sources having greater than 200 net counts in the master catalog , yielding 7 objects ( there are 13 sources with @xmath1 100 net counts ) . in this article , we include only those appearing in table [ tab : variables ] after excluding ic 10 x-1 and x-2 ( which have been reported elsewhere ) , which leaves 5 objects for the analysis in this paper . for every such object , we used the _ ciao _ script _ specextract _ to generate source and background spectra and response files ( rmf , arf ) for each observation ; and also for the merged event list containing all observations in which the source was detected . the resulting spectra are plotted in figure [ fig : spectra ] .
the spectra were examined in xspec and were fitted by the following models ( combined with the _ phabs _ absorption model ) : power - law ( pl ) , raymond - smith ( rs ) , mekal ( m ) , black - body ( bb ) , and thermal bremmstrahlung ( tb ) .
these models were selected as appropriate for x - ray spectra of accretion , magnetic flare , and thermal origin . in cases where the absorption was unconstrained by the model we froze it at 5@xmath48 @xmath22 for the column density toward ic 10 .
the resulting model parameters and goodness of fit statistics are presented in table [ tab : xrayspec ] . given the low s / n ratios , only single component fits were considered .
we caution that both quantiles ( in common with all flux ratio - based color systems ) and single - component spectral fits can be non - unique . for example
spuriously low @xmath49 values can occur when soft components remain unaccounted for @xcite .
this is a motivation to obtain deeper datasets where such components can be resolved .
cccccc source 5 + pl & 0.842 @xmath19 0.3 & - & 7.5 @xmath19 2.2 & 2.0 @xmath19 0.9 & 1.167/74 + rs & 0.442 @xmath19 0.1 & 0.28 @xmath19 0.06 & - & 2.1 @xmath19 1.2 & 0.596/74 + m & 0.000 @xmath19 0.1 & 0.55 @xmath19 0.08 & - & 0.21 @xmath19 0.09 & 0.631/74 + bb & 0.006 @xmath19 0.1 & 2.8 @xmath19 0.02 & - & 7.6 @xmath19 0.8 & 1.829/74 + tb & 0.78 @xmath19 0.5 & 0.13 @xmath19 0.04 & - & 0.02 @xmath19 0.001 & 0.934/74 + source 8 + pl & 0.565 @xmath19 0.09 & - & 1.6 @xmath19 0.2 & 1.3 @xmath19 0.2 & 1.020/53 + rs & 0.483 @xmath19 0.07 & 10.2 @xmath19 4.2 & - & 4.2 @xmath19 0.3 & 1.011/53 + m & 0.482 @xmath19 0.07 & 10.4 @xmath19 4.4 & - & 4.2 @xmath19 0.2 & 1.017/53 + bb & 0.051 @xmath19 0.06 & 0.92 @xmath19 0.005 & - & 0.061 @xmath19 0.004 & 1.049/53 + tb & 0.49 @xmath19 0.07 & 9.4 @xmath19 3.8 & - & 1.5 @xmath19 0.1 & 1.001/53 + source 12 + pl & 0.73 @xmath19 0.08 & - & 2.2 @xmath19 0.1 & 2.9 @xmath19 0.4 & 0.835/64 + rs & 0.57 @xmath19 0.06 & 4.1 @xmath19 0.5 & - & 5.8 @xmath19 0.4 & 0.866/64 + m & 0.57 @xmath19 0.06 & 4.1 @xmath19 0.6 & - & 5.7 @xmath19 0.4 & 0.883/64 + bb & 0.19 @xmath19 0.05 & 0.71 @xmath19 0.03 & - & 0.065 @xmath19 0.3 & 1.180/64 + tb & 0.58 @xmath19 0.05 & 3.6 @xmath19 0.7 & - & 2.5 @xmath19 0.3 & 0.849/64 + source 18 + pl & 0.831 @xmath19 0.2 & - & 6.2 @xmath19 2.1 & 0.41 @xmath19 0.2 & 0.993/46 + rs & 0.072 @xmath19 0.03 & 6.2 @xmath19 1.8 & - & 0.16 @xmath19 0.09 & 1.013/46 + m & 0.623 @xmath19 0.2 & 0.56 @xmath19 0.2 & - & 0.19 @xmath19 0.1 & 0.990/46 + bb & 0.000 @xmath19 0.04 & 0.32 @xmath19 0.05 & - & 0.001 @xmath19 0.0004 & 1.004/46 + tb & 0.453 @xmath19 0.1 & 0.39 @xmath19 0.1 & - & 1.17 @xmath19 0.9 & 0.9934/46 + source 23 + pl & 2.45 @xmath19 0.9 & - & 5.8 @xmath19 0.6 & 5.4 @xmath19 1.2 & 0.803/71 + rs & 3.31 @xmath19 0.5 & 0.21 @xmath19 0.06 & - & 0.002 @xmath19 0.0005 & 0.805/71 + m & 0.834 @xmath19 0.3 & 1.06 @xmath19 0.9 & - & 0.95 @xmath19 0.6 & 0.834/71 + bb & 0.001
@xmath19 0.002 & 3.31 @xmath19 0.9 & - & 0.15 @xmath19 0.08 & 0.827/71 + tb & 1.02 @xmath19 0.8 & 1.7 @xmath19 1.2 & - & 0.42 @xmath19 0.2 & 0.810/71 + a list of positional counterparts from @xcite was generated by the procedure described in @xcite .
a limiting visual magnitude of @xmath50 was used for sources without counterparts , this being the faintest magnitude of any object in the catalog .
the distance independent x - ray to optical flux ratios were found using @xmath51 where the observed flux @xmath52 was computed from the net count rate assuming an absorbed power - law with a photon index of 1.5 and @xmath53 @xmath22 .
the values in table [ tab : optical ] are the maximum measured values , remembering that @xmath52 is variable . in the following analysis , sources with optical counterparts believed to lie in ic 10 have their apparent magnitudes corrected for the foreground extinction assuming a value of @xmath54= 2.85 and distance modulus @xmath55 .
narrow - band @xmath11 and @xmath56 images from the gemini observatory were examined in order to look for sources of balmer - line emission as described in @xcite .
preliminary spectroscopic followup was reported by @xcite , although spectral type and radial velocity ( rv ) are only available for 3 of the variable sources .
four sources have variability @xmath57 and @xmath58 , these are src 46 ( ic 10 x-2 ) , src 90 , src 28 , & src 50 ( listed in order of decreasing amplitude ) .
the light curves for these sources appear in figure [ fig : lcs_gt10 ] .
there are a further 2 sources ( src 53 and src 29 ) that we include as transients as each was detected in a single observation , with a broad - band count rate exceeding the lowest upper limit by three @xmath0 ( i.e. @xmath59 and @xmath60 , but not reaching @xmath57 ) listed in table [ tab : variables ] .
both were in the acis - s3 field of view 8 - 10 times , so that the null detections and corresponding upper limits place meaningful limits on the duty - cycle .
lightcurves for these two objects are include in figure [ fig : lcs_gt5 ]
. only one transient , src 46 , which is ic 10 x-2 , @xcite provided enough counts to produce pulse - height spectra and appears to be an x - ray binary with a supergiant be companion and similarities to supergiant fast x - ray transients ( sfxts ) . for reference
we include its lightcurve in figure [ fig : lcs_gt10 ] and it appears on the quantile diagram ( figure [ fig : quantiles ] ) at ( -0.18,1.17 ) which is consistent with a very hard power - law spectrum .
the case of ic 10 x-2 highlights the need for optical spectra , since its photometric colors are unremarkable unless one knows the appropriate correction , and its nature as an extragalactic hmxb required the unambiguous doppler - shift to place it firmly in ic 10 .
src 28 was observed 8 times yielding 3 b - band ( and one s - band only ) detections with a variability ratio of 15.6 , a 9.3@xmath0 variation above the most constraining upper limit .
quantile values for the 4 observations were all consistent with the merged value .
lying at [ -0.9,1.1 ] in figure [ fig : quantiles ] the source is consistent with a soft power law ( @xmath61 ) or a @xmath23few kev thermal plasma . in both cases
the absorption is small ( @xmath62@xmath22 ) suggesting a foreground galactic location for the object ( subject to the caveats in section [ sect : quantile ] . )
there are two optical counterparts ( table [ tab : optical ] from @xcite ) and the source is not in the field of the gemini images , so no followup spectrum was obtained .
only one of these stars is consistent with the x - ray detection having the smallest error circle .
both stars have color indices suggestive of spectral type m. taking the most likely counterpart , whose b - v is already at the un - reddened value for m0v , we calculate distance modulus @xmath63 .
the distance is then @xmath23900 pc , yielding peak @xmath64erg s@xmath3 , in the range of isolated dme stars , which is supported by the distance independent ratio @xmath65=-0.53 being lower than seen for active binaries .
note that in the absence of a spectral type or rv value , ic 10 membership can not be entirely ruled out .
if located at 660 kpc , and de - reddened by @xmath66=2.85 the star would have an extreme absolute magnitude @xmath67 , which favors the galactic dme interpretation .
src 29 was detected in _
obsid 08458 _ with 22 net counts which placed it a factor of 5 ( or 3.6@xmath0 ) above its most sensitive non - detection .
the detection is in the b and h ( and not s ) bands .
the quantile value [ -0.13,1.2 ] lies in the hard power law region ( @xmath68 ) of figure [ fig : quantiles ] on the contour for @xmath69 @xmath22 .
there are 2 optical counterparts in m07 having near identical photometry ( @xmath70 21.7 , @xmath71 ) , gemini @xmath11 line and continuum images indicate that two stars are blended in the @xcite images , with one catalog position on top of the brighter star , and the other midway between the two .
the source lies in a region containing clumpy diffuse @xmath11 emission , but neither star is a strong @xmath11 point - source .
a gmos spectrum of the brighter star ( which is near the center of the x - ray error circle ) shows a strong continuum without obviously identifiable features other than telluric lines .
this rules out a foreground m - dwarf and suggests a reddened early - type star in ic 10 . applying the distance modulus and
extinction correction assumed throughout this paper gives @xmath72 , and @xmath73 .
the star is also blue in u - b and v - r color indices compared to the other stars in table [ tab : optical ] , and once de - reddened looks like a b spectral type .
this object is likely an hmxb in ic 10 , based on x - ray and optical properties .
src 50 was observed 8 times , yielding 6 b - band detections and two upper limits .
the most constraining upper limit comes from one of the pair of 45 ksec exposures made in nov 2006 providing the evidence for a large variability ratio ( @xmath74=12 , @xmath36=5.7 ) .
the quantile values varied with time , the spectrum being harder in _ obsids 11081,11082,11084_. the merged value plotted in figure [ fig : quantiles ] lies at [ -0.52,1.05 ] which is dominated by photons from its softer apparition in 2006 .
the model grid indicates a power law ( @xmath75 ) spectrum with typical ic 10 absorption ( @xmath23 5@xmath48@xmath22 ) .
there is an optical counterpart in @xcite but it lies outside the gemini field and no followup spectrum has been obtained .
if the object is in ic 10 , then its de - reddened photometry gives @xmath76 , @xmath77 .
src 53 was detected only in _
obsid 08458 _ with 27 net counts , at a count - rate a factor of 6.6 ( 3.7@xmath0 ) above the most constraining non - detection _
obsid 07082 _ which was of equal exposure - time ( 45 ksec ) and pointing , and occurred only a day before .
a point - source is clearly visible in the image of _ obsid 08458 _ and absent in _ obsid 07082_. the detection was made in the s and b bands only .
the very rapid variability and soft spectrum implied by the s - band detection are indicative of a magnetic stellar flare but not conclusive .
the quantile value [ -0.8,0.5 ] lies outside our model grids and has large errors so we can say little about its spectral shape .
there is an optical counterpart ( v=21.4 , table [ tab : optical ] ) but it lies outside the gemini field and no followup spectrum has yet been obtained .
src 90 was covered by the acis - s field of view 9 times , but was detected only once , in _
obsid 11082 _ in the s , b & h bands .
it is clearly visible as a point source in the adaptively smoothed image .
the observed count - rate was a factor of 18 ( 4.7@xmath0 ) above the most constraining upper limit .
the energy quantile ( figure [ fig : quantiles ] ) lies at [ -0.67,1.23 ] within the power law model grid at a location consistent with @xmath78 , @xmath79 @xmath22 .
there is an optical counterpart ( [ tab : optical ] ) which we identify as an @xmath11 emission line star in gemini narrow - band imaging .
we obtained a gemini gmos optical spectrum of this star @xcite , which shows a red continuum with molecular bands characteristic of spectral type m. prominent lines of @xmath11 , @xmath80 in emission enable us to measure the radial velocity rv=-165 km / s , rv=-171 km / s for the respective lines .
thus the star belongs to the milky way since the mean rv of ic10 is -340 km / s .
photometry from @xcite ( v=21.348 , b - v=1.8 ) indicates the color - index is consistent with early m. for example spectral type m0ve ( @xmath81 would require only e(b - v)@xmath230.3 , @xmath82 .
the distance modulus is then @xmath83 , for a distance of @xmath232 kpc , implying @xmath84erg s@xmath3 .
the distance - independent flux ratio during the flare reaches @xmath65= + 0.3 .
stellar flares in dme stars reach @xmath85 , with the by dra binary type being particularly luminous @xcite .
the luminosity , quantile , and duty - cycle point to a hard and energetic event on a magnetically active m - dwarf .
such interlopers are frequent bi - catch of x - ray surveys , and we have seen one in the foreground of the smc @xcite . many of the brighter sources in our study show significant variability ( @xmath38 ) as expected given that astrophysical x - ray emission mechanisms are dynamic in nature . the light curves of these candidates are presented in figure [ fig : lcs_gt5 ] for those varying by a factor of 5 - 10 , and in figure [ fig : lcs_lt5 ] for those not exceeding a factor of 5 .
all of these objects have sufficient counts in the merged event - list to provide energy quantiles with reasonable s / n . in figure
[ fig : quantiles ] we plot the quantiles for all unique objects in the master catalog having at least 10 counts in their merged event list , with the candidate variable sources picked out in larger points . a group of 10 candidate variables clusters along the @xmath49 contour corresponding to @xmath86@xmath22within the power law model grid region indices in the range @xmath87 . based on this evidence
they are consistent with being hmxbs in ic10 . in the following section we present individual analyses for the remaining variability candidates listed in table [ tab : variables ] , beginning with the five candidate variables to exceed 200 net counts . for these objects we performed spectral fitting in xspec testing
pl , bb , tb , nei models combined with the photo - absorption model _ phabs_. the best fit for each of these sources is displayed in table [ tab : xrayspec ] .
source 1 was observed in half of the acis - s observations and is detected in 4 of the 5 .
the are enough counts for the quantile diagram to place it firmly in the xrb locus . the optical counterpart
@xmath88 if adjusted for distance ( @xmath55 ) and reddening ( @xmath89 ) gives @xmath90 which implies an early b spectral type .
thus the object is an hmxb , subject to optical spectral confirmation of the precise subtype .
the light - curve for src 5 ( figure [ fig : lcs_gt5 ] ) shows a large amplitude variable with a 100% duty cycle , being detected in all observations for which it was in the fov .
it also shows that virtually all the counts are in the soft band .
the hardness - intensity plot ( figure [ fig : hardint ] ) shows src 5 is at once the softest variable , and one of the brightest objects in the survey , pointing to a foreground object .
quantile analysis suggests a thermal spectrum with median energy of just 0.5 kev that did not vary between detections .
the @xmath91 net - count spectrum was best fit by a black - body model ( fig .
[ fig : spectra ] , table [ tab : xrayspec ] ) .
a power - law model can provide an acceptable fit but the spectral index of 7.7 implies the cool thermal plasma is more physically motivated .
no optical counterpart is found in @xcite and the source lies outside the gemini field .
src 8 is a very low amplitude variable with 100% duty cycle in the chandra survey .
the summed 863 net - count spectrum was fit with a power - law model of @xmath92 , and typical ic 10 absorption ( fig .
[ fig : spectra ] , table [ tab : xrayspec ] ) .
there is an optical counterpart in @xcite ( lies outside the gemini field ) for which we can estimate @xmath93 , @xmath94 which would fit a supergiant or be - x interpretation .
src 12 is a persistent source with 100% duty cycle and a lightcurve that suggests very slow x - ray variability .
the source has the second highest number of total x - ray counts ( @xmath11000 ) in the survey , so spectral fitting was feasible .
the quantiles lie at the upper end of the cluster of points in figure [ fig : quantiles ] , and accordingly a powerlaw fit to the spectrum yields @xmath95 powerlaw with typical absorption for the ic 10 field ( figure [ fig : spectra ] , table [ tab : xrayspec ] ) . turning to physically motivated models
, acceptable fits are obtained for the thermal plasma models ( tb , rs , mekal ) , all giving temperatures consistent with @xmath96 , and @xmath97@xmath22 similar to the powerlaw fit .
a blackbody fit yields a significantly lower column density and temperature ( kt@xmath981 kev ) .
the lack of an optical counterpart in m07 provides the following constraints on distance - independent flux ratio ( @xmath65 @xmath99.9 ) and absolute magnitude ( @xmath100 ) if located in ic 10 .
the object could be an agn on the basis of the spectral index , large @xmath65 and slow variability .
a foreground dme star seems unlikely due to the lack of flares and persistently high @xmath65 .
an xrb can not be ruled out .
given the persistent nature of this source it will be possible to obtain much higher s / n x - ray spectra and timing in the future .
a transient x - ray source appearing to exhibit rapid variability , src 13 was detected ( at a very low flux ) in only one of the deep ( 45 ksec ) observations separated by 2 days in 2006 .
subsequently the source flared up in a single one of the 5 monitoring observations of 2009/10 which are spaced approx . 6 week apart .
these two independent positive detections confirm the reality of the object .
the limiting fluxes for those four nulls are at or above the flux detected in the deep 2006 exposure .
both positive detections were dominated by soft - band counts , with no detection in the hard band .
the quantile value is consistent with a plasma temperature of @xmath101 , or a @xmath102 powerlaw .
there is an optical counterpart within the 95% confidence radius .
the photometry ( table [ tab : optical ] : @xmath103 ) would be consistent with an un - reddened m0v star at a distance of 2 kpc , or a hotter spectral type at larger distance .
it is impossible to make the star closer since the observed color index imposes the limit of @xmath104 .
if located in ic 10 we obtain @xmath105=-6.4 .
an optical spectrum is needed to resolve the nature of this source .
src 14 has an interesting lightcurve that contains an outburst of duration @xmath232 months during 2010 .
quantile analysis reveals a persistently hard source , having the highest median photon energy ( e50 = 3.64 kev ) of any object in table [ tab : variables ] , and lying in the high - absorption hard power - law region of figure [ fig : quantiles ] .
the lack of an optical counterpart down to a limiting magnitude of @xmath10623.8 provides us with a distance independent flux ratio @xmath65@xmath107 + 1.24 for the brightest recorded x - ray flux .
if located in ic10 , the absolute magnitude would be @xmath108 if we assume the typical extinction correction .
a more luminous star would require more extinction to remain hidden .
spectral type b0v has @xmath109 so the data support a be - xrb .
if a dme star located in the milky way , then the spectral type would have to be late m ( for m0v the luminosity distance implied by @xmath50 is greater than 9 kpc ) and the x - ray quantiles are too persistently hard . in summary the long duration outburst , persistently hard spectrum , and high @xmath65 point to an xrb .
src 18 is a low amplitude ( less than a factor of 2 ) variable with 100% duty cycle in the chandra survey .
the total 435 net - count spectrum is soft , equivalent to powerlaw @xmath110 or cool thermal plasma for any reasonable @xmath49 ( fig .
[ fig : spectra ] , table [ tab : xrayspec ] ) .
there is no optical counterpart in m07 , providing a lower limit on @xmath65 @xmath1 + 1.12 .
the x - ray position lies at the center of a faint compact @xmath11 shell identified in narrow - band gemini images .
a gmos slit placed across the ring did not attain sufficient s / n to detect any lines .
src 23 is a variable source with a high duty cycle ( 7/10 ) and no optical counterpart in m07 which puts a lower limit on @xmath65 @xmath1 + 1.2 . with 203 net counts summed over the 6 observations where it was detected we obtained an acceptable x - ray spectral fit to the thermal bremsstrahlung spectral model with @xmath111=1.7 kev , @xmath112 @xmath22 ( fig .
[ fig : spectra ] , table [ tab : xrayspec ] ) . due to the low number of counts all simple models obtained similar @xmath113 values , with the powerlaw model confirming a soft spectrum ( @xmath110 ) and the blackbody fit unable to constrain the @xmath49 to a significantly non - zero value .
indicative of a foreground galactic location .
transient source with duty cycle @xmath114 and no optical counterpart , if located in ic 10 the limiting magnitude @xmath115 implies @xmath65@xmath116 subject to a typical @xmath49 .
the quantile value is also consistent with an hmxb .
another high duty - cycle , hard , variable source without optical counterpart . as with src 42
, the object could be an hmxb or agn with its high @[email protected] , and limiting @xmath115 implying @xmath117 depending on absorption .
hard transient source with duty cycle @xmath114 and no optical counterpart , the lightcurve is reminiscent of a be - xrb .
if located in ic 10 the limiting magnitude @xmath115 implies @xmath65@xmath118 subject to a typical @xmath49 .
the quantile value is also consistent with an hmxb .
this is a persistent variable source with a hard x - ray spectrum ( median energy = 2.7 kev ) , and no optical counterpart in m07 .
table [ tab : variables ] reports 9 broad - band detections out of 10 observations , however the x - ray lightcurve ( fig [ fig : lcs_lt5 ] ) shows a hard - band only detection in that observation ( final point in 2010 ) , so the source has a 100% duty cycle .
there are not enough counts to analyze the spectrum , however the quantile diagram fig [ fig : quantiles ] shows src 42 ( the point at @xmath119 ) is consistent with a heavily absorbed powerlaw @xmath120@xmath22 .
the limiting magnitude of @xmath121 constrains the optical counterpart to @xmath122 for typical ic10 absorption , however the likely much higher @xmath49 could hide an o star .
the flux ratio @xmath65@xmath1 + 1.2 together with the hard spectrum and variability make an hmxb the most likely interpretation , although it could conceivably be a background agn .
variable source on all timescales probed , and seen in all observations except for the 2003 exposure ( which places a limit on its quiescent flux ) , although since it appears as [ w05]24x we know the source was seen by _ xmm - newton _ in 2003 .
the lack of an optical counterpart limiting magnitude @xmath115 implies @xmath65@xmath123 if located in ic10 .
the energy quantile is hard and lies in the hmxb region .
this object is a large amplitude variable with 3 detections out of 9 observations .
two of these detections are mere days apart , and show a factor of 5 change in flux ( perhaps a flare ) , with the fainter point only seen due to the long observation ( 45 ksec ) , while a 3rd and final detection was seen in the soft - band ( only ) in a single one of the 2009/10 monitoring series .
there is a bright optical counterpart in m07 , which at @xmath124=17.3 , @xmath125=+1.6 is almost certainly a galactic star ( the alternative being an extreme luminosity @xmath105=-9.6 hyper - giant if located in ic 10 ) .
assuming spectral type m0 ( based on the color index ) the distance modulus is @xmath126 giving a distance of 450 pc .
the distance independent flux ratio @xmath65= -1.5 supports this interpretation but an optical spectrum is needed to confirm it .
lllllllllll 1 & j002008.68 + 591340.7 & 0.15 & 0.43 & 22.384 & 0.700 & -0.087 & 0.762 & 0.855 & 0.77 & n + 8 & j002012.68 + 591501.4 & 0.1 & 0.32 & 22.974 & 1.242 & & 0.735 & 1.144 & 1.45 & n + 13 & j001958.69 + 591527.7 & 0.37 & 0.53 & 20.473 & 1.525 & & 0.932 & 0.962 & -0.33 & n + 20 & j002029.08 + 591651.7 & 0.30 & 0.28 & 21.722 & 0.905 & -1.094 & 0.137 & 0.754 & 2.43 & + 20 & j002029.12 + 591651.8 & 0.17 & 0.28 & 22.478 & 0.017 & -0.916 & 0.851 & 0.805 & 2.74 & y + 28 & j002046.55 + 591731.7 & 0.61 & 0.53 & 18.793 & 1.521 & 1.327 & 0.957 & 0.958 & -0.53 & + 28 & j002046.36 + 591732.4 & 0.55 & 0.80 & 20.199 & 1.786 & 99.999 & 1.191 & 1.395 & -0.53 & n + 29 & j002009.14 + 591738.3 & 0.32 & 0.39 & 21.777 & 1.059 & -0.820 & 0.536 & 1.052 & -0.09 & n + 29 & j002009.15 + 591738.4 & 0.37 & 0.39 & 21.732 & 0.932 & -0.588 & 0.494 & 1.049 & -0.1 & n + 46 & j002020.94 + 591759.3 & 0.28 & 0.31 & 19.954 & 1.211 & -0.259 & 1.043 & 0.820 & 0.54 & y + 50 & j002047.28 + 591935.7 & 0.89 & 0.66 & 22.539 & 1.823 & & 1.292 & 1.564 & 0.6 & + 52 & j002033.29 + 592135.4 & 1.16 & 0.75 & 17.277 & 1.551 & 1.365 & 0.946 & & -1.5 & + 53 & j002013.70 + 592236.1 & 0.49 & 0.96 & 21.404 & 1.872 & & 1.354 & 1.603 & -0.2 & + 90 & j002028.42 + 592024.0 & 0.34 & 0.54 & 21.348 & 1.805 & & 1.294 & 1.579 & 0.37 & y + 18 & * & & & & & & & & & y +
from a total of 110 unique point sources detected in 10 _ chandra _ observations ( years 2003 - 2010 ) , we found 21 that are variable at the 3@xmath0 level .
the x - ray light curves exhibit a range of qualitatively different behaviors , including transients , large amplitude variables , and persistent sources with duty cycles ranging up to 100% .
analysis in the x - ray quantile diagram reveals a split between soft thermal sources and harder power - law like spectra that cluster along the model grid contour consistent with the average column density towards ic 10 .
while this separation is not a complete diagnostic of source classes , it supports to a large extent the hypothesis of @xcite that ic 10 hosts numerous hmxbs .
our variability study , better positions , and the increased spectral s / n ratios made possible by analyzing a cumulative 225 ksec of chandra exposure ( vs the previous 30 ksec ) have provided a richer and more detailed view of this population . in this paper , we were able to confidently classify most of the variable sources and to usefully constrain the nature of the remaining sources .
deeper observations are needed to increase the number of individual sources for which x - ray spectral fits can be obtained .
variability is a characteristic of all known astrophysical x - ray production mechanisms , and the fact that the majority of the catalog sources do not make the 3@xmath0 cut is due to their faintness ( and consequent difficulty in establishing variability ) of most objects encountered .
the hmxb candidates presented in this paper are all characterized by power - law spectra , strong x - ray variability , and either the presence of an optical counterpart consistent with an o / b star in ic10 ; or the lack of an optical counterpart to a limiting magnitude of @xmath127 which still allows for the existence of an early type star ( b0v or fainter , @xmath100 for minimal ic 10 extinction @xmath89 ) .
substantially higher extinction values do exist in parts of ic 10 as there is a high degree of structure in the interstellar medium ( ism ) of that galaxy . in the former category ,
we find src 1 , src 8 , and src 29 , in addition to the well - studied src 20 = ic 10 x-1 and src 46 = ic 10 x-2 , for a total of 5 hmxbs with optical counterparts , only one of which has a known orbital period and three have no optical spectrum . of the sample in this paper ,
only 12 x - ray sources have positional optical counterparts in @xcite , among which we identify 4 sources with @xmath11 line emission .
optical spectra have been obtained for just 4 counterparts to x - ray variables .
these were the known bh+wr binary ic 10 x-1 ; the previously reported hmxb ic 10 x-2 ; a reddened early - type star associated with the hard absorbed transient source # 29 ; and a foreground dme star associated with the soft transient source # 90 . given that highly luminous stars exist in ic 10 , spectroscopic followup is required to positively identify the majority of the optical counterparts to these sources .
the photometric colors of the counterparts are expected to be somewhat degenerate between extinction and spectral type due to the large line - of - sight depth through the foreground galactic plane .
future efforts must be directed at obtaining optical spectra of the counterparts which can be efficiently done with multi - object spectrographs on 8-m telescopes thanks to the compact size of ic10 .
positive spectroscopic identification of the x - ray binaries requires spectral types , confirmed ic 10 membership via radial velocity , and ultimately orbital elements from rv and/or photometry .
the distance and extinction values to ic 10 indicate that only the upper - end of the spectral sequence can be resolved from ground - based telescopes , however this is not an insurmountable problem since hmxbs are formed from precisely these stars .
this paper reports objects with x - ray properties similar to known hmxbs , both with and without candidate optical counterparts .
no transients were seen at super - eddington luminosity , thus no candidate bh x - ray novae occurred during any of the observations .
transient x - ray sources are conventionally identified by a variability ratio in the @xmath110 range and transient candidates in the 5 - 10 range ( e.g. @xcite ) .
we are aware of powerful and innovative approaches that rely on bayesian analysis ( e.g. @xcite ) to obtain the probability that the variability ratio for an individual source lies within the range of interest .
these techniques reduce bias and hence offer two advantages : ( a ) enhanced sensitivity to extremely faint transients even in the case of a very small number of observations , ( b ) reduced tendency to falsely identify positive photon - counting fluctuations as variability .
our selection criteria are based on the conventional statistical prescription for comparing data points .
in particular the 3@xmath0 cut is conservative ( 99% significance ) and avoids sweeping up low probability transient candidates . with 110 point sources detected , the expectation is of order 1 false variable candidate out of the 21 variables identified . admittedly , the cost of this conservative approach is reduced sensitivity to fainter transients
. there are 36 single - detection b - band sources , only 2 of which made it into the sample presented here .
as we showed in figure [ fig : hardint ] all of those 34 other objects are very faint , therefore they are not hmxbs in outburst . as to the nature of the new hmxbs reported here
, we can already assert that they are a different population than that in the magellanic clouds .
the young age of the underlying stellar population and the elevated production of massive stars and stellar remnants noted by other authors ( e.g. @xcite ) have implications for the expected x - ray binary population .
the smc with its episodic starburst history contains at least 100 known or suspected hmxbs , the vast majority of which are be+ns systems , and all earlier than b3 @xcite .
it was first proposed by @xcite that the narrow spectral type range seen in be - hmxb is an evolutionary signature .
this idea is supported by the groundbreaking work of @xcite , who showed the smc be - hmxb are associated with distinct stellar populations of age 40 - 70 myr and @xcite who found a similar age - association for hmxbs in ngc 300 and ngc 2403 . in ic 10
there has not been sufficient time to form these objects , therefore we expect to see completely different hmxb species . following this logic , we assume that the be phenomenon is not yet widespread in ic 10 , hence accretion - powered x - rays must involve mass donors with stronger winds and/or smaller orbital separation in order to deliver the requisite mass - transfer rate .
supergiants and wr stars are the obvious candidates , although we find only one x - ray source ( x-1 ) coincident with the wr catalog of @xcite ; weak - lined wr stars could be present but as - yet unidentified .
some very young hmxbs have recently been reported : @xcite found an age of @xmath45 myr for sn 2010da ( an hmxb in ngc 300 ) ; in own galaxy , cir x-1 has been revealed as the youngest known related example at @xmath44600 yr @xcite . however , there is currently debate about the nature of the optical companion in sn 2010da ( lbv , yellow supergiant , etc . ) , and the companion in cir x-1 has proven infamously difficult to classify .
@xcite proposed that the recent sn in cir x-1 has shock - heated the envelope of the companion , puffing it up , which can reconcile the strangely red color for an hmxb .
consequently we are wary of excluding red objects from the ic 10 sample and we must obtain spectra of all counterparts . in summary , our _ chandra _ monitoring study has found a heterogeneous population of exotic hmxbs in ic 10 .
we are witnessing a starburst at an early stage , even perhaps before its peak @xcite , so these unfamiliar systems will play an important role in understanding the age / metallicity dependence of massive binaries , the x - ray binary luminosity function , and ulx sources .
we are most grateful to the anonymous referee for their constructive improvements to the manuscript .
this project was made possible by the support of sao grant nas8 - 03060 and the physics department of university of massachusetts , lowell .
sl also thanks the gemini observatory for its support .
gemini is operated by the association of universities for research in astronomy , inc .
, on behalf of the international gemini partnership of argentina , australia , brazil , canada , chile , the united kingdom , and the united states of america . | we have monitored the cassiopeia dwarf galaxy ( ic 10 ) in a series of 10 _ chandra _ acis - s observations to capture its variable and transient x - ray source population , which is expected to be dominated by high mass x - ray binaries ( hmxbs ) .
we present a sample of 21 x - ray sources that are variable between observations at the 3@xmath0 level , from a catalog of 110 unique point sources .
we find 4 transients ( flux variability ratio greater than 10 ) and a further 8 objects with ratio @xmath1 5 .
the observations span years 2003 - 2010 and reach a limiting luminosity of @xmath2erg s@xmath3 , providing sensitivity to x - ray binaries in ic 10 as well as flare stars in the foreground milky way .
the nature of the variable sources is investigated from light - curves , x - ray spectra , energy quantiles , and optical counterparts .
the purpose of this study is to discover the composition of the x - ray binary population in a young starburst environment .
ic 10 provides a sharp contrast in stellar population age ( @xmath410 my ) when compared to the magellanic clouds ( 40 - 200 my ) where most of the known hmxbs reside .
we find 10 strong hmxb candidates , 2 probable background active galactic nuclei , 4 foreground flare - stars or active binaries , and 5 not yet classifiable sources .
complete classification of the sample requires optical spectroscopy for radial velocity analysis and deeper x - ray observations to obtain higher s / n spectra and search for pulsations .
a catalog and supporting dataset are provided . |
compounds of the form cxcl@xmath0 exhibit an interesting phase behavior whose complexity depends on the nature of the x substituent . for example , cbrcl@xmath0 and ccl@xmath1 have an intermediate plastic phase between the liquid and the pure crystalline phases @xcite . on the other hand , crystalline chcl@xmath0 melts to its liquid form without the intermediate plastic step @xcite .
these compounds , along with other tetrahedral molecules , have received considerable attention in the recent years @xcite . besides the importance that these molecules have due to their application in several industries and their environmental impact
, they represent an interesting academic problem @xcite .
in particular , the plastic phases could be regarded as simplified models for liquids that lead to the formation of simplified glass systems .
the comparative study of the local order and orientational correlations of these compounds allows us to understand the different microscopic scenarios in the liquid and solid phases @xcite . in this letter
we concentrate our attention on the kinetic and thermodynamic behavior of chcl@xmath0 , complementing our previous studies for x = cl @xcite and x = br @xcite . here
we perform a molecular dynamics study ( md ) of solid and liquid chloroform , along with nuclear quadrupolar resonance ( nqr ) measurements of the molecular relaxation times in the solid phase . our new findings , together with our previous knowledge about the other two systems of the cxcl@xmath0 family ,
allow us to interpret the role of the substituent x in determining the stable phases and their molecular kinetics signature .
nqr experiments were performed on a sample of chcl@xmath0 ( fluka - sigma - aldrich # 25670 ) placed in a glass ampoule and sealed under vacuum .
the @xmath2cl nqr transition frequencies and spin lattice relaxation times ( t@xmath3 ) were measured between 77 k and melting temperature 210 k by means of a conventional homemade fast fourier transform ( fft ) pulsed spectrometer @xcite .
the relaxation times were measured by using the ( @xmath4/2-@xmath5-@xmath4/2 ) sequence .
the temperature was controlled within 0.1 k using a homemade cryostat .
we performed md simulations of chcl@xmath0 in solid and liquid conditions in the npt ensemble using the gromacs v.4.5.5 simulation package @xcite .
the dynamics equations were integrated using the leap - frog algorithm with a time step of 0.001 ps .
the simulated system consisted of 3920 molecules ( 19600 atoms ) , periodic boundary conditions were applied in all three cartesian directions and a spherical cutoff was imposed at 2 nm for all interactions .
the temperature was controlled with the velocity rescale algorithm @xcite with a time constant of 0.2 ps , and the pressure maintained at 1 atm with a berendsen anisotropic barostat with a time constant of 0.5 ps and compressibility of 7 @xmath6 bar@xmath7 .
the md simulations of the solid phase of chcl@xmath0 were performed using as initial configuration the structure determined by x - ray diffraction@xcite .
the crystalline structure of chloroform is orthorhombic and corresponds to the space group pnma 62 , with four molecules per unit cell .
the melting temperature at ambient pressure is 210 k. the lattice parameters , which were determined at 185 k , are @xmath8 , @xmath9 and @xmath10 @xcite . the initial configurations for the simulations in the solid phase consist of perfect crystalline systems of @xmath11 unit cells , which include @xmath12 molecules .
these configurations were relaxed at each temperature with short nvt runs of 100 ps .
several models for chcl@xmath0 consisting of different sets of lennard - jones parameters , charges and slightly different interatomic distances can be found in the literature @xcite , and a comparison between them and several parameter variations can be found in ref .
@xcite for liquid chloroform at ambient conditions . here ,
after tests in the liquid and solid phases , we adopted one variation of a model originated by dietz and heinzinger @xcite with atomic charges resulting from an optimized hf / aug - cc - pvdz geometry @xcite . this five sites chcl@xmath0 model is based on a rigid , distorted tetrahedron with the c atom at the center , three cl atoms on three vertices , and a h atom located in the remaining vertex .
the intermolecular interactions were described by a lennard - jones ( l - j ) potential plus a coulombic term .
the l - j parameters ( @xmath13,@xmath14 ) for interactions between different atoms were determined using the combination rules @xmath15 and @xmath16 , where @xmath17 , @xmath18 represent the three atomic species c , cl and h. the details of the parameters and model architecture are summarized on table [ tabla1 ] .
l | c c c atom & @xmath13 [ nm ] & @xmath14 [ kj / mol ] & @xmath19 [ e ] + c & 0.340 & 0.427 & 0.569 + cl & 0.344 & 1.255 & -0.177 + h & 0.280 & 0.084 & -0.038 + + + +
in order to determine the correlation times corresponding to the rotational jumps of chcl@xmath0 we measured the nqr spin lattice relaxation times , @xmath20 , between 77 k and the melting point at 210 k. in the solid phase , the chcl@xmath0 spectra displays only two nqr lines with an intensity ratio of 2:1 , see top - right inset of figure [ nqr ] .
these two lines correspond to the the two nonequivalent cl atoms in the asymmetric unit as revealed by the crystallographic studies of fourme and renaud @xcite .
the nqr relaxation times of the two lines are shown in figure [ nqr ] .
below 180 k the @xmath20 data shows a weak temperature dependence ( @xmath21 ) while above that temperature @xmath20 has the behavior characteristic of thermally activated process . in molecular crystals and for rigid molecules ,
the quadrupolar relaxation is usually due to two types of mechanisms : a phonon or molecular libration contribution @xcite and a rotational contribution resulting from the slow reorientational jumps of the molecules between equivalent positions @xcite .
then , the relaxation rates for each nucleus that participates in the reorientation can be described by the sum of these two independent processes , and for a three - fold symmetric molecule @xcite @xmath22 .
therefore , subtracting the phonon contribution , the reorientational contribution to the relaxation process , @xmath23 , can be extracted from the data and the resulting reorientational correlation time @xmath5 is shown in the bottom - left inset of figure [ nqr ] .
we began our md study by estimating the melting temperature of the model . in order to have an estimation of the model s melting temperature we performed several tests , starting from checking the stability of the crystalline phase as a function of the temperature .
it was found that at 240 k the system remains as a solid for up to 50 ns of simulation . at 250
k the system melts in less than 1.5 ns . since a small crystalline sample such as the simulated one
is expected to persist as a solid under superheated conditions , 250 k represents an upper limit for the melting temperature of the model chcl@xmath0 .
next we study the stability of a system consisting of a solid slab in contact with a liquid slab , and it was found that the system melts at 180 k but remains stable for up to 50 ns at 160 k and lower temperatures . in this case , the later temperature represents a lower bound for the melting temperature since the fluctuations , which are relatively large due to the intrinsic small size of the simulated system , tend to favor the phase with larger entropy .
therefore , the model s melting temperature is between these two limits , and although this is a large interval , it suffices for the purpose of the analysis that we present below .
.[tabla]molecular rotations in solid chcl@xmath0 .
@xmath24 is the observed number of molecules that rotated at least once in a simulation of @xmath25 ns , @xmath26 is the corresponding probability according with the experimental relaxation time ( see text ) and @xmath27 is the relaxation time of the c - cl molecular bonds calculated from the md simulations . [ cols="<,^,^,^,^",options="header " , ] the nqr measurements show that the orientational relaxation time is @xmath28 ms at 205 k. this time is essentially infinity for an atomistic simulation so we can not perform a direct comparison between experiments and simulations .
however , in simulations of 50 ns we are able to see some molecular reorientations that become more frequent as the temperature is increased .
these molecular motions are in all cases sudden jumps between equivalent positions achieved by rotations of 120@xmath29 about the c - h bond . as a result
, the auto correlation function of the c - h bonds does not show any decay with time besides a sharp small drop at very short times due to the molecular librations .
the sudden 120@xmath29 rotations are reflected by a slow decay of the auto correlation function of the c - cl bonds , see figure [ acfs ] .
then , we analyze the rotational jump dynamics in a total time of @xmath25 ns . this is done assuming that the rotations of a single molecule follow a poisson distribution characterized by the experimental rate of rotations , _ i.e. _ @xmath30 . the probability that a molecule jumps at least once , @xmath31 , is taken as the probability of success in a binomial distribution considering @xmath32 independent trials .
next we calculate the probability of having @xmath24 molecules jumping at least once during the simulated time @xmath33 , which is shown in table [ rots ] .
the table also includes the relaxation times calculated from the rotational autocorrelation functions of the c - cl bond vector , denoted by @xmath27 .
these correlation times , much longer than the total simulation time , were estimated from the initial slope of autocorrelation functions assuming an exponential decay .
the expectation value for the number of molecules having at least one rotation is @xmath34 , which approximately coincides with the number of molecules that jump at least once observed in the simulation at 200 k. moreover , as the temperature increases above 210 k the number of molecular jumps dramatically increases and therefore can not be explained in terms of the experimental relaxation time at pre - melting conditions .
this collection of results , the lower and upper bounds described above together with the probability analysis , point to a melting temperature of the simulated model very close to the experimental melting temperature of 210 k. the molecular rotations in the solid phase are essentially about the c - h bond vector .
this is clear from the auto - correlation functions of the c - cl and c - h bond vectors displayed in figure [ acfs ] for the _ overheated _ solid at 240 k. both correlations show a fast initial decay from 1 to @xmath350.95 due to the molecular librations , but the c - h bond remains in the same direction while the c - cl bonds relax due to the sporadic molecular jumps .
the question that arises is whether this anisotropy in the rotations is preserved upon melting or not . in figure [ acfl ]
we show the corresponding curves for a liquid system at a stable 240 k , and at _ supercooled _ conditions at 190 k. the two systems show a similar rate for the rotations about the two distinct bond vectors .
the disappearance of the rotational anisotropy is possible due to a decrease in the density of the system of approximately 10 % as it melts . in order to understand what are the molecular rearrangements upon melting that lead to the decrease in density we analyze the pair distribution functions , @xmath36 , for the different atom pairs .
in particular , we plot on figure [ gofr ] the @xmath36 for c - h and c - cl pairs .
the curves corresponding to the solid phase at 190 k display the signature multi - peak pattern of the lattice structure .
the pair correlation functions for the liquid systems reveal the local structure of the fluid that dies off for @xmath37 nm .
the picture that emerges is that , as the system melts , the strong c@xmath38h association in the solid phase is lost resulting in an overall expansion of the volume .
this is clear from @xmath39 that has a distinctive peak at @xmath40 nm in the solid structure that completely disappears in the liquid , moving the peak to @xmath41 nm .
the cc pair correlation functions also reflect the overall expansion of the volume , which in turn , provides sufficient space for the molecules to rotate with no preferential axis .
the strong c@xmath38h association is the result of the electrostatic interactions at short distances , which has also been considered to be the microscopic origin of polar phases in chloroform and bromoform at higher pressure @xcite .
a more detailed analysis of the local structure of the liquid , based on a geometrical criteria , has been applied recently to several tetrahedral molecular systems of the form zy@xmath42 , where the four y atoms may be of different type @xcite .
the idea is based on considering the number of y atoms between the two parallel planes containing the z atoms of neighboring molecules . depending on this number and to which molecules these y atoms
belong , six different classes of molecular arrangements can be defined : @xmath43 , @xmath44 , @xmath45 , @xmath46 , @xmath47 and @xmath48 .
the numbers refer to the number of atoms from each molecule located between the two parallel planes .
the probabilities of these different classes , given that the distance between the two z atoms is @xmath49 , is denoted by @xmath50 . in a recent study of cbrcl@xmath0 @xcite we claimed that a more meaningful quantity is the product @xmath51 , which produces partial radial distribution functions , since it directly removes those configurations that have negligible contribution to the overall structure of the system . in figure [ rey ] we show @xmath51 for the liquid system at 190 k and 300 k. there are no major differences for the two cases , only the expected softening of the liquid structure as the temperature is increased . at short distances ,
the _ edge - to - edge _ and _ edge - to - face _ configurations are the most probable .
_ face to - face _ arrangements are meaningful at very close distances , and _ corner - to - face _ and _ corner - to - edge _ have their maximum contribution at slightly larger separation between molecules . this general pattern and relative order of the different molecular orientations
is remarkably similar to one observed for the case ccl@xmath0br , that is , with a large substituent atom .
further insight on the local molecular ordering can be gained by looking at the probability @xmath52 of finding @xmath53 substituent atoms ( @xmath54 , @xmath55 or @xmath56 ) between the two parallel planes defined by the c atoms of neighboring molecules . moreover
, the size of the substituent atom may play a role in this probability and therefore we compare our new results for chcl@xmath0 with our previous calculations for cbrcl@xmath0 . in figure [ reyka ]
we show @xmath52 for the two systems at 300 k. while for a h substituent at closest approach @xmath57 is maximum , for the br case is @xmath58 the most probable configuration .
this can be understood from pure geometrical considerations .
since the h atom is much smaller than the cl atoms , the closest approach between two molecules happens with a _ corner - to - corner _
configuration directly placing the h atoms facing each other . since the br is considerably larger than the cl atom the closest approach results from a _ face - to - face _ arrangement with the br atom on the opposite sides of the facing planes .
interestingly , this difference is washed away when we introduce the @xmath59 as a weighting factor , as shown in figure [ reyk ] for the liquid system at 190 k and 300 k , which shows remarkable similarity with the corresponding curves for cbrcl@xmath0 . for large separations
, the probabilities go to 0.5 for @xmath60 and 0.25 for @xmath54 and @xmath61 , which are the expected values from the relative number of the different atomic species in the system . at short range ,
the most probable configurations are those with @xmath60 . a comparative analysis of the findings of this work with those corresponding to ccl@xmath1 and cbrcl@xmath0 allow us to draw some general conclusions .
the essential difference among these three molecules is the size of the substituent atom .
the perfect tetrahedral structure has an intermediate fcc plastic phase between the solid and the liquid .
substituing a cl atom by a larger br atom preserve the intermediate plastic phase .
however , the substitution with a much smaller h atom suppresses completely the plastic phase .
the dynamics of the chcl@xmath0 solid phase is _ qualitatively _ different than that of the other two compounds : while rotations about all the bonds are observed in the cbrcl@xmath0 and ccl@xmath1 , only rotations about the c - h bond are possible in the present case .
nb , mz and ps and would like to acknowledge secyt - unc , conicet , and mincyt crdoba for partial financial support of this project . | we present a detailed analysis of the molecular kinetics of chcl@xmath0 in a range of temperatures covering the solid and liquid phases . using nuclear quadrupolar resonance
we determine the relaxation times for the molecular rotations in solid at pre - melting conditions .
molecular dynamics simulations are used to characterize the rotational dynamics in the solid and liquid phases and to study the local structure of the liquid in terms of the molecular relative orientations .
we find that in the pre - melting regime the molecules rotate about the c - h bond , but the rotations are isotropic in the liquid , even at supercooled conditions . nuclear quadrupolar resonance , molecular dynamics , chloroform , melting point |
the lhcb detector@xcite is a single - arm forward spectrometer designed for the study of heavy flavor hadrons .
many results have been published by the lhcb collaboration , in particular on cp violation in decays of @xmath2 and @xmath3 hadrons .
in contrast , until recently lhcb had made no measurements on cpt violation in these decays . in these proceedings , a new result@xcite from the lhcb collaboration on cpt violation in @xmath0 mixing and @xmath1 mixing is highlighted .
violation of cpt symmetry implies a breaking of lorentz invariance in a local , interacting quantum field theory.@xcite this means that any cpt - violating parameter must also violate lorentz invariance . the standard - model extension ( sme )
is an effective field theory , where cpt- and lorentz - violating terms are added to the standard - model lagrangian.@xcite this framework provides the experimental opportunity to measure the coupling coefficients in these terms .
the lhcb result@xcite presented here is given in terms of these sme coefficients .
in the past , there have been many experimental searches for cpt violation in neutral - meson systems.@xcite the majority of these searches have been done without any assumption on the breaking of lorentz invariance , referred to as the classical approach . in the last 15 years , more searches have been performed within the sme framework , placing tight constraints on its coefficients.@xcite
the particle - antiparticle mixing between neutral - meson states creates an interferometric system that enhances the sensitivity to cpt violation enormously .
conservation of cpt symmetry implies equal mass and lifetime of particles and antiparticles .
the cpt - violating observable in the mixing process is given by @xmath4 where @xmath5 and @xmath6 are the ( cpt - violating ) mass and decay width differences between the particle and antiparticle states .
the high sensitivity to @xmath7 comes through the small values in the denominator of the eigenvalue differences , @xmath8and @xmath9 , of the two - state system . in the sme framework ,
the @xmath7 observable becomes@xcite @xmath10 where @xmath11 is the four velocity of the neutral meson and is a real four - vector vacuum expectation value that describes the coupling with the mesons .
the complex parameter @xmath7 can be determined directly from the decay rates as function of the decay time of the neutral meson.@xcite there are four systems of neutral mesons . in all of them ,
the mixing formalism is identical , however , their phenomenology is very different owing to the different values of @xmath8and @xmath9 , and number of decay modes . in the @xmath12 system
, there have been many searches for cpt violation by dedicated kaon experiments ( kloe , ktev , cplear , and na48 ) following the classical approach .
an experimental overview is given in ref . .
strong constraints on the sme coefficients have been made using data from kloe , ktev and e773.@xcite it will be difficult for lhcb to compete with these dedicated kaon experiments due to the lower statistics and worse kaon lifetime acceptance .
the situation is already different in the @xmath13 system . only a single measurement exists , by the focus collaboration,@xcite using a sample of 35k @xmath14 decays .
lhcb would be able to improve this measurement significantly owing to the 50 m @xmath14 decays , collected during run 1.@xcite in the following , i will focus on the two remaining neutral - meson systems : the and systems .
in both the system and system , @xmath9is negligibly small compared to @xmath8 .
the standard model predicts that @xmath9is about a factor 200 smaller than @xmath8,@xcite which is already confirmed in the system . since is real , it follows from eq .
( [ eq : z ] ) that @xmath15is a factor 400 smaller than @xmath16 . therefore , to constrain the sme coefficients , decays to cp eigenstates are used , which are more sensitive to @xmath16 in comparison to using decays to flavor - specific final states.@xcite the golden decay modes to cp eigenstates , @xmath17 and @xmath18 , have been used due to their relatively large branching fraction .
these modes have been studied already at lhcb to measure @xmath19 and @xmath20.@xcite for the present analysis , the fit models have been extended to allow for possible cpt violation.@xcite the results@xcite are shown in table [ tab : cptresults ] .
no significant sidereal variation and no violation of cpt symmetry are observed . in the system
, there is a large improvement of three orders of magnitude with respect to the previous best result@xcite from babar . in the system
, there is an order of magnitude improvement with respect to the previous best result@xcite from .
the improvements are primarily attributed to the large boost of the mesons at lhcb ( i.e. , @xmath21 versus @xmath22 at babar and @xmath23 at ) .
[ tab : cptresults ]
in summary , interferometry with neutral mesons provides a sensitive method to test violations of cpt symmetry and lorentz invariance . the lhcb experiment is well suited to improve the sme bounds , in particular due to the high boost of the particles produced at the lhc .
there are plans to measure @xmath7 and in the system , which aim to improve the current bounds by a factor 40 . as highlighted in these proceedings ,
greatly improved limits on cpt violation and lorentz symmetry breaking in mixing have been published by lhcb.@xcite these results are based on an integrated luminosity of 3obtained in run 1 of the lhc . at the end of run 2 in 2019 , an expected @xmath24-@xmath25 will be added . due to the larger cross sections at the new center - of - mass energy of 13@xmath26 , the heavy flavor yields are almost a factor two higher in run 2 .
furthermore , the meson boost will also be about 30% higher .
together this means that the uncertainties will reduce by a factor two .
a further improvement can be expected from the upgraded lhcb detector that will start data taking after 2019 : with a projected 50the uncertainties will drop by more than a factor of six .
this work is supported by the netherlands organisation for scientific research ( nwo vidi grant 680 - 47 - 523 ) and the foundation for fundamental research on matter ( fom ) . | recent measurements of cpt violation and lorentz symmetry breaking in @xmath0 mixing and @xmath1 mixing , obtained from data taken by the lhcb experiment , are highlighted .
the results are expressed in terms of the standard - model extension ( sme ) coefficients , which incorporate both cpt and lorentz violation . due to the large boost of the mesons at lhcb
, the sme coefficients can be determined with high precision .
the bounds on these coefficients are improved significantly compared to previous measurements . |
we have seen many presentations in this conference illustrating the enormous richness of emission line phenomenology in agn , as well as the great progress which has been made in understanding their excitation mechanisms and locating their excitation regions . in this
talk i wish to step back for a moment and ask what the study of emission lines has gained us with respect to answering the most fundamental questions about agn .
innumerable proposals requesting support for agn emission line work have been justified by the assertion that these lines may , much as emission lines from fusion plasmas do , serve as diagnostics of the entire system .
i wish to examine to what extent this assertion has been borne out so far , and evaluate the prospects for success of this program in the foreseeable future .
the natural place to begin is to list the most important questions of agn studies . to my mind , there are six that stand out clearly as being truly fundamental : 0.1 in @xmath2 1 ) what is the power source ?
is it really accretion onto a massive black hole ?
@xmath2 2 ) what triggers nuclear activity in galaxies ? put another way , what regulates the accretion flow ? why was powerful nuclear activity so much more common at @xmath3 3 than it is today ?
@xmath2 3 ) why is the continuum emission so broad - band ? what balances the emission so that ( to order of magnitude ) the flux is constant per logarithmic frequency from the mid - infrared through hard x - rays ?
@xmath2 4 ) how are jets accelerated and collimated
? why do they exist at all ?
@xmath2 5 ) how can we relate the different varieties of agn ?
what controls the choices in type of activity ?
for example , what creates the correlation between host morphology and radio ( i.e. jet ) power ?
@xmath2 6 ) does the existence of an active nucleus affect the evolution of the host ? 0.1 in despite more than thirty years of intense activity in this field , most of these questions remain almost as open today as they were decades ago .
we do have partial answers to questions 1 and 5 , and emission line studies have contributed significantly to providing those answers , but we have made little headway with the others .
the central issue before us is how emission line studies may be used to provide more complete answers to all these questions .
my review of the contributions of emission line work will be ordered from the outside in , moving from the narrow line region inward toward the black hole s event horizon .
historically most of the effort devoted to analyzing narrow emission lines has been devoted to inferring the physical conditions in which they are excited .
however , those details have told us little about other , more important issues .
it is , rather , the shape of the region from which they are emitted that has revealed the greatest information .
as reviewed , for example , by wilson ( 1994 ) , in a great many seyfert galaxies the narrow line region is a double cone whose apex coincides with the galactic nucleus .
the edges of these cones are often so sharply defined and so straight that we can imagine hardly any other explanation for them but the projection of light rays ( a particularly clear example is displayed in figure 1 ) .
that is , the existence of these cones tells us that there is a sharp - edged toroidal collimating structure close to the nucleus which permits ionizing photons to escape only within a certain angle of a preferred axis .
several important conclusions follow immediately .
the first is that agn appear very different , depending on one s viewing angle .
if an observer s line of sight lies within the favored cone , a strong ionizing continuum can be seen ; if outside , the ionizing continuum ( and likely the non - ionizing optical continuum also ) will appear much weaker .
thus , there must be categories of agn which are separated by empirical classification schemes but are intrinsically the same .
some of the now almost universal support for the unification of type 1 and type 2 seyfert galaxies ( antonucci 1993 ) stems from these observations of conical narrow emission line regions .
another conclusion is that the surprisingly weak impact agn appear to have on their hosts is due in part to the collimation of ionizing photons so close to the nucleus .
if the absorbing structure obscures @xmath4 of solid angle ( as indicated by both the opening angles of the cones and the statistics of type 1 and type 2 seyferts ) , then most of the host galaxy is protected from exposure to the agn s ionizing continuum .
we are beginning to have some grounds for hope that narrow emission lines may also help us learn about the dynamics of the accretion flow at distances @xmath5 1000 pc from the nucleus .
nelson & whittle ( 1996 ) have discovered a strong correlation between the integrated width of the narrow lines and the velocity dispersion in their hosts bulges .
this correlation strongly suggests that narrow line dynamics are dominated by gravity , although the spatial separation between the narrow line region and the portion of the bulge where the stellar dispersions are generally measured makes the nature of this connection somewhat obscure .
detailed images of the line profiles , as provided , for example , by fabry - perot spectroscopy ( cecil et al .
1990 ) , have the potential to reveal more about the nature of the forces directing gas towards the center .
care will be necessary to successfully carry out this program , however . if one wishes to interpret gas kinematics as solely due to gravity , it is vital to omit those regions for which the character of the emission lines suggests shock excitation rather than photoionization .
contrasting the gas s kinematics with the kinematics of adjacent stars should go a long way towards helping us make these distinctions .
in a classic paper , antonucci & miller ( 1985 ) demonstrated that broad balmer emission lines , while invisible in the total flux of the classic type 2 seyfert galaxy ngc 1068 , are as strong in its polarized spectrum as in any type 1 seyfert . based on this finding , they inferred that the nucleus of ngc 1068 ( and by the astronomical version of mathematical induction , the nuclei of most , if not all , other type 2 seyfert galaxies ) are in fact identical to the nuclei of type 1 seyferts but for the accident of very optically thick obscuration that blocks our line of sight .
the simplest geometry for this obscuration is toroidal , as confirmed by the conical shape of the narrow line region .
we see the intrinsic spectrum of the nucleus only because a small fraction of the light is reflected our way by electrons within the illuminated cone but above the obscuration , thus having unencumbered lines of sight to both the nucleus and our telescopes .
the polarization created by the asymmetric disposition of these electrons on the sky allows us to identify the light as coming from the nucleus . clearly , the presence of strong broad emission lines in the polarization spectrum served in this case as a marker of nuclear reflected light .
many others ( again see the review by antonucci 1993 ) have told the story of how polarized emission lines showed us the path to unification of the different seyfert types , as well as ( more speculatively ) fr2 radio galaxies and radio - loud quasars . here
i wish to emphasize a less well - known way in which emission lines from the mirror tell us more about agn . because the primary scattering mechanism is electron scattering , the profile of any reflected emission line with line - center frequency @xmath6 is both broadened [ by @xmath7 and shifted [ by @xmath8 , where @xmath9 is the electron temperature , @xmath10 the electrons bulk velocity , and @xmath11 are the initial and final directions of the scattered photons .
thus , the detailed profile properties of the polarized emission lines function as diagnostics of the physical properties of the mirror gas . for example , in ngc 1068 , miller , goodrich , & mathews ( 1991 ) were able to show that @xmath12 k , and that the h@xmath13 line is shifted 600 km s@xmath14 from systemic in the sense that the scattering medium is moving outward .
x - ray spectroscopy of the mirror gas has the potential to tell us even more . when our line of sight is in the equatorial plane of the obscuration ,
we see the mirror gas by virtue of its intrinsic emission and its reflection of the nuclear light ( the specific spectrum predicted by a particular model is shown in figure 2 ) . when this gas is warm ( i.e. @xmath15
@xmath16 k as in ngc 1068 ) , electron scattering is the only scattering mechanism effective in the optical and ultraviolet bands , but numerous resonance lines enhance the gas s albedo in the soft x - ray band ( band et al . 1990 ; krolik & kriss 1995 ) .
exactly which lines are present , and at what strength , depends on details of the physical conditions .
in particular , we have good reasons , both observational ( the line shift seen in ngc 1068 described in the previous section ) and theoretical ( krolik & begelman 1986 , balsara & krolik 1993 ) to believe that this gas is flowing outward supersonically .
in such a state , the gas absorbs somewhat more heat from radiation than it emits , so its temperature is not exactly the one given by thermal balance at a fixed ionization parameter .
consequently , to specify its ionization state one must define both the ionization parameter and the temperature .
dynamical models ( e.g. the two papers just cited ) suggest that @xmath17 ( the ionization parameter defined as a ratio between radiation and gas pressure ) is @xmath18 and the temperature is between @xmath19 and @xmath16 k. that is , the gas is slightly to the net heating side of the equilibrium curve near the region of that curve where it is marginally stable with respect to isobaric perturbations . in figure 2
the spectrum has been smoothed with a gaussian filter having a fractional frequency width of 0.05 in order to illustrate the richness of the structure which might be seen when x - ray spectrographs are able to obtain both resolution on this scale , and adequate signal / noise to make use of it .
the features seen are actually blends of numerous lines , both intrinsic and reflected , whose shapes and strengths depend strongly on the state of motion of the mirror gas .
supersonic motion has an especially powerful effect on the reflected lines because it can move them from the flat part of the curve - of - growth to the linear part : the ratio between the velocity width of a line formed in a supersonic flow and the velocity width that line would have in a static gas is @xmath20 , where @xmath21 is the atomic mass of the ion and @xmath22 is the mach number of the flow . for those ions with resonance lines in the soft x - ray band , @xmath21 is typically a few tens , so a supersonic velocity gradient can widen their lines by an order of magnitude .
a polar view of the mirror ( and nucleus ) provides us with a different view that is potentially equally informative in a complementary way . in this case
the features are more likely to be absorption than emission .
because the gas is only partially ionized , photons are removed from the nuclear continuum both by photoionization ( true absorption ) and by resonance scattering ( as just discussed ) .
because the mirror covers at most @xmath23 of the solid angle around the nucleus , reemission by radiative recombination and scattering from other lines of sight into ours replace only a small part of the photons removed by absorption and scattering .
viewed down the axis , then , the mirror acts as an absorption filter in the soft x - ray band .
it is fair to ask whether these predicted absorption features have ever been seen .
if the mirror covers most of the solid angle left open by the obscuration , they must occur in many , if not most , type 1 seyfert galaxies .
estimates of the fraction of the nuclear light reflected ( e.g. miller et al .
1991 ) indicate a column density @xmath24
@xmath25 @xmath26 , and , as we have seen , other measurements and theoretical arguments suggest a temperature in the range @xmath19
@xmath16 k and an ionization parameter slightly larger than the value at which marginally stable thermal equilibria are found .
when radiative equilibrium models are fit to warm absorber spectra , very similar column densities and temperatures are found , and the ionization parameter is estimated to be very close to the marginally stable value ( e.g. netzer 1997 ) .
the mirror is , therefore , an excellent candidate to _ be _ the warm absorber .
x - ray absorption line studies may then provide another vehicle for learning about the character of the reflection region .
emission lines from the obscuring material are also useful . covering @xmath27 of solid angle , it should be a source of substantial fe k@xmath1 emission if its column density is @xmath28 @xmath26 or more ( krolik et al . 1994 ; ghisellini et al . 1994 ) .
since much of the obscuring matter is located 1 10 pc from the nucleus ( for seyfert - scale luminosities ) , the flux in fe k@xmath1 produced in the obscuring torus can only change on timescales of at least a few years .
therefore , very long duration ( 10 yr or more ) reverberation mapping experiments might be able to map out the location of k@xmath1-emitting material in much the same way as shorter duration reverberation mapping has given us much information about the location of broad emission line gas ( maoz 1997 )
. in at least one object ( ngc 2992 ) , incomplete time sampling shows line and continuum variations of just the sort that could be used for this kind of experiment ( weaver et al .
1996 ) .
the obscuring torus houses a still more powerful emission line diagnostic in the form of 22 ghz h@xmath0o masers ( braatz et al .
because interferometric techniques are so well developed at radio frequencies , and the angular scale of these masers is comparatively great ( relative to other agn structures ) , we are able to obtain genuine _ images _ of their location , without any sort of inferential guesswork . in the case of ngc 1068 , we can see them outlining the inner edge of the obscuring torus ( greenhill 1997 ) ! combining direct positional information with line of sight velocities , we can conduct detailed investigations of the torus dynamics using these maser spots .
when , as in the case of ngc 4258 ( miyoshi et al . 1995 ) , those dynamics are simple , it is possible to make very robust and precise inferences about the overall gravitational potential .
because the maser spots in that object are very closely confined to a thin ( albeit slightly warped ) plane , the spots travel in orbits that are very close to circular , and the magnitude of their orbital speed scales @xmath29 , it is very hard to escape the conclusion that their motions are simple tracers of a point - mass potential created by a central mass of @xmath30 whose size must be rather smaller than 0.1 pc . on the other hand , there are also examples , such as ngc 1068 ( greenhill 1997 ) , in which the dynamics are complicated .
interesting inferences may still be made , although the arguments are much less clean . in that case , there is a very large velocity dispersion at the innermost radii , and the upper envelope of the line of sight velocity scales much more slowly with radius than @xmath31 ; in fact , it is more like @xmath32 .
if one nonetheless makes the simplistic assumption that the motions are due primarily to gravitational forces , the orbital speed at the innermost radius may be used to yield an order of magnitude indicator of the central mass , i.e. @xmath33 . compared to the bolometric luminosity of @xmath34 erg s@xmath14 ( pier et al .
1994 ) , this means that the luminosity relative to eddington is @xmath35 . inside the torus , the spectrum peaks in the mid - infrared where , for a normal dust - to - gas ratio , the opacity per unit mass is about ten times thomson . since the radiation force is the flux times the opacity ,
these observations demonstrate that radiation force _ must _ be important to the dynamics of the torus , as predicted by pier & krolik ( 1992 ) .
these arguments illustrate an important general principle : simple sub - systems ( e.g. the maser spots in ngc 4258 ) can be so thoroughly understood that they may be reliably used for inferences about the larger system in which they are embedded .
but complicated systems ( e.g. the maser spots in ngc 1068 ) are never well enough understood that they may be used as springboards for inferences about larger issues .
inevitably we find ourselves fully occupied refining what we know about their _ intrinsic _ properties .
that activity may be fascinating and well worth the effort , but it does not lead to information bearing on larger questions .
having said that , we are now equipped to consider the role of broad emission line studies in the context of the broad sweep of agn work .
these lines offer a tremendous wealth and quality of data . in any one object
there are numerous different lines ; the flux and profile of each one may be measured ; and they vary in time in a way which is correlated with the continuum . at the same time , because the atomic physics of optical and ultraviolet line emission is well understood , it is possible to make very detailed calculations modelling their excitation and have some fair confidence that the results are meaningful . by comparing the relative strengths of different lines
one may infer the ionizing flux and pressure of the region where the lines are made ; using the methods of reverberation mapping one may also constrain the distance from the central continuum source to the line - emitting region , thereby fixing all the unknowns of the problem .
the pressure in the line - emitting regions might be taken as a measure of the pressure in the surrounding gas , which , if we were lucky , might be identified with the accretion flow through that range of radii
. the line profiles give measures of bulk velocities ; again , with good fortune , we might hope that these velocities indicate at least something about the velocities of surrounding matter . on this basis , many people have long expected that the study of agn broad emission lines would cast a bright light on the field as a whole . unfortunately , there are numerous obstacles that may prevent us from ever reaching these goals .
reverberation mapping experiments ( again see the review by maoz 1997 ) have shown that the broad line region is a very complex place .
the material responsible for individual lines is spread over at least an order of magnitude dynamic range in radius , and maybe more ( done & krolik 1996 ) .
the ranges of radius corresponding to different lines are systematically offset from each other by sizable amounts ( e.g. clavel et al .
1991 , krolik et al .
1991 ) . and
the whole structure may change in timescales as short as a few years ( peterson & wanders 1996 ) . moreover , even obtaining a believable reverberation map is subtler than might be apparent at first sight . the effective signal to
noise ratio for a monitoring experiment is _ not _ the ratio of mean flux to rms error , but rather the ratio of rms flux change ( that is , compared to the long - time mean of the flux ) to rms error . in most monitoring experiments
so far this ratio has been only a few ; even in the best ( measurements of the total flux in the civ 1549 line in the _ hst _ campaign on ngc 5548 reported by korista et al .
1995 ) it rose only to @xmath36 .
blending is frequently a problem , corrupting otherwise easily - measured lines .
it is particularly frustrating that nv 1240 almost always strongly contaminates ly@xmath1 .
even when the data are very clean , inversion of the line and continuum time series to obtain the response function _ requires _ insertion of some _ a priori _ information in order to achieve numerical stability .
this automatically means that _ all _ response functions are model - dependent on some level ( figure 3 ) .
it is possible to make this model - dependence relatively benign , but it is always present . as figure 3 makes plain , while there are certain features that are robust with respect to changes in the model assumptions , there are definitely others which are clearly model - dependent . to go from the 1-d map given by the response function to a true 3-d picture of the emission line region one inserts far greater model - dependence .
a response function is no more than a projection of the marginal response of line emissivity onto the surfaces of constant delay _ relative to our line of sight_. we certainly can not expect our line of sight to be a symmetry axis for any but a tiny minority of agn , so we must make some choice about the real symmetry in order to transform the response function into a 3-d description .
this procedure is obviously highly non - unique .
the potential pay - off from reverberation mapping becomes still greater when separate maps are made not just for different emission lines , but for different portions of the lines , separated by line - of - sight velocity .
there is no change in the technical issues ; the only respect in which velocity - resolved reverberation mapping is more difficult than integrated flux reverberation mapping is the reduction in signal / noise due to chopping up each line into several segments .
what makes the scientific yield so much more interesting is that the solution provides a constraint on the _ kinematics _ of the line - emitting gas , which we may hope will ultimately provide clues about their _
dynamics_. unfortunately , the remarks made above about non - uniqueness are underlined by the current state - of - the - art in velocity - resolved reverberation mapping .
four dramatically different models for the ngc 5548 civ 1549 response have been put forward , all claiming qualitative consistentency with the data ( in order of publication , wanders et al .
1995 ; done & krolik 1996 ; chiang & murray 1996 ; bottorff et al .
the first argues that the velocities are all random , but that a shelf " exists in the total flux response similar to that appearing in the solid curve of figure 3b .
they interpret this shelf as due to strong anisotropy of the ionizing continuum , rejecting the numerous other possible interpretations of such a feature .
the second finds no evidence against spherical symmetry and isotropy in the emission . on the other hand , while their analysis shows ( like that of the first paper ) that the line s red and blue cores respond similarly , they differ from wanders et al . in discovering that the red wing responds significantly more quickly than the other line segments , especially the blue wing . on this basis
they suggest that the line motions are predominantly random , but there is a net radial infall .
the third posits a rotating wind driven off a flat accretion disk by radiation pressure , and argues that the anisotropy of line emission in this sort of situation creates leading response in the red wing despite the outflow .
the fourth also favors a disk wind , but driven by rotating magnetic fields .
the only hope we have to break this degeneracy is to impose the twin thresholds of physical self - consistency and quantitative ( e.g. as measured via a statistic such as @xmath37 ) consistency with the data .
none of the above models passes both tests fully : the first two are physically _ ad hoc _ , while the second two are based on incompletely worked out dynamical models ; only the second has even computed the @xmath37 for its predictions .
let us suppose that over the next few years we are able to arrive at a generally - accepted picture of the kinematics of the broad line region .
how much will we able to learn then ? for a number of years
, it has been commonplace in the literature to assume that the dynamics of the line - emitting material are dominated by the gravitational force of a massive central object , and that its mass could be estimated simply by computing @xmath38 .
unfortunately , these statements have as yet little or no justification .
first of all , we do not yet have any clear indication regarding which forces dominate the broad emission line gas . in any gas
optically thick over much of the continuum bandwidth , the ratio between the mean outward radiation force and the inward gravitational force due to the central object is @xmath39 where @xmath40 is the luminosity in eddington units and @xmath41 is the optical thickness of the gas in thomson units .
we do not know with any confidence either of the two dimensionless numbers taken in ratio on the right hand side of this equation .
both have been suggested to lie anywhere in the range @xmath42 to @xmath43 .
we can hardly state with confidence , therefore , that radiation force is clearly negligible relative to gravity .
other forces may also play a role .
while we have good reason to believe that the broad emission line gas occupies only a very small fraction of the volume of its region , we have no idea what sort of gas fills the rest of the volume .
friction against this other gas could be very important to the emission line gas .
similarly , we have even less idea about the state of magnetic fields running through the region , so we have no sensible way to even begin to limit @xmath44 relative to gravity ( see , for example , rees 1987 or emmering et al .
1992 ) .
indeed , contrary to what is often said , @xmath38 does not even provide a limit .
if other forces dominate gravity , the number it gives for the central mass is an overestimate ; if other forces are comparable to gravity and partially cancel it , this number gives an underestimate .
even if in the future we are able to eliminate significant roles for forces other than gravity , there are several complications that stand between us and using broad line kinematics to measure the mass of the central black hole .
first of all , it is essential to have a complete picture of the kinematics , not just measures of the characteristic size and velocity .
as we have just finished discussing , one of the most important results of the reverberation mapping program has been to demonstrate that emission line material exists over a very broad range of radii .
simply inserting some sort of weighted - mean radius and ( differently ) weighted - mean velocity in the formula @xmath45 can be extremely misleading .
in fact , one way we might some day demonstrate that the motions are due to the gravitational force of a point - mass would be to use velocity - resolved reverberation maps to verify that @xmath46 ( intriguingly , done & krolik 1996 suggest that the scaling of @xmath47 with @xmath48 is a good deal slower than this ) .
furthermore , although @xmath38 does provide an order of magnitude estimator of the central mass in the event that the motions are purely orbital , it is no better than that without additional information about the shapes of the orbits .
when the orbits are exactly circular , the central mass is exactly @xmath45 ; but if the velocities at any given location are distributed randomly in direction , the central mass is three times greater . finally ,
if , as has long been hoped , we are to eventually use the broad emission lines as accretion diagnostics , we must find some way to relate the gas which emits them to the accretion flow .
if the broad emission line material is _ not _ moving inward , it can not itself be taking part in the accretion process , although it might perhaps be nearby .
learning how to draw this relation is unlikely to precede a clear understanding of the line - emitting matter s dynamics . on the other hand , if the broad line material is going in , we might speculate that it comprises the accretion flow , or is at least a part of it . to feed a typical seyfert galaxy with a luminosity @xmath49 erg s@xmath14 ,
an accretion rate of @xmath50 @xmath51 yr@xmath14 is needed , where @xmath52 is the accretion efficiency in rest mass units . if one interprets the reverberation mapping of ngc 5548 as showing net inflow , the associated mass inflow rate is @xmath53 m_{\odot}$ ] yr@xmath14 , where the ( not well - determined ) column density of the gas has been scaled in units of @xmath54 @xmath26 , the radius in units of 10 lt - d , the velocity in units of 1000 km s@xmath14 ( roughly the best - fit number for the net radial speed found by done & krolik 1996 ) , and @xmath55 is the covering factor of the gas around the central source of radiation .
thus , even if one accepts the net inward flow interpretation of the reverberation mapping , the line - emitting stuff may not be more than a small part of the total accretion .
if that is so , we are still faced with the problem of understanding its dynamics well enough to know how to use it as a tracer of the bulk of the accretion .
broad emission lines , therefore , despite ( or perhaps because of ) their phenomenological richness appear to be a prime example of the general principle enunciated in the previous section .
they are so complicated that we may never go beyond studying their _ intrinsic _ properties ; using them to answer fundamental questions about the global agn system may forever be beyond our reach .
finally we come to the innermost observable portion of any agn , the accretion disk . by far the strongest emission line that we can expect to be emitted from the accretion disk
is the fe k@xmath1 fluorescence line , but numerous other , weaker lines are also possible , mostly with energies @xmath43 kev ( e.g. as described in ycki et al .
improvements in both the energy resolution of x - ray spectrometers and their s / n might eventually allow us to study the weaker lines , but in the meanwhile , the fe k@xmath1 line , which has an equivalent width that is commonly a few hundred ev , is already very informative ( nandra 1997 ) .
the discovery , enabled by the asca satellite , that this line commonly has a fwhm between 20,000 and 70,000 km s@xmath14 is a dramatic signal that relativistic motions are taking place deep inside agn .
that the profile is frequently offset to the red also suggests a significant gravitational redshift , and the plausible supposition that those relativistic motions are simply orbital motions in a relativistically deep gravitational potential . moving from qualitative to quantitative results will take a great deal more effort , however .
several detailed calculations must be performed in order to adequately model the k@xmath1 emission in these circumstances , and we do not yet know how to carry out every step in these calculations .
first , the line is excited by fluorescence , so this entails a radiation transfer calculation of modest scope to determine how many k - edge photons reach how deep into the material . if the opacity of the material to k - edge photons were independent of ionization state , this would be a straightforward calculation . however , because a significant part of the opacity at energies above 7.1 kev ( the edge energy for neutral fe ) is due to other elements , one must also calculate the ionization structure self - consistently .
this , in turn , requires a model of the underlying disk structure .
unfortunately , we do not yet have a satisfactory description for the equilibrium state of accretion disks around black holes .
the standard shakura - sunyaev ( 1973 ) equilibrium is thermally unstable when the accretion rate is more than @xmath56 of the eddington rate ( shakura & sunyaev 1976 ) , and we do not know what other equilibrium replaces it .
nonetheless , if one assumes a model for the ( un - irradiated ) density and temperature in the disk , it is possible to then compute in a self - consistent manner its density , temperature , and ionization structure when irradiated by x - rays ( see , for example , ross & fabian 1993 and ycki et al .
it is important to know the ionization distribution of fe as a function of depth because the energy of the k@xmath1 photon depends on ionization state .
the dependence is weak until the atom is mostly stripped , but can be significant if highly - ionized stages are reached : the k@xmath1 line in h - like fe has @xmath57 greater energy than the line emitted by neutral fe , a difference comparable to the frequency shifts expected from relativistic effects .
after the spectrum of the initial photons is calculated , one must next find the angular distribution and spectrum with which they emerge from the disk surface .
note that the angular distribution with respect to the system axis may be different from the angular distribution with respect to the local surface normal if the disk surface is curved . some of the line photons may be absorbed by photoionizing lower - z atoms before escaping from the disk , others may change direction and energy as a result of electron scattering . with the quality of modern x - ray spectroscopy ,
the last effect is significant : compton recoil losses are about 1% per scatter ; thermal broadening is about @xmath58k@xmath59 per scatter .
if the abundance of fe ionization stages with l - shell vacancies is significant , resonance scattering can also occur . in the ionization stages with at least one l - shell electron ,
the k@xmath1 photon has a high probability of being lost as a result of resonance scattering because the excited atom is more likely to deexcite by auger ionization than by reradiation .
the net result of all these processes is moderately strong limb - darkening , but the degree of limb - darkening is a strong function of photon energy ( matt et al .
1996 ) .
thirdly , the relativistic effects must be evaluated ( as in matt et al .
1993 and laor 1991 ) .
there are several .
the photon energies and directions must be transformed from the rest frames of the disk s fluid elements to the observer s frame .
in addition , if we are considering regions not far outside the black hole s event horizon , the photon trajectories suffer substantial bending by the gravitational field .
all these effects clearly depend strongly on the initial directions of the photons , a fact which emphasizes the importance of a good calculation of their fluid rest - frame angular distribution .
finally , each of these previous steps is implicitly a local calculation .
each must be performed separately for different rings in the accretion disk , and then added with weights appropriate to the run of emissivity with radius . that distribution is , obviously , almost totally unconstrained _ a priori _ , so our procedure will usually be to guess some simple form , compute its predictions , and compare to the data . to date
, there is no model calculation in which all these elements have been combined .
perhaps the single most complete is that of matt et al .
( 1996 ) , in which the ionization balance as a function of depth , local angular distribution , and relativistic effects are all computed for two sample emissivity laws
. however , even in such a complex model there are still unresolved issues that can be expected to change the predicted profiles at the order unity level .
first , of course , is the underlying disk model
they chose the standard shakura - sunyaev solution , despite its instability .
second , the ionizing radiation transfer is treated in the diffusion approximation , despite the fact that few ionizing photons penetrate to thomson depths of more than @xmath43 . finally
, their relativistic calculation is limited to the schwarzschild geometry , whereas realistic black holes are likely to be much better described by the kerr metric .
the line profiles and angular distributions predicted by the two metrics are substantially different when there is much line emission from within a few gravitational radii ( laor 1991 ) .
thus , it is still a bit premature to expect quantitative conclusions from fits of measured line profiles to disk emission models .
another line of approach may offer additional insight : variability analysis in the spirit of reverberation mapping .
if the line emission comes from a region comparable in size to the continuum source ( as would be predicted by the disk models ) , one of the key assumptions of ordinary reverberation mapping that the line - continuum delay depends only on the position of the line - emitting material is no longer valid .
the continuum source distribution is also important because the light travel times from different parts of the continuum source to a particular part of the line source are significantly different .
nonetheless , it is still possible to work out relationships between light travel times and source sizes that are analogous to those used in conventional reverberation mapping . with adequate data , constraints may then be put on the source size and geometry .
the difficulties in carrying out this monitoring program have to do with the very short timescales that are relevant .
if the fe k@xmath1 photons are emitted from within 10 gravitational radii of the black hole , the light travel times across the system are only @xmath60 s. because x - ray variability power spectra are generally rather red " , the amplitudes of variations on such short timescales are quite small .
even the very large effective area detectors on _ xte _ may provide only marginal signal / noise on even the brightest sources .
the much smaller effective area of _ axaf _ at 6 kev will severely handicap its use for such studies .
moreover , the low earth orbit in which many x - ray telescopes are placed in order to avoid radiation backgrounds poses another problem : most sources are occulted by the earth for @xmath61 s out of every 5400 s. a window function of this sort makes observations of variability on 1000 s timescales extremely difficult .
only if a sufficiently bright source can be found near enough the pole of the orbit to be continuously in view does this sort of experiment become feasible .
from this quick survey of the achievements of agn emission line studies we may extract a list of how the different sub - areas have propelled us along the path toward a fundamental understanding of agn .
ironically , the greatest qualitative advances have come from two areas in which efforts to date have been comparatively small - scale : maser spot kinematics , and fe k@xmath1 spectroscopy .
study of the former has yielded by far the cleanest measurement of a central mass concentration in a galactic nucleus , as well as the tightest upper bound on its size ; work with the latter technique has produced the strongest evidence we have for the existence of relativistically deep potentials in agn .
we may hope for more quantitative results from fe k@xmath1 spectroscopy in the future , but they will require improvements in both data and models . a variety of other uses of emission lines
maps of the region radiating the narrow emission lines , spectropolarimetry revealing reflected emission lines have been of great aid in the construction of unification models for otherwise apparently dissimilar classes of agn . now that we realize
that agn can look very different depending on the observer s viewing angle , the presence of broad emission lines can be used as a crude indicator of our angle relative to the system axis . however , despite the enormous amount of work lavished on studies of broad emission line flux ratios , profiles , and variability , and the substantial advances in our knowledge of how and where these lines are produced that have come from these studies , we have not gained from them any significant information bearing on the most important questions of agn research .
it remains to be seen whether further efforts can break through the restraints imposed by the complexity of their emission , and reach our true goals of solving the problems posed in 1 of this review .
antonucci , r.r.j .
1993 , , 31 , 473 antonucci , r.r.j . and miller , j.s .
1985 , , 297 , 621 balsara , d.s . and
krolik , j.h .
1993 , , 402 , 109 band , d.l . ,
klein , r.i . , castor , j.i . , and nash , j.k . 1990 , , 362 , 90 bottorff , m. et al .
1997 in emission lines in active galaxies : new methods and techniques , ed .
peterson , f .- z .
cheng , and a.s .
wilson ( san francisco : astronomical society of the pacific ) , in press braatz , j.a . ,
wilson , a.s . , and menkel , c. 1994 , , 437 , l99 cecil , g. , bland , j. , and tully , r.b .
1990 , , 355 , 70 chiang , j. and murray , n. 1996 , , 466 , 704 clavel , j. et al .
1991 , , 366 , 64 done , c. and krolik , j.h .
1996 , 463 , 144 emmering , r.t . ,
blandford , r.d . , and shlosman , i. 1992 , , 385 , 460 ghisellini , g. , haardt , f. , and matt , g. 1994 , 267 , 743 greenhill , l. 1997 , in emission lines in active galaxies : new methods and techniques , ed . b.m .
peterson , f .- z .
cheng , and a.s .
wilson ( san francisco : astronomical society of the pacific ) , in press korista , k. et al .
1995 , , 97 , 285 krolik , j.h . and begelman , m.c .
1986 , , 308 , l55 krolik , j.h . ,
horne , k.d . ,
kallman , t.r . ,
malkan , m.a . , edelson , r.a . , and kriss , g.a . 1991 , , 371 , 541 krolik , j.h . and
kriss , g.a .
1995 , , 447 , 512 krolik , j.h .
, madau , p. , and .
zycki , p. 1994
, , 420 , l57 laor , a. 1991 , , 376 , 90 maoz , d. 1997 , in emission lines in active galaxies : new methods and techniques , ed .
peterson , f .- z .
cheng , and a.s .
wilson ( san francisco : astronomical society of the pacific ) , in press matt , g. , perola , g.c . , and stella , l. 1993 , , 267 , 643 matt , g. , fabian , a.c . , and ross , r.r .
1996 , , 278 , 1111 miller , j.s . ,
goodrich , r.w . , and mathews , w.g .
1991 , , 378 , 47 miyoshi , m. , moran , j. , herrnstein , j. , greenhill , l. , nakai , n. , diamond , p. , and inoue , m. 1995 , nature , 373 , 127 nandra , k. 1997 , in emission lines in active galaxies : new methods and techniques , ed .
peterson , f .- z .
cheng , and a.s .
wilson ( san francisco : astronomical society of the pacific ) , in press netzer , h. 1997 , in emission lines in active galaxies : new methods and techniques , ed .
peterson , f .- z .
cheng , and a.s .
wilson ( san francisco : astronomical society of the pacific ) , in press peterson , b.m . and wanders , i. 1996 , , 466 , 174 pier , e.a . and krolik , j.h .
1992 , , 399 , l23 pier , e.a . ,
antonucci , r.r.j . , hurt , t. , kriss , g.a . , and krolik , j.h .
1994 , , 428 , 124 rees , m.j .
1987 , , 228 , 47p shakura , n.i . and sunyaev , r.a .
1973 , , 24 , 337 shakura , n.i . and sunyaev , r.a . 1976 , , 175 , 613 wanders , i. , goad , m.r . ,
korista , k.t . ,
peterson , b.m . , horne , k. , ferland , g.j . ,
koratkar , a.p .
, pogge , r.w . , and shields , j.c .
1995 , , 453 , l87 weaver , k. , nousek , j. , yaqoob , t. , mushotzky , r.f . , makino , f. , and otani , c. 1996 , , 458 , 160 wilson , a.s .
1994 , in proceedings of the oxford torus workshop , m.j .
ward , ed .
( oxford university : oxford uk ) , p. 55
ycki , p. , krolik , j.h .
, zdziarski , a.a . , and kallman , t.r .
1994 , , 437 , 597 | a global view is given of how emission lines have been , and may in the future , be used to enhance our understanding of agn .
lines from the microwave to the x - ray bands all contribute .
although we have a deep understanding of the physical processes by which line photons are generated , when the circumstances of line emission are complicated , models become too unreliable to provide strong inferences about the rest of the agn system . at present , the lines which appear the most promising for helping answer the most important questions about agn are the 22 ghz h@xmath0o rotational transition and the 6.4 kev fe k@xmath1 fluorescence .
# 1 1.25 in .125 in .25 in = 8.5 in |
it is fair to say that tetrad calculations are generally considered superior to classical coordinate methods for the calculation of curvature in spacetime .
experiments by campbell and wainwright @xcite now dating back many years showed that tetrad methods are faster than coordinate methods by factors of @xmath0 .
even larger factors have been obtained by maccallum @xcite in a euclidean context . within the well known system sheep , for example
, the advice to the beginner is to always use frame versions of the metric ( e.g. maccallum and skea @xcite p. 23 ) . on the commercial side , within the system macsyma2 @xcite
the demonstration ctensor4 begins with an explanation that `` frame fields '' ( orthonormal bases ) allow the computations to run much more quickly .
the demonstration calculates the bases components of the ricci tensor for the kerr - newman spacetime in boyer - lindquist coordinates , and is a good place to begin our discussion .
+ in table 1 we have reproduced this demonstration within the system grtensorii @xcite@xcite running under maplev release 3 @xcite , and have included the calculation of the weyl tensor .
the table demonstrates some interesting properties .
the theoretical advantage of the frame approach is clearly demonstrated in the boyer - lindquist coordinates ( column bkn )
. however , under the elementary coordinate transformation @xmath1 this advantage fails to deliver superior performance ( column bknu ) .
the importance of strategic application of simplification at intermediate steps is illustrated in column bkns .
for this test simplification of components has been carried out only after the components of the ricci and weyl tensors are calculated .
it is worth noting that without some optimization in the simplification strategy ( e.g. post - calculation simplification only ) this calculation can not be executed in maplev on a 32 bit machine . [
cols="<,>,>,>",options="header " , ]
the following sections list tetrads and metrics used as inputs for the tests listed in tables 13 .
this list has been produced directly from the input files which were used in the tests ( listed in appendix [ app : a ] ) and converted to latex using maplev s latex output facility with only minor modifications to improve readability .
for this set of tests ( whose output is given in table 1 ) , the kerr - newman spacetime is described by a frame consisting of four independent covariant vector fields whose inner product is the constant matrix @xmath2 .
the basis vectors and corresponding line element for each case are given below .
+ ' '' '' + spacetime : kerr - newman ( boyer - lindquist coordinates ) ( bkn and bkns ) @xcite + ' '' '' + input file = frame + ' '' '' + bkn.mpl , @xmath3=@xmath4 $ ] + @xmath5@xmath6 $ ] + @xmath7@xmath8 $ ] + @xmath9@xmath10 $ ] + corresponding line element : + @xmath11 + @xmath12 + @xmath13 + @xmath14 + @xmath15 + ' '' '' + spacetime : kerr - newman ( boyer - lindquist coordinates , @xmath16 ) ( bknu ) @xcite + ' '' '' + input file = frame + ' '' '' + bknu.mpl @xmath17 @xmath18 $ ] + @xmath19 @xmath20 $ ] + @xmath21 @xmath22 $ ] + @xmath23 @xmath24 $ ] + corresponding line element : + @xmath25 + @xmath26 + @xmath27 + ' '' '' + this section lists the set of null tetrads for the test cases used to generate table 2 . for each spacetime , four forms of input were used .
the times listed in column @xmath28 of table 2 are obtained using a contravariant null tetrad as input . in columns @xmath29 and @xmath30 the metric is calculated from this tetrad and used for subsequent calculation . in column @xmath31 a covariant tetrad is loaded , and in @xmath32 and @xmath33 , its corresponding metric is used .
though the metric of column @xmath32 is , of course , equivalent to that of @xmath29 , there are often differences in representation which can in principle alter calculation times ( though in practice we have found this effect to be minimal ) .
thus the line elements used as inputs for each tests are listed along with the tetrad used to calculate them . ' '' ''
+ spacetime : griffiths ( * grif * ) @xcite + ' '' '' + input file = contravariant tetrad + ' '' '' + npupgrif.mpl @xmath34 $ ] + @xmath35 = @xmath36 $ ] + @xmath37 @xmath38 + @xmath39 \mbox{\hspace{74pt}}$ ] + @xmath40 @xmath41 + @xmath42 \mbox{\hspace{79pt}}$ ] + corresponding line element : + @xmath43 + @xmath44 + @xmath45 + ' '' '' + input file covariant tetrad + ' '' '' + npdngrif.mpl @xmath46 @xmath47 $ ] + @xmath48 @xmath49 $ ] + @xmath50 @xmath51 $ ] + @xmath52 @xmath53 $ ] + corresponding line element : + @xmath54 + @xmath55 + @xmath56 + ' '' '' + spacetime : lewis - papapetrou ( * l - p * ) @xcite + ' '' '' + input file contravariant tetrad + ' '' '' + npuplew.mpl @xmath57 @xmath58 $ ] + @xmath59 @xmath60 $ ] + @xmath61 @xmath62 $ ] + @xmath63 @xmath64 $ ] + corresponding line element : + @xmath65 + @xmath66 + ' '' '' + input file covariant tetrad + ' '' '' + npdnlew.mpl @xmath46 @xmath67 $ ] + @xmath68@xmath69 $ ] + @xmath70@xmath71 $ ] + @xmath52 @xmath72 $ ] + corresponding line element : + @xmath73 + @xmath66 + ' '' '' + spacetime : bondi ( * bondi1 * , * bondi2 * ) @xcite + ' '' '' + input file contravariant tetrad + ' '' '' + npupbon.mpl @xmath57@xmath74 $ ] + @xmath59@xmath75 $ ] + @xmath76@xmath77 $ ] + @xmath40@xmath78 $ ] + corresponding line element : + @xmath79 + @xmath80 + ' '' '' + input file covariant tetrad + ' '' '' + npdnbon.mpl @xmath81@xmath82 $ ] + @xmath48@xmath83 $ ] + @xmath50@xmath84 \mbox{\hspace{352pt}}$ ] + @xmath85@xmath86 \mbox{\hspace{106pt}}$ ] + corresponding line element : + @xmath87 + @xmath88 + ' '' '' + spacetime : debever ( * deb * ) @xcite + ' '' '' + input file contravariant tetrad + ' '' '' + npupdeb.mpl @xmath89 @xmath90 + @xmath91 $ ] + @xmath92@xmath93 + @xmath94 $ ] + @xmath95@xmath96 + @xmath97 $ ] + @xmath63@xmath98 + @xmath99 $ ] + corresponding line element : + @xmath100 + @xmath101 + @xmath102 + ' '' '' + input file covariant tetrad + ' '' '' + npdndeb.mpl @xmath81@xmath103 $ ] + @xmath68@xmath104 $ ] + @xmath50@xmath105 $ ] + @xmath106@xmath107 $ ] + corresponding line element : + @xmath100 + @xmath108 + @xmath109 + ' '' '' + spacetime : debever - mclenaghan - tariq ( * dmt1 * ) @xcite + ' '' '' + input file contravariant tetrad + ' '' '' + npupsdmt.mpl @xmath57@xmath110 + @xmath111 $ ] + @xmath112@xmath113 + @xmath114 $ ] + @xmath115@xmath116 + @xmath117 $ ] + @xmath118@xmath119 + @xmath120 $ ] + corresponding line element : + @xmath121 + @xmath122 + @xmath123 + ' '' '' + input file covariant tetrad + ' '' '' + npdnsdmt.mpl @xmath124@xmath125 $ ] + @xmath68@xmath126 $ ] + @xmath70@xmath127 $ ] + @xmath106@xmath128 $ ] + corresponding line element : + @xmath121 + @xmath122 + @xmath123 + ' '' '' + spacetime : debever - mclenaghan - tariq ( modified ) ( * dmt2 * ) @xcite + ' '' '' + input file contravariant tetrad + ' '' '' + npupdmt.mpl @xmath129@xmath130 + @xmath131 $ ] + @xmath59@xmath132 + @xmath133 $ ] + @xmath115@xmath134 + @xmath135 $ ] + @xmath118@xmath136 + @xmath137 $ ] + corresponding line element : + @xmath138 + @xmath139 + @xmath140 + @xmath141 + ' '' '' + input file covariant tetrad + ' '' '' + npdndmt.mpl @xmath81@xmath142 $ ] + @xmath48@xmath143 $ ] + @xmath70@xmath144 $ ] + @xmath106@xmath145 $ ] + corresponding line element : + @xmath138 + @xmath146 + @xmath147 + @xmath148 + ' '' '' + spacetime : kerr - newman ( @xmath149 ) ( * kn - euc1 * ) + ' '' '' + input file contravariant tetrad + ' '' '' + npupkn1.mpl @xmath150 @xmath151 + @xmath152 \mbox{\hspace{188pt } } $ ] + @xmath112@xmath153 + @xmath154 \mbox{\hspace{184pt } } $ ] + @xmath61@xmath155 $ ] + @xmath63@xmath156 $ ] + corresponding line element : + @xmath157 + @xmath158 + @xmath159 + ' '' '' + input file covariant tetrad + ' '' '' + npdnkn1.mpl @xmath160@xmath161 + @xmath162 \mbox{\hspace{178pt } } $ ] + @xmath48@xmath163 + @xmath164 \mbox{\hspace{187pt } } $ ] + @xmath165@xmath166 $ ] + @xmath52@xmath167 $ ] + corresponding line element : + @xmath157 + @xmath158 + @xmath168 + ' '' '' + spacetime : kerr - newman ( @xmath149 ) ( * kn - euc2 * ) + ' '' '' + input file contravariant tetrad + ' '' '' + npupkn2.mpl @xmath129@xmath169 + @xmath152 \mbox{\hspace{188pt } } $ ] + @xmath112@xmath169 + @xmath170 \mbox{\hspace{184pt } } $ ] + @xmath61@xmath171 $ ] + @xmath63@xmath172 $ ] + corresponding line element : + @xmath157 + @xmath158 + @xmath173 + ' '' '' + input file covariant tetrad + ' '' '' + npdnkn2.mpl @xmath160@xmath163 + @xmath174 \mbox{\hspace{178pt } } $ ] + @xmath175@xmath163 + @xmath176 \mbox{\hspace{187pt } } $ ] + @xmath177@xmath178 $ ] + @xmath52@xmath179 $ ] + corresponding line element : + @xmath157 + @xmath158 + @xmath173 + ' '' '' + spacetime : kerr - newman ( boyer - lindquist coordinates ) ( * kn - bl1 * ) @xcite + ' '' '' + input file contravariant tetrad + ' '' '' + npupkn3.mpl @xmath180@xmath181 $ ] + @xmath59@xmath182 $ ] + @xmath115@xmath183 + @xmath184 \mbox{\hspace{180pt } } $ ] + @xmath185 + @xmath186 \mbox{\hspace{196pt}}$ ] + corresponding line element : + @xmath187 + @xmath188 + @xmath189 + @xmath190 + ' '' '' + input file covariant tetrad + ' '' '' + npdnkn3.mpl @xmath81@xmath191 $ ] + @xmath192@xmath193 $ ] + @xmath194 @xmath195 + @xmath196 \mbox{\hspace{84pt}}$ ] + @xmath52@xmath197 + @xmath198 \mbox{\hspace{104pt } } $ ] + corresponding line element : + @xmath199 + @xmath200 + @xmath201 + @xmath190 + ' '' '' + spacetime : kerr - newman ( eddington - finklestein coordinates ) ( * kn - ef1 * ) @xcite + ' '' '' + input file contravariant tetrad + ' '' '' + npupkn4.mpl @xmath150 @xmath202 $ ] + @xmath203 @xmath204 $ ] + @xmath76 @xmath205 $ ] + @xmath206@xmath207 $ ] + corresponding line element : + @xmath208 + @xmath209 + @xmath210 + @xmath211 + @xmath212 + ' '' '' + input file covariant tetrad + ' '' '' + npdnkn4.mpl @xmath213 @xmath214 $ ] + @xmath215 @xmath216 + @xmath217 $ ] + @xmath177 @xmath218 $ ] + @xmath85 @xmath219 $ ] + corresponding line element : + @xmath208 + @xmath209 + @xmath220 + @xmath211 + @xmath221 + ' '' '' + spacetime : kerr - newman ( boyer - lindquist coordinates , @xmath16 ) ( * kn - bl2 * ) @xcite + ' '' '' + input file contravariant tetrad + ' '' '' + npupkn5.mpl @xmath57@xmath222 $ ] + @xmath112 @xmath223 $ ] + @xmath76 @xmath224 + @xmath225 \mbox{\hspace{195pt } } $ ] + @xmath206 @xmath226 + @xmath227 \mbox{\hspace{212pt } } $ ] + corresponding line element : + @xmath228 + @xmath229 + @xmath230 + @xmath231 + ' '' '' + input file covariant tetrad + ' '' '' + npdnkn5.mpl @xmath124@xmath232 $ ] + @xmath215 @xmath233 $ ] + @xmath234 @xmath235 + @xmath236 \mbox{\hspace{22pt } } $ ] + @xmath85 @xmath237 + @xmath238 \mbox{\hspace{67pt } } $ ] + corresponding line element : + @xmath239 + @xmath229 + @xmath230 + @xmath240 + ' '' '' + spacetime : kerr - newman ( eddington - finklestein coordinates ) ( * kn - ef2 * ) @xcite + ' '' '' + input file contravariant tetrad + ' '' '' + npupkn6.mpl @xmath150 @xmath202 $ ] + @xmath203 @xmath241 $ ] + @xmath76 @xmath242 $ ] + @xmath206 @xmath243 $ ] + corresponding line element : + @xmath208 + @xmath244 + @xmath210 + @xmath211 + @xmath212 + ' '' '' + input file covariant tetrad + ' '' '' + npdnkn6.mpl @xmath160 @xmath245 $ ] + @xmath215 @xmath246 + @xmath217 $ ] + @xmath177 @xmath247 $ ] + @xmath85@xmath248 $ ] + corresponding line element : + @xmath208 + @xmath249 + @xmath210 + @xmath211 + @xmath212 + ' '' '' + for this set of tests , not only are the basis vectors varied , but also their inner product .
the input used for these tests is listed below . ' '' ''
+ spacetime : mixmaster ( * mix * ) @xcite + ' '' '' + input file = frame + ' '' '' + mix.mpl inner product of basis vectors : + @xmath250 + basis vectors : + @xmath251 = @xmath252 $ ] + @xmath253 $ ] + @xmath254@xmath255 $ ] + @xmath256 $ ] + corresponding line element : + @xmath257 + @xmath258 + @xmath259 + @xmath260 + ' '' '' + spacetime : mixmaster ( * mix1 * ) @xcite + ' '' '' + input file frame + ' '' '' + mix1.mpl inner product of basis vectors : + @xmath261 + basis vectors : + @xmath262@xmath263 $ ] + @xmath264@xmath265 $ ] + @xmath266@xmath267 $ ] + @xmath268@xmath269 $ ] + corresponding line element : + @xmath270 + @xmath271 + @xmath272 + @xmath273 + @xmath274 + ' '' '' + spacetime : mixmaster ( * mix2 * ) @xcite + ' '' '' + input file = frame + ' '' '' + mix2.mpl inner product of basis vectors : + @xmath275 + basis vectors : + @xmath276 $ ] + @xmath264@xmath277 $ ] + @xmath278@xmath279 $ ] + @xmath280@xmath281 $ ] + corresponding line element : + @xmath282 + @xmath258 + @xmath283 + @xmath260 + ' '' '' + spacetime : mixmaster ( * mix3 * ) @xcite + ' '' '' + input file = frame + ' '' '' + mix2.mpl inner product of basis vectors : + @xmath275 + basis vectors : + @xmath262@xmath284 $ ] + @xmath285@xmath286 $ ] + @xmath287 $ ] + @xmath288 $ ] + corresponding line element : + @xmath289 + @xmath290 + @xmath272 + @xmath273 + @xmath274 + ' '' '' +
the following is the input file : .... ndim _ : = 4 : x1 _ : = t : x2 _ : = r : x3 _ : = u : x4 _ : = phi : eta12 _ : = 1 : eta34 _ : = -1 : b11 _ : = ( r^2+a^2)/(r^2 - 2*m*r+a^2+q^2 ) : b12 _ : = 1 : b14 _ : = a/(r^2 - 2*m*r+a^2+q^2 ) : b21 _ : = 1/2*(r^2+a^2)/(r^2+u^2 ) : b22 _ : = -1/2*(r^2 - 2*m*r+a^2+q^2)/(r^2+u^2 ) : b24 _ : = 1/2*a/(r^2+u^2 ) : b31 _ : = 1/2*(i*(a^2-u^2)^(1/2)*r+(a^2-u^2)^(1/2)*u)*2^(1/2)/(r^2+u^2 ) : b33 _ : = 1/2*(-(a^2-u^2)^(1/2))*(r - i*u)*2^(1/2)/(r^2+u^2 ) : b34 _ : = 1/2*i*(r - i*u)*a/(a^2-u^2)^(1/2)*2^(1/2)/(r^2+u^2 ) : b41 _ : = 1/2*(-i*(a^2-u^2)^(1/2)*r+(a^2-u^2)^(1/2)*u)*2^(1/2)/(r^2+u^2 ) : b43 _ : = 1/2*(-(a^2-u^2)^(1/2))*(r+i*u)*2^(1/2)/(r^2+u^2 ) : b44 _ : = -1/2*i*a*(r+i*u)/(a^2-u^2)^(1/2)*2^(1/2)/(r^2+u^2 ) : info_:=`contravariant nptetrad for kerr - newman metric ( u = a*cos(theta ) to boyer - lindquist coordinates ) ` : .... unfortunately , the output obtained from each of these methods are not in exactly the same form . since p2 is smaller as measured in maple _ words _
, this must be considered the fully simplified form .
in fact , by performing the maplev operation factor on the denominators of @xmath335 ( an operation requiring less than @xmath336 cpu seconds ) , it can be reduced to the required fully simplified form . in all tests listed in appendix [ app : a ]
the final results of calculation are either presented in exactly equivalent forms , or equivalent within some simplification operation requiring a negligible amount of time . | we examine the relative performance of algorithms for the calculation of curvature in spacetime .
the classical coordinate component method is compared to two distinct versions of the newman - penrose tetrad approach for a variety of spacetimes , and distinct coordinates and tetrads for a given spacetime . within the system grtensorii
, we find that there is no single preferred approach on the basis of speed .
rather , we find that the fastest algorithm is the one that minimizes the amount of time spent on simplification .
this means that arguments concerning the theoretical superiority of an algorithm need not translate into superior performance when applied to a specific spacetime calculation . in all cases
it is the global simplification strategy which is of paramount importance .
an appropriate simplification strategy can change an untractable problem into one which can be solved essentially instantaneously . |
a gamma - ray burst ( grb ) is the most luminous event in the universe . in a few seconds it produces an enormous amount of electromagnetic energy , which is comparable to the integrated emission of the sun over ten billion years ( meszaros 2006 ) .
most of this energy is released in the first few seconds in the form of gamma rays ( kev to mev ) , known as the prompt emission phase . in some grbs
very high energy ( gev ) emission is also observed during the prompt phase .
later the afterglow of a grb can be observed at all electromagnetic frequencies from the radio wavelengths up to x - rays and even gamma rays .
numerous satellite experiments have been performed to pin down the mechanism of grbs . for ten long years ( 1991 - 2000 )
, a great wealth of prompt emission data was collected by the burst and transient source experiment ( batse ; fishman et al .
1994 ) onboard the compton gamma - ray observatory ( _ cgro _ ) . with
( barthelmy et al .
2005 ; gehrels et al . 2004 ) and _ fermi _ ( meegan et al . 2009 ) , launched respectively in 2004 and 2008 , we have entered into the modern era of grb research . _ swift _ , with its fantastic slewing capability , an unprecedented localization accuracy of the primary grb detector ( burst alert telescope bat , a large field of view gamma ray monitor ) and the high spatial resolution instruments for the afterglow studies has enabled the measurement of the redshift for many grbs .
_ fermi _ , on the other hand , has enabled a detailed spectral study during the prompt emission phase , and discovered the very high energy ( gev ) emission both during the prompt and afterglow phase for many bursts
( very few such events were known before _
fermi_. see e.g. , gonzlez et al .
both these satellites have supplied extremely valuable data for the identification of the working mechanism of grbs .
but , even after nearly fifty years of discovery , various issues remain puzzling , most notably the prompt emission mechanism . in the preliminary model of a grb , known as the `` standard fireball model '' ( goodman 1986 ; paczynski 1986 ) , the radiation is supposed to come from a photosphere .
the instantaneous spectrum is predicted to be a blackbody ( bb ) and the light curve ( lc ) is a simple pulse .
though we do see a grb having a single pulse , there is a variety of lc profiles .
there are also differences in the predicted and observed spectrum . while the bb spectral shape in the photon space is a hump with a narrow peak , which can be approximated with a low energy power - law with a + 1 index and a steep power - law with a negative index at high energy , grb prompt emission spectra usually have non - thermal spectral shape characterized with negative power - law indices ( @xmath4 ) .
though there is a provision for geometric broadening ( goodman 1986 ) , the spectrum is still far from what is observed . to overcome this difficulty
, it is assumed that the radiation is not coming from the photosphere , but it is produced in the internal shocks ( is ) of the ejecta via synchrotron radiation ( rees & meszaros 1994 ; woods & loeb 1995 ; sari & piran 1997 ; kobayashi et al .
this model is phenomenologically represented by the featureless band function ( band et al .
1993 ) , which is a smoothly joined power - laws with two indices @xmath3 as the lower energy index and @xmath5 as the higher energy index . in the ef(e )
representation , the function peaks at an energy ( @xmath6 ) , provided that @xmath7 , which can be violated for very hard spectra . the prompt emission data , whether time - integrated or time - resolved , is adequately fit by the band function ( kaneko et al .
2006 ; nava et al . 2011
; zhang et al . 2011 ) .
however , the band function is a phenomenological function and the actual spectrum may have rich structure . in many studies , additional components to the band function ( preece et al . 1996 ; meszaros & rees 2000 ; gonzlez et al .
2003 ; shirasaki et al .
2008 ; abdo et al .
2009 , ackermann et al . 2010 ; ackermann et al . 2011
; guiriec et al . 2011 ; axelsson et al . 2012 ; guiriec et al .
2013 ) and even alternative models ( ghirlanda et al . 2003
; ryde 2004 ; ryde & peer 2009 ; peer & ryde 2011 ) have been suggested .
the reason for resorting to alternative description is that the internal shock model demands that in the electron slow cooling regime , the photon index ( @xmath3 ) should be less than -2/3 ( preece et al .
1998 ) , the single particle synchrotron limit .
grbs are usually expected to be in the fast cooling regime where the photon index should be near -3/2 ( cohen et al .
these two limits are referred to as the `` synchrotron lines of death '' .
the value of @xmath3 is often found to be higher than the -2/3 ( crider et al . 1999 ) .
moreover , the is model has the following issues .
the efficiency of converting the kinetic energy to the internal energy and then radiating is rather low ( less than 20% ; piran 1999 ; but , see also nemmen et al .
the situation of an inner shell moving faster than an outer shell , which is a necessary condition for the is to produce , is unstable ( waxman & piran 1994 ) . these difficulties of the is model have instigated the search for alternate models .
some of them involve looking back into the prediction of the original fireball model .
for example , ryde ( 2004 ) has shown for a few batse grbs , having single pulse , that the instantaneous spectrum is consistent with either a bb or a bb with a power - law ( bbpl ) spectrum ( see also ryde & peer 2009 ) .
the reasons to choose single pulse are two - fold ( a ) the original fireball model predicts single pulse and instantaneous bb radiation ; ( b ) if the spectral evolution is a pulse property rather than a burst property , a single pulse is an ideal system to study the spectral evolution and then one can apply the knowledge of a single pulse to a more complex grb .
the bbpl model has sometimes shown superiority over the band function in the batse data .
this is a very promising result because , this is the first step towards identifying the physical mechanism of the radiation .
however , later it was found that more complex models may be required to fit a wider spectrum provided by fermi ( ryde et al .
2010 ; burgess et al . 2011 ; guiriec et al .
2013 ; basak & rao 2013 br13 , hereafter ) .
it is shown that a model consisting of a band and a blackbody model ( band+bb ) is required to fit some of the long ( guiriec et al .
2011 ; axelsson et al .
2012 ) and short ( guiriec et al . 2013 ) grbs .
interestingly , in a band+bb fit , the parameters of the band function are compatible with the synchrotron model prediction , while they are not for a fit with a band only function . band+bb
scenario with a subdominant bb is not compatible with the genuine fireball model and requires a highly magnetized outflow close to the source and a low magnetization at large radii to explain the observations .
in addition , the magnetization seems to vary from burst to burst ( guiriec et al . 2013 ) . in this context , it is important to examine the fermi grbs having single pulses employing various models .
( 2012 ; lu12 hereafter ) have analyzed a set of fermi grbs .
they have paid special attention to the single pulses .
their motivation , however , was to find the variation of @xmath6 of band function . in this study
, we examine various models and their parameter evolution for grbs having single pulses .
the organization of this paper is as follows . in section 2
, we describe our sample , followed by data analysis techniques . in section 3
, we show our results .
the major findings are summarized and discussed in section 4 .
[ cols="^,^,^,^,^,^,^,^,^,^,^ " , ] [ t2 ] @xmath8 deviation of the mean value of @xmath3 from the -3/2 `` line of death '' ( fast cooling ) in the units of @xmath9 is shown .
@xmath10 is the reduced @xmath11 of the fit of @xmath3 values assuming the model @xmath12 .
higher @xmath10 denotes higher deviation from the synchrotron model predicted photon index ( for fast cooling ) .
+ @xmath13 band function compared to the bbpl model .
f - test is performed to check the superiority of band function over the bbpl model for all the time - resolved spectrum of each grb .
the reported values are the mean ( and standard deviation ) of the confidence level of f - test .
we fit the lcs of all the grbs using equation 1 .
these are shown in figure [ lc ] .
we note that norris model adequately captures the pulse profile of all the grbs .
however , note that the @xmath10 are generally not good .
this is due to the fact that the pulses , other than the broad structure , have very rapid variations ( see rao et al .
these finer time variability can not be captured in a simple pulse structure . in the current study , we are interested to quantify broad pulse properties e.g. , width and asymmetry of the pulses .
hence , the use of norris model is adequate for our purpose . in table
[ t1 ] , we report the parameters of norris model fit .
the errors in the derived parameters are calculated by propagating the mean error of the norris model parameters ( see norris et al .
uncertainty in p , however , is obtained by noting down the @xmath14 for different values of p because the correlated errors in @xmath15 and @xmath16 gives incorrect results , particularly for symmetric grbs . the width
( w ) and asymmetry ( @xmath17 ) , however , are well determined .
though these grbs have broad single pulses , we find that for grb 081207 , which is a very long grb compared to all others , at least two heavily overlapping pulses are present during the main bursting episode .
also , for grb 110721a , we found that fitting two pulses is significantly better than a single pulse fitting ( an improvement of @xmath18 is obtained to the expense of 4 degrees of freedom ) .
lu12 have suggested that the hts pulses are more asymmetric than the it pulses .
but , from table [ t1 ] we find that there are some exceptions to this consensus for our single pulse sample .
for example , grb 100612a , which is a hts grb , is very symmetric ( @xmath17 is only @xmath19 ) .
grb 090922a , despite being an it grb , is very asymmetric ( as high as @xmath20 ) .
the highest asymmetry is shown by single pulse grb 100707a .
we visually inspect this grb and find that the rising part is in fact very steep .
hence , in order to get more data point in the rising part , we analyze the lc using 0.1 s bin size .
we find the asymmetry , @xmath21 , which is indeed the highest among all the grbs .
though we found two pulses in grb 110721a , one of these pulses shows even higher asymmetry @xmath22 .
one of the two pulses of grb 081207 , an it grb , show high asymmetry ( @xmath23 ) .
hence , we can not differentiate between hts and it grbs according to their asymmetric lcs . as described in section 2.2 ,
we extract the time - resolved spectra of the individual grbs by requiring a minimum background subtracted count per time bin . the time interval ( @xmath24 to @xmath25 ) and the number of bins ( n ) are listed in table [ t1 ] . for grb 081207 , which has another pulse after 80 s ,
the data is taken till 76 s. this selection , however , will not affect any conclusion regarding this grb .
we also report the detectors used in our analysis in the last column of table [ t1 ] .
the results of the time - resolved spectroscopy for all the grbs are shown graphically in figure [ grb1 ] through figure [ grb9 ] .
the upper panels show the kt variation of the bbpl and @xmath6 variation of the band function . we have shown the lcs of the individual grbs in this panel to compare the @xmath6 variation with the count rate variation . from this variation
we can identify that the first 6 grbs are the hts grbs and the rest 3 are it grbs .
note that the peak of bb occurs at @xmath263kt , and hence , in this panel , we have plotted 3kt instead of kt to compare the peak position given by the bbpl and band function fit .
we note that there is a general agreement of evolution of these peaks , however , the peak of the bbpl fit is always found to be lower than the corresponding @xmath2 peak of the band function . in the second panels from the top
, we have shown the variation of @xmath3 of the band function and @xmath27 of the bbpl model .
the parameter @xmath5 is often unconstrained or reaches the hard bound of -10 .
hence , we do not show this parameter . in this panel
, we show the `` synchrotron lines of death '' by two straight lines one at -2/3 ( slow cooling regime ) and another at -3/2 ( fast cooling regime ) .
the general trend of @xmath3 is higher at the beginning and lower in the later part .
note that the value @xmath3 is always greater than the line at fast cooling regime ( -3/2 ) . to quantify the deviation of @xmath3 from the predicted -3/2 line , we device two methods : ( a ) the deviation of the mean value of @xmath3 from -3/2 in the units of @xmath9 and ( b ) the @xmath11 value of the fit to @xmath3 , assuming the model predicted @xmath3=-3/2 . in table [ t2 ] ( column 4 ) , we have shown the mean values of @xmath3 for each grb .
the error in the mean is 1@xmath9 , and it is calculated by using the two tailed errors of @xmath3 ( at nominal 90% confidence ) .
the deviation of the mean from -3/2 value in the units of @xmath9 is shown in column 5 .
we see a high deviation in each case .
the @xmath14 ( shown in parenthesis ) is also high in each case denoting the significance of the trend . in some cases , it s value is greater than -2/3 , which is disallowed by synchrotron model for slow cooling electrons .
this deviation from -2/3 line , however , is not very significant .
for some hts grbs the deviations are significant : 40@xmath9 ( grb 100707a ) , 5.8@xmath9 ( grb 081224 ) and 5.0@xmath9 ( grb 110817a ) . for it grbs
the deviations are consistent with zero ( within 1@xmath9 ) . for it grb 100528a , the mean value of @xmath3 is @xmath28 which is even within the slow cooling `` line of death '' at 8.8@xmath9 .
we find two extreme cases .
these are grb 110817a ( hts ) and grb 100528a ( it ) . in the former case ,
@xmath3 values are always above the line at -2/3 and in the latter case @xmath3 values are always below the line . in the third panels
, we have shown the evolution of the photon flux for the two components of bbpl model .
note that , the pl flux lags behind the bb flux .
this behaviour is particularly significant for grb 081224 , grb 100707a and grb 110721a . in a recent paper , basak & rao ( 2013b )
have shown that very high energy ( gev ) emission is expected for grbs which have delayed pl emission . for the three grbs in the present study , we also find reported lat detection ( lle data ) at 3.1@xmath9 , 3.7@xmath9 and 30.0@xmath9 , respectively ( ackermann et al .
though we have not used the lat data of these grbs for spectroscopy , the lat detection is consistent with our earlier claim .
finally , the fourth panels show the @xmath14 of various model fits . to compare the fit statistics of bbpl model with other models
, we perform f - test between bbpl and band function . the confidence level ( cl ) that the alternative model ( band ) is better than the original hypothesis ( bbpl ) is calculated for all the time - resolved bins of the individual grbs .
we then compute the mean and standard deviation of the cl .
the corresponding values are reported in table [ t2 ] .
it is evident that the it pulses are adequately fit by bbpl model .
the hts pulses , on the other hand , require a different model to fit the spectrum . however , band , mbbpl and 2bbpl are all equally good for this purpose .
hence , one needs physical arguments to distinguish between these models .
the results of all the grbs are summarized in table [ t2 ] .
we note that the grbs can be classified in 2 categories : hts and it .
we note the following trends for hts and it grbs .
first , note that a lat detection is reported only for hts class . for hts
grbs , the value of @xmath3 is greater than -2/3 in 63.4% cases .
the exception of this trend is shown by grb 110721a . excluding this grb
, we obtain 74.6% of cases of hts grbs where @xmath3 is greater than -2/3 .
the deviation of @xmath3 from -3/2 line are systematically high ( large @xmath9 deviation and high values of @xmath14 ) . the confidence level ( cl ) that the band model is better than the bbpl model has a mean value of 79.3% . in case of it
grbs , the value of @xmath3 is greater than -2/3 in only 44% cases . as pointed out the mean values
are greater only within 1@xmath9 deviation , and for one it grb ( namely , 100528a ) it is even less than -2/3 at 8.8@xmath9 . also , the cl that the band model is better than bbpl model has a mean value of only 67% . the only exception of the behaviour of @xmath3 is noted for the hts grb 090809b for which @xmath3 is greater than -2/3 in only 53.3% cases . however , note in figure [ grb2 ] ( upper panel ) that the @xmath29 evolution is not strictly hard - to - soft . in figure [ alpha_hts_it ] (
left panel ) , we have plotted the @xmath3 values for hts and it grbs , with y - axis as time sequence . in the same plot ,
we have indicated the -3/2 and -2/3 lines of death .
note that @xmath3 values are always greater than the -3/2 line .
in many cases the @xmath3 values are greater than the -2/3 line , especially for the hts grbs . the mean value of @xmath3 is -0.42 and -0.68 for hts and it grbs with dispersion of 0.72 and 0.50 , respectively .
also there is a clear trend of the evolution of @xmath3 the values of @xmath3 decreases with time i.e. , the spectrum becomes softer .
the four models used in our analysis are taken from br13 , who applied them to the brightest grbs having separable pulses ( grb 081221 and grb 090618 ) .
the present results generally agree with the results obtained for those two bright grbs . in the present case
too we find that the decomposition of the spectrum in terms of 2bbs or mbb and a power - law components significantly improves the fit over single bb with a pl .
though the 2bbpl model gives better @xmath10 in all cases , the present analysis can not conclusively show whether the 2bbpl model is an approximation for some underlying complex spectral distribution . exploring this model further , basak & rao ( 2013b ) ,
have shown some very promising results like prediction of lat emission from gbm only analysis .
they have fitted the mev gbm data , without invoking the lat data , with the 2bbpl model and have shown that the pl flux , independent of the lat data , mimics the lat gev emission .
the pl has a delayed onset and lingers at the later part , just as the same way as the lat flux .
moreover , the pl fluence correlates with the lat fluence . in the present analysis
, we find lat detection for three grbs ( grb 081224 , grb 100707a and grb 110721a ) which show similar behaviour ( a delayed onset of the power - law component which lingers at the later phase ) .
another interesting feature we find that the 2bbs of the 2bbpl model are highly correlated . in figure [ kt_corr ] , we have plotted the temperatures of the two bbs of the 2bbpl model .
it is clear that the two temperatures are highly correlated .
this correlation shows that either there is some underlying physical reason for the two black body components or they are connected to each other because they happen to be an approximation for some more complex spectral shape .
to summarize , we have collected a sample of gbm grbs having single pulse .
we have studied the individual pulses in the time and energy domain and found some interesting results .
we found that the band model always gives better fit to the data compared to the bbpl model .
the @xmath10 of the band model is comparable to mbbpl and 2bbpl model .
however , the values of @xmath3 , in many cases , are greater than -2/3 , the limit due to the synchrotron emission from electrons in the slow cooling regime .
hence , the spectrum can not be fully synchrotron . hence
, neither bbpl nor band can be considered consistent with the underlying physical mechanism of the prompt emission .
it is possible that both thermal and non - thermal emissions contribute to the total emission . in this regard , the 2bbpl and the mbbpl model are preferred over bbpl model .
we found that the hts grbs have generally larger values of @xmath3 compared to the it grbs .
also , the @xmath3 has a decreasing trend with time , making the spectrum softer .
this signifies that the spectrum has a different origin than synchrotron at the beginning , but later synchrotron emission may dominate .
we found that the peak energy of the bbpl model gives lower value than the corresponding band function .
as the peak energy of the spectrum is used for the correlation studies in the prompt emission ( amati et al .
2002 ; yonetoku et al .
2004 ; ghirlanda et al . 2004 ; basak & rao et al .
2012a ) , it is essential to determine the correct peak energy of the spectrum .
if the underlying reason for the @xmath6 and isotropic equivalent energy correlation ( @xmath6-@xmath30 : the amati correlation ) is some basic physical process , a correct model description and a proper identification of @xmath6 would be required to improve the amati - type correlation ( note that @xmath6 and 3 kt are note strictly correlated with each other and the difference varies between grbs ) . investigating a pulse
- wise amati correlation ( e.g. , see krimm et al .
2009 ; basak & rao 2013c ) with different models to derive the @xmath6 can provide inputs to strengthen the amati correlation as well as to identify the physical process responsible for the correlation .
note that a hts evolution , by its nature violate time - resolved @xmath6-@xmath30 correlation , because the high values of @xmath6 is obtained even when the flux ( and hence @xmath30 ) is low .
for example , axelsson et al . (
2012 ) have found @xmath31 mev during the initial bins of grb 110721a .
hence , time - resolved amati correlation is likely to fail .
basak & rao ( 2012b ) have studied amati correlation by replacing time - resolved @xmath6 and @xmath30 by pulse average @xmath6 and total @xmath30 of a pulse .
a pulse average @xmath6 , due to the crucial averaging over a pulse , will not show the effect of hts evolution history , thus restoring the amati correlation .
note that , pulses , rather than time - resolved bins are preferred as the pulses behave as independent entities in grbs . in this context
one reasonable choice is peak energy at the zero fluence ( @xmath32 ) as done by basak & rao ( 2012b ) . unlike @xmath6 , which is a pulse average quantity ,
@xmath32 is a constant for each pulse and correlates better with the total @xmath30 .
consequently , the strong pulse - wise @xmath32-@xmath30 correlation indicates that a pulse with high initial peak energy may have started with a low flux , but eventually it will give rise to high total flux of the pulse .
however , one can not use the it class for such correlation study , as it grbs , by their nature will not give such @xmath32 .
hence , one can use pulse average peak energy for such correlation study , which does not depend on hts or it evolution ( basak & rao 2013c ) finally , we want to remind that the 2bbpl model has some interesting features like highly correlated bb temperatures .
the pl component of this model fitted to the prompt kev - mev emission has been shown to correlate with the high energy gev emission ( basak & rao 2013b ) .
these definitely point towards some underlying physical process .
hence , it is important to find the existence of the 2bbs in the prompt emission sepctrum .
making similar spectral analysis for grbs with better low energy measurement ( like that from _ swift_/x - ray telescope ( xrt ) ) would clarify these further .
as the temperature of the two correlated bbs decrease with time , the lower bb may show up in the low energy detector like xrt during the late prompt emission phase
. however , such detailed study needs good knowledge of detector systematics and cross - detector calibration . in basak & rao ( 2012b )
, we have shown that spectral analysis using bat and gbm gives systematics difference in the results .
though one needs to find out ways to perform joint fitting , the procedure is quite involved and beyond the scope of the present paper .
this research has made use of data obtained through the heasarc online service , provided by the nasa / gsfc , in support of nasa high energy astrophysics programs .
we thank the referee for valuable comments and numerous suggestions especially to make quantitative statements of the results . | we analyze gamma - ray bursts ( grbs ) detected by _ fermi_/gamma - ray burst monitor ( gbm ) and having single pulse .
we fit the light curves with a model having exponential rise and decay parts .
we perform a detailed time - resolved spectroscopy using four models : band , blackbody with a power - law ( bbpl ) , multicolour blackbody with a power - law ( mbbpl ) and two blackbodies with a power - law ( 2bbpl ) .
we find that models other than the bbpl give better @xmath0 for the `` hard - to - soft '' ( hts ) pulses , while for the `` intensity tracking '' ( it ) pulses , the bbpl model is statistically as good as the other models .
interestingly , the energy at the peak of the spectrum resulting from the bbpl model ( @xmath1 ) , is always lower than that of the @xmath2 spectrum of the band function .
the values of the low energy photon index ( @xmath3 ) of the band function are often higher than the fundamental single particle synchrotron limit , especially for the hts pulses .
specifically we find two extreme cases for grb 110817a ( hts grb ) @xmath3 is always higher , while for grb 100528a ( it grb ) @xmath3 is always within the synchrotron regime .
the pl component of the bbpl model always starts with a delay compared to the bb component , and it lingers at the later part of the prompt emission phase . for three hts grbs , namely , grb 081224 , grb 100707a and grb 110721a this behaviour is particularly significant and interestingly there are reported lat detections for them . finally , we argue that various evidences hint that neither bbpl nor band model is acceptable , while 2bbpl and mbbpl are the most acceptable models for the set of grbs we have analyzed .
[ firstpage ] radiation mechanisms : non - thermal radiation mechanisms : thermal methods : data analysis methods : observational methods : statistical gamma - ray burst : general . |
charmonium production has long been considered as a clean probe for the quantum chromodynamics ( qcd ) matter formed in heavy ion collisions at relativistic energy @xcite , due to the color screening of heavy quark potential at finite temperature @xcite . when the temperature of the medium becomes higher than the charmonium dissociation temperature @xmath4 , the initially produced charmonia will be destroyed .
it is called the anomalous suppression @xcite .
therefore the comparison of the final yield of charmonia from heavy ion collisions to that from the corresponding nucleon - nucleon collisions can be used to extract the medium properties in high - energy nuclear collisions @xcite .
this anomalous suppression induced by the hot nuclear matter effect explains well the experimentally observed charmonium suppression at the super proton synchrotron ( sps ) @xcite . however , color screening is not the only hot medium effect in high - energy nuclear collisions . once the energy is sufficiently high , such as the collisions at the relativistic heavy ion collider ( rhic ) and the large hadron collider ( lhc ) ,
many charm quarks are produced @xcite .
the initially uncorrelated charm quarks @xmath5 and @xmath6 can recombine into a charmonium . this process is called regeneration @xcite . in high - energy nuclear collisions , these two effects are co - existing and their relative strength is energy dependent . at rhic
it turns out that both suppression and regeneration processes are important for understanding the charmonium production @xcite . even before the formation of quark - gluon plasma ( qgp ) , the cold nuclear matter effects affect the charmonium production . usually three effects are considered in the literatures : ( 1 ) the change in the parton distribution function in nucleus which controls the initial parton behavior and strongly depends on the collision kinematics . in the small @xmath7 region , the nuclear parton distribution function
is clearly suppressed compared to that of nucleon .
that is called the shadowing effect @xcite ; ( 2 ) cronin effect @xcite which describes the initial gluon multi - scattering with the nucleons prior to any hard scattering and the quarkonium formation .
the nature of such collisions can be described as brownian motion . as a result the transverse momentum distribution of the produced charmonia is broadened ; and ( 3 ) nuclear absorption via interaction between charmonia and the primary nucleons which leads to the normal suppression of charmonia @xcite . both cronin effect and nuclear absorption
have been studied in p+a collisions @xcite , while a deep understanding of shadowing effect requires e+p and e+a collisions . in order to understand the charmonium production and
extract properties of the medium created in high - energy nuclear collisions , one must take into account both cold and hot nuclear matter effects .
lhc provides new data on the charmonium production in pb+pb collisions at colliding energy @xmath8 tev @xcite . at lhc
the initially produced charmonia are strongly suppressed by the hotter , larger and longer lived fireball , hence the regeneration becomes the dominant production source .
different from rhic where the parton longitudinal momentum fraction @xmath9 is located in the interface between shadowing and anti - shadowing @xcite regions and therefore the shadowing and anti - shadowing effect is not yet very strong ( the ratio of parton distribution function in a+a and p+p collisions is between 0.96 and 1.08 for charmonium transverse momentum @xmath10 gev / c , see fig.7 of ref.@xcite ) , the collisions at lhc with much smaller @xmath7 are in the strong shadowing region @xcite , and the shadowing effect plays an important role in the study of charmonium production . in the present study , we will extend our detailed transport approach for charmonium motion in qgp to self - consistently including both the cold and hot nuclear matter effects , and apply it to the charmonium production at lhc energy .
the suppression and regeneration mechanisms in the hot medium are reviewed in section [ s2 ] , and the shadowing effect , the gaussian smearing treatment for the cronin effect , and the initial charmonium distribution are focused in section [ s3 ] .
the comparison between the theoretical calculations and the experimental data are shown in section [ s4 ] .
we summarize our study in section [ s4 ] .
in order to extract information about the nature of the medium from charmonium production in heavy ion collisions , the medium created in the initial stage and the charmonia produced in the initial stage and in the medium should be treated both dynamically .
we employ the well tested 2 + 1 dimensional version @xcite of the ideal hydrodynamic equations @xmath11 to simulate the evolution of the almost baryon - free medium created at rhic and lhc , where @xmath12 is the energy - momentum tensor of the medium .
the solution of the hydrodynamic equations provides the local temperature and fluid velocity of the medium which will be used in the charmonium regeneration and suppression . to close the hydrodynamical equations one needs to know the equation of state of the medium .
we follow ref.@xcite where the deconfined phase at high temperature is an ideal gas of gluons and massless @xmath13 and @xmath14 quarks plus 150 mev massed @xmath15 quarks , and the hadron phase at low temperature is an ideal gas of all known hadrons and resonances with mass up to 2 gev @xcite .
there is a first order phase transition between these two phases . in the mixed phase
, the maxwell construction is used .
the mean field repulsion parameter and the bag constant are chosen as @xmath16=450 mev @xmath17 and @xmath18=236 mev to obtain the critical temperature @xmath19 mev @xcite at vanishing baryon number density .
taking the initialization in ref .
@xcite for the hydrodynamic equations , we get the maximum temperature @xmath20 and @xmath21 mev of the medium at the initial time @xmath22 fm / c , corresponding respectively to the observed charge number density @xmath23 in mid rapidity and @xmath24 in forward rapidity by the alice collaboration @xcite . since a charmonium is so heavy , its equilibrium with the medium can hardly be reached , one usually uses a transport approach to describe the charmonium motion in the medium .
the charmonium distribution function @xmath25 in the phase space @xmath26 at time @xmath27 in heavy ion collisions with impact parameter @xmath28 is controlled by the boltzmann - type equation @xmath29 considering the contribution from the feed - down of the excited states @xmath30 and @xmath31 to the finally observed @xmath2 @xcite , we should take into account the transport equations for all the charmonium states @xmath32 and @xmath31 , when we calculate the @xmath2 distribution .
the lose and gain terms @xmath33 and @xmath34 describe the charmonium dissociation and regeneration , and the elastic scattering is neglected since the charmonium mass is much larger than the typical medium temperature . introducing proper time @xmath35 , space rapidity @xmath36 $ ] ,
momentum rapidity @xmath37 $ ] and transverse energy @xmath38 to replace @xmath39 and @xmath40 , the transport equation can be rewritten as @xmath41f_\psi = - \alpha_\psi f_\psi+\beta_\psi , \label{trans2}\ ] ] where the third term in the square bracket arises from the free streaming of @xmath42 with transverse velocity @xmath43 which leads to a strong leakage effect at sps energy @xcite , and the loss term @xmath44 and gain term @xmath45 are respectively the charmonium dissociation and regeneration rates .
considering the gluon dissociation @xmath46 as the dominant dissociation process in the hot qgp , @xmath47 can be calculated by the gluon momentum integration of the dissociation cross section @xmath48 multiplied by the thermal gluon distribution @xmath49 and the flux factor @xmath50 , @xmath51 where @xmath52 is the gluon energy , the cross section @xmath48 in vacuum can be derived through the operator production expansion with perturbative coulomb wave function @xcite , and the cross section at finite temperature is estimated by taking the geometrical relation between the cross section and the average size of the charmonium state , @xmath53 .
the averaged radius square @xmath54 for the bound state @xmath42 can be obtained from the potential model @xcite , its divergence self - consistently defines a mott dissociation temperature @xmath4 which indicates the melting of the bound state due to color screening .
the local temperature @xmath55 and the fluid velocity @xmath56 of the medium in the cross section and the gluon thermal distribution come from the solution of the ideal hydrodynamic equations ( [ hydro ] ) .
the regeneration cross section is connected to the dissociation cross section @xmath48 via the detailed balance between the gluon dissociation and reversed process @xcite . to obtain the regeneration rate @xmath57 ,
we need also the charm quark distribution @xmath58 in the medium . from the experimental data at rhic and lhc ,
the observed large quench factor @xcite and elliptic flow @xcite for charmed mesons indicate that the charm quarks interact strongly with the medium .
therefore , one can reasonably take , as a good approximation , a kinetically thermalized phase space distribution @xmath59 for charm quarks , @xmath60 . neglecting the creation and annihilation for charm - anticharm pairs inside the medium , the spacial density of charm ( anti - charm ) quark number
@xmath61 satisfies the conservation law @xmath62 with the initial density determined by the nuclear geometry @xmath63 where @xmath64 and @xmath65 are the thickness functions at transverse coordinate @xmath66 defined in the glauber model @xcite , and @xmath67 is the rapidity distribution of charm quark production cross section in p+p collisions @xcite .
the initial condition for the charmonium transport equation ( [ trans2 ] ) can be obtained from a geometrical superposition of p+p collisions , along with the modification from the cold nuclear matter effect . for the charmonium production ,
the cold nuclear matter effect includes usually the nuclear shadowing @xcite , cronin effect @xcite and nuclear absorption @xcite . at lhc energy ,
the collision time for two heavy nuclei to pass through each other is much shorter than the charmonium formation time and the qgp formation time , and one can safely neglect the nuclear absorption and just take into account the nuclear shadowing and cronin effect .
the cronin effect broadens the momentum of the initially produced charmonia in heavy ion collisions . before two gluons fuse into a charmonium
, they acquire additional transverse momentum via multi - scattering with the around nucleons , and this extra momentum would be inherited by the produced charmonium . inspired from a random - walk picture , we take a gaussian smearing@xcite for the modified transverse momentum distribution @xmath68 @xmath69 where @xmath70 is the path length of the two gluons in nuclei before their fusion into a charmonium at @xmath71 , @xmath72 and @xmath73 are the longitudinal coordinates of the two nucleons where the two gluons come from , @xmath74 is the averaged charmonium transverse momentum square obtained from the gluon scattering with a unit of length of nucleons , and @xmath75 is the momentum distribution for a free p+p collision .
the length @xmath76 is calculated from the nuclear geometry , and the cronin parameter @xmath74 is usually extracted from corresponding p+a collisions where the produced charmonia suffer from only cold nuclear matter effect . considering the absence of p+a data at lhc energy , we take @xmath77 gev@xmath78/fm from empirical estimations @xcite . assuming that the emitted gluon in the gluon fusion process @xmath79 is soft in comparison with the initial gluons and the produced charmonium and can be neglected in kinematics , corresponding to the picture of color evaporation model at leading order @xcite , the longitudinal momentum fractions of the two initial gluons are calculated from the momentum conservation , @xmath80 where @xmath81 is the charmonium rapidity .
the free distribution @xmath82 can be obtained by integrating the elementary partonic processes , @xmath83 where @xmath84 is the gluon distribution in a free proton , @xmath85 is the emitted gluon s rapidity , @xmath86 is the charmonium momentum distribution produced from a fusion process , and @xmath87 is the factorization scale of the fusion process .
now we consider the shadowing effect .
the distribution function @xmath88 for parton @xmath89 in a nucleus differs from a superposition of the distribution @xmath90 in a free nucleon .
the nuclear shadowing can be described by the modification factor @xmath91 . to account for the spatial dependence of the shadowing in a finite nucleus
, one assumes that the inhomogeneous shadowing is proportional to the parton path length through the nucleus @xcite , which amounts to consider the coherent interaction of the incident parton with all the target partons along its path length .
therefore , we replace the homogeneous modification factor @xmath92 by an inhomogeneous one @xmath93 @xcite @xmath94 with the definition @xmath95 .
we employ in the following the eks98 package @xcite to evaluate the homogeneous ratio @xmath96 , and the factorization scale is taken as @xmath97 .
replacing the free distribution @xmath49 in ( [ fg ] ) by the modified distribution @xmath98 and taking into account the cronin effect ( [ cronin ] ) , we finally get the initial charmonium distribution for the transport equation ( [ trans2 ] ) , @xmath99 where @xmath100 and @xmath101 are the nucleon distribution functions in the two colliding nuclei .
now the only thing left is the distribution @xmath102 in a free p+p collision which can be fixed by data or by some empirical estimations . since the charmonia in heavy ion collisions at lhc
are measured by the cms collaboration at mid rapidity @xmath103 @xcite and by the alice collaboration at mid rapidity @xmath104 @xcite and forward rapidity @xmath105
@xcite , one should consider the rapidity dependence of the free distribution @xmath102 .
the charmonium states measured by the alice are reconstructed down to @xmath106 via the @xmath107 decay channel at forward rapidity and via the @xmath108 channel at mid rapidity . from the measurement @xcite ,
the averaged cross section is @xmath109b at @xmath104 and @xmath110b at @xmath105 .
the alice has also measured the transverse momentum distribution in p+p collisions @xcite , the combined @xmath81 and @xmath111 dependence for inclusive @xmath2s at @xmath105 can be parameterized as @xmath112 with @xmath113 and @xmath114(gev / c)@xmath78 , shown as the lower solid line in fig.[fig1 ] . from the alice @xcite and rhic @xcite data and cem(color evaporation model ) @xcite calculation , the free distribution @xmath115 for inclusive @xmath2s at rapidity @xmath103 can also be described by the parametrization ( [ ppjpsi ] ) with @xmath116 , @xmath117 ( gev / c)@xmath78 and @xmath118b , see the other solid line in fig.[fig1 ] .
production cross section as a function of transverse momentum in 2.76 tev p+p collisions .
the thick-(red online ) and thin-(black online ) lines represent mid rapidity and forward rapidity , respectively .
the data at forward rapidity are from the alice collaboration @xcite.,scaledwidth=45.0% ] transverse momentum in p+p collisions .
the data are from the cdf @xcite and cms @xcite collaborations , and the straight line is our linear parametrization . , scaledwidth=42.0% ] the inclusive @xmath2s measured by alice include the prompt part and the contribution from the b decay .
the former consists of direct production and feed down from the excited states , @xmath119 from @xmath31 and @xmath120 from @xmath30 , and the latter comes from the decay of bottomed hadrons .
the b decay contributes about @xmath120 @xcite to the total inclusive yield .
since the alice does not separate the two parts from each other , we have to take into count the b decay contribution in our model to compare our theoretical calculation with the alice data .
cms experiment has measured the prompt charmonia at high transverse momentum @xmath121 gev / c and at mid rapidity @xmath103 @xcite . since at the moment there are no p+p data for prompt charmonia in the cms rapidity region , we employ @xmath122b from the alice data @xcite for inclusive @xmath2s in mid rapidity , and then eliminate the b decay contribution by multiplying the inclusive cross section by a factor of @xmath123 , where @xmath124 is the b decay fraction .
fig.[fig2 ] shows the recent data on the b decay fraction in p+p(@xmath125 ) collisions as a function of @xmath126 @xmath111 @xcite .
the data can be well parameterized as @xmath127/(gev / c ) .
note that the linear parametrization is rapidity independent in the region we considered .
the prompt @xmath2 distribution at mid rapidity @xmath103 is shown as the upper dashed curve in fig.[fig1 ] .
since the b decay fraction is approximately rapidity independent , the distribution @xmath102 for inclusive @xmath2s at mid rapidity @xmath104 can be obtained from the prompt distribution at mid rapidity @xmath103 and the b decay fraction @xmath124 shown in fig.[fig2 ] .
we now turn to the rapidity dependence of the charm quark production cross section @xmath128 in ( [ rhoc ] ) . from the alice data
@xcite , there is @xmath129 mb at mid rapidity . by taking the fonll @xcite scaling ,
we extract @xmath130 mb at forward rapidity . considering the uncertainty in the experimental data and theoretical calculations ( see fig.5 of ref.@xcite ) , the upper limits of fonll @xcite and pqcd are often used in models @xcite , to estimate the maximum quarkonium production .
we take in the following numerical calculations @xmath131 between @xmath132 and @xmath133 mb at mid rapidity and @xmath134 and @xmath135 mb at forward rapidity .
note that the shadowing effect changes not only the initial @xmath2 distribution , but also the in - medium @xmath2 regeneration and the non - prompt contribution from the b decay , by reducing the number of charm and bottom quarks . in principle
, the shadowing should be centrality dependent . to simplify the numerical calculations , we take in the following a reduction of @xmath136 for the charm and bottom quark production cross sections , estimated from the centrality averaged eks98 evolution @xcite .
we start with the @xmath2 nuclear modification factor @xmath138 as a function of the number of participants @xmath139 , where @xmath140 and @xmath141 are respectively the numbers of measured @xmath2s in p+p and a+a collisions , and @xmath142 is the number of nucleon - nucleon collisions at fixed @xmath139 .
our model calculations for 2.76 tev pb+pb collisions and the comparison with the inclusive alice data are shown in fig.[fig3 ] at forward rapidity ( upper panel ) and mid rapidity ( lower panel ) .
nuclear modification factor as a function of centrality for 2.76 tev pb+pb collisions at forward rapidity and in different transverse momentum bins .
the dot dashed lines are the initial fraction , the thick and thin dashed lines are the regeneration fraction with charm quark cross section @xmath143 and @xmath134 mb , and the bands are the full result .
the data are from the alice collaboration @xcite.,scaledwidth=50.0% ] nuclear modification factor as a function of centrality for 2.76 tev pb+pb collisions at mid rapidity and in a high transverse momentum bin .
the line is the model calculation and the data are from the cms collaboration @xcite.,scaledwidth=45.0% ] with increasing centrality , the initial contribution ( dot - dashed lines ) drops down and the regeneration ( dashed lines ) goes up monotonously .
considering the uncertainty in the charm quark production cross section , we take the range of cross - sections @xmath144 and @xmath135 mb at forward rapidity and @xmath132 and @xmath133 mb at midrapidity , corresponding to the thin and thick dashed lines in fig.[fig3 ] .
this uncertainty in the regeneration results in a band for the full result .
different from the collisions at sps energy where the regeneration can be neglected @xcite and at rhic energy where the initial production is still a dominant component and the regeneration becomes equivalently important only in very central collisions @xcite , the regeneration at lhc energy becomes the dominant source of charmonium production in a wide centrality bin .
the competition between the strong dissociation and regeneration leads to a flat structure for the total charmonium production at both forward rapidity and mid rapidity , see fig.[fig3 ] . at forward rapidity where high statistic data are available , our model results
well explain the data in semi - central and central collisions with @xmath145 , and the deviation from the data in small @xmath139 region is probably due to the invalidation of hydrodynamics for the medium and the canonical limit for the charmonium regeneration @xcite . at mid rapidity where the data are with large error bars ,
the model results seem systematically under the data .
we have seen the strong competition between the charmonium suppression and regeneration in the centrality dependence of the @xmath2 yield . in order to see the charmonium production and suppression mechanisms more clearly in heavy ion collisions , we turn to the transverse momentum dependence of the yield .
figs.[fig4 ] and [ fig5 ] show the nuclear modification factor @xmath137 as a function of centrality in different @xmath111 bins .
in addition we make comparisons with alice data @xcite ( fig.4 ) and cms data @xcite ( fig.5 ) for inclusive and prompt @xmath2 production , respectively . in our calculation
, we considered a reduction of @xmath136 for the charm quark number due to the strong shadowing effect at lhc energy .
we also took a kinetically thermalized charm quark distribution , by assuming strong interaction between charm quarks and the medium . for bottom quarks
, we took into account the same reduction for the number in the calculation of @xmath111 integrated yield .
since bottom quarks are so heavy , their thermalization is unreasonable in nuclear collisions .
while the total @xmath2 yield shown in fig.[fig3 ] is not sensitive to the bottom quark transverse momentum distribution , it may change remarkably the @xmath2 yield in a fixed @xmath111 bin , especially in a high @xmath111 bin where the bottom quark energy loss becomes important .
inspired from the cms measurement @xcite , we apply a b quench factor @xmath146 when calculating the non - prompt @xmath2s in the region of @xmath147 gev / c .
heavy quarks are produced via hard scatterings and their initial momentum distribution is hard . through interation with the medium ,
heavy quarks lose energy , and the corresponding @xmath148 distribution becomes steeper .
since the medium is hot and dense , some of the charm quarks are even thermalized .
considering that the regenerated charmonia from thermalized charm quarks in hot medium are mainly distributed in low momentum region , their contribution to the yield decreases with increasing transverse momentum , and therefore the band structure for the total result due to the uncertainty in regeneration disappears in the high @xmath111 bins @xmath147 gev / c and @xmath121 gev / c , see figs.[fig4 ] and [ fig5 ] .
as we have pointed out that the flat platform for semi - central and central collisions comes from the competition between the suppression and regeneration , it vanishes in the high @xmath111 bins due to the disappearance of the regeneration .
the yield in high @xmath111 bins is characterized by the debye screening effect on the initially produced charmonia . in this case
, there exists a kink for the nuclear modification factor which is located at @xmath149 ( fig.4 ) for @xmath147 gev / c at forward rapidity and @xmath150 ( fig.5 ) for @xmath121 gev / c in mid rapidity . before the kink , the temperature of the fireball is less than the @xmath2 dissociation temperature @xmath151 and there is @xmath152 , resulted from the fact that the decay contribution from the excited charmonium states to the finally observed @xmath2s is about @xmath153 .
starting at the kink , the temperature is higher than @xmath151 and the @xmath2 suppression becomes more and more important in more central collisions .
as one can see from figs.[fig4 ] and [ fig5 ] , the kink at mid rapidity appears at a lower value of @xmath139 than that at forward - rapidity .
this is due to the fact that the medium at mid rapidity is hotter and @xmath2 starts to melt at more peripheral collisions .
the model results underestimate the high @xmath154 results a bit , see figs.[fig4 ] and [ fig5 ] .
part of the discrepancy may due to the fact we have not include the velocity dependence of the dissociation temperature . as discussed in @xcite , fast moving charmonia tend to be melt at a higher temperature making the velocity dependence particularly important in high @xmath154 regions . nuclear modification factor as a function of transverse momentum for 2.76 tev pb+pb collisions at forward rapidity and in different centrality bins .
the dot - dashed lines are the initial fraction , the thick and thin dashed lines are the regeneration fraction with charm quark cross section @xmath143 and @xmath134 mb , and the bands are the full result .
the data are from the alice collaboration @xcite.,scaledwidth=50.0% ] nuclear modification factor as a function of transverse momentum for 2.76 tev pb+pb collisions at mid rapidity and in minimum bias event .
the dot - dashed line is the initial fraction , the thick and thin dashed lines are the regeneration fraction with charm quark cross section @xmath155 and @xmath132 mb , and the band is the full result .
the data are from the cms collaboration @xcite.,scaledwidth=45.0% ] the transverse momentum distribution of the nuclear modification factor @xmath156 in a fixed centrality bin and its comparison with the alice data are shown in fig.[fig6 ] for inclusive @xmath2s at forward rapidity .
we see clearly that , the regeneration ( dashed lines ) dominants the low @xmath154 regions at all centralities . on the other hands ,
initially produced @xmath2s ( dot - dashed lines ) become important at high @xmath154 region .
therefore , the competition between the suppression and regeneration depends strongly on the transverse momentum and collision centralities .
when both the regeneration at low @xmath111 and initial production at high @xmath111 are important , there exists a minimum structure located at intermediate @xmath111 , see the upper and middle panels of fig.[fig6 ] .
note that without the regeneration as the second production source , the nuclear modification factor @xmath137 would decrease monotonously with increasing centrality and increases monotonously with increasing transverse momentum .
the @xmath156 for prompt @xmath2s at mid rapidity is shown in fig.[fig7 ] . while the cms data are in the high @xmath111 region where there is almost no contribution from the regeneration and the system
is controlled only by the dissociation of the initially produced @xmath2s , the predicted @xmath137 at very low @xmath157 is larger than unity which is not possible without the regeneration mechanism . for heavy ion collisions in the past decades , the colliding energy from sps to lhc increases by two orders of magnitude , and the heavy quark production cross section increases dramatically . while the transverse momentum dependence of the nuclear modification factor @xmath137 tells us the importance of the regeneration mechanism , we hope to find a quantity which is more sensitive to the nature of the medium . on the other hand , in our treatment for hot nuclear matter effect
, we assumed a thermalized charm quark distribution for the calculation of the regeneration rate , how good is this assumption and how does the charm quark thermalization affect the charmonium distribution ?
the other question is how to separate the hot nuclear matter effect from the cold nuclear matter effect . to focus on the property of the hot medium , we hope to have a quantity which is sensitive to the hot nuclear matter effect but affected weekly by the cold nuclear matter effect . to these ends
, we introduced a new nuclear modification factor for the transverse momentum @xcite @xmath159 where @xmath160 and @xmath161 are averaged transverse momentum square for @xmath2 in a+a and p+p collisions , respectively . to be different from the usually used nuclear modification factor @xmath137 for the yield
, we take here @xmath3 as the ratio of the second moment of transverse momentum distributions .
since the nuclear matter effect changes not only the size of the transverse momentum but also its shape , we focus on @xmath162 instead of @xmath163 .
it is worthy to note that the second moment of the @xmath154 distribution is also sensitive to the cronin effect but not much affected by the parton distribution functions .
the calculation and the comparison with experimental data at sps , rhic and lhc energies are shown in fig.[fig8 ] for two rapidity bins . in mid rapidity where the hot nuclear matter effect is most important , the ratio changes significantly with increasing colliding energy , @xmath164 see the upper panel of fig.[fig8 ] . at sps ,
almost all the measured @xmath2s are produced through initial hard processes and carry high momentum .
the continuous increasing with centrality arises from the cronin effect and leakage effect @xcite .
the latter is described by the third term on the left hand side of the transport equation ( [ trans2 ] ) and means that the high @xmath111 @xmath2s can escape from the hot medium . at rhic ,
the regeneration starts to play a role and even becomes equally important as the initial production in central collisions @xcite , the cancelation between the suppression and regeneration leads to a flat @xmath3 around unity . at lhc ,
the regeneration becomes dominant , especially in central collisions . since the regenerated charmonia carry low momentum in comparison with the initial production , the more and more important regeneration results in a decreasing @xmath3 with increasing centrality .
this tremendous change in the @xmath111 ratio comes from the nature of the hot medium at each corresponding collision energy .
our calculation agrees well with the data at sps and rhic energies . at lhc energy
there are currently no data at mid rapidity .
the band structure at lhc is again from the uncertainty in the charm quark production cross section .
note that , different from the ratio @xmath137 for the yield where the lower and higher borders of the band correspond to the smaller and larger cross sections , the relation between the two borders of @xmath3 and the cross section is in an opposite way , namely , the lower and higher borders here correspond to the larger and smaller cross sections . intuitively this is simple to understand : in the case of larger charm quark cross section , more @xmath2s are from the regeneration and then carry low momentum . to have a comparison with the lhc data , we calculated the ratio @xmath3 in the forward rapidity where the hot medium effect becomes weaker .
in this case the trend of the ratio is still very different at rhic and lhc : with increasing centrality it goes up at rhic but drops down at lhc .
@xmath165 see the lower panel of fig.[fig8 ] .
the calculation agrees well with both the rhic and lhc data .
for @xmath2s as a function of centrality at sps , rhic and lhc energies .
the bands at lhc are due to the uncertainty in the charm quark cross section @xmath166 mb at forward rapidity ( lower panel ) and @xmath167 mb at mid rapidity ( upper panel ) , and the data are from na50 @xcite , phenix @xcite and alice @xcite collaborations . ,
scaledwidth=50.0% ] for @xmath2s as a function of centrality at lhc energy and at forward rapidity .
the upper and lower panels are the comparisons between pqcd and thermal charm quark distributions and between with and without shadowing effect .
the bands are again due to the uncertainty in the charm quark cross section @xmath166 mb , and the data are from the alice collaboration @xcite.,scaledwidth=50.0% ] after a careful inspecting the centrality dependence of @xmath3 , one can find a common feature in all cases , namely , @xmath3 in peripheral collisions is always increasing as the centrality increases . in case
the initial production is dominant , the ratio continues to increase , while strong regeneration from thermalized charm quarks pulls the ratio down below unity in more central collisions . in order to see
if charm quarks are thermalized with the hot medium , we calculate now the ratio @xmath3 at lhc energy with a pqcd simulated charm quark distribution , shown in the upper panel of fig.[fig9 ] at forward rapidity .
the thermal distribution is the limit of strong interaction between charm quarks and the medium , while the pqcd distribution taken from the simulator pythia @xcite is the limit of no interaction . for the pqcd distribution , the charm quark energy loss
is excluded , the initially produced charm quarks keep their high momentum in the medium , and therefore the regenerated charmonia will be no longer soft but carry high momentum . in this case , the averaged transverse momentum is enhanced at the forward rapidity and the ratio @xmath3 becomes larger than the one with thermal distribution , see the upper panel of fig.[fig9 ] . again
the band is due to the uncertainty in the charm quark cross section . from the comparison with the alice data ,
we conclude that charm quarks are thermalized in high - energy nuclear collisions at the lhc energy . in the above calculations we have taken a reduction of @xmath136 for the charm and bottom quark distributions in the medium , due to the shadowing effect .
this reduction leads to an extra charmonium suppression . supposing the total number of initially produced charm quarks is @xmath168 , the regenerated charmonium number is then proportional to @xmath169 if we do not take into account the shadowing effect and @xmath170 when the shadowing effect is included .
however , the averaged transverse momentum square @xmath162 is a normalized quantity , it should not be so sensitive to the shadowing effect .
the comparison between the calculations with and without considering the shadowing effect is shown in the lower panel of fig.[fig9 ] .
as we expected , the two bands are very close to each other and overlap partly .
this indicates that the nuclear modification factor @xmath3 for charmonium transverse momentum can be used to extract the hot medium information with only small correction from the early shadowing effect . for prompt ( upper panel ) and inclusive ( lower panel ) @xmath2s in @xmath8 tev pb+pb collisions .
the calculation is with impact parameter @xmath171 fm , corresponding to the minimum bias event .
the dot - dashed , dashed and solid lines represent the initial , regeneration and total contributions . in the lower panel ,
the thick and thin lines indicate the total results including thermal bottom decay and pqcd bottom decay , respectively .
the experimental data are taken from alice experiment @xcite.,scaledwidth=50.0% ] since the elliptic flow originates from the geometric anisotropy in the configuration space and develops in the evolution of the hot medium , it is significant in semi - central collisions and approaches to zero in peripheral and central collisions . as far as the charmonium concern , those that produced from the initial collisions prior to the formation of the hot medium will not present any significant elliptic flow , see the dot - dashed line in the upper panel of fig.[fig10 ] .
the regenerated @xmath2s , on the other hand , may inherit flow from the thermalized charm quarks , as shown by the dashed line . through interactions , the initial anisotropy in the configuration space
is converted into the anisotropy in momentum space for all final hadrons .
the full @xmath2 @xmath172 is shown as the solid line in the upper panel of fig.[fig10 ] .
note that a sizable @xmath172 has been developed at lhc .
unlike the light charged hadrons where @xmath172 persist with large values up to high @xmath154 region , the @xmath2 @xmath172 quickly drops to the value from initial production with @xmath173 . up to now , our discussion on @xmath2 @xmath1 is focused on the prompt @xmath2 .
once we include the @xmath2 from bottom quark decay , new scenarios arise .
the lower panel of fig.[fig10 ] shows the elliptic flow @xmath172 for inclusive @xmath2s at forward rapidity where the bottom decay is included .
as one can see in the plot , the @xmath2 @xmath1 from the decay of the initially ( or pqcd ) produced bottom quarks quickly falls to close to zero at @xmath174 gev / c .
however , the @xmath2s from thermalized bottom decay show significant value of @xmath172 at high @xmath111 region . at this point ,
the error of the experimental data is too large to draw any conclusion of bottom quark thermalization .
our calculation predicts that future precise @xmath2 @xmath172 at high @xmath154 will provide important information on bottom production in high energy nuclear collisions .
we have studied the cold and hot nuclear matter effects on charmonium production in relativistic heavy ion collisions . in the framework of a transport approach for the charmonium motion plus a hydrodynamic description for the medium evolution , we calculated the charmonium transverse momentum distribution and the corresponding ratios @xmath137 and @xmath3 in different rapidity and centrality bins .
our model calculations agree reasonably well with the experimental data at sps , rhic and lhc energies .
we found that , in comparison with the often used nuclear modification factor @xmath175 , based on the charmonium yield , the newly defined nuclear modification factor @xmath158 , based on the charmonium transverse momentum distribution , is more sensitive to the nature of the hot medium and to the degree of heavy quark thermalization . when the colliding energy increases from sps to lhc , the @xmath111 @xmath3 changes dramatically from above unity and increasing as a function of collision centrality to below unity and decreasing versus centrality .
different from the yield @xmath137 which is strongly affected by both the cold and hot nuclear matter effects , the @xmath111 @xmath3 is weekly affected by the shadowing effect .
in addition , we observed that the @xmath2 elliptic flow @xmath172 for minimum bias events evolves from almost zero at rhic to a large value similar to that of light hadrons at lhc .
the root for these observed dramatic changes is the formation of the strongly coupled quark - gluon plasma and the thermalization of charm quarks in high energy nuclear collisions at the lhc energy .
99 t.matsui and h.satz , phys . lett . *
b178 * , 416(1986 ) . m.asakawa and t.hatsuda , phys .
lett . * 92 * , 012001(2004 ) s.datta , f.karsch , p.petreczky and i.wetzorke , phys . rev . *
d69 * , 094507(2004 ) .
m.gonin , et al .
, nucl . phys . *
a610 * , 404(1996 ) .
n.s.topilskaya , et al . , [ na50 collaboration ] , nucl . phys . * a715 * , 675(2003 ) .
b.alessandro , et al.[na50 collaboration ] , eur .
j. * c39 * , 335(2005 ) .
j.adams , et al . , [ star collaboration ] , phys .
* 94 * , 062301(2005 ) .
a.adare , et al . , [ phenix collaboration ] phys
* 97 * , 252002(2006 ) .
p.braun-munzinger and j.stachel , phys . lett . *
b490 * , 196(2000 ) ; a.adronic , p.braun-munzinger , k.redlich and j.stachel , phys . lett . *
b652 * , 259(2007 ) .
l.grandchamp and r.rapp , nucl . phys . *
a709 * , 415(2002 ) .
r.l.thews m. schroedter and j. rafelski , phys
* c63 * , 054905(2001 ) .
l.yan , p.zhuang and n.xu , phys .
lett . * 97 * , 232301(2006 ) . x.zhao and r.rapp , phys
b664 * , 253(2008 ) .
a.h.mueller and j.w.qiu , nucl . phys . *
b268 * , 427(1986 ) .
j.w.cronin , et al .
, phys . rev .
* d11 * , 3105(1975 ) ; j.hufner y.kurihara and h.j.pirner , phys . lett . *
b215 * , 218(1988 ) .
c.gerschel and j.hufner , phys . lett .
* b207 * , 253(1988 ) .
e.m.c.abreu , et al .
, phys . lett . *
b466 * , 408(1999 ) .
c.silvestre , et al . , [ cms collaboration ] , j. phys . *
g38 * , 124033(2011 ) .
j.wiechula , et al . , [ alice collaboration ] , c12 - 05 - 27.2 proceedings , arxiv:1208.6566 .
cms collaboration , cms - pas - hin-12 - 014(2012 ) .
d.das , et al .
, [ alice collaboration ] , conf .
proc.*57 * , 37(2012 ) , arxiv:1212.2704 .
c.lourenco , h.k.whri , physics reports * 433 * , 127(2006 ) .
k.j.eskola , v.j.kolhinen and c.a.salgado , eur .
j. * c9 * , 61(1999 ) .
x.zhu , p.zhuang and n.xu , phys . lett . * b607 * , 107(2005 ) .
j.sollfrank , et al .
, phys . rev .
* c55 * , 392(1997 ) .
k.hagiwara , et al . , particle data group , phys . rev . *
d66 * , 010001(2002 ) .
t.hirano , h.pasi and n.yasushi , phys . rev .
* c83 * , 021902(2011 ) .
k.gulbrandsen , [ alice collaboratoin ] , j. phys . conf
. ser . * 446 * , 012027(2013 ) .
a.zoccoli , et al .
, hera - b collaboration , eur .
j. * c43 * , 179(2005 ) .
j.hufner and p.zhuang , phys . lett .
* b559 * , 193(2003 ) . g.bhanot and m.e.peskin , nucl
b156 * , 365(1979 ) ; g.bhanot and m.e.peskin , nucl . phys . *
b156 * , 391(1979 ) .
f.arleo , et al . , phys . rev . * d65 * , 014005(2002 ) .
ys.oh , s.kim and s.h.lee , phys . rev . *
c65 * , 067901(2002 ) .
x.n.wang , phys . lett . *
b540 * , 62(2002 ) .
h.satz , j. phys . *
g32 * , r25(2006 ) . l.kluberg and h.satz , helmut , arxiv:0901.3831[hep - ph ] , relativistic heavy ion physics , landolt - brnstein - group i elementary particles , nuclei and atoms * 23 * , 373(2010 ) .
b.i.abelev , et al . , [ star collaboration ] , phys . rev
. lett . * 98 * , 192301(2007 ) .
a.dainese , [ alice collaboration ] , pos .
ichep2012 , 417(2013 ) , arxiv:1212.0995 ; z.conesa del valle , [ alice collaboration ] , arxiv:1210.2163[nucl - ex ] , c12 - 05 - 27.2 proceedings .
a.adare , et al .
, [ phenix collaboration ] , phys .
lett . * 98 * , 172301(2007 ) . z.conesa del valle , [ alice collaboration ] , nucl . phys .
* a904 - 905 * , 178c(2013 ) ; bianchin , chiara , et al . , [ alice collaboration ] , arxiv:1111.6886[hep - ex ] , sqm 2011 conference proceedings .
m.l.miller , et al .
sci . * 57 * , 205(2007 ) .
b.abelev , et al . , [ alice colloboration ] , jhep * 07 * , 191(2012 ) .
m.cacciari , m.greco , and p.nason , jhep , 9805:007(1998 ) .
m.cacciari , s.frixione , and p.nason , jhep , 0103:006(2001 ) . x.zhao and r.rapp , nucl
* a859 * , 114(2011 ) ; x.zhao , e.andrew , r.rapp , nucl .
* a904 - 905 * , 611c(2013 ) .
y.liu , et al .
, phys . lett .
* b697 * , 32(2011 ) .
x.n.wang , phys .
81 * , 2655(1998 ) .
r.vogt , int . j. mod
e12 * , 211(2003 ) .
r.l.thews and m.l.mangano , phys .
rev . * c73 * , 014904(2006 ) .
h. fritzsch , phys . lett . *
b67 * , 217(1977 ) .
j.f amundson , et al .
* b372 * , 127(1996 ) .
j.f amundson , et al .
b390 * , 323(1997 ) .
s.r.klein and r.vogt , phys .
lett . * 91 * , 142301(2003 ) .
r.vogt , phys . rev . *
c71 * , 054902(2005 ) .
i.c.arsene , et al .
, [ alice collaboration ] , nucl . phys . * a904 - 905 * , 623c(2013 ) .
g.m.garca , et al . , [ alice collaboration ] , j , phys . * g38*,124034(2011 ) .
p.pillot , et al . , [ alice collaboration ] , j , phys . * g38*,124111(2011 ) .
d.das , et al . , [ alice collaboration ] , pramana*79 * , 863(2012 ) .
pereira da costa hugo , et al .
, [ alice collaboration ] , arxiv:1110.1035 , aip conf .
proc.*1441 * , 859(2012 ) .
b.alessandro , et al . , [ alice collaboration ] , j. phys . *
g32 * , 1295(2006 ) .
s.s.adler , et al .
, [ phenix collaboration ] , phys .
lett . * 92 * , 051802(2004 ) .
a.accardi , et al . , cern-2004 - 009-a , hip-2003 - 40/th .
r.aaij , et al .
, [ lhcb collaboration ] , eur .
phys . j * c71 * , 1645(2011 ) .
d.acosta , et al . , [ cdf collaboration ] , phys . rev . *
d71 * , 032001(2005 ) .
v.khachatryan , et al . , [ cms collaboration ] , eur . phys . j. * c71 * , 1575(2011 ) .
y.liu , et al . , j. phys .
* g36 * , 064057(2009 ) .
l.grandchamp and r.rapp , phys .
b523 * , 60(2001 ) .
c.m.ko , et al .
lett . * 86 * , 5438(2001 ) .
y.liu , n.xu and
p.zhuang , phys . lett . * b724 * , 73(2013 ) .
k.zhou , n.xu and p.zhuang , nucl . phys . *
a834 * , 249(2010 ) .
a.adare , et al . , [ phenix collaboration ] , phys .
* 98 * , 232301(2007 ) .
e.scomparin ( infn torino ) , [ alice collaboration ] , qm2012 presentation . t. sjstrand , et al . ,
commun*135 * , 238(2001 ) .
h.yang , [ alice collaboration ] , nucl .
* a904 - 905 * , 673c(2013 ) . | we investigate with a transport approach the cold and hot nuclear matter effects on the charmonium transverse momentum distributions in relativistic heavy ion collisions .
the newly defined nuclear modification factor @xmath0 and elliptic flow @xmath1 for @xmath2 are sensitive to the nature of the hot medium and the thermalization of heavy quarks . from sps through rhic to lhc colliding energies ,
we observe dramatic changes in the centrality dependence of @xmath3 .
we find that at lhc energy , the finally observed charmonia are dominated by the regeneration from thermalized heavy quarks . |
several multiple planet systems have been reported , including the triple planet system around upsilon andromedae ( butler et al .
1999 ) and double planet systems around gj876 ( marcy et al .
2001b ) , hd 83443 ( mayor et al .
2000 ) , hd 168443 ( marcy et al .
2001a , udry et al . 2002 ) and 47 uma ( fischer et al . 2002 ) .
double - planet systems have also been reported in a press release ( eso press release 07/01 , 2001 ) for hd 82943 and hd 74156 .
these multiple planet systems contain planets reported to range from a saturn mass to nearly 10 jupiter mass , all orbiting within 4 au .
interactions between the planets in some of these systems , notably gliese 876 , are measurable on a time scale of a few years ( lissauer & rivera 2001 ; laughlin & chambers 2001 ; rivera & lissauer 2001 ) .
doppler measurements can reveal the ongoing gravitational perturbations and constrain both the planet masses and orbital inclinations .
the interactions and orbital resonances , both mean motion and secular , provide clues about the dynamical history of the systems ( snellgrove et al .
2001 ; lee and peale 2002a ; chiang et al .
2002 ) .
a most extraordinary double planet system was reported for hd 83443 ( mayor et al .
their doppler measurements made with the coralie spectrometer indicate the existence of two saturn mass planets that both reside within 0.2 au .
the inner planet has an orbital period of 2.985 d , an eccentricity of 0.079 ( @xmath00.033 ) , a minimum ( @xmath1 ) mass of 0.34 m@xmath2 , and an orbital distance of 0.038 au .
the orbital period is the shortest known for extrasolar planets .
zero eccentricity of this inner planet is notable as planets with periods less than 5 d suffer tidal circularization ( wu & goldreich 2002 ) . with the exception of the saturn
mass planet around hd 46375 ( marcy et al . 2000 ) , all 15 of the previously discovered `` 51 peg like '' planets have spectral types of g5 or earlier .
mayor et al .
( 2000 ) report a remarkable second planet around hd 83443 .
it has an orbital period of 29.83 ( @xmath00.18 ) d , an eccentricity of 0.42 , a minimum ( @xmath1 ) mass of 0.16 m@xmath2 , and an orbital distance of 0.17 au .
this outer planet induces a velocity semiamplitude in the star of @xmath3 = 13.8 @xmath0 1 , rendering it a 14-@xmath4 detection . this planet has the smallest @xmath1yet reported , and is only the third reported planet with a semiamplitude smaller than 15 ( cf .
marcy et al . 2000 ; fischer et al . 2002 ) . both of the planets were indicated by doppler measurements obtained with the 1.2-m leonhard euler telescope at the eso la silla observatory , which feeds the coralie spectrometer ( queloz et al .
wavelength calibration is achieved by coupling the telescope and thorium lamp to the spectrometer with a double scrambled fiber .
the quoted instrumental precision is now 2 ( udry et al . 2002 ) .
as the two planets orbiting hd 83443 are crowded within 0.2 au , the system is dynamically active .
calculations by j.laskar and w.benz ( reported in mayor et al .
2000 ) and by wu & goldreich ( 2002 ) and lee & peale ( 2002b ) suggest the occurrence of significant gravitational interactions between the two planets .
the tidal circularization time scale for the inner planet to hd 83443 is estimated to be @xmath5 yr ( wu & goldreich 2002 ) , while the star is estimated to have an age of 6.5 gyr . in this context , the non
zero eccentricity of the inner planet and the apside alignment of the two orbits are understood to be due to secular interactions between the two planets and tidal interactions with the star ( mayor et al .
2000 ; wu & goldreich 2002 ) .
these in turn constrain the orbital inclination of this system , and the radius of the inner planet ( wu & goldreich 2002 ) . the dynamical evolution that led to the system may involve migration and resonances ( lee & peale 2002b ) .
section 2 of this paper will describe new doppler measurements of hd 83443 made from the keck and aat telescopes .
section 3 describes a search for the two planets , notably the interesting outer planet .
our failure to detect the outer planet is discussed in section 4 .
hd 83443 ( hip47202 ) is among the fainter g & k dwarfs surveyed by precision doppler programs with v = 8.23 and b@xmath6v=0.811 ( perryman et al .
1997 ) , consistent with the assigned spectral type , k0 v. the hipparcos derived distance is 43.5 pc .
( note that the distance of 23 pc reported in mayor et al .
( 2000 ) is incorrect . )
the star is photometrically stable at the level of hipparcos measurement uncertainty .
the metallicity of the star , [ fe / h ] @xmath7 @xmath8 ( santos et al .
2000a ) , is similar to other stars with `` 51 peg like '' planets .
the precise doppler observations presented in this paper were made with the hires echelle spectrometer ( vogt et al .
1994 ) on the 10m keck i telescope , and the ucles echelle spectrometer ( diego et al .
1990 ) on the 3.9m anglo - australian telescope .
these spectrometers are operated at a resolution of r @xmath9 80000 and r @xmath9 45000 respectively .
wavelength calibration is carried out by means of an iodine absorption cell ( marcy & butler 1992 ) which superimposes a reference iodine spectrum directly on the stellar spectra ( butler et al .
1996a ) these systems currently achieve photon limited measurement precision of 3 .
detailed information on these two systems , including demonstration stable stars , can be found in vogt et al .
2000 ( keck ) and butler et al .
2001 ( aat ) . based on our photometrically estimated metallicity , [ fe / h]=+0.31 , we added hd 83443 to the anglo
australian precision doppler survey in feb 1999 .
this is among the very faintest stars in the aat survey .
ten minute exposures on the 3.9m anglo australian telescope yield typical s / n @xmath970 , giving a median measurement uncertainty of 8.0 ( butler et al . 2001 ) . a total of 16 aat observations of hd 83443 have been made between feb 1999 and mar 2002 .
hd 83443 was added to the keck precision doppler survey ( vogt et al .
2000 ) in dec 2000 as a result of the coralie announcement of a double planet system .
a total of 20 keck observations have been obtained through mar 2002 .
the aat and keck velocity measurements are listed in table 1 .
we fit the velocities with a simple keplerian model for which the usual free parameters are @xmath10 , @xmath11 , @xmath12 , @xmath13 , @xmath3 , as well as a system velocity zero point , @xmath14 .
figures 1 and 2 show the keck and aat velocities , respectively , phased at the best
fit keplerian orbital period of 2.9856 d. the reduced @xmath15 to the keplerian fits to these data sets are 1.33 and 0.83 respectively .
figure 3 shows the combined set of velocities phased .
figure 1 shows that a single keplerian model , without invoking a second planet , yields a fit to the keck velocities with an rms of 4 . while this strongly confirms the inner planet , the low rms is surprising because the reported second planet causes a semiamplitude of 13.8 ( mayor et al . 2000 ) , but is not included in this single
keplerian fit .
similarly the aat velocities are well fit , within measurement uncertainty , with a single keplerian model , as shown in figure 2
. the combined velocities from keck and aat ( figure 3 ) can also be fit with a single keplerian model , and yield a semiamplitude k = 57 , an orbital eccentricity @xmath12 = 0.05 , and yield a minimum ( @xmath1 ) mass of 0.34 m@xmath2 .
here we have adopted a stellar mass of 0.79 m@xmath16 ( mayor et al .
the actual stellar mass is probably closer to 1.0 m@xmath16 after accounting properly for the high metalicity of the star .
the rms to this keplerian fit for the combined keck and aat observations is 8.1 , and the corresponding reduced @xmath15 is 1.39 .
figure 4 shows the residuals to the single keplerian fit to the combined keck and aat data .
the rms of the residuals are 3.8 and 10.6 respectively for the keck and aat observations .
the keplerian orbital parameters derived from the separate and combined keck and aat observations are listed in table 2 , along with the orbital parameters from the geneva web page for both of the planets 83443 announced from the coralie data ] .
the orbital parameters for the inner planet , hd 83443b , derived from the keck and aat data sets are in good agreement with the coralie result , differing primarily in that the keck
aat orbit is nearly circular , within measurement uncertainty , as are the other extrasolar planets within 0.05 au .
the median internal uncertainty of the keck observations is 2.8 .
based on the ca ii h & k lines , we measure the chromospheric diagnostic @xmath17 of hd 83443 to be ,
log(r@xmath18 ) = -4.85 .
the doppler velocity `` jitter '' associated with this level of activity for a k0 v star is 3.0 ( saar et al .
1998 ; saar & fischer 2000 , santos et al .
2000b ) . adding the doppler `` jitter '' in quadrature with the measurement uncertainty of 3 produces an expected keplerian rms to the keck data of 4.1 ,
which is consistent with the observed rms of 3.8 .
the aat and keck data sets have independent and arbitrary velocity zero points .
the velocity offset between these two data sets was thus left as an additional free parameter in the combined keplerian fit . as this velocity zero point is dependent on the model used to fit the data , it is not possible to use the combined data set to search for multiple periodicities .
as the keck data has both better phase coverage and significantly higher precision than the aat data , we have intensely searched the keck velocity set for evidence of a second planet with a period of 29.83 d. however we also searched the aat velocities for the second planet , yielding similar results as from the keck data .
periodogram analysis ( scargle 1982 ; gilliland & baliunas 1987 ) reveals a strong periodicity near 3 days for both the keck and aat data sets .
figure 5a ( top panel ) shows the periodogram for the keck data .
the highest peak is the 2.986 period .
the dotted line is the 1% false alarm level .
there remain no other significant peaks , notably near 29.83 d. as a strong primary peak can hide secondary peaks ( butler et al .
1999 ) , we have removed the primary peak by subtracting off the best
fit keplerian from figure 3 .
figure 5b ( lower panel ) shows the periodogram of the keck velocity residuals from figure 4 .
no significant peaks remain .
we considered the possibility that a 29.8 d periodicity in our velocities , caused by an outer planet , might have been missed in the keck data due to the temporal sampling of velocity measurements .
the observational window function may cause blind spots at certain periods . to test this possibility , we constructed 1000 artificial velocity sets .
the fake velocities were calculated from the keplerian orbital parameters of both planets by simply adding the motion of the star caused by each planet .
we adopted the orbital parameters for both planets from mayor et al .
( 2000 ) , listed here in table 2 as planets `` b '' and `` c '' . in the simulation , we sampled the reflex velocity of the star at the times of the 20 keck observations listed in table 1 .
in addition , random noise with an rms of 4.0 was added to each of these artificial data set to simulate the combined effects of doppler `` jitter '' and the keck measurement errors .
each of these fake data sets was then fit with a single least squares keplerian , and the rms to this single keplerian fit was recorded .
a histogram of rms to the resulting single keplerian fits is shown in figure 6 .
the rms of these fits range from 6.7 to 14.0 .
the median rms to the single keplerian fit is 10.3 . in contrast , the rms of the single keplerian fit to the actual keck data is 3.8 , as indicated by the arrow in figure 6 .
since none of our 1000 artificial velocity sets could be adequately fit with a single keplerian model , the supposed outer planet , if it exists , would similarly not permit an adequate fit with a single keplerian model . thus there is less than 0.1% probability that the outer planet can hide in our actual velocities .
we conclude that the window function of the keck observations would not prevent the detection of the outer planet to hd 83443 .
such an outer planet , if it existed , should have caused an excess rms in the velocity residuals of @xmath910 when fit by single keplerian .
such velocity residuals are not seen .
it remains possible that the period of the outer planet of hd 83443 might be slightly different from that given on the coralie web page .
if this were so and the window function of the keck observations were unfortunately aligned , it might still be possible that the outer planet could be lurking in the keck data set . to test this
, we have fit the keck data set with a double keplerian , using as the input guess the double keplerian parameters from the coralie web page ( listed in table 2 ) .
the period , eccentricity , and velocity semiamplitude of the inner and outer inner planets have been frozen at the reported values of the supposed outer planet , but the remaining keplerian parameters including time of periastron and @xmath13 have been allowed to float .
outer planet periods ranging from 28 to 32 d were systematically attempted in steps of 0.001 d. figure 7 shows the resulting best fit reduced @xmath15 for each of the trial periods .
the arrow indicates the location of the 29.83 d period , which yields best fit reduced @xmath15 of 2.31 , much worse than the single keplerian fit to the keck data set with reduced @xmath15 of 1.33 . to estimate the largest semiamplitude allowed by the keck velocities for a planet in a @xmath930 d orbit
, we again fit the keck data with a double keplerian as in figure 7 , but this time allowed the velocity semiamplitude of the outer planet to float .
figure 8 shows this best
fit semiamplitude for a potential outer planet having orbital periods ranging from 28 to 32 d. at 29.83 d , the best
fit amplitude is 2.8 .
the keck data rule out any periodicities between 29 and 31 days with an amplitude greater than 3 .
given the temporal sampling of the keck data , it remains possible to hide a 13.8 semiamplitude with a period near 27.7 d from the current keck data set .
this is @xmath910 @xmath4 removed from the coralie period of 29.83 d.
precision doppler observations made with the 10m keck and the 3.9m aat strongly confirm the existence of the inner planet orbiting hd 83443 , and indicate the orbital parameters are in very good agreement with those reported by mayor et al .
the present orbital parameters differ only marginally in that the orbit of the inner planet is circular within measurement uncertainty for the keck and aat data , similar to the other known close in planets .
but our doppler measurements did not detect the 29.8 d outer planet , despite the clear ability to do so .
the present measurements impose a limit on any such velocity periodicity at a level of no more than 3 , well below the reported velocity amplitude of 13.8 .
orbital periods within 2 days of 29.8 d would have been detected .
the supposed velocity amplitude of 13.8 is four times larger than the uncertainties in our velocity measurements rendering the outer planet immediately detectable .
various tests suggest quantitatively that the velocities should have revealed the outer planet .
the velocities from the keck and aat telescopes could have independently detected the outer planet , but neither data set revealed it .
we considered various possible reasons that we failed to detect the outer planet .
one possibility is that some interactive resonance between the two planets causes the reflex velocity of the star to mimic insidiously a single keplerian orbit .
that is , perhaps the 10:1 ratio of orbital periods , along with gravitational interactions , yields a final reflex velocity that traces a single keplerian velocity curve .
if so , we might be fooled into fitting the velocities with such a simple model .
we find this possibility unlikely . as shown by w. benz ( mayor et al .
2000 ) and by wu & goldreich ( 2002 ) , the gravitational interactions yield temporal evolution of the orbits on a time scale of @xmath91000 yr rather than a few years .
thus we expect the outer planet , if it exists , to remain in a coherent orbit during the few year duration of the present observations .
moreover , the 10:1 ratio of the two periods do not constitute a powerful fourier harmonic from which a single keplerian may be constructed ( as is the case with a 2:1 ratio of periods ) . of the 75 extrasolar planet candidates ( @xmath1@xmath19 13 m@xmath2 ) announced from precision doppler surveys ] ,
a total of 57 have been published in refereed journals .
these planets are listed in table 3 , which also notes the telescope from which the data originates , and whether the actual doppler velocities are publicly available . refereed precision velocity confirmations are also included .
substellar candidates found by other techniques such as astrometry and low precision doppler velocities are not included .
an additional 4 planet candidates have been announced in conference proceedings ( queloz et al .
2000b ; sivan et al .
doppler velocity measurements are not available for candidates that have only been published in conference proceedings .
an additional 12 claimed doppler planets , all of which were announced more than 1 year ago , have not been submitted to either a conference proceeding or a refereed journal .
while the discovery of extrasolar planets has become seemingly commonplace over the past 6 years , we still consider the detection of planets orbiting other stars as extraordinary , and as such worthy of the dictum , `` extraordinary claims require extraordinary evidence . ''
publishing discovery data in a refereed journal remains a crucial part of the process , though this is not in itself sufficient to establish the credibility of a planet claim .
it remains extremely likely that at least a handful of the reported planets do not in fact exist .
multiple confirmation both by independent precision doppler teams and by completely independent techniques remain the only means by which to ensure the veracity of extrasolar planet claims .
we acknowledge support by nsf grant ast-9988087 , nasa grant nag5 - 12182 , and travel support from the carnegie institution of washington ( to rpb ) , nasa grant nag5 - 8299 and nsf grant ast95 - 20443
( to gwm ) , nsf grant ast-9619418 and nasa grant nag5 - 4445 ( to ssv ) , and by sun microsystems .
we thank the nasa and uc telescope assignment committees for allocations of keck telescope time , and the australian ( atac ) and uk ( patt ) telescope assignment committees for allocations of aat time .
we thank debra fischer , greg laughlin , doug lin , stan peale and man hoi lee for valuable conversations .
the authors wish to extend special thanks to those of hawaiian ancestry on whose sacred mountain of mauna kea we are privileged to be guests . without their generous hospitality , the keck observations presented herein would not have been possible .
butler , r. p. & marcy , g. w. 1997 , `` the near term future of extrasolar planet searches '' , brown dwarfs and extrasolar planets , held on tenerife , 17 - 21 march 1997 , ed .
r. rebolo , e.l .
martin , and m.r .
zapatero osorio , asp conference series , vol .
134 , p. 162 .
mayor , m. , naef , d. , pepe , f. , queloz , d. , santos , n. c. , udry , s. & burnet , m. 2000 , `` hd 83443 : a system with two saturns '' , planetary systems in the universe , held at the iau general assembly , manchester uk , 7 august 2000 , ed .
penny , p. artymowicz , a.m. lagrange , and s.s .
russell , asp conference series , in press .
mazeh , t. , naef , d. , torres , g. , latham , d. w. , mayor , m. , beuzit , j. l. , brown , t. m. , buchhave , l. , burnet , m. , carney , b. w. , charbonneau , d. , drukier , g. a. , laird , j. b. , pepe , f. , perrier , c. , queloz , d. , santos , n. c. , sivan , j. p , udry , s. , zucker , s. 2000 , 532 , l55 .
queloz , d. , mayor , m. , naef , d. , pepe , f. , santos , n. c. , udry , s. & burnet , m. 2000b , `` 4 jovian extrasolar planets detected with coralie '' , planetary systems in the universe , held at the iau general assembly , manchester uk , 7 august 2000b , ed .
penny , p. artymowicz , a.m. lagrange , and s.s .
russell , asp conference series , in press .
sivan , j. p. , mayor , m. , naef , d. , queloz , d. , udry , s. , perrier bellet , c. , & benzit , j. l. 2000 , `` a planetary companion to hd 190228 '' , planetary systems in the universe , held at the iau general assembly , manchester uk , 7 august 2000b , ed .
penny , p. artymowicz , a.m. lagrange , and s.s .
russell , asp conference series , in press .
rrrr [ vel83443 ] 212.1830 & -61.0 & 10.2 & aat + 213.1756 & -5.7 & 10.6 & aat + 682.9088 & 26.5 & 8.6 & aat + 898.0961 & 26.0 & 2.7 & keck + 899.0788 & -52.3 & 2.5 & keck + 900.0854 & 39.9 & 2.5 & keck + 901.0806 & 22.4 & 2.6 & keck + 919.2047 & 4.1 & 11.6 & aat + 920.1821 & -48.0 & 9.6 & aat + 971.9566 & 50.2 & 3.2 & keck + 972.9432 & -7.0 & 3.7 & keck + 974.8502 & 47.2 & 3.3 & keck + 981.9535 & -7.0 & 3.2 & keck + 982.9366 & -46.2 & 2.9 & keck + 983.0440 & -29.4 & 9.0 & aat + 984.0236 & 40.0 & 9.8 & aat + 1003.7982 & -40.1 & 3.1 & keck + 1006.9015 & -28.6 & 2.8 & keck + 1007.8096 & 52.1 & 2.3 & keck + 1009.0816 & -32.8 & 9.7 & aat + 1060.9180 & 0.5 & 7.8 & aat + 1062.7608 & -37.2 & 3.4 & keck + 1064.7379 & 62.8 & 2.7 & keck + 1091.8643 & 55.2 & 9.7 & aat + 1092.8878 & -57.8 & 8.4 & aat + 1127.8521 & 35.1 & 13.4 & aat + 1188.2812 & -33.9 & 5.0 & aat + 1189.2733 & -11.3 & 10.6 & aat + 1219.1297 & -14.7 & 2.5 & keck + 1236.1362 & -50.4 & 2.2 & keck + 1243.1498 & -3.1 & 3.0 & keck + 1307.9118 & -56.5 & 2.3 & keck + 1333.9629 & 21.6 & 2.3 & keck + 1334.8507 & -53.7 & 2.6 & keck + 1359.1162 & -28.7 & 7.4 & aat + 1360.1546 & 67.3 & 11.1 & aat + keck & 2.98571 ( 0.001 ) & 57.0 ( 4 ) & 0.059 ( 0.06 ) & 44 ( 40 ) & 1876.99 ( 0.15 ) & 0.35 & 0.0375 & 20 & 3.81 + aat & 2.98559 ( 0.0006 ) & 52.9 ( 5 ) & 0.00 & 0 & 1213.8 ( 0.1 ) & 0.32 & 0.0375 & 21 & 8.47 + keck aat & 2.98553 ( 0.0004 ) & 57.5 ( 2 ) & 0.052 ( 0.05 ) & 46 ( 30 ) & 1211.24 ( 0.1 ) & 0.34 & 0.0375 & 36 & 8.09 + coralie b & 2.9853 ( 0.0009 ) & 56.1 ( 1.4 ) & 0.079 ( 0.033 ) & 300 ( 17 ) & 1386.50 ( 0.14 ) & 0.34 & 0.0380 & 93 & 6 + coralie c & 29.83 ( 0.18 ) & 13.8 ( 1 ) & 0.42 ( 0.06 ) & 337 ( 10 ) & 1569.59 ( 0.73 ) & 0.16 & 0.17 & 93 & 6 + rrrllllll [ candid ] 217014 & 113357 & 51 peg & mayor & queloz 1995 & 1995 & n & elodie + ... & ... & ... & marcy et al .
1997 & 1996 sep 19 & y & lick + 117176 & 65721 & 70 vir & marcy & butler 1996 & 1996 jan 22 & n & lick + 95128b & 53721b & 47 uma b & butler & marcy 1996 & 1996 feb 15 & n & lick + ... & ... & ... & fischer et al .
2002 & 2001 mo da & y & lick + 120136 & 67275 & @xmath20 boo & butler et al .
1997 & 1996 aug 12 & n & lick + 75732b & 43587b & 55 cnc b & butler et al .
1997 & 1996 aug 12 & n & lick + 9826b & 7513b & @xmath21 and b & butler et al .
1997 & 1996 aug 12 & n & lick + ... & ... & ... & butler et al . 1999 & 1999
apr 8 & y & lick , afoe + 186408 & 96895 & 16 cyg b & cochran et al .
1997 & 1996 nov 21 & y & lick , mcdonald + 143761 & & hr 5968 & noyes et al .
1997 & 1997 apr 18 & n & afoe + & 113020b & gj 876 b & marcy et al .
1998 & 1998 jul 7 & n & lick , keck + & ... & ... & delfosse al .
1998 & 1998 aug 17 & n & elodie , coralie + & ... & ... & marcy et al .
2001b & 2000 dec 27 & y & lick , keck + 187123 & 97336 & & butler et al .
1998 & 1998 sep 6 & n & keck + ... & ... & & vogt et al . 2000 & 1999 nov 15 & y & keck + 195019 & 100970 & & fischer et al .
1999 & 1998 oct 8 & y & lick + ... & ... & & vogt et al .
2000 & 1999 nov 15 & y & keck + 217107 & 113421 & hr 8734 & fischer et al .
1999 & 1998 oct 8 & y & lick + ... & ... & ... & vogt et al . 2000 & 1999 nov 15 & y & keck + ... & ... & ... & naef .
2001a & 2000 aug 30 & y & coralie + 210277 & 109378 & & marcy et al .
1999 & 1998 dec 16 & y & keck + ... & ... & & vogt et al .
2000 & 1999 nov 15 & y & keck + ... & ... & & naef .
2001a & 2000 aug 30 & y & coralie + 168443b & 89844b & & marcy et al .
1999 & 1998 dec 16 & y & keck + ... & ... & & marcy et al .
2001a & 2000 dec 13 & y & keck + 9826c & 7513c & @xmath21 and c & butler et al .
1999 & 1999 apr 8 & y & lick , afoe + 9826d & 7513d & @xmath21 and d & butler et al .
1999 & 1999 apr 8 & y & lick , afoe + 13445 & 10138 & gl 86 & queloz et al . 2000 & 1999 apr 22 & n & coralie + ... & ... & ... & butler et al .
2001 & 2000 dec 25 & y & aat + 17051 & 12653 & @xmath22 hor & kurster et al .
2001 & 1999 oct 19 & y & eso + ... & ... & ... & naef .
2001a & 2000 aug 30 & y & coralie + ... & ... & ... & butler et al .
2001 & 2000 dec 25 & y & aat + 10697 & 8159 & & vogt et al .
2000 & 1999 nov 15 & y & keck + 37124b & 26381b & & vogt et al .
2000 & 1999 nov 15 & y & keck + ... & ... & & butler et al .
2002 & 2002 may 21 & y & keck + 222582 & 116906 & & vogt et al .
2000 & 1999 nov 15 & y & keck + 177830 & 93746 & & vogt et al .
2000 & 1999 nov 15 & y & keck + 134987 & 74500 & & vogt et al .
2000 & 1999 nov 15 & y & keck + ... & ... & & butler et al .
2001 & 2000 dec 25 & y & aat + 209458 & 108859 & & henry et al .
2000 & 1999 nov 18 & n & keck + ... & ... & & mazeh et al .
2000 & 1999 dec 3 & n & elodie , coralie + 130332 & 72339 & & udry et al .
2000 & 1999 dec 2 & n & coralie + 75289 & 43177 & & udry et al . 2000 & 1999 dec 2 & n & coralie + ... & ... & & butler et al .
2001 & 2000 dec 25 & y & aat + 89744 & 50786 & hr 4067 & korzennik et al .
2000 & 2000 jan 20 & n & afoe , lick + 16141 & 12048 & & marcy et al .
2000 & 2000 mar 6 & y & keck + 46375 & 31246 & & marcy et al .
2000 & 2000 mar 6 & y & keck + & & bd -103166 & butler et al .
2000 & 2000 apr 21 & y & keck + 52265 & 33719 & hr 2622 & butler et al .
2000 & 2000 apr 21 & y & keck + ... & ... & ... & naef .
2001a & 2000 aug 30 & y & coralie + 12661b & 9683b & & fischer et al .
2001 & 2000 jul 19 & y & lick , keck + 92788 & 52409 & & fischer et al .
2001 & 2000 jul 19 & y & lick , keck + 38529b & 27253b & & fischer et al .
2001 & 2000 jul 19 & y & lick , keck + 22049 & 16537 & @xmath23 eri & hatzes et al .
2000 & 2000 aug 22 & y & mcdonald , cfht , eso + 169830 & 90485 & & naef .
2001a & 2000 aug 30 & y & coralie + 1237 & 1292 & gj 3021 & naef .
2001a & 2000 aug 30 & y & coralie + 179949 & 94645 & & tinney et al .
2001 & 2000 oct 11 & y & aat , keck + 160691 & 86796 & hr 6585 & butler et al .
2001 & 2000 dec 25 & y & aat + 27442 & 19921 & hr 1355 & butler et al .
2001 & 2000 dec 25 & y & aat + & 113020c & gj 876 c & marcy et al .
2001b & 2000 dec 27 & y & lick , keck + 80606 & 45982 & & naef et al .
2001b & 2001 may 29 & y & coralie + 95128c & 53721c & 47 uma c & fischer et al .
2002 & 2001 jun 29 & y & lick + 28185 & 20723 & & santos et al .
2001 & 2001 jul 30 & y & coralie + 213240 & 111143 & & santos et al .
2001 & 2001 jul 30 & y & coralie + 4208 & 3479 & & vogt et al .
2002 & 2001 oct 16 & y & keck + 114783 & 64467 & & vogt et al .
2002 & 2001 oct 16 & y & keck + 4203 & 3502 & & vogt et al .
2002 & 2001 oct 16 & y & keck + 68988 & 40687 & & vogt et al .
2002 & 2001 oct 16 & y & keck + 33636 & 24205 & & vogt et al .
2002 & 2001 oct 16 & y & keck + 142 & 522 & hr 6 & tinney et al .
2002 & 2001 nov 12 & y & aat + 23079 & 17096 & & tinney et al .
2002 & 2001 nov 12 & y & aat + 39091 & 26394 & hr 2022 & jones et al .
2002 & 2001 nov 29 & y & aat + 108147 & 60644 & & pepe et al .
2002 & 2002 feb 26 & y & coralie + 168746 & 90004 & & pepe et al .
2002 & 2002 feb 26 & y & coralie + 141937 & 77740 & & udry et al .
2002 & 2002 feb 26 & y & coralie + 137759 & 75458 & @xmath22 dra & frink et al .
2002 & 2002 mar 21 & y & lick + 83443b & 47202 & & this paper & 2002 apr 8 & y & aat , keck + | the geneva group has reported two saturn mass planets orbiting hd 83443 ( k0v ) with periods of 2.98 and 29.8 d. the two planets have raised interest in their dynamics because of the possible 10:1 orbital resonance and the strong gravitational interactions .
we report precise doppler measurements of hd 83443 obtained with the keck / hires and the aat / ucles spectrometers .
these measurements strongly confirm the inner planet with period of 2.985 d , with orbital parameters in very good agreement with those of the geneva group .
however these doppler measurements show no evidence of the outer planet , at thresholds of 1/4 ( 3 ) of the reported velocity amplitude of 13.8 .
thus , the existence of the outer planet is in question .
indeed , the current doppler measurements reveal no evidence of any second planet with periods less than a year . |
last paper of the opera collaboration @xcite generates a lot of reactions and comments about the observation of the advance of 60.7 ns of the neutrino beam coming from the cern compared to the time of travel given by the speed of light .
one of these comments was given by contaldi @xcite , explaining the advance by an effect of general relativity for travelling atomic clocks .
three errors and misunderstanding by the author of this article @xcite are detailed here : one is an angular definition , the second is a missing term and third a misunderstanding of the method used by opera and cern to continuously resynchronize their atomic clocks with gps devices .
thus the supposed correction of @xmath2 @xcite is shown to be irrelevant .
c. r. contaldi in his paper arxiv:1109.6160 @xcite uses a polynomial expansion up to the quadrupolar term of the geoid model .
this type of geoid model is quite usual except the fact an error was introduced in the angular definition .
contaldi states : @xmath3 such a model describes a prolate spheroid ( rugby balloon shape ) , with an equatorial radius smaller than the polar radius ( see fig [ geoidcontaldi ] ) .
unfortunately , we knew for few centuries that the surface of the earth does not have this shape ( cf newton 1687 , huyghens 1690 , la condamine , bouguer and godin 1736 , clairaut 1736 , maupertuis 1742 , maclaurin 1742 ) .
the equatorial radius of the earth is greater than the radii at the poles .
the famous equatorial flatenning of the earth gives at this order of polynomial expansion an oblate shape to the earth .
in wgs-84 referential , the adopted value of the flattening of the earth is 1:298.257223563 .
thus a correct description of the earth would be @xcite : @xmath4
an effective gravitational potential of the earth including a centripetal contribution and a quadrupolar contribution ( see eq .
2 ) @xcite is described .
contaldi states : @xmath5+{{1 \over 2}(\omega_e r)^2 } \end{aligned}\ ] ] a mistake was done in the second term of this expression , as the radius @xmath6 used here is not the distance to the center of the earth but to the axis of rotation of the earth .
the centripetal correction to the effective potential should be null at the poles and is maximal at the equator of the earth .
this can be corrected inserting a cosine of the latitude ( or @xmath7 ) in the second term of the potential as follows : @xmath8+{{1 \over 2}(\omega_e r sin(\theta))^2 } \end{aligned}\ ] ] thus , assuming the following latitude for the cern : @xmath9 and for the gran sasso laboratory ( lngs ) : @xmath10 , we find : @xmath11 so while the author of @xcite found a variation of the effective potential correction term between cern and lngs @xmath12 , taking into account theses two errors we have computed here a value : @xmath13 , which is twice the value previously found by contaldi @xcite .
and last but not least , a profound conceptual mistake was done believing that an atomic clock or any timekeeper apparatus was carried in a journey by car or plane between cern to gran sasso as it was switched on and keeping the time during the travel . as explained by the opera collaboration @xcite the atomic clocks used either at cern or at gran sasso are continuously resynchronized through two gps devices at each location with the satellites .
moreover these two gps devices were independently calibrated by metas and ptb for common view mode time transfert @xcite .
consequently , the variation of potential term applies only for the time correction of the travel of the neutrino itself .
as the cern - lngs distance is about 730 km , this time corrections applies to 2.43 millisecond instead of 12 hours + 4 days at rest as mentioned in contaldi s scenario .
thus instead of a @xmath14 correction claimed by contaldi in a travel of 12 hours plus 4 days at rest for an atomic clock , we have found a time correction to be applied only on the neutrino travel itself @xmath1 .
of course , such a time correction is far away from the sensitivity of the opera experiment , and it can not explain any advance of the detected neutrinos of 60.7 ns as observed by opera @xcite .
three important errors are found in the paper of c.r .
contaldi @xcite trying to explain some time corrections in the synchronization of clocks for the neutrino beam travelling from cern to lngs as observed by the opera experiment .
one first error is due to a wrong definition of a polar angle related to the latitude in the calculation of the geoid model of the surface of the earth , leading to rugby ball shape instead of an oblate shape with an equatorial flattening .
a second error in the paper arxiv:1109.6160 , is due to the missing of a sine term of the polar angle in the centripetal correction term . and
third , a conceptual error appeared in believing that an atomic clock has been carried at the surface of the earth by car or plane during several hours or days between cern and lngs in order to synchronize the atomic clocks . instead of that , the atomic clocks are continuously synchronized thanks to gps satellites at a high level of precision ( see @xcite ) .
thus the time correction of @xmath15 claimed by contaldi , is reduced to @xmath1 .
so the mystery of the neutrino faster the speed of light remains complete .
i must thanks j. rich , b. vallage and g. vasseur for fruitful discussions about the measurement of the speed of the neutrino from cern observed at lngs by the opera collaboration . | we found three mistakes in the article `` the opera neutrino velocity result and the synchronisation of clocks '' by contaldi @xcite .
first , the definition of the angle of the latitude in the geoid description leads to a prolate spheroid ( rugby ball shape ) instead of an oblate spheroid with the usual equatorial flattening .
second , contaldi forgot a cosine of the latitude in the centripetal contribution term . and
last but not least , a profound conceptual mistake was done in believing that an atomic clock or any timekeeper apparatus was carried in a journey by car or plane between cern and gran sasso ; instead of that atomic clocks are continuously resynchronized through a gps device , and the variation of the potential term applies only for the neutrino travel itself . thus instead of a @xmath0 correction claimed by the author in a travel of 12 hours plus 4 days at rest for an atomic clock , we have found a time correction only for the neutrino itself @xmath1 ! that means , that this paper @xcite does not give the right explanation why the neutrino is seen travelling faster than the speed of light in the opera neutrino experiment . |
in this paper we present the results of a photometric survey for be stars within and around six young clusters in the magellanic clouds .
such studies have been carried out on galactic clusters but they are limited there by the relative paucity of be stars .
the young populous magellanic cloud clusters examined in this paper present us with large samples of be stars in a range of metallicity environments and the opportunity to further examine the mechanism for formation of be stars .
searches for be stars in the magellanic clouds have been made by feast ( @xcite ) using objective prism techniques , and the more recently by grebel et al .
( @xcite ) , bessell and wood ( @xcite ) ( who gave a brief description of some of the results presented in this paper ) and grebel ( @xcite ) using ccd - based imaging photometry .
these latter searches involved imaging through an h@xmath0 filter so that be stars , which show strong h@xmath0 emission , stand out in comparison to normal stars .
photometric techniques such as these are very efficient methods of identifying be stars , in particular within dense clusters where spectroscopy is difficult .
the main purpose of the present study is to provide a sample of be stars for follow - up spectroscopy and to see if there are differences in be star populations between different clusters and between clusters and the local field population .
the results of the follow - up spectroscopy will be presented in a subsequent paper .
because the present study is ground - based , its spatial resolution is seeing limited , which means that we are unable to search for be stars in the cluster cores . in another paper , we will present the results of a search for be stars in the cluster cores using hst imaging .
the present paper provides the comparison sample of be stars in the local field and the outer cluster zones .
the complete data set of ground - based and hst images will allow an accurate comparison of field and cluster be stars properties . in this paper , however , we make some initial comparisons with the more limited data set described herein .
the photometric search method we used to detect be stars involved taking a ccd image through a narrow - band ( 15 ) h@xmath0 filter and comparing this image with a similar image obtained in the cousins @xmath4 band .
stars with strong h@xmath0 emission should appear brighter in the h@xmath0 image than in the r image .
given that a typical be star has h@xmath0 emission with a full width at half maximum of @xmath57 and peak h@xmath0 flux 5 times the local continuum flux , this sort of search with a 15 filter is readily capable of detecting be stars .
the observations were obtained using direct ccd imaging on the 1-m telescope at siding spring observatory from september to november 1991 .
all h@xmath0 images were 900 second exposures .
@xmath4 images were obtained immediately before or after each h@xmath0 image in order to avoid the possibility of stellar variability affecting our measure of emission strength .
a tektronix 1024@xmath61024 ccd with pixels of scale 0.6 was used for these observations , giving us an imaged area of typically 10@xmath610 ( compared with [email protected] in grebel @xcite ) . @xmath1 and
@xmath3 images were obtained at the same time for some fields , and in some cases on other nights using a gec 770@xmath61152 ccd .
a log of the observations is presented in table [ obslog ] .
the ccd images were processed with iraf and the photometry of the fields was done using the dophot photometry package ( mateo and schechter @xcite ) .
standard magnitudes in @xmath1 and @xmath3 were derived from the standards used in sebo and wood ( @xcite ) and from magnitudes given by walker ( @xcite ) .
the @xmath4 and h@xmath0 magnitudes were not standardised and @xmath4@xmath2h@xmath0 colours have an arbitrary zeropoint .
.the log of observations presented in the current paper . [ cols="^,^,^,^",options="header " , ] [ brat ] in order to compare the be star fraction near the main - sequence turnoff with the be fraction for less evolved stars , we have divided the stars into two groups : the brighter stars with @xmath7 16 and the fainter stars with 17 @xmath8 16 , all stars having @xmath1@xmath2@xmath3 @xmath9 0.6 .
the be star fractions for these groups are given for the clusters and their surrounding fields in table [ brat ] . from these ratios , and looking at figs .
[ fig330bef ] to [ fig2100bef ] , it seems that for most of the clusters and fields there is no statistically significant difference between the be star fractions near the main - sequence turnoff and further down the main - sequence .
there are a few distinct exceptions to this rule : the clusters ngc 330 and ngc 2004 show a significantly higher fractions of be stars near the main - sequence turnoff than at fainter magnitudes , as does the field of ngc 1818 ( where the four brightest main - sequence stars are all be stars ) .
we now examine what our observations imply about the be mechanism . in the literature ,
binarity , pulsation and rotation have been suggested as the mechanism for the establishment and maintainence of the circumstellar material seen in be stars .
harmanec and k@xmath10i@xmath11 ( @xcite ) have suggested that be stars comprise a population of interacting binaries . in this model , elaborated upon by pols et al .
( @xcite ) , emission arises from material in accretion onto the secondary component .
the difficulty that remains with this model is that the expected number of close interacting binaries is too few to account for the observed population of be stars ( vanbever and vanbeveren @xcite ) .
our observations , particularly the high be star ratio in ngc 330 , do not favour the interacting binary hypothesis as the population of stars required to be in mass exchanging binary systems near the ms turnoff ( abt @xcite ) is too high .
the non - radial pulsator model for be activity ( vogt and penrod @xcite ; baade @xcite ) has gained support from the recent work of dziembowski et al .
( @xcite ) . in their models , dziembowski et al .
have shown that an instability domain exists on the main - sequence that almost fills the gap between @xmath12sct and @xmath13cep variables .
these slowly pulsating b - type stars span a range of mass from 3 to 6 @xmath14 .
the be star @xmath15cen has been intensively studied by rivinius et al .
( @xcite ) . from their analysis of photospheric radial velocity variations
they have discerned six discrete periods between 0.27 - 0.51 days .
rivinius et al .
suggest that when the majority of these modes are in constructive interference , expulsion of material and a line emission outburst occurs .
there are problems with associating the magellanic cloud be stars observed here with the pulsation instability mechanism described above .
firstly , in conditions of reduced metallicity such as those within the magellanic clouds , the extent of the pulsation instability domain is reduced ( dziembowski et al .
@xcite ) and @xmath13cep instability is seemingly extinguished ( balona @xcite ) .
most importantly , the be stars observed here are by no means confined to a range @xmath166 m@xmath17 .
indeed our observations show that the largest proportion of be stars occurs at @xmath51415 m@xmath17 .
the most widely accepted explanation for the be phenomenon is that it is a result of the rapid rotation of the b star .
this explanation was originally put forward by struve ( @xcite ) . in its modern incarnation , due to bjorkman and cassinelli ( @xcite ) ,
the stellar wind from a rapidly rotating star is concentrated and confined to a disk around the equatorial plane of the star .
the model of bjorkman and cassinelli predicts a maximum likelihood for disk formation near spectral type b2 ( where the stellar angular velocity required for disk formation is 55% that of the critical breakup velocity ) .
observations of galactic be stars find that the highest be star fraction does indeed occur at early spectral types ( b1 according to zorec and briot @xcite ) . within our sample
spectral types are available for only five of our flagged be candidates
. we will present further spectral types in a forthcoming paper .
assuming the be phenomenon is related to the formation of an equatorial disk caused by rapid rotation , we would expect that clusters which form from gas that has a high angular momentum content would contain a high fraction of be stars .
indeed , the evidence is suggestive that galactic clusters whose stars present a higher than average rotational velocity have larger populations of be stars ( mermilliod @xcite ) . in the magellanic clouds ,
the cluster ngc 330 has a noticeably ellipsoidal shape which may suggest that the cluster as a whole formed from material of higher angular momentum and that , as a consequence , the cluster stars have a rapid rotation rate .
this would certainly be consistent with the high be star content of ngc 330 . in a future paper
we will examine the rotation velocities of samples of stars within and around the clusters ngc 330 and ngc 2004 to see if the overall be star fraction is related to average cluster star rotation velocity .
the tendency for there to be a higher be star fraction near the main - sequence turnoff in some clusters also appears consistent with the rapid rotation hypothesis for be star formation .
the evolution of rapidly rotating stars has been examined by endal and sofia ( @xcite ) and endal ( @xcite ) .
these models explicitly follow the radial exchange of angular momentum with evolution . in the models of endal and sofia , stars commencing their main - sequence lives with relatively slow angular velocity remain slow rotators throughout the whole of the main - sequence evolutionary phase
( although there is a marked increase in angular velocity of all stars during the core contraction phase associated with exhaustion of hydrogen in the core , this evolutionary phase is far too short to account for the observed proportion of be stars ) .
thus slow rotators will never evolve into be stars .
in contrast to the slow rotators , the models of endal and sofia show that stars commencing their lives with an angular velocity greater than 56 - 76% of the critical breakup velocity @xmath18 spin up to the critical velocity over a moderate fraction of the main - sequence lifetime . in a cluster where some fraction of the stars is formed with angular velocity greater than @xmath550% of @xmath18 , those stars nearer the main - sequence turnoff are more likely to be be stars since they will have been more spun up by evolution than the less luminous stars that have not evolved as far through the main - sequence phase .
this process could explain the concentration of be stars to the main - sequence tip seen in ngc 330 , ngc 2004 and ngc 1818 .
we add the cautionary note that endal and sofia ( @xcite ) and endal ( @xcite ) made a detailed study of only a single star which had a mass of 5 m@xmath17 , much less than the turnoff masses of @xmath515 m@xmath17 which is appropriate for the clusters observed here .
for comparison with magellanic cloud data , models such as those of endal and sofia need to be made at higher masses and lower metallicities .
other factors which might influence the be star fraction are metallicity and age . however , neither of these factors seem to influence the be star fractions in our data . to examine abundance dependence
, we note that the two smc clusters are probably more metal poor than the lmc clusters by a factor of about two ( russell and bessell @xcite ) .
however , in the smc , ngc 330 has a high be star content while ngc 346 does not . at the higher abundance of the lmc , clusters with a range of be star content are present , and the range is similar to the range in the smc in spite of the abundance difference between smc and lmc . turning to age , we note that the clusters studied here have ages of @xmath513@xmath610@xmath19 years , although these ages are uncertain by factors of 23 because of uncertainties in the amount of convective core overshoot which occurs in real stars . for comparison of
be star fractions with age , we take from the literature the following ages : ngc 330 , 20 myr ( chiosi et al .
@xcite ) ; ngc 346 , 13 myr ( massey et al .
@xcite ) ; ngc 1818 , 30 myr ( will et al .
@xcite ) ; ngc 1948 , 15@xmath205 myr ( vallenari et al .
@xcite ) ; ngc 2004 , 20 myr ( caloi and cassatella @xcite ) ; ngc 2100 , 15 myr ( cassatella et al .
there does not appear to be any correlation of be star fraction with age : the cluster ngc 330 with the highest be star fraction has an intermediate age and the oldest cluster ( ngc 1818 ) has an intermediate be star fraction .
this result contrasts with the results of mermilliod ( @xcite ) who noted a trend of decreasing be fraction with increasing age amongst young galactic clusters .
we have presented the results of a photometric survey for be stars in six fields centred on ngc 330 and ngc 346 in the smc , and ngc 1818 , ngc 1948 , ngc 2004 and ngc 2100 in the lmc . in all six of these young populous clusters
we find large numbers of be stars .
the be stars are shown to be redder in @xmath1@xmath2@xmath3 than main - sequence stars of similar @xmath21 .
furthermore , the displacement to the red increases with increasing h@xmath0 emission line strength .
some suggested causes for the redness of be stars were examined using the widely accepted model that be stars are main - sequence stars with a surrounding disk .
our photometric observations suggest that the redness of the stars is due to emission from the circumstellar envelope , in line with the results of detailed spectrophotometric studies based upon galactic samples .
the fraction of main - sequence stars that are be stars varies significantly between the various clusters from @xmath50.100.34 .
the average value is close to the average fraction of @xmath50.17 seen in the galaxy , and the maximum value is also similar to the maximum galactic value , which occurs at spectral type b1 .
it was noted that the cluster with the maximum be star fraction , ngc 330 , is distinctly elliptical , suggesting that the stars in it formed out of rapidly rotating material and that there is a connection between rapid rotation and the be phenomenon .
no connection between be star fraction and age or metallicity was evident . in some clusters ,
the be stars tend to be concentrated toward the main - sequence turnoff .
this is consistent with evolutionary models which predict that stars which form with rotation velocities more than @xmath550% of the breakup value spin up as they evolve , reaching the breakup velocity over a moderate fraction of the main - sequence lifetime .
we thank alistair walker for making available his photometry prior to publication .
sck acknowledges the support of an australian postgraduate award scholarship .
we thank our referee eva grebel for her useful comments .
abt h.a . , 1987 , in iau
colloquium 92 , the physics of be stars , eds .
a. slettebak and t.p .
snow ( cup : cambridge ) , p.470 arp h. , 1959 , aj 64 , 245 baade d. , 1987 , in iau colloquium 92 , the physics of be stars , eds .
a. slettebak and t.p .
snow ( cup : cambridge ) , p.361 balona l.a .
1992 , mnras 256 , 425 balona l.a . , jerzykiewicz m. , 1993 , mnras 260 , 782 bessell m.s . , castelli f. , plez b. , 1998 , a&a 333 , 231 bessell m.s . , wood p.r . , 1992 , in new aspects of magellanic cloud research , eds
b. baschek , g. klare , j. lequeux ( springer : berlin ) , p.271 bjorkman j.e .
, cassinelli j.p . , 1993 , apj 409 , 429 caloi v. , cassatella a. , 1995 , a&a 295 , 63 cassatella a. , barbero j. , brocato e. , castellani v. , geyer e.h . , 1996 , a&a 306 , 125 chiosi c. , vallenari a. , bressan a. , deng l. , ortolani s. , 1995 , a&a 293 , 710 collins g.w . , truax r.j . , cranmer s.r . , 1991 , apjs 77 , 541 dachs j. , poetzel r. , kaiser d. , 1989 , a&as 78 , 487 dachs j. , engels d. , kiehling r. , 1988 , a&a 194 , 167 dziembowski w.a .
, moskalik p. , pamyatnykh a.a .
, 1993 , mnras 265 , 588 elson r.a.w .
, fall s.m . , freeman k.c . , 1987 ,
apj 323 , 54 endal a.s . , 1982 , in iau symposium 98 ,
be stars , eds .
m. jaschek , h .-
groth ( reidel : dordrecht ) , p. 299
endal a.s .
, sofia s. , 1979 , apj 232 , 531 feast m.w . , 1972 ,
mnras 159 , 113 grebel e.k .
, richtler t. , de boer k.s . , 1992 ,
a&a 254 , l5 grebel e.k . , 1997 ,
a&a 317 , 448 hanuschik r.w .
, dachs j. , baudzus m. , thimm g. , 1993 , a&a 274 , 356 harmanec p. , k@xmath10i@xmath11 s. , 1976 , in iau symposium 70 , be and shell stars , ed .
a. slettebak ( dordrecht : reidel ) , p.3 kaiser d. , 1989 , a&a 222 , 187 kjeldsen h. , baade d. , 1994 , in iau symposium 162 , pulsation , rotation and mass loss in early - type stars , eds .
balona , h.f .
hendrichs , j.m .
le contel ( kluwer : dordrecht ) , p.271 massey p. , parker j.w .
, garmany c.d .
, 1989 , aj 98 , 1305 mateo m. , schechter p. , 1989
, in proc .
1st eso / st - ecf data analysis workshop , eds .
f. murtagh , r.h . warmels ( eso : garching ) , p.69 mermilliod j .- c . ,
1982 a&a , 109 , 48 pols o.r . , cote j. , waters l.b.f.m . , heise j. , 1991 , a&a 241 , 419 rivinius t. , baade d. , stefl s. , wolf b. , kaufer a. , 1998 , a&a 336 , 177 robertson j.w . , 1974 ,
a&as 15 , 261 russell s.c .
, bessell m.s . , 1989 ,
apjs 70 , 865 sebo k.m . ,
wood p.r . , 1994 , aj , 108 , 932 struve o. , 1931 , apj 73 , 94 taylor b.j . , 1986 , apjs 60 , 577 tucholke h
.- j . , de boer k.s . , seitter w.c . , 1996 , a&as 119 , 91 vallenari a. , bomans d.j . ,
de boer k.s . , 1993 , a&a 268 , 137 van bever j. , vanbeveren d. , 1997 , a&a 322 , 116 vogt s.s , penrod g.d . , 1983 , apj 275 , 661 walker a. , 1996 , private communication will j - m . ,
bomans d.j .
, de boer k.s . , 1995 , a&a 295 , 54 zorec j. , briot d. , 1991 , a&a 245 , 150 zorec j. , briot d. , 1997 , a&a 318 , 443 | we present the results of a search for be stars in six fields centered on the young clusters ngc 330 and ngc 346 in the smc , and ngc 1818 , ngc 1948 , ngc 2004 and ngc 2100 in the lmc .
be stars were identified by differencing r band and narrow - band h@xmath0 ccd images .
our comparatively large images provide substantial be star populations both within the clusters and in their surrounding fields .
magnitudes , positions and finding charts are given for the 224 be stars found .
the fraction of be stars to normal b stars within each cluster is found to vary significantly although the average ratio is similar to the average be to b star ratio found in the galaxy . in some clusters ,
the be star population is weighted to magnitudes near the main sequence turn - off .
the be stars are redder in @xmath1@xmath2@xmath3 than normal main - sequence stars of similar magnitude and the redness increases with increasing h@xmath0 emission strength . |
multiresolution analysis of deterministic signals by the use of wavelets has been extensively studied in recent years . however , in the context of stochastic processes , general wavelet approximations has not yet been fully investigated . in the majority of cases developed deterministic methods and used error measures and metrics may not be appropriate to investigate wavelet representations of stochastic processes .
it indicates the necessity of elaborating special stochastic techniques . from a practical point of view
, multiresolution analysis provides an efficient framework for the decomposition of random processes .
wavelet representations could be used to convert the problem of analyzing a continuous - time random process to that of analyzing a random sequence , which is much simpler .
this approach is widely used in statistics to estimate a curve given observations of the curve plus some noise , in time series analysis for smoothing functional data , in simulation studies of various functionals defined on realizations of a random process , etc .
recently , a considerable attention was given to wavelet orthonormal series representations of stochastic processes .
numerous results , applications , and references on convergence of wavelet expansions of random processes in various spaces can be found in @xcite , just to mention a few .
figures 1 and 2 illustrate wavelet expansions of stochastic processes . a simulated realization of the process @xmath2 and its two wavelet reconstructions with different numbers of terms are plotted in figure 1 .
figure 2 displays boxplots of mean - square approximation errors for 500 simulated realizations of @xmath2 for each reconstruction .
figure 2 suggests that empirical probabilities of large errors become smaller when the number of terms in the wavelet expansion increases . although the mentioned effect is well - known for deterministic functions , it has to be established theoretically for different stochastic processes and probability metrics .
numerical simulation results need to be confirmed by theoretical analysis .
it is also important to obtain theoretical estimations of the rate of convergence for various stochastic wavelet expansions .
[ fig1 ] our focus in this paper is on convergence in @xmath0)$ ] of wavelet expansions of @xmath1-sub - gaussian random processes .
the paper extends the recent results on uniform convergence in the papers @xcite to new classes of stochastic processes and probability metrics .
the analysis and the approach presented in the paper contribute to the investigations of wavelet expansions of random processes in the former literature .
the approach is of a special interest if @xmath3 as it extends the available @xmath4 results . in that sense ,
theorems [ conv_lp_sub_ner]-[str_sub_covlp_nonstation ] are of special importance .
the results are obtained under simple assumptions which can be easily verified .
the paper deals with the most general class of wavelet expansions in comparison with particular cases considered by different authors , see , for example , @xcite .
the organization of the article is the following .
in the second section we introduce the necessary background from the theory of @xmath1-sub - gaussian random variables and processes .
section 3 discusses wavelet expansions and approximations of stochastic processes . in
4 we present the main results on convergence in @xmath0)$ ] of wavelet expansions of @xmath1-sub - gaussian random processes . in this section
we also obtain the rate of convergence of the wavelet expansions and discuss some specifications for which the assumptions in the theorems can be easily verified .
in this section , we review the definition of @xmath1-sub - gaussian random processes and their relevant properties . the space of @xmath1-sub - gaussian random variables was first introduced in the paper @xcite .
more information about the space of @xmath1-sub - gaussian random variables and processes can be found in @xcite .
( @xcite ) a continuous even convex function @xmath5 @xmath6 is called an orlicz n - function , if it is monotonically increasing for @xmath7 , @xmath8 @xmath9 when @xmath10 and @xmath11 when @xmath12 * assumption q * let @xmath13 be an orlicz n - function and @xmath14 the constant @xmath15 can be equal to @xmath16 ( @xcite ) every n - function @xmath13 can be represented as @xmath17 where @xmath18 is a monotonically nondecreasing , right - continuous function , such that @xmath19 and @xmath20 when @xmath21 the function @xmath18 is called a density of @xmath22 ( @xcite ) let @xmath23 be an orlicz n - function .
the function @xmath24 @xmath6 defined by the formula @xmath25 is called the young - fenchel transform of @xmath22 let @xmath26 be a standard probability space and @xmath27 denote a space of random variables having finite @xmath28-th absolute moments .
( @xcite ) let @xmath13 be an orlicz n - function satisfying the assumption @xmath29 a zero mean random variable @xmath30 belongs to the space @xmath31 ( the space of @xmath1-sub - gaussian random variables ) , if there exists a constant @xmath32 such that the inequality @xmath33 holds for all @xmath34 ( @xcite ) the space @xmath31 is a banach space with respect to the norm @xmath35 a gaussian centered random variable @xmath30 belongs to the space @xmath36 where @xmath37 and @xmath38 ( @xcite ) let @xmath39 be a parametric space .
a random process @xmath40
@xmath41 belongs to the space @xmath42 if @xmath43 for all @xmath44 and @xmath45 let us denote @xmath46 ( @xcite ) a family @xmath47 of random variables @xmath48 is called strictly @xmath42 if there exists a constant @xmath49 such that for any finite set @xmath50 @xmath51 @xmath52 and for arbitrary @xmath53 @xmath54 @xmath55 @xmath56 is called a determinative constant .
[ def_strictly sub ] ( @xcite ) @xmath1-sub - gaussian random process @xmath57 @xmath58 is called strictly @xmath42 if the family of random variables @xmath59 is strictly @xmath60 the determinative constant of this family is called a determinative constant of the process and denoted by @xmath61 .
gaussian centered random process @xmath57 @xmath58 is a strictly @xmath42 process , where @xmath37 and the determinative constant @xmath62 [ ner_xsub_exp](@xcite ) let @xmath63 be a measurable space and @xmath57 @xmath58 be a random process from the space @xmath60 if @xmath64 for some @xmath65 then the integral @xmath66 exists with probability 1 and the following inequality holds @xmath67 for each non - negative @xmath68 where @xmath69 and @xmath18 is a density of @xmath22 if @xmath70 @xmath71 then @xmath72 and @xmath73 where @xmath74 and @xmath75 hence , @xmath76 therefore , inequality ( [ epsilon_ner ] ) holds if @xmath77 if @xmath57 @xmath58 is a gaussian centered random process and @xmath78 then inequality ( [ epsilon_ner ] ) holds true for @xmath79 where @xmath80 for such @xmath81 the following estimate is valid @xmath82 if @xmath57 @xmath58 is a strictly @xmath42 random process , then @xmath83
in this section we introduce wavelet representations and approximations of non - random functions and stochastic processes .
let @xmath84 @xmath85 be a function from the space @xmath86 such that @xmath87 and @xmath88 is continuous at @xmath89 where @xmath90 is the fourier transform of @xmath91 suppose that the following assumption holds true : @xmath92 there exists a function @xmath93)$ ] , such that @xmath94 has the period @xmath95 and @xmath96 in this case the function @xmath97 is called the @xmath98-wavelet .
let @xmath99 be the inverse fourier transform of the function @xmath100 then the function @xmath101 is called the @xmath102-wavelet .
let @xmath103 it is known that the family of functions @xmath104 is an orthonormal basis in @xmath86 ( see , for example , @xcite ) .
an arbitrary function @xmath105 can be represented in the form @xmath106 @xmath107 the representation ( [ 2.5 ] ) is called a wavelet representation .
the series ( [ 2.5 ] ) converges in @xmath86 i.e. @xmath108 the integrals @xmath109 and @xmath110 may also exist for functions from @xmath111 and other function spaces .
therefore it is possible to obtain the representation ( [ 2.5 ] ) for function classes which are wider than @xmath86 ( see , for example , @xcite ) .
* assumption s * ( @xcite ) for the function @xmath97 there exists a decreasing function @xmath112 @xmath113 such that @xmath114 @xmath115 ( a.e . ) , and the integral @xmath116 is finite .
let @xmath117 @xmath118 be a random process such that @xmath119 for all @xmath120 .
if sample trajectories of this process are in the space @xmath86 with probability one , then it is possible to obtain the representation ( wavelet representation ) @xmath121 with mean - square integrals @xmath122 however , the majority of random processes does not possess the required property .
for example , sample paths of stationary processes are not in the space @xmath86 ( a.s . ) . however
, in many cases it is possible to construct a representation of type ( [ x(t ) ] ) for @xmath123 consider the approximants of @xmath2 defined by @xmath124 where @xmath125 for the sake of simplicity , we will omit the index @xmath126 in @xmath127 and will use the notation @xmath128 to denote that @xmath129 @xmath130 and @xmath131 for all @xmath132 theorem [ 8103 ] below guarantees the mean - square convergence of @xmath133 to @xmath123 [ 8103 ] ( @xcite ) let @xmath117 @xmath118 be a random process such that @xmath134 @xmath135 for all @xmath118 and its covariance function @xmath136 is continuous . let the @xmath98-wavelet @xmath137 and the @xmath102-wavelet @xmath138 be continuous functions and the assumption s hold true for both @xmath137 and @xmath139 suppose that there exist a function @xmath140 and @xmath141 such that @xmath142 for all @xmath143 assume that there exists an even non - decreasing on @xmath144 function @xmath145 @xmath85 with @xmath146 such that @xmath147 for all @xmath148 then 1 .
@xmath149 2 .
@xmath150 in mean square when @xmath151 it was shown that the integrals in ( [ intksieta ] ) exist if @xmath2 satisfies the assumptions of theorem [ 8103 ] , see @xcite .
though there are other sufficient conditions in the literature which guarantee the existence of the integrals in ( [ intksieta ] ) and the convergence in the mean - square sense in theorem [ 8103 ] , the assumptions in terms of the functions @xmath152 and @xmath153 can be easily verified in many practical examples .
in this section we present the main results on convergence in @xmath154 @xmath155,$ ] @xmath156 of the wavelet expansions of @xmath1-sub - gaussian random processes .
the rate of convergence in the space @xmath0)$ ] is obtained .
we also present some specifications of the general results for which the assumptions can be easily verified . in @xcite
, it was shown that there always exists a constant @xmath157 such that @xmath158 therefore , random processes from the space @xmath42 belong to the space @xmath159 the following theorem is a corollary of theorem [ 8103 ] and the estimate @xmath160 [ meansq_subfi ] let @xmath117 @xmath118 be a random process such that @xmath161 and @xmath2 is continuous in the norm @xmath162 let the functions @xmath163 @xmath164 @xmath165 @xmath166 and @xmath152 satisfy the assumptions of theorem [ 8103 ] . if @xmath167 for all @xmath118 then 1 .
@xmath149 2 .
@xmath150 in mean square when @xmath151 a random process @xmath57
@xmath168 belongs to the space @xmath0)$ ] if @xmath169 [ conv_lp_sub_ner ] let @xmath117 @xmath118 be a random process such that @xmath161 and @xmath2 is measurable and continuous in the norm @xmath162 let the assumptions of theorem [ meansq_subfi ] are satisfied .
suppose that @xmath170 for some @xmath171 and @xmath172 if @xmath173 when @xmath174 then @xmath175 in the space @xmath0)$ ] in probability , when @xmath176 i.e. for all @xmath177 @xmath178 when @xmath179 . furthermore , the following estimate holds @xmath180 where @xmath181 and @xmath182 _ proof .
_ first , we will show that the random variables @xmath183 and @xmath184 are from the space @xmath60 we prove it only for @xmath185 the case of @xmath183 can be dealt with similarly .
we will need the following generalization of the minkowski inequality .
[ gmink ] if @xmath117 @xmath118 is a measurable @xmath1-sub - gaussian random processes , then @xmath186 _ proof .
_ without loss of generality we may assume that @xmath187 for all @xmath120 and @xmath188 let us denote @xmath189 notice that @xmath190 is a convex function and @xmath191 hence , by jensen s inequality , we obtain @xmath192 therefore , @xmath193 by the definition of @xmath194 we get @xmath195 by inequalities ( [ i ] ) and ( [ est ] ) we obtain @xmath196 finally , the statement of the lemma follows from the definition of @xmath162 @xmath197 by lemma [ gmink ] we get @xmath198 since @xmath199 for sufficiently large @xmath200 . thus , the processes @xmath201 belongs to the space @xmath60 to prove the theorem , it is enough to show that for all @xmath177 @xmath202 when @xmath203 and @xmath204 indeed , by theorem [ ner_xsub_exp ] and ( [ inttau ] ) , @xmath205)$ ] with probability 1
. if the condition ( [ p_int_xn - xm1 ] ) is satisfied then there exists a process @xmath206)$ ] such that @xmath207 when @xmath208 and by theorem [ meansq_subfi ] we get @xmath209 with probability 1 . to prove ( [ p_int_xn - xm1 ] ) we denote @xmath210 by theorem [ ner_xsub_exp ] the following inequality holds @xmath211 for @xmath212 the condition ( [ inttau_xn - xm ] ) implies that @xmath213 as @xmath214 the function @xmath215 is right - continuous and @xmath216 hence @xmath217 therefore , for arbitrary @xmath177 and sufficiently large @xmath218 inequality ( [ intxn - xm_ner_cnm ] ) holds true . by definitions 1 and 2 the right - hand side of the inequality vanishes when @xmath219 which implies ( [ p_int_xn - xm1 ] ) .
notice , that @xmath220 in @xmath0)$ ] with probability 1 and @xmath221 when @xmath204 hence , we obtain inequality ( [ p_int_xn - xm < exp ] ) if we let @xmath222 in ( [ intxn - xm_ner_cnm ] ) approach infinity
. @xmath197 [ cor1_str_sub ] let @xmath117 @xmath118 be a strictly @xmath42 random process with a determinative constant @xmath223 then the condition ( [ inttau ] ) holds true if the integral @xmath224 is convergent . if @xmath225 then the condition ( [ inttau_xn - xm ] ) holds when @xmath226 the inequality ( [ p_int_xn - xm < exp ] ) is valid for @xmath227 _ proof . _
the statement of corollary [ cor1_str_sub ] follows from theorem [ conv_lp_sub_ner ] , definition [ def_strictly sub ] , and the fact that a linear closure of a family of strictly @xmath42 random variables is a family of strictly @xmath228 see @xcite . @xmath197
[ cor2_str_sub ] the statement of corollary [ cor1_str_sub ] is valid if instead of ( [ int_xn - xm^2 ] ) the series @xmath229 converges uniformly on the interval @xmath230.$ ] then the inequality ( [ p_int_xn - xm < exp ] ) holds for @xmath231 _ proof . _
the statement of corollary [ cor2_str_sub ] follows from ( [ c_n^infty ] ) and the estimate @xmath232 @xmath233 @xmath234 @xmath197 [ lem_ner_phi ] if @xmath235 is an m - wavelet or an f - wavelet satisfying the assumption @xmath236 then @xmath237 and @xmath238 for @xmath239 _ proof
. _ inequality ( [ sum|delta| ] ) is a simple modification of an inequality from @xcite .
therefore we only prove ( [ supsum|delta| ] ) . by the assumption @xmath236 @xmath240 where @xmath241 is an even function .
as @xmath153 is a decreasing function on @xmath242 and @xmath243 we obtain @xmath244 @xmath197 let @xmath117 @xmath118 be a strictly @xmath42 random process for which the assumptions of theorem [ 8103 ] hold true and @xmath245 if @xmath246 then @xmath247 in the space @xmath248 in probability , when @xmath249 furthermore , for @xmath250 inequality ( [ p_int_xn - xm < exp ] ) is valid when @xmath251 where @xmath252 _ proof . _ notice that @xmath253 when @xmath249 by the choice of @xmath254 and ( [ supsumksieta^2 ] ) we get @xmath255 when @xmath249 the statement of the theorem follows from theorem [ conv_lp_sub_ner ] , corollary [ cor2_str_sub ] , lemma [ lem_ner_phi ] , and the fact @xmath256 @xmath197 [ str_sub_covlp4 ]
let @xmath117 @xmath118 be a strictly @xmath42 stationary short - memory random process with the covariance function @xmath257 let the assumptions of theorem [ 8103 ] are satisfied and @xmath245 suppose that there exists the fourier transform @xmath258 and for some @xmath259 @xmath260 if @xmath261 is a lipschitz function of order @xmath262 then ( [ supsumksieta^2 ] ) holds .
_ by parseval s theorem and the representation @xmath263 we deduce @xmath264 @xmath265 by the lipschitz conditions , ( [ st ] ) , and @xmath266 we obtain that for all @xmath267 @xmath268 where @xmath269 is the lipschitz constant .
similarly to ( [ st ] ) we deduce @xmath270 the integral in ( [ st2 ] ) is finite because the assumption s implies @xmath271 and therefore @xmath272 is bounded . by estimates ( [ st1 ] ) and ( [ st2 ] ) the series in ( [ supsumksieta^2 ] ) is convergent .
@xmath197 [ str_sub_covlp_nonstation ]
let @xmath117 @xmath118 be a strictly @xmath42 random process with the continuous covariance function @xmath273 suppose that @xmath274 there exist @xmath275 and for some @xmath259 @xmath276 if the assumptions of theorem [ 8103 ] are satisfied and @xmath261 is a lipschitz function of order @xmath277 then ( [ supsumksieta^2 ] ) holds true .
_ by parseval s theorem , @xmath278 by properties of the @xmath102-wavelet @xmath99 we have @xmath279 therefore , using the lipschitz conditions , we obtain @xmath280 where @xmath269 is the lipschitz constant . analogously to ( [ st1_2 ] ) we deduce that @xmath281 due to the boundedness of @xmath282 an application of the above estimates gives the convergence of ( [ supsumksieta^2 ] )
. @xmath197 conditions of theorems [ str_sub_covlp4 ] and [ str_sub_covlp_nonstation ] on the random process @xmath2 are formulated in terms of its spectral density .
these conditions are related to the behavior of the high - frequency part of the spectrum .
such assumptions are standard in the convergence studies of stochastic approximations .
if the assumptions of theorem [ str_sub_covlp4 ] or [ str_sub_covlp_nonstation ] are satisfied , then ( [ p_int_xn - xm < exp ] ) holds true for @xmath283 given by ( [ cn+ ] ) .
all the terms in ( [ cn+ ] ) can be easily computed in practice for specific stochastic processes and wavelet bases .
the obtained results may have various practical applications for the approximation and simulation of random processes .
the analysis of the rate of convergence provides a constructive algorithm for determining the number of terms in the wavelet expansions to ensure the approximation of stochastic processes with given accuracy . the developed methodology and results are important extensions of the recent findings in the wavelet approximation theory of stochastic processes to the space @xmath0)$ ] and the class of @xmath1-sub - gaussian random processes .
this class plays an important role in generalizations of various theoretical properties of gaussian processes .
in addition to classical applications of @xmath1-sub - gaussian random processes in signal processing , the results can also be used in new areas , like compressed sensing and actuarial modelling , consult , for example , @xcite . * to adopt and specify the results for different wavelet bases , which satisfy the assumptions of the theorems , for example , daubechies , battle - lemarie and meyer wavelets ; * to examine the tightness of the estimates by simulations .
vershynin r ( 2012 ) introduction to the non - asymptotic analysis of random matrices . in eldar
y , kutyniok g ( ed ) compressed sensing , theory and applications .
cambridge university press , cambridge , pp 210 - 268 | the article presents new results on convergence in @xmath0)$ ] of wavelet expansions of @xmath1-sub - gaussian random processes .
the convergence rate of the expansions is obtained .
specifications of the obtained results are discussed . |
the monitor project is a large - scale photometric survey of galactic open clusters and star forming regions . we intend to measure high - cadence time series photometry for @xmath0 cluster members over @xmath1 of sky , aiming to find the first transiting planets in open clusters , and tens - hundreds of low - mass eclipsing binary systems , possibly including brown dwarfs . for more details of the project s scientific goals and the results of simulations giving likely numbers of detected systems ,
the reader is referred to @xcite , hereafter paper i. a brief summary of the project is also given in @xcite .
data processing in this project is challenging . in a typical night
we obtain @xmath9 of imaging data using the wide field camera ( wfc ) on the isaac newton telescope ( int ) , and this can be as large as @xmath10 for some of the other instruments we are using ( for example megacam on the canada - france - hawaii telescope , hereafter cfht ) . since our survey covers @xmath11 clusters over @xmath12 nights per cluster , this is a multi - terabyte project . @xcite
give a detailed discussion of differential photometry problems and techniques , from the point of view of attempting to detect low - amplitude stellar oscillations , but many of their arguments apply equally to transit surveys .
using ccd cameras , one can perform differential photometry on very large numbers of stars simultaneously , using non - variable stars in the field as comparison sources to remove transparency ( and other ) variations in the atmosphere .
differential photometric precision at the sub-@xmath13 level can be readily achieved using this method , even in somewhat non - photometric conditions .
our methodology is based on experience gained by members of our group from the university of new south wales extrasolar planet survey @xcite , and much of the pipeline code is now shared between the two projects .
we describe the observations in [ obs_section ] and the basic ccd data reduction in [ dr_section ] . [ dp_section ] gives an overview of the steps required to produce differential photometry , and hence lightcurves , from these data , and the practical details of their implementation are discussed in [ implement_section ] and [ lc_section ] . in
[ noise_section ] we examine the noise properties of our data , with particular attention given to correlated ( ` red ' ) noise , which can be a serious problem in differential photometry @xcite . [ seeing_section ] examines one particular source of correlated noise , namely seeing - correlated variations induced in the lightcurves by blending of flux from neighbouring sources into the photometric apertures , and in
[ sysrem_section ] we apply the sysrem algorithm @xcite to search for any further sources of correlated noise in the data .
finally , we summarise our conclusions in [ conc_section ] .
we are using wide - field mosaic cameras on several telescopes to perform the survey , principally : the wide field camera ( wfc ) on the @xmath14 int ( @xmath15 2k@xmath164k ccds , @xmath17 field - of - view ) and megacam on the @xmath18 cfht ( @xmath19 [email protected] ccds , @xmath20 fov ) in the northern hemisphere , and the eso / mpg @xmath21 wide field imager ( wfi ) ( @xmath22 2k@xmath164k ccds , @xmath23 fov ) and mosaic ii on the @xmath6 ctio blanco telescope ( @xmath22 2k@xmath164k ccds , @xmath24 fov ) in the southern hemisphere . due to the enormous quantity of data , a uniform strategy for observing ( where possible ) and data processing is essential .
the peculiarities of scheduling for each of these telescopes limit our flexibility in observing strategy so this will be discussed only briefly .
we observe in @xmath25 or @xmath26 , since this maximises signal - to - noise for our faint , red objects of interest , and minimises any colour - dependent atmospheric extinction , which can be difficult to correct in the lightcurves .
the wide - field mosaic instruments we are using typically suffer from fringing in red bandpasses , so the sdss - like @xmath25 filter @xcite is preferred where available , since this minimises fringing due to its sharp red cut - off at @xmath27 , compared to the long red tail of the standard @xmath26 filters .
exposure times are selected to give good signal to noise on the largest possible number of cluster members , while keeping the targets sufficiently bright that medium - resolution follow - up observations on @xmath6 class telescopes and high - precision radial velocities on @xmath28 class telescopes remain feasible .
typically our exposures are in the range @xmath29 , so the survey efficiency is overhead dominated with the slow readout times for the mosaic instruments we are using ( most are @xmath30 ) . in several cases we cycle between multiple fields in a single cluster to increase our spatial coverage , or between multiple clusters , but we aim to obtain an observing cadence no worse than @xmath31 for clusters where we are primarily searching for eclipsing binaries , and @xmath32 for planet searches , or where short - term stellar variability is a problem , ie . the youngest clusters ( see paper i for more details ) .
accurate flat fielding is of critical importance in differential photometry , so we take extra care to ensure that this is done as well as possible .
we find that twilight flat fields provide superior results compared to dome flat fields for all the instruments we are using , provided sufficient signal can be accumulated . for a typical detector with gain of a few @xmath33 , and a typical twilight flat illumination level of @xmath34 ,
the poisson noise is @xmath35 , ie
. a signal - to - noise ratio of @xmath36 , which is equivalent to @xmath37 photon noise per pixel . averaged over a typical photometric aperture of @xmath38 radius
this gives @xmath39 ie .
a significant contribution . over a typical one week observing run , we can readily obtain at least @xmath40 flat field frames , which reduces the poisson noise to @xmath41 , a level which is perfectly acceptable for our purposes .
a related issue is that of positioning the telescope . even using the flat fielding procedure described , small errors of the order of @xmath42 remain in the flat field frames , and fringing in the detectors , even after correction , can reach amplitudes of @xmath43 .
the effects can be divided into low spatial frequencies , dominated by non - uniform illumination of the flat field frame , and high spatial frequencies , eg .
fringing , or differential variations in the quantum efficiency of the pixels ( eg . as a function of wavelength , since the spectra of the flat field source and target star are different ) .
the combination of these effects typically limits the achievable photometric precision to a few mmag depending on the instrument , in our experience . in order to minimise these effects
we therefore aim to reposition each star on _ exactly _ the same pixel of the detector in each exposure .
this is done by using the telescope guiding system to correct for pointing errors , where available .
we note in passing that this procedure may introduce correlated noise ( see [ noise_section ] ) , particularly in the event that any positioning errors are periodic or result in a slow drift across a few pixels of the detector .
it has been suggested that an intentional random jitter in the telescope positions may prove beneficial to convert this source of correlated noise to a source of random noise .
however , due to the need to move over a larger region of the detector , doing this is likely to introduce greater effects due to flat fielding errors , fringing , and other effects operating over short spatial scales .
it therefore carries an inherent risk of raising the overall noise level , and thus would require more data , so we have been unable to explore it further as telescope time is always at a premium when using large international facilities . equatorial standard star fields ( from the catalogue of @xcite )
are observed regularly during our observing runs , to provide calibrated photometry on a standard zero - point system .
the need for a uniform data processing strategy was highlighted in [ obs_section ] .
we employ a modified version of the int / wfc data reduction pipeline , developed for the int wide field survey ( wfs ) and originally described in @xcite .
this has been successfully applied to data from all of the instruments mentioned in [ obs_section ] at the time of writing .
two of the instruments we are using ( int / wfc and ctio mosaic ) suffer from electrical cross - talk between the detector readouts , the effect of which is illustrated in figure [ m34_crosstalk ] .
for the int / wfc the maximum level is @xmath44 , typically a sufficiently low level to be ignored , but for the ctio mosaic the level is @xmath45 . therefore , before starting the standard ccd reduction procedure this must be corrected , and is done in a simple manner by subtracting a fraction @xmath46 of the detected counts on detector @xmath47 from detector @xmath48 . ) from ccd 4 to ccd 3 ( eg . at the position in the top panel corresponding to the pair of bright stars visible at the right of the bottom panel , marked with arrows ) , and negative cross - talk ( @xmath49 ) from ccd 3 to ccd 4 ( eg . at the position in the bottom panel corresponding to the brightest star on the left hand side of the top panel , marked with an arrow).,width=307 ] we then follow the standard ccd reduction scheme of bias correction , trimming of overscan and non - illuminated regions , non - linearity correction , flatfielding and gain correction , followed by defringing , catalogue generation , astrometric and photometric calibration described in @xcite .
we use the point source catalogue ( psc ) from the two - micron all sky survey ( 2mass ) as an astrometric reference catalogue , which we find gives typical rms residuals of @xmath50 .
in the discussion that follows , we use aperture photometry . the technique we use is similar to standard aperture photometry , except our apertures are ` soft - edged ' , and overlapping sources are fitted simultaneously using circular top - hat functions as the ` psf ' .
we have found that for our open cluster fields , this technique is sufficient to obtain a photometric precision of @xmath51 for the brightest stars , without need to invoke more exotic techniques such as point spread function fitting ( psf - fitting ; eg .
@xcite ) or difference image analysis ( dia ; @xcite , @xcite ) , although these are discussed briefly in [ seeing_section ] . robust , repeatable background estimation is of vital importance in aperture photometry .
we use a variant of the technique discussed in @xcite for background estimation in our aperture photometry , which has been found empirically to work at least as well as the standard technique of using an annulus around the photometric aperture , for fields with slowly - varying sky backgrounds .
a brief description of the method is given here , and the reader is referred to @xcite , @xcite and @xcite for a more detailed discussion .
briefly , the image is divided into a coarse grid of @xmath52 bins ( @xmath53 on sky ) .
the background level in each bin is estimated using a robust @xmath54 clipped median of the counts in that bin , using the robust median of absolute deviations ( mad ; eg .
@xcite ) estimator to calculate @xmath55 , and rejecting bad pixels using the confidence maps ( see @xcite ) .
the resulting map is filtered using 2-d bilinear and median filters to avoid problems due to single bins dominated by bright stars .
the background in a given image pixel can then be estimated using bilinear interpolation over the coarse background map .
differential photometry is very sensitive to small positioning errors when placing photometric apertures on the science images . for a gaussian psf
, the error in the derived fluxes is given to first order by : @xmath56 where @xmath57 is the positioning error , @xmath58 is the radius of the aperture , and @xmath55 describes the psf size ( ie .
seeing , @xmath59 ) .
see appendix [ applac_deriv ] for a derivation .
typically we set @xmath60 , ie .
an aperture radius equal to the image fwhm , so : @xmath61 taking for example a typical value @xmath62 , this implies a flux error of @xmath63 .
( [ applac_equation ] ) also confirms the intuitive result that using a larger aperture reduces the effect of centroid errors , at the cost of increased noise from the sky background .
we therefore first consider the question of how best to determine the correct locations for the apertures . the ` default ' technique used by existing source extraction software , as included in our pipeline ( @xcite , @xcite ) , or sextractor @xcite , is to find the centroid of each star on the ccd frame in question , to place an aperture at this position , and measure the flux .
the accuracy to which this can be done for a star measured with signal to noise ratio @xmath64 improves in proportion to @xmath65 ( eg .
@xcite ) , giving the general ` rule of thumb ' that the error in the image centroid is @xmath66 where @xmath67 is the sampling interval ( pixel scale ) , implying in general a decrease in the accuracy of aperture placement moving to fainter stars .
a further problem is that as the seeing changes , the amount of blending in very close sources will also vary , to the point that they could become resolved in frames with good seeing , and unresolved in frames with poor seeing .
this causes the centroid to shift in the unresolved ( or poorer seeing ) image toward the companion star , and hence results in a serious error in the aperture flux measurements . the standard method for solving these problems , which we call `
co - located aperture photometry ' , is therefore to use as many stars as possible to determine the aperture positions , in two stages .
the first is to determine accurately the _ relative _ centroid positions of all the stars on the frame , which will be the same for all frames in the time - series ( provided the stars do not move ) .
this can be done using a stacked image to increase @xmath64 ( we typically stack the @xmath68 frames with the best seeing , providing a @xmath69 four - fold improvement in @xmath64 over a single frame ) and thus obtain an improved master catalogue with more accurate relative positions .
furthermore , since the placement of the apertures remains consistent , the effects of varying seeing are limited purely to varying flux loss from the apertures , which can be corrected to a good approximation by a global normalisation over the frame . in the second stage , a transformation is computed between this _
master frame _ and each frame in the time - series on a per - detector basis , using a standard 6-coefficient linear transformation , derived using a least - squares fit to a large number of bright stars . in this case
, the error for the bright stars is dominated by the error in the transformation , and assuming sufficiently large numbers of stars were used , this is in turn dominated by errors in the model , eg . due to radial distortions or other similar effects .
moreover , any errors in the mapping from the master frame to the individual frames will typically either affect all stars in the same way , or will be a smoothly - varying function of position .
such effects are readily removed using a simple polynomial fit ( see [ poly_section ] ) .
figure [ cent_comp ] illustrates this for our m50 data . in this case we have used a simple constant multiplier to normalise each frame to the photometric system of the master frame , using an iterative @xmath70 clipped fit ( derived from the objects classified as stellar ) to remove any variable stars . in monitor data , although there is little to no improvement using the ` co - located apertures ' technique for the majority of sources , it is still necessary to eliminate the problem of centroid shifts in blended sources , as we have suggested .
we suspect that this is the origin of the spurious variable sources seen in the upper panel of figure [ cent_comp ] .
furthermore , another advantage is clear at the faint end , where it provides much more complete sampling , since we can still place an aperture and measure the flux even if the object does not pass the detection threshold on that particular frame , whereas in the upper panel , the object must be detected and the centroid computed before this can be done .
-band observations of m50 , showing all objects of stellar morphological classification .
the upper plot shows the results obtained by placing the photometric apertures at the centroid positions of the stars , as determined on each frame , and the lower plot shows the same using the ` co - located apertures ' technique .
the plots have been truncated at @xmath71 since we require the sources to be detected in at least @xmath72 of the images for the upper diagram , and the detections start to become substantially incomplete for fainter magnitudes . in both cases ,
a simple zero point correction of the individual frames to the master frame has been used ( see [ poly_section ] ) .
the diagonal dashed lines show the expected rms from poisson noise the object , the diagonal dot - dashed lines show the rms from sky noise in the photometric aperture , and the dotted lines show an additional @xmath73 contribution added in quadrature to account for presumed systematic effects .
the solid lines show the overall predicted rms , combining these contributions.,title="fig:",width=288 ] -band observations of m50 , showing all objects of stellar morphological classification .
the upper plot shows the results obtained by placing the photometric apertures at the centroid positions of the stars , as determined on each frame , and the lower plot shows the same using the ` co - located apertures ' technique .
the plots have been truncated at @xmath71 since we require the sources to be detected in at least @xmath72 of the images for the upper diagram , and the detections start to become substantially incomplete for fainter magnitudes . in both cases ,
a simple zero point correction of the individual frames to the master frame has been used ( see [ poly_section ] ) .
the diagonal dashed lines show the expected rms from poisson noise the object , the diagonal dot - dashed lines show the rms from sky noise in the photometric aperture , and the dotted lines show an additional @xmath73 contribution added in quadrature to account for presumed systematic effects .
the solid lines show the overall predicted rms , combining these contributions.,title="fig:",width=288 ] for under - sampled data , the required fractional accuracy relative to the pixel scale is much more stringent , and the noise - induced centroid errors alone can become highly significant , eg . giving a @xmath74 improvement in rms scatter for
significantly under - sampled data from the university of new south wales extrasolar planet search @xcite .
it is straightforward to show that for the majority of images , an aperture with radius approximately equal to the fwhm of the stellar images achieves the optimal balance between flux loss ( and consequently , increased poisson noise in the counts ) and integrated noise in the sky background ( which increases with the area of the aperture ) .
however , for bright sources , this wastes flux since the relative size of the sky noise contribution is much smaller , and a much larger aperture can be used .
our aperture photometry procedure computes the flux in a sequence of apertures of radii @xmath75 , @xmath76 , @xmath77 , etc .
( doubling the area each time ) where the ` core radius ' @xmath75 is set equal to the typical fwhm of stellar images ( and kept fixed for all the data ) .
we use @xmath78 ( @xmath79 ) for the ctio-4m+mosaic data .
we employ a simple procedure to make use of these measurements .
the lightcurve is computed for each aperture separately , and the root mean square scatter ( computed using a robust median - based estimator ) compared for each source .
we simply choose the aperture with the smallest rms for that star .
this procedure ensures that larger apertures are used where they give an improvement for bright sources , but also accounts for blending , where using a larger aperture results in increased contamination of the flux measurement by neighbouring stars , and introduces modulations into the lightcurve as the seeing ( and hence the amount of contaminating flux in the aperture ) changes . in order to place all the stars onto the same zero point system , this procedure necessitates using aperture corrections , to account for the differing amounts of flux lost from the different sized apertures .
these are computed as simple ratios of the flux measured in the different apertures , for non - variable stars .
the dominant effect of this procedure is to produce a small improvement in the achieved rms scatter for the bright stars in the sample .
figure [ aper_comp ] shows a comparison between the results of using this procedure , and using only the @xmath75 ( smallest ) aperture .
we have used a simple constant multiplier to normalise each frame to the photometric system of the master frame , via an iterative @xmath80 clipped fit to remove any variable stars .
-band observations of m50 , showing all objects of stellar morphological classification .
the upper plot shows the results obtained using a single photometric aperture ( radius @xmath81 ) , and the lower plot shows the same using multiple apertures , selected on a per - star basis .
lines as figure [ cent_comp ] .
in both cases , a simple zero point correction of the individual frames to the master frame has been used ( see [ poly_section]).,title="fig:",width=288 ] -band observations of m50 , showing all objects of stellar morphological classification .
the upper plot shows the results obtained using a single photometric aperture ( radius @xmath81 ) , and the lower plot shows the same using multiple apertures , selected on a per - star basis .
lines as figure [ cent_comp ] .
in both cases , a simple zero point correction of the individual frames to the master frame has been used ( see [ poly_section]).,title="fig:",width=288 ] the dominant effect of the atmosphere in ground - based differential photometry is a time - variable shift in the photometric zero point of each frame in the time - series .
this can result from the combination of several effects , and is dominated by variations in transparency and overall extinction ( including the airmass - induced change in the extinction seen on the frame ) .
nightly zero point correction using photometric standard star fields , as is commonly done for measuring absolute photometry , is sufficient to reach the level of a few percent down to @xmath82 .
considerable progress can be made for the purposes of differential photometry , especially over small fields of view , by using non - variable stars in the field of interest to compute zero point shifts for each frame in the time - series . for wide - field instruments such as the ones we are using , higher - order effects start to become significant . in particular , over a @xmath83 diameter field ( eg . int+wfc or ctio+mosaic from corner - to - corner ) , differential variations in airmass across the frame are no longer negligible . assuming the approximation for the airmass @xmath84 where @xmath85 is the zenith distance , and differentiating , @xmath86 substituting a typical value of @xmath87 , @xmath88 . for a typical @xmath89-band atmospheric extinction of @xmath90 ,
this contribution is @xmath91 , and becomes larger moving away from the zenith .
figure [ ext_plot ] shows the difference in extinction across a @xmath92 field as a function of zenith distance .
field as a function of zenith distance , for an assumed atmospheric extinction of @xmath90.,width=288 ] since there are other slowly - varying effects as a function of position on the frame ( eg . some flatfielding problems , astrometric errors inducing position - dependent loss of flux from the apertures , etc . )
we have opted for a generalised approach of fitting 2-d polynomials to the magnitude residuals for each non - variable reference star on each frame , rather than enforcing the particular airmass dependence for atmospheric extinction ( and in our experience this technique does indeed give better results ) .
we have found a quadratic of the form : @xmath93 to be sufficient for all our wide - field data thus far , where @xmath94 and @xmath95 are the pixel coordinates ( with the means @xmath96 and @xmath97 subtracted to give a zero - mean coordinate system , which improves the stability of the least - squares solution ) , @xmath98 are the polynomial coefficients ( fit for each frame from a number of non - variable reference stars ) and @xmath99 is the zero point offset at the position @xmath100 on the frame .
non - variable stars can be identified automatically by using the rms of the lightcurves to reject any variable sources .
we have found that it is often possible to compute this directly from the uncorrected light curve to obtain the initial fit of ( [ quad_equation ] ) , and the refine the solution iteratively by rejecting the most variable stars at each stage .
this technique selects @xmath101 non - variable bright stars on each ccd of the mosaic for the monitor data .
figure [ m50_rms_comp ] compares the effects of applying no zero point correction , a simple zero point shift , and the full quadratic fit , for our ctio-4m+mosaic m50 data .
the best precision reached was @xmath102 for the first case , @xmath103 with the zero point shifts , and @xmath104 with the quadratic fit .
-band observations of m50 , showing all objects of stellar morphological classification .
the upper plot shows the results with no zero point correction , the centre plot the effect of applying the zero - order correction only , and the lower plot shows the full quadratic correction .
lines as figure [ cent_comp].,title="fig:",width=288 ] -band observations of m50 , showing all objects of stellar morphological classification .
the upper plot shows the results with no zero point correction , the centre plot the effect of applying the zero - order correction only , and the lower plot shows the full quadratic correction .
lines as figure [ cent_comp].,title="fig:",width=288 ] -band observations of m50 , showing all objects of stellar morphological classification .
the upper plot shows the results with no zero point correction , the centre plot the effect of applying the zero - order correction only , and the lower plot shows the full quadratic correction .
lines as figure [ cent_comp].,title="fig:",width=288 ] scintillation provides a fundamental limit to the noise performance which can be reached in ground - based photometry .
conventional results for the scintillation level have typically assumed that one star is observed at a time , and we might expect that some of the scintillation would be cancelled out in ccd photometry due to the availability of simultaneous observations of comparison stars .
however , @xcite show that the typical coherence length is @xmath105 , so over the fields of view we are considering , the single star result should apply to a good approximation .
therefore , we can adopt the usual expression ( see @xcite ) of : @xmath106 where @xmath107 is the rms scintillation ( in flux units ) , @xmath108 is the object flux , @xmath109 is the airmass , @xmath110 is the telescope aperture in centimetres , @xmath111 is the exposure time in seconds , @xmath112 is the telescope altitude , and @xmath113 is a turbulence weighted atmospheric altitude , taken here to be @xmath114 . for the int+wfc survey @xmath47-band observations this value is @xmath115 , and for ctio+mosaic @xmath116 . in both cases ,
scintillation is negligible compared to the dominant noise sources in the data .
this is nearly always the case for moderate exposure times on large telescopes .
we present here some details of our actual implementation , as based on the discussion in
[ dp_section ] , for completeness .
the frame - to - frame astrometric transformations are computed using a full astrometric model including radial distortions , by performing an internal astrometric refinement . a single data frame ,
typically the one taken in best seeing and sky conditions , with a good absolute astrometric solution ( against 2mass ) , is used as a reference .
the pipeline - generated object catalogue for this frame is used to produce an astrometric reference catalogue , using the measured positions for all bright , stellar sources ( we use sources down to @xmath117 below saturation ) . the astrometric solution for each data frame in the field
is then refined against this reference .
the internal accuracy after this procedure is typically @xmath118 or better .
we generate the master catalogue by stacking the @xmath68 data frames taken in the best seeing and sky conditions , and use the standard pipeline source detection and morphological classification software .
the classification software ( see @xcite for a more detailed description ) uses the flux of each object , measured in a series of apertures of increasing radii : @xmath119 , @xmath75 , @xmath76 , @xmath77 and @xmath120 , where the default @xmath75 is set approximately equal to the fwhm of the stellar images . by comparing these flux measures ( including also the peak height ) ,
the locus of stellar objects ( which all have approximately the same psf and hence the same flux ratios between apertures ) is defined in planes of flux ratio as a function of magnitude formed from several combinations of the measures .
this is used to define a mean and standard deviation of the flux ratio for the stars , as a function of magnitude , and a normalised statistic is generated from this measuring how ` stellar - like ' each image is .
a classification flag is subsequently derived by defining a boundary in the statistic , and also factoring in the measured image ellipticities .
we use a simple procedure for lightcurve production . the first stage is to convert all the flux measurements to magnitudes .
all of the remaining stages of the processing are performed in magnitudes rather than flux units for convenience .
points with null or negative fluxes ( ie . below sky ) are excluded from the lightcurves .
each ccd of the mosaic is processed separately ( there are always enough stars to do this in our fields of interest , otherwise we would have to use another procedure ) . the median and rms flux of each object is calculated over all the differential photometry measurements , using a robust mad estimator scaled to the equivalent gaussian standard deviation ( ie .
@xmath121 ) .
we apply the procedure of [ poly_section ] to fit and subtract a 2-d quadratic surface from the residuals as a function of @xmath94 and @xmath95 coordinates on each frame . in order to reduce contamination
, the 2-d surface fits use inverse variance weighting ( using the rms flux of each object calculated earlier ) , and we exclude objects flagged as possible blends , saturated datapoints , and all objects with non - stellar morphological classifications .
we estimate expected per - datapoint photometric errors as the quadrature sum of components from poisson noise in the object counts , poisson noise in the sky , rms of the sky background fit ( multiplied by the square root of the number of pixels in the aperture ) , and a constant component of @xmath122 ( as in figure [ m50_rms_comp ] , for example ) to account for systematic errors .
see [ noise_section ] for a more detailed analysis of this last component .
the lightcurves for each field are written into fits binary tables in multi - extension fits files , with one extension per detector ( this convention is also used for the images , object catalogues and differential photometry output ) .
these tables have one row per input object from the _ master catalogue _ , and the lightcurve itself , the photometric error on each lightcurve point and the heliocentric julian date of observation , are stored in columns of the table . our lightcurve generation software , and this file format ,
have been specifically designed to efficiently handle very large data - sets , for example we have also successfully used them on data from the superwasp transit search project @xcite . at this stage ,
the data are ready for lightcurve analysis .
our analysis software , including period finding algorithms , an implementation of the transit search algorithm of @xcite and a number of other programs , interface directly to the lightcurve fits files , and write their results out to additional columns in the files for convenient storage .
typically the full reduction of one week of data from the int+wfc or ctio-4m+mosaic takes @xmath123 including manual checking of the pipeline results .
often the most time - consuming stage of the entire process is reading the data onto disk , which ranges from relatively fast ( @xmath124 ) using external iee-1394 hard disks ( eg . for eso wfi data ) , to very slow ( up to @xmath125 ) for dlt tapes .
we stress the increasing importance of this issue as data rates from astronomical facilities continue to increase , and the enormous savings in time and cost afforded by using internet transfers ( where possible ) or efficient media such as external hard disks or lto-2 tapes .
lightcurves from ground - based transit surveys are invariably found to show significant correlated , or ` red ' noise ( see @xcite for a very detailed discussion ) .
these correlations mean that , averaging over @xmath126 data points , the error in the mean drops less quickly than the ` white ' ( uncorrelated ) noise prediction : @xmath127 where @xmath128 is the error in the mean of @xmath126 data points , and @xmath129 is the error in a single data point ( where we have assumed , for simplicity , that the uncertainties are equal for all the data points ) . throughout this analysis , we assume a value of @xmath126 corresponding to @xmath130 , an appropriate timescale for a hot jupiter transiting a solar - like star , but also comparable to timescales for eclipses in low - mass ebs .
we have tried to maintain consistent notation with @xcite throughout this section .
the least - squares problem of finding the best - fitting box - shaped transit model for a given lightcurve reduces to simply finding the inverse variance weighted mean of the in - transit data points ( eg .
@xcite ) , giving the transit depth if the mean of the out - of - transit data points is subtracted . in order to evaluate the significance of a given detection
, we use the detection statistic @xmath131 of @xcite , repeated here : @xmath132 where the summations run over all in - transit data - points @xmath47 , @xmath133 , the difference between the @xmath47th measured flux @xmath134 and the average flux @xmath135 over all measurements , and @xmath136 is the uncertainty on the @xmath47th flux measurement .
the presence of correlated noise in the lightcurves tends to give larger values of @xmath131 in the absence of transits .
consequently , to maintain a low false alarm rate , we must use a higher detection threshold in @xmath131 , reducing sensitivity to shallow transits , or those with few in - transit data points . furthermore ,
if the level of correlated noise in each lightcurve is known , eq .
( [ q_eqn ] ) can be modified to account for this in the transit detection process ( see @xcite ) .
we have examined the noise properties of our data using a method based on that of @xcite .
we present results based on the m50 lightcurves as a ` best case ' where we believe that our data reduction is closest to optimal .
it should be noted that the prescription we follow for evaluation of red noise will not work at very faint magnitudes , where random noise sources dominate over the correlated noise .
we have therefore analysed lightcurves of the brightest non - saturated stars in our sample , where the effects of red noise are much more significant .
figure [ m50_rednoise_1 ] shows the rms scatter as a function of magnitude for a sample of lightcurves chosen to be approximately ` flat ' ( small reduced @xmath137 ) , which should be noise - dominated .
we have calculated @xmath129 and @xmath128 from ( [ whitenoise_eqn ] ) for @xmath138 , corresponding to @xmath8 with the sampling of these data , and compared @xmath128 with @xmath139 calculated as the rms of means over a @xmath140 window moved along the lightcurve .
this measures the correlated noise over a transit - length window , and in general is larger than @xmath128 if there are correlations on this time - scale .
the results indicate that the level of correlated noise on these time - scales is @xmath7 at the bright end .
other teams have found instances of an increase in the level of correlated noise at faint magnitudes , and figure [ m50_rednoise_1 ] shows that the same is true here for the majority of the stars , where the @xmath139 values never converge to the @xmath128 values .
two likely causes of such effects are residuals in the sky background determinations , and blending , both of which are likely to affect faint stars close to sky more than bright stars . , the rms scatter per data point , filled triangles are @xmath139 , the rms scatter of averages over @xmath140 windows , and asterisks are @xmath128 , the predicted rms scatter over the @xmath140 window assuming white noise .
the filled triangles lie between the other symbols , indicating the presence of correlated noise at the @xmath7 level over @xmath8 for the brightest stars , where the correlations dominate over random ( white ) photometric noise.,width=288 ] in order to make a quantitative estimate of the level of correlations in the noise , we have attempted to measure how rapidly the noise ` averages out ' as a function of the number of data points observed in - transit .
figure [ m50_rednoise_single ] shows the result for a single ` flat ' lightcurve at the bright end of the rms diagram ( @xmath141 ) . in order to generate the diagram ,
a @xmath142 window was moved over the data in @xmath143 time intervals ( approximately @xmath144 of the sampling ) , counting the number @xmath145 of data points lying in the interval , and recording the mean of the data points .
we then computed @xmath146 as the variance of the means at each value of @xmath145 ( where more than one mean was available ) . for uncorrelated ( white ) noise
we expect @xmath147 , where @xmath148 is the standard deviation of the white noise . in general
there is an additional red noise component , which does not average out as the number of data points is increased , ie .
@xmath149 where @xmath150 is the standard deviation of the red noise component .
( ie . the standard deviation ) plotted for a single ` flat ' lightcurve , with @xmath151 , @xmath152 , and @xmath153 .
the solid line shows the white noise prediction @xmath147 , and the dashed line shows the fit of eq .
( [ var_equation ] ) to the data , with parameters @xmath154 and @xmath155 .
the scatter ( especially at large @xmath145 ) is caused by the limited number of @xmath140 windows in the lightcurve containing these particular numbers @xmath145 of data points .
the dot - dashed line shows the predicted curve derived from the autocorrelation function of this lightcurve , using eq .
( [ var_acf_equation]).,width=288 ] figure [ m50_rednoise_2 ] shows the values of @xmath150 as a function of magnitude for all the lightcurves in figure [ m50_rednoise_1 ] .
the upper envelope of derived values increases toward the faint end , ie .
the red noise level is higher at faint magnitudes , as discussed earlier .
we note that the increased random noise level at the faint end affects the determination of the values of @xmath150 ( and @xmath129 ) , and hence introduces scatter as seen in figure [ m50_rednoise_2 ] . from fitting eq .
( [ var_equation ] ) plotted as a function of magnitude .
the upper envelope of derived values increases toward the faint end , which suggests that the red noise level increases for fainter stars.,width=288 ] an alternative method to investigate correlations among the time - sampled data points is to compute the autocorrelation function .
figure [ m50_acf ] shows the autocorrelation function @xmath156 of a representative ` flat ' m50 lightcurve , defined as : @xmath157 where the outer sum is over nights of data @xmath145 , and the inner sum over data points within the night , up to the total @xmath158 taken in that night .
@xmath159 is the magnitude of the star in measurement @xmath47 of night @xmath145 , and @xmath160 is the mean magnitude of the star in night @xmath145 .
the summations were performed in this manner to avoid the nightly gaps influencing the results for short time - scales .
the results indicate that the characteristic coherence timescale of the correlations we see is @xmath161 ( or @xmath162 data points ) , which is typical of the ` flat ' lightcurves in the m50 data - set . ) .
the sampling is approximately one data point every @xmath162 minutes , and the level of correlation is negligible for @xmath163.,width=288 ] it is straightforward to show that the expected @xmath146 can be expressed in terms of the autocorrelation function as : @xmath164 \label{var_acf_equation}\ ] ] this function is shown as the dot - dashed line figure [ m50_rednoise_single ] , for an example lightcurve from the m50 data - set , and provides a better approximation to the observed functional form for @xmath165 than the simple single - parameter description of eq .
( [ var_equation ] ) .
note that eq .
( [ var_acf_equation ] ) is not expected to exactly reproduce the calculated @xmath146 because for a given value of @xmath145 , @xmath146 counts only 2.5 hour windows containing @xmath145 data points , ie . for small @xmath145
the function is dominated by the behaviour at the end of the night , or at the end of observing windows interrupted by the weather , whereas the acf calculates the correlated noise over the entire lightcurve for all @xmath145 .
we find overall levels of ` red noise ' at the low end of the range spanned by other surveys ( eg . see @xcite , @xcite ) , of @xmath122 at the bright end .
since telescope time is at a premium , we have only been able to use one observing strategy throughout , so it is difficult to quantify the factors contributing to @xmath150 from the present data - set .
however , since our levels of red noise are comparable to the existing ground - based surveys , we suggest that we may be obtaining close to the best achievable performance for a ground - based survey over a @xmath166 field using @xmath167 class telescopes , and that the strategy of trying to keep the positions of the sources on the detector as close as possible to constant , appears to be successful
. nevertheless , it would be interesting to investigate the possibility of using small random offsets to attempt to randomise the noise .
we performed a search for correlations in the lightcurves with a number of external parameters , including the image fwhm , sky level ( both globally and local to the sources ) , airmass , hour angle and image morphology ( major axis , ellipticity , position angle ) .
the dominant effect was found to be seeing - correlated variations induced by image blending .
variations in the seeing cause an increase in the amount of blended flux in the photometric apertures as the fwhm of the stellar images increases , so therefore we expect to find a correlation between the measured fwhm and the magnitude , for lightcurves of blended objects . this can be used both for flagging blended objects , and as we shall see , for removing some of the variations induced by blending .
our source detection software flags any objects where the deblending algorithm ( eg .
@xcite ) was invoked , and this flag is propagated into the lightcurves to assist with identifying blended objects . we have found empirically that the flag is often set for objects which do not exhibit any obvious blending effects in the lightcurves , since a greater degree of overlap is required before the object lightcurve becomes sufficiently contaminated .
we therefore developed an empirical technique to characterise the level of blending - induced effects in each lightcurve , by looking for seeing - correlated shifts of the object from its median magnitude .
this is done by fitting a simple quadratic polynomial to the shift as a function of the measured fwhm of the stellar images on the corresponding frame .
some examples are shown in figures [ blend_comp_1 ] and [ blend_comp_2 ] .
we use the following statistic to quantify the level of blending : @xmath168 where @xmath137 is defined as @xmath169 for lightcurve points @xmath170 with uncertainties @xmath136 , and @xmath171 is the median magnitude in the lightcurve .
@xmath172 is the same statistic measured with respect to the quadratic model .
@xmath173 implies that @xmath137 was improved by the model fit
ie . increasing values of @xmath174 to the maximum @xmath175 imply progressively greater amounts of seeing correlation in the lightcurve , or increasing levels of blending .
figure [ blend_hist ] shows a histogram of the blend index , indicating the presence of a peak at @xmath176 , corresponding to objects showing clear seeing - correlated features due to blending , and another peak at @xmath177 corresponding to lightcurves without seeing correlations .
the deblending flag from the source detection software appears to work well for selecting lightcurves with no blending , and hence no seeing correlation , but also flags relatively large number of objects showing little or no seeing - correlated behaviour , due to varying degrees of overlap . for all lightcurves of stellar morphological classification in the ctio m50 data .
the solid line includes all objects , and the dashed line only those objects flagged as blended by the source detection software .
the lower panel shows an expanded version of the upper panel.,title="fig:",width=288 ] for all lightcurves of stellar morphological classification in the ctio m50 data .
the solid line includes all objects , and the dashed line only those objects flagged as blended by the source detection software .
the lower panel shows an expanded version of the upper panel.,title="fig:",width=288 ] a natural progression from the analysis we have described is to attempt to remove some of the seeing - correlated features in the lightcurves by subtracting the fit .
figures [ blend_comp_1 ] and [ blend_comp_2 ] show the results of doing this for two typical lightcurves : one showing significant seeing correlated behaviour ( @xmath178 ) and the other showing little seeing correlated behaviour ( @xmath179 ) . in both cases ,
the procedure significantly reduces the amount of seeing correlated features , and importantly , does not introduce significant additional correlated features . in both cases
the lightcurve rms was reduced , as expected .
figures [ blend_rednoise_single ] and [ blend_rednoise ] show that this corresponds to a reduction in the level of correlated noise as measured in [ noise_section ] .
we have used this simple approach to produce a filter which can be optionally applied to our lightcurves before embarking on transit searches and other similar analyses . for this lightcurve , and the rms was reduced from @xmath180 to @xmath181 after subtracting the fit . in this case the results could be improved further by using a higher degree for the polynomial ( eg .
a quartic).,width=288 ] .
the statistic @xmath182 for this lightcurve , and the rms was reduced from @xmath183 to @xmath184 after subtracting the fit.,width=288 ] for the object in figure [ blend_comp_1 ] before ( top ) and after ( bottom ) subtracting the polynomial fit .
the value of @xmath150 changed from @xmath185 to @xmath186 , and @xmath148 from @xmath187 to @xmath188 , indicating a significant reduction in the levels of white and correlated noise for this lightcurve.,title="fig:",width=288 ] for the object in figure [ blend_comp_1 ] before ( top ) and after ( bottom ) subtracting the polynomial fit .
the value of @xmath150 changed from @xmath185 to @xmath186 , and @xmath148 from @xmath187 to @xmath188 , indicating a significant reduction in the levels of white and correlated noise for this lightcurve.,title="fig:",width=288 ] , the rms scatter of averages over @xmath140 windows , for all lightcurves flagged as possible blends on a single detector in the m50 data - set , before ( dashed line ) and after ( solid line ) the correction for seeing - correlated lightcurve features , showing the reduction in rms resulting from the correction .
the dotted line shows the @xmath189 prediction for white noise.,width=288 ] it is important to note that this approach to removing the effects of blending , in reality , addresses the _ symptom _ , rather than the _ cause _ of the problem .
since aperture photometry ( using multiple apertures ) is a simple approximation to full point spread function fitting ( hereafter , psf - fitting ) , it is not surprising that heavily overlapping images are not well - fit .
a conventional method for reducing the effects of image blending is to move to psf - fitting photometry ( eg .
@xcite ) , using analytical or empirical psfs , or a mixture of the two .
the use of psf - fitting brings with it a significant problem : that of accurately estimating the psf , which is particularly problematic over the wide fields of view we are using due to the presence of significant psf variations .
difference image analysis ( dia ; @xcite , @xcite ) is a popular alternative , and is combined with aperture photometry , or even psf - fitting .
briefly , in this method , the _ master image _ is subtracted from each of the images in the time series .
the resulting difference image should contain mostly noise , and only sources which have varied in flux compared to the _ master image _ will remain . in reality
, the psf varies from frame to frame on any real system , which would leave residuals on the difference images , so it is necessary to use an adaptive kernel @xcite , which is convolved with the master image to degrade the psf to match each target image , before subtraction .
dia considerably simplifies the task of measuring photometry , since the flux from blended stars is cancelled out if they do not vary ( this is nearly always the case ) and therefore does not contribute to the sums over the photometric apertures .
however , the method also suffers from the problem of psf estimation when computing the adaptive kernel . in most cases ,
psf variations require a spatially - varying kernel @xcite to produce good results and avoid leaving residuals on the subtracted images for the non - variable stars .
thus far , our attempts to use dia have not produced superior results to aperture photometry , although the work is still ongoing , particularly in the onc where extensive nebulosity limits the photometric precision available from aperture photometry . particularly in the case of our int data , where the images have variable ellipticities
, we have found that the subtracted images contain significant residuals due to poor psf matching , and these introduce extra ( correlated ) noise into the lightcurves . in these data
, the method does give some measurable improvement for blended stars , but overall higher levels of correlated noise and occasional serious lightcurve ` glitches ' in some objects .
we have therefore chosen to continue using aperture photometry , until we can resolve these issues .
this very popular method for finding ( unknown ) systematic effects in time - series photometry was presented by @xcite .
the sysrem algorithm resembles a generalised form of principal component analysis ( pca ) , where the principal components are a set of generalised ` extinction ' and ` airmass ' terms .
mathematically , the technique searches for the best two sets of coefficients @xmath98 and @xmath190 , to minimise the expression ( in the notation of @xcite ) : @xmath191 where the @xmath126 is the number of measurements in each lightcurve , @xmath192 is the number of lightcurves , @xmath193 is the residual ( mean - subtracted ) flux of object @xmath47 on frame @xmath48 , and @xmath194 is the corresponding uncertainty .
the products @xmath195 can then be subtracted from the lightcurves to remove this principal component , and the technique repeated for subsequent components , deriving progressively smaller corrections to the lightcurves .
since the coefficients are not constrained to be the actual extinction and airmass , the technique also works for other forms of systematic effect . by examining the coefficients , it is possible to determine the origin of the particular effect found by sysrem . in particular the terms @xmath190 ,
representing the correction applied on each frame @xmath48 in the time series , are often correlated with the parameters of the images ( eg . the seeing ) , pointing to the true cause of that particular systematic effect .
we have therefore undertaken such an analysis to find any residual effects in our data .
figure [ sysrem_a ] shows a plot of @xmath190 for the first three sysrem components , and for comparison , plots of several important image parameters .
figure [ sysrem_c ] shows the coefficients @xmath98 for each star , plotted as a function of @xmath196 colour .
the first component seems to show its largest values on a few non - photometric nights , during periods of cloud .
there is no clear correlation with @xmath196 colour .
for the first three sysrem components plotted as a function of @xmath196 colour.,width=288 ] the second component is clearly correlated with the image fwhm .
this indicates that sysrem has found some residual effects of image blending , not corrected by the method described in [ seeing_section ] .
this component is also mildly correlated with @xmath196 colour ( see figure [ sysrem_c ] ) , which indicates that a wavelength - dependent effect ( eg . extinction ) has been detected .
the third component shows very little structure , and gives a correction of very small amplitude ( @xmath197 ) , with only one or two frames having significantly non - zero values of @xmath198 , and no correlation with @xmath196 colour is apparent .
the effect of this component is very small , and we conclude that for these data , the use of two sysrem components appears to be sufficient .
figure [ sysrem_rednoise ] shows the result of subtracting off these two components on the rms over @xmath140 intervals ( approximately the transit timescale ) . , the rms scatter of averages over @xmath140 windows , for all lightcurves on a single detector in the m50 data - set , before ( dashed line ) and after ( solid line ) removing the first two sysrem components .
the dotted line shows the @xmath189 prediction for white noise.,width=288 ] the dotted line in figure [ sysrem_rednoise ] indicates that this method has not detected all of the red noise sources present in the data .
this conclusion is in agreement with the work of other authors ( pont , private communication ) , and suggests that we still can not fully describe the sources of correlated noise in time - series data using the sysrem method .
this is most likely to arise for effects which are not correlated between large samples of stars ( including the case where the effects are present in multiple stars , but at different times ) .
we also note that some of the apparent ` red noise ' could be due to very low - amplitude stellar variability .
@xcite find a very high occurrence of variability at the few mmag level , which is included in our ` red noise ' estimates if it occurs on a transit timescale .
it should be noted that we do _ not _ at present apply the lightcurve corrections derived by sysrem ( or the method of [ seeing_section ] ) to our standard lightcurve output .
instead , the application of these filters is left to the user .
specifically , they have not been used for our rotation work ( eg .
@xcite ) or for visual transit searches , since at this level the systematics corrected tend only to introduce ( small numbers of ) false positives , which can be easily eliminated at the visual inspection stage , whereas the subtraction of the sysrem corrections carries with it the risk of introducing spurious variability from the residuals .
we have developed a software pipeline for processing the high - cadence time - series photometric data generated by the monitor project , using aperture photometry , to achieve rms accuracy down to below @xmath199 at the bright end , typically with rms @xmath3 over @xmath200 ( eg .
@xmath201 for the int / wfc using @xmath202 exposures , @xmath203 for the ctio-4m / mosaic using @xmath204 exposures ) .
our lightcurves are stored in a convenient fits binary table format , designed for efficient storage of multiple lightcurves , and able to handle very large data - sets .
noise properties of the data were investigated in [ noise_section ] , finding correlated ( ` red ' ) noise at the level of @xmath7 over a @xmath140 transit - length timescale .
these effects are important for transit searches since they reduce the effective signal to noise ratio of the transit detection statistic ( here @xmath131 as defined by @xcite ) , thus leading to reduced sensitivity to low - amplitude transits and those with few measured in - transit data points .
@xcite examined the effect of the level of correlated noise on the yield of hot jupiter detections , finding that a level of @xmath104 gave a yield of @xmath69 half the value for no correlated noise , as compared to @xmath205 for example , where the yield was @xmath206 .
therefore , we conclude that the effects of correlated noise on the yield of our survey are acceptable at the present level , but nevertheless we will continue to pursue avenues for improvement such as psf - fitting photometry .
we have investigated seeing - correlated systematic effects in our lightcurves induced by image blending .
a simple blend index was developed to quantify the level of these effects seen in a given lightcurve , based on the @xmath137 of a polynomial fit to the lightcurve magnitudes as a function of the measured image fwhm ( used as an estimate of the seeing ) .
subtracting the fit was found to be an effective method for the removal of these seeing correlations , in lieu of the use of techniques to properly eliminate the effects of blending , such as psf - fitting photometry and difference image analysis .
finally , the sysrem algorithm of @xcite was applied to the data , and the effect of each component examined , to look for further systematic effects .
the removal of two components was found to be sufficient , with the first component removing some systematic effects mostly associated with what appear to be particularly poor - quality frames , and the second removing a seeing - correlated effect , most likely due to residual image blending .
the second component is also mildly correlated with @xmath196 colour , suggesting that this effect has some wavelength - dependence , and may be related to atmospheric extinction .
the isaac newton telescope is operated on the island of la palma by the isaac newton group in the spanish observatorio del roque de los muchachos of the instituto de astrofisica de canarias . based on observations obtained at cerro tololo inter - american observatory , a division of the national optical astronomy observatories , which is operated by the association of universities for research in astronomy , inc . under cooperative agreement with the national science foundation .
this publication makes use of data products from the two micron all sky survey , which is a joint project of the university of massachusetts and the infrared processing and analysis center / california institute of technology , funded by the national aeronautics and space administration and the national science foundation .
jmi gratefully acknowledges the support of a pparc studentship , and sa the support of a pparc postdoctoral fellowship .
we also thank frdric pont for useful discussions of correlated noise , and richard alexander , dan bramich and patricia verrier for assistance with the observing .
sections [ noise_section ] and [ seeing_section ] are based in part on discussions from meetings of the international team on transiting planets of the international space science institute ( issi ) , university of bern .
finally , we would like to express our gratitude to the staff of both observatories the isaac newton group and cerro tololo inter - american observatory for their support .
aigrain s. , irwin m.j .
, 2004 , mnras , 350 , 331 aigrain s. , hodgkin s. , irwin j. , hebb l. , irwin f. , favata f. , moraux e. , pont f. , 2006 , mnras submitted alard c. , 2000 , a&as , 144 , 363 alard c. , lupton r.h . , 1998 , apj , 503 , 325 bertin e. , arnouts s. , 1996 , a&as , 117 , 393 fukugita m. , ichikawa t. , gunn j.e . , doi m. , shimasaku k. , schneider d.p . , 1996 ,
aj , 111 , 1748 hidas m.g . , et al . , 2005 , mnras , 360 , 703 hoaglin d.c . , mosteller f. , tukey j.w . ,
_ understanding robust and exploratory data analysis _ , wiley series in probability and mathematical statistics , wiley , new york hodgkin s.t . , irwin j.m . , aigrain s. , hebb l. , moraux e. , irwin m.j . , 2006 , an , 327 , 9 kjeldsen h. , frandsen s. , 1992 , pasp , 104 , 413 irwin m.j . , 1985 , mnras , 214 , 575 irwin m.j . , 1996 , in _
7th canady islands winter school _ , ed .
espinosa j.m .
irwin j.m . , aigrain s. , hodgkin s. , irwin m. , bouvier j. , clarke c. , hebb l. , moraux e. , 2006 , mnras , 370 , 954 irwin m. , lewis j. , riello m. , hodgkin s. , gonzales - solares e. , evans , d. , bunclark p. , mnras , in prep irwin m.j .
, lewis j.r . , 2001 ,
newar , 45 , 105 landolt a.j . , 1992 ,
aj , 104 , l340 pollacco d.l . ,
et al . , pasp in press , astro - ph/0608454 pont f. , zucker s. , queloz d. , mnras in press , astro - ph/0608597 ryan p. , sandler d. , 1998 , pasp , 110 , 1235 smith a.m.s . ,
et al . , mnras in press , astro - ph/0609618 stetson p.b . , 1987 , pasp , 99 , 191 tamuz o. , mazeh t. , zucker s. , 2005 , mnras , 356 , 1466 tonry j.l .
, howell s.b . , everett m.e . , rodney s.a .
, willman m. , vanoutryve c. , 2005 , pasp , 829 , 281
in order to derive a simple analytic expression , let us consider a source with a gaussian psf , centred on the origin , with total flux @xmath207 and standard deviation @xmath55 .
suppose that we use an aperture of width @xmath208 in the @xmath94-direction , but integrate out to @xmath209 in the @xmath95-direction .
the flux measured , if this aperture is perfectly centred , is given by @xmath210 differentiating with respect to the shift @xmath211 yields : @xmath213 \\ & = & { f_0\over{\sqrt{2 \pi \sigma^2 } } } \left[{\rm e}^{-(r - \delta)^2/2\sigma^2 } - { \rm e}^{-(r + \delta)^2/2\sigma^2}\right]\end{aligned}\ ] ] the expression will be non - analytic for a circular aperture with finite extent in the @xmath95-direction , but the method we have used gives a simple scaling relation to obtain an order of magnitude estimate of the effect of mis - centring . | we have begun a large - scale photometric survey of nearby open clusters and star - forming regions , the monitor project , aiming to measure time - series photometry for @xmath0 cluster members over @xmath1 of sky , to find low - mass eclipsing binary and planet systems .
we describe the software pipeline we have developed for this project , showing that we can achieve peak rms accuracy over the entire data - set of better than @xmath2 using aperture photometry , with rms @xmath3 over @xmath4 , in data from @xmath5 and @xmath6 class telescopes with wide - field mosaic cameras .
we investigate the noise properties of our data , finding correlated ` red ' noise at the @xmath7 level in bright stars , over transit - like timescales of @xmath8 .
an important source of correlated noise in aperture photometry is image blending , which produces variations correlated with the seeing .
we present a simple blend index based on fitting polynomials to these variations , and find that subtracting the fit from the data provides a method to reduce their amplitude , in lieu of using techniques such as point spread function fitting photometry which tackle their cause .
finally , we use the sysrem algorithm to search for any further systematic effects .
methods : data analysis techniques : photometric surveys |
within the past decade , direct imaging of exoplanets has advanced from initial discoveries of favorable low contrast , planetary mass companions like 2m1207b @xcite and ab pic b @xcite to high - contrast exoplanets around bright young stars like the hr8799 planetary system @xcite , @xmath1 pic b @xcite , and hd 95086b @xcite , and gj 504b @xcite .
directly imaged exoplanets form a critical subset of the exoplanet population for which it is possible to characterize atmospheric properties from thermal emission .
given the steep , monotonic decline in planet brightness with time ( e.g. * ? ? ?
* ) , the currently known directly imaged exoplanets ( companions to young stars ) , enable investigations into the early evolution of exoplanets
. imaged young atmospheres also present a valuable comparison to the transmission(e.g . * ? ? ?
* ; * ? ? ?
* ) and emission ( e.g. * ? ? ?
* ; * ? ? ?
* ) spectra of intensely irradiated planets orbiting older stars .
_ hst _ has been employed successfully in characterizing exoplanet systems for over a decade @xcite .
we present wide field camera 3 imaging of the multiple planet system , hr8799 .
the system has four young , massive planets ( @xmath247 m@xmath3 ) orbiting a @xmath230 myr @xcite a5 star , amenable to both photometric and spectroscopic follow - up .
this is the first time that planets with high contrasts have been successfully imaged with wfc3/ir , a camera without a coronagraph and undersampled point spread function ( psf ) .
the techniques described in the paper should enable future studies of planetary systems with _ hst_. in this paper , we characterize the outer two planets , hr8799 _ b _ and _ c _ , by combining the new space - based near - infrared photometric data to the existing suite of ground - based data available for the two planets and discuss the implications of the new photometry .
observations for this study were taken from nov 2011 to dec 2012 with the near - ir channel of wfc3 on the _ hst_. due to the brightness of the primary star ( j=5.3 mag ) , all combinations of sub - array and detector readout time results in saturated zeroth read . in the absence of a coronagraph we designed the observations to saturate the central @xmath20.@xmath4 of the stellar psf using 2.2s long observations , focusing the program on the _ b _ and _ c _ planets .
the hr8799 system was observed in three medium - band filters , f098 m , f127 m , and f139 m over 15 orbits .
the orbits were grouped contiguously in sets of three to increase psf correlation .
the observing sequence was designed to maximize the total roll angle and minimize variations in the psf of the observations over the duration of the program . to maximize the rotation ,
the star was observed over two different roll angles in each orbit , resulting in observations taken over 30 separate roll angles covering @xmath2270 degrees ( not uniformly sampled ) to enable angular differential imaging ( adi ) reduction @xcite . within each roll
, the telescope was dithered using a customized nine - point spiral dither pattern covering a 0.@xmath513@xmath60.@xmath513 region , with half - pixel dithers ( @xmath264 mas ) , ensuring a well sampled psf .
this observing procedure enhanced stability , but resulted in reduced total exposure time on the object and a limited range of roll angles within each block of three orbits . at the conclusion of the observing program ,
the full dataset comprised 270 images in each of the three filters .
the wfc3/ir images are undersampled data at our wavelengths of interest .
the ` drizzlepac ` software is the standard tool to improve the resolution of _ hst _ data , but it not optimal for high contrast imaging since it corrects the geometric distortion prior to improving the resolution of the images , the opposite being more appropriate here .
we therefore modified the _
pipeline @xcite to include the calibration of dithered wfc3 images .
5-@xmath7 hr8799 detection sensitivity with @xmath8f127 m magnitudes as a function of angular separation from the star .
the black circles are our photometry values for hr8799bc .
the solid red line shows the contrast estimated from a global klip reduction and the red dashed segments are contrast levels obtained in the optimized zones .
this figure is intended to help preparation of future observations at high - contrast with _ hst / wfc3 _ with the green line indicating the photon noise floor for the full dataset . ]
the program used a 9-pt spiral dither pattern with 0.5 pixel dither steps .
_ hst _ pointing stability within a single orbit for small dithers is @xmath225 mas . using the wcs information and the `` interleave method '' @xcite
, the dithered frames were combined to improve the resolution by a factor of 2 ( @xmath264 mas / pixel ) .
the selected dither pattern does not permit reconstruction of two fully independent images using interleaving and one dithered frame has to be shared between them . the additional time - sampling provided by generating two images per 9-pt dither meant a deeper contrast and was thus preferred over the one image per 9-pt dither case .
the upsampled images are then corrected from the detector distorted frame to an undistorted frame using correction maps ( j. anderson , private comm ) .
finally , the images were aligned by cross - correlating on the diffraction spikes in the data , resulting in 60 aligned and upsampled images .
the data were reduced by the karhunen - love image projection ( klip ) algorithm based on principal component analysis ( pca , * ? ? ?
* ) used in the alice pipeline . for each of the 60 upsampled images ,
the reference psf library is assembled imposing a minimum rotation of 2@xmath6fwhm at planet location to reduce self - subtraction .
we assumed knowledge of the location of the planets and reduced data in local zones around the expected positions for planets _ b _ and _ c _ , similar to what was done in @xcite .
the reduction was performed in annular sections , over a parameter space exploring two radial sizes and three azimuthal widths . with @xmath225
klip reductions per image our parameter space includes @xmath2150 images , giving @xmath29000 images for each reduction of a given planet / filter combination .
we explored a number of geometries and in particular the location of the planet with respect to the zone with best results when the planet is close to the inner edge of the zone .
each reduced image is corrected for the klip throughput loss using forward modeling , estimated by projecting a _ tinytim _ model psf @xcite onto the kl modes , as described in @xcite .
we then compute a data quality criterion as the ratio of the algorithm throughput to a noise estimate with matched - filtering in a local region around the planet .
this criterion is therefore proportional to the true snr , but without introducing any of the planet signal in the calculation so that it does not bias the results by amplifying speckles .
the reduced data are ordered according to this criterion and cumulatively - combined using a median .
the data quality criterion is then recalculated to determine the optimal number of images in the final image , shown in fig .
[ fig : detection ] .
this approach permits exploration of any algorithm parameter space to produce a single final image automatically by identifying the best combination of algorithm throughput and speckle suppression .
photometry was obtained using matched - filtering with a truncated model psf ( 5 pix ) to reduce potential contamination from local speckles .
this matched filter combined with partial truncation of the planet psf by the reduction zone leads to an incomplete fraction of the planet flux @xmath9 .
all these effects were carefully calibrated using aperture photometry on archival wfc3 data of white dwarfs ( gd153 , g191b2b ) from _ hst _ calibration programs , and of similar spectral type isolated brown dwarfs .
the overall photometric correction precision is of the order of 3 - 4% , significantly smaller than the final photometric error bars on the planets .
the psf library is not rigorously free of companion contribution , however the contribution of the companion at other roll angles is limited , since the data is not a true adi sequence ( the orients span the 270 degrees over multiple epochs ) and the planets are very faint compared to the psf wings .
in addition , the noise is zero - mean within the reduction zone from klip but is not necessarily zero - mean within the matched - filter equivalent aperture , and small biases may remain .
the validity of the forward modeling approach for this particular dataset was investigated by injecting _
psfs in the data at the planets radial separation over a range of azimuthal angles ( 25 psfs for _ b _ and 20 psfs for _ c _ ) .
the injected psf flux was iteratively adjusted to result in the same snr as each true planet after reduction .
the mean photometric error on the synthetic planets was used as a bias correction to account for the two effects discussed above and the error on the photometry was estimated from the standard deviation in the measured signal for each of the fakes .
upper limit detections for hr 8799 _ c _ in f098 m and f139 m were obtained by adjusting the injected psfs to detect almost all of them . to ensure the absence of speckle amplification we injected `` zero - flux synthetic planets '' using the exact same pipeline .
no significant detection could be noticed at the location of these injected zero - flux psfs .
various combinations of reduced images were tested for the presence of flux at given positions as a function of telescope roll angle and reduction parameters , see @xcite for a detailed description .
we estimated the false positive fraction ( fpf ) of 7.6@xmath10 ( snr @xmath210 ) and 3.8@xmath11 ( [email protected] ) for b and c in the f127 m filter assuming gaussian statistics .
we also calculated an fpf of 2.7@xmath12 ( snr@xmath27 ) and 1.3@xmath13 ( snr@xmath24 ) for b and c in f127 m filter assuming a small number of resolution elements using the student s t - test relation presented in @xcite .
contrast curves for the data in the f127 m filter are shown in fig .
[ fig : contrast ] , the figure indicates the contrast achieved using different analysis techniques and the contrast in the global - klip and small - zone klip reductions .
the global - klip reduction is not used in our analysis , we present this contrast curve estimate to guide future possible studies at high - contrast with _ hst / wfc3-ir_. the measured ` vegamag ` photometry for the detections and 1@xmath7 limits is presented in table [ tab : phot ] using revised zero - points .
the photometry of the planets is compared with brown dwarfs measured in the same _ hst _ filters in the color - magnitude diagram in fig .
[ fig : cmd ] . the positions of the planets in fig .
[ fig : cmd ] , verifies that the flux measurement for the planets do not deviate unrealistically from the brown dwarfs .
the spectral type derived for each planet is dependent on the contrast between @xmath0 and water band fluxes being similar to that measured for brown dwarfs . using a relation converting _ hst _ photometric colors to a spectral type from @xcite
, we calculate a spectral type of l9.5 @xmath14 0.5 for hr8799b and @xmath15l7 for hr8799c .
the errors are the quadrature sum of the intrinsic scatter in the relation and the error in the photometric color
. the larger role of clouds in giant planets with temperatures comparable to cloud - free brown dwarfs complicates the conversion of the color - magnitude position into spectral types .
the following two subsections model the combined _ hst _ photometry with measurements from the literature at different wavelengths to determine the atmospheric properties .
_ hst _ near - infrared color - magnitude diagram .
the hr8799bc planets ( red star , shaded region ) are plotted with field l ( blue circles ) and t ( green circles ) brown dwarfs from recent _ hst _ studies @xcite .
the black points are background field targets .
the red rectangle for _ c _ represents our upper - limit estimate . ]
the _ hst _ data in f139 m provide the first detection of the planet _
b _ in the water absorption region at 1.4@xmath16 . additionally , the photometric points in f098 m and the peak of the @xmath0 band ( f127 m ) are consistent with measurements taken across similar wavelengths with ground based instruments @xcite .
we combine the new _ hst _ data with photometric data available in the literature and fit the fluxes to synthetic model spectra from @xcite ( cited as b15 in the rest of the paper ) .
the models include clouds located at the intersection of the pressure - temperature profile and the chemical equilibrium condensation curve with a parameterized thickness but a homogeneous distribution in latitude and longitude across the planet .
the models also include non - equilibrium chemistry for all important c , n and o bearing molecules , as well as updated line lists for ch4 and nh3 ( see b15 for details ) .
the particle size - distribution is centered on 5@xmath16 following a log - normal distribution .
figure [ fig : bestfitmodel ] ( top ) shows two fits for the hr8799b photometry , the best fitting model to the f139 m ( black line ) , and the best fit to the [ 4.05 ] point ( * ? ? ? * green line ) .
no individual model was able to fit the pair of neighboring fluxes at either f127 m and f139 m or @xmath17 and [ 4.05 ] .
the discrepancy may indicate missing physical processes in the model and the data is complicated by the non - contemporaneous measurements .
although the [ 4.05 ] band covers the br-@xmath18 line , accretion is unlikely due to the lack of br-@xmath19 emission in the k - band spectrum @xcite .
the ratio of f139 m flux to f127 m flux may be useful for estimating the combined effects of cloud size distribution , cloud thickness and coverage . however , effective temperature , gravity and composition will need to be better constrained to avoid degeneracies in the model fitting and better agreement with evolutionary models .
fitting a combination of near - ir spectroscopy and ir photometry , b15 found a best matching model with t@xmath20 = 1000k and log(g ) = 3.5 , with potentially subsolar water abundance and enhanced c / o , however , the uncertain surface gravity results in a wide spread of values including solar metallicity . in this paper ,
the cooler best - fit model has solar abundances and we find t@xmath20 = 10001100 k and log(g ) = 3.04.5 , indicating consistency between photometry - based fits and those including spectroscopy .
our measured water band flux , relative to @xmath0-band , however , is brighter than predicted by all of the non - solar models explored by b15 , suggesting that perhaps clouds play an important role . for hr8799c we measured the photometry in the f127 m band and upper limits
were determined for the f098 m and f139 m filters .
the f127 m flux is consistent with values obtained by integrating the spectrum from p1640 over the same wavelength range @xcite , but substantially lower than the reported @xmath0-band photometry @xcite .
the new _ hst _ data were combined with similar and longer wavelength fluxes from the literature to construct the spectral energy distribution given in fig . [
fig : bestfitmodel ] ( bottom ) . plotted with the data is a model with t@xmath20 = 1100k , log(g ) = 3.5 and elevated c / o that fit the k - band spectrum from @xcite .
the fits agree well with the _ hst _ photometry and the longer wavelength data with the exception of a significant difference with the earliest measurement of the @xmath0-band photometry and an @xmath22.5@xmath7 difference with the [ 4.05 ] data point .
our models for hr8799c with homogeneous cloud model agree , within the uncertainties , with the _ hst _ photometry and longer wavelength data out to @xmath23.5@xmath16 with the exception of the first epoch of @xmath0-band photometry .
the difficulty in fitting all available photometry with reported uncertainties persists even when using patchy cloud models formed by combining clouds at different temperatures @xcite or opacity @xcite and the _ hst _
m measurement is more consistent with the patchy cloud models .
our attempts with a linear combination of models did not provide a better fit than the homogeneous cloud model with higher c / o ratio presented in fig . [ fig : bestfitmodel ] ( bottom ) .
.hr8799bc absolute photometry [ cols="^,^,^,^ " , ] p1640 spectrum through the f127 m filter gives , hr8799b = @xmath21 and hr8799c = @xmath22 .
among the _ hst _ filters , the f127 m is the most analogous to a ground - based filter and can be directly compared to previous results reported in the @xmath0-band .
the hr8799b f127 m photometry is consistent with previous results @xcite .
there are two prior reported values for hr8799c - a p1640 spectrum @xcite and a keck @xmath0-band flux @xcite .
the ground - based @xmath0-band photometry for hr8799c is approximately twice as bright as the measured _ hst _ flux . integrating the hr8799c flux calibrated p1640 spectrum through the f127 m filter matches the _ hst _ photometry to within 2-@xmath7 . a potential solution for this discrepancy might be intrinsic photometric variability caused by heterogeneous cloud layers , which we find is not required in our model fits .
the early groundbased photometry might also have suffered from calibration issues which combined by the intrinsic variability of the star might explain some of the difference in flux measurement .
efforts to match synthetic spectra to the ensemble of photometric data for hr8799b result in high effective temperatures and corresponding radii that are smaller by @xmath250% ( 0.69 - 0.92 r@xmath3 ) , than predicted by theoretical brown dwarf and giant planet cooling tracks ( see * ? ? ? * for a summary ) .
difficulty finding a model spectrum that simultaneously matches the near and mid - ir photometry could be due to the non - contemporaneous nature of the observations or perhaps the model - observation inconsistencies at multiple bands are an indication of large - scale flux variations .
large variability would bias such model comparisons .
for example , the bright @xmath0-band flux from @xcite is more consistent with higher effective temperatures than our _ hst _ f127 m flux .
the deep water absorption demonstrates that the atmospheres of both _ b _ and _ c _ are not enshrouded in high altitude hazes or clouds many pressure scale heights thick , important for many transiting exoplanets @xcite .
nonetheless , clouds are important in shaping the overall sed of the planets .
both best - matching models plotted in fig . [ fig : bestfitmodel ] ( top panel ) have clouds composed primarily of iron and magnesium - silcate grains , located in the near - infrared photosphere .
the cooler model has a cloud located at pgas @xmath2 1 bar and extending upward 1 pressure scale - height .
the warmer ( and higher gravity ) model has a cloud base near 10 bar , extending upward 2 pressure scale heights .
using single models to reproduce the observations assumes global cloud coverage which is probably an overestimation . with spectral energy distributions
well sampled observationally and with model spectra that match reasonably well , the bolometric luminosities of both planets can be estimated , and we determine l@xmath23 values of [email protected] for _ b _ and [email protected] for _ c _ consistent with @xcite .
in this study , we have demonstrated that wfc3 is capable of investigating the atmospheres of planets requiring high contrast at wavelengths either inaccessible from the ground or at which the extreme adaptive optics systems perform poorly . with its photometric stability , _ hst _ provides a valuable resource to explore exoplanet atmospheres and will enable measurements such as the search for atmospheric variability .
the data from the current program can be used to build a psf library for future high contrast imaging programs . based on experience from the current program
future studies would benefit from obtaining more exposures rather than performing the time - consuming in - orbit roll . in this study , we were able to detect hr8799b in three wfc3 filters . in this work and in other studies , matching all of the data simultaneously has not been possible .
in particular , some of the shortest wavelength data favor warmer conditions while some of the longer wavelength ir data favor cooler conditions , possibly caused by clouds . for hr8799c , we find very good agreement between the model fit with the f127 m , resolving the long standing difficulty of fitting the previously reported @xmath0-band photometry .
a possible explanation for the f127 m and @xmath0-band flux difference is variability .
future spectrophotometric monitoring of hr8799b and hr8799c in the near and mid - ir might resolve whether the planets are variable over the duration of a single rotation period .
based on observations made with the hubble space telescope , associated with program # 12511 .
this research was supported in part by nasa cooperative agreements hst - go-12511.04 , hst - ar-12652.01 , nsf ast-1411868 . , . ,
pueyo , l. , hagan , j. b. , gofas - salas , e. , rajan , a. , chen , c. , perrin , m. d. , debes , j. , golimowski , d. , hines , d. c. , ndiaye , m. , schneider , g. , mawet , d. , marois , c. , & soummer , r. 2014 , in society of photo - optical instrumentation engineers ( spie ) conference series , vol . 9143 , society of photo - optical instrumentation engineers ( spie ) conference series , 57 , j. e. , hook , r. n. , & stoehr , f. 2011 , in society of photo - optical instrumentation engineers ( spie ) conference series , vol . 8127 , society of photo - optical instrumentation engineers ( spie ) conference series , 0 , m. , tamura , m. , kudo , t. , janson , m. , kandori , r. , brandt , t. d. , thalmann , c. , spiegel , d. , biller , b. , carson , j. , hori , y. , suzuki , r. , burrows , a. , henning , t. , turner , e. l. , mcelwain , m. w. , moro - martn , a. , suenaga , t. , takahashi , y. h. , kwon , j. , lucas , p. , abe , l. , brandner , w. , egner , s. , feldt , m. , fujiwara , h. , goto , m. , grady , c. a. , guyon , o. , hashimoto , j. , hayano , y. , hayashi , m. , hayashi , s. s. , hodapp , k. w. , ishii , m. , iye , m. , knapp , g. r. , matsuo , t. , mayama , s. , miyama , s. , morino , j .-
i . , nishikawa , j. , nishimura , t. , kotani , t. , kusakabe , n. , pyo , t .- s . , serabyn , e. , suto , h. , takami , m. , takato , n. , terada , h. , tomono , d. , watanabe , m. , wisniewski , j. p. , yamada , t. , takami , h. , & usuda , t. 2013 , , 774 , 11 , b. r. , baranec , c. , beichman , c. , brenner , d. , burruss , r. , cady , e. , crepp , j. r. , dekany , r. , fergus , r. , hale , d. , hillenbrand , l. , hinkley , s. , hogg , d. w. , king , d. , ligon , e. r. , lockhart , t. , nilsson , r. , parry , i. r. , pueyo , l. , rice , e. , roberts , j. e. , roberts , jr . , l. c. , shao , m. , sivaramakrishnan , a. , soummer , r. , truong , t. , vasisht , g. , veicht , a. , vescelus , f. , wallace , j. k. , zhai , c. , & zimmerman , n. 2013 , , 768 , 24 , l. , soummer , r. , hoffmann , j. , oppenheimer , r. , graham , j. r. , zimmerman , n. , zhai , c. , wallace , j. k. , vescelus , f. , veicht , a. , vasisht , g. , truong , t. , sivaramakrishnan , a. , shao , m. , roberts , jr .
, l. c. , roberts , j. e. , rice , e. , parry , i. r. , nilsson , r. , lockhart , t. , ligon , e. r. , king , d. , hinkley , s. , hillenbrand , l. , hale , d. , dekany , r. , crepp , j. r. , cady , e. , burruss , r. , brenner , d. , beichman , c. , & baranec , c. 2015 , , 803 , 31 , a. j. , hinz , p. m. , esposito , s. , burrows , a. , leisenring , j. , skrutskie , m. , desidera , s. , mesa , d. , arcidiacono , c. , mannucci , f. , rodigas , t. j. , close , l. , mccarthy , d. , kulesa , c. , agapito , g. , apai , d. , argomedo , j. , bailey , v. , boutsia , k. , briguglio , r. , brusa , g. , busoni , l. , claudi , r. , eisner , j. , fini , l. , follette , k. b. , garnavich , p. , gratton , r. , guerra , j. c. , hill , j. m. , hoffmann , w. f. , jones , t. , krejny , m. , males , j. , masciadri , e. , meyer , m. r. , miller , d. l. , morzinski , k. , nelson , m. , pinna , e. , puglisi , a. , quanz , s. p. , quiros - pacheco , f. , riccardi , a. , stefanini , p. , vaitheeswaran , v. , wilson , j. c. , & xompero , m. 2012 , , 753 , 14 , a. j. , marley , m. s. , hinz , p. m. , morzinski , k. m. , skrutskie , m. f. , leisenring , j. m. , close , l. m. , saumon , d. , bailey , v. p. , briguglio , r. , defrere , d. , esposito , s. , follette , k. b. , hill , j. m. , males , j. r. , puglisi , a. , rodigas , t. j. , & xompero , m. 2014 , , 792 , 17 , r. , perrin , m. d. , pueyo , l. , choquet , .
, chen , c. , golimowski , d. a. , brendan hagan , j. , mittal , t. , moerchen , m. , ndiaye , m. , rajan , a. , wolff , s. , debes , j. , hines , d. c. , & schneider , g. 2014 , , 786 , l23 , h. , apai , d. , marley , m. s. , saumon , d. , morley , c. v. , buenzli , e. , artigau , e. , radigan , j. , metchev , s. , burgasser , a. j. , mohanty , s. , lowrance , p. l. , showman , a. p. , karalidi , t. , flateau , d. , & heinze , a. n. 2014 , arxiv e - prints | we present results from a _ hubble space telescope _ ( _ hst _ ) program characterizing the atmospheres of the outer two planets , in the hr8799 system .
the images were taken over 15 orbits in three near - infrared medium - band filters
f098 m , f127 m and f139 m using the wide field camera 3 .
one of the three filters is sensitive to water absorption band inaccessible from ground - based observations , providing a unique probe of the thermal emission from the atmospheres of these young giant planets .
the observations were taken at 30 different spacecraft rolls to enable angular differential imaging , and the full data set was analyzed with the karhunen - love image projection ( klip ) routine , an advanced image processing algorithm adapted to work with _ hst _ data . to achieve the required high contrast at sub arcsecond resolution , we utilized the pointing accuracy of _ hst _ in combination with an improved pipeline designed to combine the dithered , angular differential imaging data with an algorithm designed to both improve the image resolution and accurately measure the photometry .
the results include f127 m ( @xmath0 ) detections of the outer planets , hr8799 _ b _ and _ c _ and the first detection of hr8799b in the water - band ( f139 m ) filter .
the f127 m photometry for hr8799c agrees well with fitted atmospheric models resolving a long standing difficulty to model the near - ir flux for the planet consistently . |
a solid majority of observed galaxies have disk - dominant morphology ; despite wide variance in methods of sampling and classification , roughly 70% of galaxy - sized dark matter halos in the universe host late - type systems @xcite .
moreover , @xcite find that about one - third of all local disk galaxies have no observable pressure - supported component ( whether a classical " bulge formed by the central starburst associated with a merger event , or a pseudobulge " having arisen from the secular transport of angular momentum towards the galactic center ) , and another one - third host systems with only pseudobulges , a conclusion supported by spheroid - disk decomposition of large galaxy samples @xcite .
the vast majority of disk stars in the milky way reside in the thin disk component , with an exponential scale height of @xmath5 pc ( * ? ? ? * and references therein ) and
a total velocity dispersion of @xmath6 km s@xmath7 @xcite . whether the scale height of the galactic disk is typical for galaxies of its size is a topic of vital interest .
unfortunately , firm measurements for a statistical sample of galaxies have been limited by dust obscuration , which present a problem even in k - band imaging @xcite . aside from the considerable challenges associated with forming disk galaxies in @xmath1cdm cosmologies ( e.g. , * ? ? ?
* ) , hierarchical models must also self - consistently maintain thin , rotationally - supported systems against the constant barrage of merging subhalos .
though the former endeavor has enjoyed some recent advances @xcite , the survival of disk galaxies during the often - violent mass accretion history of their dark host remains a concern @xcite and has been the target of numerous studies aimed at quantifying the resilience of galactic disks to satellite accretion events @xcite .
both numerical simulations @xcite and purely analytic calculations @xcite indicate that mass delivery into dark matter halos of mass @xmath8 is dominated by the accretion of objects with mass @xmath9 .
@xcite find that @xmath10 of @xmath11 galaxy - sized halos have accreted a system of mass @xmath2 into their virial radii in the last 10 gyr , with associated disk impacts within the last 8 gyr .
@xcite also find that less massive accretions are virtually ubiquitous , and that the merger fraction falls off quickly for satellites larger than @xmath12 .
overall , these results suggest that @xmath13 satellite accretion events represent the primary concern for disk survival in @xmath1cdm . along these lines ,
recent analytic work by @xcite suggests that orbital energy deposition via merger is less destructive to a disk than was often previously surmised @xcite , claiming that the galaxy could have undergone @xmath14 independent mergers of this kind since @xmath15 while maintaining a thin disk .
recently , @xcite utilized dissipationless @xmath4-body simulations to investigate the response of thin galactic disks subject to a @xmath1cdm - motivated satellite accretion history .
these authors showed that the thin disk component survives , though it is strongly perturbed by the violent gravitational encounters with substructure .
however , @xcite focused on infalling systems with masses in the range @xmath16 , ignoring the most massive accretion events expected over a galaxy s lifetime . in this letter
, we expand upon this initiative by investigating the morphological and dynamical evolution of initially - thin galaxy - type disks during cosmologically common @xmath13 accretion events involving two - component ( stars and dark matter ) satellites of mass @xmath2 . working in a similar mass regime
, @xcite simulated the formation of thick disks via the infall of satellite galaxies with virial masses @xmath17 that of the host , using both a @xmath18 galactic primary system and a scaled version at @xmath19 in order to show that realistic thick disks result from these impacts .
though our preparation is similar , our goals and techniques are different .
we aim to determine whether _ any _ thin , dynamically cold component can survive such an event , and conservatively use a primary disk that is as massive as the milky way disk _
today_. past investigations into the stability of galactic disks against the infall of satellites have often suffered from the necessities of numerical limitations or from analytic axioms later deemed incompatible with standard cosmological models ; for example , the modeling of one or more structural components as rigid potentials @xcite , the initialization of a disk much thicker than the old , thin stellar disk of the galaxy @xcite , the infall of satellites with only a concentrated baryonic component @xcite , and the imposition of subhalo infalls with orbital parameters inconsistent with @xmath1cdm cosmological models @xcite .
analytic arguments , meanwhile , have historically been forced to assume simplifications such as the local deposition of a satellite s orbital energy @xcite , or the absence of global heating modes @xcite which are analytically shown to dominate disk heating by @xcite , although the latter authors employ a rigid satellite model and perfectly radial polar orbits for their simulated experimental tests .
fortunately , advances both in computational power and in our understanding of @xmath1cdm expectations allow us to address these concerns directly .
our contribution improves upon earlier studies in several important respects .
first and foremost , we examine the response of galactic disks to accretion events that represent the primary concern for disk survival in @xmath1cdm cosmologies .
secondly , we employ galaxy and satellite models that are constructed in equilibrium from fully self - consistent distribution functions and which have the resolution in force and mass to study the heating of a disk that is as thin as the old thin stellar disk of the milky way ( @xmath20 pc ) ; in synergy with the high mass and force resolution we adopt , this quality allows us to construct equilibrium @xmath4-body models of disk galaxies that are as thin as the _ old _ , thin stellar disk of the galaxy .
lastly , the masses , density structure , stellar content , and orbital configurations of our infalling satellites are directly motivated by the prevailing @xmath1cdm paradigm of structure formation .
all simulations are performed using the multi - stepping , parallel , tree @xmath4-body code pkdgrav @xcite , in which we set the gravitational softening length to @xmath21 100 pc and 50 pc for dark matter and stellar particles , respectively .
we construct @xmath4-body realizations of primary disk galaxies and satellites using the method of @xcite .
this technique produces self - consistent , multi - component galaxy models that are ideal for studying complex dynamical processes associated with the intrinsic fragility of galactic disks such as gravitational interactions with infalling subhalos .
we explore two initial models for the primary galaxy in our satellite - disk encounter simulations : _ galaxy 1 _ ( hereafter g1 ) , a milky - way - analog system drawn from the set of self - consistent equilibrium models that best fit galactic observational parameters as produced by @xcite ; and _ galaxy 2 _ ( hereafter g2 ) , an identical system save for the absence of a central bulge , i.e. , the two models have stellar disks and dark halos with equivalent _ initial _ properties . in each case the dark matter halo of the primary galaxy was populated by @xmath22 particles following the ( * ? ?
* hereafter nfw ) density profile with scale radius @xmath23 kpc , and the bulge in g1 ( comprised of @xmath24 particles ) contained a stellar mass @xmath25 following a srsic profile with effective radius @xmath26 kpc and index @xmath27 .
the stellar disks , comprised of @xmath28 particles each , contained a mass @xmath29 following an exponential distribution in cylindrical radius with scale length @xmath30 kpc , while the vertical distribution of stars was described by a sech@xmath31 function with @xmath32 kpc being the vertical scale height .
we note that the choice of numerical and physical parameters minimize secular evolution ( e.g. , strong bar formation , artificial heating through interactions with massive halo particles ) on the timescales of relevance to our investigation , which could interfere with the interpretation of our results . in the left panel of figure [ fig : maps ]
, we show the edge - on surface brightness map for both primary galaxy models , having assumed a stellar mass - to - light ratio @xmath33 .
the satellite galaxy in each case was initialized with @xmath34 dark particles representing a mass @xmath35 within the virial radius of a halo which is well - fit by an nfw profile with a concentration @xmath36 at @xmath37 .
we populate this satellite with a stellar mass @xmath38 , roughly corresponding to the upper-@xmath39 limit derived by @xcite for @xmath40 at our subhalo s virial mass , and we distribute these @xmath41 stellar particles in a central spheroid with srsic index @xmath42 according to the distribution of shape parameters versus dwarf elliptical galaxy magnitudes found by @xcite in their survey of virgo cluster members .
our initial subhalo velocity vectors are motivated by cosmological investigations of substructure accretions , where the distributions of radial and tangential velocity components ( @xmath43 and @xmath44 ) peak respectively at 90% and 60% of the virial velocity of the satellite s host halo @xcite . in our case
this corresponds to an initial subhalo velocity with @xmath45 km / s and @xmath46 km / s .
we initiate the infall of each simulation s subhalo at a relatively large radius of approximately @xmath47 kpc to ensure that the disk does not suffer substantial perturbations due to the sudden presence of the satellite s potential well .
we simulate an array of orbital inclinations ( @xmath48 , defining @xmath49 as the angle between the angular momentum axes of the disk and the orbit ) in order to assess the consequence of this parameter on the evolution of the galactic disk . in the polar infall ( @xmath50 )
, we eliminate the tangential velocity component of the subhalo , sending the satellite on a direct - impact trajectory into the center of the primary galaxy ; this case is somewhat unrealistic , but provides an interesting experimental benchmark .
all but one of the non - polar subhalo orbits are initialized as prograde with respect to the primary galaxy s rotation ; we also simulate a retrograde orbit for g1 with @xmath51 in order to investigate whether the heating effects are reduced ( as conjectured by * ? ? ?
all simulations were evolved for a total of 5 gyr , after which the subhalo has fully coalesced into the center of the host halo and the stellar disk has relaxed into stability ; although there are certainly remnant features in the outer disk and halo that will continue to phase - mix and virialize on a much longer timescale , our investigations indicate that the disk - evolution process has reached a quasi - steady state by this point in the encounter s evolution .
= 3.5 in = 3.2 in [ cols="<,^,^,^,^,^,^,^ " , ]
edge - on surface brightness profiles for remnants of the @xmath52 and 90@xmath53 impacts are shown in the middle and left panels of figure 1 , where the upper and lower renderings correspond to primary cases g1 and g2 , respectively ( with and without initial bulge ) .
it is clear from these images that the resultant disks are considerably thicker than the initial case .
we note that while the stars in the accreted satellite end up in the final disk remnant ( c.f .
villalobos & helmi 2008 ) , primary disk stars dominate these images , even high above the plane .
figure [ fig : diskheight ] shows the minor - axis surface brightness profiles for the g1 simulations using @xmath33 .
the left panel shows a vertical slice at a projected radius of @xmath54 kpc and the right panel shows a similar slice at radius @xmath55 .
black solid lines show the initial disk and different color / line types represent the remnants as indicated .
clearly , the resultant disks are dramatically thicker than the initial galaxy model in each case . in order to conservatively compare our disks to the milky way
, we allow for thick _ and _ thin components by fitting a double - sech@xmath31 profile at @xmath56 .
the fitted scale heights are shown in table [ finalscaleheights ] and compared directly in figure [ fig : scaleheight ] to the galactic values obtained by @xcite .
profiles that fall by one mag / arcsec@xmath31 at the same height as the exponential fits .
this multiplicative factor is more appropriate for thin - disk comparisons near the peak of the profile than the widely - used factor of 2 that matches exponential and sech@xmath31 profiles at large heights above the disk plane . ]
though we initially employ a disk that is thicker ( @xmath57 kpc ) and therefore conservatively more robust to accretion events ( kazantzidis et al .
2009 , in preparation ) than the galactic value of @xmath58 kpc from @xcite , the final systems all have thin - disk components with @xmath59 larger by a factor of @xmath60 than the milky way . moreover , the low - surface - brightness thick component in our remnant disks is also considerably thicker than the galactic thick disk , with scale heights so large ( @xmath61 kpc ) that this material would likely be considered a stellar halo component .
a second relevant measure of disk survival is the stellar velocity dispersion ; we therefore compare the velocity ellipsoid of our final disks to that observed in the solar neighborhood by the geneva - copenhagen survey ( * ? ? ?
* see also @xcite ) . in figure
[ fig : sigma ] , we show these values for velocity dispersion in each coordinate , where the indicated range spans the stellar population age , and the point is placed at the median age @xmath62 gyr according to @xcite for this local sample . shown also are the corresponding dispersion components for our initial and final stellar disks measured within an 0.3-kpc box centered on the disk plane at @xmath63kpc . as summarized in table 2 , each of our simulated merger remnants
are substantially enhanced in all three components of velocity dispersion ( @xmath64 corresponding to @xmath65 ) .
the total dispersion @xmath66 increases by a factor of @xmath67 compared to that of the initial disk .
using fully self - consistent @xmath4-body simulations of satellite - disk interactions we have quantitatively demonstrated for the first time that cosmologically common accretion events of mass ratio @xmath13 do not preserve thin , dynamically cold stellar disks like the old , thin stellar disk of the milky way .
this has potentially serious ramifications for models of galaxy formation and evolution .
it is possible that our benchmark case of the milky way is not representative , and that the galaxy sits within a rare halo that has not experienced in the last @xmath68 gyr a disk impact associated with a significant accretion event , as posited in the observationally - motivated suggestion of @xcite , in which the galaxy is shown to have remarkably low angular momentum and stellar mass compared to local spiral galaxies in host halos of similar mass .
future investigations may help quantify the range of thin - disk scale heights in the local universe .
otherwise , the addition of gas physics may play a role in explaining the apparent discrepancy .
gas can cool and reform a thin disk , and its presence may stabilize the stellar disk ( e.g. , * ? ? ?
the regrowth of the massive thin disk _ after _ a satellite accretion may cause heated stars to contract and lose kinetic energy .
accurate treatment of the various aspects of hydrodynamics will therefore play a crucial role in the capacity of simulated galaxy evolution to reproduce thin disks such as those that dominate observed galaxy catalogs . in a recent paper @xcite
have argued that disk heating is less effective than previously thought and that the expected merger histories of @xmath1cdm halos are compatible with the high thin - disk fraction seen in the universe .
it is important , therefore , to investigate this point of disagreement .
their result , a reshaping of the arguments presented in @xcite ( updated to reflect the more realistically radialized orbits of a @xmath1cdm cosmology ) , relied primarily on an analytic formula , normalized to simulations with much lower mass and force resolution than those explored here , to map the ratio ( @xmath69 ) to a disk heating parameter @xmath70 , where @xmath71 is the median scale height of the resultant disk and @xmath72 is the radius where the height is measured ( and must be within a factor of two of the disk half - mass radius @xmath73 ) . for the @xmath13 mass - ratio accretion events we explore here , the @xcite formula predicts a disk thickening of @xmath74 .
our simulations typically exhibit significantly more heating ; at @xmath75 we measure @xmath76 and , because of the impact - induced flaring , we measure even larger values @xmath77 at @xmath78 .
it is perhaps not surprising that our results disagree with first - order analytic expectations .
in addition to direct heating , the resultant disk structure is affected by global modes such as bending and density waves excited in the disk as the interaction occurs @xcite , and not included in the simple analytic scalings is a dependence on the orbital inclination of the encounter that is likely associated with resonant coupling .
finally , though @xcite normalized their results to numerical simulations , those initial disks were significantly thicker than the galactic - type disk we have simulated , and were therefore more robust to tidal perturbations .
direct numerical experiments involving satellite - disk encounters indicate that mass - ratio , orbital inclination , initial disk scale height , and relative dark matter fraction are all crucial in determining the degree to which galactic disks are perturbed by infalling subhalos ( figures [ fig : scaleheight ] , [ fig : sigma ] , and kazantzidis et al .
2009 , in preparation ) .
more detailed analysis is forthcoming of the morphological and dynamical effects experienced by our disks ; among other concerns , we defer for future work the issues of stellar - halo / thick - disk distinguishability and the reinforcement of central bulges by accreted stars .
we thank joachim stadel for providing the pkdgrav code .
we would like to thank charlie conroy , phil hopkins , kyle stewart , and andrew zentner for useful discussions as well as larry widrow and john dubinski for kindly making available the software used to set up the primary galaxy models .
cwp and jsb are supported by national science foundation ( nsf ) grants ast-0607377 and ast-0507816 , and the center for cosmology at uc irvine .
sk is supported by the center for cosmology and astro - particle physics at the ohio state university .
the numerical simulations were performed on the ia-64 cluster at the san diego supercomputing center . | most galaxy - sized systems ( @xmath0 ) in the @xmath1cdm cosmology are expected to have interacted with at least one satellite with a total mass @xmath2 in the past @xmath3 gyr .
analytic and numerical investigations suggest that this is the most precarious type of accretion for the survival of thin galactic disks because more massive accretion events are relatively rare and less massive ones preserve thin disk components .
we use high - resolution , dissipationless @xmath4-body simulations to study the response of an initially - thin , fully - formed milky - way type stellar disk to these cosmologically common satellite accretion events and show that the thin disk does not survive .
regardless of orbital configuration , the impacts transform the disks into structures that are roughly three times as thick and more than twice as kinematically hot as the observed dominant thin disk component of the milky way . we conclude that if the galactic thin disk is a representative case , then the presence of a stabilizing gas component is the only recourse for explaining the preponderance of disk galaxies in a @xmath1cdm universe ; otherwise , the disk of the milky way must be uncommonly cold and thin for its luminosity , perhaps as a consequence of an unusually quiescent accretion history . |
the interstellar medium ( ism ) within a few hundred parsecs of the solar system has been a topic of considerable observational and theoretical investigation over recent years ( e.g. breitschwerdt , freyberg & trmper 1998 ; frisch 1996 ; cox & reynolds 1987 ) .
this region , often referred to as the local interstellar medium ( lism ) , is known to contain several major features in the form of bubbles , supernova shells and clouds .
the solar system itself is thought to reside in a low - density , x - ray emitting cavity of a mean radius @xmath0 100 pc , a region usually referred to as the local bubble ( e.g. cox & reynolds 1987 ) .
it is believed to be the remnant of a supernova explosion that occurred @xmath1 yr ago .
prominent amongst other features in the lism are the four large - diameter , almost circular rings seen in the all - sky distribution of radio continuum emission .
these are appropriately named radio loops i to iv ( haslam , khan & meaburn 1971 ) and are thought to be the projections of quasi - spherical bubbles . of these , the loop i bubble the largest in angular extent ( solid angle @xmath2 steradian ) and also the brightest of the four loops is of special interest ( see salter ( 1983 ) for a review ) .
its close proximity to the sun allows an in - depth study ; see for example , the work of egger ( 1993 ) using data from the rosat soft x - ray sky survey , and that of nishikida ( 1999 ) who combined the rosat pspc data with iras sky survey and radio 21-cm data . in terms of origin ,
loop i is thought to be an expanding super - bubble triggered by an epoch of star formation in the sco - cen association some @xmath3 years ago .
further , it has been postulated that the properties of the lism may well be conditioned by the outer shock wave of this supernova remnant ( frisch 1981 , 1996 ; lallement 1998 ) .
besides these radio loops , the other well - known examples of nearby bubbles that may be important in understanding the lism are : the eridanus bubble ( distance from the sun , d @xmath0 100150 pc ) , the gum nebula ( d@xmath0200250 pc ) , the orion bubble ( d @xmath0 455 pc ) and the monogem ring ( d @xmath0 1001300 pc ) .
recent x - ray and uv data from rosat have led to several new insights into the structure of the lism .
a noteworthy result comes from rosat pspc data suggesting an ongoing interaction between the local bubble and the loop i bubble . by comparing the diffuse x - ray background maps ( in the 0.12.0 kev band ) from the rosat all - sky survey ( trmper 1983 ) and data over a @xmath4 region centered at ( @xmath5)=(@xmath6 ; the apparent centre of the radio loop i ) , egger & aschenbach ( 1995 ; hereafter ea95 ; see also egger 1993 , 1998 ) recognize a `` ring - like '' structure between the local bubble and loop i , where the column density ( ) is even higher than that of the intervening dense shell known to exist in the direction of the sco - cen ob association ( centurion & vladilo 1991 ) . from absorption - line studies of nearby stars ,
the distance to this interaction feature ( where jumps by nearly an order of magnitude ) is estimated to be @xmath0 70 pc ( see ea95 ) , quite comparable to the distance to the neutral wall ( @xmath7 pc ) inferred from rosat wfc star counts ( warwick et al .
1993 ) and optical and uv spectral line data ( centurion & vladilo 1991 ) .
interestingly , the formation of such a ring and the inferred density enhancement ( by a factor @xmath0 25 ) are in good agreement with the numerical simulations of colliding interstellar bubbles ( yoshioka & ikeuchi 1990 ) . in this picture ,
loop i is considered to be an active super - bubble , with at least one bubble having already formed a dense cool shell prior to collision
. however , alternative interpretations do exist ( e.g. , frisch 1996 , 1998 ) , wherein the local bubble is pictured as an appendix of loop i. more data and modeling may help to distinguish clearly between the different scenarios .
while much of our understanding of the structure of the lism comes primarily from x - ray and uv data , nearby pulsars are promising tools for enhancing this understanding .
studies of dispersion and scattering of the radio signals from pulsars are useful means for probing the intervening ism .
of these two , interstellar scintillation ( iss ) effects are more likely to be influenced by the peculiar distribution of the ionized plasma along the line of sight ( los ) such as clumps of enhanced density superposed on a uniform distribution of ionized material .
this is mainly because of ( a ) the nonlinear relation(s ) of the scintillation properties to the electron density , and ( b ) the fact that scintillation effects depend critically on the relative location of the scatterer ( or more generally , the actual distribution of scattering plasma along the sight line ) .
for instance , if the medium is inhomogeneous , in order to produce noticeable effects in the dispersion measure ( dm ) , the density at the clumps has to be considerably larger than to produce similar effects in the scintillation data .
however , if the pulsar happens to lie within or near the clumped region , its effect on the scintillation properties could be significantly reduced .
thus , in some sense , dispersion and scintillation data can provide information complementary to each other .
nevertheless , ( b ) would imply that the interpretation of scintillation data is less immune to distance errors , and hence offers a better handle .
it is quite plausible that the distribution of ionized plasma in and around large local features such as the local bubble and loop i can considerably influence the dispersion and scintillation of nearby pulsars , and in some cases even scintillation of extra - galactic radio sources .
the large - scale distribution of free electrons and turbulent scattering plasma in the galaxy has been modeled by taylor & cordes ( 1993 ; hereafter tc93 ) .
while sophisticated compared to its predecessors , the tc93 model takes very little account of the peculiar properties of the lism due to individual interstellar features .
the recent years have seen an accumulation of observational evidence for the effects of such features in the lism .
early investigations include the work of phillips & clegg ( 1992 ) , who proposed that the scattering of radiation from the nearby pulsar psr b0950 + 08 is probably dominated by weakly turbulent plasma present in the interior of the local bubble , and that of hajivassiliou ( 1992 ) , who invoked an ellipsoidal envelope of highly turbulent plasma to explain the directional anisotropy seen in the turbulent intensity maps derived from interplanetary scintillation studies of radio sources .
even earlier , rickard & cronyn ( 1979 ) had suggested scattering from the outer shell of loop i as the plausible cause of a statistically significant lack of interplanetary scintillators seen in a band some @xmath8 outside the brightest section of the loop i radio emission , the north polar spur ( nps ) . in a previous paper ( bhat , gupta & rao 1998 ; hereafter bgr98 ) , we presented a detailed , systematic study of the lism using pulsar scintillation data from the ooty radio telescope ( ort ) .
our results and analysis strongly support the view that the scattering in the lism is probably dominated by turbulent plasma at the boundaries of the local bubble .
we proposed a simple model , wherein the solar system is surrounded by an ellipsoidal shell - like structure , with a size of @xmath0100 pc in the galactic plane and @xmath0500 pc in a plane perpendicular to this ( fig . [ plotfour ] ) .
the scattering structure has its center located at @xmath02035 pc from the sun towards @xmath9 , @xmath10 . in this picture , the interior of the bubble is filled with plasma of relatively low turbulence ( characterized by a scattering strength , @xmath11 ) , whereas the shell material ( thickness @xmath0110 pc ) has a scattering strength @xmath050800 times larger than that in the ambient ism ( integrated strength of scattering , scattering measure , given by 0.11@[email protected] ) .
the contribution of the shell thus dominates the total scattering , which would imply that the scattering geometry towards many sight lines can be approximated by a `` thin screen '' placed at the bubble boundary ( location in the range @xmath020200 pc ) .
this model successfully explained the enhanced level of scattering measured towards a number of nearby pulsars .
the ooty experiment covered only a few pulsars that would be useful for a study of loop i. however , the recent results from parkes observations of a large number of southern pulsars ( johnston , nicastro & koribalski 1998 ) has significantly improved the iss data available for probing the ism in and around loop i. this motivated us to take a more detailed look at the distribution of scattering material towards loop i and beyond , a study which forms the main theme of this paper . the paper is organized as follows :
[ s : data ] describes observational data and our analysis techniques ; [ s : data - loop1 ] and [ s : data - spiral ] present the modeling of the scattering plasma associated with loop i. the uncertainties relevant to our analysis are discussed in [ s : res - loop1 ] . in [ s : res - zone ] , we consider some possible alternative models , while later sections of [ s : res ] discuss some general implications of our results for scintillation data and for the lism . in [ s : conc ] , we summarize our conclusions .
for the present analysis , we are interested in pulsars whose scintillation properties are likely to be influenced by the structure of loop i. in order to pre - select pulsars useful for this purpose , we adopt a geometry and size for loop i from the published literature .
specifically , berkhuijsen , haslam & salter ( 1971 ) derived an angular diameter of @xmath13 , and a centre at @xmath14 ; see also ea95 ) .
further , on the basis of the plausible connection between the origin of loop i and the sco - cen ob association ( e.g. egger & aschenbach 1995 ) , it is fair to assume the loop center to be near the center of mass of the association ( @xmath0170 pc ) . based on the above , we model the loop i bubble as a spherical shell of size @xmath0290 pc , with the center located at @xmath0170 pc towards ( @xmath15 ) .
this yields 52 pulsars within @xmath0 2 kpc of the sun whose sight lines intersect the projected area of loop i on the sky . on carrying out a literature search
, we find scintillation measurements to be available for only 20 of these ( see table [ tableone ] ) .
the majority of these scintillation measurements are from recent observations with the parkes and ooty radio telescopes , while a few are from arecibo ( data from johnston et al .
1998 ; bhat et al .
1999b , gothoskar & gupta 2000 , and cordes 1986 ) .
the measurements of decorrelation bandwidth ( ) and the observing frequencies ( ) are listed in columns ( 5 ) and ( 6 ) of table [ tableone ] , respectively .
the distance estimates in column ( 4 ) are derived from dispersion measures ( dms ) and the model of galactic electron density ( ) by tc93 , except for psrs j14566843 and j17441134 .
for these two pulsars , we use independent distance estimates derived from the measurement of annual trigonometric parallax ( bailes et al .
1990 ; toscano et al .
1999 ) .
although the decorrelation bandwidth measurements listed in table [ tableone ] are at different observing frequencies , we have scaled them to a common frequency of 327 mhz , assuming a kolmogorov scaling law ( i.e. , a wavenumber spectrum with a slope of 11/3 over the spatial scales of interest ; @xmath16 ) .
we note that this can potentially produce some errors , as the exact nature of the electron density wavenumber spectrum is still debated .
however , there is substantial observational evidence in favor of an @xmath17 spectrum towards many sight lines in the lism ( e.g. bhat , gupta & rao 1999a ) .
even if this were not strictly correct , most measurements from parkes will be only marginally affected by an incorrect frequency scaling .
this scaling bias may be significant for values at 1.5 ghz ; however , these are for relatively distant objects ( [email protected] kpc ) , and hence less critical for the investigation of scattering in the lism .
there are three different methods by which pulsar scintillation measurements can be used to investigate the distribution of scattering material in the galaxy : ( 1 ) using the decorrelation bandwidth , ( or its equivalent , the temporal pulse broadening time , ) this quantifies the scattering measure ( sm ) which characterizes the total amount of scattering along the line of sight .
this technique has been used extensively to investigate the large - scale distribution of in the galaxy ( cordes , weisberg & boriakoff 1985 ; cwb85 hereafter ; cordes et al .
1991 ; tc93 ) : ( 2 ) using measurements of angular broadening ( ) in conjunction with wherein the differing weighting functions of the two observables can be used to determine a more exact distribution of along the los ( e.g. gwinn , bartel & cordes 1993 ) : ( 3 ) using the hybrid method recently proposed by cordes & rickett ( 1998 ) wherein the diffractive scintillation measurements ( decorrelation bandwidth and scintillation timescale , ) , in conjunction with the pulsar proper motion and distance , can be used to obtain the distribution of along the los ( e.g. chatterjee et al .
2000 ) . of these methods , ( 2 ) is not relevant in our case , as measurements of and do not exist for bulk of the objects in table [ tableone ] . to date , proper motion measurements are known for 9 of the 20 pulsars in table [ tableone ] .
however , uncertainties are too large for the values to be useful in most cases , and we were unable to derive any meaningful results on the distribution of scattering in and around loop i. we therefore restricted ourselves to method ( 1 ) for the analysis described in this paper ( see appendix a for a detailed description of the method ) .
we first examine how well the scintillation data in table [ tableone ] agree with the predictions of the local bubble model of bgr98 , as described in
[ s : intro ] .
[ plotone ] shows a plot of the ratio of the measured to predicted decorrelation bandwidths ( ) against the pulsar distance estimates .
we introduce a quantity , @xmath19 , that is a measure of the degree of agreement between and .
this is expressed as @xmath20 ^ 2\ ] ] where @xmath21 is the total number of pulsars for which the comparison is being made .
the logarithm has been taken to give equal weight to discrepancies that are below and above unity when computing . as is obvious from fig .
[ plotone ] , there are large discrepancies ( ranging from a factor 2 to as much as @xmath050 ) for the case of the local bubble model ( shown by crosses ) . considering that most measurements in table [ tableone ] are _ not _ from observations averaged over many epochs ,
a potential explanation for part of these discrepancies may be errors due to long - term refractive interstellar scintillation ( riss ) effects ( e.g. gupta , rickett & lyne 1994 ; bhat et al .
however , most discrepancies are considerably larger than the worst case factor of 35 that can be accounted for by riss effects .
moreover , it is striking that all the ratios are less than unity , implying that the scattering strengths are systematically larger than that can be accounted for by the local bubble model .
further , a systematic trend with distance is also evident in fig .
[ plotone ] , whereby ( for the local bubble model ) tends to lie in the range 0.010.1 for pulsars beyond 1.2 kpc , but between 0.1 and 1 for those nearer than this .
in contrast , for a sample of pulsars within 1 kpc of the sun in the complementary sky ( shown by @xmath22 in fig . [ plotone ] , cf .
bgr98 ) , most ratios fall within a factor 23 of unity . the value of @xmath19 for this sample ( @xmath23 0.07 ) is about 16 times lower than that for the local bubble model applied to the current data set ( @xmath24 , as shown in table [ tablethree ] ) .
the simplest interpretation is that there is yet another source of excess scattering ( in addition to the local bubble shell ) in the general direction of loop i. a number of previous studies have revealed the existence of strong excess scattering towards several sight lines in the galaxy ( e.g. moran et al .
1990 ) . in the context of modeling the large - scale distribution of in the galaxy ,
cordes et al .
( 1991 ) propose a ` clump ' component , possibly associated with supernova shocks or regions , to explain such unusually high scattering .
these are regions of intense turbulence , with the strength of scattering many orders of magnitude larger than that in the typical ism .
however , the mean free paths expected for such clumps are rather large ( typically @xmath0510 kpc ) .
there seems to be no observational evidence for the presence of any such regions in the lism .
strong excess scattering may also arise if the sight line to a pulsar intercepts an region ( e.g. gupta et al .
1994 ) . to the best of our knowledge
, none of the pulsars in table [ tableone ] represents such a situation .
even if the lism contains strong scattering regions that are hitherto unknown , these could be relevant only for a few lines - of - sight , whereas the data in fig .
[ plotone ] suggest a common source of enhanced scattering for many sight lines .
given the current understanding of the lism , and also on the basis of the results from our earlier work , we postulate that this enhanced scattering is probably caused by the shell of the loop i bubble .
here we try to explain the observed discrepancies in decorrelation bandwidths for nearby pulsars in the direction of loop i in terms of enhanced scattering from a combination of the local bubble and the loop i bubble .
a full modeling of the problem involves the determination of the parameters characterizing their size and location , and the distribution of scattering in and around them .
this was done by bgr98 for the case of the local bubble alone .
however , in the present case , since we already have a model for the scattering from the local bubble and there are very good models for the physical dimensions of loop i , we do not really need to do such full - blown modeling .
instead , we take the following simpler approach . first , we assume a scattering model for loop i which is very similar to the multi - component model for the local bubble , i.e. , an interior filled with a plasma of relatively low turbulence ( the los - averaged strength of scattering , @xmath25 ) and a dense , highly turbulent shell in which the scattering strength is many times larger than in the ambient ism that lies beyond loop i. for this ambient medium , we assume @xmath26 ( see bgr98 ) .
second , as described in
[ s : data - sample ] , the size and structure of loop i are fairly well constrained by previous work ( e.g. berkhuijsen et al .
we simply adopt these values for characterizing the physical dimensions and location of the loop i shell in our analysis .
third , we simplify the problem by assuming the thickness of the loop i shell ( ) to be small compared with the pulsar distances involved . the problem
then essentially reduces to obtaining the best estimate of the scattering measure due to the loop i shell ( @xmath27 , where denotes the scattering strength inside the shell ) that will minimize the discrepancies between and .
note that the spherical shell geometry of loop i implies that should be a function of the los , or more precisely , of the angular distance of the los from the loop centre ( @xmath28 ) .
this is because the actual path length through the shell intercepted by the los varies with @xmath28 .
the effect will be more prominent for sight lines that are close to being `` tangential . '' in the analysis that follows , we will constrain sm as defined for a los that is normal to the shell ( i.e. , along @xmath29 ) .
fourth , instead of obtaining a solution that will give a best case minimum for all the pulsars in table [ tableone ] , we prefer to solve for the sm of the loop i shell using only two pulsars psrs j17441134 and j14566843 .
this is justified on the grounds that ( a ) for many pulsars in table [ tableone ] , the distance uncertainties and the large errors in the estimates of the decorrelation bandwidth ( due to riss effects ) , make it less meaningful to do a global fit , and ( b ) for several pulsars at distances 1 kpc , loop i may _ not _ necessarily be the _ only _ source of enhanced scattering and a global fit for could therefore be perturbed by this effect .
the choice of the above two pulsars is dictated by the following arguments .
both pulsars are heavily scattered ( johnston et al . 1998 ) and have precise , independent distance estimates from parallax measurements @xmath30 pc for psr j17441134 ( toscano et al .
1999 ) and @xmath31 pc for psr j14566843 ( bailes et al .
interestingly , the estimate for psr j17441134 is over twice the value of 166 pc derived from the model of tc93 .
the independent distance estimates allow the scattering geometry along these lines of sight to be fairly well constrained .
this is illustrated in fig .
[ plottwo ] , where the expected locations of the local bubble ( ) and loop i boundaries ( and ) in these directions are clearly marked .
the farther boundary of the loop i shell ( ) lies very close to the mid - point to both objects , and is thus optimally placed for maximum contribution to the observed decorrelation bandwidth .
such a geometry will be consistent with the enhanced level of scattering inferred for these two pulsars .
appendix a describes our method for computing the expected decorrelation bandwidth ( ) for a given distribution of the scattering material along the los .
we note that if homogeneously distributed scattering material ( i.e. , equation [ a1 ] , with the integral replaced by @xmath32 , where denotes the los - averaged strength of scattering ) were to explain the observed scattering for the lines of sight of these pulsars , the implied level of scattering strength ( = @xmath33 towards psr j17441134 , obtained from the value reported by johnston et al . [
1998 ] , after scaling for the new distance ) would be several times larger than that expected for the typical ambient ism .
this seems quite unlikely .
the lism as relevant to the local bubble model of bgr98 can be treated as an inhomogeneous scattering medium , which can be best represented by multiple components of different strengths of scattering ( ) along the los ( see equation [ a2 ] , for example ) .
this can be represented by a simpler , modified version of equation ( a2 ) ( i.e. , the right hand side consisting only the first two integrals and the last integral with the lower limit , + , replaced with + ) . if we were to attribute the entire excess scattering to material inside the local bubble shell ( located at @xmath03060 pc in this direction ; see fig .
[ plottwo ] ) , then would have to be over an order of magnitude larger than that derived by bgr98 ; alternatively , the shell would have to be much more extended in this direction ( say , by an order of magnitude ) .
again , either of these possibilities is rather unlikely , as it would require a drastic change in the properties of the local bubble in this direction .
hence the possibility that the loop i shell is the source of the enhanced scattering is most plausible .
we now estimate the optimal value of by modeling the scattering towards psrs j17441134 and j14566843 , taking into consideration the scattering due to the local and loop i bubbles , as well as that due to the distributed plasma ( i.e. , ambient ism ) beyond loop i. our model also takes into consideration the characteristic fall - off with z - height for the ambient scattering plasma ( scale height @xmath0500 pc ; see tc93 ) , and a gaussian fall - off ( scale height @xmath0125 pc ; see bgr98 for arguments ) for the shell material .
the interiors of both bubbles are taken to be filled with weakly turbulent plasma ( @xmath34 ) .
the shell of the local bubble is taken to have a scattering measure ( ) of 0.2 ( i.e. , @xmath0 a few 100 times larger than that in the ambient medium ) .
the method for computing the expected decorrelation bandwidth ( ) for such a geometry is described by equation ( a2 ) in appendix a. to determine , we start with an initial value of 0.2 , and vary this from 0.02 to 2.0 in steps of 0.01 .
interestingly , the best agreement between the predicted values and those measured is obtained for @xmath23 0.29 , a value quite comparable to that inferred for the local bubble shell . using the sm of the loop i shell , as constrained by the above method
, we compute new values for the ratios .
these are shown by the filled star symbols ( @xmath35 ) in fig .
[ plotone ] .
it is quite remarkable that the above modeling successfully removes the discrepancies for pulsars out to a distance of a few 100 pc .
the @xmath19 value for this model ( @xmath360.55 ) is a significant improvement over that for the earlier model of the local bubble alone , as listed in table [ tablethree ] .
further , as can be seen from this table , there is a dramatic improvement in ( i.e. , for pulsars within 1 kpc ) , and a much smaller one in exactly as expected for the local bubble + loop i bubble model .
next we address the systematic downward trend for the ratio in fig .
[ plotone ] for pulsars beyond @xmath01.2 kpc
. a closer examination of these loss suggests material within the sagittarius spiral arm to be the most plausible source of enhanced scattering for several of the pulsars .
we show here that these results can be better explained by taking into account the enhanced level of scattering that can be expected due to this spiral arm as per the existing model of tc93 . with the arm parameters of the tc93 model and the distance
estimates based on this model , the sight lines of six of our pulsars pass through this spiral arm . to model the effect of this ,
we have simply adopted the arm locations and density parameters of tc93 .
they consider a squared hyperbolic secant for the z - dependence and a gaussian fall - off for the radial dependence for the arm density , with @xmath37 pc for the scale height and 300 pc for the half width . with an arm density of 0.08
i.e. , several times larger than the typical in the lism , and a fluctuation parameter ( @xmath38 ) of @xmath39 , the implied level of scattering is some 2 orders of magnitude larger inside the arm ( @xmath00.05 ) than in the lism .
hence , substantial amounts of enhanced scattering are expected for these pulsars . on incorporating this ,
the discrepancies are further reduced ( see the open circles in fig . [ plotone ] ) , with a significant improvement in overall agreement ( @xmath40 ) as well as for ( see table [ tablethree ] ) .
the best agreement is seen with the lower and upper limits allowed ( by the model of tc93 ) for the values of the fluctuation parameter and the scale height , respectively .
we have shown that the combined effect of enhanced scattering from the local bubble shell , the loop i bubble shell and the sagittarius arm goes a long way towards explaining the observed scattering properties of pulsars in the general direction of loop i. our new model consists of : ( a ) an ellipsoidal shell of [email protected] to account for scattering due to the local bubble , ( b ) a spherical shell of [email protected] to characterize the scattering due to loop i , ( c ) the ambient ism ( @xmath26 ) in the inter - arm region , and ( d ) the sagittarius spiral arm ( @xmath0 0.047 ) .
we now discuss some of the implications of these results .
the agreement between the measurements and the predictions of our new model is such that the scattering properties of most pulsars out to a distance of 1 kpc are successfully explained ( in the range 0.52.0 for 8 out of 10 objects ) .
the major outliers are psrs j1713 + 0747 and j17302304 , with residual discrepancies of @xmath41 and @xmath42 times respectively between the predicted and measured values ( see fig .
1 ) that remain to be explained . these discrepancies are possibly due to riss , or distance errors , or perhaps as yet un - modeled source(s ) of enhanced scattering . for more distant pulsars ( d 1.2 kpc ) , although the discrepancies are better accounted for , there still seems to be a systematic downward trend with distance , which is accounted for by extra enhanced scattering due to the sagittarius arm .
the results thus clearly substantiate the role of loop i as a source of enhanced scattering in the lism .
we now address the various sources of uncertainties relevant to our analysis .
these include : ( a ) the uncertainties in size and geometry of ( i ) the local bubble , and ( ii ) loop i , ( b ) the uncertainty in the ( integrated ) scattering strength of the local bubble ( ) , ( c ) distance uncertainties , and ( d ) errors on the measurements of .
first , we note that the local bubble geometry itself is not very well determined ( see bgr98 ) , and the expected location of the boundary is in the range @xmath02458 pc for psr j17441134 , and @xmath04474 pc towards psr j14566843 ( see fig .
[ plotfour ] ) . taking this into consideration ,
could be in the range 0.250.3 .
however , the uncertainties in the angular size and location of loop i itself are small ( see berkhuijsen et al . 1971 ) , hence we do not expect them to affect appreciably . turning to ( b ) , the strength of scattering of the local bubble shell ( ) itself has a significant uncertainty : 0.11@[email protected] ( see bgr98 )
. however , this seems to affect only marginally ( 0.240.32 ) .
the distances to psrs j17441134 and j14566843 have @xmath010% uncertainties , which translate in to a variation of 0.280.33 for . as for the measurements of , the dominant source of errors are due to their apparent variations caused by riss effects on time scales from days to weeks
. however , these uncertainties are hard to estimate .
nevertheless , if we consider a typical case of a factor of two variation in the measured values of , may lie in the range 0.150.57 .
hence , the riss - induced errors in measured decorrelation bandwidths are likely to be the dominant source of errors in the final estimate of the strength of scattering of the loop i shell material . in the above analysis
, we have attempted to understand the observations in terms of enhanced scattering caused by turbulent plasma within the shells of the two bubbles .
although this two - bubble model seems to successfully explain the scintillation measurements for most nearby pulsars , there is an interesting alternative that deserves consideration . as described in
[ s : intro ] , the work of ea95 and egger ( 1998 ) suggest that the two bubbles are probably undergoing a collision process ; the observational evidence for a dense interaction feature at @xmath070 pc towards sco - cen ( fig .
4 of ea95 ) supports this .
interestingly , the expected location of the closer boundary of loop i ( ) is either within the range of or near the local bubble boundary ( ) as constrained by bgr98 ( see fig .
[ plotfour ] ) .
in particular , for several lines - of - sight , lies in between the smaller and larger boundaries ( and respectively ) of the local bubble .
the possibility of enhanced scattering from an `` interaction zone '' between the two bubbles is therefore worthy of consideration . to examine this possibility
, we consider a model where the measured level of enhanced scattering is _ entirely _ attributable to an `` interaction wall '' or `` zone '' whose effective location is at the mean distance of the lb and loop i boundaries .
the two extreme possible geometries for the lb would thus imply two possible locations for the interaction zone ( iz ) : = 0.5(+ ) and = 0.5(+ ) , which correspond to the cases of nearer and farther boundaries of the lb , respectively .
these are listed in column ( 5 ) of table [ tabletwo ] ( = /d ) . in the discussion
below , we will refer to these as the iz - a and iz - b models , respectively .
we also assume that the thickness of this zone ( ) is comparable to that of the lb or li shell , and also that it is much smaller than the pulsar distances .
further , like the case of the loop i shell , the path length of the sight line through a possible interaction zone region , and consequently its contribution to the scattering measure , will be a function of the angular distance of the los from the loop centre . in order to determine the scattering measure of the iz ( ) , we performed a rather similar analysis to that described in
[ s : data - model ] , in which psrs j17441134 and j14566843 are used to estimate the value for . for the iz - a model ( =) , the best agreement between the measured and predicted values for
is obtained for @xmath231.12 , which is @xmath04 times larger than the value of sm for the loop i shell , and @xmath0410 times larger than that for the lb .
the required sm for the iz - b case is somewhat smaller : @xmath03 times the value for , and @xmath039 times .
this is mainly due to the relatively further location of the zone in this case . using these values for , we re - computed new values for .
the results for the iz - a model are shown in fig .
[ plotthree ] .
the agreement is _ not _ as good as that achieved with the two - bubble model , as indicated by somewhat larger values for the three @xmath43 parameters that quantify the level of agreement ( see table [ tablethree ] ) . on a closer comparison of figs .
[ plotone ] and [ plotthree ] , it is obvious that the agreement is somewhat poorer for several objects at d 1 kpc
. nevertheless , the iz model(s ) may still be considered as a possible alternative to the two - bubble model . however ,
unless there are well studied objects located in the loop i interior , it is not easy to distinguish conclusively between these two scenarios . as a natural extension of the iz model(s )
, we incorporate scattering due to the sagittarius spiral arm ( as discussed in [ s : data - spiral ] ) in order to improve upon the agreement for pulsars beyond 1 kpc with sight lines intercepting that arm .
the results for the iz - a+sa case are shown in fig .
[ plotthree ] .
the overall agreement , as indicated by = 0.27 and 0.25 respectively for the iz - a+sa and iz - b+sa cases , is somewhat poorer than that achieved with the two - bubble+sa scenario ( = 0.22 , see table [ tablethree ] ) . _
( a ) interpretation of the scintillation data _ from the results of this work
it is quite evident that the structure of the lism plays an important role in the interpretation of the scintillation data .
the outer shells of the local bubble and loop i may very well be the dominant sources of scattering towards many directions in the lism , especially at higher galactic latitudes .
interestingly , independent evidence in favor of similar close - by scattering screens ( @xmath025250 pc ) comes from recent observations of cm - wave iss and intra - day variability ( idv ) of quasars ( e.g. , dennett - thorpe & de bruyn 2000 ; rickett 2000 ) .
in particular , the screen location of @xmath025 pc and strength of scattering , @xmath230.2 , inferred by dennett - thorpe & de bruyn ( 2000 ) towards the idv quasar j1819 + 3845 is in very good agreement with the location of the local bubble boundary ( @xmath02550 pc ) and the scattering strength ( ) expected in this direction .
further , from the work of hjellming & narayan ( 1986 ) , the bulk of the scintillation ( riss ) of the radio source pks 1741038 at 1.49 ghz is caused by a single screen located at @xmath0140 pc from the sun .
interestingly , the los to this object ( @xmath44 ) lies very close to the `` inner ridge feature '' of the nps ( beginning at @xmath45 ) . also , hjellming & narayan infer that there is significant excess scattering ( @xmath46 ) along this los . recently , lazio et al .
( 2000 ) report multi - epoch vlbi observations of this object as it underwent an extreme scattering event ( ese ) ; the inferred level of scattering for the ese lens is orders of magnitude larger than that in the ambient medium . given the close proximity of the object to loop i
, it is quite probable that the structure(s ) that caused the ese are located at the loop i shell . similarly ,
if an interaction wall between the two bubbles exists , it could potentially act as a thin scattering screen for nearby pulsars with sight lines within @xmath47 and @xmath48 .
such close - by scattering screens may also explain the shorter than expected timescales seen with the slow intensity variations of pulsars and the low frequency variables ( gupta , rickett & coles 1993 ; spangler et al .
in addition , the enhanced scattering from such bubble shells may be responsible for some of the unusual scattering effects . in this context
, it is interesting to note that the location of the scatterer that caused the multiple imaging event of psr b1133 + 16 in the ooty data ( gupta , bhat & rao 1999 ) was found to match well with the expected location of the local bubble shell ( @xmath23 0.77d ) in this direction .
similarly , it has also been suggested that the the structures that cause extreme scattering events in the radio light curves of some quasars are probably associated with the galactic loops ( fiedler et al .
recent work by toscano et al .
( 1999 ) and chatterjee et al .
( 2000 ) provides some evidence for enhanced scattering in the third galactic quadrant probably associated with an interface region between the local bubble and the gsh 238 + 00 + 09 super - bubble .
all of the above arguments clearly support the view that the structure of the lism needs to be taken into account for a proper interpretation of dispersion and scintillation data .
_ ( b ) on the nature of the bubble shells _ in our earlier work ( bgr98 ) , we argued that the elongated ellipsoidal shell of turbulence derived from the ooty data is possibly associated with the local bubble .
the analysis described in [
s : data - loop1 ] presents quite convincing evidence for an outer turbulence shell for loop i also ( the existence of a turbulent interaction zone is an alternative possibility , though ) .
thus , from observational data , it appears that interstellar bubbles in general may have turbulent outer shells . on the theoretical front , a number of authors
have dealt with the evolution of bubbles in the ism ( ikeuchi 1998 and references therein ) ; in particular , special attention has been paid to the major local features such as the local and loop i bubbles .
interestingly , it turns out that turbulent outer shells can indeed be expected for interstellar bubbles .
specifically , the recent model of breitschwerdt , freyberg & egger ( 2000 ) predicts turbulence to exist at the interface region of the local and loop i bubbles as a natural consequence of the hydro - magnetic rayleigh - taylor instability caused by the interaction of the two bubbles .
a similar mechanism may also apply for the loop i shell . in an earlier work , breitschwerdt & kahn ( 1989 ; see also kahn & breitschwerdt 1988 ) showed that turbulent shells could also be expected for stellar - wind blown bubbles as a consequence of acoustic instability .
so , in general , the shells around bubbles may be expected to be turbulent .
there are no explicit predictions available for the level of turbulent intensity expected at the shell or in interaction zone regions , or for a relation connecting the turbulent intensity to the physical parameters governing the process(es ) . nevertheless , the basic results from our investigations using pulsar data appear to be in accordance with some of the theoretical models of bubble evolution .
the strength of scattering we derived for the outer shell of loop i is substantially higher ( @xmath0100500 times depending on the shell thickness ) than that measured towards typical loss in the lism .
the regions of enhanced electron density ( ) are also thought to be plausible sites for enhanced scattering ( e.g. the gum nebula ) , though the exact relationship between these two is not clearly known .
if we consider to be uniform within the shell or the interaction zone regions ( i.e. , .@xmath49 , @xmath49 , where and represent the path lengths through the shell and interaction zone regions respectively ) , and assume a simple relation ( @xmath50 ) , then an indirect estimate of electron density in the shell ( ) is possible for a given shell thickness . considering the loop i shell ( [email protected] ) , for a shell thickness @xmath0110 pc , this yields a density @xmath01030 times larger than the ambient ism value .
somewhat higher values ( @xmath02050 ) result for the interaction zone models described in [ s : res - zone ] .
interestingly , a similar level of density enhancement is seen for the neutral gas at the annular interface region of the two bubbles ( ea95 ; egger 1998 ) . for the above densities
, the effects on dispersion due to the shell region will be significant only for pulsars within a distance of a few 100 pc . taking this into consideration would , however , result in a _ much lower _ ambient density ( @xmath230.007 ) towards psr j17441134 , the disk pulsar with the lowest known . for a similar density enhancement ( @xmath010 times ) for the shell region towards psr j14566843 ( the only pulsar in quadrant 4 with a measured parallax ) , consideration of shell dispersion would result in an ambient density of @xmath23 0.016 for the two - bubble model , and @xmath23 0.013 for the interaction zone model .
these values are comparable to measured towards several objects in quadrant 1 with known parallaxes ( see toscano et al .
therefore , it is quite plausible that the loop i shell ( or the interaction zone ) contributes significantly to the dispersion of nearby pulsars in quadrant 4 .
it is interesting to note that for the two pulsars psrs j14566843 and j17441134 with independent distance estimates , the mean electron densities along the sight lines differ significantly : towards the former is almost twice that for the latter , despite their comparable distances and z - heights ( 67 and 57 pc respectively )
. a closer look at their loss _ vis - a - vis _ the geometries of the local and loop i bubbles ( figs .
[ plottwo ] and [ plotfour ] ) indicates the following : ( 1 ) the local bubble cavity extends out almost twice as far towards psr j14566843 as towards psr j17441134 ( @xmath04474 pc and @xmath02458 pc respectively , for the geometry in fig . [ plotfour ] ) , ( 2 ) the interior of loop i covers a much longer section of the los towards psr j14566843 ( @xmath49 @xmath23 233 pc for j14566843 , compared to 157 pc for psr j17441134 ; see fig .
[ plottwo ] ) , and ( 3 ) the extent of the ambient medium beyond loop i is not much larger along the sight line towards psr j14566843 ( @xmath0192 pc compared to @xmath0159 pc for psr j17441134 ) .
all of these make it harder to explain the larger towards psr j14566843 .
there appears to be two plausible explanations for the above : ( a ) the ambient ism in quadrant 4 is simply much denser than that in quadrant 1 .
however , in the absence of other pulsars in quadrant 4 with measured parallax , there seems to be no independent way to confirm this .
( b ) the existence of a dense interface region between the local and loop i bubbles , as described in
[ s : res - zone ] . on a closer inspection of the sight lines of the two pulsars , case ( b ) seems the more likely .
the loss towards psr j17441134 intercepts the `` annular ring , '' as recognized by ea95 ( see also egger 1998 ) , whereas the los toward psr j14566843 lies well within the `` interaction zone region '' confined by this ring - like feature .
if the interaction zone region is filled with a denser , highly turbulent plasma , then it could potentially account for much of the dispersion towards psr j14566843 .
if we assume an ambient density of @xmath23 0.01 ( i.e. , a value comparable to that in quadrant 1 ) , the interaction zone density ( ) will have to be @xmath040 times larger in order to account for the excess dispersion .
we note that this is quite comparable to the value @xmath02050 derived from the sm estimate as constrained by our modeling .
the consistency of the density estimates from the two independent methods can be argued , in some sense , to favor the existence of a dense interaction wall .
measurements of parallax for other nearby pulsars in quadrant 4 will help to test for the existence of a wall .
the most promising candidates for this seem to be low - dm ( say , 20 ) objects with loss intercepting the interaction zone region .
for instance , psr j17514657 would be located much farther ( @xmath01765 pc ) than its tc93 distance ( @xmath01080 pc ) , if the ambient in this quadrant is comparable to that in quadrant 1 ( @xmath00.01 ) .
psr j17302304 ( close - by to psr j17441134 ) is another interesting test case , whose distance will only be slightly more ( @xmath23555 pc ) if a denser zone is present , while it will be located much farther away ( @xmath01 kpc ) if the entire sight line was uniformly filled at @xmath00.01 .
similarly , the distance of psr j14553330 will be @xmath0950 pc ( compared to tc93 value of @xmath0740 pc ) if we consider a dense zone and a low - density ( @xmath230.01 ) ambient medium . while our arguments are based on a fairly simple picture , it is amply clear that measurements of a few interesting test cases will be valuable for understanding the lism in this quadrant . from the work presented here and related work in the recent past ( see [ s : res - impli ] )
, it is clear that pulsars can be used quite effectively to probe large - scale features in the lism .
conversely , it is also obvious that detailed interpretation of pulsar dispersion and scintillation data will need to take the structure of the lism into account .
there is growing observational evidence that the local bubble is surrounded by numerous bubbles of similar properties , many of which are likely to be filled with hot x - ray emitting gas .
while this work has concentrated on loop i , our most prominent such neighbor , other nearby examples include the eridanus bubble , the gum nebula and possibly radio loops ii and iii .
further , there are also observations indicating possible interactions of some of these with the local bubble . in this paper
, we have considered the possibility of an interaction region between the local and loop i bubbles .
toscano et al .
( 1999 ) and chatterjee et al .
( 2000 ) have provided some evidence for interaction of the local bubble with the gsh 238 + 00 + 09 super - bubble in galactic quadrant 3 .
the closeness between the location of the near side of the eridanus bubble ( 159@xmath5116 pc ; guo et al .
1995 ) and the expected location of the local bubble shell in this direction ( @xmath23130 pc ) suggests a possible interaction between these two bubbles .
the vela supernova remnant has been shown to be embedded in a hot bubble confined by the shell of the gum nebula ( aschenbach , egger & trmper 1995 ) . with the recently revised estimate for its distance ( 250@xmath5130 pc by cha , sembach & danks 1999 ) ,
the near side of the `` gum - bubble '' is quite close to the local bubble boundary ( @xmath23100 pc ) ; this is also supported by the observation that the absorbing column density towards the vela snr is only about @xmath52 .
an interaction between these two bubbles therefore seems quite likely . of the 163 pulsars known within @xmath02 kpc of the sun , scintillation data
are presently available for only 73 , and independent distance estimates for only 12 .
clearly , a lot more needs to be done here .
the new generation of large telescopes , such as the gmrt and the gbt , could potentially extend the available scintillation data for nearby weak pulsars .
for many such cases , measuring pulse - broadening times at low frequencies ( 100 mhz ) may prove to be a more viable technique than decorrelation bandwidth measurements .
observations of sources along well - chosen lines of sight where the scattering is dominated by the shells of these bubbles may also result in more detections of unusual scattering effects , such as multiple imaging and extreme scattering events .
further , iss studies of radio sources through cm - wave iss and idv may also provide useful insights into the nature of the lism at higher galactic latitudes .
many such observations hold great promise for extending the rather simplistic models derived from the present investigations into a more realistic picture of the lism .
we have investigated the distribution of scattering plasma in the lism in the general direction of the loop i bubble by combining recent pulsar scintillation measurements .
many pulsars within @xmath02 kpc , and located towards @xmath53 and @xmath54 , show enhanced levels of scattering , detected as a significantly reduced value of the ratio of the measured to expected decorrelation bandwidths .
the discrepancies can not be explained by the local bubble model ( bgr98 ) alone , and we interpret them as being due to enhanced scattering associated with the shell of the loop i bubble and the sagittarius spiral arm . using data from two heavily scattered pulsars with precisely known distances , psrs j17441134 and j14566843
, we have placed useful constraints on the scattering strength associated with the loop i shell .
the scattering measure inferred for the plasma inside this shell is @xmath00.3 , quite similar to that for the local bubble shell . assuming a shell thickness @xmath0110 pc , this implies an average turbulence level @xmath0100500 times larger than that in the ambient ism . adopting this , our earlier model of the lism
is extended by incorporating an explicit scattering component for plasma in and around loop i , in addition to that due to the local bubble .
this two - bubble model successfully explains many of the observed decorrelation bandwidth discrepancies for nearby pulsars .
the alternative possibility of enhanced scattering from an interaction zone between the local bubble and loop i is also considered .
however , the level of turbulence in the interaction zone would have to be several times larger than that for the shells in the two - bubble model in order to explain the observations .
even then , the final agreement with observations for the interaction zone model is not as good as that for the two - bubble model . assuming a simple relation between the scattering strength and the free electron density ( @xmath55 ) yields a shell density that is @xmath01030 times larger than the ambient value , and a somewhat higher density ( @xmath02050 ) for the interaction zone model . for several low - latitude pulsars at d@xmath012 kpc ( @xmath56300 pc ) ,
we find that the observed scattering discrepancies are consistent with additional enhanced scattering from the sagittarius spiral arm .
we have discussed implications of our results for the interpretation of scintillation data as well as for the structure of the lism . in the light of our results and those from several other recent works , we conclude that the structure of the lism needs to be considered in the interpretation of pulsar dispersion and scintillation data , and may also be relevant for observations of cm - wave iss .
further , the general picture that emerges from the investigations in support of turbulent outer shells and/or interface regions of interstellar bubbles appears to be in qualitative agreement with the expectations of the models that describe the evolution of bubbles in the ism .
_ acknowledgments : _ we thank c. salter for several fruitful discussions and valuable comments on the earlier versions of this paper which helped us to improve upon the content .
we also thank p. b. preethi for help with the analysis software during early stages of this work .
here we describe the method adopted for computing the decorrelation bandwidth ( ) .
[ plottwo ] depicts typical scattering geometries relevant in our analysis : is the expected location of the local bubble boundary , and denote the nearer and farther boundaries , respectively , of the loop i bubble .
d is the pulsar distance . where is a model dependent constant , denotes the frequency of observation , and @xmath58 is the slope of the electron density wavenumber spectrum . in this scheme ,
the observer is at @xmath59 and the pulsar at @xmath60 . if we adopt the canonical value of 11/3 for @xmath58 ( i.e. , kolmogorov - like spectrum ) , @xmath61 . the lism model corresponding to fig .
[ plottwo ] can be treated as an inhomogeneous scattering medium , with multiple components of different strengths of scattering ( , also called turbulent intensity or scattering strength ) , located at different points along the line of sight .
the observable is determined by the path length differences of scattered rays ; therefore , contributions to it from need to be appropriately weighted in such a way that scattering regions near the source or the observer produce smaller path - length differences than those that are mid - way ( cwb85 ) .
hence , for the case of our interest , equation ( a1 ) can be rewritten as where and are the thickness of local bubble and loop i shells , respectively .
the turbulent intensity in the local bubble interior and in its shell are represented by and , respectively , and and are the equivalent quantities for the loop i bubble .
the scattering plasma in the ambient ism is characterized by .
the symbol is the ` weighting function ' for , and is given by this simple function is symmetric with respect to the mid - point between observer and pulsar , which means an inherent ambiguity is involved in the interpretation of the underlying scattering geometry .
for instance , in the simple case of a thin screen placed between us and a pulsar , a screen closer to pulsar ( say , at @xmath64 ) will be equivalent to the one placed at @xmath65 from us . in the case of an extended , inhomogeneous medium ( as in fig .
[ plottwo ] ) , the ` inverse geometry ' is indistinguishable from the actual geometry .
aschenbach , b. , egger , r. j. , trumper , j. , 1995 , nature , 373 , 587 + bailes , m. , manchester , r. n. , kesteven , m. j. , norris , r. p. , & reynolds , j. e. 1990 , nature , 343 , 240 + berkhuijsen , e. , haslam , c. g. t. , salter , c. , 1971 , a&a , 14 , 252 + bhat , n. d. r. , gupta , y. , rao , a. p. , 1998 ,
apj , 500 , 262 ( bgr98 ) + bhat , n. d. r. , gupta , y. , rao , a. p. , 1999a
, apj , 514 , 249 + bhat , n. d. r. , rao , a. p. , gupta , y. , 1999b , apjs , 121 , 483 + breitschwerdt , d. , freyberg , m. j. , egger , r , j. , 2000 , a&a , 361 , 303 + breitschwerdt , d. , freyberg , m. j. , trmper , j. , 1998 , _ lecture notes in physics _ , 506 + breitschwerdt , d. & kahn , f. d. , 1989 , mnras , 235 , 1011 + centurion , m. , & vladilo , g. , 1991 , apj , 372 , 494 + cha , a. n. , sembach , k. r. , danks , a. c. , 1999 , apj , 515 , l25 + chatterjee , s. , cordes , j. m. , lazio , t. j. w. , goss , w. m. , fomalont , e. b. , benson , j. m. , 2000 , apj , submitted .
+ cordes , j. m. , 1986 , apj , 311 , 183 + cordes , j. m. , weisberg , j. m. , & frail , d. a. , et al .
1991 , nature , 354 , 121 + cordes , j. m. & rickett , b. j. 1998 , apj , 507 , 846 + cordes , j. m. , weisberg , j. m. and boriakoff , v. 1985 , apj , 288 , 221 , ( cwb85 ) + cordes , j. m. , wolszczan , a. , dewey , r. j. , blaskiewicz , m. , stinebring , d. r. , 1990 , apj , 349 , 245 + cox , d. p. , & reynolds , r. j. 1987 , ara&a , 25 , 303 + dennett - thorpe , j. & de bruyn , a. g. 2000 , apj , 529 , l65 + egger , r. j. , 1993 , ph.d .
thesis , tu muenchen + egger , r. j. 1998 , lecture notes in physics , 506 , 287 + egger , r. j. , & aschenbach , b. 1995 , a&a 294 , l25 , ( ea95 ) + fiedler , r. l. , dennison , b. , johnston , k. j. , waltman , e. b. and simon , r. s. 1994 , apj , 430 , 581 + frisch , p. c. 1981 , nature , 293 , 377 + frisch , p. c. 1996 , space sci .
78 , 213 + frisch , p. c. 1998 , lecture notes in physics , 506 , 269 + gothoskar , p. & gupta , y. 2000 , apj , 531 , 345 + guo , z. , burrows , d. n. , sanders , w. t. , snowden , s. l. , penprase , b. e. , 1995 , apj , 453 , 256 + gupta , y. , bhat , n. d. r. & rao , a. p. , 1999 ,
apj , 520 , 173 + gupta , y. , rickett , b. j. and coles , w. a. 1993 , apj , 403 , 183 + gupta , y. , rickett , b. j. , & lyne , a. g. , 1994 , mnras , 269 , 1035 + gwinn , c. r. , bartel , n. , cordes , j. m. , 1993 , apj , 410 , 673 + hajivassiliou , c. a. 1992 , nature , 355 , 232 + haslam , c. g. t. , khan , f. d. & meaburn , j. , 1971 , a&a , 12 , 388 + haslam , c. g. t. , large , m. i. & quigley , m. j. s. , 1964 , mnras , 127 , 273 + heiles , c. , 1998 , apj , 498 , 689 + hjellming , r. m. & narayan , r. , 1986 , apj , 310 , 768 + ikeuchi , s. , 1998 , lecture notes in physics , 506 , 399 + johnston , s. , nicastro , l. , & koribalski , b. 1998 , mnras , 297 , 108 + kahn , f. d. & breitschwerdt , d. , 1988 , mnras , 242 , 209 + a&a , 287 , 470 + lallement , r. 1998 , lecture notes in physics , 506 , 19 + lazio , t. j. w. , fey , a. l. , dennison , b. , et al .
, 2000 , apj , 534 , 706 + milne , d. k. , 1968 , australian j. phys . , 21 , 201 + moran , j. m. , greene , b. , rodriguez , l. f. & backer , d. c. , 1990 , apj , 348 , 147 + nishikida , k. , 1999 , ph.d .
thesis , the pennsylvania state university + phillips , j. a. & clegg , a. w. 1992 , nature , 360 , 137 + rickard , j. j. , & cronyn , w. m. 1979 , apj , 228 , 755 + rickett , b. j. , 1990 , ara&a , 28 , 561 + rickett , b. j. , 2000 , proceedings of iau 182 coll .
, kluwer academic publishers + salter , c. 1983 , basi , 11 , 1 + spangler , s. r. , eastman , w. a. , gregorini , l. , et al . 1993 , a&a , 267 , 213 + stinebring , d. r. , faison , m. d. & mckinnon , m. m. , 1996 , apj , 460 , 460 + taylor , j. h. , & cordes , j. m. 1993 , apj , 411 , 674 , ( tc93 ) + toscano , m. , britton , m. c. , manchester , r. n. , et al . 1999 ,
apj , 523 , l171 + trmper , j. , adv .
space res .
1983 , 4 , 241 + warwick , r. s. , barber , c. r. , hodkin , s. t. , pye , j. p. , 1993 ,
mnras , 262 , 28 + yoshioka , s. , & ikeuchi , s. 1990 , apj , 360 , 352 + | recent pulsar scintillation measurements from ooty , in conjunction with those from parkes and other large radio telescopes , are used for a systematic investigation of the local interstellar medium ( lism ) towards the general direction of the loop i bubble . for several pulsars ,
clear evidence is found for an enhanced level of scattering which is over and above what can be accounted for by the enhanced scattering model for the local bubble .
these results are interpreted in terms of enhanced scattering due to turbulent plasma associated with the loop i shell .
useful constraints are obtained for the scattering properties of the shell . the inferred value for the scattering measure for the loop i shell is found to be @xmath00.3 .
assuming a shell thickness @xmath0510 pc , this implies an average strength of scattering in the shell that is @xmath0100200 times larger than that in the ambient ism .
an alternative explanation , where the enhanced level of scattering is due to a possible `` interaction zone '' between the local bubble and the loop i bubble , is also considered ; it is found to be somewhat less satisfactory in explaining the observations .
the best fit value for the scattering measure for such an interaction zone region is estimated to be @xmath01.1 .
further , several pulsars beyond @xmath01 kpc are found to show enhanced levels of scattering over and above that expected from this `` two - bubble model . '' for some of the low - latitude pulsars , this is found to be due to enhanced scattering from plasma inside the intervening sagittarius spiral arm .
we discuss the implications of our results for the interpretation of scintillation data and for the general understanding of the lism . |
living matter exhibits a broad spectrum of unique phenomena which emerge as a consequence of its active constituents .
examples of such systems range from the macroscopic scale of flocks of birds and mammalian herds to the microscopic scale of bacterial suspensions @xcite . specifically , active systems exhibit remarkable nonequilibrium phenomena and emergent behavior like swarming @xcite , turbulence @xcite , and activity - induced clustering and phase transitions @xcite .
the understanding of these collective phenomena requires the characterization of the underlying physical interaction mechanisms .
experiments and simulations indicate that shape - induced interactions , such as inelastic collisions between elongated objects or of active particles with surfaces lead to clustering , collective motion , and surface - induced aggregation @xcite . for micrometer - size biological unicellular swimmers , e.g. , bacteria ( _ e .
coli _ ) , algae ( _ chlamydomonas _ ) , spermatozoa , or protozoa ( _ paramecium _ ) , hydrodynamic interactions are considered to be important for collective effects and determine their behavior adjacent to surfaces @xcite .
generic models , which capture the essential swimming aspects , are crucial in theoretical studies of microswimmers . on the one hand
, they help to unravel the relevant interaction mechanisms and , on the other hand , allow for the study of sufficiently large systems .
a prominent example is the squirmer model introduced by lighthill @xcite and revised by blake @xcite .
originally , it was intended as a model for ciliated microswimmers , such as _ paramecia_. nowadays , it is considered as a generic model for a broad class of microswimmers , ranging from diffusiophoretic particles @xcite to biological cells ( _ e .
coli , chlamydomonas _ , etc . ) and has been applied to study collective effects in bulk @xcite , at surfaces @xcite , and in thin films @xcite . in its simplest form
, a squirmer is represented as a spherical rigid colloid with a prescribed surface velocity @xcite . restricting the surface velocity to be tangential ,
the spherical squirmer is typically characterized by two modes accounting for its swimming velocity and its force - dipole .
the latter distinguishes between pushers , pullers , and neutral squirmers .
the assumption of a spherical shape is adequate for swimmers like _ volvox _ , however , the shape of bacteria such as _ e. coli _ or the time - averaged shape of cells such as _ chlamydomonas _ is nonspherical .
hence , an extension of the squirmer concept to spheroidal objects is desirable . in 1977 ,
keller and wu proposed a generalization of the squirmer model to a prolate - spheroidal shape , which resembles real biological microswimmers such as _ tetrahymenapyriformis _ , _ spirostomum ambiguum _ , and _ paramecium multimicronucleatum _ @xcite .
however , that squirmer model accounts for the swimming mode only and does not include a force - dipole mode .
this is unfortunate , since the force - dipole mode determines swimmer - swimmer and swimmer - wall interactions @xcite .
a route to incorporate the force - dipole mode into the spheroidal squirmer model was proposed in ref . .
however , to the best of our knowledge , the resulting hydrodynamic model is not solvable analytically . in this article
, we propose an alternative model for a spheroidal squirmer , taking into account both , a swimming and a force - dipole mode .
the major advantage of our approach is that the flow field can be determined analytically ( cf . fig . [ fig : theo_sf_bf ] ) .
various mesoscale simulation techniques have been applied to study the dynamics of squirmers embedded in a fluid , comprising stokesian dynamics @xcite , the boundary - element method @xcite , the multiparticle collision dynamics ( mpc ) approach @xcite , lattice boltzmann simulations @xcite , the smoothed profile method @xcite , and the force - coupling approach @xcite . in the following
, we will apply the mpc method .
mpc is a particle - based simulation technique which incorporates thermal fluctuations @xcite , provides hydrodynamic correlations @xcite , and is easily coupled with other simulation techniques such as molecular dynamics simulations for embedded particles @xcite .
the method has successfully been applied in various studies of active systems underlining the importance of hydrodynamic interactions for microswimmers @xcite . here
, we implement our spheroidal squirmer model in mpc . more specifically , we study the resulting flow field and compare it with the theoretical prediction . moreover
, we present results for the cooperative swimming behavior of two spheroidal squirmers in thin films .
two pullers exhibit a long - time stable configuration , where they swim together in a wedge - like conformation with a constant small angle due to the hydrodynamic interaction between the anisotropic squirmers as well as squirmers and walls .
the cooperative and collective swimming motion of spheroidal squirmers in stokes flow has been addressed in ref . by an adopted boundary - element method .
this approach neglects thermal fluctuations and tumbling of the squirmers completely ; only hydrodynamic and excluded - volume interactions determine the squirmer motion .
in contrast , our simulation approach includes thermal fluctuations , which affects the stability of the cooperative swimming motion due to the rotational diffusion of a spheroid .
flow field of a spheroidal puller with @xmath0 , ( a ) in the laboratory frame , and ( b ) in the body - fixed frame .
the logarithm of the magnitude of the velocity field is color coded . ]
sketch of normal and tangent vectors of a spheroidal ( left ) and spherical ( right ) squirmer . in the squirmer model , self - propulsion ( in @xmath1-direction )
is achieved by a prescribed tangential surface velocity in direction of the tangent vector @xmath2 . ]
we describe a nonspherical squirmer as a prolate spheroidal rigid body with a prescribed surface velocity @xmath3 . in cartesian
coordinates @xmath4 , the surface equation of a spheroid , or ellipsoid of revolution , is @xmath5 with @xmath6 and @xmath7 the semi - major and semi - minor axis , respectively , and @xmath8 ( cf .
[ fig : sketch_spheroid ] ) . we denote half of the focal length by @xmath9 , which yields the eccentricity @xmath10 .
furthermore , we define a swimmer diameter as @xmath11 . in terms of prolate ( @xmath12 ) spheroidal coordinates @xmath13 , the cartesian coordinates are given by @xmath14 where @xmath15 , @xmath16 , and @xmath17 .
all points with @xmath18 lie on the spheroid s surface .
the intersection of the spheroid and a meridian plane , where @xmath19 is constant , is an ellipse .
the normal @xmath20 and tangent @xmath2 to this ellipse are given by the unit vectors @xmath21 and @xmath22 , respectively , which follow by partial derivative of eqs .
( [ eq : xyz_of_tau_zeta_phi ] ) with respect to the coordinates @xmath23 and @xmath24 . for @xmath25 ,
the spheroid becomes a sphere .
the spherical coordinates @xmath26 are obtained from eq .
( [ eq : xyz_of_tau_zeta_phi ] ) for @xmath27 , and @xmath28 . in this limit ,
the unit vectors turn into @xmath29 and @xmath30 ( cf . fig .
[ fig : sketch_spheroid ] ) .
the lam metric coefficients for prolate spheroidal coordinates are @xmath31 , @xmath32 and @xmath33 . + the squirmer is immersed in an incompressible low - reynolds - number fluid , which is described by the incompressible stokes equations @xmath34 here , @xmath35 is the fluid velocity field , @xmath36 the pressure field at the position @xmath37 , and @xmath38 the viscosity . in an axisymmetric flow , the velocity field
can be expressed by the stream function @xmath39 as @xcite @xmath40 the stream function itself satisfies the equation @xcite @xmath41 with the operator @xcite @xmath42 each function @xmath39 in the kernel of @xmath43 can be represented as @xcite @xmath44 with constants @xmath45 and the functions @xmath46 here , @xmath47 and @xmath48 are gegenbauer functions of the first and second kind , respectively ( see appendix [ app : gegenbauer ] ) .
the velocity components follow from the stream function via @xcite @xmath49 an important feature of a squirmer is the hydrodynamic boundary condition at its surface , which demands @xmath50 . for the squirming velocity @xmath3 we propose @xmath51 here , @xmath2 is the tangent vector , @xmath52 is the unit vector in @xmath1-direction , @xmath53 and @xmath54 are the two surface velocity modes , and @xmath55 ( cf .
[ fig : sketch_spheroid ] ) . @xmath53 determines the swimming velocity , while the @xmath54 term introduces a force - dipole , or pusher ( @xmath56 ) and puller ( @xmath57 ) mode .
note that the spherical squirmer introduced by lighthill and blake with modes @xmath53 and @xmath54 @xcite is recovered for the spherical limit of a spheroid , where @xmath58 .
+ for @xmath59 , this model of a spheroidal squirmer was already introduced and analysed in refs . and .
an additional force - dipole mode has been introduced in refs . and as @xmath60 .
however , we prefer the squirming velocity introduced in eq .
( [ eq : squirming_vel_spheroid_analytical_new ] ) , since it yields an analytically solvable boundary value problem for the stokes equation .
the two approaches provide a somewhat different flow field in the vicinity of the squirmer , but both yield the model of lighthill and blake in the limit of zero eccentricity . in the swimmer s rest frame , and with eq .
( [ eq : squirming_vel_spheroid_analytical_new ] ) , the boundary value problem becomes @xmath61 equation ( [ eq : spheroid_squirmer_bc1 ] ) implies a constant background flow @xmath62 infinitely far from the squirmer , eq .
( [ eq : spheroid_squirmer_bc2 ] ) guarantees @xmath63 at the spheroid surface , and eq .
( [ eq : spheroid_squirmer_bc3 ] ) demands @xmath64 . due to linearity of the stokes stream function equation ( [ eq : stokes_stream_function_equation ] )
, we can solve this boundary value problem for @xmath59 first , which yields the stream function @xmath65 .
subsequently we solve the problem @xmath66 equation ( [ eq : spheroid_squirmer_b2_bc1 ] ) imposes a vanishing velocity field infinitely far from the squirmer , eq .
( [ eq : spheroid_squirmer_b2_bc2 ] ) again guarantees @xmath63 at the spheroid surface , and eq .
( [ eq : spheroid_squirmer_b2_bc3 ] ) demands @xmath67 .
we denote the solution of the problem eqs .
( [ eq : spheroid_squirmer_b2_bc1])-([eq : spheroid_squirmer_b2_bc3 ] ) by @xmath68 .
finally , @xmath69 solves the initial problem ( [ eq : spheroid_squirmer_bc1])-([eq : spheroid_squirmer_bc3 ] ) for arbitrary @xmath53 and @xmath54 .
the boundary value problem eqs .
( [ eq : spheroid_squirmer_bc1])-([eq : spheroid_squirmer_bc3 ] ) for @xmath59 can be solved by the ansatz @xmath70 here , the third term is found by the separation ansatz @xmath71 for eq .
( [ eq : stokes_stream_function_equation ] ) . equation ( [ eq : spheroid_squirmer_bc1 ] )
directly yields @xmath72 .
the remaining coefficients @xmath73 and @xmath74 are determined by eqs .
( [ eq : spheroid_squirmer_bc2 ] ) and ( [ eq : spheroid_squirmer_bc3 ] ) , keeping in mind that @xmath59 .
this yields @xmath75 \label{eq : alpha3 } \alpha_3&=c^2 \frac{b_1 \tau_0(\tau_0-(\tau_0 ^ 2 - 1)\coth^{-1 } \tau_0 ) -u_0 } { ( \tau_0 ^ 2 + 1 ) \coth^{-1 } \tau_0 -\tau_0}.\end{aligned}\ ] ] the boundary value problem eqs .
( [ eq : spheroid_squirmer_b2_bc1])-([eq : spheroid_squirmer_b2_bc3 ] ) can be solved by the ansatz @xmath76 as before , the third term follows by a separation ansatz @xmath77 for eq .
( [ eq : stokes_stream_function_equation ] ) .
equation ( [ eq : spheroid_squirmer_b2_bc1 ] ) yields @xmath78 .
the coefficients @xmath79 and @xmath80 are determined by eqs .
( [ eq : spheroid_squirmer_b2_bc2])-([eq : spheroid_squirmer_b2_bc3 ] ) such that @xmath81 \alpha_6&=c^2 b_2 \tau_0 \frac { 2/3-\tau_0 ^ 2+\tau_0(\tau_0 ^ 2 - 1)\coth^{-1 } \tau_0 } { 3 \tau_0+(1 - 3 \tau_0 ^ 2)\coth^{-1 } \tau_0}. \label{eq : alpha_6_equals}\end{aligned}\ ] ] the total stream function @xmath69 can be transformed to the laboratory frame ( cf .
[ fig : theo_sf_bf ] ) by adding the background flow @xmath82 , which yields @xmath83 the force by the fluid on the spheroid is given by @xcite @xmath84 where @xmath85 and @xmath86 . as expected
, @xmath68 does not contribute to the force , since it assumes a constant value at infinity . since
a swimmer must be force free , @xmath87 , which implies @xmath88 .
then , eq . ( [ eq : alpha3 ] )
yields the swimming velocity of the squirmer ( @xmath89 ) @xmath90 which was already found by keller and wu for the case @xmath59 @xcite . as a consequence , @xmath73 in eq .
( [ eq : alpha2 ] ) simply becomes @xmath91 .
+ the flow field of a point - like force - dipole is given by @xcite @xmath92 with the dipole strength @xmath93 , whereas the flow field of a source doublet is @xcite @xmath94 with the source - doublet strength @xmath95 . comparing the corresponding stream functions with eq .
( [ eq : stream_function_lab_frame ] ) far from the origin , we find @xmath96 for our model .
as expected , in the spherical limit ( @xmath97 , where @xmath98 is the radius ) we obtain @xmath99 and @xmath100 .
examples of fluid velocity fields of a spheroidal squirmer are presented in figs .
[ fig : theo_sf_bf ] and [ fig : theo_b1_b2 ] .
fluid velocity fields of a spheroidal squirmer in the laboratory frame for ( a ) @xmath101 , and ( b ) @xmath102 .
the corresponding stream function is given by eq .
( [ eq : stream_function_lab_frame ] ) .
the logarithm of the magnitude of the velocity field is color coded .
note that the pusher velocity field with @xmath103 is not shown , since it follows from that of the puller with @xmath104 by inverting the arrows . ]
multiparticle collision dynamics ( mpc ) is a stochastic , particle based mesoscale hydrodynamic simulation method @xcite .
thereby , a fluid is modeled by @xmath105 point particles with equal mass @xmath106 , undergoing subsequent streaming and collision steps . in the streaming step ,
the particle positions @xmath107 , are updated according to @xmath108 where @xmath109 are the particle velocities and @xmath110 is denoted as collision time step . in the subsequent collision step ,
the particle velocities are changed by a stochastic process , which mimics internal fluid interactions . in order to define the local collision environment ,
particles are sorted into cells of a cubic lattice with lattice constant @xmath111 .
different realizations for this stochastic process have been proposed.@xcite we employ the stochastic rotation dynamics ( srd ) approach of mpc with angular momentum conservation ( srd+a ) @xcite , which updates the particle velocities in a cell according to @xmath112 .
\end{split}\end{aligned}\ ] ] here , @xmath113 , where @xmath114 is the center - of - mass position of the particles in the cell , and similarly , @xmath115 , with the center - of - mass velocity @xmath116 .
@xmath117 is the rotation matrix , which describes a rotation around a randomly oriented axis by the angle @xmath118 .
the angle @xmath118 is a constant , and the axis of rotation is chosen independently for each cell and time step .
finally , @xmath119 is the moment - of - inertia tensor of the particles in the center - of - mass reference frame of the cell .
partition of the system into collision cells leads to a violation of galilean invariance . to reestablish galilean invariance ,
a random shift of the collision - cell lattice is introduced at every collision step @xcite .
since energy is not conserved in the collision step , we apply a cell level canonical thermostat at temperature @xmath120 @xcite .
the latter ensures maxwell - boltzmann distributed velocities .
the mpc algorithm is embarrassingly parallel .
hence , we implement it on a graphics processing unit ( gpu ) for a high performance gain @xcite . the following simulations are performed with the mean number of particles per collision cell @xmath121 , the rotation angle @xmath122 , and the time step @xmath123 , which yields a fluid viscosity of @xmath124 .
a spheroidal squirmer is a homogeneous rigid body characterized by its mass @xmath125 , center - of - mass position @xmath126 , orientation @xmath127 , translational velocity @xmath128 , and angular momentum @xmath129 .
thereby , @xmath130 is a rotation quaternion and can be related to the rotation matrix @xmath131 , which transforms vectors from the laboratory frame to the body - fixed frame @xcite .
we distinguish vectors in the laboratory frame and body - fixed frame by a superscript , i.e. , @xmath132 is a vector in the laboratory ( or space - fixed ) frame while @xmath133 is the corresponding vector in the body - fixed frame . for vectors in the laboratory frame
, we will frequently omit the superscript .
the orientation vector of a spheroid is @xmath134 .
the moment of inertia tensor in the body - fixed frame @xmath135 is a constant diagonal matrix with diagonal elements @xmath136 and @xmath137 . when needed , the angular velocity is calculated as @xmath138 . + for all simulations
we choose a neutrally bouyant spheroid , i.e. , @xmath139 , where @xmath140 is the fluid mass density . during the streaming step ,
a spheroid will collide with several mpc particles . since the total change in ( angular ) momentum of a spheroid during one streaming step is small
, we perform the collisions with mpc particles in a coarse - grained way @xcite : for the streaming step at time @xmath141 , we determine the spheroid s position , velocity , orientation , and angular velocity at times @xmath142 and @xmath143 , under the assumption that there is no interaction with mpc particles .
however , steric interactions between spheroids , as well as spheroids and walls are taken into account as described in sec .
[ sec : rbd_for_spheroids ] .
subsequently , all mpc particles are streamed , i.e. , their positions are updated according to @xmath144 .
thereby , a certain fraction of mpc particles penetrates a spheroid . to detect those particles in an efficient way
, possible collision cells intersected by the spheroid are identified first . for this purpose ,
we select all those cells , which are within a sphere of radius @xmath6 enclosing the spheroid instead of the spheroid itself , which is more efficient , since it avoids rotating candidate cells into the body - fixed frame during selection .
a loop over all particles in respective collision cells identifies those particles , which are inside the spheroid and they are labeled with the spheroid index .
then , each particle @xmath145 inside a spheroid at time @xmath143 is moved back in time by half a time step and subsequently translated onto the spheroid s surface .
the translation can be realized in different ways .
one possibility is to constructing a virtual spheroid with semi - axes @xmath146 and @xmath147 on its surface .
the particle is then translated along the normal vector of the virtual spheroid until it is on the real spheroid s surface .
alternatively , the difference vector @xmath148 can be scaled such that the particle position lies on the spheroidsurface .
we tried both approaches and found no significant difference . once the mpc particle at time @xmath142 is located on the spheroid s surface , the momentum transfer @xmath149 \right\}\end{aligned}\ ] ] at time @xmath142 is determined , taking into account the squirmer surface fluid velocity @xmath3 of eq .
( [ eq : def_usq_of_zeta ] ) @xcite .
thereby , a useful identity to determine @xmath2 is given in eq .
( 8) of ref . , and @xmath23 is given by @xmath150 the velocity of the mpc particle is updated according to @xmath151 .
subsequently , the position @xmath152 is obtained by streaming the mpc particle for the remaining time @xmath153 with velocity @xmath154 , i.e. , @xmath155 . as a consequence of the elastic collisions , the center - of - mass velocity and rotation frequency of a spheroid are finally given by @xmath156 where @xmath157 is total momentum transfer by the mpc fluid and @xmath158 is the respective angular momentum transfer . in a first step
, ghost particles are distributed inside each spheroid @xcite .
the number density and mass are equal for ghost and fluid particles .
the ghost particle positions @xmath159 are uniformly distributed in the spheroid and their velocities are given by @xmath160 the cartesian components of @xmath161 are gaussian - distributed random numbers with zero mean and variance @xmath162 .
the squirming velocity @xmath163 is determined by eq .
( [ eq : def_usq_of_zeta ] ) , with the ghost particle position projecting onto the spheroid s surface ( cf . sec .
[ sec : streaming ] ) . as a result of mpc collisions , a spheroid s linear and angular momenta change by @xmath164 and @xmath165 , where @xmath166 and @xmath167 are the ghost particle s velocity after and before the mpc collision .
hence , the spheroid velocity and angular velocity become @xmath168 during the streaming step , the spheroids move according to rigid - body dynamics , governed by @xcite @xmath169 , \\
\dot{\bm{q}}&=\frac{1}{2 } \bm{q}(\bm{q } ) \left ( \begin{matrix } 0 \\ \bm{\omega}^b \end{matrix } \right ) , \label{eq : qdot_equals } \\ \frac{d \omega_\alpha^b}{dt}&=i_\alpha^{-1 } \left [ t_\alpha^b+(i_\beta - i_\gamma ) \omega_\beta^b
\omega_\gamma^b \right ] .
\label{eq : eulerrotation}\end{aligned}\ ] ] here , @xmath170 is defined in eq .
( [ eq : def_q_of_q ] ) , and @xmath171 and @xmath172 are the force and torque acting on the spheroid . forces and torques are derived from steric interaction potentials as presented in appendix [ app : steric_interactions ] .
equations ( [ eq : eulerrotation ] ) are euler s equations for rigid body dynamics and hold for @xmath173 , and @xmath174 .
whenever necessary , body - fixed and laboratory - frame quantities can be related by the rotation matrix @xmath131 which is given in terms of the quaternion @xmath127 in eq .
( [ eq : rotation_matrix_for_rbd ] ) . for the numerical integration of the equations of motion ,
the widely applied leap - frog method @xcite is not useful , since velocity , angular momentum , position , and orientation are required at the same point in time for the coupling to the mpc method . hence , we employ the verlet algorithm for rigid - body rotational motion proposed in ref . .
integration for a time step @xmath24 is performed as follows : * update @xmath126 and @xmath127 according to ( cf .
( [ eq : qdot_equals ] ) and ( [ eq : eulerrotation ] ) ) @xmath175 the parameter @xmath176 is introduced to guarantee @xmath177 . *
calculate forces and torques @xmath178 and @xmath179 .
* update @xmath128 and @xmath180 according to @xmath181,\\ \bm{l}^s(t+\tau ) & = \bm{l}^s(t)+\frac{\tau}{2}[\bm{t}^s(t)+\bm{t}^s(t+\tau)].\end{aligned}\ ] ]
orientation correlation functions @xmath182 for passive spheroids with @xmath183 ( bottom blue line ) and @xmath184 ( top black line ) .
the plot shows the simulation data ( blue and black solid lines ) , an exponential fit to that data ( red dashed ) , and the theoretical prediction according to eq .
( [ eq : exp_decay_of_orientational_correlation_passive ] ) ( green dotted ) . ] for the passive spheroidal colloid ( @xmath185 ) , we perform equilibrium simulations and determine @xmath186 as well as @xmath187 for @xmath188 . due to the equipartition of energy
, we expect @xmath189 we fix the aspect ratio @xmath190 and vary @xmath7 in the range @xmath191 $ ] .
the simulation results agree very well with the theoretical values ( [ eq : equi_u ] ) and ( [ eq : equi_omega ] ) .
as expected , the deviations from theory decrease with increasing spheroid size , due to a better resolution in terms of collision cells . in general , the relative error @xmath192 is larger for @xmath193 than for @xmath194 .
we find the largest relative error for @xmath195 , namely @xmath196 and @xmath197 for @xmath198 and @xmath199 .
hence , we choose the minor axis @xmath200 in the following .
in addition , we determine the orientation correlation function @xmath182 .
the theory of rotational brownian motion @xcite predicts @xmath201 where @xmath202 , @xmath203 , @xmath204 , and @xmath205 and @xmath206 are the parallel and perpendicular rotational friction coefficients of a prolate spheroid with respect to the major semi - axis ; explicitly @xcite @xmath207 simulation results for the orientational auto - correlation function are shown in fig .
[ fig : orientation_correlation ] for two spheroids of different eccentricity .
the correlation functions decay exponentially .
however , for the spheroid with the smaller eccentricity , we find a somewhat faster decay than predicted by theory , whereas good agreement is found for the larger spheroid .
we attribute the difference to finite - size effects related to the discreteness of the collision lattice . for larger objects ,
discretization effects become smaller .
mean swimming velocity as function of the eccentricity @xmath208 for a spheroidal squirmer with @xmath209 and @xmath59 .
the solid line shows the theoretical prediction of eq .
( [ eq : swimming_velocity_spheroid_squirmer ] ) .
black dots are simulation results .
the eccentricity was varied by changing @xmath6 and keeping @xmath210 constant .
for the red triangle , we simulated a larger spheroid with @xmath211 , which shows a better agreement with theory . ]
fluid flow fields of a spheroidal squirmer in the laboratory frame with @xmath210 , @xmath212 , @xmath213 , and @xmath0 ( ( a),(b),(c ) ) , and with @xmath214 ( ( d),(e),(f ) ) .
the logarithm of the magnitude of the velocity field ( in units of @xmath215 ) is color coded .
the plots ( a ) , ( d ) show theoretical results , ( b ) , ( e ) simulation results , and ( c ) , ( f ) relative errors .
the relative error of the flow field is defined as @xmath216 $ ] .
note , due to the discrete representation of the velocity field , some streamlines end abruptly . ]
we determine the steady state swimming velocity of a squirmer via @xmath217 , which should be equal to @xmath218 ( cf .
( [ eq : swimming_velocity_spheroid_squirmer ] ) ) .
results for various eccentricities are displayed in fig .
[ fig : u_of_tau0 ] .
the velocity @xmath218 increases with increasing eccentricity @xmath208 in close agreement with the theoretical prediction of eq .
( [ eq : swimming_velocity_spheroid_squirmer ] ) .
we confirm that the force - dipole parameter @xmath219 does not affect the velocity of the squirmer , as long as the reynolds number @xmath220 is low , i.e. , @xmath221 .
we also determine the orientational correlation function and find that a squirmer exhibits the same orientational decorrelation as the corresponding passive particle ( cf .
[ fig : orientation_correlation ] ) .
moreover , we calculate the flow field from the simulation data and compare it with the theoretical prediction . as shown in fig .
[ fig : squirmer_sim_theo ] , the two fields are in close agreement .
the two - dimensional flow field of the mpc fluid , averaged over the rotation angle @xmath19 , is determined at the vertices of a fine resolution mesh .
the velocities at these vertices include averages over time of an individual realization as well as ensemble averages over various realizations . by the latter , we determine an estimate for the error of the mean velocity .
the median ( over vertices ) of this error is approximately @xmath222 for the parameters of fig .
[ fig : squirmer_sim_theo ] ( b ) and @xmath223 for that of fig .
[ fig : squirmer_sim_theo ] ( e ) .
note that we choose a smaller swimming mode @xmath53 for the puller ( fig .
[ fig : squirmer_sim_theo ] ( b ) ) than for the neutral squirmer ( fig .
[ fig : squirmer_sim_theo ] ( e ) ) .
the reason is that the agreement with theory was not satisfactory for the puller with @xmath224 , which we attribute to nonlinear convective effects . in figs .
[ fig : squirmer_sim_theo ] ( c ) and ( f ) , we observe lines of high relative errors ( yellow in the color code ) .
they appear because theory predicts @xmath225 or @xmath226 for these lines , which is difficult to achieve in simulations .
hence , the overall agreement between simulations and theory is very satisfactory , and the implementation is very valuable for the simulation of squirmer - squirmer and squirmer - wall interactions , where the details of the flow field matter .
average surface - to - surface distance @xmath227 and orientation of squirmers , where @xmath228 , as function of time .
the solid blue , dashed black , and dotted red lines correspond to pullers @xmath229 , neutrals @xmath230 and pushers @xmath231 .
the standard deviation of the blue line ( @xmath229 ) is indicated by the cyan shaded region . ]
we simulate the cooperative swimming behavior of two squirmers in a slit geometry .
the slit is formed by two parallel no - slip walls located at @xmath232 and @xmath233 . the no - slip boundary condition is implemented by applying the bounce - back rule and ghost particles of zero mean velocity in the walls @xcite .
steric interactions between two squirmers and between a squirmer and a wall are taken into account by the procedure described in appendix [ app : steric_interactions ] .
the initial positions and orientations of the two squirmers ( @xmath234 ) are @xmath235 here , @xmath236 is the initial center - of - mass distance and @xmath237 , where @xmath238 is the inital angle between @xmath239 and @xmath240 . the swimming mode is chosen as @xmath209 and the force dipole mode @xmath241 .
we choose @xmath236 such that the squirmers are well separated and vary @xmath238 .
the squirmers major and minor axes are @xmath210 and @xmath212 , respectively , and the simulation box size is @xmath242 , and @xmath243 .
note that @xmath244 which keeps the swimming orientation essentially in the @xmath245-@xmath1 plane .
results for the mean surface - to - surface distance between squirmers @xmath246 and the mean alignment @xmath247 are shown in fig .
[ fig : d_cos ] for pushers , pullers , and neutral swimmers with an inital angle @xmath248 . due to the setup
, the squirmers initially approach each other and collide at @xmath249 .
the ( persistence ) pclet number @xmath250 is sufficiently high , such that the squirmer orientation has hardly changed before collision .
when the neutral swimmers collide , they initially align parallel ( @xmath251 at @xmath252 in fig .
[ fig : d_cos ] ) , but their trajectories start to diverge immediately thereafter .
pushers remain parallel for an extended time window , which is expected as pushers are known to attract each other @xcite , but at @xmath253 ( cf .
[ fig : d_cos ] ) their trajectories diverge as well .
this is probably due to noise , since we observe several realizations where pushers remain parallel .
interestingly , pullers , which are known to repel each other when swimming in parallel@xcite , swim cooperatively and reach a stable orientation with @xmath254 shortly after they collided ( at @xmath252 ) .
thereby , their cooperative swimming velocity is about @xmath255 .
the flow field of this stable state , determined by mpc simulations , is shown in fig .
[ fig : coop_flow_field ] .
note that the velocity field in the swimming plane is left - right symmetric , and that there is a stagnation point in the center behind the swimmers .
figure [ fig : coop_flow_field ] reveals that this point actually corresponds to a line normal to the walls .
time dependence of the average alignment @xmath256 of two pullers with @xmath229 , @xmath190 , and various initial angles @xmath257 .
] figure [ fig : initial_conditions_to_fixed_point ] shows that the fixed point of cooperatively swimming pullers is reached for nearly all simulated initial conditions @xmath258 . only pullers that are nearly parallel initially ( @xmath259 , @xmath260 in fig .
[ fig : initial_conditions_to_fixed_point ] ) , repel each other such that they will not reach the fixed point . for pclet numbers @xmath261 ,
the fixed point remains at @xmath254 .
however , it becomes more likely for the swimmers to escape ( or never reach ) the fixed point .
a detailed study reveals that the fixed point vanishes , when the walls are replaced by periodic boundary conditions .
this is even true when we apply three - dimensional periodic boundaries , but keep the wall potential implemented , i.e. , the squirmers are still confined in a narrow slit . in addition
, we studied the swim behavior of spherical squirmers . here
, we observe diverging trajectories for all squirmer types , i.e. , pushers , neutral squirmers , and pullers .
such diverging trajectories have already been reported in ref . for spherical squirmers in bulk .
hence , the stable close - by cooperative swimming of pullers is governed by the squirmer anisotropy , by the hydrodynamic interactions between them and , importantly , between pullers and confining surfaces .
this conclusion is in contrast to results presented in ref . , where a monolayer of spheroidal squirmers is considered , with their centers and orientation vectors fixed in the same plane , however , without confining walls .
the study reports a stable cooperative motion for pullers with angles @xmath262 by nearest - neighbor two - body interactions , where all angles between @xmath263 and @xmath264 are stable .
the difference to our study is that in ref .
cooperative features were extracted from a simulation of many swimmers , whereas we explicitly studied two swimmers .
furthermore our study explicitly models no - slip walls and includes thermal noise . to shed light on the stability of the cooperative puller motion , we varied the puller strength @xmath219 , the aspect ratio @xmath265 , and the width of the slit @xmath266 .
thereby , we started from our basic parameter set @xmath267 and @xmath229 . with decreasing @xmath219 ,
the stable alignment disappears , i.e. , the pullers distance increases after collision . for increasing @xmath219
the fixed point remains , but the value of @xmath268 decreases , i.e. , the squirmers form a larger angle . with increasing wall separation , the fixed - point value of @xmath268 decreases , i.e , the angle between the swimmers increases . for @xmath269 and higher
, the fixed point disappears .
an increase of the aspect ratio @xmath265 from @xmath270 to @xmath271 and @xmath272 increases the fixed - point value of @xmath273 from @xmath274 to @xmath275 and @xmath276 .
the more elongated shape leads to a more parallel alignment of the squirmers .
the minimal value of @xmath219 required to achieve cooperative motion depends weakly on the aspect ratio . for @xmath277 and @xmath272 the critical values for @xmath219
are @xmath278 and @xmath279 .
hence , a large aspect ratio is beneficial for cooperative swimming .
we have introduced a spheroidal squirmer model , which comprises the swimming and force - dipole modes .
it is a variation of previously proposed squirmer models . on the one hand , it includes the force - dipole mode as an extension to the model of ref . .
on the other hand , it is an alternative approach compared to refs . and , with the major advantage that our model allows for the analytical calculation of the flow field .
in the present calculations we employed the stokes stream function equation .
very recently a full set of solutions to stokes equations in spheroidal coordinates were given in ref .
, which opens an alternative approach to derive the flow field for our choice of boundary conditions .
furthermore , we have presented an implementation of our spheroidal squirmer in a mpc fluid .
in contrast to other frequently employed simulation approaches , mpc includes thermal fluctuations .
the comparison between the fluid flow profile of a squirmer extracted from the simulation data with the theoretical prediction yields very good agreement . as a consequence of the mpc approach with its discrete collision cells ,
the minor axis of the spheroid has to be larger than a few collision cells to avoid discretization effects .
the analysis of the squirmer orientation correlation function shows that very good agreement between theory and simulations is already obtained for @xmath280 ( major axis ) and @xmath210 ( minor axis ) . to shed light on the cooperative swimming motion and on near - field hydrodynamic interactions , we investigated the collision of two spheroidal squirmers in a slit geometry .
we found stable stationary states of close - by swimming for spheroidal pullers , which is determined by hydrodynamic interactions between the anisotropic squirmers , and , even more important , by squirmers and surfaces .
this stationary state disappears for low puller strengths and low eccentricities .
we expect the stable close - by swimming of pullers to strongly enhance clustering in puller suspensions in thin films .
our studies confirm that spheroidal squirmers can accurately be simulated by the mpc method .
the proposed implementation opens an avenue to study collective and non - equilibrium effects in systems of anisotropic microswimmers .
even large - scale systems can be addressed by the implementation of mpc and the squirmer dynamics on gpus .
we thank a. wysocki , a. varghese , and b.u .
felderhof for helpful discussions . support of this work by the dfg priority program spp 1726 on `` microswimmers from single particle motion to collective behaviour '' is gratefully acknowledged .
the rotation matrix @xmath131 introduced in eq .
( [ eq : vb_equals_dvs ] ) is given in terms of the rotation quaternion @xmath127 as @xmath281 the matrix @xmath170 in eq .
( [ eq : qdot_equals ] ) is given by @xmath282 q_0 & -q_1 & -q_2 & -q_3 \\
q_1 & q_0 & -q_3 & q_2 \\ q_2 & q_3 & q_0 & -q_1
\\ q_3 & -q_2 & q_1 & q_0 \end{pmatrix*}.\end{aligned}\ ] ]
for @xmath283 and @xmath284 the gegenbauer functions of the first and second kind @xmath285 and @xmath286 , are defined in terms of the legendre functions of the first and second kind @xmath287 and @xmath288 as @xcite @xmath289 for @xmath290 , they are defined as @xmath291 for the reader s convenience , we give the formula for the gegenbauer functions of the first kind for @xmath292 , and @xmath284 @xmath293 furthermore , the gegenbauer functions of the second kind for @xmath292 , and @xmath294 are given by @xmath295 here , we used @xmath296/[x-1])/2 $ ] .
here , we illustrate our implementation of the excluded - volume interactions between spheroids and walls following the approach provided in ref . . the spheroid s surface in the laboratory frame is given by the quadratic form @xmath297 where the orientation matrix @xmath298 can be expressed as @xmath299 for the steric interactions , we introduce a virtual safety distance @xmath300 , which is small compared to @xmath7 and @xmath6
when computing steric interactions , we replace @xmath7 and @xmath6 by @xmath301 and @xmath302 , respectively . in this paper
we used @xmath303 for all simulations . we introduce a repulsive interaction potential between spheroids to prevent their overlap .
the potential is given by @xmath304.\end{aligned}\ ] ] here , @xmath305 and @xmath306 correspond to a length and energy scale , respectively .
we choose @xmath307 and @xmath308 .
the directional contact distance @xmath309 between two spheroids , with orientation matrices @xmath310 , @xmath311 and center positions @xmath312 , @xmath313 , is an approximation to their true distance of closest approach and is defined by @xmath314 here , @xmath315 , @xmath316 , and @xmath317 is the elliptic contact function , defined as @xcite @xmath318 minimization with respect to @xmath319 demands @xmath320 , and hence , @xmath321 the critical value @xmath322 that maximizes @xmath323 can be found by the root finding problem @xmath324 we implement brent s root finding approach @xcite .
the forces and torques arising from the potential ( [ eq : potential_of_dr ] ) can be calculated analytically and are given by @xcite @xmath325 \\ & \times \left ( \frac{\bm{r}}{r}(f^{-1/2}-1 ) -\frac{r}{2}f^{-3/2 } \bm{x}_c\right),\end{aligned}\ ] ] and @xmath326 \\ & \times f^{-3/2 } ( \bm{x}_c-\bm{c})\times \bm{x}_c\end{aligned}\ ] ] for the first spheroid , where @xmath327 .
the force and torque on the second spheroid follow by newton s action - reaction law , namely @xmath328 we restrict ourselves to short - rang repulsive interactions by setting the potential @xmath329 to a constant value for @xmath330{2}-1)\sigma_0 $ ] , which implies that @xmath331 and @xmath332 are zero for this range of @xmath309 values .
note that an upper bound to @xmath309 is @xmath333 , which means that two spheroids will not interact if @xmath334{2}-1)\sigma_0 $ ] .
this inequality is checked before a numerical calculation of @xmath309 is employed .
we assume that two parallel walls are positioned at @xmath335 , which taking into account the safety distance @xmath300results in the effective wall positions @xmath336 and @xmath337 .
we propose an interaction between a spheroid and a wall in the style of the spheroid - spheroid interaction presented in ref . .
first , we find the point @xmath319 on the spheroid s surface that is closest to a wall .
for the wall at @xmath336 , this is achieved by minimizing the height @xmath338 under the constraint @xmath339 . using the method of lagrange multipliers
, we have to minimize @xmath340 . the necessary condition for a minimum @xmath341 yields @xmath342 and hence , @xmath343 substitution of eq .
( [ eq : x_equals_c - stuff_over_2_lambda ] ) into @xmath339 yields @xmath344 finally , we obtain the point closest to the wall as @xmath345 here , the minus sign has to be chosen , which can be visualized by the example of a sphere of radius @xmath98 , for which @xmath346 .
this finally yields the height @xmath347 we employ the lennard - jones potential @xmath348\end{aligned}\ ] ] for a repulsive wall , and @xmath349 assumes a constant value for all @xmath350{2}-1)\sigma_0 $ ] .
we can derive the force @xmath351 and torque @xmath352 acting on the spheroid analytically .
for the force , we find @xmath353 \bm{e}_y\end{aligned}\ ] ] and for the torque @xmath354 with @xmath355 here , we use the relation @xmath356 which holds for an invertible matrix @xmath357 depending on a scalar parameter @xmath141 , and eq .
( c9 ) from ref . . for the wall at @xmath358
, we have to minimize @xmath359 , with @xmath319 on the spheroid s surface .
this yields @xmath360^{-1/2 } .\end{aligned}\ ] ] the formulas for torque and force do not change , except that we have to insert @xmath361 and need to change the sign of the force . | we propose a hydrodynamic model for a spheroidal microswimmer with two tangential surface velocity modes .
this model is analytically solvable and reduces to lighthill s and blake s spherical squirmer model in the limit of equal major and minor semi - axes .
furthermore , we present an implementation of such a spheroidal squirmer by means of multiparticle collision dynamics simulations .
we investigate its properties as well as the scattering of two spheroidal squirmers in a slit geometry . thereby we find a stable fixed point , where two pullers swim cooperatively forming a wedge - like conformation with a small constant angle . |
in many realistic systems , dynamics undergo a pronounced slow - down , a feature characteristic of `` complex '' landscapes @xcite .
such phenomena arise for instance in physical systems ( thermal relaxation @xcite ) , in optimisation problems ( diminishing returns on search efforts @xcite ) , and in evolutionary dynamics ( punctuated equilibria , i.e. , long periods of stasis in the evolutionary record @xcite ) . in this work , inspired by the theoretical framework of statistical physics for glassy systems , we reconsider a simple model of evolutionary dynamics , namely rna secondary structure evolution @xcite .
we find : ( i ) slow evolutionary dynamics , whereby the time to find an advantageous phenotypic change has a distribution with a fat ( power - law ) tail ; ( ii ) non - self - averaging behavior , _
i.e. _ , even for long rna molecules , directional selection for some targets will lead to significantly slower dynamics than for other targets ; ( iii ) only weak out - of - equilibrium effects : the genotypes visited by an evolutionary trajectory are quite similar to those arising in equilibrium under stabilizing selection , so except near the ends of periods of stasis , the genotypes produced during an evolutionary trajectory do not have anomalously high or low mutational robustness .
the paper is organised as follows . in sect .
[ sect : rnaevolution ] we introduce the evolutionary model . in sect .
[ sect : slow ] we exhibit the non - exponential nature of the relaxation process ; empirically , relaxation seems to follow an inverse power law .
we also show that the relaxation curves remain sensitive to the target used for directional selection even in the limit of very long molecules . finally in sect .
[ sect : innovation ] we give evidence that at nearly all times the evolutionary trajectory is in quasi - equilibrium ; more precisely , the `` innovative '' genotypes produced by a transition to a new period of stasis are only a bit different from random genotypes as measured by the phenotypic effects of mutations ; furthermore , during the periods of stasis , the initial genotype seems to be quickly forgotten , so no significant trace of innovation seems to be maintained .
concluding remarks are given in the final section .
for the purpose of this study , rna molecules will be thought of as chains or strings of @xmath0 nucleotides , taken from an alphabet of four possible nucleotides ( a , c , g and u ) .
chemically , the bases can pair via hydrogen bonds .
in addition to watson - crick pairings ( a - u and g - c ) , the u - g pairing is also possible , though it is weaker .
the pairings between bases give rise to an rna secondary structure , as illustrated in fig.[fig : structure ] .
apart from this graphical representation , the secondary structure of an rna molecule can be specified by the more convenient dot - bracket notation where a non - paired base of the sequence is denoted by a dot `` . '' and a paired base is denoted by a left or right parenthesis .
this representation allows one to reconstruct the pairings as long as the secondary structure is `` planar '' @xcite .
( planarity means that if the @xmath0 bases are positioned on a line and the pairings are represented by arcs between the bases , these arcs can be drawn in the plane without any crossings . )
an rna molecule will spontaneously ( _ i.e. _ , via thermodynamic forces only ) fold into the structures with lowest free energies . to simulate this folding in silico
, we use the `` vienna rna package '' @xcite to find , for any given sequence , the pairings ( secondary structure ) which lead to the _ minimum _ free energy . in effect
, this procedure produces a map from sequences ( genotypes ) to secondary structures ( phenotypes ) @xcite .
the cpu time needed to determine a minimum free energy structure is @xmath1 for a chain of @xmath0 bases .
, width=226,height=302 ] all genotypes that have the same phenotype form a _
neutral network _ @xcite .
more precisely , given a genotype to phenotype map , the neutral network associated with the phenotype @xmath2 is a graph whose nodes are all the genotypes having that phenotype ; the edges of that graph connect genotypes only if they are nearest neighbors .
it is thus necessary to introduce a notion of neighborhood in genotype space . by convention , two rna genotypes ( sequences )
will be considered as nearest neighbors if and only if they differ by a single base .
it is also possible to define a distance @xmath3 between two _ phenotypes_. in our case , we take the distance between two rna secondary structures @xmath4 and @xmath5 to be @xmath6 times the hamming distance between their dot - bracket string representations . from this , one can introduce a fitness landscape @xcite , where genotype and phenotype spaces both have a distance metric , and where each phenotype can be assigned `` fitness '' .
we take the fitness of a phenotype to be a monotonically decreasing function of its distance to some `` target '' phenotype @xmath7 . for our purpose
, we will extend the notion of a neutral network to that of a _
neutral ensemble _ : we call neutral ensemble the union of all neutral networks whose corresponding phenotype has a given fitness , regardless of the specific phenotype .
this definition reflects that fact that secondary structures are relevant in the evolutionary search mainly via their distance to @xmath7 .
rna sequences with their associated fitness form a fitness landscape with hills , valleys and passes ( saddles ) .
analogous landscapes also arise in other systems ; for example spin glasses @xcite are physical materials in which energy is often identified with fitness ; the landscape s ruggedness and many valley structure are important for the associated dynamical properties .
we take the unit of time to be the expected waiting time @xmath8 between two mutations .
thus if @xmath9 is the mutation rate per rna molecule and per generation , our unit of time is @xmath10 . for our evolutionary dynamics
, we simulate the process of rna evolution towards a target structure by allowing a genotype to change by a point mutation at each unit of time . the directional selection associated with this evolutionary search then proceeds as follows . at each step , the genotype is mutated at one base taken at random ; the corresponding phenotype ( secondary structure for that genotype ) is determined , and selection is applied : if the distance to the target structure has increased , the mutation is refused and the previous genotype is reinstated , otherwise the new genotype is accepted .
this process is usually referred to as `` blind ant '' dynamics @xcite or as an adaptive walk @xcite .
( if one were to keep track only of the accepted moves that are neutral , _
i.e. _ , that do not change the distance to the target phenotype , one would have a random `` neutral '' walk performing myopic ant dynamics @xcite . ) in practice , we continue the random walk until the target structure is reached , or until a maximum number of trial mutations is reached . in our work , we allow only neutral or improving moves in the landscape , which corresponds to zero temperature dynamics in physical systems , or hill climbing in optimisation theory .
the trajectories are stochastic and are influenced both by the initial genotype chosen ( e.g. an arbitrary rna _ sequence _ ) , and by the target ( an arbitrary _ secondary structure _ ) .
we thus need to average over many trajectories , and consider the dependence of our results on these choices . in a more general context
, one can consider the evolution of a _ population _ in the fitness landscape .
if @xmath9 is the mutation rate and @xmath11 the population size , then the effective number of genotypes in a population scales approximately linearly with @xmath12 .
when @xmath13 , the population remains essentially monomorphic , and this is the case we focus on in this work .
if instead @xmath12 were large , the evolutionary dynamics would occur in a polymorphic population with many different genotypes . to test whether the out - of - equilibrium evolutionary dynamics generates atypical genotypes
, one must define the `` null hypothesis '' ; clearly we want to compare to _ random _ genotypes in the fitness landscape .
this means that we need to sample _ uniformly _ the fitness landscape for each given distance to the target phenotype ; this means we focus on the genotypes that have a given fitness .
the algorithmic procedure to do so is to start with any genotype in the fitness landscape with the specified fitness and then produce a long random walk with importance sampling using the metropolis monte carlo algorithm @xcite . in our rna neutral ensemble context , this simply corresponds to using the blind ant dynamics , accepting only the mutations that do not change the distance to the target ; this is then identical to the dynamics under stabilizing selection . from this sampling
, we shall obtain equilibrium averages in this space .
we will be particularly interested in the mean mutational robustness , where the mutational robustness of a genotype is defined as the fraction of the single - base mutations that do not change the fitness ; this is the same thing as the number of neighbors of this genotypye that belong to the neutral ensemble divided by its total number of neighbors .
consider a typical evolutionary trajectory starting from a random initial genotype . at the beginning
, there is a high frequency of advantageous mutations , so the phenotypic distance @xmath3 to the target structure initially decreases fast .
but at long times , as first realized in @xcite , the frequency of favorable mutations becomes small and long periods of `` stasis '' appear where the fitness remains constant @xcite .
this is illustrated for two typical evolutionary trajectories in fig .
[ fig : trajectory ] .
successive plateaus in fitness are separated by small changes in the hamming distance to the target : the steps decrease this distance by 1 or 2 units typically , rarely more than that .
also , the time of a stasis period typically increases as the distance to the target decreases : one can speak of diminishing returns for the effects of mutations in approaching the target . to claim that the convergence to the target is particularly slow ,
it is appropriate to have a comparison benchmark . for this , consider the situation where directional selection is for a target genotype instead of a target phenotype .
just as for phenotypic distances , we define the genotypic distance of two genotypes @xmath14 and @xmath15 as @xmath16 where @xmath17 is the hamming distance bewteen the two strings defining the sequences for @xmath14 and @xmath15 .
if @xmath18 is then the distance between the _ current _ genotype and the target genotype , a mutation has a probability @xmath19 to produce a strictly better genotype .
it follows that the convergence to the target is exponential in time : @xmath20 consider now the times @xmath21 when a favorable mutation arises ; a transition between one stasis period and the next generates such an event which diminishes @xmath18 by @xmath6 .
if @xmath22 is the density of these times , we have @xmath23 we then obtain @xmath24 .
this shows that selection for a target genotype leads to fast ( exponential ) relaxation .
the situation is completely different in our system where selection is instead for a target phenotype .
guided by the above analysis , we have determined the corresponding distribution @xmath22 in this case .
[ fig : fat_tails ] shows the result when averaging over target and initial genotypes at @xmath25 .
this distribution has a very clear fat tail which seems compatible with a power law .
this same behavior is observed for the other values of @xmath0 tested ( data not shown ) .
one can thus say that convergence to the target under phenotypic selection is `` slow '' , much slower than when compared to genotypic selection .
this difference between the selection types can also be seen by looking at the average distance to the target as a function of evolutionary time . in landscapes associated with disordered systems such as glasses or spin glasses ,
the relaxation processes encountered typically have non - exponential dynamics @xcite .
empirically , two families of functions have been used to perform fits : the stretched exponential family which for our purposes corresponds to @xmath26 $ ] , and the shifted power law family for which @xmath27 for small @xmath0 , both types of fits lead to satisfactory results , but at larger @xmath0 , the data favor a shifted power law behavior . in fig .
[ fig : loglogaveragetrajectory ] , we show that @xmath28 decreases rather slowly , roughly as an inverse power of time .
this leads to a nearly straight line on a log - log plot ; also shown in that figure are the fits to shifted power laws .
this overall behavior should be contrasted with the law ( cf .
[ eq : dg_exp ] ) found when using selection for a target genotype : there the approach to the target was much faster .
, width=415,height=302 ] to further analyse the slow approach towards a target structure , we investigated how the evolutionary dynamics changes with the target . first , we ran simulations for genotypes of length 100 for different targets . in these simulations , for each target , we averaged the relaxation curves for different randomly generated initial genotypes .
[ fig : diff60tar ] illustrates how the relaxation is different for different target structures .
we conclude that there is slow dynamics whose speed depends on the target structure , a conclusion that holds for all the values of the chain length @xmath0 we have investigated .
although we can not deal with arbitrary @xmath0 because of computational limitations , it is nevertheless relevant to ask whether this dependence on targets survives for arbitrarily large @xmath0 .
it seems possible that the relaxation behavior is not self - averaging , _
i.e. _ , that fluctuations associated with different targets do not become negligible when @xmath0 grows . to test this , we carried out simulations with multiple targets for different lengths of rna chains , to determine whether the relaxation curves have smaller dispersion for the different targets when the chain length @xmath0 increases .
we carried out simulations for thirty different targets , and averaged the relaxation curve for each target over 1000 evolutionary trajectories with random initial genotypes .
we then measured the standard deviation @xmath29 of the relative distance to the target phenotype at times where the _ mean _ distance ( averaged over all curves ) to the target was @xmath30 percent of the chain length .
data are summarised in the following table . [ cols="<,<,<",options="header " , ] here @xmath31 denotes the mean time at which the average distance to the target phenotype reaches the value @xmath32 , @xmath0 being the chain length .
the simulations were quite time - consuming which prevented us from testing lengths larger than 120 .
however , from the table we see that the dispersion @xmath29 decreases initially , but then remains practically unchanged .
the relaxation process towards different targets therefore is compatible with a non self - averaging behavior .
[ fig : rmu_two ] the transition from one period of stasis to the next is due to an evolutionary `` innovation '' .
it is appropriate to ask whether the associated transition genotypes can be considered to be atypical according to some measurable quantity .
here we shall study the _ short term evolvability _ of the entry genotype in a period of stasis and compare it to that of _ random _ genotypes at the same distance to the target . for a given genotype
we define its short - term evolvability @xcite as the fraction of single base mutations that lead to a phenotype with higher fitness .
we found that the mean fraction of such beneficial mutations is approximately 40% _ lower _ for the entry genotype than for genotypes randomly sampled from the neutral ensemble .
thus with our definition , the entry genotype has an atypically low evolvability .
since it is close to a neutral ensemble of less fit phenotype , this result is expected at a qualitative level .
however , we also find that the entry genotype has a smaller fraction of deleterious mutations than the neutral ensemble average .
( these mutations produce a phenotype with increased distance to the target . )
this result should be contrasted with what is naively expected .
indeed , by construction the entry genotype has one particular mutation which is known to be deleterious ( taking it back to the previous plateau ) . neglecting all other effects
, one would predict that on average the entry genotype would have its fraction of deleterious mutations be @xmath33 above that of genotypes in the neutral ensemble .
instead , the effect is four times larger and in the opposite direction . during evolution on a plateau ,
one goes from the entry genotype ( which we just saw is atypical according to some objective measure ) to more random genotypes : to some extent , one looses the memory of the entry genotype through successive mutations .
is it a slow change of genotypes that is responsible for the stasis periods ? to address this question , we monitored the distance between a mutating genotype and the entry genotype of the corresponding period of stasis . because @xmath0 is moderately large in our simulations ,
the initial growth in distance is linear in the number of accepted mutations . at larger times , we see a saturation effect caused by multiple substitutions , as displayed in fig . [
fig : meangeno ] .
overall , the evolutionary dynamics on the neutral ensemble shows that even in the absence of phenotypic change , genotypes diffuse on a neutral ensemble at a rate not much slower than if their diffusion was not constrained by this set , until they discover a new phenotype closer to the target . to understand the expected dynamics in the absence of selection consider that a given site mutates with probability @xmath6 at each step , so the relaxation time scales as @xmath0 . at long times
one approaches the average distance @xmath34 .
this is the behavior , translated mathematically in the jukes - cantor formula @xcite , that the uppermost curve in fig.[fig : meangeno ] represents . in summary ,
because of the intricate relation between genotype and phenotype , there can be slow dynamics in the approach to the target phenotype despite rapid evolutionary change of genotypes .
in fact , as displayed in fig . [
fig : meangeno ] , the rate of change of genotypes does not seem to be significantly different when comparing easy and difficult targets for the directional selection .
mappings from genotypes to phenotypes play a central role in biology , from the molecular scale up to whole organisms . working at the level of rna allowed us to use a framework for such mappings that is not biologically arbitrary , even though it is clearly idealised .
one of this mapping s main advantages is that it is computationally tractable . within this mapping
, we showed a rich phenomenology of the evolutionary dynamics towards an optimum phenotype : ( 1 ) the `` relaxation '' towards the target undergoes severe slowing down as the target is approached ; ( 2 ) this slowing down gives rise to stasis periods with fat tails , typical of what is expected in complex fitness landscapes ; ( 3 ) the relaxation curves remain sensitive to the choice of the target , even in the limit of long rna sequences ; ( 4 ) the diffusion in genotype space during the periods of stasis is not slow , in obvious contrast with what happens at the level of phenotypes .
we observed that the probability of generating a favorable mutation decreases severely as one approaches the optimum , a property of the fitness landscape itself , _
i.e. _ , the fraction beneficial mutations goes down in this limit , much faster than in systems undergoing exponential relaxation . why this leads to inverse power laws remains open , just as in many other landscape problems coming from other fields .
furthermore , because the stasis periods are long , the few innovative genotypes appearing in an evolutionary trajectory represent a tiny fraction of the whole , and thus most genotypes visited have little trace of the out - of - equilibrium dynamics .
99 s. gavrilets , _ fitness landscapes and the origin of species _ , princeton university press , 2004 a. p.
young , _ spin glasses and random fields _ , world scientific publishing , 1998 h. h. hoos , t. sttzle , _ stochastic local search : foundations and applications _ , elsevier , 2005 n. eldredge , j. n. thompson , p. m. brakefield , s. gavrilets , d. jablonski , j. b. c. jackson , r. e. lenski , b. s. lieberman , m. a. mcpeek , w .
miller iii , _ the dynamics of evolutionary stasis _ , paleobiology v. 31 ; no .
2_suppl , june 2005 m. huynen , p. stadler w. fontana : _ smoothness within ruggendess : the role of neutrality _ , proc .
usa , vol .
93 , 397 - 401 ( 1996 ) .
w. fontana , p. schuster : _ slow evolutionary dynamics of rna structures and evolvability _ , science 29 , vol .
5368 , pp .
1451 - 1455 ( 1998 ) w. fontana , p. schuster , _ shaping space : the possible and the attainable in rna genotype - phenotype mapping _
, j. theor .
biology , vol .
194 , issue 4 , 491 - 515 ( 1998 ) gesteland , r.f .
, atkins , j.f .
( ed . ) : _ the rna world _ , cold spring harbor laboratory press , 1993 http://www.tbi.univie.ac.at/~ivo/rna/ [ http://www.tbi.univie.ac.at/~ivo/rna/ ] w. fontana p. f. stadler , e. g. bornberg - bauer , t. griesmacher , i. l. hofacker , m. tacker , p. tarazona , e. .d weinberger , p .
schuster , _ rna folding and combinatory landscapes _ , phys .
e 47 , 1993 a. wagner , _ robustness and evolvability : a paradox resolved _ , proc .
, vol 275 , 91 - 100 , 2008 p. f. stadler , _ landscapes and their correlation functions _ j.math.chem .
20 : 1 - 45 , 1996 c. monthus , j .- p .
bouchaud , _ models of traps and glass phenomenology _ , j. phys .
29 , 3847 - 3869 , 1996 j .- p .
bouchaud , _ weak ergodicity breaking and aging in disordered systems _ , j. phys .
i ( france ) vol . 2 , 1705 - 1713 , 1992 b. d. hughes , _ random walks and random environments _ , oxford : clarendon , 1996 s. kauffman , s. levin , _ toward a general theory of adaptive walks on rugged landscapes _
, j. theor . biol .
1987 , 128 , 11 - 45 e. van nimwegen , j. p. crutchfield , m. huynen , _ neutral evolution of mutational robustness _ ,
usa , vol .
96 , 9716 - 9720 , 1999 k. binder , d. w. heermann,_monte carlo simulation in statistical physics : an introduction _ , springer , 2002 m. zuker , p. stiegler , _ optimal computer folding of large rna sequences using thermodynamics and auxiliary information _ ,
nucleic acids res . , jan 10;9(1):133 - 48 , 1981 m. kimura , t. ohta , _ on the stochastic model for estimation of mutational distance between homologous proteins _
, j. mol .
evol . , 87 - 90 , 1972 m. a. huynen , p. stadler , w. fontana , _ smoothness within ruggedness : the role of neutrality in adaptation _ , proc .
usa , vol . 93 , 1996 m. kirschner , j. gerhart , proc . natl
. acad .
evolvability _ , vol .
95 , 8420 - 8427 , 1998 l. a. meyers , f. d. ancel , m. lachmann , _ evolution of genetic potential _ , plos computational biology , vol . 1 , 2005 m. c. cowperthwaite , l. a. meyers , _ how mutational networks shape evolution : lessons from rna models _ , annu . | we re - examine the evolutionary dynamics of rna secondary structures under directional selection towards an optimum rna structure .
we find that the punctuated equilibria lead to a very slow approach to the optimum , following on average an inverse power of the evolutionary time .
in addition , our study of the trajectories shows that the out - of - equilibrium effects due to the evolutionary process are very weak . in particular , the distribution of genotypes is close to that arising during equilibrium stabilizing selection . as a consequence ,
the evolutionary dynamics leave almost no measurable out - of - equilibrium trace , only the transition genotypes ( close to the border between different periods of stasis ) have atypical mutational properties . |
the advent of large radial velocity surveys and the @xmath9 mission @xcite have revealed planets with masses several times that of earth in orbits that may permit the existence of liquid surface water @xcite .
perhaps the most favorable objects for follow - up studies in the next decade are planets in the habitable zone ( hz ) around low - mass stars , dwarfs of spectral type m0v and later ( also referred to as m - dwarfs or dm stars ) .
these planets have received considerable theoretical attention , and a consensus is emerging that neither synchronous rotation nor close proximity to the host star rules out the possibility of biological development on these worlds @xcite .
the heating and chemistry of an earth - like atmosphere depends strongly on the characteristics of the host star , both spectrally and temporally .
the photospheric ultraviolet ( uv ) continuum of m - stars is very small relative to solar - type stars due to their lower effective temperature .
bright chromospheric and transition region emission lines dominate the uv spectrum of m - dwarfs , containing a much larger fraction of the stellar luminosity than for solar - type stars .
these differences influence the atmospheric temperature profile and the production of potential biomarkers ( e.g. ozone , o@xmath10 ; segura et al .
2003 ; segura et al . 2005 ) .
h@xmath5o , ch@xmath11 , and co@xmath5 are highly sensitive to far - uv radiation , with photo - absorption cross - sections that overlap with ly@xmath0 , the strongest emission feature in the uv spectrum of an m - star .
m - stars are known to display frequent and intense flare activity @xcite , and the combination of uv photon and proton bombardment from a large flare may produce @xmath12 90% o@xmath10 depletions in an earth - like atmosphere @xcite . despite the importance of understanding these host m - stars , their actual spectral and temporal behavior is not well known outside of a few active flare stars @xcite . gj 876 is an m4v star ( @xmath13 = 0.33 @xmath14 ) hosting a multi - planet system @xcite .
the four planets have masses ranging from super - jovian to super - earth ( @xmath15 = 2.64 @xmath16 , @xmath17 = 0.21 au ; @xmath18 = 0.83 0.86 @xmath16 , @xmath19 = 0.13 au ; @xmath20 @xmath2 6 @xmath21 , @xmath22 = 0.02 au ; @xmath23 = 12 @xmath21 , @xmath24 = 0.33 au ) , with gj 876b orbiting near the center of the hz .
conditions favorable for the sustained presence of liquid surface water might therefore exist on any satellites of gj 876b .
there is no present evidence for a bright remnant debris disk around gj 876 @xcite , and taken together with the slow stellar rotation ( 40 @xmath25 @xmath26 @xmath27 97 days ; rivera et al .
2005 , 2010 ) , gj 876 is estimated to have an age between 0.1
5 gyr @xcite . on the basis of its h@xmath0 absorption spectrum , gj 876
is considered to be an inactive m - dwarf . however , gj 876 has a measureable level of coronal emission in soft x - rays ( @xmath28 @xmath3 3 @xmath4 10@xmath29 erg s@xmath30 in the 0.2 2.0 kev band ; poppenhaeger et al .
2010 ) .
@xcite have further challenged the assumed link between optical activity indicators and uv flux , using low - resolution @xmath31-acs prism observations to demonstrate that gj 876 has levels of chromospheric near - uv ( nuv ) flux comparable to the well - studied active m - dwarf ad leo .
other than the low - resolution spectra presented by @xcite , there are very few nuv observations of gj 876 , and none that clearly detect emission from the star at far - uv wavelengths ( fuv ; 1150 @xmath32 @xmath33 @xmath32 2000 ) . as current m - dwarf models do not properly account for chromospheric and transition region emission , the lack of observational basis means that a crucial input to atmospheric models of planets orbiting low - mass stars is missing .
the situation is sufficiently unclear that atmospheric modelers have taken widely varying approaches to estimating the uv radiation field incident upon the hz of m - dwarfs , from assuming the uv spectrum of a young flare star ( ad leo ; wordsworth et al .
2010 ) to assuming an effective uv flux of zero @xcite . as suggested by the latter authors , constraints from direct observation
are required . in order to address this situation ,
we have begun to acquire muscles ( measurements of the ultraviolet spectral characteristics of low - mass exoplanetary systems ) . in this letter , we present our first results from muscles ; nuv and fuv spectra of gj 876 obtained with the _ hubble space telescope_-cosmic origins spectrograph ( cos ) and space telescope imaging spectrograph ( stis ) .
gj 876 was observed with cos and stis on 05 january 2012 and 12 november 2011 , respectively ( @xmath31 program i d 12464 ) .
multiple central wavelengths and focal - plane offset ( fp - poss ) settings with the cos g130 m and g160 m modes were used to create a continuous fuv spectrum from 1143 1798 .
these modes provide a point - source resolution of @xmath34@xmath35 @xmath2 17 km s@xmath30 with 7 pixels per resolution element @xcite . the total exposure times in the g130 m and g160 m modes are 2017s and 2779s , respectively . because cos is a slitless spectrograph , observations at ly@xmath0 are heavily contaminated by geocoronal emission .
we therefore observed the stellar ly@xmath0 profile with stis , using the g140m / cenwave 1222 mode ( @xmath34@xmath35 @xmath2 30 km s@xmath30 ) mode through the 52 @xmath4 0.1 slit for a total of 1138s . using the g230l mode on stis
, we also acquired the 1800 3140 spectrum of gj 876 at @xmath34@xmath35 @xmath27 400 km s@xmath30 .
in figure 1 , we display the full 1150 3140 uv spectrum of gj 876 ( @xmath36 ) . we observe emission from atomic and ionic species tracing a range of formation temperatures , including ly@xmath0 , , ( @xmath6 @xmath3 1 3 @xmath4 10@xmath37 k ) ; , , , ( @xmath6 @xmath3 6 8 @xmath4 10@xmath37 k ) ; , , and ( @xmath6 @xmath3 1.0
1.6 @xmath4 10@xmath38 k ) .
these emission lines clearly indicate that there is an active chromosphere and transition region on this star . as suggested by @xcite ,
gj 876 is @xmath39 an inactive star when observed in the uv .
gj 876 may be in the class of `` weakly active '' stars whose optical spectra show h@xmath0 in absorption while displaying h & k emission lines @xcite .
the stis g230l observations show emission lines from and as well as a weak nuv continuum at @xmath33 @xmath7 2200 . in figure 1 , we plot the flux observed at the orbit of gj 876b , roughly centered in the hz of gj 876 ( @xmath40 = @xmath41 @xmath4 ( @xmath42/@xmath17)@xmath43 ; @xmath17 = 0.21 au ) . the hz spectrum of gj 876 is compared with the solar spectrum ( woods et al . 2009 ) .
the gj 876 data have been convolved with a 2 fwhm gaussian kernel for comparison with the solar data .
table 1 presents the band- and line - integrated hz fluxes for several key regions / species . in order to produce an accurate estimate of the ly@xmath0 flux incident on the planets in the gj 876 system , the ly@xmath0 profile observed by stis
must be corrected for attenuation by interstellar hydrogen and deuterium ( e.g. , ehrenreich et al .
2011 ) . in order to do this , we approximate the intrinsic ly@xmath0 emission line as a two - component gaussian ( as suggested for transition region emission lines from late - type stars by wood et al .
1997 ) and create a simultaneous least - squares fit to the stellar emission and interstellar absorption components .
the ly@xmath0 emission components are characterized by an amplitude , fwhm , and velocity centroid ( @xmath44 ) .
interstellar absorption is parameterized by the column density of neutral hydrogen ( @xmath45(hi ) ) , doppler @xmath46 parameter , d / h ratio , and the velocity of the interstellar absorbers ( @xmath47 ) .
we assumed a fixed d / h ratio ( d / h = 1.5 @xmath4 10@xmath48 ; linsky et al .
1995 ; 2006 ) .
a two - component fit with fwhms = 133 and 303 km s@xmath30 , and an interstellar atomic hydrogen absorber characterized by log@xmath49(@xmath45 ( ) ) = 18.06 and @xmath46 = 8.0 km s@xmath30 , is able to produce an adequate fit ( @xmath50 = 1.73 ) to the observations .
the best model has @xmath51 = + 2.4 km s@xmath30 , @xmath52 = @xmath538.0 km s@xmath30 , and @xmath47 = @xmath533.0 km s@xmath30 ; the ly@xmath0 emission is somewhat blueshifted relative to the + 8.7 km s@xmath30 radial velocity of the star .
the ism velocity from our best - fit model is consistent with the expected local interstellar cloud velocity towards gj 876 ( @xmath54 = -3.56 km s@xmath30 ; redfield & linsky 2008 ) , supporting our ly@xmath0 reconstruction analysis .
a single gaussian ly@xmath0 emission line is unable to reproduce the observed profile ( @xmath50 @xmath7 6 ) .
the reconstructed stellar ly@xmath0 profile is shown in figure 2 .
integrated over the model profile , the total intrinsic ly@xmath0 flux is 4.38 @xmath4 10@xmath55 erg @xmath56 s@xmath30 .
the uncertainty on the integrated intrinsic flux is @xmath2 5 10 % , while the uncertainty on any individual fit parameter is most likely @xmath2 20 % .
we note that the correction for interstellar scattering does not take into account higher order opacity sources such as the stellar wind and heliosphere , which are not expected to significantly affect our results @xcite .
c|ccc @xmath1(ly@xmath0 ) & ( 1210 1220 ) & 5.6 @xmath4 10@xmath57 & 9.5 ( @xmath58 0.7 ) @xmath4 10@xmath57 + @xmath1 ( ) & ( 1235
1245 ) & 1.8 @xmath4 10@xmath59 & 2.0 ( @xmath58 0.1 ) @xmath4 10@xmath30 + @xmath1 ( ) & ( 1390
1410 ) & 5.6 @xmath4 10@xmath59 & 1.9 ( @xmath58 0.1 ) @xmath4 10@xmath30 + @xmath1 ( ) & ( 1545 1555 ) & 1.2 @xmath4 10@xmath30 & 4.2 ( @xmath58 0.3 ) @xmath4 10@xmath30 + @xmath1 ( ) & ( 1635
1645 ) & 2.4 @xmath4 10@xmath59 & 1.2 ( @xmath58 0.1 ) @xmath4 10@xmath30 + @xmath1(fuv ) & ( 1150 1210 ) + & 1.8 @xmath4 10@xmath60 & 4.1 ( @xmath58 0.6 ) @xmath4 10@xmath57 + & ( 1220 1790 ) & & + @xmath1(nuv ) & ( 2200 3140 ) & 2.2 @xmath4 10@xmath37 & 9.8 ( @xmath58 0.6 ) @xmath4 10@xmath57 + @xmath1(ly@xmath0)/@xmath1(fuv ) & & 0.3 & 2.3 + @xmath1(ly@xmath0)/@xmath1(fuv+nuv ) & & 2.6 @xmath4 10@xmath61 & 0.7 + a surprising result from the cos g160 m observations is the detection of relatively strong molecular hydrogen fluorescence . we clearly detect six h@xmath5 emission lines pumped by ly@xmath0 through the ( 1
2 ) r(6 ) @xmath331215.73 and ( 1
2 ) p(5 ) @xmath331216.07 transitions .
the lines pumped through ( 1
2 ) r(6 ) are : [ ( 1 7)r(6 ) @xmath331500.24 and ( 1 7)p(8 ) @xmath33 1524.65 ] .
the lines pumped through ( 1
2 ) p(5 ) are : [ ( 1 6)r(3 ) @xmath331431.01 , ( 1 6)p(5 ) @xmath33 1446.12 ,(1 7)r(3 ) @xmath331489.57 , and ( 1 7)p(5 ) @xmath33 1504.76 ] .
the ly@xmath0-pumping mechanism requires a significant population of molecules in the @xmath35 = 2 level of the ground electronic state , typically implying rotational temperatures of @xmath6(h@xmath5 ) @xmath7 2000k ( see herczeg et al . 2004 and france et al . 2010a ) .
the relative h@xmath5/atomic line ratios are large compared with other m - stars ( ad leo , proxima cen , ev lac , and au mic ) . since the s / n of the individual h@xmath5 lines is relatively low , we coadded the six cleanly detected lines ( figure 3 ) .
we present a discussion of the origin of the h@xmath5 emission in 4.2 . during the cos g130
m observations of gj 876 , we observed a flare in several lines formed in the active upper atmosphere of the star .
the light curves were extracted from the calibrated three - dimensional data by exploiting the time - tag capability of the cos microchannel plate detector @xcite .
we extract a [ @xmath62,@xmath63,@xmath64 photon list ( where @xmath33 is the wavelength of the photon , @xmath65 is the cross - dispersion location , and @xmath66 is the photon arrival time ) from each exposure @xmath67 and coadd these to create a master [ @xmath33,@xmath65,@xmath66 ] photon list .
the total number of counts in a [ @xmath34@xmath33,@xmath34@xmath65 ] box is integrated over a timestep @xmath34@xmath66 .
the instrumental background level is integrated over the same wavelength interval as the emission lines , but offset below the active science region in the cross - dispersion direction .
figure 4 shows the emission line count rates evaluated in @xmath34@xmath66 = 40s timesteps .
the count rate in the emission lines at exposure times @xmath70 @xmath7 4800s is decaying from the flare maximum .
the peak of the flare most likely occurred while the target was occulted by the earth ( the break in the observations between @xmath70 @xmath2 1300
4800 s ) , which prevents us from measuring the actual peak or the full decay time of the flare .
however , the @xmath71 flare maximum is approximately 10@xmath4 the quiescent level , with an exponential decay time scale @xmath3 400s , implying a peak flare / quiescent fuv flux ratio of @xmath8 10 .
our data do not have the temporal coverage to determine the fuv flare frequency on gj 876 , but one could expect flares of this magnitude on timescales of hours to days @xcite .
we observe a weak fuv continuum associated with the flare event .
integrating the total number of counts over an emission line - free 15 region centered near @xmath33 @xmath3 1380 , we find an orbit - averaged continuum / background ratio of 1.21 @xmath58 0.08 .
figure 1 and table 1 show that the fuv / nuv ratio is much larger for m - dwarfs than for solar - type stars , and that the intrinsic ly@xmath0 flux dominates the hz uv spectrum of gj 876 . comparing the intrinsic ly@xmath0 flux to the rest of the fuv ( 1150
1210 + 1220 1790 ) , we see that the ly@xmath0/(fuv ) ratio is @xmath3 8@xmath4 the solar value .
the reconstructed ly@xmath0 emission line has roughly the same integrated flux as the rest of the uv spectrum ( 1150
3140 ) combined .
this is striking when compared to the relative contribution of ly@xmath0 to the total uv luminosity ( ly@xmath0/(fuv+nuv ) ) from the sun .
gj 876 displays a ly@xmath0/(fuv+nuv ) ratio @xmath3 2500@xmath4 larger than the solar value .
because most of the uv luminosity from m - dwarfs is emitted in emission lines , molecular species in the atmospheres of orbiting planets will be `` selectively pumped '' ; only species that spectrally coincide with stellar uv emission lines will be subject to large energy input from the host star . in order to estimate of the impact of ly@xmath0 irradiation , we computed the total flux absorbed by several important atmospheric constituents , @xmath72(@xmath73 ) = @xmath74(gj 876 ) @xmath4 ( 1
@xmath75 ) , where ( @xmath76(@xmath73 ) ) is calculated from published photoabsorption cross - sections for @xmath73 = co@xmath5 @xcite , h@xmath5o @xcite , or ch@xmath11
we find that ly@xmath0 photons account for @xmath7 40% of the energy absorbed by co@xmath5 and @xmath7 85% of the energy absorbed by h@xmath5o and ch@xmath11 .
other chromospheric and transition region emission lines contribute only a few percent to the energy absorbed by h@xmath5o and ch@xmath11 .
, , and each contribute @xmath3 10% to the fuv stellar flux absorbed by co@xmath5 . while we are not presenting a rigorous treatment of the radiative transfer of ly@xmath0 photons incident on a planetary atmosphere ( this will be done in future modeling work ) , this rough calculation highlights the importance of ly@xmath0 emission to the total energy budget available to drive atmospheric chemistry on these worlds .
strong fuv flux should be included in models of terrestrial planets in the hz of `` weakly active '' and perhaps `` inactive '' m - dwarf host stars .
stellar fuv flux will drive co@xmath5 photolysis , leading to an enhancement of atmospheric o@xmath5 ( and o@xmath10 ) , although the exact enhancement factor must be determined .
@xcite showed that o@xmath5 and o@xmath10 column densities are enhanced by a factor of @xmath3 3 in co@xmath5-rich atmospheres when the fuv - bright young solar analog ek dra is used as the illuminating radiation field . in a terrestrial atmosphere illuminated by an older solar - type star
, the much stronger nuv flux would serve to balance the possible enhancement of o@xmath10 ( the o@xmath10 photolysis threshold is at @xmath33 @xmath3 2500 ) .
however the weak nuv field of gj 876 should not destroy significant amounts of o@xmath10 .
we therefore suggest that the relatively strong fuv flux may increase the amount of atmospheric o@xmath10 present relative to what is predicted by models of terrestrial hz planets orbiting `` inactive '' m - dwarfs @xcite .
this highlights the need for measured uv spectra of the host star when modeling the atmospheric properties of an exoplanet .
the paucity of uv observations of low - mass exoplanetary host stars may bias our interpretation of potential biomarkers when direct spectroscopy of extrasolar terrestrial planets becomes feasible , possibly in the coming decade with the _ james webb space telescope _
@xcite .
the cos spectra of gj 876 show surprisingly strong emission from fluorescent h@xmath5 , excited by ly@xmath0 photons .
jupiter s dayglow and auroral spectra are dominated by collisionally - excited and ly@xmath77-pumped h@xmath5 @xcite .
however , emission from hot h@xmath5 pumped by ly@xmath0 photons was observed at the shoemaker - levy 9 impact site @xcite .
if there were a heat source capable of producing high molecular temperatures ( @xmath6(h@xmath5 ) @xmath7 2000 k ) in the atmosphere of gj 876b or gj 876c , then the observed fluorescence could have an exoplanetary , rather than a stellar origin .
similar emission was searched for from the well - studied transiting planet hd 209458b , but not conclusively detected @xcite .
we can not rule out the possibility that some of the h@xmath5 emission is associated with gj 876b or gj 876c , but a calculation of the available energy budget argues for a stellar origin for this fluorescence .
from the reconstructed ly@xmath0 profile ( 3.2 ) , we can calculate the ly@xmath0 surface flux , @xmath78(ly@xmath0 ) = @xmath41(ly@xmath0 ) @xmath4 ( @xmath42/@xmath79)@xmath43 , where @xmath79 is the stellar radius @xmath79 @xmath2 0.3 @xmath80 . the ly@xmath0 flux absorbed by h@xmath5
is then @xmath81(h@xmath5 ) = @xmath78(ly@xmath0 ) @xmath4 ( 1
@xmath82 ) , where @xmath76(h@xmath5 ) is the optical depth of h@xmath5 , calculated for @xmath45(h@xmath5 ) = 10@xmath83 @xmath56 ( the absorbed flux only increases by a factor of @xmath3 2.5 when @xmath45(h@xmath5 ) = 10@xmath84 @xmath56 ) and @xmath6(h@xmath5 ) = 2500 k. if we assume that the h@xmath5 resides in an optically thin circular cloud , the emitted and absorbed luminosities will be equal , @xmath85(h@xmath5 ) = @xmath86@xmath87@xmath81(h@xmath5 ) = @xmath88(h@xmath5 ) .
focusing on the [ @xmath89,@xmath90 = [ 1,4 ] progression , @xmath88(h@xmath5 ) = 4@xmath86@xmath91 @xmath4 ( @xmath92/@xmath93 ) , where @xmath92 is the flux in the h@xmath5 ( 1
7)p(5 ) @xmath331504.76 emission line ( @xmath92 = ( 2.2 @xmath58 0.4 ) @xmath4 10@xmath94 erg @xmath56 s@xmath30 ) and @xmath93 is the @xmath89 @xmath95 @xmath96 = 1 @xmath95 7 branching ratio ( @xmath93 = 0.115 ) . equating @xmath88(h@xmath5 ) and @xmath85(h@xmath5 ) , we find @xmath97 = 0.32 @xmath58 0.13 @xmath80 , consistent with the h@xmath5 being located in the stellar atmosphere .
a similar calculation at the semi - major axis of gj 876b would require @xmath97 @xmath7 400@xmath98 , much larger than observational or theoretical estimates of the size of hot jupiter envelopes @xcite .
we conclude that the observed fluorescence is produced by hot photospheric molecules illuminated by the strong chromospheric ly@xmath0 radiation field .
+ + k.f .
thanks sarah levine for graphic support of project muscles .
we acknowledge enjoyable discussions with john stocke , hao yang , and brian wolven .
these data were obtained as part of the @xmath31 guest observing program # 12464 .
this work was supported by nasa grants nnx08ac146 and nas5 - 98043 to the university of colorado at boulder . , w. j. , koch , d. g. , basri , g. , batalha , n. , brown , t. m. , bryson , s. t. , caldwell , d. , christensen - dalsgaard , j. , cochran , w. d. , devore , e. , dunham , e. w. , gautier , iii , t. n. , geary , j. c. , gilliland , r. , gould , a. , howell , s. b. , jenkins , j. m. , latham , d. w. , lissauer , j. j. , marcy , g. w. , rowe , j. , sasselov , d. , boss , a. , charbonneau , d. , ciardi , d. , doyle , l. , dupree , a. k. , ford , e. b. , fortney , j. , holman , m. j. , seager , s. , steffen , j. h. , tarter , j. , welsh , w. f. , allen , c. , buchhave , l. a. , christiansen , j. l. , clarke , b. d. , das , s. , dsert , j .-
m . , endl , m. , fabrycky , d. , fressin , f. , haas , m. , horch , e. , howard , a. , isaacson , h. , kjeldsen , h. , kolodziejczak , j. , kulesa , c. , li , j. , lucas , p. w. , machalek , p. , mccarthy , d. , macqueen , p. , meibom , s. , miquel , t. , prsa , a. , quinn , s. n. , quintana , e. v. , ragozzine , d. , sherry , w. , shporer , a. , tenenbaum , p. , torres , g. , twicken , j. d. , van cleve , j. , walkowicz , l. , witteborn , f. c. , & still , m. 2011 , , 736 , 19 , j. l. , draine , b. t. , moos , h. w. , jenkins , e. b. , wood , b. e. , oliveira , c. , blair , w. p. , friedman , s. d. , gry , c. , knauth , d. , kruk , j. w. , lacour , s. , lehner , n. , redfield , s. , shull , j. m. , sonneborn , g. , & williger , g. m. 2006 , , 647 , 1106 , r. , parafita , r. , giuliani , a. , hubin - franskin , m. , lourenco , j. m. c. , garcia , g. , hoffmann , s. v. , mason , n. j. , ribeiro , p. a. , raposo , m. , & limao - vieira , p. 2005 , chemical physics letters , 416 , 152 , s. , green , j. , froning , c. , bland , s. , burgh , e. , france , k. , penton , s. , delker , t. , ebbets , d. , sahnow , d. , bacinski , j. , kimble , r. , andrews , j. , wilkinson , e. , mcphate , j. , siegmund , o. , ake , t. , aloisi , a. , biagetti , c. , diaz , r. , dixon , w. , friedman , s. , ghavamian , p. , goudfrooij , p. , hartig , g. , keyes , c. , lennon , d. , massa , d. , niemi , s. , oliveira , c. , osten , r. , proffitt , c. , smith , t. , & soderblom , d. 2011 , , 157 , b. y. , wheatley , j. m. , seibert , m. , browne , s. e. , west , a. a. , siegmund , o. h. w. , barlow , t. a. , forster , k. , friedman , p. g. , martin , d. c. , morrissey , p. , small , t. , wyder , t. , schiminovich , d. , neff , s. , & rich , r. m. 2007 , , 173 , 673 | extrasolar planets orbiting m - stars may represent our best chance to discover habitable worlds in the coming decade . the ultraviolet spectrum incident upon
both earth - like and jovian planets is critically important for proper modeling of their atmospheric heating and chemistry . in order to provide more realistic inputs for atmospheric models of planets orbiting low - mass stars ,
we present new near- and far - ultraviolet ( nuv and fuv ) spectroscopy of the m - dwarf exoplanet host gj 876 ( m4v ) . using the cos and stis spectrographs aboard the _ hubble space telescope _ , we have measured the 1150
3140 spectrum of gj 876 .
we have reconstructed the stellar ly@xmath0 emission line profile , and find that the integrated ly@xmath0 flux is roughly equal to the rest of the integrated flux ( 1150
1210 + 1220 3140
) in the entire ultraviolet bandpass ( @xmath1(ly@xmath0)/@xmath1(fuv+nuv ) @xmath2 0.7 ) .
this ratio is @xmath3 2500@xmath4 greater than the solar value .
we describe the ultraviolet line spectrum and report surprisingly strong fluorescent emission from hot h@xmath5 ( @xmath6(h@xmath5 ) @xmath7 2000 k ) .
we show the light - curve of a chromospheric + transition region flare observed in several far - uv emission lines , with flare / quiescent flux ratios @xmath8 10 .
the strong fuv radiation field of an m - star ( and specifically ly@xmath0 ) is important for determining the abundance of o@xmath5 and the formation of biomarkers in the lower atmospheres of earth - like planets in the habitable zones of low - mass stars . |
few bodies problems have been studied for long time in celestial mechanics , either as simplified models of more complex planetary systems or as benchmark models where new mathematical theories can be tested .
the three body problem has been a source of inspiration and study in celestial mechanics since newton and euler , in particular the restricted three body problem ( r3bp ) has demonstrated to be a good model of several systems in our solar system such as the sun
jupiter asteroid system , and with less accuracy the sun
earth moon system , in these systems the r3bp was used to know preliminary orbits in some space missions . in analogy with the r3bp , in this paper
we study a restricted problem of four bodies consisting of three primaries moving in circular orbits keeping an equilateral triangle configuration and a massless particle moving under the gravitational attraction of the primaries .
it is known that in our solar system we can find such configurations , the so called trojan asteroids of jupiter , mars and neptune form approximately an equilateral configuration with their respective planet and the sun , saturn tethys
telesto , saturn tethys
calypso or saturn dione
helen are good examples of such configuration .
several authors @xcite , @xcite , have considered the restricted four body body problem to model the dynamics of a spacecraft in the sun - jupiter - asteroid - spacecraft system .
+ + g. w. hill developed his famous lunar theory @xcite as an alternative approach for the study of the motion of the moon .
as a first approximation , this approach consider a kepler problem ( earth - moon ) with a gravitational perturbation produced by a far away massive body ( sun ) , some orbital elements such as the eccentricities of the orbits of the moon and the earth and the inclination of the moon are supposed to be zero .
previously to the hill s work , the approach to study the dynamics of the moon consisted on considering two kepler problems , one for the motion of the earth and the moon around their center of mass and other for the motion of the sun and such center of mass .
however , this approach had several difficulties because of the solutions were given in terms of formal power series of orbitals elements , the principal inconvenience was due to the poor convergency of these series in terms of the ratios of the mean motions of the earth and the moon , the so called critical parameter .
the success of the hill s approach was given by using his model to obtain a periodic orbit of the trajectory of the moon and then he included the orbital elements to correct it , in such a way , he avoided the computation of expansions in terms of the critical critical parameter . in a four body problem context , the smallness of one primary creates complicated equations of motion where an analytical study is extremely difficult to make and even there are technical inconveniences in the accuracy of numerical simulations . in the next sections we develop a model as a first approximation of the dynamics of a masses particle in a sun - planet - asteroid system , as possible applications of this model we can consider the massless body like a spacecraft or a small satellite like the moon of the trojan asteroid 624 hektor @xcite . in future works we may include relevant effects produced by inclinations and librations of the asteroids or perturbations due to other bodies for example .
consider three point masses , called @xmath2 , moving in circular periodic orbits around their center of mass under their mutual newtonian gravitational attraction , forming an equilateral triangle configuration .
a fourth massless particle is moving under the gravitational attraction of the primaries , this problem is known as the equilateral restricted four body problem or simply as the restricted four body problem ( r4bp ) .
the equations of motion in the usual dimensionless coordinates of the massless particle referred to a synodic frame of reference , where the primaries remain fixed , are : , width=240 ] @xmath3 @xmath4 @xmath5 where @xmath6 and @xmath7 , for @xmath8 .
the general expressions of the coordinates of the primaries in terms of the masses of the three primaries are given by @xmath9 @xmath10 @xmath11}{2k\sqrt{m_{2}^{2}+m_{2}m_{3}+m_{3}^{2}}},\ ] ] @xmath12 @xmath13 @xmath14 where @xmath15 and the three masses satisfy the relation @xmath16 .
it can be proved that the equations of motion have a first integral @xmath17 where @xmath18 is a constant .
it is worth noting that when we make @xmath19 and @xmath20 we recover the coordinates of the restricted three body problem ( r3bp ) @xmath21 , @xmath22 and @xmath23 , now the `` phantom '' mass @xmath0 is located in the so called equilibrium point @xmath24 of the r3bp . in the following
, it will be necessary to consider the hamiltonian of the system @xmath25
in this section we will discuss the how to compute the limit when @xmath1 for the r4bp .
we use a similar procedure as shown in @xcite by considering a symplectic scaling of the hamiltonian and expansions in taylor series in a neighborhood of the small mass @xmath0 .
the resulting hamiltonian will be a three degrees of freedom system depending on a parameter @xmath26 which is the mass of the primary @xmath27 .
the limit @xmath1 of the hamiltonian ( [ originalhamiltonian ] ) restricted to a neighborhood of @xmath0 exists and gives rise to a new hamiltonian @xmath28 @xmath29 where @xmath30 and @xmath20 . _
proof_. we consider the hamiltonian of the restricted four body problem ( r4bp ) in the center of mass coordinates @xmath31 where @xmath32 and @xmath33 denotes the position of the primary @xmath34 for @xmath35 .
we make the change of coordinates @xmath36 ,
@xmath37 , @xmath38 , @xmath39 , @xmath40 , @xmath41 , therefore in these new coordinates the hamiltonian ( [ originalhamiltonian ] ) becomes @xmath42 where now we have @xmath43 for @xmath35 .
we expand the terms @xmath44 and @xmath45 in taylor series around the new origin of coordinates , if we ignore the constant terms we obtain the following expressions @xmath46 @xmath47 where @xmath48 is a homogenous polynomial of degree @xmath49 for @xmath50 in order to take the limit as @xmath1 , we perform the following symplectic scaling @xmath51 , @xmath52 , @xmath53 , @xmath54 , @xmath55 @xmath56 with multiplier @xmath57 , therefore @xmath58 @xmath59 a straightforward computation shows @xmath60 @xmath61 where @xmath62 it is important to note that the the first partial derivative is given by @xmath63 for @xmath64 therefore we obtain@xmath65 and @xmath66 now if we recall that the three masses are in equilateral configuration and we use the relation @xmath67 we obtain @xmath68x\ ] ] @xmath69y,\ ] ] in terms of the coordinates of the primaries ( [ coordinatesprimaries ] ) , we can write @xmath70=-m_{3}^{2/3}m_{2}s_{1}(m_{1},m_{2},m_{3})+m_{3}^{2/3}y_{3},\ ] ] where @xmath71 and @xmath14 a similar computation shows that the coefficient @xmath72 $ ] can be written in terms of a positive power of @xmath0 .
therefore , the hamiltonian ( [ scaled ] ) looks like @xmath73 we have defined @xmath74 .
now , we are allowed to take the limit @xmath1 in the expression ( [ unlimitedhamiltonian ] ) , the final expression looks like @xmath75 where @xmath76 , @xmath20 and @xmath30 .
* remarks * * the expression @xmath77 is the quadratic part of the hamiltonian of the restricted three body problem centered in the so called equilibrium point @xmath24 . * because of the above remark , we can consider the mass parameter in the range @xmath78 $ ] , the case where @xmath79 corresponds to the equal massive bodies case . *
it will proved in the next section that this system have 4 equilibrium points in a neighborhood of @xmath0 and such equilibrium points will posses the same stability properties as in the full r4bp when @xmath0 is small but non zero .
now the gravitational and effective potential are @xmath80 @xmath81 respectively .
the equations of motion can be written as in the full problem @xmath82 @xmath4 @xmath5 but @xmath83 is given by the equation ( [ hillefectivepotential ] ) .
in this section we prove that the system has 4 equilibrium points and we will be able to compute them explicitly in terms of the mass parameter @xmath26 .
so , in order to find the equilibrium points of the limit case , as usual , we need to find the critical points of the effective potential ( [ hillefectivepotential ] ) , an easy computation shows that @xmath84 the equation @xmath85 implies that @xmath86 so the equilibrium points of the system are coplanar .
therefore , it is enough to study the critical points of the planar effective potential @xmath87 in matrix notation .
@xmath88 where @xmath89 and @xmath90 is the matrix @xmath91 the above matrix has eigenvalues @xmath92 @xmath93 with respective eigenvectors @xmath94 and @xmath95 where @xmath96 .
the eigenvectors have been chosen such that @xmath97 .
the equation to be solved is @xmath98 , or explicitly @xmath99 we can use the invertible matrix @xmath100 to solve the above equation , if we consider the linear change of variables @xmath101 , substitute in the equation ( [ gradient ] ) and multiply by @xmath102 , we obtain @xmath103 or equivalently @xmath104 where @xmath105 is given by the diagonal matrix @xmath106 in terms of coordinates the equation ( [ gradientdiagonal ] ) is equivalent to the system @xmath107 @xmath108 it is clear that in the above equations the case @xmath109 corresponds to a singularity and the case @xmath110 , @xmath111 gives rise a contradiction , therefore when @xmath112 we have @xmath113 or equivalently @xmath114{\lambda_{1}}\vert v_{1}\vert},\ ] ] on the other hand , when @xmath115 we have @xmath116 or equivalently @xmath117{\lambda_{2}}\vert v_{2}\vert},\ ] ] but @xmath97 , therefore we obtain four equilibrium points given by @xmath118{\lambda_{2 } } } ) , l'_{2}=(0,-\frac{1}{\sqrt[3]{\lambda_{2 } } } ) , l'_{3}=(\frac{1}{\sqrt[3]{\lambda_{1}}},0 ) , l'_{4}=(-\frac{1}{\sqrt[3]{\lambda_{1}}},0),\end{aligned}\ ] ] or in the original coordinates we have @xmath119 for @xmath120 .
it is easy to see that @xmath121{\lambda_{2}}}v_{2 } , l_{3}=\frac{1}{\sqrt[3]{\lambda_{1}}}v_{1},\end{aligned}\ ] ] and @xmath122 , @xmath123 . in the previous subsection we obtained explicit expression of the four equilibrium points in terms of the parameter @xmath26 so
, we can analyze the stability in the whole range @xmath78 $ ] , we will perform such analysis for the planar case @xmath86 .
we need to linearize the equations of motion , i.e. , we need to study the linear system @xmath124 , where @xmath125 and @xmath126 is the matrix @xmath127 where the partial derivatives @xmath128 @xmath129 and @xmath130 need to be evaluated in each @xmath131 for @xmath132 however we must observe that because of the symmetry of the equilibrium points , we just need to study the equilibrium points @xmath133 and @xmath134 .
it is well known that the characteristic polynomial of the matrix ( [ matrixlinearization ] ) is given by the expression @xmath135 where @xmath136 , @xmath137
. therefore the four eigenvalues are @xmath138 with @xmath139
. a equilibrium point will be linearly stable if only if @xmath126 , @xmath140 and @xmath105 are non - negatives .
because of we know explicit expressions in term of the mass parameter @xmath26 for each equilibrium point , the referred coefficients of the characteristic polynomial are functions of @xmath26 so , using techniques of calculus we can study the behavior of such coefficients . in the figures [ stabilityl1 ] and [ stabilityl3 ] we can observe the behavior of @xmath126 , @xmath140 and @xmath105 as functions of the parameter @xmath26 . the coefficient @xmath140 in negative for @xmath78 $ ] so the equilibrium point @xmath133 is unstable for this range of values of the mass parameter , in fact , the eigenvalues are given by @xmath141 and @xmath142 with @xmath143 and @xmath144 .
there exists a value @xmath145 such that @xmath146 , as a consequence , this equilibrium point has the following properties : for @xmath147 the eigenvalues are @xmath148 and @xmath149 , for @xmath150 we have a pair of the eigenvalues @xmath142 of multiplicity 2 , finally when @xmath151 $ ] the eigenvalues are @xmath152 with @xmath153 and @xmath144 . in the figure [ limithillregions ]
we show the so called hill s regions for the planar case and for @xmath154 that corresponds to mass ratio of the sun - jupiter system , in the first two figures of the first row we show the hill s regions for the limit problem and for the full r4bp when @xmath155 , the mass ratio of the asteroid 624 hektor , the lines in the second figure are imaginary lines that connect @xmath0 with the remaining masses .
we have marked the position of the fixed mass with a black dot and the positions of the four equilibrium points with red dots .
99 baltagiannis , a.n . ,
papadakis , k.e . ; periodic solutions in the sun - jupiter - trojan asteroid - spacecraft system .
planetary and space science . * 75 * , 148157 ( 2013 ) .
ceccaroni m. , biggs j. ; extension of low - thrust propulsion to the autonomous coplanar circular restricted four body problem with application to future trojan asteroid missions . in : 61st int .
congress iac 2010 prague , czech republic ( 2010 ) .
, researches in the lunar theory.american journal of mathematics .
* 1 * 5 - 26 ( 1878 ) .
marchis , f. , et al : the puzzling mutual orbit of the binary trojan asteroid ( 624 ) hektor .
astropysical letters , apj * 783 * , l37 .
meyer k. ; introduction to hamiltonian dynamical systems and the n - body problem .
springer verlag . | the restricted four - body problem studies the dynamics of a massless particle under the gravitational force produced by three masses ( primaries ) in an equilateral configuration .
one primary , say @xmath0 , is considered too small compared with the other ones . in a similar way as in the classical hill s problem , we study the limit case @xmath1 in the hamiltonian of the r4bp . in this paper we prove that such limit exists and the resulting limit problem produces a new hamiltonian that inherits some basic features of the restricted three and four body problems .
we analyze some dynamical aspects of this new system that can be considered as a generalization of the hill s problem .
* keywords : * four body problem , hill
s problem , equilibrium points , stability , trojan asteroids .
* ams classification : * 70f10 , 70f15 |
in the standard model ( sm ) , interactions among elementary particles are mediated by the vector bosons of the strong , weak and electomagnetic forces .
experimental evidence for the existence of those bosons is compelling and precise measurements of their properties have been accumulated in the past decades .
new forces can have escaped detection so far , either if their associated bosons are very heavy or if their couplings to ordinary matter are weak enough . the latter case has been advocated , among others , in models which try to explain and reconcile among them several puzzling astrophysical observations performed in recent years @xcite .
they are sometimes also used to reconcile the measured value of the muon anomaly to the sm prediction , which differ by approximately by 3.5 @xmath3 ( see , for instance @xcite ) . if new , light , neutral bosons ( which from now on will be called @xmath0 or _ dark photons _ ) exist and if they are measurably , albeit weakly , coupled with sm particles they can be produced and observed at colliding - beams and fixed targed experiments @xcite .
in fact there have been several attempts to observe evidence for such particles , using data from running facilities @xcite or data - mining old experiments @xcite . since no evidence for their existence was found , limits have been set as a function of the @xmath0 mass and of its coupling to ordinary matter . in the near future , new experiments under construction are expected to extend those limits in a region of couplings and/or masses so far unexplored .
all of them are designed to exploit the radiative production of the @xmath0 by a very intense electron or positron beam on a properly built high - z target @xcite .
the purpose of the present letter is to show that comparable results can be obtained by high luminosity , low energy electron - positron colliders , such as those under construction or under study in several laboratories in the world @xcite .
these facilities will take advantage of two main construction features which coherently conspire to enhance their discovery potential : their very high goal luminosity , and the usage of very compact beams ( these two features are in fact strongly correlated ) .
actually , high luminosity translates into the possibility of probing lower production cross sections , i.e. lower effective couplings between the @xmath0 and ordinary matter . on the other hand ,
low couplings translate into longer @xmath0 decay paths , especially for low @xmath0 masses .
thus , the usage of beams of very small dimensions allows one to obtain a clear @xmath0 signal by observing secondary vertices of a well defined invariant mass , well separated by the beams interaction point . in the paper , this case will be discussed for three different possible choices of the machine center - of - mass energy , corresponding , respectively , to the mass of the @xmath4(1020 ) , the j@xmath5(1s ) and the @xmath6(4s ) mesons .
this choice is motivated by the projects mentioned above .
it will be shown that higher energy machines are favoured , not only because they are expected to deliver larger data sets , but also because the @xmath0 therein produced have longer decay paths , ceteris paribus .
instrumental effects play however a relevant role in the actual detection strategy , and can in some cases dramatically reduce the discovery potential of the method .
still , in particular for the case of a high luminosity @xmath7-charm factory , it remains high enough to be competitive to the fixed target experiments mentioned above .
the paper is organized as follows .
firstly , the theoretical framework of the paper is discussed , together with a short presentation of the experimental limits on the existence of dark photons obtained so far . the case for the searches at low energy , high luminosity e@xmath1e@xmath2 colliders
is discussed in section 3 , followed by some considerations on the actual implementation of the proposed method to existing , or planned , facilities .
radiative vector meson decays are discussed in section 5 .
conclusions are given in section 6 .
in many new physics scenarios , the sm is extended by simply adding an additional u(1)@xmath8 symmetry , under which sm particles are uncharged at first order @xcite .
the gauge boson associated to the new symmetry , the @xmath0 , can still interact with ordinary matter via kinetic mixing described by an effective interaction lagrangian @xmath9 where @xmath10 denotes the @xmath0 field . the kinetic mixing factor @xmath11 parametrizes the coupling strength relative to the electric charge and is predicted in various models to be in the range 10@xmath12 - 10@xmath13 .
the mass of the dark photon ( m@xmath14 ) rests unpredicted . on phenomenological grounds , however
, masses in the mev - gev range , which are of interest for the present work , are favoured .
there might exist non - sm matter particles which are sensitive to the new u(1)@xmath8 interaction .
often they are postulated to be the main constituent of the yet undiscovered dark matter component of the universe ( dm ) , and must therefore be electrically neutral and stable . if kinematically allowed , the @xmath0 will decay preferably into pairs of these particles , thus its decay becomes `` invisible '' .
the case for detecting invisible decays is discussed , for instance , in @xcite . on the other hand ,
if the dark photon is lighter than dm , it is forced to decay into a pair of sm particles , with a width regulated by eq .
[ eq : kinmix ] . in this case , its proper time is approximately given by @xcite @xmath15 where n@xmath16 is the number of sm decay channels allowed by kinematics .
there have been several attempts to experimentally observe a @xmath0 signal , using many different techniques .
figure [ fig : sarah ] , taken from reference @xcite , shows the exclusion plot in the plane @xmath17-@xmath11 , resulting from the above mentioned searches .
electron beam dump experiment cover the region of low masses and very low couplings , down to @xmath1810@xmath19 . for higher masses and lower couplings limits come mainly from meson decays , electron - nucleon scattering experiments and to b - factories data .
important information , not shown in figure [ fig : sarah ] , can also be deduced by astrophysical observations ( see for instance @xcite and references therein ) . for m@xmath2010 - 20 mev ,
the region with @xmath2110@xmath22 remains largely unexplored .
resulting from presently available data .
electron beam dump experiment cover the region of low masses and very low couplings , down to @xmath1810@xmath19 . for higher masses and lower couplings limits come mainly from meson decays , electron - nucleon scattering experiments and to b - factories data ( plot courtesy of s. andreas ) .
the plot reports also the projections for the experiments presently under construction . for details on single experiments see @xcite . ]
there are currently various experiments either running or under construction aiming at probing part of this region .
all of them are designed to exploit the radiative production of the @xmath0 by a very intense electron or positron beam on a properly built high - z target .
in particular the hps experiment at thomas jefferson laboratory ( usa ) is designed to cover the region @xmath11=(10@xmath23,10@xmath24 ) , m@xmath14=(20,300 ) mev @xcite ( see figure [ fig : sarah ] ) . in the following , the case for the search for dark photons in the same parameters space region at a very high luminosity e@xmath1e@xmath2 collider will be discussed .
in the last decades large amount of data have been collected at high luminosity e@xmath1e@xmath2 flavour factories operating at different center of mass energies .
these data range from the @xmath252 fb@xmath26 delivered at the @xmath4(1020 ) peak by the italian collider da@xmath27ne , to the 0.5 - 1 ab@xmath26 produced by the b - factories at pep - ii ( usa ) and kek - b ( japan ) . in the near future a consistent increase of the above statistics is expected both at da@xmath27ne and at kek - b which aim at increasing their data sample by a factor of 10 and 50 , respectively .
an option to increase the center of mass energy of da@xmath27ne up to 2.5 gev , has been taken into consideration @xcite .
finally , studies for the construction of a collider capable of delivering @xmath25 1 ab@xmath26 around the charm threshold are under consideration in italy , russia and china ( see for instance @xcite ) . as of today
, searches for dark photons at e@xmath1e@xmath2 colliders have been pursued mainly by studying the process e@xmath1e@xmath2@xmath28 with the subsequent decay of the @xmath0 into a @xmath29 pair .
this limits the search to m@xmath202 m@xmath30 , which , as a consequence of equation [ eq : lifet ] , results in its lifetime being unmeasurably short .
therefore the signal can be separated by the more copious and otherwise indistinguishible qed background , only by observing a sharp peak in the invariant mass distribution of the final state lepton pair .
the question arises whether it would be possible to extend the search also to the region with m@xmath312 m@xmath30 and in particular with @xmath2110@xmath22 .
the main message of the present paper is that the foreseen increase of the potentially available data sample allows one to give a positive answer to the above question , not only because of the increased statistical sensitivity but also because it opens the doors to the possibility of observing a clear signal for a long - lived @xmath0 , which is only marginal with the presently available data . here and in the following , for the sake of simplicity ,
only the case of symmetric machines is considered .
also , since we are interested at the case with m@xmath31 2m@xmath30 , the dark photon can decay only into a e@xmath1e@xmath2 pair .
the differential cross section for radiative @xmath0 production in e@xmath1e@xmath2 collisions , is given by @xcite @xmath32 where @xmath33 is the angle between the incoming positron and the outcoming photon and e@xmath34 denotes the center of mass energy of the event . by integrating the above equation between cos(@xmath33)= @xmath35 0.8 , one obtains a total cross section of approximately 0.9@xmath36 @xmath37b , 0.1@xmath36 @xmath37b and 3.6@xmath36 nb for e@xmath34 = 1.02 , 3.1 , 10.5 gev respectively .
thus , in the presently available data samples , for @xmath11=10@xmath23 a few dozens @xmath0 can be found . due to the two body kinematics
, the dark photon is boosted in the laboratory frame by a factor e@xmath34/m@xmath14 .
therefore for small kinetic mixings and for low enough dark photon masses , its lifetime in the laboratory frame becomes sizeable .
for instance ,
for @xmath11=10@xmath23 and m@xmath14=20 mev , the mean decay path of a dark photon is @xmath251 , 3 , 10 cm for e@xmath34 = 1.02 , 3.1 , 10.5 gev respectively .
can these long decay paths be exploited to separate a potential @xmath0 signal from the qed background ? clearly , although the _ secondary _ vertex can be determined with standard vertexing techniques , the _ primary production _ one can not on an event by event basis . on the other hand ,
the actual position and size of the collision envelope can be determined _ on a statistical basis _ using known processes , such as bhabha scattering or muon pair production .
interestingly , at all of the e@xmath1e@xmath2 facilities under consideration , one of the strategies used to maximize luminosity is to keep the transverse beam dimensions at the interaction point as small as possible , typically @xmath381 mm .
therefore , assuming a perfectly gaussian distribution of the beam spot , with a maximum transverse dimension of 1 mm , the probability for observing an e@xmath1e@xmath2 vertex from standard qed processes at a transverse distance of 1 cm or more from the center of the collision spot is practically zero . on the other hand
, the number of @xmath0 decay events with transverse decay path larger than 1 cm , n@xmath39 , can be as large as several thousands , depending on the actual value of @xmath11 , m@xmath14 , e@xmath34 and of the luminosity integrated by the machine , @xmath40 .
decay events with transverse decay path larger than 1 cm , as a function of @xmath11 , for different values of m@xmath14 .
three different experimental conditions are taken into consideration : a ) e@xmath34 = 1.02 gev , l@xmath41=20 fb@xmath26 ; b ) e@xmath34 = 3.1 gev , l@xmath41=1 ab@xmath26 ; c ) e@xmath34 = 10.5 gev , l@xmath41=50 ab@xmath26 .
note that the horizontal scale is slightly different for the three plots . ]
figure [ fig : fig2 ] shows the variation of n@xmath39 as a function of @xmath11 , for different values of m@xmath14 and for three different experimental conditions : a ) e@xmath34 = 1.02 gev , l@xmath41=20 fb@xmath26 ; b ) e@xmath34 = 3.1 gev , l@xmath41=1 ab@xmath26 ; c ) e@xmath34 = 10.5 gev , l@xmath41=50 ab@xmath26 .
the choosen values for l@xmath41 correspond to the target performance for the facilities under construction or under study mentioned above .
the behaviour of the curves is easily explained .
for @xmath4210@xmath23 the mean decay path of a dark photon is much larger than 1 cm , and therefore n@xmath39 increases with @xmath43 , independently of @xmath17 .
it eventually reaches a peak and drops rapidly towards zero , as long as , with increasing @xmath11 , the lifetime becomes shorter and shorter .
the position of the peak is determined by the proper balance between the effect of the production cross section , which increases with @xmath43 , and that of the lifetime which decreases with it .
it depends also on the value of m@xmath14 , the decay path decreasing again quadratically with it .
it is seen that , despite the lower production cross section , the largest expected integrated luminosity combined with the higher boost factors , favour the b - factory option ( case c ) ) . in this case , however , the peak of the distribution , especially for lower masses , is obtained for values of the kinetic mixing @xmath2510@xmath22 .
it can be noted also that in case a ) the number of observable dark photons with masses greater than @xmath25 30 mev becomes hopelessly small .
this is not only due to the lower luminosity but also to the reduced lorentz boost , consequence of the lower center of mass energy of the collision .
although the results obtained so far look very encouraging on general grounds , there are two main limitations coming from the implementation of the above search strategy into a real experiment . on the one hand , for specific values of the parameters ,
the @xmath0 lifetime becomes so long that a relevant part of the decays would escape detection of an apparatus of realistic dimensions .
for instance , for e@xmath34= 10.5 gev , @xmath11=5@xmath4410@xmath24 and m@xmath14=10 mev , the @xmath0 mean decay path is about 1.5 m. more importantly , a very dangerous _ instrumental _
background comes into operation , namely photon conversions on the detector material induced by e@xmath1e@xmath2@xmath45 events .
this is particularly relevant since experiments are often designed to have beam pipes with very small radii at the interaction point .
although one can try to minimize the conversion probability on the detector elements , by properly choosing type and dimensions of the materials , the cross section of the e@xmath1e@xmath2@xmath45 process is so much larger than the signal one ( in fact it is larger by a factor @xmath46 ) that this background rapidly becomes unbeatable . , with the further request that the transverse decay path is lower than 5 cm .
] the simplest way to cope with this problem is to allow for a large enough empty region around the interaction point , where photons can not interact with matter and dark photons can at least partly undergo their decay .
it would then be reasonable to accept only events with decay vertices occuring before the beam pipe but still far ( 1 cm ) from the nominal beam spot center . assuming a beam pipe of 5 cm radius , such as that presently used by the kloe-2 experiment at da@xmath27ne , the number of events thus obtained within acceptance , n@xmath47 , is shown in figure [ fig : fig3 ] for the three cases under consideration . for high values of @xmath11 this acceptance cut does not observably affect the previous distributions .
in fact , in this case , the lifetime is so short that almost all of the dark photons decay much before 5 cm . for lower values of @xmath11 ,
instead , the consequence of the cut in acceptance are more visible and can decrease the number of accepted events by an order of magnitude , especially for very low @xmath0 masses .
however , and this is one the main messages of this paper , the number of potentially observable events remains still considerable for a wide region of the parameter space , especially for the higher energy machine options . in particular , also allowing for some further detection inefficiency , it can be seen that kinetic mixings down to few times 10@xmath24 and masses up to @xmath25 200 mev can be probed .
based only on signal statistics ( i.e. without taking into consideration possible detector resolution effects and other possible instrumental backgrounds ) , this translates into the explorable regions shown in figure [ fig : fig5 ] , for the three cases under consideration .
while case a ) covers almost entirely a region already excluded by previous beam - dump experiments , cases b ) and c ) can potentially probe a relatively wide unexplored region ( cfr .
figure [ fig : sarah ] ) . on the other hand , it has also to be underlined that this same region is expected to be covered by the aforementioned future fixed target experiments ( see again figure [ fig : sarah ] ) .
no instrumental backgrounds are taken into consideration , as well as potential efficiency and resolution effects for the detector . above @xmath25210 mev
, the opening of the 2-muon channel drastically reduces the effectiveness of the method . ]
it is worthwhile stressing once more that the requirement for observing a cm - scale decay path , ideally rejects every possible _ physical _ background to our signal .
still , other instrumental effects need to be taken carefully into account , as discussed in the following section .
it is of interest understanding how difficult would it be to practically implement on real experimental facilities the ideas discussed so far .
this requires a detailed knowledge of the actual machine and detector s design and expected ( or measured ) performance . only specific studies based on these figures can in the end
determine whether the method is applicable or not , to which extent and on which machine .
an obvious difference between our simplified models and reality can be found for instance in case c ) ; both the old and the future b - factories are in fact asymmetric machines , the electron beam being of higher energy with respect to the positron one .
although this might somewhat change the specific acceptance requirements with respect to the symmetric option discussed in this paper , it is however reasonable to assume that similar conclusions can still be drawn . on general grounds , there are four parameters that have to be taken into consideration : the primary @xmath0 production rate , the dimensions of the beams , those of the beam pipe , and the vertexing capabilities of the detector . as for the first point , despite the higher production cross section , the @xmath4-factory option ( case a ) ) is less performing than the other two , not only because of the much lower integrated luminosity , but also , as noted before , because of the intrinsic limitation due to the lower lorentz boost factors .
the proposal for running da@xmath27ne at higher energies is , under this respect , particularly interesting because this would allow increasing the decay paths proportionally to e@xmath34 .
as previously remarked , for all of the machines under consideration , the dimensions of the beams are kept very small in the transverse direction .
for instance , at da@xmath27ne the beams have @xmath481.5 mm , @xmath490.02 mm , and much lower dimensions at the other machines .
note that both fig [ fig : fig2 ] and fig [ fig : fig3 ] , on which we base our search strategy , refer to transverse decay paths , so the beam dimensions in the longitudinal direction are irrelevant for our conclusions .
obviously , non gaussian tails of the collision envelope can to some extent increase the background contamination .
however , if not completely suppressed , they can still be studied using other known processes , as for the gaussian part .
a real concern are the actual beam - pipe dimensions . among the existing facilities ,
kloe-2 at da@xmath27ne is the only one having a beam pipe radius at the interaction point of 5 cm . as for the machines running at the charm threshold ,
the only one presently in operation , the chinese collider bepc , has a beam pipe at the interaction point of 3.5 cm in radius ; however its present luminosity is about a factor 100 lower than required by our arguments at that energy . for the b - factories , the beam pipe radii range from the 2.5 cm for the babar detector to 1 cm for the future superbelle .
while a reconsideration of the inner region of superbelle is probably very unlikely , in the case of a future high luminosity @xmath7-charm factory it is conceivable that the interaction region can be designed such as to maximise the sensitivity for the dark photon search under consideration .
it is worth noticing here that the request for minimizing the beam pipe dimensions comes to first order from the experiment more than from the machine .
in fact , they are somehow anti - correlated with the detector s vertexing capabilities . actually , the resolution of a decay length measurement for a generic detector is approximately proportional to the single point resolution of the most internal tracking device and inversely proportional to its distance from the decay point . under this respect
the less favourable situation is that of kloe-2 , whose first tracking device , a triple - gem cylindrical detector , has an internal radius of 12 cm and single point resolution of @xmath25200@xmath37 m .
still , its estimated vertex resolution for @xmath50 events is of 1 - 2 mm @xcite .
the use of silicon detectors , which can have single point resolution of order 10 @xmath37 m , would definitely improve with respect to the kloe-2 case .
noticeably , all the lep experiments , which had beam pipes of 5.5 cm , could reach typical decay length resolution of @xmath25250@xmath37 m for b decay events , thanks to the use of silicon detectors @xcite .
this implies that , considering the cm scale decay length we have been interested so far , vertexing resolution should not be a major issue . on the other hand
, it can play a relevant role in the considerations discussed in the following section .
electron - positron colliders provide a useful @xmath0 production mechanism via radiative vector meson decays , too .
actually , for each observed @xmath51 decay ( @xmath52 and @xmath53 being a vector and a pseudoscalar meson , respectively ) , there could be a @xmath54 process , suppresed by a factor @xmath43 with respect to the former one @xcite .
this fact has actually been exploited by the kloe-2 collaboration which has searched for the dark photon using the @xmath55 e@xmath1e@xmath2 process in @xcite . as for the searches in the e@xmath1e@xmath2@xmath56 channel ,
the signal is separated by the sm ( dalitz decay ) background by looking for a peak in the invariant mass distribution of the final state lepton pair . however , once the @xmath0 lifetime becomes sizeable , these events are characterised as well by the presence of e@xmath1e@xmath2 vertices clearly separated by the collision point , thus allowing the usage of the search strategy described in the previous sections .
the number of produced @xmath0 is given by @xmath57 where n@xmath58 is the number of produced vector mesons , and br@xmath59 is the branching ratio for the corresponding standard radiative decay . with decay paths larger than 1 cm and lower than 5 cm , as a function of @xmath11 and for various values of m@xmath14 .
an integrated luminosity of 1 ab@xmath26 is considered . ]
let us firstly consider the above mentioned @xmath60 process . at a @xmath4-factory ,
@xmath253@xmath6110@xmath62 @xmath4 mesons are produced every fb@xmath26 delivered by the machine . since [email protected]@xmath6110@xmath13 @xcite , it is easily seen that the number of produced signal events , assuming l@xmath41=20 fb@xmath26 , becomes negligible for @xmath6410@xmath23 . on the other hand , for higher values of @xmath11 , the @xmath0 mean decay path becomes unmeasurably ( as compared with millimiter scale vertex resolutions ) short , but for very low masses .
for instance , for @xmath11=2@xmath6110@xmath23 it is already 0.8(0.2 ) cm , for m@xmath14=10(20 ) mev . unless , therefore , one integrates luminosities largely exceeding those expected from the presently considered machine , the method is hardly applicable to this decay channel .
let us now turn our attention to the @xmath65 transition .
this process has been already studied in @xcite , where , however , the case for short lived dark photons only is considered . as before , one has n@xmath66 3@xmath6110@xmath62/fb@xmath26 , if running at the j/@xmath67 peak .
considering that br@xmath685@xmath6110@xmath22 @xcite , one obtains @xmath25150 events for @xmath11=10@xmath23 and l@xmath41=1 ab@xmath26 . on the other hand , for this value of kinetic mixing , the @xmath0 mean decay path is of the right order of magnitude only for a limited range of mass values .
it is , for instance , 11.2 , 2.8 , 0.45 cm for m@xmath14=10 , 20 , 50 mev respectively .
the effect of this is seen in figure [ fig : fig4 ] where the number of @xmath0 decays occurring at a distance between 1 and 5 cm from the interaction point is plotted as a function of @xmath11 , for different values of m@xmath14 . differently from the e@xmath1e@xmath2@xmath69 case , here it is the effect of the lifetime being short to be dominant , at least for the kinetic mixing values of interest , because for higher dark photon masses and low enough @xmath11 , almost all of the @xmath0 survive for less than 1 cm .
note , also , that some further reduction in the number of observable events must be considered due to geometrical acceptance considerations .
still , there remain a small region of the parameter space for which one can hope to observe a reasonable number of @xmath0 decays within acceptance .
there is however a further benefit specific of @xmath70 events : these processes can in fact be used also to measure _ on an event - by - event basis _ the actual dark photon decay path , provided that the final state meson decays into at least a pair of charged particles . in this case , the position of the latter particles determines the _ primary production _ vertex , while the @xmath0 decay position is determined , as usual , by the e@xmath1e@xmath2 one .
for instance , for @xmath65 events , one can use the process @xmath71 , where the @xmath72 pair allows determining exactly the collision point , at the price of reducing the total amount of useful events by a factor @xmath250.43 @xcite .
since we are interested in millimeter scale decay paths , the photon conversion background should not be a problem anymore .
however there is a physical background to be kept now into consideration , namely the dalitz decay of the j@xmath67 , @xmath73e@xmath1e@xmath2 .
its branching ratio can be estimated to be approximately br(@xmath73e@xmath1e@xmath2 ) @xmath25 br(@xmath74)@xmath440.01 , so that the process is @xmath2510@xmath75 times more frequent than the signal , if @xmath11=10@xmath23 .
however , in this case , the e@xmath1e@xmath2 and @xmath72 vertices must coincide within the detector s resolution , @xmath76 .
therefore a reduction of the background by a factor 10@xmath77 can be achieved by accepting only events with measured decay paths larger than 4@xmath76 .
moreover , background events are expected to have a broad e@xmath1e@xmath2 invariant mass distribution , while @xmath0 decays are a narrow resonance in such channel .
not considering form factor effects , the number of background events , n@xmath78 , in a window of @xmath79 m around m@xmath80=m@xmath14 is given approximately by @xcite @xmath81 note that the 1/m@xmath14 dependence of equation [ eq : eqnb ] favours the observation of higher mass dark photons .
on the other hand , since the decay path scales as 1/m@xmath82 , vertex resolution effects point in the opposite direction . in a given experiment ,
therefore , two parameters should ideally be kept as small as possible , @xmath76 and @xmath79 m .
take , for instance , @xmath11=10@xmath23 , m@xmath14=50 mev . according to eq .
[ eq : eqnb ] , the number of background events in the interesting mass bin would in this case be @xmath252@xmath4410@xmath83 for @xmath84=1 mev . by applying the 4@xmath76 cut discussed above , this number reduces to @xmath2520 .
therefore , for @xmath76=1(0.5 ) mm , the signal significance ( i.e. the number of signal events divided by the square root of the background ) would be @xmath254(6 ) .
it is important to underline that in this case the dimensions of the beam pipe are to first order irrelevant , since we are dealing with relatively short decay lengths . on the other hand
, it has also to be noted , that we are here assuming full detection efficiency which might reveal an overoptimistic assumption . as for the continuum events , only detailed studies based on realistic detector parameters can finally assess the potentials of the method .
experimental searches for a new , light , neutral boson , the `` dark photon '' or @xmath0 , are being pursued in many laboratories in the world , using different detection techniques .
if the @xmath0 is light enough and if its couplings with sm particles are suppressed by a factor @xmath3810@xmath22 with respect to those of the ordinary photon , it can acquire a relatively long lifetime .
this fact can be exploited at e@xmath1e@xmath2 colliders by searching a @xmath85 decay vertex well separated by the primary production one , in @xmath86 events .
the paper shows that new generation e@xmath1e@xmath2 colliders have the potentials to fully exploit this technique , and can explore effective couplings down to few times 10@xmath24 and @xmath0 masses in the range 10 - 100 mev approximately . in general higher energy machines
are favoured , because the higher lorentz boost of the produced ( light ) bosons allows a better separation of the secondary vertices .
it is also seen , however , that the implementation of this method to real facilities requires a proper design of the interaction region and a wise choice of the tracking detector . to our knowledge , unfortunately , the beam pipe dimensions of superbelle are far from being optimal under this respect . on the other hand , since all of the future high luminosity @xmath7-charm factories are still in a preliminary design study phase , it is conceivable that in this case the interaction region can be designed such as to maximise the sensitivity for the proposed @xmath0 search technique .
such a machine can also provide a complementary search method by the observation of displaced e@xmath1e@xmath2 vertices in fully reconstructed @xmath87 events , therefore enhancing the interest for the construction of such a facility .
i would like to thank s. andreas and d. babusci for help and useful discussions .
99 n. arkani - hamed , d. p. finkbeiner , t. r. slatyer and n. weiner , phys .
d * 79 * ( 2009 ) 015014 m. pospelov and a. ritz , phys .
b * 671 * ( 2009 ) 391 c. boehm and p. fayet , nucl .
phys b * 683 * ( 2004 ) 219 n. borodatchenkova , d. choudhury and m. drees , phys .
* 96 * ( 2006 ) 141802 m. pospelov , phys .
d * 80 * ( 2009 ) 095002 p. fayet , phys .
d * 75 * ( 2007 ) 115017 j. d. bjorken , r. essig , p. schuster and n. toro , phys .
d * 80 * ( 2009 ) 075018 b. batell , m. pospelov and a. ritz , phys .
d * 79 * ( 2009 ) 115008 m. reece and l. .t .
wang , jhep * 0907 * ( 2009 ) 051 b. batell , m. pospelov and a. ritz , phys .
d * 80 * ( 2009 ) 095024 r. essig , p. schuster and n. toro , phys .
d * 80 * ( 2009 ) 015003 l. barze et al . ,
j. c * 71 * ( 2011 ) 1680 h. merkel et al . ,
[ a1 collaboration ] , phys .
* 106 * ( 2011 ) 251802 s. abrahamyan et al . , [ apex collaboration ] phys .
* 107 * ( 2011 ) 191804 j.p .
lees et al . , [ babar collaboration ] phys .
* 108 * ( 2012 ) 211801 b. echenard , mod .
lett . a * 27 * ( 2012 ) 1230016 f. archilli et al . , [ kloe-2 collaboration ] phys .
b * 706 * ( 2012 ) 251 d. babusci et al .
, [ kloe-2 collaboration ] phys .
b * 720 * ( 2013 ) 111 p. adlarson et al .
, [ wasa - at - cosy collaboration ] phys .
b * 726 * ( 2013 ) 187 s. andreas , c. niebuhr and a. ringwald , phys .
d * 86 * ( 2012 ) 095019 j. blumlein and j. brunner , phys .
b * 701 * ( 2011 ) 155 s. n. gninienko , phys .
d * 85 * , ( 2012 ) 055027 s. n. gninienko , phys .
b * 713 * , ( 2012 ) 244 r. essig , p. schuster , n. toro and b. wojtsekhwski , jhep * 1102 * ( 2011 ) 009 p.
hansson adrian et al . , [ hps collaboration ] https://confluence.slac.edu/display/hpsg/project+overview t. beranek .
h. merkel and m. vanderhaeghen , arxiv:1303.2540 [ hep - ph ] m freytsis , g. ovanesyan and j. thaler , jhep * 1001 * ( 2010 ) 111 b. wojtsekhowski , d. nikolenko and i. racheck , arxiv:1207.5089 ( 2012 ) s. n. gninienko , arxiv:1308.6521 [ hep - ph ] d. alesini et al . ,
lnf-06/33(ir ) ( 2006 ) m. e. biagini et al .
, infn-13 - 13(clab ) ( 2013 ) y. ohnishi et al . , ptep * 2013 * ( 2013 ) 03a011 p. fayet , nucl .
b * 347 * ( 1990 ) 743 b. holdom , phys .
b * 166 * ( 1986 ) 196 y. kahn and j. thalerm phys .
d * 86 * ( 2012 ) 115012 p. de niverville , m. pospelov and a. ritz , phys .
d * 84 * ( 2011 ) 075020 e. izaguirre , g. krnjaic , p. schuster and n. toro , arxiv:1307.6554 [ hep - ph ] r. essig et al .
, arxiv:1309.5084 [ hep - ph ] ( 2013 ) s. andreas , desy - thesis-2013 - 024 ( 2013 ) h. k. dreiner , j. f. fortin , c. hanhart and l. ubaldi arxiv:1310.3826 [ hep - ph ] d. babusci et al . ,
lnf-10/17(p ) ( 2010 ) f. archilli et al .
, lnf-10/3(p ) ( 2010 ) p. coyle and o. schnider , c. r. physique * 3 * ( 2002 ) 1143 j. beringer et al .
, [ particle data group ] phys .
d * 86 * ( 2012 ) 010001 j. fu , h. b. li , x. qin and m. z. yang , mod .
lett a * 27 * ( 2012 ) 1250223 | the existence of a new , photon - like , massive particle , the @xmath0 or dark photon , is postulated in several extensions of the standard model .
these models are often advocated to explain some recent puzzling astrophysical observations , as well as to solve the unsofar unexplained deviation between the measured and calculated values of the muon anomaly .
dark photons can be produced at e@xmath1e@xmath2 colliders both in continuum events and in vector meson transitions and can eventually decay into an electron - positron pair . for a proper choice of the parameters of the theory , a @xmath0 can have a relatively long lifetime and can therefore be observed as an e@xmath1e@xmath2 vertex well separated by the primary interaction point .
this case is discussed in reference to very high luminosity e@xmath1e@xmath2 colliders either in construction or under study in several laboratories in the world .
it is shown that a search strategy based on the detection of displaced vertices can be in principle very effective in covering a rather wide and to date unexplored region of the theoretical parameters space |
an interesting generalization of the boltzmann - gibbs entropy form has been recently proposed by tsallis @xcite , for a recent review see @xcite , for a full bibliography see @xcite .
this new entropy , that we discuss below , posseses the usual properties of positivity , concavity and irreversibility and generalize the additivity in a non - extensive way . in the later years
, a lot of work has been done in order to show either that many relationships involving energy and entropy in the usual scheme holds in the new one or in order to search for suitable generalizations for them .
we should mention , among others , the work of curado and tsallis @xcite , that shows that the whole mathematical structure of thermodynamics legendre
transform based becomes invariant under a change from extensive to non - extensive statistics ( nes ) .
this , in fact , is a property recently proved to be valid for any entropic functional form @xcite , making then important to have a partition function within the formalism .
boltzmann - gibbs thermostatics constitutes very complete and powerful techniques in situations whenever thermodynamics extensivity holds . but this formalism provide divergent partition , energy and entropy functions whenever the physical system includes long range forces or long memory effects . in this regard , the connection with quantum mechanics and with information theory has been stablished @xcite and further applications to self - gravitating objects @xcite and astrophysics @xcite , to lvy flights @xcite , turbulence of fluids @xcite and to the problem of solar neutrinos @xcite have also been studied . in general , systems which present a fractal like or unconventional structure in their space - time description or in their phase space evolution could develop hardly tractable mathematical problems within the standard formalism @xcite . very recently , grigera @xcite derived a molecular dynamic test for systems of long range interacting particles with potentials of the form @xmath1 .
he found that the potential energy per particle do not converge for some cases of @xmath2 and sizes of the system , whereas non - extensive quantities do .
also see refs .
@xcite .
non - extensive thermostatistics was also used to study the cosmic blackbody radiation @xcite .
this was the setting in which powerful bounds upon the non - extensive parameter were obtained .
we shall review some of these results in our final section . a new analysis of nes blackbody cosmological radiation showed that , independently of the degree of non - extensivity , the temperature of the universe varies as the inverse of the scale factor of a flat friedmann - robertson - walker metric @xcite .
this result suggested cosmology of the early universe as a very profitable arena for searching observational consequences of nes .
as far as we know , the first of such studies were presented in @xcite , where a stringent bound on non - extensivity , of comparable order of magnitude as those obtained via the cosmic blackbody radiation , was derived .
there , it was analized how a first order variation in the non - extensive parameter would affect the amount of primordial helium production , a quantity which can be compared with observational results , via a rather easy and straightforward approximate technique .
we now turn to the issue of early universe cosmology and particularly , to the computation of primordial neutron to baryon ratio in the framework of nes .
we shall also analyze cosmological situations of great importance , such as pair annihilation , temperature relationships , and comoving quantities .
we thus refine the previous study and enhance it by a sharpening of the model .
although we shall present a very brief review of the main formulae of nes ; in first approach , this work should be understood as the search of deviations from the standard model of the early universe provided by a different statistical description .
what we want to use is a direct analytical computation of helium production , following the work of bernstein , brown and feimberg @xcite for the standard model , to allow for an approximate study .
it is a matter of fact , that even when nucleosynthesis processes may be as complicated as one can afford , some simple assumptions makes results agree , within a typical few percent , with those generated by a numerical code , thus allowing for the isolation of the essential physics .
that is exactly what it is needed here in order to study how , for instance , a different formula for the number of particles influence the evolution of the early universe and how can this be used to extract testeable predictions .
nucleosynthesis has given a strong basis to do such a thing when alternatives theories of gravitation are concerned , see for instance @xcite , and also for particle physics @xcite and we hope to show , the same is applicable for alternative statistics .
the rest of the work is organized as follows .
firstly , some useful concepts and formulae of nes are presented .
afterwards , the analytical computation of the neutron to baryon ratio is done , and finally we study the necessary approach to handle with neutron decay corrections . in the final section , we can use our previous results to estimate the mass fraction of primordial helium produced in the context of nes .
although we shall not have at that moment every ingredient needed to do such a thing with great precision , the bound which we obtain hardly constrain the degree of non - extensivity in the very early universe and is already comparable with previous results .
finally , we state our conclusions and mention other recently obtained related bounds .
let us first recall some formulae of nonextensive statistic ( nes ) , in tsallis s approach .
the formalism starts by postulating @xcite * postulate 1*.- _ the entropy of a system that can be found with probability @xmath3 in any of @xmath4 different microstates @xmath5 is given by _ s_q = k _ i=1^w [ p_i - p_i^q]= k ( 1-_i=1^w p_i ^q ) , [ sq ] .
we have a different statistics for every possible @xmath6value .
in ( [ sq ] ) we have used , of course , that _ i p_i = 1 . in general ,
[ sq2 ] s_q = k .
* postulate 2*.- _ an experimental measurement of an observable @xmath7 , whose expectation value in microstate @xmath5 is @xmath8 , yields the @xmath0- expectation value _ ( generalized expectation value ( gev ) ) [ gev ] < a>_q=_i=1^wp_i^q a_i = tr ^q a , @xmath7 .
these two statements have the rank of _ axioms_. as such , their validity is to be decided exclusively by the conclusions to which they lead , and ultimately by comparison with experiment .
the entropy given in ( [ sq ] ) is non - negative and reproduces the boltzmann - gibbs one ( @xmath9 ) in the limit @xmath10 .
in fact , note that ( [ sq ] ) can be obtained from the boltzmann - gibbs expression by writing @xmath11 and then dropping the limit operator .
it satisfies the pseudoadditivity property which states that if @xmath7 and @xmath12 are two independent systems ( @xmath13 ) then , = + + ( 1-q ) , which shows that @xmath14 is a measure of the degree of non - extensivity of the system .
the optimization of @xmath15 given by ( [ sq2 ] ) together with the constraints @xmath16 and @xmath17 yield to the canonical ensemble equilibrium distribution @xcite , @xmath18^\frac{1}{1-q},\ ] ] and to the generalized partition function , @xmath19^\frac{1}{1-q}.\ ] ] here , as usual , @xmath20 and @xmath21 is the hamiltonian of the system .
we shall focus in the @xmath22 limit , in which a first order expansion allows analytical computations .
the expression of a general mean value of an operator was computed in this limit by tsallis , sa barreto and loh @xcite . when applied to particle number operators ,
the result is : @xmath23 \right],\ ] ] where @xmath24 stands for @xmath25 ( @xmath26 is the energy of a simple particle ) and the symbol @xmath27 means _ to be computed in boltzmann - gibbs statistics_. with the standard values of @xmath28 and @xmath29 , for fermions and bosons , we obtain the main corrections ( the second terms in equation ( 8) ) , @xmath30 , as : @xmath31 @xmath32 when @xmath24 is such that @xmath33 , then , @xmath34 ; while for @xmath24 big enough we get @xmath35 .
we turn now our attention to the description of the model for the evolution of the neutron abundance as the universe evolves .
as stated in the introduction , we shall follow the leading ideas of bernstein , brown and feimberg @xcite .
we shall denote by @xmath36 the rate for the weak processes to convert protons into neutrons and by @xmath37 the rate for the reverse ones .
@xmath38 will be , as usual , the number of neutrons to the total number of baryons .
for it , a kinetic equation may be built , @xmath39 the solution to the previous equation reads : @xmath40 with the integrating factor given by , @xmath41 where , @xmath42 note that the form of this solution does not depend on the statistical mechanics used .
however , changing the statistical mechanics entails modifications in the rates , and this will make the solution differs from the standard .
expression ( [ x2 ] ) is simplified by setting the initial value @xmath43 .
we expect that @xmath44 and @xmath45 be very large at early times and high temperatures
. then , the integration factor @xmath46 will be very small for times order @xmath47 and therefore the term @xmath48 can be omitted .
besides , we also expect that due to the large interaction rates , the integral in the first term will be not sensitive to the change of @xmath49 by 0 . with this choice , we now have , @xmath50 finally , we may note that @xmath51 and integrate by parts to obtain @xmath52 to explicitly compute expression ( [ x4 ] ) , we need to know the functional form for the rates .
let us first consider the rate @xmath53 .
this is the sum of the rates of three individual processes , @xmath54 which are given by @xcite , @xmath55 @xmath56 @xmath57 where @xmath7 is an overall effective constant fixed by the experimental value of @xmath58 , @xmath59 stands for the magnitudes of the neutrino and electron momentum and @xmath60 for their energies .
the recoil energy of the nucleons may be neglected , and this enable us to write the energy conservation equation @xmath61 for ( [ r1 ] ) and @xmath62 for ( [ r2 ] ) .
these equations must be used in order to explicitly compute the integrals . in ( [ r3 ] ) , @xmath63 , with @xmath64 and this gives the upper limit in the integration range . finally , @xmath65 are distributions functions and @xmath66 are blocking pauli factors . in the usual scheme , @xmath65 are given by fermi distributions , whereas in our case , by the @xmath67 of the previous section .
in general , the electron and neutrino temperatures , @xmath68 and @xmath69 , may differ because at the end of the freezing out period , electrons and positrons annihilate , heating only the photons and mantaining with them thermal equilibrium .
the magnitude of this difference will be discussed below .
as we shall see , this difference will be of a few percent during the period of consideration and our first approximations , following bernstein et .
al . , is to set all temperatures equal , @xmath70 .
when working in the boltzmann - gibbs framework this unique assumption ensure that the rates for reverse reactions , such as @xmath71 , obey the principle of detailed balance ( pdb ) , that we analize below in this general setting .
let us consider , in the boltzmann - gibbs ( bg ) statistics and as an example , the reverse rate for ( [ r1 ] ) , that is , @xmath72 using energy conservation and the fact that @xmath73 we have , neglecting the recoil energy of the nucleon , @xmath74 , and thus : @xmath75 as @xmath76 are fermi - dirac distributions , it is easily seen that @xmath77 and that @xmath78 .
replacing this into ( [ reverse2 ] ) and recalling that @xmath79 , we get @xmath80 which is the expression of the pdb .
when nes is concerned , and correspondingly change from the fermi - dirac distributions to the @xmath81 ones , the pdb is no longer valid . in this framework
, it may be seen that @xmath82 where the correction is = ( 1-n_e , bg ) c _ ( + ) - + n_,bg c_e ( + ) and the @xmath83 factors are as in ( [ correc - fer ] ) . in order to get a simplified picture , and following the bg case , we further approximate the model -exactly as was made in the standard case- by assuming that during the period of freezing , the temperature @xmath84 is low in comparison of the typical energies @xmath85 that contributes in the integrals for the rates .
hence , we may replace the fermi - dirac distributions by boltzmann weights ( i.e. @xmath86 ) and consistently neglect the the pauli blockings ( i.e. @xmath87 ) . even in the case of nonextensive statistics ,
pauli blockings corrections are @xmath88 , which are neglectable for @xmath89 and a first order deviation from @xmath90 .
in addition , doing this allows for the analytical study that follows and which is the objective of the work . in this @xmath89 regime , the one which we are going to use thorough , we obtain @xmath91 , and we recover for the reverse rates the expression given in ( [ pdb ] ) of the pdb . before we finally compute the integrals for the rates we neglect the electron mass in equations ( [ r1 ] ) and ( [ r2 ] ) in comparison to the energies @xmath60 . when this is made , the two rates become identical , exactly as they were in bg statistics .
then we obtain : @xmath92 where , @xmath93 concerning the rate of free decay of neutrons , we note that nes does not present any change within the context of previous approximations .
that is due to the disappearence of all distributions functions from ( [ r3 ] ) . in that equation
we can not neglect the mass of the electron and the standard result holds , @xmath94 this enable to eliminate @xmath7 in favor of the measured quantity @xmath95 , @xmath96 all along in this section , we shall neglect the free neutron decay rate when computing the total @xmath44 .
thus , we have @xmath97 finally , by using the variable @xmath98 , we get @xmath99,\ ] ] for the neutron - proton rate and @xmath100 for the reverse reaction . in equation ( [ x4 ] ) ,
we change now variables to the scaled temperature @xmath101 and obtain , @xmath102 the integrating factor now becomes , @xmath103 to evaluate the jacobian @xmath104 , we need to recall that the scale factor of the universe , @xmath105 , in a friedmann - robertson - walker metric , behaves as @xmath106 , independently of the statistics @xcite .
therefore , @xmath107 , and the right hand side is given by einstein equations : @xmath108 here @xmath109 is the energy density for relativistic particles in nes .
this may be easily computed , if the distributions functions are known , by @xmath110 where @xmath111 is the degeneracy factor .
the result is @xmath112 where @xmath113 . at high enough temperatures the energy density of the universe
is essentially dominated by @xmath114 , @xmath115 , @xmath116 and @xmath117 s .
interactions among these particles keep all them nearly the same temperature .
accordingly , we set , @xmath118 , @xmath119 and @xmath120 . with these values
we get , @xmath121 with this in mind , and using @xmath122 , we obtain @xmath123^{\frac 12 } t^3\ ] ] which yields , to first order in @xmath124 to , @xmath125.\ ] ] we shall call the constants , @xmath126 and @xmath127 with all above , the integrating factor becomes , @xmath128 where , @xmath129 \times \nonumber \\ ( 1+\exp{(-y)}).\end{aligned}\ ] ] this integrates to give , @xmath130 where @xmath131 stands for the exponential integral . for @xmath132 .
] introducing , @xmath133 the neutron abundance ratio reads : @xmath134}.\ ] ] the exact previous result may be expanded again to first order in @xmath124 .
this gives , x(y)=x_standard + ( 1-q ) ( bb(y)+bca(y ) ) + _
o^y dy^ x_eq(y^)^2 ) + -(1-q ) _ o^y dy^ x_eq(y^)^2 ( bb(y^)+bca(y^ ) ) , where , @xmath135 and the functions @xmath7 and @xmath12 are : @xmath136 @xmath137 we can numerically compute the integrals that appears here . using explicit values for the constants @xmath138 and @xmath139 with the mean life of the neutron given by @xmath140 we get the curves of fig .
, it is shown the standard value of the neutron to baryon ratio and the correction due to nes . in fig .
2 . are shown different explicit computations for a range of the @xmath0 parameter .
finally , it may be seen that @xmath141 asymptotes to @xmath142 for very low temperatures .
we briefly assess here the possible dependence on the number of neutrino types . to do this we note that such a variation would introduce changes in the coefficients @xmath143 and @xmath144 quoted above .
for three neutrino families , @xmath120 and the addition of one more family change it in an amount @xmath145 .
then , the standard case coefficient @xmath143 is affected as , @xmath146 and the non - extensive related as , @xmath147 introducing these new values in ( [ kk ] ) and making all numerical computations again , it is possible to obtain the corrections due to a different number of neutrino families . in the sake of conciness
, we shall explicitly skip these computations .
instead we note that , while in the standard case , the addition of one more neutrino family would involve an increment of the energy density of the universe and a speed up in its expansion , in the general case , the correction will depend on the sign of the factor @xmath124 .
we have already solved the evolution of the neutron abundance neglecting the neutron decay . as in the standard formalism , it is useful to change notation by using an overbar for the result just obtained , @xmath148 . including the effects of the free neutron decay in the rate equation we would get , @xmath149 because @xmath150 does not vary much during the period in which neutrons decay . in the capture time @xmath151 , when the temperature drops somewhat below the deuteron binding energy ( @xmath152 ) , the neutrons are captured in deuterons .
then , deuterons collide and almost all neutrons present at @xmath151 are converted into @xmath153 . inserting this time in ( [ ult ] ) and using the asymptotic value of @xmath154
just derived would yields , thus , half of the mass fraction of helium produced in the early universe . in order to obtain a precise value of @xmath151 or @xmath155 for the nonextensive result
, we have to analize the set of the basic reactions : [ reac4 ] n + p d + [ reac5 ] d+ d p + t [ reac6 ] t+ d ^4he + n. doing this in the standard case @xcite yields a value of @xmath156 such as the correction @xmath157 .
although the analysis of ( [ reac4],[reac5],[reac6 ] ) in a nes framework is beyond the scope of the present work , we shall establish here the basis on which this could be done in the future .
for this , we shall study the process of annihilation of pairs @xmath158 .
we know that , when the temperature of the universe is low compared with the electron mass , electrons and positrons annihilate heating the photons and , morover , changing the effective quantity of degrees of freedom
. then , one of the assumptions of the previous section , i.e. @xmath70 , is no longer valid and @xmath159 will be different from the decoupled neutrino temperature . we shall need the connection between these temperatures . to compute it ,
we begin by noting that , at these times , the number of neutrinos in a comoving volume is fixed , independently of the statistics .
effectively , after the moment in which the collision rate for neutrinos is lower than the expansion of the universe , @xmath160 , they can not be neither created nor destroyed . as the number of neutrinos is given by , n_(q)= _ 0^e^2 <
n>_q = ( 3 ) g_t^3 + ( q-1 ) 5.50 t^3 the relationship , n_(t ) r(t)^3 = const . , or , equivalently , [ law ] t_(t ) r(t ) = const . , continues holding .
> from the full set of einstein field equations , it may be established that d((q ) r^3 ) = -p(q ) d(r^3 ) , this may also be written as , [ rel1 ] = ( r^3 ) .
but we may compute @xmath161 , as was done with the energy density , obtaining @xmath162 for the relativistic gas . using this and ( [ rel1 ] )
, we get [ rel2 ] = 0 , as can be cheked by direct differentiation .
this means , that in general case , @xmath163 is a conserved quantity in a comoving volume .
when @xmath90 the previous expression stands for the entropy of the system .
we may now consider that pair annihilation occurs at a temperature @xmath84 , and then indicate with symbols @xmath164 and @xmath165 , quantities before and after @xmath84 . from all above , we have [ rel3 ] = . here
@xmath166 is as in ( [ rho ] ) with the corresponding values of @xmath167 and @xmath168 . due to the removal of pairs ,
we have that @xmath169
. we may now expand ( [ rel3 ] ) , to first order in @xmath124 , obtaning the relationship , t_+^3 = t_-^3 ( ) , where , = ( 40.02 g_b,- + 34.70 g_f,- ) - ( 40.02 g_b,+ + 34.70 g_f,+ ) . until the moment of @xmath170 annihilation , the gas composed by @xmath170and @xmath117 follows a law of identical form that ( [ law ] ) , as such temperatures , being @xmath171 .
but after the annihilation , where only @xmath172 remain , we have [ rel4 ] t_= t _ ( ) ^13 , where the corresponding values @xmath173 and @xmath174 were used . to finally compute the capture time @xmath156
, we use that = - ( ( q ) ) ^ 12 , from where [ t ] t= _ t_^ ( ) ^ 12 follows .
here @xmath175 adquires its low energy form _ 0=+ , where @xmath176 and @xmath177 . recalling the relationship ( [ rel4 ] )
we get , to first order in @xmath124 , _
0=g_eff t_^4 + ( q-1 ) t_^4 14.84 where @xmath178 . in the neighbourhood of the time of the neutron capture , an approximate evaluation of ( [ t ] ) is in order . using the expression for @xmath179 just derived , we get t= ( ) ^12 ( 1+(1-q ) 1.74 ) + t_0 , where @xmath49 is an additional integration constant and the corrective number 1.74 arise from the first order evaluation of @xmath180 .
recalling then the relationship between the temperatures , we may express [ t2 ] t= ( ) ^12 ( 1+(1-q ) 1.76 ) + t_0 , we may neglect @xmath49 in a first approximation or obtain it from a perturbation analysis of ( [ t ] ) , @xcite .
now , a comprehensive analysis of the reactions ( [ reac4],[reac5],[reac6 ] ) would yield a value for @xmath159 , for which almost all nucleons are converted into helium .
using then ( [ t2 ] ) , the evaluation of @xmath156 follows and thus , the mass fraction of primordial helium . in passing through nes , we have to note that a fundamental tool in this analysis , i.e. saha formulae , is modified in a non - trivial way .
this entails modifications in the abundance fractions that may in turn change their ratios , which are the magnitudes on which the set depends .
so , although phenomenological fits for the rates may be considered , we acknowledge here that further research must be done to analytically solve for the reactions , in the case this be ultimately possible . this analysis is far out from the present work and we shall not discuss this problem in deeper detail .
we have presented a comprehensive study of the primordial neutron - baryon ratio in the context of non - extensive statistics .
this work , thus , extends previous advances on the helium nucleosynthesis production problem @xcite and is the result of a detailed analysis of the weak interaction rate and the energy densities corrections at first order in the non - extensive parameter @xmath124 .
we have also studied some important cosmological scenarios such as the possible existence of conserved comoving quantities , temperatures relationships and pair annihilation , together with pointing out where further research is needed .
these other results will be used elsewhere to put new stringent bounds on the non - extensivity of the universe by using high precision satellite cosmological measurements of the cosmic microwave background and new galaxy surveys . even without getting a precise analytical value for the helium synthesis , it is possible to obtain a first insight in the problem by considering it as exactly twice of the neutron - baryon asymptotic value , weighted by the free neutron decay : y_p^theoretical ~2 x(t ) where @xmath181 is given by equation ( 56 ) .
but as mentioned above , in order to put an exact bound we would need the value of @xmath155 .
however , even considering that @xmath159 is the same as in the standard formalism ( @xmath182 ) we could use ( [ t2 ] ) to get the correction in @xmath156 . using afterwards the first order expansion of ( [ ult ] ) ( which contains equation ( 53 ) ) we would get the stringent bound : [ bou ] |1-q| < 2.6 10 ^ -4 . in order to get this bound ,
we have imposed that the non - extensive corrections ( coming from equations ( 75 ) and ( 53 ) and the theoretical value of @xmath183 as explained above ) be less than the observational error in the primordial helium production , which is doubled so as to neglect the difference between the central values of the analytical and observational magnitudes @xcite , i.e. : y_p^observational=0.23 0.02 .
this abundance is a very safe estimate of the primordial helium production , which in fact is a very complex observable magnitude , see for instance the recent review by schramm and turner @xcite for details .
besides , it encodes other possible uncertainties such as the value of the cosmological constant or the exact number of neutrino species . the bound ( [ bou ] )
is comparable with that already obtained in ref . @xcite .
although it should be recalled that such derivation was different from this and the method followed skip the uncertainties of the neutron capture time here analyzed .
in addition , the weak interaction rates were only approximately treated in @xcite while here they were studied in full detail . in a recent letter @xcite
, it was suggested that the universe as a whole could behave with a non - extensive parameter roughly equal to 0.54 .
such a value seems to be too high as to resist early universe tests as the one presented here .
therefore , this bound , obtained in the simplest situation so as to admit an analytical solution , pertains to those imposed in the context of early cosmological models , and from this point of view , enlarge the problems that non - extensivity would have as a correct description of the physical universe thorough all eras of cosmic evolution . very recently
, other works have also been devoted to study possible deviation from the boltzmann - gibbs statistics in cosmological and other situations . among them , and apart from ref .
@xcite already commented , we should note the following ones : * plastino , plastino and vucetich @xcite studied the value of the stefan - boltzmann constant in the new framework .
a comparison with the experimental value leads to @xmath184 .
they also used data from the firas experiment on the cobe satellite concening the microwave background radiation , from there they obtain @xmath185 . * tsallis , sa barreto and loh @xcite also generalized the planck radiation law and used satellite data from cobe to constrain the non - extensive parameter .
they found @xmath186 .
* tirnakli , bykkili and demirhan @xcite reworked the generalized quantal distributions functions and also obtain a similar bound : @xmath187 .
the authors acknowledge partial support from conicet and valuable discussions with a. plastino and u. tirnakli during the course of this research .
useful insights from an anonymous referee helped to improve the presentation and are gratefully acknowledged .
d.f.t . was supported by a chevening scholarship of the british council . | we study the primordial neutron to baryon ratio in a cosmological expanding background within the context of non - extensive statistics , in a fully analytical approach .
first order corrections to the weak interaction rates and energy densities in the early universe are obtained and their consequences on the helium synthesis are analized .
we find that , if the nucleosynthesis scenario should be in agreement with observation , this entails a very restrictive bound upon @xmath0 , the non - extensive parameter .
we also study some cosmological interesting situations such as the moment of pair annihilation , conserved comoving quantities and derive for the nonextensive statistics , temperature relationships , before and after the removal of pairs . _
pacs number(s ) : _ 05.30.-d , 05.70.-a , 26.35.+c , 98.80.ft |
since the earliest dynamical models of strongly irradiated exoplanetary atmospheres were published @xcite , it has been apparent that the expected large temperature gradients and the resulting high wind speeds could have such a large impact on these planets appearance that the effects could be observable even at a distance of dozens of parsecs . as the models have grown increasingly sophisticated over the past six years
, research has progressed along two distinct lines .
@xcite , @xcite , @xcite and @xcite have produced three - dimensional models , in an effort to simulate the greatest possible range of relevant physical processes .
because of the high computational costs of such models , they have been run at comparatively low resolution .
in contrast to this approach , @xcite , @xcite , @xcite , and @xcite have chosen to employ two - dimensional models which can be run at higher resolutions .
on highly irradiated planets , the radiative zone is believed to extend deep into the atmosphere , to a pressure depth of hundreds of bars @xcite .
the flow is therefore expected to be strongly stratified , with vertical motion comparatively unimportant .
the assumption inherent in a two - dimensional model is that the motion at small scales which can be captured due to the finer grid spacing is more important than the vertical flow which a two - dimensional model must neglect .
nevertheless , a serious concern for two - dimensional models is the possible existence of crucial three - dimensional processes which do not require strong vertical motion in order to become significant .
furthermore , previous attempts to develop two - dimensional models have been hampered by questions regarding the validity of the physics involved .
@xcite employ a shallow - water model which is at best a first - order approximation to realistic atmospheric dynamics . while the barotropic - equivalent model used by @xcite , formally very similar to the shallow - water equations , is on a firmer physical footing ,
there is some indication that the model becomes numerically unstable at high wind speeds @xcite . in an attempt to achieve greater realism than is possible using a shallow - water model
, @xcite developed a model employing fully - compressible two - dimensional hydrodynamics . as we will show in 2
, however , such models produce features which necessarily violate hydrostatic balance . in this paper , then , we present a two - dimensional model which must maintain hydrostatic equilibrium . it is our hope that this new model will be able to approach the rigor of a three - dimensional model , while maintaining the adaptability and speed of a two - dimensional model .
while these modeling efforts have been underway since 2002 , the ability to constrain these models using observations is a more recent development .
of particular interest , @xcite have produced a map of the planet s longitudinal temperature variation based on their _ spitzer _ observations of its flux in the 8-@xmath0 m band . from the depth of the secondary eclipse
, they find a hemispherically - averaged day - side brightness temperature of @xmath2 k , while they estimate a cooler hemispherically - averaged night - side brightness temperature of @xmath3 k , based on the flux curve .
interestingly , the hottest part of the planet is _ not _ directly beneath the star : the temperature maximum is offset some @xmath4 east of the substellar point . perhaps even more interestingly
, the temperature minimum is offset @xmath4 _ west _ of the antistellar point .
it is clear , then , the temperature distribution is strongly influenced by planetary winds ; at first glance it would appear that the flow is eastward on the day - side , while westward on the night - side .
to date , models have not reproduced this flow geometry @xcite .
additionally , these observations do not seem to be generally applicable to hot jupiters .
@xcite , for example , observed the 24-@xmath0 m flux of the giant exoplanet @xmath5 andromedae b , finding a much larger temperature contrast between the illuminated and dark hemispheres .
additionally , they found a slight _ westward _ displacement of the hot spot from the substellar point , although their observations did not exclude the possibility that there was no offset at all .
the mid - infrared observations of hd 179949 b by @xcite also appear to be consistent with zero phase offset .
it is not clear to what extent these discrepancies result from different dynamics on the planets themselves , and to what extent they are caused by the different pressure depths under observation .
in other respects , hd 189733 b remains a fairly typical hot jupiter ; its period , however , is shorter than most , with @xmath6 d. the transit has allowed determinations of several other parameters : @xmath7 , @xmath8 @xcite , @xmath9 , and @xmath10 @xcite .
the orbit is assumed to be circular .
hd 189733 b has not received the attention from modelers which has been bestowed on its more famous cousin , hd 209458 b. nevertheless , the well - resolved flux curve produced by @xcite has made it a most interesting target for simulation , and @xcite apply their atmospheric model to hd 189733 b , with mixed success .
they recover the correct eastward phase offset for the hot spot , but are unable to produce a flow pattern which causes the observed westward offset for the cold spot from the antistellar point .
their flow patterns in the upper atmosphere are in general supersonic , with very low pressures characterized by longitudinal and latitudinal flow from the day side towards the night side , while deeper layers are dominated by a supersonic eastward jet at the equator . interestingly , despite the high wind speeds ( @xmath11 km / s ) and the resulting large wind shear , no turbulence develops in their simulations .
as we shall see , this does not match the results of the two - dimensional model presented in this paper .
it is possible that their relatively low horizontal resolution ( 144x90 ) combined with the finite - difference differentiation employed by the aries / geos dynamical core used in their model conspire to produce sufficient numerical dissipation to prevent the development of turbulence ; it is also possible ( and possibly more likely ) that the development of turbulence is prevented by three - dimensional effects which our two - dimensional model is unable to capture .
an analytical study to constrain the conditions under which turbulence is expected to arise would surely be profitable ; however , this , and indeed , the more general question of the conditions necessary for planetary - scale turbulent flow , must remain a topic for future investigation . in any case , it is clear that the @xcite time - series is of sufficiently high quality to provide an excellent benchmark for both existing and future simulations of exoplanetary atmospheres .
the aries / geos core used in @xcite and @xcite has been applied to the terrestrial atmospheres of earth and mars @xcite , and the equivalent barotropic formulation of @xcite also used in @xcite has enjoyed some success in reproducing the primary features of jupiter s dynamics .
however , the conditions on many extrasolar planets are so utterly unlike anything seen in our solar system that it is advisable to include data from exoplanet observations when testing the models .
this paper is organized as follows : in 2 , we derive in some detail the hydrodynamical core of our model , as well as providing a treatment of the radiative forcing scheme . in 3 , we apply our model to the atmosphere of hd 189733 b , comparing our results to those of @xcite and to the data obtained by @xcite .
we conclude in 4 .
we begin with the equations of motion in three dimensions for an irradiated , hydrostatically - balanced ideal gas in a rotating reference frame , with an isobaric vertical coordinate : @xcite @xmath12 where @xmath13 is the geopotential at constant pressure , @xmath14 is the `` velocity '' in pressure coordinates ( while @xmath15 does not have units of velocity , the quantity @xmath16 does have the correct units of advection ) , and @xmath17 for an ideal diatomic gas .
it is also important to note that @xmath18 describes _ only _ the horizontal components of the flow ; vertical motion , associated with the upward or downward motion of an isobar , necessary to maintain hydrostatic equilibrium , is encapsulated in @xmath15 . to adapt this three - dimensional system to two dimensions ,
it is necessary to assume a particular vertical structure for @xmath19 and @xmath18 .
we assume that the temperature does not vary with pressure ; this is a reasonable first approximation , and any temperature variation is not expected to cause significant deviations in the geopotential . to see this ,
consider an atmosphere where the temperature can vary ( weakly ) with pressure : @xmath20 .
( this is , of course , a gross simplification , as a temperature that was a function solely of pressure would not give rise to any interesting dynamics if one is looking at layers of constant pressure ! )
the geopotential is found by integrating equation [ p31 ] from some pressure @xmath21 at the boundary to @xmath22 , the pressure of the single layer being simulated in our two - dimensional model .
we further assume that the isobar at @xmath21 is at constant geopotential , so that we can set @xmath23 .
then @xmath24 in the case that @xmath25 , as it would if the layer under consideration is in the upper atmosphere , then the logarithmic term accounts for the majority of the variation in @xmath26 , and we can simply take @xmath27 to good approximation . at first , we make no assumptions about the variation of flow velocity with pressure , other than separability : @xmath28 .
if we define @xmath29 according to @xmath30 , then we must have @xmath31 .
for the time being , we place no other restrictions on @xmath32 . with the assumption of a suitable boundary condition , it is now possible to determine @xmath15 . as a boundary condition
, we choose @xmath33 .
this gives @xmath34 where @xmath35 defining @xmath36 we can substitute into equations [ p33 ] and [ p34 ] to obtain a general form of the two - dimensional hydrostatically - balanced governing equations : @xmath37 these equations describe atmospheric motion at a single pressure depth .
note that isobars are in general not material surfaces ; therefore , the equations do not describe the evolution of any particular set of particles in the atmosphere .
a result of this is that mass is not generally conserved , since matter is free to flow into this layer from other layers and _ vice versa_. in principle , this makes the derivation of @xmath38 difficult , since the heating term is obtained by applying the first law of thermodynamics to a single parcel of gas .
however , the assumption that @xmath39 mitigates this concern , since the material flowing into a particular zone from above or from below is expected to be at approximately the same temperature .
it is worth noting that the vertical temperature variations expected in a real atmosphere could cause as much or more error in our model due to these energy - balance concerns than due to their direct influence over the geopotential height of an isobar .
these concerns , however , are difficult to address in the context of the single - layer approach presented here . in order to apply this model to a specific atmosphere
, it is necessary to choose a specific form for @xmath40 .
if the wind shear is not large , a reasonable approximation is to take @xmath41 , so that @xmath42 . in this case
, @xmath43 substitution into equations [ k1 ] and [ k2 ] yields @xmath44 at this point , there still seems to be little insight as to physically reasonable values for @xmath45 .
however , examination of the pressure dependence of equation [ p2v0 ] will allow for an _ ad hoc _ determination of the velocity shear , which should be sufficient for current purposes .
let us assume that the scale of the velocity varies as @xmath46 ; this is tantamount to assuming that the dominant accelerating force is the temperature gradient term .
then the velocity can be written @xmath47 here , @xmath48 is the velocity at some reference pressure @xmath49 , which we will momentarily identify with the @xmath22 that appears in equations [ p2v0 ] , [ k12 ] , and [ k22 ] . then @xmath50 a taylor expansion of equation [ vp ] about @xmath51 therefore yields @xmath52 rearranging , @xmath53 then we can identify @xmath54 and @xmath55 we are interested in the value of @xmath56 at the layer under consideration that is , at @xmath57 in the taylor expansion . making this substitution in equation [ feq ] yields the desired value of @xmath56 : @xmath58 where we have dropped the subscript from @xmath49 . defining @xmath59 and @xmath60 ,
we can write the final form of the two - dimensional hydrostatically - balanced governing equations for an atmospheric layer at constant pressure : @xmath61 these equations govern the time - evolution of @xmath19 and @xmath18 , and , when augmented by an appropriate treatment of the radiative forcing term @xmath38 , are suitable for atmospheric simulation . to derive the radiative forcing term , we assume that the layer under consideration extends from some pressure @xmath62 which corresponds to the infrared photosphere , to arbitrarily low pressures .
we assume that this layer absorbs some fraction @xmath63 of incident stellar flux .
while @xmath63 is effectively a free parameter in the model , in principle it can be determined based on atmospheric chemistry and appropriate @xmath22-@xmath19 profiles .
the remaining portion of incident radiation is absorbed deeper in the atmosphere , where it is reradiated through thermal emission .
this is modeled by an isothermal layer which always remains at some constant @xmath64 , where @xmath64 is a function of both the amount of penetrating solar radiation and internal heating due to tidal dissipation , gravitational contraction , etc .
the assumption here is that below the infrared photosphere , advective processes quickly redistribute heat and destroy any temperature variation .
this is , of course , a considerable oversimplification of the real atmosphere , but it will have to suffice for our purposes .
it is also assumed that all of the thermal emission from this interior layer is absorbed by the upper layer being simulated .
should it prove necessary to relax this second assumption for example , to model extremely low pressure depths the adjustments to accomplish this are straightforward . in this model ,
the energy input from the absorption of incident sunlight is simply @xmath65 .
the energy input due to absorption of thermal emission from deeper in the atmosphere is simply @xmath66 , since the outgoing long - wave radiation from the internal layer is completely absorbed .
@xmath64 is calculated by time - averaged energy balance , wherein the stellar flux penetrating to the deeper layer is averaged over the the course of an orbit .
the time - averaged penetrating flux is @xmath67 where the additional factor of @xmath68 in the denominator arises from the time - averaging , and represents the disproportionate contribution to the heating during the periastron passage .
in addition to the heating of the internal layer produced by stellar irradiation penetrating the upper layer , the planet itself can provide a considerable contribution to the energy budget , due to tidal heating , gravitational contraction , and possibly other effects .
these are encapsulated in the @xmath69 parameter , which is related to the planet s intrinsic luminosity by @xmath70 . with these two sources of heating
, energy balance requires that the inner layer maintain a steady - state temperature @xmath71 energy balance for the upper layer therefore requires that @xmath72 where @xmath73 is the column density of the upper layer and @xmath22 is the pressure at the @xmath74 of the upper layer .
rearranging and explicitly writing out the contribution from @xmath75 , we have @xmath76 this model can not , of course , provide either the realism or the physical insight that is achievable within the framework of genuine radiative transfer , which explicitly accounts for both the vertical structure and the chemistry of the atmosphere .
however , it offers significantly greater accuracy than newtonian heating @xcite , particularly in the case that the temperature strays far from its equilibrium value . unlike a full treatment of radiative transfer
, it imposes negligible additional computational cost .
the hydrodynamic equations [ p2v ] and [ p2 t ] , combined with the radiative forcing described by equations [ tneq ] and [ rad2 ] provide a set of equations which are suitable for numerical integration . as in our previous simulations @xcite , the numerical integration is accomplished using the vector spherical harmonic transform procedure described in @xcite . to obtain a numerical solution ,
specific values for three parameters must be chosen : the base pressure @xmath21 , the pressure at the simulated layer @xmath22 , and the fraction @xmath63 of absorbed incident light .
the dynamics are not particularly sensitive to the choice of @xmath21 ; we take @xmath77 bar , which is a reasonable value .
larger values tend to increase wind speed , but due to the fact that the winds are coupled only logarithmically to @xmath21 , the effect is fairly small .
the choice of @xmath22 , the pressure at the simulated layer , determines both the wind speed and the radiative time - scale .
the pressure depth of the 8-@xmath0 m photosphere can not be determined precisely without a fairly detailed understanding of the atmospheric chemistry and structure ; we therefore assume @xmath22 to be a free parameter , albeit one which is relatively constrained by physical considerations .
the results presented here are based upon simulations with 100 mbar @xmath78 400 mbar .
likewise , the fraction @xmath63 of incident sunlight absorbed in the simulated layer must also be treated as a free parameter .
however , @xmath63 and @xmath64 are fairly well - constrained in the case of hd 189733 b by the observed flux variation .
while a full parameter study might yield values here which produce more precise fits , we find that taking @xmath79 , leading to @xmath80 , gives sufficient accuracy for our purposes in this preliminary investigation .
we note here that @xmath64 depends upon both @xmath63 and the intrinsic temperature @xmath69 , which is produced by internal heating . for this planet ,
we take a conservative value of @xmath81 k. for such low values of the internal heating , @xmath64 is determined almost completely by the reabsorption of thermal emission from deeper atmospheric layers .
to lowest order , our simulated flows agree with those obtained by @xcite . in the uppermost portions of the atmosphere ( @xmath82 mbar ) , air flows from the substellar point towards the antistellar point , so that the flow is westward on the evening terminator and eastward on the morning terminator . at deeper pressures , however , a persistent eastward equatorial jet develops , with a characteristic wind speed of @xmath83 km / s .
( at the pressure depths considered in these simulations , the maximum wind speed typically falls in the range 3.5 km / s @xmath84 5 km / s . )
this jet is supersonic : the speed of sound on hd 189733 b varies between 2.2 km / s in the coldest regions and 3 km / s in the hottest . ,
the top right at @xmath85 , the bottom left at @xmath86 , and the bottom right at @xmath87 .
the substellar point on these equirectangular plots is at the center of the leftmost ( or rightmost ! ) edge , while the antistellar point is directly in the center of each plot .
these flows were generated assuming @xmath79 and @xmath88 mbar ; under these assumptions , the equatorial jet reaches speeds in excess of 4 km / s . ]
a key difference between the atmospheric flow produced by our model and that found by @xcite is the development of large - scale turbulence in our simulations ; the flow in the @xcite model is laminar .
this is a robust characteristic of our simulations : some degree of turbulence arises at all pressure depths considered .
as might be expected , it is most extreme at lower pressures , where the wind speed is greater . in general
, one would expect to find turbulent flow in the upper atmosphere , due to the tremendous wind shears generated by the head - on collision of eastward and westward flows that characterize the lowest pressures .
the absence of turbulent conditions in the @xcite results is therefore perhaps surprising .
the evolution of the flow can be seen in figure [ 189733flow ] , which shows four equirectangular plots showing the wind velocity superimposed on the temperature distribtion .
m from hd 189733 b on two successive orbits .
the variation between orbits is likely to be significant if the model used here is correct .
transit corresponds to @xmath89 and @xmath90 , while the secondary eclipse occurs at @xmath91.during the first orbit the blue curve the flux minimum occurs 3.6 hours after ( the center of ) transit , while the flux maximum occurs 0.36 hours after secondary eclipse . during the second orbit , shown in red , the flux minimum occurs 0.81 hours after transit , with the maximum coming 1.1 hours before the secondary eclipse .
the black points show the data obtained after transit by @xcite . ]
mbar layer of hd 189733 b over one orbital period , as seen from earth .
the four globes correspond to the following times , measured from transit : top left , @xmath92 ; top right , @xmath93 ; bottom left , @xmath94 ; bottom right , @xmath95 .
hemispherical integration of these temperature distributions leads to the light curve shown blue in figure [ 189733lc ] . ]
mbar layer of hd 189733 b over a second orbital period , as seen from earth .
the four globes correspond to the following times , measured from transit : top left , @xmath92 ; top right , @xmath93 ; bottom left , @xmath94 ; bottom right , @xmath95 .
hemispherical integration of these temperature distributions leads to the red curve shown in figure [ 189733lc ] . ]
the flux variation depends primarily on the choice of @xmath63 , only weakly varying with @xmath22 . for @xmath96
we find a maximum relative flux @xmath97 , with a minimum relative flux of @xmath98 , with variations between orbits on the order of @xmath99 .
this is in good agreement with @xcite , who find a total flux variation @xmath100 , with @xmath101 : the results of our model yield @xmath102 , with @xmath103 .
our model shows significant variation in the shape of the light - curve from one orbit to the next , as can be seen in figure [ 189733lc ] . while the overall flux variation changes by only a small amount , the phase offset of the flux extrema from the time of central transit is measurably different from one orbit to the next . in the first orbit shown in figure [ 189733lc ] ,
the flux minimum occurs 3.6 hours after ( the center of ) transit , while the flux maximum occurs 0.36 hours after secondary eclipse . during the second orbit , shown in red ,
the flux minimum occurs 0.81 hours after transit , with the maximum coming 1.1 hours before the secondary eclipse .
the evolving temperature distribution that gives rise to these light - curves is shown in figures [ 189733earth1 ] ( first orbit ; blue curve in figure [ 189733lc ] ) and [ 189733earth2 ] ( second orbit ; red curve ) . while it is currently impractical rigorously to determine the positions of the flux minima and maxima over many orbits , it is possible to estimate the range over which they can be expected to vary .
we take a `` snapshot '' of the temperature distribution twice per orbital period over thirty orbits following the establishment of a dynamical equilibrium .
we then determine the light - curve that would result if the temperature distribution were to remain constant over the course of an entire period .
this provides a zeroth - order approximation to the variation in the shapes of light - curves that might be expected if one were to observe many orbits .
from the sixty light - curves that are obtained through this method , the mean position of the flux maximum was 0.89 hours before the secondary eclipse , with results ranging from 2.6 hours prior to the eclipse to 1.1 hours after the eclipse .
the mean position of the flux minimum was 0.69 hours after transit , with results ranging from 4.2 hours before transit to 4.7 hours after transit .
in future work , a more rigorous determination of these variations will be necessary .
it is also important to note that the times for flux maximum and minimum seem to depend rather strongly on the pressure depth of the simulation , with the largest variations occurring in the highly turbulent region between 100 mbar and 200 mbar .
these results are qualitatively consistent with the shape of the light - curve determined by @xcite . in the @xcite data ,
the flux minimum is found to occur @xmath104 hours after the transit , with the flux maximum preceding the secondary eclipse by @xmath105 hours .
it is worthwhile here to note that turbulent shallow - water simulations performed by @xcite also yield potentially variable light curves .
the dynamics in the @xcite model are completely different from those posted here : wind speeds are limited to @xmath106 m / s , and the flux variation is produced by the motion of cold spots produced by persistent circumpolar vortices , rather than the more chaotic collision of supersonic jets seen in our model .
nevertheless , it is instructive that those models in which the flow is turbulent produce variable light curves , while the light curves resulting from the laminar flows seen in @xcite and @xcite appear to be essentially constant from one orbit to the next .
m. the points show the transit depth over thirty successive orbits , while the dashed line shows the average transit depth of @xmath107 . ]
recent observations of the the depth of secondary transit over several orbits by @xcite place further constraints on the allowable inter - orbit variations .
@xcite find less than about @xmath108 deviation from their mean observed transit depth of @xmath109 . in figure
[ trns ] , we show the secondary transit depth over 30 successive orbits , as predicted by our model . we obtain a slightly larger mean transit depth of @xmath107 , with a maximum variation of about @xmath110 . thus , the model appears to be fully consistent with the @xcite observations .
while this model has not yet provided an optimized fit to the light - curve of hd 189733 b obtained by @xcite , it does provide a reasonable explanation for a dynamical configuration which could give rise to the observed light curve . in general ,
other models have been able to reproduce the observed phase offset of the flux maximum , corresponding to an eastward shift of the hottest temperatures from the substellar point @xcite .
it has been more difficult to account for the apparent westward shift of the coldest temperatures from the anti - stellar point .
however , the model proposed in this paper offers a physically reasonable mechanism by which the observed light curve may arise : at low pressures @xmath111 mbar , the flow tends to run from the substellar point to the antistellar point at supersonic speeds .
this results in a collision between eastward winds and westward winds , which produces chaotic , turbulent flow on a large scale . as a result
, there is no steady - state temperature distribution ; the enormous winds cause unpredictable shifts in the temperature distribution which , in turn , alter the position of the flux minima and maxima from orbit to orbit .
the phase offset of the flux minimum observed by @xcite is not inconsistent with these results .
furthermore , this hypothesis is readily testable : turbulence on the scale predicted by our model is expected to induce shifts of several hours or more in the timing of the flux minima and flux maxima ; the light - curve obtained by @xcite indicates that it is possible to measure the timing of these flux extrema with an error of @xmath112 hour .
therefore , if subsequent observations of hd 189733 b showed significant variations in the infrared light curve from one orbit to the next , the presence of significant large - scale turbulence would be strongly supported .
conversely , the absence of such variations would imply that turbulence on a scale necessary to explain the @xcite results is suppressed by other factors for which this simple two - dimensional model is unable to account .
some support for the model presented here may be found from the recent 24-@xmath0 m phase curve obtained by @xcite .
these observations were taken over the same portion of the orbit as the earlier 8-@xmath0 m observations @xcite . in both cases ,
the flux maximum precedes secondary eclipse .
in contrast to the 8-@xmath113 m data , however , the 24-@xmath0 m curve increases monotonically throughout the observing window
there is no flux minimum following the transit .
it is unclear whether this is due to a qualitatively different flow at the 24-@xmath0 m photosphere , or due to the type of turbulent variation suggested by the model presented in this paper
. it would be useful to obtain more observations in both the 8- and 24-@xmath0 m bands so that a comparison can be made between flows on multiple orbits , but at the same atmospheric depth . in any case
, it is quite clear that extensive further observations are necessary to obtain a detailed characterization of the atmospheric behavior of hd 189733 b. the model presented here represents a significant forward step from previous two - dimensional models .
although the number of free parameters in our treatment of the radiative forcing mitigates our success in fitting the @xcite light curve , the quality of the fit is sufficient to conclude that the dynamics produced by our model offer a reasonable explanation for the observations , while a poor fit would imply that the simulated flows are excluded by the data .
we can therefore say with some confidence that our model is not excluded by the observations currently available .
it is also the first model in either two or three dimensions to provide an explanation for the unexpected phase offset of the flux minimum in the hd 189733 b 8-@xmath0 m curve . despite these encouraging results , much room for improvement exists .
a more sophisticated treatment of the radiative forcing is necessary .
however , true radiative transfer is difficult to approximate in a two - dimensional model ; among other issues , the depth to which incident radiation penetrates depends rather strongly on the wavelength .
it is therefore likely that improvements in this area will require a shift to a fully three - dimensional code .
furthermore , a three - dimensional treatment is necessary to ensure that departures from stratified flow do not significantly affect the results .
however , the ubiquity of turbulent flow obtained in two - dimensional simulations suggests that small - scale motion is non - negligible , so that the low horizontal resolutions seen current three - dimensional models are less than ideal .
an immediate goal is therefore the production of a high - resolution three - dimensional model , including a realistic multi - wavelength treatment of radiative transfer . with the recent decision to fund a non - cryogenic _
mission motivated at least in part by the promise of useful observations of exoplanets , the efforts of modelers to assist in the identification of interesting targets will be critical
. whatever uncertainties still plague these efforts , it is certain that the influx of _ spitzer _ data over the next few years should provide rigorous tests on both existing models and on those yet to be developed .
we are grateful to jonathan fortney for useful suggestions .
the code employed in this paper makes use of the spherepack 3.0 routines written by drs .
adams and swarztrauber and provided by ncar .
this research has been supported by the nsf through career grant ast-0449986 , and by the nasa planetary geology and geophysics program through grant nng04gk19 g .
bakos , h. knutson , f. pont , c. moutou , d. charbonneau , a. shporer , f. bouchy , m. everett , c. hergenrother , d. w. latham , m. mayor , t. mazeh , r. w. noyes , d. queloz , a. pl , and s. udry . .
_ , 650:0 11601171 , october 2006 .
doi : 10.1086/506316 .
f. bouchy , s. udry , m. mayor , c. moutou , f. pont , n. iribarne , r. da silva , s. ilovaisky , d. queloz , n. c. santos , d. sgransan , and s. zucker . .
_ _ , 444:0 l15l19 , december 2005 .
doi : 10.1051/0004 - 6361:200500201 . | we have developed a new two - dimensional hydrostatically - balanced isobaric hydrodynamic model for use in simulation of exoplanetary atmospheres .
we apply this model to the infrared photosphere of the hot jupiter hd 189733 b , for which an excellent 8-@xmath0 m light curve has been obtained . for reasonable parameter choices ,
the results of our model are consistent with these observations . in our simulations ,
strongly turbulent supersonic flow develops , with wind speeds of approximately 5 km s@xmath1 .
this flow geometry causes chaotic variation of the temperature distribution , leading to observable variations in the light curve from one orbit to the next . |
the double pendulum has been studied as an example of the chaotic motion in physics@xcite .
if we consider the first cycle of the pendulum , the double pendulum model holds application in sports such as golf@xcite , baseball@xcite , and tennis@xcite . since the double pendulum is not a simple linear system , the motion of the pendulum can not be optimized as a simple analytic form .
the swing pattern utilised to maximize the angular velocity of the hitting rod such as racket , bat , and club has been analyzed on the assumption that the angular velocity is the dominant factor for the speed of the rebound ball@xcite . for a tennis stroke ,
the two rods for double pendulum are an arm and a forearm for first rod and a racket for the second rod . in our model
, we added the collision process between the ball and the racket . without the collision process@xcite ,
there is no criteria to attain high speed of the rebound ball except for the angular speed of the second rod , racket .
if we set the impact angle of the first rod at which the ball hits the racket , the speed of the rebound ball is mainly dependent on the angular velocity of the second rod . on the other hand ,
if we release the impact angle of the first rod , the speed of the rebound ball does not remain as a simple function of the angular speed of the hitting rod .
we showed that the speed of the rebound ball is different even though the angular velocities of the second rod at the contact time are same .
furthermore , considering the whole stroke , the maximum angular velocity for the lower speed of the rebound ball is greater than the maximum angular velocity for the higher speed of the rebound ball . therefore , to get maximum speed of the rebound ball , it s not sufficient to set the condition to generate high angular velocity of the hitting racket .
the collision between the racket and ball has been studied in various ways @xcite . in our simple collision model
, we assumed that the racket is a simple one dimensional rod without any nodal motion and the collision occurs in one dimension .
this assumption is valid if the racket and a ball moves in the same line for the short collision time .
although , our model does not give any detailed information on the collision about the effect of the tension of the string and the mass distribution of the racket , however provides some insights on the proper swinging pattern to get maximum speed of the rebound ball in this article , we also analyzed the tame lagged torque effect for the double pendulum system . by applying time independent constant torques on first rod and the second rod
, the speed of the rebound ball can be calculated for certain initial conditions . in the same condition ,
if we simply hold the racket for a short time without enforcing a torque and with subsequent application of torque , the naive intuition estimates decrease in the speed of the rebound ball .
the reason behind this is application of less energy for the double pendulum system .
however , the speed of the rebound ball increase by @xmath0 by choosing the proper delay time .
it s mainly because the double pendulum system is not a simple linear system .
adding energy to the double pendulum system does not directly increase the speed of the rebound ball .
we also analyzed movement of the elbow to which the first rod of the double pendulum is attached . at a first glance ,
if we add extra movement towards the rebound ball s direction at contact time , the speed of the ball increases ; however , with varying results .
apparently , it becomes clear that , the double pendulum system is really a nonlinear system .
the present paper is organized as follows : in section ii , we introduce a double pendulum system including the collision process . in section iii ,
the differential equations obtained in section ii are solved , we numerically showed that the speed of the rebound ball does not simply depend on the angular speed of the racket .
for some cases , the higher angular velocity gives lower speed of the rebound ball . in section iv
, we analyzed the dependence of the racket mass and length of the first rod .
the general properties of swing system has been demonstrated .
when we applied a time dependent torque on the racket , we could increase the speed of the rebound ball .
the time lagged rotation of the racket was analyzed in section v. in section vi , the elbow movement is analyzed to add additional speed to the ball . in section vii ,
we summarize the main results and discuss the application of our results .
the geometry of the double pendulum model for the swing of a racket is shown in fig [ figracket]-(a ) .
though , this geometry is for the left - handed player if we see from the @xmath1 direction , it s originally related to the real double pendulum problem in the gravitational field @xcite . our basic model and some notations
are closely related with those in work reported by rod cross @xcite .
the elbow moves in the @xmath2 plane and the arm and the racket also moves in @xmath2 plane .
we also modeled the arm and the forearm as a simple uniform rod with mass @xmath3 and length @xmath4 .
the racket including the hands is also treated as a uniform rod with mass @xmath5 and length @xmath6 . the arm and the racket rotate in a clockwise direction in a plane at angular velocities @xmath7 and @xmath8 , respectively .
if we assume that the velocity of the elbow is @xmath9 then the velocities of the center of the first rod ( arm and forearm system ) and the second rod ( hand and racket system ) becomes @xmath10 , where @xmath11 , @xmath12 are the velocity of the center of the mass of the first rod and the second rod , respectively .
the center of masses are located in the middle of the rod , @xmath13 and @xmath14 , since we assumed uniform rods . ,
and the length of the racket is @xmath6 .
the angles @xmath15 and @xmath16 are defined from the @xmath17 direction .
( b ) the force on two rods .
@xmath18 is the force acted on the first rod at the joint between elbow and the first rod .
@xmath19 is the force acted on the second rod at the joint between the two rods .
@xmath20 is the force on the second rod by a ball .
, width=188 ] let the force from the elbow to the first rod on joint point between the elbow and the first rod be @xmath21 , and the force from the first rod on joint point between two rods be @xmath22 , then the equations of the motion for two center of masses become @xmath23 @xmath24 the two forces @xmath25 are reaction force of @xmath26 from the second rod .
we added the force @xmath27 from the ball to the second rod at the point of @xmath28 from the center of the mass .
this force from the collision between the racket and the ball changes the torque equation in @xcite and the two torque equations becomes @xmath29 @xmath30 , where @xmath31 is the torque on joint point between the elbow and the first rod and @xmath32 is the torque on the joint point between the two rods .
from the eqs .
[ eqf1 ] - [ torque1 ] , we obtain two equations for the time derivative of two angular velocities as follows @xmath33 where @xmath34 , @xmath35 , and @xmath36 \nonumber \\ q & = & c_2 + m_2 h_2 l_1 \omega_1 ^2 \sin ( \phi-\theta ) + m_2 h_2 [ a_x \cos \phi + a_y \sin \phi ] \label{eqpq}\end{aligned}\ ] ] @xmath37 and @xmath38 are similar to the results in @xcite with setting @xmath39 , and the collision force from the ball add following two terms @xmath40 \nonumber \\ s_2 & = & f_{col } [ ( 2 h_h ( i_{1,cm } + m_1 h_1 ^2 + m_2 l_1 ^2 ) \nonumber \\ & + & h_2 ( 2 i_{1,cm } + 2 m_1 h_1 ^2 + m_2 l_1 ^2 ) ) \cos \phi - h_2 m_2 l_1 ^2 \cos ( 2 \theta - \phi ) ] \label{eqs12}\end{aligned}\ ] ] the collision force defines the motion of the ball as follow , @xmath41 , where @xmath42 is the @xmath43 component of the hitting part of the racket which locates @xmath44 from the bottom of the racket , and @xmath45 is the @xmath43 component of the joint between the elbow and arm . the force between the racket and the ball should be repulsive , so that the force form becomes @xmath46 if we consider this harmonic force between the ball and the second rod , the period of the oscillation is @xcite @xmath47 since the half of this period is the collision duration between the ball and the racket , we controlled the @xmath48 value from @xmath49 to @xmath50 .
subsequently , the collision time varies from @xmath51 to @xmath52 .
however , if the contraction length is large , the hook s model is not valid for the ball and the racket system . then the force can be rewritten as @xmath53 @xcite . or @xmath54 in order to limit the maximum contraction length .
however , in our numerical calculation we restricted our system in order to follow the hook s rule . in this article
, we assumed a simple model for the ball and racket collision . in our model ,
the ball is assumed to hit the racket when the racket is parallel to @xmath55 axis ( @xmath56 ) .
in addition to this , the ball is assumed to be moving in @xmath43 axis . in this case
, we numerically calculated @xmath57 and @xmath58 till the time @xmath59 , when the racket is parallel to the @xmath55 axis @xmath60 . with these numerical results ,
we set new initial conditions just before the collision .
we assumed that the collision occurs at @xmath61 , where @xmath62 is the period of the harmonic oscillator system between the racket and the ball . at this time
, we numerically solved the ball and two rod system attached to the moving elbow numerically till the ball is rebound from the racket .
we determined the speed of the ball at that moment . based on our simple model ,
the coupling constant @xmath48 determines the collision duration , but the speed of the rebound ball is not altered very much .
the reason for such a behavior is consideration of one dimensional collision .
however , the purpose of this article was to find the optimum path for the swing of the racket , we did not extend our model to two dimension .
we only restricted our numerical conditions to render presence of our setups in the valid region .
we assumed that an arm and a forearm forms a simple rod and the moment of inertia about the center of mass is @xmath63 .
the moment of inertia about the center of mass of the hand - racket system is also assumed as @xmath64 . although this assumptions are not enough to study the tennis stroke in detail , we focused our attention on the double pendulum model , which has its application in the tennis stroke . in this section , the torques applied to the first forearm system @xmath31 and to the racket system @xmath65 were set by @xmath66 and @xmath67 as described in rod cross work @xcite .
the velocity ( @xmath68 ) of the elbow is assumed as follows @xcite , @xmath69 in fig .
[ fthetadeltalp3 ] , we plotted the speed of the rebound ball from the racket as a function of the initial angles @xmath70 .
the angle @xmath71 is defined as the angle between the first rod and the racket .
@xmath72 is the initial angle at @xmath73 .
we assumed that the length of the first rod ( an arm and a forearm system ) as @xmath74 and the mass of the first rod as @xmath75 . and the length of the racket system as @xmath76 and the mass of the racket system as @xmath77 .
we assumed that the initial velocity of the ball just before the contact is @xmath78 to the positive @xmath43 direction .
the @xmath42 , a hitting position from the center of mass assumed to be as @xmath79 .
the length of the first rod ( an arm and forearm system ) assumed to be as @xmath80 .
this length is projection length in @xmath81 planes .
we plotted initial conditions which gives maximum speed of the rebound ball .
as @xmath82 changes from @xmath83 to @xmath84 , the @xmath72 becomes smaller and reaches @xmath85 . in fig .
[ fthetadeltalp4 ] , we also plotted the speed of the rebound ball from the racket as a function of the initial angles @xmath86 .
we only changed the length of the first rod , in other words we extended the distance between the elbow and the hand . at this time
the @xmath72 is smaller as compared to the case when @xmath74 .
when the initial angle @xmath82 is @xmath84 , the racket and the first rod should be in the same line so as to get the maximum speed of the ball .
unit as a function of the initial angle of the racket and arm with respect to the @xmath17 axis .
@xmath82 : initial angle of the arm .
@xmath72 : initial angle between the racket and the first rod .
the length of the first rod is @xmath80 .
red line indicates two initial angles which gives the maximum speed .
, width=188 ] unit as a function of the initial angle of the racket and arm with respect to the @xmath17 axis .
@xmath82 : initial angle of the arm .
@xmath72 : initial angle between the racket and the first rod .
the length of the first rod is @xmath87 .
red line indicates two initial angles which gives the maximum speed , width=188 ] in our model , the speed of the rebound ball is not a simple function of the angular velocity @xmath88 . in fig .
[ fw1nw2 ] , we plotted the angular velocities @xmath89 and @xmath88 for two cases . for the case
a ) the initial angle @xmath90 is @xmath91 and the contact time @xmath92 and the angular velocity is @xmath93 . for the case
b ) , the initial angle @xmath90 is @xmath94 and the contact time @xmath95 and the angular velocity @xmath96 .
however , the angular velocities @xmath88 at the contact times have the similar value as @xmath97 .
furthermore , the angular velocity @xmath98 increases even after the collision time .
the interesting thing is that the speed of the rebound ball for the case b ) is smaller than the speed of the ball for the case a ) .
the speeds are @xmath99 and @xmath100 , respectively .
if we only check the angular velocity @xmath88 , we may conclude that the case b ) gives higher speed of the rebound ball .
but the angular velocity is not the entire factor to determine the speed of the ball .
this can be explained if we examine the angle @xmath101 of the first rod when the racket contacts the ball .
@xmath102 s are @xmath103 and @xmath104 for the cases a ) and b ) , respectively .
the speed of the ball ( @xmath105 ) is also a functions of @xmath102 as well as @xmath106 and @xmath88 . in order to get the maximum speed of the rebound ball
, the double pendulum system should be examined as a whole system . and @xmath88 for two cases .
the initial angle @xmath90 is @xmath107 for the case a ) and @xmath108 for the case b ) . @xmath109 and @xmath110 are the contact time.,width=188 ]
the mass of the racket - hand system is assumed to be @xmath111 , but the mass of the racket may be changed .
all the movements are assumed to be in the @xmath2 plane in our model , the projected length between the elbow and the hand in @xmath2 plane can be changed by controlling the angle between the forearm and arm in actual tennis stroke . in fig .
[ flxmyvout ] , we plotted the speed of the rebound ball from the racket as a function of the racket mass @xmath112 and the projected length in @xmath2 plane between the elbow and the hand @xmath4 .
if the mass of the racket is about @xmath113 , the speed of the rebound ball is decreased as the projected length between the elbow and the hand is increased . however
, if the racket mass is getting heavier , the speed of the rebound ball increases . since an actual mass of the tennis racket is around @xmath114 , and @xmath4
is limited , the speed of the ball is restricted . in fig .
[ flxmytheta ] , we plotted the angle ( @xmath102 ) of the forearm system at the impact time as a function of the racket mass @xmath112 and the projected length in @xmath2 plane between the elbow and the hand @xmath115 regardless of the racket mass ( @xmath5 ) , the forearm angle @xmath102 decrease to @xmath116 as the length @xmath115 increase .
this results shows that for the player with folded arm whose effective projected length of the forearm system is small , the contact angle ( @xmath117 ) should be about @xmath118 .
if the length @xmath4 and the contact angle @xmath119 are reduced the ball should be hit relatively close to the body .
this tendency may explain the impact point of the ball in tennis stroke and golf . and the projected length in @xmath2 plane between the elbow and the hand @xmath4 .
the unit of @xmath105 is @xmath120 , width=188 ] and the projected length in @xmath81 plane between the elbow and the hand @xmath4.,width=188 ]
in actual tennis stroke , most of the players use time lagged racket movement .
they intentionally keep the racket back as forearm rotates then they start to move the racket to get a high angular velocity @xmath88 .
the main difference of double pendulum when compared to the single pendulum , is the separate movement of the first rod ( arm and forearm system ) and the second rod ( the racket system ) . in fig .
[ fcbtau ] , we plotted the speed of the rebound ball as a function of increase in the torque @xmath32 and the delay time @xmath121 .
we set the projected length in @xmath2 plane between the elbow and the hand @xmath115 as @xmath87 .
the torque applied to the racket at the joint of two rods , varies from @xmath122 to @xmath123 .
the time delay @xmath121 is the starting time at which we applied the torque to the racket . after waiting for @xmath121 second
, the torque suddenly changes from zero to a certain value @xmath32 till @xmath124 in our numerical calculation .
this time is far behind the contact time ( around @xmath125 ) . at first glance ,
if we start to apply the torque early , the angular velocity of the racket may accelerate little bit more .
it s simply because the acceleration of the angular velocity depends on the torque .
if the torque @xmath32 is less than @xmath126 , the maximum speed of the rebound ball is obtained when the delay time @xmath121 is zero as we expected .
however , for the higher torque @xmath127,the numerical results in fig .
[ fcbtau ] are different from our simple intuition . for a given torque value @xmath128
, there always exist a certain time delay @xmath121 at which the speed of the rebound ball is maximum and the @xmath121 is not zero . in other words ,
the important thing to get high speed of the rebound is not the total amount of impulse ( torque @xmath129 ( applied time ) ) , but the timing when the torque starts . and the time lag @xmath121 .
the initial angle @xmath130 is @xmath131 , the which angle gives the maximum speed for the angle @xmath132 .
@xmath133 , width=188 ] if the torque @xmath32 is lagged , the initial angle @xmath134 to get a maximum speed of the rebound ball for the fixed initial angle @xmath135 is no more the optimum condition .
in order to see the time lag effect clearly , we set the torque @xmath136 . in fig .
[ ftimedelta ] , we plotted the speed of the rebound ball as a function of the time lag @xmath121 and initial angle @xmath72 with an initial angle @xmath135 . the new optimum condition to get maximum speed of the rebound ball is the time lag @xmath137 and the initial angle @xmath138 . and initial angle @xmath72 with an initial angle @xmath135,width=188 ] in fig .
[ ftimedelay ] , we plotted the speed of the rebound ball as a simple function of time delay @xmath121 . if the time lag @xmath121 of the torque @xmath32 is @xmath139 , the rebound ball speed is @xmath140 .
and @xmath105 has its maximum ( @xmath141 ) at @xmath142 .
the speed increases by about @xmath143 by adjusting the time delay with the same magnitude of the torque @xmath32 .
the angular velocity of the forearm system and the racket system is shown in fig .
[ ftimedelayw1w2 ] . for two time delay @xmath144 and @xmath145 ,
the difference in the contact time is less then @xmath146 . the dashed line in fig .
[ ftimedelayw1w2 ] indicates the time at which the ball collides with the racket .
we note that the racket hit the ball before the racket has its maximum angular velocity . in earlier work @xcite ,
the authors have analyzed the double pendulum system in order to get the maximum angular velocity @xmath88 .
however , this may mislead the double pendulum system .
considering fig .
[ fw1nw2 ] , it is clear that higher angular velocity gives lower speed of the rebound ball . in fig .
[ ftimedelayw1w2 ] , when the angular velocity @xmath147 has its maximum , the angular velocity @xmath106 is almost zero .
this is explained for the double pendulum model of tennis stroke .
the angular momentum and energy of the first rod ( arm and forearm system ) were totally transferred to the racket system in order to get maximum angular velocity of the racket . .
at @xmath148 , the @xmath105 has it s maximum .,width=188 ] and the racket system ( @xmath149 ) .
@xmath150 indicates the angular velocities when the time delay is @xmath144 , and @xmath151 indicates the angular velocities when the time delay is @xmath145,width=188 ] we plotted the first rod ( forearm and arm system ) and the racket system for the time delays represented as @xmath145 and @xmath144 in fig .
[ timelaggedstroke ] and [ constantstroke ] , respectively .
when the first rod starts to rotate , the racket stayed back in fig .
[ timelaggedstroke ] till time @xmath145 . after applying the torque @xmath32 ,
the racket suddenly starts to rotate , and hits the ball with the angle @xmath152 .
the racket and the first rod are shown in red at the contact time . comparing the stroke with a constant torque ( fig .
[ constantstroke ] ) , the racket rotates more rapidly at the contact time . in double pendulum model for the stroke , this phenomena was expected qualitatively . in our model
we quantitatively demonstrated that why time lagged stroke is needed and the extent to which the velocity can be increased .
the motions are captured from @xmath73 to @xmath153 evenly .
the red one is the racket position at the contact time .
, width=188 ] .
the motions are captured from @xmath73 to @xmath153 evenly .
the red one is the racket position at the contact time.,width=188 ]
in the previous section , we noted that the important thing is not the total impulse to accelerate the racket , but the timing in order to increase the speed of the rebound ball .
now , we analyze the influence of the elbow movement on the speed of the rebound ball .
we made a simple model to add an additional movement of the elbow . the added force around the time @xmath145 as follows , @xmath154 where @xmath155 is the angle measured from the @xmath156 axis counterclockwise , we also set @xmath157 , and the width @xmath158 . since the ball is rebound to @xmath159 axis , the proper direction of the force seems to be in the @xmath159 direction .
we plotted the speed of the rebound ball when the extra force is applied in the direction of @xmath155 and with extra time delay @xmath160 in fig .
[ fforcem ] .
numerical results show that we can obtain the maximum speed of the rebound ball when the angle is @xmath161 .
when the direction of the force @xmath162 , the speed of the rebound ball is almost same to the speed of the ball without the additional force .
in other words , the external movement towards @xmath159 direction does not give any additional speed to the rebound ball . in fig .
[ farmm ] , we plotted the trajectory of the elbow from the time @xmath73 till the contact time .
@xmath163 is the path when no additional force is added .
@xmath164 are elbow s trajectory when the direction of the force @xmath155 are @xmath165 and , @xmath166 , respectively .
table [ tb1 ] demonstrates the @xmath105 and the angular velocity @xmath88 for 4-cases .
the movement with @xmath167 is slightly backward and mainly perpendicular to the direction of the rebound ball . from this result , we conclude that the main factor for the increased speed of the rebound ball is not the linear momentum added to the racket but the angular momentum added to the racket system . and with a extra time delay @xmath160 , width=188 ] till the contact time .
@xmath163 is the path when no additional force is added .
@xmath168 are elbow s trajectory when the direction of the force @xmath155 s are @xmath169 , respectively.,width=188 ] .additional force dependent output [ cols="^,^,^,^,>",options="header " , ]
the double pendulum model is applied to the baseball , tennis , and golf .
it analyzes the swing pattern to maximize the angular velocity of the hitting rod such as racket , bat , and club on the assumption that the angular velocity is the dominant factor to attain speed of the rebound ball .
if we set , @xmath102 , the angle of the first rod at the impact time , it seems to be obviously reasonable . on the other hands , if we release the angle @xmath102 , the speed of the rebound ball is not a simple function of the angular speed of the hitting rod .
the speed of the rebound ball is different even though the angular velocities at the contact time are same in fig .
[ fw1nw2 ] , . furthermore
, considering the whole stroke , the maximum angular velocity @xmath88 for the lower speed of the rebound ball is greater than the maximum angular velocity @xmath88 for the higher speed of the rebound ball .
therefore , to attain maximum speed of the rebound ball , it s not sufficient to set the condition to generate high angular velocity @xmath88 of the hitting rod . in the double pendulum ,
the efficient way to generate high angular velocity of the hitting rod is the energy transfer from the first rod to the second hitting rod . in other words , when the angular velocity @xmath106 of the forearm system is zero , the angular velocity of the hitting racket has maximum value as shown in fig .
[ ftimedelayw1w2 ] .
we analyzed the time lagged torque effect for the double pendulum system . with applying constant torques @xmath170 on the forearm system and on the racket , respectively
, the speed of the rebound ball can be calculated . for the same condition , if we simply hold the racket for a short time without enforcing a torque @xmath32 , then applying the torque @xmath32 at the proper time ( @xmath171 ) , the speed of the rebound ball increases by @xmath0 as can be seen in fig .
[ ftimedelay ] .
the reason is mainly because the double pendulum system is not a simple linear system . adding the energy to the double pendulum system
does not directly increase the speed of the rebound ball .
we also analyzed the elbow movement effect .
in addition to the velocity of the elbow for the medium pace forehand , we added extra movement of the elbow . at a first glance ,
if we add extra movement towards the rebound ball s direction , the speed of the ball is increased .
but the double pendulum system is not a simple linear system .
when the direction of the elbow movement is perpendicular to the ball s direction , the speed of the rebound increases .
actually the direction is towards the center of the elbow s circular movement . in other words ,
the added centripetal force does not add not the linear momentum of the racket , but the angular velocity of the racket .
although our collision model is applied in one dimension , this collision process allows us to analyze the double pendulum system in a more realistic manner .
we showed that the speed of the rebound ball does not simply depend on the angular velocity of the racket .
the increase in the ball speed by the proper time lagged racket rotation was numerically studied . the elbow movement for adding
the ball s speed was counter intuitive .
the addition of simple linear momentum to the elbow is not important ; however , the elbow should move in order to add angular velocity to the racket . in actual tennis stroke ,
the motion occurs in three dimensions and the magnitudes of the forces and torques are dependent on the muscle shape and movement .
we did not include any bio - mechanical information such as pronation and we did not include the spin of the ball in any way .
the numerical data may also be not suitable for some players .
however , our study on the double pendulum system for tennis stroke provides some insights to attain an efficient way to stroke a tennis ball .
this study was supported by the basic science research program through the national research foundation of korea ( nrf ) funded by the ministry of education , science and technology ( nrf-2014r1a1a2055454 ) r.b .
levien , and s. m. tan , `` double pendulum : an experiment in chaos '' am . j. phys .
* 61 * , 1038 - 1044 ( 1993 ) 1 t .
t. shinbrot , c. grebogi , j. wisdom , and j. a. yorke , `` chaos in a double pendulum '' , am .
* 60 * , 491 - 499 ( 1992 ) g. vadai , z. gingl and j. mellar , `` real - time demonstration of the main characteristics of chaos in the motion of a real double pendulum '' , eur .
* 33 * , 907 ( 2012 ) d. williams , `` the dynamics of the golf swing . '' , q. j. mech .
appl . math . * 20 * , 247 - 264 ( 1967 ) c. b. daish , _ the physics of ball games " _ .
( english university press , london , 1972 ) t. jorgensen , `` on the dynamics of the swing of a golf club , '' am . j. phys .
* 38 * , 644 - 651 ( 1970 ) t. jorgensen , _ the physics of golf _ ( springer - verlag , new york , 1999 ) , 2nd ed .
r. cross , `` a double pendulum swing experiment : in search of a better bat , '' am .
* 73 * , 330 - 339 ( 2005 )
. r. cross , `` a double pendulum model of tennis strokes '' , am .
* 79 * ( 5 ) , 470 ( 2011 ) .
r. cross,``the coefficient of restitution for collisions of happy balls , unhappy balls , and tennis balls '' , am . j. phys .
* 68 * , 1025 - 1031 ( 2000 ) .
r. cross , `` impact of a ball with a bat or racket , '' am .
67 , 692 - 702 ( 1999 ) r. cross , `` oblique impact of a tennis ball on the strings of a tennis racket '' , sports engineering , * 6 * , 235 - 254 ( 2003 ) | by means of adding a collision process between the ball and racket in double pendulum model , we analyzed the tennis stroke .
it is possible that the speed of the rebound ball does not simply depend on the angular velocity of the racket , and higher angular velocity sometimes gives lower ball speed .
we numerically showed that the proper time lagged racket rotation increases the speed of the rebound ball by @xmath0 .
we also showed that the elbow should move in order to add the angular velocity of the racket . |
the floor layout problem ( flp ) , also known as the ( unequal areas ) facility layout problem , is central to the design of objects such as factory floors and very - large - scale integration ( vlsi ) computer - chips .
the designer is given a fixed rectangular floor and @xmath0 rectangular boxes to place onto the floor .
each box must sit completely on the floor , and they can not overlap .
each box has a fixed area , but the widths and heights can be varied to change the shape , subject to constraints on the area and aspect ratio of the components .
the objective is to minimize the weighted sum of the manhattan norm distances between each pair of boxes .
the flp can be naturally described as a disjunctive programming problem , which are often reformulated as mixed - integer programming ( mip ) problem such as to take advantage of state - of - the - art mip solvers . however , the flp and its various mip formulations have proven extremely difficult to solve to optimality . in this work , we take a systematic approach to generating mip formulations for the flp that unifies existing mip formulations from the literature and leads to new formulations and valid inequalities .
we also computationally compare the range of formulations , and show that the new approaches can be used to solve previously unsolved instances .
the main contributions of this work include : 1 .
* case study on systematic construction of effective mip formulations : * the number , heterogeneity , and complexity of existing mip formulations for the flp and the fact that it remains computationally challenging make it an excellent candidate for such a study . through the use of the _ embedding formulation _ approaches of @xcite and @xcite and through a systematic treatment of alternative disjunctive descriptions of the flp
, we are able to recover and unify all existing , seemingly ad - hoc , mip formulations . in addition
, we are able to derive new formulations that can provide a significant computational advantage and solve previously unsolved instances .
while the study concentrates on specific characteristics of the flp , it exemplifies generic formulation techniques and practices that should be useful for a wide range of problems . 2 . * valid inequalities for alternative formulations of the flp : * using the embedding approach , we are able to construct a variety of new valid inequalities for flp .
one key of the embedding formulation approach of this work is the flexible use of @xmath1 variables to model disjunctive constraints or unions of polyhedra .
however , such flexibility can cloud the `` interprebility '' of the @xmath1 variables , which is often needed to construct valid inequalities to strengthen formulations . in this work ,
we show how ideas from @xcite can be used to translate valid inequalities between formulations of the flp that allows us to state a broad class of valid inequalities in a generic form .
* comprehensive computational study of flp : * while the flp has been extensively studied , most existing works compare only a small subset of formulations and valid inequalities when making comparison . in this work
we collect several instances from the literature to construct a publicly available library and present a comprehensive computational study of existing and new formulations on this library .
furthermore , our systematic approach allows us to compare a host of formulations and a wide range of common valid inequalities when making our comparison . in particular , while no single formulation seems to be dominant , we may offer a small collection of techniques which prove particularly effective for the flp .
furthermore , we also study various theoretical and practical aspects of these approaches that help explain their success .
the remainder of this work is organized as follows . in section
[ sec : lit ] we present a literature review of the existing solution techniques for the flp . in section [ sec : def ] we formally define the flp and show how it can be cast as a disjunctive programming problem . in section [ sec : formulations ] we review the formulation techniques we use to transform the flp into a mip , and in section [ sec : twobox ] we use the techniques to construct formulations that are based on the interaction of two boxes at a time .
then in section [ sec : inequalities ] we develop valid inequalities that can be used to strengthen formulations and show how they can be translated from one formulation to the other . in section [ sec : inequalities ] we also restrict attention to the interaction of two boxes at a time , so in section [ sec : multibox ] we develop formulations and inequalities that are based on the interaction of larger collections of boxes .
finally , in section [ sec : computations ] we present results of our computational experiments , and in section [ sec : conclusions ] we present a brief summary of this work . complementary material and
omitted proofs are included in the appendix .
the floor layout problem can be viewed as a specific version of a general layout problem that consists of orthogonally packing rectangular pieces onto a rectangular floor ; @xcite offer a taxonomy of variations of the flp and its relatives .
originally studied primarily in the context of factory design , the emergence of the field of very - large scale integration ( vlsi ) computer - chip design saw renewed interest in layout problems such as the flp . broadly , algorithmic approaches to these layout problems
can be grouped into two classes : exact and heuristic .
exact algorithms were predominant in the earlier literature , although the boom of applications in computer - chip design require solving large scale instances beyond the reach of existing exact approaches . as a result , a bevy of work has appeared over the past three decades , proposing heuristic approaches to produce good solutions for large - scale instances .
much of the work applies existing metaheuristic frameworks to the flp , for example @xcite and @xcite .
contrastingly , many of the novel heuristics for the flp take advantage of ideas and machinery from mathematical programming : e.g. @xcite , and @xcite , albeit in a way that can not prove optimality .
we note in particular the surveys of @xcite and @xcite , which collect pointers to much of the heuristic literature . in keeping with the mip approach taken in this paper , we will survey the existing exact methods for the flp in detail .
early work can be traced back to @xcite , who studies a discretized version of the flp .
@xcite introduced a natural mip model for the flp , along with a collection of valid inequalities and techniques to help reduce solution time .
@xcite introduces novel formulations for a single pair of boxes , as well as useful computational techniques such as symmetry breaking constraints and branching priorities .
@xcite presents a new mip formulation for the flp with fewer binary variables , alongside a number of additional formulations and approaches inspired by nonlinear and mixed - integer nonlinear optimization .
@xcite presents another formulation inspired by a technique from @xcite that reduces redundancy in the solution set . as detailed in the following section , the inclusion of certain non - linear area constraints in the flp result on its formulations being second - order - cone mip ( soc - mip ) problems
given that early formulations were developed before the availability of efficient soc - mip solvers , careful attention has been paid on constructing and proving desirable properties for specific linear approximations for the nonlinear area constraints in @xcite and @xcite .
the flp has a natural one - dimensional analogue in the single - row floor ( facility ) layout problem , which asks for an optimal layout of @xmath0 boxes of fixed length in a straight line .
this problem is already np - hard , and strong formulations and cutting planes have been developed for the problem by @xcite and @xcite .
an intriguing line of research has investigated the flp from the dual perspective , attempting to construct tight lower bounds .
this is of particular interest for the flp , where relaxations typically give poor bounds , even with strengthening valid inequalities .
@xcite presents a lower bounding technique for the single - row flp .
another line of work investigates using semidefinite programming formulations to construct bounds for the flp by @xcite and the single - row flp by @xcite .
@xcite leverage the semidefinite approach to produce optimal solutions for the single - row flp using a cutting - plane approach , and to produce high - quality solutions for larger instances in @xcite .
@xcite present a combinatorial dual bounding scheme for the flp and compare it against existing techniques .
consider a rectangular floor @xmath2 \times [ 0,l^y]$ ] for @xmath3 .
there is a collection of @xmath0 boxes @xmath4 to place on the floor , each with a target area @xmath5 and maximum allowed aspect ratio @xmath6 .
denote the set of all pairs with @xmath7 where @xmath8{\mbox{\normalfont\tiny\sffamily def}}}{=}}}\{1,\ldots , n\}$ ] . with each pair of boxes @xmath9
, there is an associated nonnegative unit communication cost @xmath10 .
the floor layout problem then is to optimally lay out each box completely on the floor , such that the area and aspect ratio constraints are satisfied , and such that no two boxes overlap .
natural decision variables for each box @xmath11 are the position of its center @xmath12 and the lengths in each direction @xmath13 . the objective function used is based on the so - called `` manhattan '' norm : @xmath14 most of the constraints described are simple to describe with linear or conic inequalities .
for instance , @xmath11 lies completely on the floor iff @xmath15 the area constraints take the form @xmath16 which is second - order - cone - representable @xcite .
the aspect ratio constraints take the form @xmath17 this can be represented with two linear constraints per box , but it can also be enforced on the flp merely through bounds on the widths of the boxes . along with the area constraints , imposing the following bounds on the box widths is sufficient to impose the aspect ratio constraints : [ eqn : bounds ] @xmath18{\mbox{\normalfont\tiny\sffamily def}}}{=}}}\min\left\ { \sqrt{\alpha\beta } , l^s \right\ } \quad & \forall s \in \{x , y\ } , i \in \llbracket n \rrbracket \label{eqn : upperbound } \\
\ell^s_i & \geq lb^s_i { \mathrel{\overset{\makebox[0pt]{\mbox{\normalfont\tiny\sffamily def}}}{=}}}\frac{\beta}{ub^s_i } \quad & \forall s \in \{x , y\ } , i \in \llbracket n \rrbracket\label{eqn : lowerbound}\end{aligned}\ ] ] we note that , since @xmath19 , we have that @xmath20 for each @xmath21 and @xmath22 .
the last remaining constraint for the flp requires that the boxes can not overlap on the floor .
one natural way to formulate this is by requiring each pair @xmath11 and @xmath23 to be separated in either the @xmath24 direction or the @xmath25 direction ( or both ) .
[ defn : precede ] we say that @xmath11 _ precedes _
@xmath23 in direction @xmath26 ( denoted by @xmath27 ) if @xmath28 therefore , we can enforce the constraint that @xmath11 and @xmath23 do not overlap with the disjunctive constraint @xmath29{\mbox{\normalfont\tiny\sffamily def}}}{=}}}\bigvee_{k=1}^4 d^k_{i , j}$ ] , where @xmath30 we omit the subscript and use @xmath31 when the meaning is clear from context and we refer to each @xmath32 as a _ branch _ of the disjunction .
( left ) , @xmath33 ( center ) , and both @xmath34 and @xmath33 ( right).,title="fig : " ] ( left ) , @xmath33 ( center ) , and both @xmath34 and @xmath33 ( right).,title="fig : " ] ( left ) , @xmath33 ( center ) , and both @xmath34 and @xmath33 ( right).,title="fig : " ] then the set of all feasible layouts is given by the disjunctive set @xmath35{\mbox{\normalfont\tiny\sffamily def}}}{=}}}\left\ { ( c,\ell ) \in { \mathbb{r}}^{2n+2n } : \eqref{eqn : sitb},\eqref{eqn : bounds},\eqref{eqn : area } , \bigwedge_{(i , j ) \in { \mathscr{p } } } d^4_{i , j } \right\},\ ] ] and a ( nonlinear ) disjunctive programming formulation of the flp is given by @xmath36 .
the main objective of this work is to transform this disjunctive programming problem into a mixed - integer formulation that can be solved by off - the - shelf optimization software . to achieve this , we first focus our attention to a single pair of boxes @xmath11 and @xmath23 for @xmath9 , in which case we are interested in constructing a mip formulation of @xmath37{\mbox{\normalfont\tiny\sffamily def}}}{=}}}\left\ { ( c_i , c_j,\ell_i,\ell_j ) \in { \mathbb{r}}^8 : \eqref{eqn : sitb},\;\eqref{eqn : bounds},\;\eqref{eqn : area},\ ; d^4 \right\}$ ] if we omit the nonlinear area constraints we obtain the set @xmath38{\mbox{\normalfont\tiny\sffamily def}}}{=}}}\left\ { ( c_i , c_j,\ell_i,\ell_j ) \in { \mathbb{r}}^8 : \eqref{eqn : sitb},\;\eqref{eqn : bounds},\ ; d^4 \right\}$ ] , which can be written as the union of ( bounded ) polyhedra @xmath39 , where @xmath40 .
we can then use various techniques to construct a linear mip formulation of @xmath41 , which can be combined with the area constraints to obtain a second - order - cone mip ( soc - mip ) formulation of @xmath42 .
finally , we can combine these formulations for all pairs @xmath9 and an appropriate linearization of the objective function to obtain a soc - mip of the complete problem .
much of our analysis will focus on constructing such mip formulations of @xmath41 , which we denote the _ pairwise flp_. however , in section [ sec : objective ] and section [ ss : multibox - cuts ] we consider the strengthening of the final formulation by explicitly considering the objective function and larger collections of boxes , respectively , when formulating the disjunctive constraint .
mip formulations for unions of polyhedra such as @xmath41 can be divided into _ extended formulations _ that uses both continuous and @xmath1 auxiliary variables ( * ? ? ?
* section 5 ) and _ non - extended ( projected ) formulations _ that only use the @xmath1 auxiliary variables that are strictly necessary to build a valid formulation ( * ? ? ?
* section 6 ) .
note that , in general , it is not possible to construct a mip formulation for unions of polyhedra in the original space , so some ( @xmath1 ) variables are needed for representability .
standard extended formulations by @xcite and @xcite have the desirable property that their linear programming ( lp ) relaxations have extreme points that naturally satisfy the integrality requirements on the @xmath1 auxiliary variables ; we call such formulations _ integral _ or _ ideal_. in contrast , non - extended formulations often fail to be ideal , but can be much smaller .
for instance , the following proposition shows a non - extended formulation of @xmath41 obtained through the classical _ big-@xmath43 _ approach .
[ observation : bigm ] the following is a formulation for @xmath41 : [ eqn : unary - bigm ] @xmath44 formulation only uses four @xmath1 auxiliary variables ( and no continuous auxiliary variables ) and is about four times smaller than the standard ideal extended formulation for @xmath41 ( see appendix [ extendedformulationsec ] ) . while formulation is not guaranteed to be ideal , its smaller size can still result in a computational advantage over the ideal extended formulation .
in addition , using various techniques it is sometimes possible to strengthen non - extended formulations considerably .
for this reason , we concentrate on constructing non - extended formulations for the flp by using three techniques : ( 1 ) the flexible use of @xmath1 auxiliary variables provided by the _ embedding formulations _
approach , ( 2 ) alternative definitions of the disjunctive constraints @xmath45 , and ( 3 ) the consideration of various common linear inequalities when building the disjunctions .
we now provide some simple examples of applying these techniques to build mip formulations for disjunctive sets .
we begin our description of the embedding formulation approach of @xcite by re - interpreting as a formulation for the _ embedding _ of @xmath41 in a higher dimensional space .
indeed , @xmath46 is feasible for if and only if it belongs to the embedding of @xmath47 into @xmath48 given by @xmath49 where @xmath50 is the @xmath51-th unit vector ( @xmath52 for @xmath53 and @xmath54 ) .
we say that representation _ embeds _
@xmath41 , which lives in the space of the @xmath55 variables , into the space of the @xmath46 variables .
it achieves this by pairing each of the four polyhedra @xmath56 with a unique binary vector @xmath50 , and so _ encoding _ the disjunctive constraint .
any valid formulation for implies a valid formulation for @xmath41 , since @xmath41 is the orthogonal projection of @xmath57 onto the @xmath55 variables .
however , representation also makes explicit the role of the @xmath58 variables : @xmath59 implies that @xmath60 . in other words , the possible values @xmath61 of @xmath58 _ encode _ the selection among the polytopes @xmath56 .
the key for the flexibility of the embedding approach is noting that this encoding can use any family of pairwise distinct @xmath1 vectors in place of the unit vectors @xmath50 .
the following definition formalizes this approach in our context , where we explicitly separate the disjunctive constraint @xmath62 from the common constraints @xmath63 , which must be satisfied by all branches of the disjunction .
[ embeddingdef ] let @xmath64 , @xmath65 for each @xmath66 , @xmath67{\mbox{\normalfont\tiny\sffamily def}}}{=}}}{\left\{x\in { \mathbb r}^d\,:\ , a^0 x\leq b^0\right\}}$ ] be a polyhedron , @xmath68 $ ] be a disjunction on @xmath69 , @xmath70{\mbox{\normalfont\tiny\sffamily def}}}{=}}}{\left\{x\in q\,:\,a^k x \leq b^k \right\}}$ ] for @xmath71 be
the family of polyhedra obtained by combining @xmath63 and @xmath62 , and @xmath72 be an encoding composed of _ pairwise distinct _ @xmath1-vectors .
if the recession cones @xmath73 for all @xmath74 , then a non - extended ( linear ) mip formulation for @xmath75 , or , equivalently for @xmath76 , is any ( linear ) mip formulation for the embedding @xmath77{\mbox{\normalfont\tiny\sffamily def}}}{=}}}\bigcup_{k=1}^k p^k \times \{h^k\}$]that uses only @xmath78 variables .
standard formulation approaches are recovered when choosing the unit vectors @xmath79{\mbox{\normalfont\tiny\sffamily def}}}{=}}}\{{{\bf e}}^k\}_{k=1}^k \subseteq \{0,1\}^k$ ] , which we denote the _ unary encoding _ , as it uses one bit per branch of the disjunction .
for example , we have that is a formulation for @xmath80 with @xmath81 .
however , the real flexibility comes from the possibility of non - unary encodings , as the specific assignment of codes to branches of the disjunctions does not change the structure of the formulation .
for instance , to obtain a valid formulation for @xmath82 with @xmath83 we simply need to interchange @xmath84 and @xmath85 in .
in contrast , for other types of encodings the specific assignment can be significant in terms of the complexity of the resulting embedding object and formulations ( e.g. see section [ sec : binaryform ] and @xcite ) .
deriving ideal non - extended formulations for embeddings with any encoding can be done using a geometric construction introduced in @xcite .
however , such construction can be hard to analyze , and many choices of encodings may naturally have very large ideal formulations ( i.e. many inequalities ) .
fortunately , non - extended formulations can also be constructed using ad - hoc approaches or through simple constructions such as the generalization of the big-@xmath43 approach to arbitrary encodings introduced in @xcite . in the coming sections
we will see how this generic approach can be used to construct a range of formulations for our disjunctive set .
in particular , varying the ingredients @xmath63 , @xmath62 , and @xmath86 lead to different embedding objects @xmath87 , which in turn will necessitate different formulations . for examples beyond the flp where varying the encoding choice @xmath86 results in different formulations we refer the reader to @xcite and @xcite . in the following subsections , we provide such examples for varying inputs @xmath62 and @xmath63 .
the new formulation for the flp proposed in section [ sec : refined - disjunction ] hinges on a logical refinement of the disjunction @xmath31 that removes many redundant solutions from the resulting formulation .
to illustrate this idea , we provide a simple example , which is independent of the flp . consider the disjunctive constraint @xmath88{\mbox{\normalfont\tiny\sffamily def}}}{=}}}{\left[x_1+x_2\leq 1\right ] }
\vee { \left [ x_2\leq x_1\right]}$ ] and common linear constraints @xmath89{\mbox{\normalfont\tiny\sffamily def}}}{=}}}{\left\{x\in { \mathbb r}^2\,:\ , 0\leq x_i \leq 1 \quad \forall i\in \llbracket 2 \rrbracket \right\}}$ ] , for which @xmath90{\mbox{\normalfont\tiny\sffamily def}}}{=}}}\{x \in q_1 : d_1^a\}$ ] is depicted in figure [ fig1 ] . because the two alternatives of @xmath91 intersect , we can define an alternative disjunction @xmath92{\mbox{\normalfont\tiny\sffamily def}}}{=}}}{\left [ x_1+x_2\leq 1 , x_1\leq x_2\right ] } \vee { \left [ x_1+x_2\leq 1 , x_2\leq x_1\right]}\vee { \left[x_1+x_2\geq 1 , x_2\leq x_1\right]},\ ] ] for which @xmath93 . using @xmath94 instead of @xmath95 could lead to larger formulations , since more branches on the disjunction will require longer codes to satisfy the distinctness property .
however , it also reduces redundancy or symmetry , phenomena which are known to reduce the effectiveness of mixed - integer solvers .
in particular , we note that the point @xmath96 satisfies both branches of @xmath91 , but only one branch of @xmath97 . using the embedding approach for some encoding @xmath86 ,
this gives to two feasible points in @xmath98 which correspond to @xmath96 and differ only in their assigned code .
one advantage of the embedding framework as described in definition [ embeddingdef ] is that it allows us to exploit the fact that the flp has many constraints beyond the disjunctive constraint @xmath31 . in particular ,
we can pick which constraints we include in the ground set @xmath63 to combine with the disjunctive constraint @xmath62 and build a formulation , and which constraints are added after the formulation process . such choice can significantly change the strength of the final formulation . to illustrate this effect
consider the following simple example , which is independent of the flp .
take the disjunctive constraint @xmath99{\mbox{\normalfont\tiny\sffamily def}}}{=}}}{\left [ x_1+x_2\leq 1\right ] } \vee { \left [ 1+x_2\leq x_1\right]}$ ] and the set of linear inequalities [ alllinearineq ] @xmath100 suppose we want to construct a mip formulation for @xmath101{\mbox{\normalfont\tiny\sffamily def}}}{=}}}{\left\{x\in { \mathbb r}^2\,:\,\eqref{alllinearineq } , d_2\right\}}$ ] depicted by the dark shaded region in figure [ fig3 ] . if we let @xmath102{\mbox{\normalfont\tiny\sffamily def}}}{=}}}{\left\{x\in { \mathbb r}^2\,:\ , \eqref{linearineq1 } \right\}}$ ] we have that both @xmath103 and @xmath104 are bounded and hence satisfy the conditions of definition [ embeddingdef ]
. then , we can at first ignore linear inequality and construct a formulation for @xmath105 ( depicted by the and light shaded region in figure [ fig3 ] ) and then impose on the resulting formulation . for instance , an ideal formulation of @xmath106 is given by [ partialform ] @xmath107 a formulation of @xmath108 is then given by and . however , a second option is to include all inequalities into @xmath109{\mbox{\normalfont\tiny\sffamily def}}}{=}}}{\left\{x\in { \mathbb r}^2\,:\ , \eqref{linearineq1 } \text{--}\eqref{linearineq}\right\}}$ ] and directly construct a formulation of @xmath110 .
for instance , an ideal formulation of @xmath111 is given by with strengthened to @xmath112 we can check that @xmath113 is feasible for the lp relaxation of and , but it does not satisfy . hence , the formulation obtained by considering all common linear inequalities is stronger that the one obtained by first ignoring .
a similar strengthening effect can occur when auxiliary variables and linear inequalities used to model other aspects of a mathematical programming problem are included in the common constraints .
this is the case with the flp and the nonlinear objective function , which may be linearized with auxiliary variables and constraints . for a simple motivating example ,
let @xmath114{\mbox{\normalfont\tiny\sffamily def}}}{=}}}{\left[x_1\leq 1\right ] } \vee { \left [ x_1 \geq 3\right]}$ ] , @xmath115{\mbox{\normalfont\tiny\sffamily def}}}{=}}}{\left\{x_1\in { \mathbb r}\,:\ , 0\leq x_1\leq 4\right\}}$ ] and @xmath116{\mbox{\normalfont\tiny\sffamily def}}}{=}}}{\left\{x_1\in q_3^a\,:\ , d_3\right\}}$ ] , and suppose we want to solve @xmath117 .
an ideal formulation for @xmath118 is given by @xmath119 which together with a standard lp modeling trick to linearize the absolute value in the objective leads to the mip formulation of the complete problem given by @xmath120 alternatively , we could instead include the linearization trick in the common constraints to obtain @xmath121{\mbox{\normalfont\tiny\sffamily def}}}{=}}}{\left\{{\left(x_1,y_1\right)}\in { \mathbb r}^2\,:\ , 0\leq x_1\leq 4,\quad x -2 \leq y_1,\quad -x_1 + 2 \leq y_1 \right\}}$ ] and @xmath122{\mbox{\normalfont\tiny\sffamily def}}}{=}}}{\left\{{\left(x_1,y_1\right)}\in q_3^b\,:\ , d_3\right\}}$ ] , depicted in figure [ fig4 ] . an integral formulation for @xmath123 is given by plus @xmath124 which leads to the mip formulation of the complete problem given by @xmath125 we can check that the optimal value of the lp relaxation of is equal to one . in contrast
, we can also check that the optimal value of the lp relaxation of is zero .
that is , we have constructed a stronger mip formulation for minimizing a nonlinear objective over a union of polyhedra by directly including the linearization of the objective in our construction procedure . given that incorporating additional structure in the ground set can allow us to construct stronger formulations , it seems at first that the optimal approach will be to simply add all constraints .
however , this can quickly lead to embedding objects @xmath87 that are very complex or difficult to study ; if @xmath63 is restricted to some minimal `` interesting '' substructure , we will see that we are better equipped to study and construct strong formulations .
we start by analyzing a simple , yet nontrivial , substructure for which we are able to construct a strong ( i.e. ideal ) formulation .
take @xmath126{\mbox{\normalfont\tiny\sffamily def}}}{=}}}\left\{(c_i ,
c_j,\ell_i,\ell_j ) \in { \mathbb{r}}^8 : \eqref{eqn : sitb},\;\eqref{eqn : lowerbound } \right\}$ ] ; that is , the set that imposes that the boxes lie completely on the floor and lower bounds on the box widths . using this set of common constraints , disjunction @xmath31 and the unary encoding we can construct the following small ideal formulation . throughout , we will use the notation @xmath127 as enumeration over the two orderings @xmath128 and @xmath129 .
[ thm : unary ] the following is a formulation for @xmath130 : [ eqn : unary - formulation ] @xmath131 if @xmath132 for both @xmath21 , then this formulation is ideal . see appendix [ app : unary - proof ] .
we dub the _
unary formulation_. however , we do not use the same naming convention for the binary variables as in formulation .
instead we rename @xmath84 , @xmath85 , @xmath133 and @xmath134 to @xmath135 , @xmath136 , @xmath137 and @xmath138 .
the reason for this is that the @xmath1 variables from have the nice interpretation that @xmath139 .
this interpretation forms the basis for the flp2 formulation in @xcite and @xcite .
in fact , the unary formulation is very similar to flp2 , but with the addition of the tightened stay - on - the - floor constraints .
in the sequel we use similar naming conventions for the @xmath1 variables when they have helpful interpretations .
the unary encoding uses codes of four bits to differentiate between four choices . if we instead use a binary encoding , we only need two bits ( i.e. codes of length two ) to impose this same decision .
in contrast to unary encodings , the specific assignment of codes to branches for binary encodings can result in significantly different formulations @xcite .
however , because of symmetry , for binary encodings of length two we may restrict our attention to two possible choices .
the first encoding corresponds to the unique ( up to symmetry ) gray code @xcite with two bits given by @xmath140{\mbox{\normalfont\tiny\sffamily def}}}{=}}}\left\ { ( 0,0 ) , ( 1,0 ) , ( 0,1 ) , ( 1,1 ) \right\}$ ] , and the second corresponds to the codes @xmath141{\mbox{\normalfont\tiny\sffamily def}}}{=}}}\left\ { ( 0,0 ) , ( 1,1 ) , ( 1,0 ) , ( 0,1 ) \right\}$ ] .
both choice of codes and their corresponding encodings can be used to reinterpret existing formulations from the literature .
the following proposition shows that the big-@xmath43 approach of @xcite can be used to construct a simple formulation for the gray encoding , which can be seen as the basis of formulation flp - sp introduced in @xcite .
[ graybigm ] a valid formulation for @xmath142 is : [ eqn : gray - binary - formulation ] @xmath143 see appendix [ app : gray - bigm ] .
formulation flp - sp is obtained from by adding the `` sequence - pair '' inequalities for the @xmath0-box formulation introduced in @xcite , which , while strictly breaking validity , can reduce redundancy in the feasible region set without removing any valid layouts . for completeness , we present the sequence pair inequalities in appendix [ app : sequence - pair ] .
we will see in section [ sec : computations ] that the flp - sp is the most competitive formulation from the literature on our computational benchmarks .
if instead we attempt to construct a formulation for @xmath144 , we can easily reconstruct the bldp1 formulation from @xcite , which we present in appendix [ app : bldp1 ] . while the disjunction @xmath45 is sufficient to enforce that @xmath11 and @xmath23 do not overlap
, its simplicity has a downside when used in a mip framework .
the disjunction is not sufficiently refined in the sense that there exist many feasible layouts that satisfy multiple branches at once .
for example , in figure [ fig:8-configurations ] , we see that @xmath11 precedes @xmath23 in both the @xmath24 and @xmath25 directions .
therefore , in any embedding constructed using @xmath45 , there exist two points that project down to the same layout ( that is , they differ only in their assigned codes ) . in practice , this redundancy can hamper the progress of branch - and - bound solvers , which must explicitly enumerate these solutions ( and all nodes preceding them in the tree ) to prove optimality .
precedes @xmath23 , @xmath145 , and @xmath146 in direction @xmath25 .
it also precedes @xmath23 in direction @xmath24 , does not precede @xmath145 in direction @xmath24 , and both precedes and does not precede @xmath146 in direction @xmath24 .
+ ( right ) the eight branches of the disjunction @xmath147 , illustrated via the relative position of @xmath23 to @xmath11.,title="fig : " ] precedes @xmath23 , @xmath145 , and @xmath146 in direction @xmath25 .
it also precedes @xmath23 in direction @xmath24 , does not precede @xmath145 in direction @xmath24 , and both precedes and does not precede @xmath146 in direction @xmath24 .
+ ( right ) the eight branches of the disjunction @xmath147 , illustrated via the relative position of @xmath23 to @xmath11.,title="fig : " ] to help remove this redundancy from the feasible set , we present a refined disjunction that is logically equivalent to @xmath31 . in definition
[ defn : precede ] , we presented a linear inequality that enforces that @xmath11 precedes @xmath23 . for our refined disjunction , we will need a description of the opposite .
we say that @xmath11 _ does not precede _
@xmath23 ( denoted by @xmath148 ) if @xmath149 .
referring back to figure [ fig:8-configurations ] , we see that @xmath11 precedes @xmath145 in direction @xmath25 , but does not precede @xmath145 in direction @xmath24 ( and vice versa ) .
note in particular that , if @xmath150 ( as with @xmath11 and @xmath146 in figure [ fig:8-configurations ] ) , we have that both @xmath27 and @xmath148 simultaneously . with the two definitions , we can construct a refinement of @xmath45 given by @xmath151{\mbox{\normalfont\tiny\sffamily def}}}{=}}}\bigvee_{k=1}^8 br^k_{i , j}$ ] where @xmath152 we have taken a refinement of @xmath31 by splitting the regions satisfying two branches at once into the new branches @xmath153 , and @xmath154 , and shrinking the other branches to exclude these new regions .
see figure [ fig:8-configurations ] for an illustration . with 8 branches in the disjunction
we need codes of length at least @xmath155 .
however , in lieu of chasing the formulation with the smallest number of @xmath1 variables ( i.e. a binary formulation ) , we instead take the encoding @xmath156{\mbox{\normalfont\tiny\sffamily def}}}{=}}}\left\ { { { \bf e}}^1 , { { \bf e}}^1+{{\bf e}}^2 , { { \bf e}}^2 , { { \bf e}}^2+{{\bf e}}^3 , { { \bf e}}^3 , { { \bf e}}^3+{{\bf e}}^4 , { { \bf e}}^4 , { { \bf e}}^4+{{\bf
e}}^1 \right\ } \subseteq \{0,1\}^4 $ ] .
intuitively , we have taken the codes from the unary encoding for the regions @xmath157 , and @xmath158 , and taken the codes for the new branches as the sum of the codes assigned to the two branches in @xmath31 the region satisfies .
for example , we get @xmath159 by taking the intersection of @xmath160 ( code @xmath161 ) and @xmath162 ( code @xmath163 ) , so we take the corresponding code as @xmath164 .
we see in the following proposition that by `` shadowing '' the unary embedding in this way , we are able to construct a formulation for @xmath165 that is very similar to the ideal formulation for @xmath166 .
[ prop : refined ] the following is a valid formulation for @xmath165 : [ eqn : refined - unary - formulation ] @xmath167 see appendix [ app : refined ] .
we dub formulation the _ refined unary formulation _ , and conjecture that it is the strongest possible for @xmath165 .
formulation is ideal .
finally , we note that our choice of codes induce the following nice interpretation for the @xmath1 variables : [ eqn : z - variable - implication ] @xmath168 this is a refinement of the interpretation for unary formulation as the variables for the unary formulation do not satisfy .
in section [ sec : formulations ] we have seen how the embedding approach can be used to construct valid formulations for substructures of the pairwise flp . in particular , we chose a subset of variables and constraints ( @xmath169 ) for which the analysis is tractable . however , in section [ sec : common - constraints ] we have seen that incorporating more of the common constraint structure in the ground set @xmath63 can allow us to construct much stronger formulations . therefore , in this section we explore embeddings of more complex substructures @xmath63 . however , since the facial structure of the embedding objects grows considerably more complex , we only focus on constructing valid inequalities for these new embeddings . in the remaining sections , we will express all inequalities for the refined unary encoding .
fortunately , as described in @xcite , it is sometimes possible to translate valid inequalities between different encodings .
we now present a self contained description of such translations for the flp and refer the reader to @xcite for a more general treatment .
[ prop : map - from - refined ] let @xmath170{\mbox{\normalfont\tiny\sffamily def}}}{=}}}\{(c_i , c_j,\ell_i,\ell_j ) \in { \mathbb{r}}^{8 } : \eqref{eqn : sitb } , \eqref{eqn : area } , \eqref{eqn : bounds}\}$ ] and consider an inequality @xmath171 with @xmath172 that is valid for @xmath173 .
then * @xmath174 is valid for @xmath175 , and * @xmath176 is valid for @xmath177 , where @xmath178 is the affine mapping that identifies @xmath179 with @xmath180 and @xmath181{\mbox{\normalfont\tiny\sffamily def}}}{=}}}\begin{pmatrix } -1 & -1 \\ 1 & -1 \\ 1 & 1 \\ -1 & 1 \end{pmatrix }
w + \begin{pmatrix } 1 \\ 0 \\ -1 \\ 0 \end{pmatrix}.\ ] ] is the affine mapping that identifies @xmath182 with @xmath183 . we prove the second , as the first follows in the same way . consider a feasible layout @xmath184 and take the corresponding feasible codes @xmath185{\mbox{\normalfont\tiny\sffamily def}}}{=}}}\left\{w \in gb^4 : ( \hat{c},\hat{\ell},w ) \in { \operatorname{em}}(q^{flp},d^4,gb^4 ) \right\}$ ] .
choose some @xmath186 .
then there exists some code @xmath187 such that @xmath188 and @xmath189 .
therefore , we have that , since @xmath172 , @xmath190 therefore , the given inequality holds for @xmath177 . [ prop : map - to - refined ] if inequality @xmath191 is valid for @xmath175 , then @xmath171 is valid for @xmath173 if either @xmath192 or @xmath193 .
consider a feasible layout @xmath184 and take the corresponding feasible codes @xmath194{\mbox{\normalfont\tiny\sffamily def}}}{=}}}\left\{u \in u^4 : ( \hat{c},\hat{\ell},u ) \in { \operatorname{em}}(q^{flp},d^4,u^4 ) \right\}$ ] and @xmath195{\mbox{\normalfont\tiny\sffamily def}}}{=}}}\left\{z \in c^8 : ( \hat{c},\hat{\ell},z ) \in { \operatorname{em}}(q^{flp},d^8,c^8 ) \right\}$ ] . from the construction of the codes @xmath196 discussed in section [ sec : refined - disjunction ] , for each @xmath197 , there exist some @xmath198 such that @xmath199 and @xmath200 ( if @xmath201 , then @xmath202 ) .
for example , if @xmath203 , we see that @xmath204 and @xmath205 . therefore , if @xmath206 , we have that @xmath207 if @xmath208 , then the same inequality holds with @xmath209 in place of @xmath210 .
therefore , any inequality valid for @xmath211 is valid for @xmath212 . in the previous section we chose the base set @xmath169 such that only lower bounds on the widths were included in the formulation
this was to make the formulation analysis tractable , but enforcing the aspect - ratio constraints via naturally includes upper bounds as well .
therefore , we can consider the set @xmath213 induced by @xmath214{\mbox{\normalfont\tiny\sffamily def}}}{=}}}\{(c,\ell ) \in q^{lb } : \eqref{eqn : upperbound}\}$ ] .
[ prop : ub - cuts ] for any assignments @xmath215 and @xmath216 , then @xmath217 is a valid inequality for @xmath213 . if @xmath218 , @xmath219 is valid for both @xmath211 and @xmath212 .
see appendix [ app : ub - cuts ] .
the objective is nonlinear but is straightforward to linearize in the usual fashion with auxiliary variables @xmath220 and the constraints @xmath221 even though this type of linearization is a very common mip formulation technique , it is often not incorporated into polyhedral studies explicitly . to do this for the pairwise flp , consider the augmented base set @xmath222 .
the resulting encoding @xmath223 leads to a collection of inequalities that serve to lower bound the auxiliary objective variables @xmath224 .
[ prop : obj - cuts ] choose @xmath21 and some assignment @xmath216
. then the following are valid inequalities for @xmath223 : @xmath225 see appendix [ app : obj - cuts ] .
note that we are now adding both constraints and variables to our ground set @xmath226 .
these inequalities are especially significant , since they explicitly incorporate the objective function , and the mip relaxation lower bounds for the flp are quite poor ( see section [ sec : lowerbounds ] ) .
thus far we have only considered representations for @xmath41 , the relationships between a single pair of boxes . in this section
we address how to use the results derived for the pairwise formulations to construct strong formulations for the original @xmath0-box floor layout problem .
since all the constraints for the flp involve at most two boxes , it suffices to consider each pair of boxes separately , construct a pairwise formulation , and identify all repeated variables across these pairwise formulations as follows .
[ prop : pairwise - to - nbox ] consider pairwise formulations @xmath227 for each pair of boxes @xmath9 over the variables @xmath228 .
if @xmath229{\mbox{\normalfont\tiny\sffamily def}}}{=}}}\sum_{(i , j ) \in { \mathscr{p } } } m_{i , j}$ ] , then @xmath230 is a formulation for @xmath231 . in particular , if we take the refined unary formulation for each pair of boxes , we construct the following formulation for the @xmath0-box flp . [ prop : pairwise - to - nboxc ] take @xmath232 .
then @xmath233{\mbox{\normalfont\tiny\sffamily def}}}{=}}}\left\ { ( c,\ell , v ) \in { \mathbb{r}}^{8n } \times \{0,1\}^{2n(n-1 ) } : ( c_i , c_j,\ell_i,\ell_j , v^{i , j } ) \in f^{ru}_{i , j } \ : \forall ( i , j ) \in { \mathscr{p}}\right\}\ ] ] is a valid formulation for @xmath231 . while this approach is sufficient to construct a valid formulation , constructing disjunctive and mip formulation for multiple pairs of boxes can lead to stronger formulations
however , such formulations can be significantly larger and/or more complicated .
for this reason we instead concentrate on identifying valid inequalities for such multi - pair or multi - box formulations to strengthen the single - pair formulation from propositions [ prop : pairwise - to - nbox ] and corollary [ prop : pairwise - to - nboxc ] . when working with more than two boxes at once , the notion of spatial transitivity appears ; that is , for any @xmath234 we have @xmath235 .
we can use this property to generalize many of the pairwise valid inequalities introduced thus far to the multi - box setting , in a similar way to section 3 of @xcite . consider a pair of boxes @xmath9 and take an arbitrary path @xmath236 , where @xmath237 and @xmath238 .
we define an affine function of the form @xmath239{\mbox{\normalfont\tiny\sffamily def}}}{=}}}1 + \sum_{\xi=1}^{m+1 } \left(z^s_{t^{\xi-1},t^\xi } - 1\right)$ ] .
our function enjoys the following property : @xmath240 this will function as an ( underestimator for the ) indicator function for when we have a particular chain of boxes @xmath241 along direction @xmath26 .
we can use this to extend the logic of the pairwise inequalities we have developed . for a simple example , if @xmath242 , then we know that @xmath11 and @xmath23 are separated in direction @xmath26 by at least the smallest width @xmath146 can take along that direction , and so @xmath243 .
this tightening can be exploited in the inequalities derived previously , leading a host of new valid inequalities for the multi - box flp .
[ prop : multibox - cuts ] consider the pair @xmath244 and an arbitrary path @xmath245 , where @xmath246 and @xmath247 and @xmath248 .
choose assignments @xmath215 and @xmath216 and define @xmath249{\mbox{\normalfont\tiny\sffamily def}}}{=}}}\sum_{\xi=1}^m lb^s_{t^\xi}$ ] .
then the following are valid inequalities for @xmath250 : @xmath251 first , we note that inequalities ( [ eqn : multi1]-[eqn : multi7 ] ) are variations on ( [ eqn : obj1]-[eqn : obj4],[eqn : ru - tight - sitb],[eqn : ru - tight - sitb],[eqn : ru - nonoverlap ] ) , respectively , with an additional @xmath252 term appearing . we can use to see that @xmath253 implies @xmath254 for all @xmath255 to take a telescoping sum and derive @xmath256 in the case where @xmath253 for all @xmath255 . combining this with property , we derive that @xmath257 is valid for any feasible solution for @xmath250 .
this can be used to directly derive ( [ eqn : multi5]-[eqn : multi7 ] ) .
for example , can be derived by summing the valid inequalities @xmath258 after noting that @xmath259 implies that @xmath260 . for ( [ eqn : multi1]-[eqn : multi4 ] )
, we first observe that , due to property , the case analysis in appendix [ app : obj - cuts ] will only differ in the case where @xmath253 for all @xmath255 .
however , under the assumption that @xmath261 for all @xmath262 , this also implies that @xmath263 .
therefore , we may use the tightened inequality ( * * ) in lieu of ( * ) in the case analyses , yielding the result .
proposition [ prop : multibox - cuts ] provides an exponential number of valid inequalities for the @xmath0-box flp . for small paths ( e.g. @xmath264 )
, these inequalities can be added to the formulation directly ; this is the approach we take in the computational trials .
we compare four formulations and four levels of cuts in the computational trials .
the unary formulation , denoted ` u ` , is based on the pairwise unary formulation ; this is a strengthened version of the flp2 formulation from @xcite .
the ` bldp1 ` formulation from @xcite is also tested ; see appendix [ app : bldp1 ] .
the third formulation is the sequence - pair formulation ( ` sp ` ) from @xcite , derived by adding global constraints to a formulation derived from ; see appendix [ app : sequence - pair ] .
finally , we compare with our new refined unary formulation , which we denote ` ru ` .
we note that we observe a slight computational advantage for using the simple stay - on - the - floor constraints rather than the tightened versions and , since we may aggregate them and add a single copy , rather than one for each pair @xmath9 ; we instead add these tightened constraints as valid inequalities .
we will compare each of these formulations with one of four levels of valid inequalities added to the formulation ( that is , they are not separated dynamically ) .
the first will be no valid inequalities .
the second will use the v2 and b2 families of inequalities appearing in @xcite ; the formulation name will be appended with ` + ` if these inequalities are added .
we present these inequalities in appendix [ app : literature - cuts ] for completeness .
the ` vi ` tag will be used for formulations with the new inequalities derived in this work added . in particular , we use ( [ eqn : obj1]-[eqn : obj4 ] ) , , and ( [ eqn : multi5]-[eqn : multi7 ] ) for paths @xmath265 . for
the ` ru ` formulation , we also add and .
adding all of these to the formulation proved impractical in the branch - and - bound setting , so we instead add an @xmath266 subset of these inequalities with ` vi ` .
finally , we consider adding ( an @xmath266 subset of ) the `` multi - box objective cuts '' ( [ eqn : multi1]-[eqn : multi4 ] ) for paths @xmath241 of length 3 , and denote this by appending a ` 3 ` .
there is a standard symmetry - breaking approach presented in @xcite that we will use in all the computational examples to follow ( except for the relaxation gap discussion in section [ sec : lowerbounds ] , where we will discuss the effect of the symmetry - breaking explicitly ) .
we present the symmetry - breaking scheme in appendix [ app : symmetry - breaking ] for completeness . for our benchmarks ,
we use the ` hp ` , ` apte ` , and ` xerox ` benchmarks from the mcnc benchmark collection @xcite .
additionally , we will use the ` armour62 - 1 ` and ` armour62 - 2 ` instances from @xcite , the ` bazaraa75 - 1 ` and ` bazaraa75 - 2 ` instances from @xcite , the ` camp91 ` instance from @xcite , the ` bozer91 ` instance from @xcite , and the ` bozer97 - 1 ` and ` bozer97 - 2 ` instances from @xcite instances .
to the best of our knowledge , none of these instances have been solved to optimality before in the literature . to construct the formulations and interface with the solver
, we use the jump algebraic modeling language from @xcite ; jump is written in the julia programming language ( see @xcite ) .
we performed the experiments on an intel i7 - 3770 3.40ghz linux workstation with 32 gb of ram .
all trials use cplex v12.6 with a maximum runtime of 4 hours .
we also performed trials with gurobi v6.0 , but the performance was not competitive .
we force cplex to use the linearization of the second - order cone constraints ( ` cpx_param_miqcpstrat ` set to 2 ) , as solving the nonlinear problem at the nodes was far slower .
all trials are performed with maximum aspect ratio @xmath267 .
the code used for these computational studies , as well as the benchmark instances in mps format are available at https://github.com/joehuchette/floor-layout .
first , we compare the lower bound produced by solving the continuous relaxation of the formulations , with and without valid inequalities added .
we present the relative gap percentage @xmath268 , where @xmath269 is the given relaxation lower bound and @xmath270 is the cost of the best known feasible solution .
note in particular that a relative gap percentage of @xmath271 implies that the relaxation bound of @xmath272 , which is the worst possible for the flp .
we observe that the formulations naturally fall into two groups with respect to the quality of their relaxation bound , and we summarize the results below ( we include a table in appendix [ app : relaxation - gap - table ] for completeness ) .
first , we observe that the `` two - bit '' formulations ( binary gray , bldp1 , flp - sp ) have a relaxation gap of @xmath271 , even with _ all inequalities _ discussed in the previous subsection added to the formulation ( ` + vi3 ` ) . with the symmetry - breaking constraints added to the formulation ,
the relaxation lower bound is no longer zero , and so the relaxation gap improves slightly ( mean 89.3% , with standard deviation @xmath273 ) .
the `` four - bit '' formulations ( unary and refined unary ) also produce a trivial relaxation gap of @xmath271 , but adding the valid inequalities helps improve the lower bound considerably ( mean @xmath274 with standard deviation @xmath275 ) . in particular , we can isolate the b2 inequalities from @xcite and as the crucial additions to the improvement in the gap .
we also compare the relative gap attained at the root node ( with respect to the best known feasible solution ) after cplex is able to apply advanced techniques such as general purpose cuts and preprocessing .
this improves the gap by roughly 1%-5% for most trials
. however , there is still an appreciable difference in gap between the `` four bit '' formulations ` u ` and ` ru ` and the `` two bit '' formulations ` sp ` and ` bldp1 ` .
a complete table is available in appendix [ app : relaxation - gap - table ] .
the rationale for introducing the refined unary formulation was that many feasible layouts will have multiple corresponding points in the encoding constructed using @xmath31 .
the refined partition removes many of these redundant solutions , which helps in the branch - and - bound setting , much in the same way symmetry breaking removes equivalent feasible solutions that would otherwise have to be explicitly enumerated in the optimization procedure .
qualitatively , we observe this change of behavior in figure [ fig : branching ] , where we compare the progress of the ` sp+ ` and ` ru ` formulations as a function of node count on the ` xerox ` benchmark .
that is , we compare both the upper and lower bound for both formulations ; when they are equal , the solver has proven optimality .
we see that the ` ru ` formulation requires fewer nodes to prove optimality , as expected . more broadly
, this illustration shows the typical trajectory when solving an instance of the flp : finding a good ( often near - optimal ) feasible solution early in the procedure , and then steadily improving the lower bound which is far from optimal , until the gap is finally closed and optimality is proven . while the ` ru ` formulation offers advantages when looking at the progress with respect to the node count , the advantage is not so clear - cut when looking at solution time .
table [ table : solve - time ] shows the solution time for each of our benchmarks and formulation and inequality combination .
first , we see that the ` u ` and ` bldp1 ` formulations are never competitive on the benchmark set . some variant of the sequence pair or refined unary formulations is the best performer on all instances , with ` sp ` winning six and ` ru ` winning three . however , four of the wins for the ` sp ` used the inequalities derived in this work .
in particular we note the ` bozer97 - 2 ` instance , which the sequence pair along with the new inequalities is able to solve the instance in the time limit , whereas none of the other approaches were able to close the gap beyond 3% .
that is , the instance is previously unsolved , and would have remained unsolved without the use of the techniques from this work . [ 1](#1% ) the results in table [ table : solve - time ] suggests that none of the approaches herein will be a clear winner on all instances , but that some combination of them can be used to tackle difficult problems . in particular , we recommend trying both the sequence pair and refined unary formulations , along with some subsets of the inequalities derived in this work to solve the flp instance most efficiently .
in this work , we presented a case study on systematically building strong formulations for disjunctive sets ; namely , for the floor layout problem .
we used the embedding approach of @xcite and @xcite to generate mip formulations of the flp and have observed how , by varying our inputs to the procedure , we are able to reconstruct all existing mip formulations for the problem , produce new formulations , and discover valid inequalities .
we also showed how valid inequalities generated for one formulation can often be translated to seemingly unrelated formulations .
finally , we presented computational results showing how the developed techniques can been used to solve previously unsolved benchmark instances .
the following corollary gives an extended formulation for @xmath41 using a standard approach by balas , jeroslow and lowe .
[ balascoro ] the following is an ideal extended formulation for @xmath41 : [ eqn : balasform ] @xmath276 this follows from proposition 4.2 in @xcite .
first , we note that is just formulation with a tightened form of constraints . for validity , first we want to show that the stay - on - the - floor constraints can be tightened to by the following case analysis .
consider @xmath277 , @xmath278 , and @xmath279 ; the other constraints follow analogously .
* reduces to the linear constraints in defining @xmath169 . *
the first inequality is unchanged . for the second
, we note that @xmath280 means that @xmath281 * same argument as the case before , but with the branch @xmath282 and the first inequality in . *
not feasible by . to show the formulation is ideal , we want to show that every extreme point of its relaxation has integral value for @xmath210 .
consider the following system , which can be thought of as the projection of the relaxation of onto just the @xmath25 variables ( the argument for @xmath24 variables is identical ) . under the assumption that @xmath283 , the following is full - dimensional : [ eqn : one - dimensional - unary ] @xmath284 take some arbitrary feasible @xmath285 where @xmath286 is fractional .
we wish to show that this is not an extreme point .
the argument for fractional @xmath287 follows in the same way . we will consider a partition of all possible cases in the following way : 1
. is active 1 .
is active 2 .
is not active 2 .
is not active 1 . and are both active 1 .
is active 2 .
is not active 2 . at most one of and
are active 1 .
is not active 2 .
is active note in particular that , since @xmath288 , constraint immediately implies that @xmath289 and that ( [ eqn:50]-[eqn:51 ] ) can not be active . in each of these cases
, we will argue that the solution is not extreme , either because there are not the requisite six constraints active , or because the selection leads to a contradiction .
also , note that and both being active implies @xmath290 under the assumptions on lower bounds ( @xmath291 ) and on @xmath292 fractional ; similarly for and together for @xmath293 .
therefore , these pairs can not be active together when these conditions are met .
we will use this observation in the following .
[ [ section ] ] first consider the case that and are active . clearly and can not be active
. then we must have and active as well : take their sum , which must hold with equality as it is equivalent to when @xmath294 .
however , this also implies that at most three of these four active constraints are linearly independent .
we can not have or active under our assumptions on the lower bounds ( take their sum with or , respectively ) .
the only remaining possibility is if , , and are all active .
however , then summing and and reducing leads to @xmath295 a contradiction .
therefore , the point is not extreme .
[ [ section-1 ] ] now assume that is not active and is .
this implies that at most one of and can be active , as their sum is equal to ( when @xmath294 ) as mentioned in the previous case . from our note above , at most one of and can be active at once .
given this , the only possibility that the point is extreme is if one of or , one of or , and , , and are all active . if and are both active in this setting , their sum implies @xmath296 a contradiction .
if and are simultaneously active in this setting , then the sum @xmath297 yields @xmath298 a contradiction .
therefore , is not active . if and are both active in this setting , then we get a similar result for @xmath299
this implies that we can have at most one of ( [ eqn:42]-[eqn:45 ] ) active , leaving us with too few active constraints for the point to be extreme .
[ [ section-2 ] ] now consider the case that @xmath300 ; clearly can not be active .
if , , and all are active , they imply @xmath301 which means that and can not both be active by the assumption on lower bounds .
if both and are active together , this implies that is active ; therefore , is not active .
moreover , the discussed active constraints have rank five at this point . by the notes above , and being active imply that and can not be active .
the ( six ) active constraints are linearly dependent , and so the point is not extreme .
the same argument holds if both and are active .
if and both are active , then necessarily is inactive ; summing and yields @xmath302 which says that can not be active from . presume then that is active , leaving us with , , , , , and active .
taking the resulting description for the extreme point represented by this system of active constraints yields @xmath303 , a contradiction .
the same argument holds if and are both active .
[ [ section-3 ] ] now assume that is not active and that and are active .
then at most one of and can be active , else we imply that @xmath304 which contradicts not being active .
the same holds for and .
also , we can not have both and active , since along with and active they imply @xmath305 therefore , at most five constraints are active , and we are not extreme .
[ [ section-4 ] ] now @xmath306 and w.l.o.g .
is not active .
our statement at the beginning implies that at most two of ( [ eqn:42]-[eqn:45 ] ) are active .
if is not active , this leaves at most five active constraints , and so the point is not extreme .
[ [ section-5 ] ] now assume that is not active but is . by the argument in 2.b.i , there are at most six constraints ( , , , , and two of ( [ eqn:42]-[eqn:45 ] ) ) that could be active .
in particular , , , and must be active at an extreme point , so presume they are .
assume for the first case that both and are active .
computing @xmath297 and reducing using the other active constraints yields @xmath307 a contradiction . alternatively , summing , , and yields @xmath308 also a contradiction .
similarly for , , and , as for , , and .
this exhausts all possible combinations of active constraints , and so the point is not extreme . [
[ piecing - together - direction - wise - formulations ] ] piecing together direction - wise formulations + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + now that we have established that the formulation , when restricted to a single direction , is ideal , it remains to show that the cartesian product of formulations for both directions , along with the restriction @xmath309 , is also ideal . to see this , consider a potential fractional extreme point @xmath285 for the relaxation of the original formulation .
then at least 12 active constraints at such an extreme point , one of which will be .
of the 11 remaining that must exist , there will be one direction for which there are at least six active constraints .
consider three cases . 1 .
the direction with six active constraints has a fractional component in @xmath210 .
then this implies an extreme point for the auxiliary system with fractional component , a contradiction .
2 . the direction with fractional component ( w.l.o.g .
@xmath25 ) has five active constraints . then those five active constraints , along with , induce a fractional extreme point for the auxiliary system for @xmath25 , a contradiction .
3 . the direction with fractional component ( w.l.o.g .
@xmath25 ) has fewer than five active constraints .
this implies that there are at least seven linearly independent active constraints for the auxiliary system for @xmath24 , a contradiction ( since its dimensionality is only six ) .
therefore , any fractional extreme point induces a fractional extreme point on , which we have shown is impossible , and so we are done .
while validity of most formulations in this work can be checked directly , the following generic big-@xmath43 formulations approach was instrumental for their construction .
we refer the reader to @xcite for a more detailed exposition of the approach , but include a proof of its validity for completeness .
[ thm : bigm ] take @xmath63 as a compact convex set , @xmath68 $ ] , and @xmath310 distinct vectors @xmath311 .
take any mip formulation @xmath312 for the set @xmath313 , and some affine functions @xmath314 such that @xmath315 where @xmath316{\mbox{\normalfont\tiny\sffamily def}}}{=}}}\{x \in q : a^sx \leq b^s\}$ ] .
then @xmath317 is a valid formulation for @xmath87 ( and , hence , for @xmath75 ) .
it is clear from the definition of @xmath62 that for any @xmath318 , there is some branch @xmath51 such that @xmath319 , and so @xmath320 is feasible for the mip formulation by the construction of the @xmath321 . to show that any feasible solution for the mip formulation lies in @xmath75 ,
consider some feasible @xmath322 for the mip formulation .
then @xmath323 and @xmath324 implies that @xmath325 for some @xmath51 .
then @xmath326 implying that @xmath24 satisfies the corresponding branch @xmath51 of the disjunction @xmath62 .
apply theorem [ thm : bigm ] with @xmath45 , @xmath327 , and @xmath328 the general proof technique is as follows .
first , we will construct the components needed to apply theorem [ thm : bigm ] : namely , a ground set describing shared constraints across all feasible layouts , a disjunction we are interested in modeling , a valid formulation for the codes , and some big-@xmath43 functions that encapsulate the logic between the codes and the branches of the disjunction .
this will leave us with a valid formulation for our set @xmath165 .
we will then do ad - hoc tightening of some of the resulting constraints , giving the system described in .
first , we choose the ground set @xmath169 and the disjunction @xmath147 .
we see that @xmath329{\mbox{\normalfont\tiny\sffamily def}}}{=}}}\left\{(z^y_{i , j},z^x_{i , j},z^y_{j , i},z^x_{j , i } ) \in \{0,1\}^4 : z^x_{i , j}+z^x_{j , i}+z^y_{i , j}+z^y_{j , i } \geq 1 , \ : z^x_{i , j}+z^x_{j , i } \leq 1 , \ : z^y_{i , j}+z^y_{j , i } \leq 1\right\}\ ] ] is a valid formulation for @xmath196 .
choose big-@xmath43 functions @xmath330 based on the disjunction @xmath147 in the following way .
take @xmath331 as the @xmath332-th clause defining @xmath333 in section [ sec : refined - disjunction ] ( recall that @xmath334 ) .
for example , @xmath335 . then take @xmath336 note that , when @xmath331 is a statement of the form `` @xmath337 precedes @xmath338 in direction @xmath26 '' , we get the same big-@xmath43 functions as appeared in the unary formulation for the same logic
. now apply theorem [ thm : bigm ] and recover the valid formulation @xmath339 , where @xmath340 note that , since the branches of the disjunction share constraints ( and corresponding big-@xmath43 functions @xmath330 ) , many of the resulting constraints will be equivalent and duplicates can be removed .
we now wish to tighten some of these constraints by lifting them in an ad - hoc manner . by the same argument as in the proof for theorem [ thm : unary ] , we may tighten to . to tighten the new constraints
, we can do a case analysis .
consider @xmath341 , @xmath342 , and @xmath343 ; the others follow analogously .
* reduces to the linear constraint @xmath344 , which follows from @xmath345 .
* we have that in this case @xmath346 and adding @xmath347 to this gives the desired inequality @xmath348 * in this case , we have that @xmath349 we can add @xmath347 to this to get @xmath350 the desired inequality . *
not feasible by .
when using the alternative codes @xmath351 , we may construct a formulation that is quite similar to the one presented in , but with slightly different big-@xmath43 terms : @xmath352 this formulation , with the addition of the area constraints , forms the basis for the bldp1 formulation from @xcite .
the sequence - pair formulation flp - sp from @xcite may be constructed from with the addition of global constraints on the 0/1 variables , based on observations made by @xcite . in particular , consider an @xmath0 box instance of the flp and the corresponding formulation derived from proposition [ prop : pairwise - to - nbox ] where each pairwise formulation @xmath227 is given by the gray binary formulation .
then the addition of the following constraints yields the flp - sp formulation : @xmath353 where notationally @xmath354{\mbox{\normalfont\tiny\sffamily def}}}{=}}}\begin{cases } w^{p , q}_k & p < q \\ 1-w^{p , q}_k & \text{o.w . }
\end{cases } \quad \forall k \in \{1,2\ } , p , q \in \llbracket n \rrbracket : p \neq q.\ ] ]
we prove by enumerating the possible values for the components of @xmath355 having support over the constraint , noting in particular that @xmath356 is always infeasible .
recall also that @xmath357 * sum the constraints @xmath358 and @xmath359 to get the constraint @xmath360 . * using and rearranging gives @xmath361 adding the constraint @xmath362 gives the desired result @xmath363 * want @xmath364 .
since @xmath365 , we must have that either @xmath366 or @xmath367 ; w.l.o.g .
choose the second .
then rearranging from gives @xmath368 * want that @xmath369 , which follows immediately from the upper bounds on @xmath370 .
we prove by enumerating the possible values for the components of @xmath355 having support over the constraint , noting in particular that @xmath356 is always infeasible .
* want to show that @xmath371 but since @xmath372 necessarily ( sum ) it suffices to show that @xmath373 , which follows immediately from summing constraints @xmath374 with @xmath375 to get that @xmath376 . applying this
also for @xmath377 and summing the resulting inequalities gives the result .
* want to show that @xmath378 we have from that @xmath379 ; adding the appropriate constraint in to this gives the result . *
same argument as the previous case .
* want to show that @xmath380 which follows immediately from and the fact that @xmath381 for any feasible solution . * want to show that @xmath382 rearranging gives @xmath383 now summing the relation from with one of gives @xmath384 summing these two derived inequalities with @xmath385 gives the desired result . *
want to show that @xmath386 take the sum of one of and the inequality from to derive @xmath387 furthermore , using our big-@xmath43 value , we have @xmath388 summing the two derived inequalities along with the lower bounds on @xmath370 gives the result .
* want to show that @xmath389 , which is immediate from . *
want to show that @xmath390 from an argument above , @xmath391 for this particular setting , and so we are done by summing this with @xmath392 implied by and using the lower bounds on @xmath370 . * want to show that @xmath393 which just follows from the fact that @xmath372 and @xmath381 . *
want to show that @xmath394 which follows immediately from the inequality ( valid for this particular setting for @xmath355 ) @xmath387 derived previously . *
same argument as the previous case .
we now present the b2 and v2 inequalities from @xcite in the notation used in the present work .
the b2 inequalities for the unary formulation are of the form @xmath395 the v2 inequalities for the unary formulation are @xmath396 which is equivalent with a potentially tightened coefficient @xmath397 . using proposition [ prop : map - to - refined ] and proposition [ prop : map - from - refined ] , these inequalities may be applied to all the formulations discussed in this work .
the symmetry - breaking described in @xcite works by restricting the possible relative layout between a single pair of components in a modification of the so - called _
position @xmath398 method _ from @xcite .
the scheme chooses a single pair @xmath399 ; in this work , we follow @xcite and choose @xmath400 .
we then may add the following constraints to the refined unary formulation : @xmath401 using proposition [ prop : map - from - refined ] , these inequalities may be applied to all the formulations discussed in this work .
.relative gap of the relaxation lower bound , with respect to the best known feasible solution .
group # 1 includes ` u ` , ` bldp1 ` , ` bldp1 + ` , ` sp ` , ` sp+ ` , ` sp+vi ` , ` sp+vi3 ` , and ` ru ` .
group # includes ` u+ ` , ` ru+vi ` , and ` ru+vi3 ` .
symmetry breaking from @xcite is added to group # 1 for comparison ( # w/ sb ) ; it does not affect the values for group # 2 . [ cols=">,^,^,^",options="header " , ] | this material is based upon work supported by the national science foundation graduate research fellowship under grant no .
1122374 and grant cmmi-1351619 . |
10 c. e. aguiar , y. hama , t. kodama , t. osada , nucl .
* a698 * , 639 - 642 ( 2002 ) . s. a. voloshin , phys .
* b632 * , 490 - 494 ( 2006 ) .
* a749 * , 287 - 290 ( 2005 ) . j. takahashi et al . phys .
lett . * 103 * , 242301 ( 2009 ) .
r. p. g. andrade , f. grassi , y. hama , w. -l .
qian , nucl .
* a854 * , 81 - 88 ( 2011 ) .
a. p. mishra , r. k. mohapatra , p. s. saumia , a. m. srivastava , phys .
rev . * c81 * , 034903 ( 2010 ) .
p. sorensen , j. phys .
g * g37 * , 094011 ( 2010 ) .
d. teaney , l. yan , [ arxiv:1010.1876 [ nucl - th ] ] .
s. a. voloshin , w. e. cleland , phys . rev . *
c53 * , 896 - 900 ( 1996 ) .
phys . rev .
* c54 * , 3212 - 3217 ( 1996 ) .
s. voloshin , r. lednicky , s. panitkin and n. xu , phys .
* 79 * , 4766 ( 1997 ) .
d. hardtke , s. a. voloshin , phys .
rev . * c61 * , 024905 ( 2000 ) .
m. a. lisa , s. pratt , r. soltz , u. wiedemann , ann .
nucl . part .
* 55 * , 357 - 402 ( 2005 ) .
b. zhang , c. m. ko , b. -a .
li , z. -w .
lin , phys .
rev . * c61 * , 067901 ( 2000 ) . | dipole , triangular , and higher harmonic flow that have an origin in the initial density fluctuations has gained a lot of attention as they can provide additional important information about the dynamical properties ( e.g. viscosity ) of the system .
the fluctuations in the initial geometry should be also reflected in the detail shape and velocity field of the system at freeze - out . in this
talk i discuss the possibility to measure such fluctuations by means of identical and non - identical particle interferometry .
the initial spatial conditions in heavy ion collisions are not azimuthally symmetric .
due to particle rescatterings it leads to the anisotropic flow
anisotropies in particle momentum distributions with elliptic flow being the well known example .
recently , a significant progress has been reached in understanding the role of the initial density fluctuations .
in particular it was realized that such fluctuations lead to odd harmonic anisotropic flow , studying of which brings new insights into dynamics of the system evolution .
below , i briefly summarize recent developments .
i apologize for not providing the full list of references many of which can be found in other proceedings of this conference . an importance of the initial state fluctuations for flow development has been noticed in @xcite , though the exact relation between fluctuation and the final event anisotropies were not clear . in
@xcite an observation was made that non - zero two - particle rapidity correlations ( as observed in @xmath0 collisions ) in conjunction with radial flow ( accepted part of the system evolution in @xmath1 collisions ) lead to a narrow in azimuth and long ranged in rapidity correlations .
such correlations were observed experimentally and called _
ridge_. this mechanism ( non - zero `` primordial '' rapidity correlations + radial flow @xmath2 ridge ) have been exploited later in several other models that try to address in more detail a ( more subtle ) question of the origin of the `` primordial '' rapidity correlations .
although the appearance of the ridge and anisotropic flow fluctuations look as totally unrelated phenomena , they appeared to be different views on the same thing
reaction of the system to the fluctuation in the initial density . in
@xcite it was shown explicitly that the fluctuations in the initial density distribution that extends over large rapidity range lead to the ridge structure in two particle correlations .
it was also noticed that after subtraction of the expected contribution from elliptic flow , two - bump structure appears on the away - side .
further studies @xcite with a single hot spot
embedded into otherwise smooth background appeared quite revealing .
contrary to an expectation @xcite that such a hot spot moved out by the radial flow would create a `` bump '' in azimuthal distribution it appeared that that the hot spot actually `` blocks '' the development of the radial flow and leads to a `` dip '' in particle distribution at the corresponding azimuth , accompanied by side - splashes from both sides .
remarkably , in terms of the correlation function the `` bump '' and the `` dip '' leads to the same structure - the ridge , as both means _ positive _ correlations . at the same time , the exact shape of the correlation function ( e.g. the strength of the third harmonic , which is important below ) is different .
it was noticed in @xcite that the fluctuating initial conditions generate anisotropic flow of different harmonics . that followed by understanding @xcite that the perturbations due to each of the hot spots can be treated independently , which allows to reformulate the problem from a different perspective
decompose the initial density into multipoles and study the system responce to each of the multipoles - the approach widely discussed at this conference as flow ( dipole , triangular , quadrangular , etc . )
fluctuations .
one can envision such an approach as rotating each of the event to a given harmonic symmetry plane , such that at the end one could have a smooth initial conditions but with a shape corresponding to a given harmonic .
the question addressed in this talk is if with the help of femtoscopy ( identical and non - identical two - particle correlations ) one can directly observe such triangular etc .
shapes of the system . in this study
i follow an approach of the first papers where the azimuthally sensitive femtoscopic analyses have been proposed @xcite , and blast wave and ampt models calculations . for typical values of parameters used in this work . ,
title="fig:",width=207 ] for typical values of parameters used in this work .
, title="fig:",width=192 ] for typical values of parameters used in this work .
, title="fig:",width=226 ] intensity interferometry ( identical particle femtoscopy ) , measures particle space - time distribution at freeze - out . due to the probe on - mass constraint
, the full space - time information can not be obtained , and one can access only the distribution in @xmath3 , where @xmath4 is the particle position at time @xmath5 ( after freeze - out ) , and @xmath6 is the particle velocity . for a gaussian source , the correlation function appears to be a gaussian c ( , ) 1 + _ i , j r_ij^2 q_i q_j , where @xmath7 is the pair relative momentum , @xmath8 is the particle average momentum , and @xmath9 are the hbt radii .
the deviation from a gaussian form can be studied by evaluating the higher moments of the distribution ( for the effect of non - gaussiness on hbt radii , see @xcite ) .
detailes of femtoscopic analyses can be found in a review @xcite .
we use a standard side - out - long system @xcite ( see figure [ fig : coordinates ] ) and concentrate on azimuthal dependence of the side , out , and side - out radii . for a _ stationary _ ( not expanding ) source
the radii azimuthal dependence can be expressed as r_side^2= = ^2+^2 - 2 , which has only @xmath10 harmonic .
higher harmonics azimuthal dependence appears only as deviation from the gaussian , e.g. in @xmath11 and @xmath12 for triangular and quadrangular shapes respectively .
the picture changes if one considers azimuthal variation in the expansion velocity . to study this effect
we employ a blast wave model .
we assume longitudinally boost invariant source and freeze - out at a constant temperature @xmath13 mev .
expansion velocity profile is parametrized in a form _ t(r,)=_t , max[1+a_boost ( n ) ] , + r_max = r[1+a_space ( n ) ] , with direction of the collective velocity being perpendicular to the line @xmath14 $ ] going through the emission point .
the parameter @xmath15 which corresponds to an average expansion velocity @xmath16 and @xmath17 gev . @xmath18 and @xmath19 parameters define the spatial eccentricity at freeze - out and modulation in the expansion velocity . for the first we assume that the system spatial eccentricity is reduced about two times during the system evolution , which leads to @xmath20
; @xmath19 can be estimated from @xmath21 dependence and is taken to be in the range 0.03 - 0.07 .
the azimuthal dependence of the obtained hbt radii , see figure [ fig : bwradii ] , indicates that the higher harmonics shape effects become clearly visible ( and measurable ) .
note the difference in the amplitudes of @xmath22 terms .
the radii modulation strongly depends on the parameters of the model , and the correspondng measurements will be very important in evaluating the velocity profiles at freeze - out and testing models . the ampt model @xcite has been extensively used in studying the effect of the initial state fluctuations .
we use this model to study au+au collisions at = 200 gev .
figure [ fig : amptradii ] shows the hbt radii variation as function of the azimuth relative to a given harmonic _
event plane _ which includes the effects of finite reaction plane resolution .
once again , the radii dependence on the azimuth is clearly seen . a detailed investigation of the side - out relative phases , dependece on the particle transverse momentum , etc .
requires additional investigations . , at @xmath23 .
, title="fig:",width=219 ] , at @xmath23 . , title="fig:",width=219 ] , at @xmath23 . ,
title="fig:",width=219 ] using the blast wave and ampt models we have demostrated the possibility of femtoscopic higher harmonics system shape analysis that promices new insights into the physics of heavy ion collisions . |
galois ( monodromy ) groups of problems from enumerative geometry were first treated by jordan in 1870 @xcite , who studied several classical problems with intrinsic structure , showing that their galois group was not the full symmetric group on the set of solutions to the enumerative problem .
others @xcite refined this work , which focused on the equations for the enumerative problem .
earlier , hermite gave a different connection to geometry , showing that the algebraic galois group coincided with a geometric monodromy group @xcite in the context of puiseaux fields and algebraic curves .
this line of inquiry remained dormant until a 1977 letter of serre to kleiman @xcite .
the modern , geometric , theory began with harris @xcite , who determined the galois groups of several classical problems , including many whose galois group is equal to the full symmetric group .
in general , we expect that the galois group of an enumerative problem is the full symmetric group and when it is not , the geometric problem possesses some intrinsic structure . despite this , there are relatively few enumerative problems whose galois group is known . for a discussion , see harris @xcite and kleiman @xcite .
the schubert calculus of enumerative geometry @xcite is a method to compute the number of solutions to , which are a class of geometric problems involving linear subspaces .
the algorithms of schubert calculus reduce the enumeration to combinatorics .
for example , the number of solutions to a schubert problem involving lines is a kostka number for a rectangular partition with two parts .
this well - understood class of problems provides a laboratory with which to study galois groups of enumerative problems .
the prototypical schubert problem is the classical problem of four lines , which asks for the number of lines in space that meet four given lines . to answer this ,
note that three general lines @xmath0 , and @xmath1 lie on a unique doubly - ruled hyperboloid , shown in figure [ f : four_lines ] .
@xmath2{pictures/4lines.eps } } \put(230 , 3){\red{$\ell_1 $ } } \put(288 , 85){\blue{$\ell_2 $ } } \put(266,147){\forestgreen{$\ell_3 $ } } \put ( 3,118){$\ell_4 $ } \put(263 , 34){\magenta{$m_1 $ } } \put(194,180){\magenta{$m_2 $ } } \thicklines \put(30,30.5){\white{\line(3,2){66.6 } } } \thinlines \put ( 22 , 25){\brown{$p$}}\put(30,30.5){\brown{\vector(3,2){66.6 } } } \end{picture}\ ] ] these three lines lie in one ruling , while the second ruling consists of the lines meeting @xmath0 , and @xmath1 .
the fourth line @xmath3 meets the hyperboloid in two points .
each of these points determines a line in the second ruling , giving two lines and which meet our four given lines . in terms of kostka numbers , enumerating the solutions
is equivalent to enumerating the tableaux of shape @xmath4 with content @xmath5 .
there are two such tableaux : @xmath6 when the field is the complex numbers , hermite s result gives one approach to studying the galois group by directly computing monodromy .
for instance , the galois group of the problem of four lines is the group of permutations which are obtained by following the solutions over closed paths in the space of lines @xmath7 .
rotating @xmath3 about the point @xmath8 ( shown in figure [ f : four_lines ] ) gives a closed path which interchanges the two solution lines @xmath9 and @xmath10 , showing that the galois group is the full symmetric group on the two solutions .
leykin and sottile @xcite followed this approach , using numerical homotopy continuation @xcite to compute monodromy for a few dozen so - called _ simple _ schubert problems , showing that in each case the galois group was the full symmetric group on the set of solutions .
( the problem of four lines is simple . )
this included a problem involving 2-planes in @xmath11 with 17589 solutions .
billey and vakil @xcite used elimination theory to compute lower bounds of galois groups , and they showed that a few enumerative problems on grassmannians with at most 10 solutions have galois group equal to the full symmetric group .
when the ground field is algebraically closed , vakil @xcite gave a combinatorial criterion , based on classical special position arguments and group theory , which can be used recursively to show that a galois group contains the alternating group on its set of solutions .
he used this and his geometric littlewood - richardson rule @xcite to show that the galois group of every schubert problem involving lines in projective space @xmath12 for @xmath13 had galois group that was at least alternating .
our main result is based on this observation .
[ th : one ] the galois group of any schubert problem involving lines in @xmath12 contains the alternating group on its set of solutions .
this nearly determines the galois group for a large class of schubert problems . in subsection
[ s : symmetric ] , we present two infinite families of schubert problems of lines , both of which generalize the problem of four lines , and show that each schubert problem in these families has galois group the full symmetric group on its set of solutions .
we conjecture this is always the case for schubert problems of lines .
any schubert problem involving lines in @xmath12 has galois group the full symmetric group on its set of solutions .
this conjecture ( and the result of theorem [ th : one ] ) does not hold for schubert problems in general .
vakil , and independently , derksen , gave a schubert problem in the grassmannian of 4-planes in 8-dimensional space whose galois group is not the full symmetric group on its set of solutions @xcite . in @xcite a schubert problem with such a deficient galois group
was found in the manifold of flags in 6-dimensional space .
both examples generalize to infinite families of schubert problems with deficient galois groups . by
vakil s criterion and a special position argument of schubert , theorem [ th : one ] reduces to a certain inequality among kostka numbers of two - rowed tableaux . for most cases ,
the inequality follows from a combinatorial injection of young tableaux . for the remaining cases , we use representation theory to rewrite these kostka numbers as certain trigonometric integrals . in this way , the inequalities of kostka numbers become inequalities of integrals , which we establish using only elementary calculus . in section [ s : one ]
we give some background on galois groups , vakil s criterion , and the schubert calculus of lines .
section [ s : two ] we explain schubert s recursion and formulate our proof of theorem [ th : one ] , showing that it follows from an inequality of kostka numbers , which we prove for most schubert problems .
we study kostka numbers in section [ s : kostka ] , giving combinatorial formulas for some and the integral formula .
the technical heart of this paper is section [ s : three ] in which we use these formulas for kostka numbers to establish the inequality when @xmath14 , which completes the proof of theorem [ th : one ] .
we summarize vakil s presentation in @xcite .
suppose that @xmath15 is a dominant morphism of ( generic ) degree @xmath16 between irreducible algebraic varieties of the same dimension defined over an algebraically closed field @xmath17 .
we will assume here and throughout that @xmath18 is generically separable in that the corresponding extension @xmath19 of function fields is separable .
consider the following subscheme of the fiber product @xmath20 where @xmath21 is the big diagonal .
let @xmath22 be a point where @xmath23 consists of @xmath16 distinct points , @xmath24 .
then the fiber of @xmath25 over @xmath26 consists of all permutations of those points , @xmath27 where @xmath28 is the symmetric group on @xmath16 letters .
the is the group of permutations @xmath29 for which @xmath30 and @xmath31 lie in the same component of @xmath25 .
the galois group @xmath32 is if it is not the full symmetric group @xmath28 , and it is if it is @xmath28 or its alternating subgroup .
vakil s criterion addresses how @xmath32 is affected by the galois group of a restriction of @xmath15 to a subvariety @xmath33 .
suppose that we have a fiber diagram @xmath34 where @xmath35 is the closed embedding of a cartier divisor @xmath36 of @xmath37 , @xmath37 is smooth in codimension one along @xmath36 , and @xmath38 is a generically separable , dominant morphism of degree @xmath16 .
when @xmath39 is either irreducible or has two components , we have the following . 1 .
if @xmath39 is irreducible , then there is an inclusion @xmath40 into @xmath32 .
2 . if @xmath39 has two components , @xmath41 and @xmath42 , each of which maps dominantly to @xmath36 of respective degrees @xmath43 and @xmath44 , then there is a subgroup @xmath45 of @xmath46 which maps surjectively onto each factor @xmath47 and which includes into @xmath32 ( via @xmath48 ) .
vakil s criterion follows by purely group - theoretic arguments including goursat s lemma . in case @xmath49 ,
if @xmath40 is at least alternating , then @xmath32 is at least alternating . in case @xmath50 , if @xmath51 and @xmath52 are at least alternating , and if either @xmath53 or @xmath54 , then @xmath32 is at least alternating .
[ r : vakil ] this criterion applies to more general inclusions @xmath35 of an irreducible variety into @xmath37 .
all that is needed is that @xmath37 is generically smooth along @xmath36 , for then we may replace @xmath37 by an affine open set meeting @xmath36 and there are subvarieties @xmath55 with each inclusion @xmath56 that of a cartier divisor where @xmath57 is smooth in codimension one along @xmath58 .
let @xmath59 ( or simply @xmath60 ) be the grassmannian of lines in @xmath61-dimensional projective space @xmath12 , which is an algebraic manifold of dimension @xmath62 .
a is the set of lines incident on a flag of linear subspaces @xmath63 , @xmath64 a asks for the lines incident on a fixed , but general collection of flags @xmath65 . this set of lines is described by the intersection of schubert varieties @xmath66 schubert @xcite gave a recursion for determining the number of solutions to a schubert problem in @xmath59 , when there are finitely many solutions .
the geometry behind his recursion is central to our proof of theorem [ th : one ] , and we will present it in subsection [ s : schubertrecursion ] .
[ r : special_only ] when @xmath67 , we may omit @xmath68 and write @xmath69 , which is a .
note that @xmath70 , the latter considered as a subvariety of @xmath71
. given @xmath72 and @xmath73 , if we set @xmath74 and @xmath75 , then @xmath76 the latter intersection taking place in @xmath77 . given a schubert problem , if @xmath78 and @xmath79 , for @xmath80 , then we may rewrite as an intersection in @xmath71 , @xmath81 we will show that it suffices to study intersections of special schubert varieties .
suppose that @xmath82 .
a general line in @xmath83 determines and is determined by its intersections with @xmath84 and with a fixed hyperplane @xmath45 not containing @xmath84 .
thus @xmath83 has dimension @xmath85 and so it has codimension @xmath86 in @xmath60 .
if @xmath87 are general linear subspaces of @xmath12 with @xmath88 for @xmath80 , and @xmath89 , then the intersection @xmath90 is transverse and therefore zero - dimensional . over fields of characteristic zero ,
transversality follows from kleiman s transversality theorem @xcite while in positive characteristic , it is theorem e in @xcite . by this transversality
, the number of points in the intersection does not depend upon the choice of general @xmath87 , but only on the numbers @xmath91 .
we call @xmath92 the of the schubert intersection . observe that we do not need to specify @xmath61
. given positive integers @xmath93 whose sum is even , set @xmath94 .
henceforth , a schubert problem will be denoted by a list @xmath95 of positive integers with even sum .
it is if @xmath96 ( this is forced by @xmath97 ) , which is equivalent to the numbers @xmath98 being the sides of a ( possibly degenerate ) polygon .
if @xmath95 is a valid schubert problem , then we set to be the number of points in a general intersection of type @xmath95 , and if @xmath95 is invalid , we set @xmath99 . this intersection number @xmath100 is a , which is the number of young tableaux of shape @xmath101 and content @xmath102 @xcite .
if @xmath95 is invalid , then there are no such tableaux , which is consistent with our declaration that @xmath103 .
these are arrays consisting of two rows of integers , each of length @xmath104 such that the integers increase weakly across each row and strictly down each column , and there are @xmath105 occurrences of @xmath106 for each @xmath107 . let be the set of such tableaux .
for example , here are the five young tableaux in @xmath108 , showing that @xmath109 . @xmath110{pictures / yt55.eps } } \put ( 0,11.5){{\small1}}\put(11.75,11.5){\small1 } \put(23.5,11.5){\small2}\put(35.25,11.5){\small2 } \put(47,11.5){\small3 } \put ( 0,-0.3){\small4}\put(11.75,-0.3){\small4 } \put(23.5,-0.3){\small5}\put(35.25,-0.3){\small5 } \put(47,-0.3){\small5 } \end{picture}}\qquad { \begin{picture}(58,25)(-2.9,-2.5 ) \put(-3.4,-3){\includegraphics[height=24pt]{pictures / yt55.eps } } \put ( 0,11.5){{\small1}}\put(11.75,11.5){\small1 } \put(23.5,11.5){\small2}\put(35.25,11.5){\small2 } \put(47,11.5){\small4 } \put ( 0,-0.3){\small3}\put(11.75,-0.3){\small4 } \put(23.5,-0.3){\small5}\put(35.25,-0.3){\small5 } \put(47,-0.3){\small5 } \end{picture}}\qquad { \begin{picture}(58,25)(-2.9,-2.5 ) \put(-3.4,-3){\includegraphics[height=24pt]{pictures / yt55.eps } } \put ( 0,11.5){{\small1}}\put(11.75,11.5){\small1 } \put(23.5,11.5){\small2}\put(35.25,11.5){\small3 } \put(47,11.5){\small4 } \put ( 0,-0.3){\small2}\put(11.75,-0.3){\small3 } \put(23.5,-0.3){\small5}\put(35.25,-0.3){\small5 } \put(47,-0.3){\small5 } \end{picture}}\qquad { \begin{picture}(58,25)(-2.9,-2.5 ) \put(-3.4,-3){\includegraphics[height=24pt]{pictures / yt55.eps } } \put ( 0,11.5){{\small1}}\put(11.75,11.5){\small1 } \put(23.5,11.5){\small2}\put(35.25,11.5){\small4 } \put(47,11.5){\small4 } \put ( 0,-0.3){\small2}\put(11.75,-0.3){\small3 } \put(23.5,-0.3){\small5}\put(35.25,-0.3){\small5 } \put(47,-0.3){\small5 } \end{picture}}\qquad { \begin{picture}(58,25)(-2.9,-2.5 ) \put(-3.4,-3){\includegraphics[height=24pt]{pictures / yt55.eps } } \put ( 0,11.5){{\small1}}\put(11.75,11.5){\small1 } \put(23.5,11.5){\small3}\put(35.25,11.5){\small4 } \put(47,11.5){\small4 } \put ( 0,-0.3){\small2}\put(11.75,-0.3){\small2 } \put(23.5,-0.3){\small5}\put(35.25,-0.3){\small5 } \put(47,-0.3){\small5 } \end{picture}}}\ ] ] it suffices to consider only certain types of schubert problems .
let @xmath95 be a ( valid ) schubert problem with @xmath111 and set @xmath112 .
suppose that @xmath113 are general linear subspaces with @xmath88 for @xmath80 .
since @xmath114 , the subspaces @xmath115 and @xmath116 are disjoint , and so every line @xmath117 in @xmath118 is spanned by its intersections with @xmath115 and @xmath116 . thus @xmath117 lies in the linear span @xmath119 , which is a proper linear subspace of @xmath12 .
let @xmath68 be a general hyperplane containing @xmath119 .
if we set @xmath120 for @xmath80 , then we have @xmath121 the latter intersection in @xmath122 . for @xmath123 , we have @xmath124 and so @xmath125 and if @xmath126 , then @xmath127 thus the righthand side of is a schubert problem of type @xmath128 , and so we have @xmath129 where @xmath130 .
we may also see this combinatorially : the condition @xmath111 implies that the first column of every tableaux in @xmath131 consists of a 1 on top of a 2 . removing this column gives a tableaux in @xmath132 , and this defines a bijection between these two sets of tableaux .
we say that a schubert problem @xmath95 is if @xmath133 for any @xmath134 . applying the previous procedure
recursively shows that every schubert problem may be recast as an equivalent reduced schubert problem .
given a schubert problem @xmath95 , let @xmath135 , and set @xmath136 which is a product of grassmannians , and hence smooth .
consider the of the schubert problem @xmath95 , @xmath137 the projection map @xmath138 to the first coordinate realizes @xmath139 as a fiber bundle of @xmath60 with irreducible fibers . as @xmath60 is irreducible , @xmath139 is irreducible .
let @xmath15 be the other projection .
its fiber over a point @xmath140 is @xmath141 in this way , the map @xmath15 contains all intersections of schubert varieties of type @xmath95 .
as the general schubert problem is a transverse intersection containing @xmath100 points , @xmath18 is generically separable , and it is a dominant ( in fact surjective ) map of degree @xmath100 . the galois group of the schubert problem of type @xmath95 is the galois group @xmath32 , where @xmath142 is the projection @xmath18 defined above .
[ r : reductionsok ] these two reductions , that a general schubert problem on @xmath60 is equivalent to one that only involves special schubert varieties ( remark [ r : special_only ] ) , and is furthermore equivalent to a reduced schubert problem ( subsection [ ss : reduced ] ) , do not affect the corresponding galois groups .
the reason is the same for both reductions , so we only explain it for that of subsection [ ss : reduced ] .
suppose that @xmath143 is a valid , but non - reduced schubert problem with @xmath144 .
let @xmath15 be the family of all instances of schubert problems of type @xmath95 as above .
fix a hyperplane @xmath145 and let @xmath146 be @xmath147 which is smooth . by setting @xmath148
we obtain a fiber diagram as in where @xmath149 is the family of all schubert problems of type @xmath150 in @xmath71 , as in subsection [ ss : reduced ] .
as @xmath39 is irreducible we have inclusions @xmath151 and therefore @xmath152 .
thus , if @xmath153 is at least alternating , then @xmath154 is at least alternating .
moreover , these galois groups coincide .
note that @xmath155 acts on @xmath12 and thus diagonally on @xmath37 and the orbit of @xmath36 is dense in @xmath37 .
this action extends to @xmath142 and to @xmath156 .
thus if @xmath157 is a point where @xmath158 consists of @xmath16 distinct points , @xmath24 and @xmath159 is a permutation , then the points @xmath30 and @xmath31 lie in the same connected component of @xmath160 if and only if they lie in the same connected component of @xmath25 .
we explain how a special position argument of schubert together with vakil s criterion reduces the proof of theorem [ th : one ] to establishing an inequality of kostka numbers . in many cases ,
the inequality follows from simple counting .
the remaining cases are treated in section [ s : three ] .
we also give two infinite families of schubert problems whose galois groups are the full symmetric groups .
we begin with a simple observation due to schubert @xcite .
[ l : schubert ] let @xmath161 be positive integers with @xmath162 , and suppose that @xmath163 are linear subspaces with @xmath164 for @xmath123 .
if @xmath165 and @xmath166 are in special position in that their linear span is a hyperplane @xmath167 , then @xmath168 where @xmath169 is any linear subspace of dimension @xmath170 of @xmath12 with @xmath171 .
furthermore , the intersection @xmath172 is generically transverse , is its irreducible decomposition , and the second intersection of schubert varieties is also generically transverse .
the reason for this decomposition is that if @xmath117 meets both @xmath165 and @xmath166 , then either it meets @xmath173 or it lies in their linear span ( while also meeting both @xmath165 and @xmath166 ) .
this lemma , particularly the transversality statement , is proven in ( * ? ? ?
* lemma 2.4 ) .
[ r : schubert_recursion ] suppose that @xmath95 is a reduced schubert problem .
set @xmath135 .
let @xmath87 be linear subspaces with @xmath174 which are in general position in @xmath12 , except that @xmath175 and @xmath176 span a hyperplane @xmath68 .
by lemma [ l : schubert ] we have @xmath177 where @xmath178 , and so @xmath179 has dimension @xmath180 .
the first intersection on the righthand side of has type @xmath181 and the second , once we apply the reduction of remark [ r : special_only ] , has type @xmath182 .
this gives schubert s recursion for kostka numbers @xmath183 as @xmath95 is reduced , the two schubert problems obtained are both valid .
this recursion holds even if @xmath95 is not reduced .
the first term in may be zero , for @xmath184 may not be valid ( in this case , @xmath185 ) .
we consider this recursion for @xmath186 .
the first tableau in ( [ eq11:tableaux ] ) has both 4s in its second row ( along with its 5s ) , while the remaining four tableaux have last column consisting of a 4 on top of a 5 .
if we replace the 5s by 4s in the first tableau and erase the last column in the remaining four tableaux , we obtain @xmath187{pictures / yt55.eps } } \put ( 0,11.5){{\small1}}\put(11.75,11.5){\small1 } \put(23.5,11.5){\small2}\put(35.25,11.5){\small2 } \put(47,11.5){\small3 } \put ( 0,-0.3){\small4}\put(11.75,-0.3){\small4 } \put(23.5,-0.3){\small4}\put(35.25,-0.3){\small4 } \put(47,-0.3){\small4 } \end{picture}}\qquad\qquad \kosfo{1}{1}{2}{2}{3}{4}{5}\qquad \kosfo{1}{1}{2}{3}{2}{3}{5}\qquad \kosfo{1}{1}{2}{4}{2}{3}{5}\qquad \kosfo{1}{1}{3}{4}{2}{2}{5}\ \,\ ] ] which shows that @xmath188 .
we sometimes use exponential notation for the sequences @xmath95 , e.g. @xmath189 . in subsection [ s : counting ] , we use this recursion to prove the following lemmas .
[ l : small_m ] suppose that @xmath95 is a valid schubert problem
. then @xmath190 and @xmath191 .
if @xmath192 or @xmath193 , then @xmath194 . if @xmath195 , then @xmath196 there are no reduced schubert problems with @xmath197 .
if @xmath95 is reduced and @xmath195 , then @xmath198 , and we have @xmath199 .
[ l : m=5 ] let @xmath200 with @xmath201 be even .
then @xmath202 we will use vakil s criterion and schubert s degeneration to deduce theorem [ th : one ] from a key combinatorial lemma .
a of a schubert problem @xmath91 is simply a listing of the integers @xmath91 in some order .
[ l : induction ] let @xmath95 be a reduced schubert problem involving @xmath203 integers .
when @xmath204 , it has a rearrangement @xmath91 such that @xmath205 and both terms are nonzero .
when @xmath206 , this@xmath207 is an equality with both terms equal to @xmath208 .
the proof of lemma [ l : induction ] will occupy part of this section and section [ s : three ] .
we use it to deduce theorem [ th : one ] , which we restate in a more precise form .
* theorem [ th : one ] . * _ let @xmath95 be a schubert problem on @xmath59 .
then @xmath154 is at least alternating . _
we use a double induction on the dimension @xmath61 of the ambient projective space and the number @xmath209 of conditions .
the initial cases are when one of @xmath61 or @xmath209 is less than four , for by lemma [
l : small_m ] , @xmath210 and the trivial subgroups of these small symmetric groups are alternating . only in case @xmath206 with @xmath211
is @xmath212 . given a non - reduced schubert problem , the associated reduced schubert problem is in a smaller - dimensional projective space , and so its galois group is at least alternating , by hypothesis .
we may therefore assume that @xmath95 is a reduced schubert problem , so that for @xmath213 , we have @xmath214 , where @xmath135 .
let @xmath15 be as in subsection [ ss : galois ] , so that fibers of @xmath18 are intersections of schubert problems .
recall that @xmath37 is smooth .
define @xmath33 by @xmath215 this subvariety is proper , for if @xmath216 are general and @xmath217 , they span @xmath12 .
let be the pullback of the map @xmath15 along the inclusion @xmath35 . by remark [ r :
schubert_recursion ] , @xmath39 has two components @xmath41 and @xmath42 corresponding to the two components of .
the first component @xmath41 is the total space of the schubert problem @xmath181 , and so by induction @xmath51 is at least alternating . for the second component @xmath218 ,
first replace @xmath36 by its dense open subset in which @xmath216 span a hyperplane @xmath219 .
observe that under the map from @xmath36 to the space of hyperplanes in @xmath12 given by @xmath220 the fiber of @xmath218 over a fixed hyperplane @xmath68 is the total space of the schubert problem @xmath221 in @xmath71 .
again , our inductive hypothesis and case ( a ) of vakil s criterion ( as elucidated in remark [ r : vakil ] ) implies that @xmath52 is at least alternating .
we conclude by an application of vakil s criterion that @xmath32 is at least alternating , which proves theorem [ th : one ] . while theorem [ th : one ] asserts that all schubert problems involving lines have at least alternating galois group , we conjectured that galois groups of schubert problems of lines are always the full symmetric group .
we present some evidence for this conjecture .
the first non - trivial computation of a galois group of a schubert problem that we know of was for the problem @xmath222 in @xmath223 where @xmath224 .
byrnes and stevens showed that @xmath154 is the full symmetric group @xcite and @xcite . in @xcite problems @xmath225 for @xmath226 were shown to have galois group the full symmetric group . both demonstrations used numerical methods .
we describe two infinite families of schubert problems , each of which has the full symmetric group as galois group .
both are generalizations of the problem of four lines . in @xcite ,
the schubert problem @xmath227 in @xmath59 was studied to find solutions in finite fields .
it involves lines meeting a fixed line @xmath117 and @xmath61 codimension - two planes in @xmath12 . fixing the line @xmath117 and
all but one codimension - two plane , the lines meeting them form a rational normal scroll @xmath228 , parametrized by the intersections of these lines with @xmath117 .
a general codimension - two plane will meet the scroll in @xmath229 points , each of which gives a solution to the schubert problem .
these points correspond to @xmath229 points of @xmath117 , and thus to a homogeneous degree @xmath229 form on @xmath117 .
the main consequence of @xcite is that every such form can arise , which shows this schubert problem has galois group the full symmetric group .
the other infinite family is @xmath230 , which is described in @xcite .
we use a slightly different description of it in the grassmannian of two - dimensional linear subspaces of @xmath231-dimensional space , @xmath232 ( which is identical to @xmath233 ) .
it involves the @xmath234-planes meeting four general @xmath86-planes in @xmath232 . if the @xmath86-planes are @xmath235 , then any two are in direct sum .
it follows that @xmath236 and @xmath237 are the graphs of linear isomorphisms @xmath238 .
if we set @xmath239 , then @xmath240 .
the condition that these four planes are generic is that @xmath241 has distinct eigenvalues and therefore exactly @xmath86 eigenvectors @xmath242 , up to scalar multiples .
then the solutions to the schubert problem are @xmath243 every element @xmath240 with distinct eigenvalues may occur , which implies that the galois group is the full symmetric group .
we remark that one may also apply vakil s remark 3.8 @xcite to these problems to deduce that their galois group is the full symmetric group .
we give a combinatorial injection on sets of young tableaux to establish lemma [ l : induction ] when we have @xmath244 for some @xmath245 .
[ l : unequal_inequality ] suppose that @xmath246 is a reduced schubert problem where @xmath247 with @xmath248 .
then @xmath249 to see that this implies lemma [ l : induction ] in the case when @xmath244 , for some @xmath245 , we apply schubert s recursion to obtain two different expressions for @xmath100 , @xmath250 by the inequality ( [ eq : unequal_inequality ] ) , at least one of these expressions involves unequal terms
. since all four terms are from valid schubert problems , none are zero , and so this implies lemma [ l : induction ] when not all @xmath105 are identical .
we establish the inequality ( [ eq : unequal_inequality ] ) via a combinatorial injection @xmath251 which is not surjective .
let @xmath252 be a tableau in @xmath253 and let @xmath254 be its sub - tableau consisting of the entries @xmath255 .
then the skew tableau @xmath256 has a bloc of @xmath257s of length @xmath86 at the end of its first row and its second row consists of a bloc of @xmath257s of length @xmath258 followed by a bloc of @xmath259s of length @xmath260 . form the tableau @xmath261 by changing the last row of @xmath256 to a bloc of @xmath257s of length @xmath262 followed by a bloc of @xmath259s of length @xmath263 . since @xmath264 , this map is well - defined , and gives the inclusion .
we illustrate this schematically .
@xmath265 to show that @xmath266 is not surjective , set @xmath267 , which is a valid schubert problem . hence @xmath268 and @xmath269 . for any @xmath270
, we may add @xmath271 columns to its end consisting of a @xmath272 above a @xmath273 to obtain a tableau @xmath274 .
as @xmath275 has more than @xmath276 @xmath257s in its first row , it is not in the image of the injection @xmath266 , which completes the proof of the lemma .
we prove lemmas [ l : small_m ] and [ l : m=5 ] using schubert s recursion and give an integral formula for kostka numbers coming from the weyl integral formula .
we show that if @xmath95 is a valid schubert problem , then @xmath190 , and we also compute @xmath100 for @xmath277 .
observe that there are no valid schubert problems with @xmath278 ( as we require that each component @xmath105 is positive ) .
[ [ section ] ] when @xmath192 , valid schubert problems have the form @xmath279 with @xmath280
. the corresponding geometric problem asks for the lines meeting two general linear spaces of dimension @xmath281 , that is , the lines meeting two general points .
thus @xmath282 . [
[ section-1 ] ] let @xmath283 be a valid schubert problem .
we may assume that @xmath284 so that @xmath285 by .
iterating this will lead to a schubert problem with @xmath192 , and so we see that @xmath286 .
[ [ section-2 ] ] suppose that @xmath287 is a valid schubert problem , and suppose that @xmath288 .
if it is reduced , then we have @xmath289 implying that the four numbers are equal , say to @xmath86 .
write @xmath290 in this case . by , @xmath291 as @xmath292 and @xmath293 , by .
since @xmath294 , as this is the problem of four lines , we obtain @xmath295 , which proves by induction on @xmath86 when @xmath95 is reduced and therefore equal to @xmath296 .
now suppose that @xmath95 is not reduced , and set @xmath297 since @xmath95 is not reduced and @xmath288 , we have @xmath298 and gives @xmath299 set @xmath300 .
we prove by showing that @xmath301 note that @xmath302 .
since @xmath303 , we have @xmath304 unless @xmath305 , in which case @xmath306 for some @xmath307 .
thus @xmath308 , and so @xmath309 .
but then @xmath310 , which proves when @xmath311 .
since @xmath312 , we have @xmath313 , unless @xmath314 , in which case @xmath315 for some @xmath307 . thus @xmath316 , and so @xmath317 .
but then @xmath318 , which proves when @xmath319 , and completes the proof of lemma [ l : small_m ] .
let @xmath200 be positive and even . by schubert s recursion ,
@xmath320 if we apply schubert s recursion to the last term and then repeat , we obtain @xmath321 since @xmath200 and @xmath322 , lemma [ l : small_m ] implies that @xmath323 if @xmath324 , then this minimum is @xmath325 , and if @xmath326 , then this minimum is @xmath327 .
writing @xmath328 when @xmath329 , we have @xmath330 which completes the proof of lemma [ l : m=5 ] .
let be the irreducible representation of @xmath331 with highest weight @xmath86 .
then @xmath332 is the multiplicity of the trivial representation @xmath333 in the tensor product @xmath334 .
if @xmath335 is the character of @xmath336 , then @xmath337 the integral with respect to haar measure on @xmath331 , as @xmath338 .
the weyl integral formula rewrites this as an integral over the torus @xmath339 of @xmath331 .
first note that for @xmath340 , @xmath341 then the weyl integral formula gives @xmath342 as the integrand @xmath343 satisfies @xmath344 .
we prove lemma [ l : induction ] in the remaining case when @xmath14 .
we use to recast the the inequality of lemma [ l : induction ] into the non - vanishing of an integral , which we establish by induction . it will be convenient to write @xmath346 for the quotient @xmath347 .
we complete the proof of theorem [ th : one ] by establishing the inequality of lemma [ l : induction ] for schubert problems not covered by lemma [ l : unequal_inequality ] . for these
, every condition is the same , so @xmath348 .
if @xmath349 , then we may use the hook - length formula @xcite .
if @xmath350 is even , then the kostka number @xmath351 is the number of young tableaux of shape @xmath352 , which is @xmath353 when @xmath354 is even , the inequality of lemma [ l : induction ] is that @xmath355 .
we compute @xmath356 and so @xmath357 when @xmath358 , but when @xmath359 both kostka numbers are @xmath208 , which proves the inequality of lemma [ l : induction ] , when each @xmath360 .
we now suppose that @xmath361 where @xmath362 and @xmath363 is even .
( we write @xmath364 to reduce notational clutter . )
the case @xmath365 is different because in the inequality , @xmath366 the left - hand side is negative for @xmath367 and otherwise positive .
this is shown in table [ t : inequality_a=2 ] .
we established the case when @xmath349 by direct computation in .
lemma [ l : inequality_a=2 ] covers the case when @xmath365 as @xmath379 , and lemma [ l : equal_inequality ] covers the remaining cases .
this completes the proof of lemma [ l : induction ] and of theorem [ th : one ] . by the computations recorded in table [ t : inequality_a=2 ] , we only need to show that @xmath380 for @xmath373 . using
, we have @xmath381 the integrand @xmath343 of the last integral is symmetric about @xmath382 in that @xmath383 .
thus it suffices to prove that if @xmath373 , then @xmath384 to simplify our notation , set @xmath385 we graph these functions and the integrand in for @xmath386 in figure [ f : graphs_a=2 ] .
we have @xmath403 we prove lemma [ l : inequality_a=2 ] by showing that for @xmath373 , we have @xmath404 we estimate the right - hand side . on @xmath405
$ ] , the function @xmath387 is decreasing and negative , so @xmath406 .
similarly , the function @xmath395 increases from @xmath391 at @xmath396 to @xmath208 at @xmath389 .
thus @xmath407 it is therefore enough to show that @xmath408 for @xmath373 .
this inequality holds for @xmath409 , as @xmath410 for @xmath414 $ ] , @xmath415 and @xmath416 as @xmath387 is decreasing on @xmath417 $ ] .
thus @xmath418 similarly , for @xmath419 $ ] , @xmath420 and @xmath421 , so @xmath422 from the induction hypothesis and equations and , we have @xmath423 this completes the proof of lemma [ l : inequality_a=2 ] . by the integral formula for kostka numbers ,
@xmath437 is equal to @xmath438 recall that @xmath439 and write @xmath440 these functions have symmetry about @xmath382 , @xmath441 thus if @xmath363 is odd , the integral vanishes , and it suffices to prove that @xmath442 as in subsection [ ss : a=2 ] , we show this inequality by breaking the integral into two pieces .
this is based on the following lemma , whose proof is given below .
the statement for @xmath449 is immediate from its definition . for @xmath450 ,
we use elementary calculus . recall that @xmath451 , which equals @xmath452 since the first term is everywhere nonnegative and the second nonnegative on @xmath453 $ ] ( and @xmath454 ) , we only need to show that @xmath450 is nonnegative on @xmath455 $ ] . since @xmath456
, it will suffice to show that @xmath457 is nonnegative on @xmath455 $ ] . as @xmath458 , we have @xmath459 , and so it will suffice to show that @xmath460 is nonnegative on @xmath455 $ ] .
since @xmath461 , we have @xmath462 , and so @xmath463 but this last expression is nonnegative on @xmath455 $ ] .
the first two inequalities hold as the sine function is concave on the interval @xmath417 $ ] , and the last is standard .
the quadratic upper bound is derived in @xcite .
the cubic lower bound for sine is the @xcite .
we illustrate these bounds .
@xmath472 suppose that we have @xmath476 for some number @xmath477 .
we use the mercer - caccia inequality at @xmath478 to obtain @xmath479 for @xmath480 $ ] , we have @xmath481 and @xmath482 , and therefore @xmath483 this last number is the constant in lemma [ l : inductive_step ] , which we now denote by . by lemma [ l : inductive_step ] , our induction hypothesis , and , we have @xmath484 which completes the induction step of lemma [ l : final_estimate ] . our proof of lemma [ l : inductive_step ] uses some linear bounds for @xmath449 . to gain an idea of the task at hand , in figure [ f : integral ]
we show the integrand @xmath485 and @xmath449 on @xmath486 $ ] , for @xmath487 and @xmath427 .
( 185,239)(-2,-2 ) ( 2,0 ) and @xmath488.,title="fig : " ] ( 6,217)@xmath489 ( 6,169)@xmath490 ( 6,121)@xmath491 ( 6 , 73)@xmath234 ( -2,2)@xmath392 ( 63,130)@xmath492 ( 75,13)@xmath493 ( 137,13)@xmath494 ( 170,13)@xmath389 ( 160,239)(-3,-1 ) ( 0,0 ) and @xmath488.,title="fig : " ] ( -3,228)@xmath495 ( -3,187)@xmath491 ( -3,145)@xmath397 ( -3,103)@xmath234 ( -3 , 61)@xmath208 ( -3 , 19)@xmath496 ( 57,200)@xmath488 ( 77,8)@xmath497 ( 149,8)@xmath493 we need some information about the derivatives of @xmath508 .
first observe that @xmath509 from this , we see that @xmath510 and @xmath511 is negative on @xmath512 .
moreover , @xmath513 is a sum of terms of the form @xmath514 , for @xmath515 .
thus @xmath513 is increasing on @xmath486 $ ] , as each term is increasing on that interval .
since @xmath500 has negative slope and @xmath510 , we have @xmath516 for @xmath517 $ ] near @xmath496 .
we compute @xmath518 . since @xmath519 we have @xmath520 as @xmath521 .
thus at @xmath522 , we have @xmath523 and @xmath524 and so @xmath516 for @xmath517 $ ] near @xmath525 .
if @xmath526 at some point @xmath527 , then we would have @xmath528 for at least two points @xmath506 in @xmath512 .
since @xmath528 at the endpoints , rolle s theorem would imply that @xmath513 has at least two zeroes in @xmath512 , which is impossible as @xmath513 is increasing . by lemma [ l : ell_a ] , we have @xmath529 and so it suffices to prove @xmath530 this is equivalent to showing that @xmath531 as @xmath532 is linear , this is the difference of two integrals of positive functions .
we establish the inequality by estimating each of those integrals .
the function @xmath533 is a line with slope @xmath534 and zero at @xmath535\,.\ ] ] the inequality is equivalent to @xmath536 for @xmath537 $ ]
, the linear inequalities of proposition [ p : sine_ineqs ] give @xmath538 and thus @xmath539 since @xmath540 is nonnegative on @xmath541 $ ] and @xmath542 , we have @xmath543 we may exactly compute this last integral to obtain @xmath544\ , .
\end{aligned}\ ] ] as @xmath362 , we have @xmath545 and @xmath546 . substituting these values into this last formula and multiplying by @xmath547
gives a lower bound for the integral on the left of , @xmath548 for the integral on the right of , consider the line through the points @xmath505 and @xmath549 , @xmath550 we claim that @xmath551 in the interval @xmath552 $ ] . to see this , first note that the slope of a secant line through @xmath505 and a point @xmath553 on the graph of @xmath449 is @xmath554 as observed in proposition [ p : sine_ineqs ] , @xmath555 is bounded above by the parabola , @xmath556 we use this and the mercer
caccia inequality for @xmath557 to bound the slope , @xmath558 with the second equality holding as the minimum of the denominator @xmath559 on the interval @xmath552 $ ] occurs at @xmath522 .
when @xmath375 we have , @xmath560 which so it follows that @xmath561 on @xmath552 $ ] . using this and the easy inequality @xmath562 , we bound the integral on the right of , @xmath563 the last integral is not hard to compute , @xmath564}{\pi^{\mu-1}(\mu+1)(\mu+2 ) ( 3a^2 + 6a-1)^2}\;.\ ] ] we claim that @xmath565 , which will complete the proof of lemma [ l : inductive_step ] and therefore the induction step for lemma [ l : final_estimate ] .
for this , we observe that if multiply @xmath566 by their common ( positive ) denominator , we obtain an expression of the form @xmath567 , where @xmath568 is a polynomial of degree six in @xmath86 and two in @xmath477 . after making the substitution @xmath569
, we obtain a polynomial in @xmath26 and @xmath570 in which every coefficient in positive , which implies that @xmath571 when @xmath572 , and completes the proof .
we establish the inequality of lemma [ l : final_estimate ] when @xmath573 , which is the base case of our inductive proof .
this inequality is @xmath574 we establish this inequality by replacing each integral by one which we may evaluate in elementary terms , and then compare the values .
we first find an upper bound for the integral on the right .
recall that @xmath575 since @xmath576 and @xmath577 for @xmath480 $ ] , we have @xmath578 for @xmath375 , we have @xmath579 .
as we observed in proposition [ p : sine_ineqs ] , this implies that @xmath580 , and so @xmath581 .
since @xmath582 , we have @xmath583 we now find a lower bound for the integral on the left of .
we use the estimate from lemma [ l : ell_a ] , that for @xmath443 $ ] , we have @xmath584 using this gives the lower bound , @xmath585 this may be evaluated in elementary terms to obtain @xmath586 for @xmath375 , @xmath587 , we have the bound from proposition [ p : sine_ineqs ] of @xmath588 . thus the expression is bounded below by @xmath589 then the difference @xmath566 of the expressions from and is a rational function of the form @xmath590 where @xmath591 is a polynomial of degree seven . if we expand @xmath592 in powers of @xmath26 , then we obtain a polynomial of degree seven in @xmath26 with positive coefficients .
this establishes the inequality for all @xmath375 , which is the base case of the induction proving lemma [ l : final_estimate ] .
this completes the proofs of lemma [ l : final_estimate ] , lemma [ l : equal_inequality ] , and ultimately of theorem [ th : one ] .
s. billey and r. vakil , _ intersections of schubert varieties and other permutation array schemes _ , algorithms in algebraic geometry , i m a vol .
146 , springer , new york , 2008 , pp . 2154 .
byrnes , _ pole assignment by output feedback _ , three decades of mathematical systems theory ( h. nijmeijer and j. m. schumacher , eds . ) , lecture notes in control and inform .
135 , springer - verlag , berlin , 1989 , pp .
byrnes and p.k .
stevens , _ global properties of the root - locus map _ , feedback control of linear and non - linear systems ( d. hinrichsen and a. isidori , eds . ) , lecture notes in control and inform .
39 , springer - verlag , berlin , 1982 .
to3em , _ intersection theory and enumerative geometry : a decade in review _ , algebraic geometry , bowdoin 1985 ( spencer bloch , ed . ) , proc .
pure math .
46 , part 2 , amer .
soc . , 1987 , pp . | we show that the galois group of any schubert problem involving lines in projective space contains the alternating group .
this constitutes the largest family of enumerative problems whose galois groups have been largely determined . using a criterion of vakil and a special position argument due to schubert , our result follows from a particular inequality among kostka numbers of two - rowed tableaux . in most cases , a combinatorial injection proves the inequality . for the remaining cases
, we use the weyl integral formulas to obtain an integral formula for these kostka numbers .
this rewrites the inequality as an integral , which we estimate to establish the inequality . |
inspiraling binaries are one of the most promising sources for the first detection of gravitational waves ( gw ) .
the post newtonian approximation methods accurately describe the phasing of the waveform - about a cycle in a wave train @xmath0 cycles long .
this makes it amenable for matched filtering analysis .
the best available estimates suggest that the expected number of neutron star ( ns)-ns binary coalescence seen per year by ground based interferometers is @xmath1 for initial detectors and @xmath2 for advanced detectors @xcite . in recent years , a number of ground based detectors are producing sufficiently interesting sensitive data and analysis of network data is highly advisable .
the advantages of multi - detector search for the binary inspiral is that , not only does it improve the confidence of detection , it also provides directional and polarisational information about the gw source .
two strategies currently exist in searching for inspiraling binary sources with a network of detectors : the coherent and the coincident .
the coherent strategy involves combining data from different detectors phase coherently , appropriately correcting for time - delays and polarization phases and obtaining a single statistic for the full network , that is optimized in the maximum likelihood sense . on the other hand
, the coincident strategy matches the candidate event lists of individual detectors for consistency of the estimated parameters of the gw signal .
however , the phase information is ignored and as also the detectors are considered in isolation .
the question arises as to which strategy performs better . on simple waveforms
the analysis has been performed by finn and arnaud et al@xcite .
both these works have shown that the coherent strategy performs better than the coincident strategy .
we consider here the astrophysically important source , namely , the inspiraling binary and report mainly the results of our work which has been described in detail in the papers @xcite .
we compare the strategies by plotting the _ receiver operating characteristic _
( roc ) curves , which is the plot of detection efficiency versus the false alarm rate .
we broadly consider the two cases of co - located aligned detectors and geographically separated misaligned detectors . in the co - located case we further consider two subcases of ( i ) uncorrelated noise , and ( ii ) correlated noise .
for the inspiraling binary , in the fourier domain we assume the spinless restricted post - newtonian ( pn ) waveform @xmath3 at detector @xmath4 : ^i ( f ) = e^i f^-7/6 i [ ( f ; t_c , _ c , _ 0 , _ 3 ) + 2 f t^i ] , [ eq : stilde ] where we take the phase @xmath5 given by the 3 pn formula .
the extended beam pattern functions @xcite @xmath6 encode the orientation and direction parameters of the source and detectors ; the parameters @xmath7 , are respectively , the time of coalescence , phase of coalescence ; the quantities @xmath8 are the chirp time parameters which are independent functions of the two masses of the stars comprising the binary , the @xmath9 denote time - delays at the detector @xmath4 with respect to a fiducial detector and @xmath10 is the amplitude depending on the masses and the distance to the source . in matched filtering , it is natural to define the scalar product @xmath11 of two real functions by , ( a , b ) = 2 _ f_l^i^f_u^i df , [ eq : scalar ] where , we use the hermitian property of fourier transforms of real functions .
@xmath12 is the one sided power spectral density ( psd ) of the noise which is assumed to be the same for all the detectors .
then the normalised templates @xmath13 corresponding to the intrinsic parameters @xmath14 are defined via the equation , ^i(f;_i , t_c , _ c ) = ^i ( _ 0 ( f ; _ i , t_c ) _ c + _ /2 ( f ; _ i , t_c ) _ c ) .
[ eq : quadrature ] we have the relation @xmath15 where @xmath16 represents a grid point in the intrinsic parameter space and the normalisation of templates requires that the scalar products @xmath17 .
we then define the complex correlation @xmath18 where , the real correlations @xmath19 and @xmath20 are obtained by taking the scalar products of the data @xmath21 with @xmath22 and @xmath23 respectively ; that is , @xmath24 .
then the single detector statistic for detector @xmath4 is just @xmath25 . to decide detection in a given single detector @xmath4 ,
@xmath26 is maximised over the template parameters @xmath27 and compared with a preassigned threshold .
for coincidence detection , the procedure is as follows : * choose the same threshold @xmath28 for the two detectors .
* prepare two candidate event lists such that @xmath29 .
look for pairs of candidate events , each candidate event coming from a different list , such that the sets of estimated parameters @xmath30 match - @xmath31 where the @xmath32 denotes the difference in the measured parameter @xmath33 , where @xmath33 stands for any of the parameters @xmath30 .
the allowed error box is denoted by @xmath34 . a quadratic sum of the noises is taken to determine the error box .
we fix this box by performing simulations , so that the final probability of not losing an event is @xmath35 - on each parameter ,
we atmost allow 1 % loss in events . for geographically separated detectors , for fixing the window size in @xmath36 , the light travel time between the detectors
is taken into account and added in quadratures to the errors due to noise in each detector .
on the other hand , coherent detection involves combining data streams in a phase coherent manner so as to effectively construct a single , more sensitive detector .
for the case of two misaligned detectors the network statistic is given by @xcite : = ||c||^2 = |c^1|^2 + |c^2|^2 = ( c^1_0)^2 + ( c^1_/2)^2 + ( c^2_0)^2 + ( c^2_/2)^2 , [ twoindep ] where @xmath37 is the complex correlation of the @xmath4-th detector ( @xmath4=1,2 ) .
for two aligned colocated detectors the statistic is different @xcite and is given by : = |c^1 + c^2|^2 . for aligned colocated detectors we consider two subcases : ( i ) uncorrelated noise , ( ii ) correlated noise
we first consider the case of co - located aligned detectors .
this case would be of significance to the two ligo detectors at hanford and other similar topologies envisaged elsewhere in the future such as lcgt .
we compare the performances of the two strategies by plotting the roc curves for uncorrelated noise ( left ) and correlated noise ( right ) in fig .
( [ fig1 ] ) . for plotting these curves the false alarm and the detection probabilities must be computed as a function of the threshold @xmath28 and then plotted versus each other parametrically by varying the parameter @xmath28 .
the correlation parameter @xmath38 is taken as the weighted average of a frequency dependent correlation @xmath39 where @xmath40 , @xmath41 being the noise in the @xmath42 detector .
details may be found in @xcite .
while performing the simulations there are many subtleties such as estimating the number of independent templates , the error window size computation etc .
the discussion of which we have omitted here although it is nonetheless important .
cc it is clear from the figures that the coherent strategy is far superior in this case .
we now consider the case of geographically separated detectors which are then also normally misaligned . when the detectors are misaligned , the sky coverage for the usual coincidence detection is very poor which leads to intolerable false dismissal .
thus another coincidence strategy is devised which we call _ enhanced _ coincidence .
the usual coincidence strategy we then call _ naive _ coincidence .
enhanced coincidence strategy is formulated as follows : * choose a low threshold @xmath43 ( we choose @xmath44 ) , and prepare two candidate event lists such that @xmath45 . *
look for a pair of candidate events , the events coming from separate lists , such that the sets of estimated parameters match within the error - window .
the procedure is the same as the co - located case , except for the parameter @xmath36 in which the distance between the detectors enters .
* choose the final ( high ) threshold @xmath46 and construct the final statistic @xmath47 and register detection if @xmath48 .
cc note that although this statistic looks formally like the coherent case , the mass parameters for the templates in the two detectors do not have to be the same ( they however must be close enough so that they lie in a error window ) .
thus this is not matched filtering while coherent detection is . in this strategy
the sky coverage is better than naive coincidence . as seen from the fig .
( [ fig2 ] ) , coincidence strategy performs far better than the naive coincidence strategy , but we see that the detection probability for the coherent strategy is still superior by around 5% for the same false alarm rate .
although the coherent strategy is superior to coincident strategies , the difference between the coherent and enhanced coincident strategies is small
. only a relative improvement of about 5% in the detection probability is obtained with the coherent strategy .
one may ask whether there is any practical advantage in using the coherent strategy in the case of two misaligned detectors .
note however that the coherent method is not so computationally expensive compared with two coincident methods , since we do not take cross correlation of two detectors data in the coherent strategy .
thus overall , we conclude that the coherent strategy is a good detection method . however , the above results assume stationary gaussian noise .
but we know that the current real data are neither stationary nor gaussian . in coincidence detection , the requirement of the consistency of estimated parameters in an error window acts as a powerful veto @xcite to veto out fake events generated from non - gaussian noise . on the ther hand in coherent detection as yet no such obvious veto has been developed .
this however does not rule out the possibility that a powerful veto can not be constructed for the coherent strategy . in future
we propose to work on this aspect of the problem .
perhaps a judicious combination of the two methods might be an effective way of dealing with this problem .
the authors would like to thank the dst , india and jsps , japan for the indo - japanese cooperative programme for scientists and engineers under which this work has been carried out . | we compare two multi - detector detection strategies , namely , the coincidence and the coherent , for the detection of spinless inspiraling compact binary gravitational wave ( gw ) signals .
the coincident strategy treats the detectors as if they are isolated - compares individual detector statistics with their respective thresholds while the coherent strategy combines the detector network data _ phase coherently _ to obtain a single detection statistic which is then compared with a single threshold . in the case of geographically separated detectors
, we also consider an _ enhanced _ coincidence strategy because the usual ( naive ) coincidence strategy yields poor results for misaligned detectors . for simplicity
, we consider detector pairs having the same power spectral density of noise , as that of initial ligo and also assume the noise to be stationary and gaussian .
we compare the performances of the methods by plotting the _ receiver operating characteristic _
( roc ) for the two strategies .
a single astrophysical source as well as a distribution of sources is considered .
we find that the coherent strategy performs better than the two coincident strategies under the assumptions of stationary gaussian detector noise . |
this paper is supported by the nsfc of china under the grant no.90103028 , 90306016 , and no.10174098 . 1 , l. jacak , p. hawrylak , a. wjs , _ quantum dots _ ( springer , berlin , 1998 ) + 2 , t. chakraborty , _ quantum dots _ ( elsevier , amsterdam , 1999 ) _ _ + 3 , m.s .
kushwaha , surface science reports , * 41 * , 1 ( 2001 ) + 4 , s.m reimann and m. manninen , rev .
mod . phys . * 74 * , 1283 ( 2002 ) + 5 , g.w .
bryant , phys .
lett . * 59 * , 1140 , ( _ 1987 _ ) + 6 , p.a .
maksym , h. imamura , g.p .
mallon , and h. aoki , j. phys .
: condens .
matter * 12 * , r299 ( 2000 ) + 7 , l.p .
kouwenhoven and p.l .
mceuen , _
nanotechnology _ , edited by g. timp ( springer , new york , 1999 ) , p471 + 8 , f. nakajima , y. ogasawara , j. motohisa , t. fukui , physica , e * 13 * , 703 ( 2002 ) + 9 , i.h .
chan , p. fallahi , r.m .
westervelt , k.d .
maranowski , a.c .
gossard , physica e * 17 * , 584 ( 2003 ) + 10 , e.b .
foxman , et al .
b * 47 * , 10020 ( 1993 ) 11 , m. wagner , u. merkt , and a.v .
chaplik , phys .
b * 45 * , 1951 ( 1992 ) + 12 , d. pfannkuche , v. gudmundsson , and p.a .
maksym , phys .
b * 47 * , 2244 ( 1993 ) + 13 , p.hawrylak , phys . rev . lett . *
71 * , 3347 ( 1993 ) + 14 , m.pustilnik , l.i .
glazman , and w. hofstetter , phts , rev .
b * 68 * , 161303r ( 2003 ) + 15 , a. fuhrer , t.ihn , k. ensslin , w.wegscheider , m. bichler , phys . rev . lett . *
91 * , 206802 ( 2003 ) + 16 , a. kogan , g. granger , m.a .
kastner , d. goldhaber - gordon , and h. shtrikman , phys .
b * 67 * , 113309 ( 2003 ) + 17 , j. kyriakidis , m. pioro - ladriere , m. ciorga , a.s .
sachrajda , and p. hawrylak , phys .
b * 66 * , 035320 ( 2002 ) + 18 , c.g .
bao , j. phys .
: conden . matter , * 14 * , 8549 , ( 2002 ) + 19 , g.m .
huang , y.m .
liu , c.g .
bao , phys . rev .
b * 68 , * 165334 ( 2003 ) + 20 , c.g .
bao , w.f .
xie , and c.d .
lin , j. phys .
b : at . mol .
* 27 * , l193 ( 1994 ) + 21 , w.y .
ruan , y.y .
liu , c.g .
bao and z.q .
zhang , _ phys .
rev_. * b51 * 7942 ( 1995 ) .
+ 22 , c.g .
bao , _ phys . rev . lett_.
* 79 * , 3475 ( _ 1997 _ ) . | based on symmetry constraint that leads to the appearance of nodes in the wave functions of 3-electron systems at regular triangle configurations , it was found that , if the parameters of confinement are skillfully given and if a magnetic field is tuned around the first critical point of the singlet - triplet transition , a 2-electron quantum dot can be used as a highly sensitive switch for single - electron transport . the corresponding author .
modern experimental techniques , e.g. , using electrostatic gates and etching , allow a certain number of electrons ( from a few to a few thousand ) to be confined in quantum dots.@xmath0 as a kind of systems different from those existing in nature , rich physics is contained .
therefore , they attract certainly the interest of academic research . on the other hand , the properties of the dots can be controlled , e.g. , by changing the gate voltage or by applying an external magnetic field , etc .
therefore , these systems have a great potential in application . in particular
, the dots can be used as a single - electron switch to control the electric microcurrent . making use of the coulomb blockade @xmath1 , current switches have already been designed by a number of authors ( refer to ref.8 and 9 , and references therein ) .
this kind of efforts is crucial for developing microtechniques . in general , there are two factors harmful to the sensitivity of this kind of device , one is thermal fluctuation , another is quantum tunneling .
the former can be effectively reduced by lowering the temperature , the latter can not . in order to eliminate the quantum tunneling
, we propose an idea of a new mechanism in this paper .
related theoretical evaluation has been performed , however , technical details are not involved .
let us review briefly the main physics of coulomb blockade .
let a dot be connected to a source and a drain by tunnel barriers .
a gate - voltage @xmath2 is applied on the dot , while bias - voltages @xmath3 and @xmath4 are applied on the source and drain , respectively .
let @xmath5electrons be confined in the dot with the ground state energy @xmath6 .
if an extra electron comes in from the source , the lowest energy of the ( n+1)-electron system is denoted as @xmath7 .
the extra electron is driven by the bias - voltage @xmath8 .
thus the condition of coulomb blockade is @xmath9 @xmath10 ( where the constant @xmath11 is introduced by foxman et al.@xmath12 , however the actual value of @xmath13 is not essential to the following discussion ) .
whereas if @xmath14 , the blockade is released and the extra electron would go through the dot to the drain . by increasing @xmath15 @xmath16 can be reduced as shown schematically in fig.1a . when @xmath2 increases and arrives at a critical value , the condition @xmath17 holds , and resonant transmission occurs .
it is noted that @xmath16 is changed continuously with @xmath2 .
therefore , even when @xmath18 , the leak of current might occur due to quantum tunnelling .
thus , the blockade is not strict , and the sensitivity of the device as a switch is more or less spoiled .
[ htbp ] in the usual device of coulomb blockade ( a ) , and in an improved device ( b ) . when @xmath19 , resonant transmission occurs .
x in ( b ) is a quantity to control the variation of @xmath16 ( not yet specified)@xmath20,title="fig : " ] on the other hand , if one can design a device so that @xmath16 varies abruptly as shown in fig.1b , the quantum tunnelling would be remarkably suppressed and the sensitivity would be greatly improved . to achieve the abrupt change
, we suggest a source - dot - drain device with the following three features .
\(1 ) the dot is axial symmetric and contains two electrons .
the eigenstates states have the orbital angular momentum @xmath21 and total spin @xmath22 to be conserved .
in particular , the ground state has @xmath23 , @xmath24 , and energy @xmath25 .
\(2 ) @xmath2 is fixed at a skillfully prescribed value ( see below ) , while an adjustable external magnetic field @xmath26 is applied and tuned around a critical point @xmath27 so that the ground state undergoes singlet - triplet transitions@xmath28 .
accordingly , @xmath29 jumps to @xmath30 , or reversely , while the corresponding ground state energies @xmath25 and @xmath31 are identical at the critical point . for convenience , the region @xmath32 is called region i ( ii ) .
\(3 ) the bias - voltage is low and the extra electrons are controlled so that they are coming slowly one - by - one , the incoming electron is so slow that it is , relative to the center of the dot , in s - wave .
let us see how this device works .
when @xmath26 is tuned in region i , if the third electron can come in and form a 3-electron system , then the quantum numbers of the 3-body final state are denoted as @xmath33 and the energy as @xmath34 .
once the partial wave is limited to s - wave , 0nly @xmath35 and @xmath36 are allowed , accordingly @xmath37 @xmath38(if more than one final states are allowed , the higher states are not involved due to the low bias - voltage ) .
when @xmath39 is in region ii , since the quantum numbers of the 2-electron ground state has already been changed , @xmath33 are accordingly changed to @xmath40 with the energy @xmath41 .
thus , after the transition , we have @xmath42 .
although @xmath43 at the critical point , however in general @xmath44 , therefore @xmath45 .
this leads to a jump of @xmath16 occurring at the critical point as expected .
the realization of the abrupt variation of @xmath16alone is not sufficient .
in addition , two more conditions @xmath46 and @xmath47 are required so that quantum tunnelling is nearly zero and @xmath16 jumps down exactly to the right place to initiate the resonant transmission as plotted in fig.1b .
it was found that , due to the help of symmetry constraint as explained below , these two conditions can be realized if the parameters of confinement are specially designed .
this is shown by a theoretical study of a model as follows .
[ htbp ] the label @xmath29 is marked by the associated curve@xmath20,title="fig : " ] for simplicity , the model dot is assumed to be 2-dimensional .
the potential of confinement is an axial - symmetric square well @xmath48 if @xmath49 , or @xmath50 elsewhere , where @xmath51 is positive .
furthermore , a uniform magnetic field @xmath26 is applied perpendicular to the x - y plane , the hamiltonian reads @xmath52 @xmath53 where @xmath54 , @xmath5 is the number of electrons ( equal to 2 or 3 ) , the zeeman term is included .
@xmath55=0.067@xmath56 , the dielectric constant @xmath57=12.4 ( for a gaas dot ) , and the units @xmath58 , @xmath59 , and @xmath60 are used through out this paper .
the hamiltonian is diagonalized to obtain the eigenenergies by using the methods as outlined in 18 , 19 .
[ htbp ] , the critical point of singlet - triplet transition , as a function of @xmath51 when @xmath61 is fixed ( for a gaas dot)@xmath20,title="fig : " ] the ground state energies @xmath25 calculated with @xmath62 and @xmath63 ( just as an example ) are plotted in fig.2 as a function of @xmath26 .
the qualitative feature of fig.2 would remain unchanged if @xmath51 and @xmath61 vary inside a large reasonable region .
one can see that the increase of @xmath26 causes a number of singlet - triplet transitions and totally three critical points appear .
it was found that the vicinity of the first critical point , denoted as @xmath27 ( equal to @xmath64 in fig.2 ) is suitable for our purpose .
we shall concentrate in the vicinity of @xmath27 . in region
i ( @xmath65 ) and ii , ( @xmath66 = @xmath67 and ( 1,1 ) , respectively .
it was found that , when @xmath61 is given , the value of @xmath27 depends on @xmath51 nearly linearly .
e.g. , when @xmath68 when @xmath69 . the larger the @xmath51 , the larger the @xmath70 , while the larger the @xmath61 , the smaller the @xmath27 .
when @xmath26 is given in region i , if the third electron with s - wave can come in , ( @xmath71 ) should be ( 0,1/2 ) . selected values of @xmath72
are calculated and plotted as functions of @xmath26 as shown in fig.3 . on the other hand , from dynamical consideration ,
if the three electrons form a regular triangle ( rt ) , the total potential energy can be minimized .
thus the rt is a favorable configuration .
however , from the study of symmetry constraint @xmath73 , it was known that the spatial wave function of a 3-electron state @xmath74 would be zero at the rts if @xmath75 and @xmath76 , or if @xmath77 and @xmath78 , where @xmath79 = 0 , 1 , 2 , @xmath80 .
this is called a prohibition of regular triangle ( prt ) , and will definitely cause serious effect .
let us prove the prt when @xmath78 . in this case
@xmath81 is totally antisymmetric .
when the three electrons form an rt , a rotation of the system by @xmath82 is equivalent to a cyclic permutation of particles .
the rotation leads to an extra factor @xmath83 , while the permutation is equivalent to two interchanges and thus causes nothing .
thus , the equivalence leads to @xmath84 this equation holds only at the rt configurations . evidently ,
when @xmath85 , the first factor of ( 2 ) is nonzero , therefore @xmath86 must be zero at the rt .
it implies the appearance of an inherent node at the rt , and the prt holds .
the case @xmath76 can be similarly proved@xmath87 .
[ htbp ] and @xmath88 with respect to @xmath26 in the vicinity of @xmath27 .
@xmath61 is fixed at @xmath89 , and @xmath51 is given at a number of values ( in mev)@xmath20,title="fig : " ] when prt holds , the inherent node at the rt will remarkably increase the kinetic energy , and at the same time push the wave function away from the area of lower potential energy .
therefore , the associated energy @xmath90 is considerably higher than those free from the prt . for this reason ,
@xmath91 is in general high . by adjusting @xmath51 or @xmath61 , @xmath91
can be easily much higher than @xmath92resulting in a large @xmath93 as confirmed by fig.3 . in region
ii , the s - wave limitation leads to two choices ( @xmath94 ) = ( 1,1/2 ) and ( 1,3/2 ) .
due to the prt , @xmath95 is considerably higher and therefore can be neglected .
whereas , due to being free from the prt , @xmath96 is remarkably lower , this leads to a small @xmath97 @xmath98as plotted in fig.3 .
thus the prt assures @xmath46 , this leads to the great jump shown in fig.3 .
the second condition @xmath47 can also be satisfied by adjusting the parameters of confinement . in what follows @xmath99
is assumed due to the very low bias - voltage . when @xmath63 , we know fom fig.3 that @xmath100 holds if @xmath101 . in general , for a given @xmath61 , there is a corresponding @xmath102 so that the pair of parameters @xmath103 and @xmath61 gives @xmath104 at @xmath105 .
based on our numerical results ( @xmath106 ) , a nearly linear relation @xmath107 was found .
the larger the @xmath61 , the smaller the @xmath103 . when other types of confinement ( e.g. , a parabolic potential with a limited height ) are used , special set of parameters can also be found to assure @xmath100 at @xmath105 ( not shown here ) .
thus we can conclude that , if the parameters of confinement are skillfully chosen , due to the symmetry constraint and making use of the singlet - triplet transition , when @xmath26 is tuned in the vicinity of its first critical point ( @xmath27 ) , the two - electron dot can work as a highly sensitive switch at very low temperature .
the regions i and ii are associated with off and on , respectively . for a further discussion , a typical low - lying spectrum of the 2- and 3-electron dots in the vicinity of @xmath27
is given in fig.4 ( where @xmath51 is a little larger than @xmath103 ) .
for these systems the levels not appearing in the figure are much higher . if @xmath51 is given considerably larger than @xmath103 , @xmath96
would become considerably lower and lead to @xmath108 . in this case the intruding electron may find some way to release its energy ( e.g. , by emitting a photon or a phonon ) and falls into the level @xmath96 .
if this occurs , the intruding electron can not go out again but remain inside , thereby the previous mechanism of transport would be spoiled .
therefore @xmath51 is better given at @xmath109 this leads to @xmath47 .
in this case a resonant transmission takes place and the intruding electron has no chance to remain inside .
[ htbp ] versus @xmath61 .
@xmath110the parameters @xmath103 and the associated @xmath61 together assure the condition of resonant transmission , i.e. , @xmath111 , ( here @xmath112 is assumed ) @xmath113,title="fig : " ] if the confinement potential contains non - symmetric component @xmath114 , @xmath21 is no more strictly conserved . in region
i ( in the status off )
if the 2-body ground state contains a small @xmath115 component , this component can absorb the incoming s - wave electron and goes to the @xmath96 level , thereby a leak may occur . alternatively ,
if the 2-body ground state contains a small @xmath116 component , this component can not absorb the extra s - wave electron because the level @xmath117 is quite high ( cf .
fig.4 ) .
this fact implies that only the kind of @xmath114 with odd parity would cause a leak , this kind should be diminished as far as possible ( e.g. , in the technical aspect , the two pipes connected to the dot are placed by the two sides as symmetric as possible ) .
a representative @xmath118 with odd parity can be written as @xmath119 , where @xmath120 is the original one in eq.(1 ) .by using the perturbation theory , for two representative cases , the ratio of the weights of the @xmath121 and @xmath122 components in the 2-body ground state is @xmath123 if @xmath63 , @xmath124 and @xmath125 ; and @xmath126 if @xmath127 , @xmath128 , and@xmath129 . from this evaluation ,
if @xmath130 and @xmath131 are in the order of 10@xmath132 , @xmath118 would not cause a serious problem . besides , a smaller @xmath61 would lead to a larger level - spacing @xmath133 , and therefore a smaller @xmath134 . the kondo effect would do no harm in this device .
this effect ceases to exist in the status off because the spin of the dot is meanwhile zero , while this effect is promotive in on .
however , the thermal fluctuation might spoil the sensitivity of the device . in the status
off , if the 2-electron state has been raised up from the ground state to the level @xmath135 by thermal fluctuation , the third electron may come in and the system may fall to the level @xmath96 , and thus the off is no more strict .
let us define@xmath136 @xmath137 which is the ratio of the probabilities of staying in the levels @xmath135 and @xmath138 , respectively , and is plotted in fig.5 as a function of @xmath26 , where @xmath139 is given at a number of values . from this figure , we know that , if @xmath140 0.01k , when @xmath141 the thermal fluctuation is negligible . therefore , in this temperature , a change of @xmath26 in the order of 0.01 around @xmath27 is sufficient to initiate the on or off of the switch .
when @xmath142 is much smaller than 0.01k , a very small change of @xmath26 is sufficient to initiate the on or off , thus the switch becomes highly sensitive .
the s - wave limitation is a basic requirement . to evaluate in a semi - classical way ,
let us assume that a classical electron comes in with a velocity @xmath143 along a straight pipe with a radius @xmath144 aiming at the center of the dot .
then the angular momentum is @xmath145 and the s - wave limitation is @xmath146 , or @xmath147 , where @xmath144 is in @xmath59 , and @xmath148 is the velocity of light in vacuum . on the other hand , the thermal velocity @xmath149 , where @xmath142 is the temperature in @xmath150 .
thus , the requirement can be rewritten as @xmath151 . if @xmath152 even @xmath143 is as large as @xmath153 , the requirement , namely @xmath154 can be realized with the present technics of fabrication . as final remarks ,
two distinguished features of the suggested device as a sensitive switch are noticeable .
firstly , the quantum tunneling has been greatly suppressed .
secondly , the device is highly sensitive to the variation of @xmath26 at low temperature , this is the superiority .
however , since very low temperature together with the s - wave limitation and good axial symmetry of the dot are required , the effectiveness of the switch depends on the technical aspect , and remains to be checked .
nonetheless , it is believed that the idea proposed in this paper , i.e. , making use of the effect of symmetry constraint and phase transition , is in general useful in the design of various microdevice . |
after a long period of experimental work , large erenkov detectors for high energy neutrinos are going to open a new observational window to the sky .
their main goal is to extend the volume of the explored universe by neutrinos , to obtain a complementary view of astronomical objects and to learn about the origin of high energy cosmic rays .
they are the successors of underground neutrino detectors which have turned out to be too small to detect the faint fluxes of neutrinos from cosmic accelerators .
the new detectors are large , expandable arrays of photomultipliers constructed in open water or ice .
the photomultipliers span a three - dimensional coarse grid and map the erenkov light of secondary particles produced in neutrino interactions .
actually , the basic idea for this detection method goes back to the early 60 s @xcite .
pioneering attempts towards its realization have been made in the course of the dumand project @xcite . in 1993 ,
the baikal collaboration @xcite succeeded to built the first deep underwater erenkov neutrino detector , which has been stepwise upgraded to its present stage , nt-200 .
the amanda collaboration @xcite has built a erenkov detector in the south pole ice .
other collaborations ( antares @xcite , nestor @xcite ) are constructing underwater neutrino detectors of similar size .
since a few years , the nemo collaboration @xcite is performing an intensive r&d program aiming at the construction of a km@xmath0 erenkov neutrino telescope in the mediterranean sea .
another cubic kilometer detector , icecube @xcite is planned at the south pole .
the cubic kilometer scale is set by various predictions on the extremely low fluxes of high energy neutrinos expected from astrophysical sources . in underwater erenkov neutrino telescopes
, water acts not only as a target but also as radiator of erenkov photons produced by relativistic charged particles .
the detection volume , as well as the angular and energy resolutions strongly depend on the water transparency . the transparency of water as a function of photon wavelength @xmath1 , is described by the so called inherent optical properties , like the coefficients for absorption @xmath2 , for scattering @xmath3 , for attenuation @xmath4 , and by the phase scattering function @xmath5 ( also referred to as volume scattering function ) which represents , for a photon , the probability to be diffused at an angle @xmath6 @xcite .
another parameter commonly used in literature is the effective scattering coefficient @xmath7 , where @xmath8 is the average cosine of the phase scattering function at a given @xmath1 .
the optical properties of natural water have to be measured _ in - situ _ in order to allow an unbiased knowledge of light transmission properties in the medium .
the baikal collaboration has been investigating the fresh water deep in lake baikal since 1980 .
the inherent optical properties have been measured with a series of specially designed devices .
it was shown that the water transparency at depths between 900 m and 1200 m is adequate to operate a neutrino telescope .
put into operation at april 6@xmath9 , 1998 , the neutrino telescope nt-200 incorporates a long - term monitoring system which performs continuous measurements of the water parameters .
this information serves as input for monte - carlo simulations of the detector response to atmospheric muons which represent a well - known calibration source for neutrino telescopes .
the muon fluxes measured with nt-200 are in very good agreement with simulation results .
this fact confirms that the custom - made devices and the methods to extract the relevant information on optical parameters yield reliable results .
the nemo collaboration has been investigating oceanographic and optical properties of several deep sea marine sites close to the italian coast , with the aim to select the optimal site for the construction of a km@xmath0 detector in the mediterranean sea .
absorption and attenuation coefficients for light in the wavelength region between 412@xmath10715 nm @xcite have been measured with a set - up based on commercial devices .
optical measurements in deep water are extremely difficult , and possible systematic errors related to these measurements suggest careful cross checks of results by complementary methods . for these reasons , during march - april 2001 , the nemo and baikal collaborations have started a joint campaign to measure the optical properties of deep water in lake baikal using two different devices .
one set - up is based on the transmissometer _
ac9 _ , operated by the nemo group , the other device , _
( absorption , scattering and phase function meter ) , was developed and operated by the baikal collaboration .
the cross check of experimental results has been crucial for both devices , since both have an excellent sensitivity in measuring water optical properties , however , they can be affected by different sources of systematic errors which could deteriorate the absolute accuracy .
the measurements reported in the following sections have been carried out during march - april 2001 , from the ice camp above the neutrino telescope nt-200 .
the _ ac9 _ , manufactured by _ wetlabs _ @xcite , is a transmissometer capable to measure absorption and attenuation coefficients at nine different wavelengths in the range @xmath11 nm . using an accurate calibration procedure ,
the nemo collaboration has achieved an accuracy of about @xmath12 m@xmath13 in @xmath14 and @xmath15 measurements @xcite . and @xmath16 ) , as functions of depth , obtained by two measurements ( red and black dots ) with the _ ac9_-ctd set - up in lake baikal during march 2001 .
the nt-200 telescope is located between 1100 m and 1170 m depth.,width=302 ] during the measurements in lake baikal , the _
ac9 _ device and a ctd ( a probe that measures water conductivity , temperature and pressure ) were operated through an electro - mechanical cable . with this set - up
we have obtained two vertical profiles of the water column ( 50 m@xmath17 depth @xmath171100 m ) , collecting about ten data - sets per meter of depth .
each data - set consists of a measurement of temperature and optical properties , @xmath2 and @xmath18 , over the nine wavelengths . in figure
[ fig : tempacbaikal ] we show , as a function of depth , the water temperature together with the values for absorption and attenuation coefficient at @xmath19 nm measured during the first and the second deployment in lake baikal ( for discussion see section [ sec : results ] ) .
the _ asp-15 _ device ( see figure [ fig : asp-15 ] ) has two receiving channels : one with a wide aperture to measure @xmath14 and @xmath20 and another one with a rotating mirror and a narrow angle collimator to measure the phase scattering function .
two photomultipliers ( type feu-130 ) and 15 interference light filters are assembled in a cylindrical container .
the filters wavelengths are ranged from 369 nm to 691 nm .
both photomultipliers operate in photon counting mode .
two isotropic point - like light sources and two screens are assembled on a frame , which can be moved by a stepping motor over distances ranging from 0.4 to 15 meters with respect to the milk glass window .
measurements were carried out separately for each source and controlled via cable by a computer on shore or at the ice camp .
the device is described in detail in @xcite .
the principle of the @xmath14 measurement with _
_ is described in @xcite .
we measure the dependence of the luminosity @xmath21 on the distance @xmath22 between source and receiver with each of the 15 light filters and approximate the absorption coefficient by @xmath23 where @xmath24 and @xmath25 are the luminosity at distances @xmath26 and @xmath27 , respectively .
monte - carlo simulations @xcite have shown that for an isotropic or a lambertian ( cosine ) light source in water , the difference between approximation ( [ abs ] ) ( which is an exact definition of @xmath14 in a case of a medium without scattering and isotropic point like source ) and the exact value of @xmath14 is less than @xmath28 , provided a strongly anisotropic phase scattering function , low scattering and @xmath29 .
the scattering coefficient @xmath20 is approximated @xcite by @xmath30 where @xmath31 and @xmath21 are the luminosity at distance @xmath22 from the screened and unscreened source , respectively .
our estimation of the total uncertainties due to approximations ( 1,2 ) and systematic errors is : @xmath32 5% for @xmath33 m@xmath13 and @xmath34 10 % for @xmath35 m@xmath13 . in all figures below ,
only statistical errors are shown for the _ asp-15 _ data .
in this section we discuss the results of light absorption measurements performed with both devices _ ac9 _ and _ asp-15_. in table [ tab : abcoeff200 ] and figure [ fig : ablength200 ] we present respectively the absorption coefficients and absorption lengths ( @xmath36 ) as a function of wavelength .
_ asp-15 _ data have been taken at a depth of 200 m , the _ ac9 _ values are the average of data collected at depths between 180 m and 220 m. .absorption coefficients measured during two deployments of _ ac9 _ ( march 28@xmath9 ) and during one deployment of _
( march 28@xmath9 ) at a depth of 200 m._ac9 _ data are averaged over the depth interval 180@xmath10220 m. [ cols="^,^,^,^ " , ] figures [ fig : scatt200 ] and [ fig : scatt1000 ] show the comparison between the scattering coefficients , @xmath3 , measured directly by _ asp-15 _ ( tables [ tab : scatt200 ] and [ tab : scatt1000])and evaluated from the absorption and attenuation coefficients measured by _ ac9 _ : ( @xmath2 from tables [ tab : abcoeff200],[tab : abcoeff1000 ] and @xmath18 from tables [ tab : atcoeff200],[tab : atcoeff1000 ] ) .
figure [ fig : scatt1000 ] shows good agreement between results obtained with _ ac9 _ and _ asp-15 _ at 1000 m depth . at 200 m depth ( figure [ fig : scatt200 ] ) there are discrepancies which confirm the different optical properties of the water layers measured by _ ac9 _ and _ asp-15 _ , already indicated by the results of the absorption measurements at the same depth ( see section [ sec : absorption ] ) . given the strong water currents at shallow depth and inhomogeneous distribution of biologically active substances , a strong variation of optical parameters within one day appears to be realistic .
at last we show in figures [ fig : atlength200 ] and [ fig : atlength1000 ] the values of the attenuation lengths ( @xmath37 ) obtained at depths of 200 m and 1000 m. the values of @xmath18 for _ asp-15 _ are obtained adding the absorption and scattering coefficients reported in tables [ tab : abcoeff200],[tab : abcoeff1000],[tab : scatt200],[tab : scatt1000 ] , while for _ ac9 _ they are measured directly ( see tables [ tab : atcoeff200],[tab : atcoeff1000 ] ) . to evaluate the _
asp-15 _ results at 200 m depth we have used the absorption data measured in march 28@xmath9 and the scattering data measured on march 27@xmath9 . figure [ fig : atlength1000 ] shows good agreement between results from _
ac9 _ and _
asp-15_. similarly to the absorption coefficient , the attenuation coefficient has its smallest value in the region of @xmath38 490 nm for both depths . ) and _ asp-15 _ at 200 m depth .
the values of @xmath18 for _ asp-15 _ are the sum of the values of @xmath2 measured in march 28@xmath9 and the values of @xmath3 measured in march 27@xmath9.,width=340 ] ) and _ asp-15 _ at 1000 m depth .
the values @xmath18 for _ asp-15 _ are the sum of @xmath2 and the values of @xmath3 measured during two runs on april 4@xmath9 and on april 8@xmath9.,width=340 ]
measurements of the optical water properties in lake baikal confirm that the nt-200 telescope is located at optimal depth , where light absorption and attenuation processes are the smallest .
data have been collected with two instruments , which use different measurement principles and have different sources of systematic errors .
data show that , at a depth of 1000 m , the highest transparency is observed for @xmath39 488 nm .
the measured values for absorption length @xmath40 , scattering length @xmath41 and attenuation length @xmath42 at 1000 m depth are : @xmath43 m , @xmath44 m as measured with _
ac9 _ and @xmath45 m , @xmath46 m as measured with _
asp-15_. the depth profile of the absorption coefficient measured by
( see figure [ fig : tempacbaikal ] ) shows the effect of biologically active substances and mineral particulate suspended in water .
this effect is very conspicuous in the depth range 0 @xmath10 400 m ( above the boundary depth of penetration of solar radiation ) , and starts to be visible again for depth higher than 1150 m , near the lake bed .
+ the obtained results demonstrate that the systematic errors are rather small for both instruments and validate the use of both devices to characterize _ in situ _ the inherent optical properties of underwater sites .
this work was supported by the russian ministry of industry , science and technology ( contract 102 - 11(00)-p ) , the russian ministry of education , the german ministry of education and research , russian fund of basic research ( grants 99 - 02 - 1837a , 01 - 02 - 31013 and 00 - 15 - 96794 ) , the russian federal program _ integration _ , the program _ universities of russia _ and unesco chair of water researches and the italian istituto nazionale di fisica nucleare ( infn ) . | measurements of optical properties in media enclosing erenkov neutrino telescopes are important not only at the moment of the selection of an adequate site , but also for the continuous characterization of the medium as a function of time .
over the two last decades , the baikal collaboration has been measuring the optical properties of the deep water in lake baikal ( siberia ) where , since april 1998 , the neutrino telescope nt-200 is in operation .
measurements have been made with custom devices .
the nemo collaboration , aiming at the construction of a km@xmath0 erenkov neutrino detector in the mediterranean sea , has developed an experimental setup for the measurement of oceanographic and optical properties of deep sea water .
this setup is based on a commercial transmissometer . during a joint campaign of the two collaborations in march and april 2001 , light absorption , scattering and attenuation in water
have been measured .
the results are compatible with previous ones reported by the baikal collaboration and show convincing agreement between the two experimental techniques . , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , absorption , ac9 , attenuation , baikal , asp-15 , nemo , neutrino telescope 95.55.vj , 29.40.ka , 92.10.pt , 07.88.+y |
in this section , we aim to prove the following identity by induction , @xmath119 the construction of the walsh functions is simple in terms of the elementary sequences known as rademacher functions @xmath120 $ ] .
the dyadic ordering of the walsh functions allow them to be defined in terms of the rademacher functions as @xmath121 when @xmath69 is expressed as a binary number @xmath122 and @xmath123 or @xmath30 . with this definition , it is easy to see that choosing the index @xmath124 for the walsh function means that all the @xmath125 coefficients are @xmath30 .
we now prove the base case , @xmath126 .
@xmath127 for the inductive step , assume that ( [ eq : walsh : coefficients ] ) is true and look at the @xmath128 case : @xmath129 for the next step , note that for an integer @xmath130 , @xmath131 , which allows the expression to be written as , @xmath132 in the next step , the substitution @xmath133 is made and we take advantage of the fact that @xmath134 for @xmath135 . @xmath136 which is zero by assumption since @xmath137 is a constant , thus concluding the proof .
if the radamacher functions are considered to be elementary sequences , then the walsh filter is optimized in this resource for the task of suppressing the errors in the spin - dependent force operation that are discussed in this paper . | we demonstrate a simple pulse shaping technique designed to improve the fidelity of spin - dependent force operations commonly used to implement entangling gates in trapped - ion systems .
this extension of the mlmer - srensen gate can theoretically suppress the effects of certain frequency and timing errors to any desired order and is demonstrated through walsh modulation of a two - qubit entangling gate on trapped atomic ions .
the technique is applicable to any system of qubits coupled through collective harmonic oscillator modes .
the use of spin - dependent forces to create entangled quantum systems has become widespread @xcite and is currently the technique used for the highest fidelity multi - qubit operations @xcite .
this powerful technique , first proposed in @xcite , has been used to implement quantum algorithms @xcite , create large entangled states @xcite , test quantum fundamentals @xcite , and perform simulations of quantum magnetism @xcite and quantum field theory @xcite . as these types of experiments are scaled to larger numbers of qubits and more complex algorithms , the accumulation of gate errors
will eventually require quantum error correction . because of the large overhead required for quantum error correction , it is important that qubit operations be optimized passively in terms of speed and robustness to non - ideal control environments . in this paper
, we show how ideas similar to the spin - echo pulse sequence @xcite and those developed in the context of dynamical decoupling @xcite can be used to optimize the mlmer - srensen ( ms ) gate that is based on the spin - dependent force .
spin - dependent force gates operate by coupling the qubit states to a collective external degree of freedom referred to as a quantum bus .
the coupling is switched on for an amount of time that introduces a particular phase between the spin states and at the same time leaves them disentangled from the external degree of freedom at the end of the gate .
while the relative spin phase is fairly robust due to geometric features @xcite , the disentanglement of the qubit space and the quantum bus at the end of the operation may be more susceptible to experimental errors and is equally crucial to achieving a high fidelity gate .
imperfect timing caused by noise on the energy splitting of the qubit can be suppressed by the insertion of an additional swapping pulse on the qubit states in the middle of a two qubit gate @xcite or , as proposed in @xcite , by a @xmath0 phase shift in the drive field . in this letter , we present new analytic results which generalize these ideas and show how frequency and timing errors can theoretically be suppressed to any desired order with an optimized gate sequence that does not rely on the insertion of additional @xmath0 pulses within the gate . furthermore , the technique is demonstrated using atomic hyperfine qubits driven by a stimulated raman process and shown to perform much better than the standard operation described in the original proposal @xcite .
similar to the single - qubit composite pulses @xcite originally designed for error suppression in nmr experiments now being in widespread use in other quantum information systems , this composite pulse should be applicable to any system of qubits coupled to a driven harmonic oscillator such as superconducting flux qubits @xcite or cavity qed @xcite . in trapped - ion systems , the spin - dependent force couples internal atomic states of neighboring ions through the collective modes of motion generated by the coulomb interaction . in the ms scheme ,
a spin - dependent force is created by off - resonantly driving the first - order red and blue sideband transitions simultaneously .
the interaction hamiltonian takes the form @xmath1 where @xmath2 is the sideband transition frequency , @xmath3 are the raising and lowering operators for the qubit , @xmath4 are the creation and annihilation operators for the collective harmonic oscillator mode , and @xmath5 is the symmetric detuning of the drive field from the sidebands @xcite .
the sum phase @xmath6 of the red and blue sideband drive fields determines the eigenstates of the spin operator in @xmath7 .
the difference phase @xmath8 determines the phase of the time - dependent displacement of the motional state . for the general case of @xmath9 ions ,
the time - evolution operator is given by , @xmath10}\\ & = & e^{\hat{s}_n(\alpha(t)\hat{a}^{\dagger}-\alpha^*(t)\hat{a})}e^{-i\phi(t)\hat{s}_n^2 } , \label{eq : time : evolution}\end{aligned}\ ] ] where the total spin operator is given by @xmath11 , the time - dependent displacement coefficient is @xmath12 and @xmath13 is a time - dependent phase that depends only on @xmath2 and @xmath14 .
when a collection of trapped ions that are each identically prepared in an eigenstate of @xmath15 evolves according to ( [ eq : time : evolution ] ) , the spin - dependent displacement operator splits the motional wavepacket into @xmath16 pieces that execute circular trajectories in phase space according to the definition of @xmath17 . the term in ( [ eq : time : evolution ] ) proportional to @xmath18 imprints a relative phase on the various spin states , allowing the operation to be used as an entangling operation . in order to prepare a pure spin state with this type of operation ,
the entanglement between the spin and motion must disappear at the end of the gate .
when the gate time @xmath19 is not equal to @xmath20 , where @xmath21 is any non - zero integer , the motional wavepackets do not trace out closed trajectories in phase space and therefore will not be fully disentangled from the spin state . the required level of precision grows with higher temperatures since the overlap between two states separated in phase space decreases exponentially with temperature . to see this , consider a qubit under the influence of the time evolution operator in ( [ eq : time : evolution ] ) .
if the initial motional state is assumed to be a gaussian state @xmath22 with an uncertainty in position @xmath23 and we describe a small timing or detuning error in the gate operation as an unintentional momentum displacement @xmath24 , then the overlap between the two motional states is given by @xmath25 $ ] .
for a harmonic oscillator in a thermal state , @xmath23 increases approximately as @xmath26 for @xmath27 meaning that the overlap between the two states decreases exponentially . as shown in fig .
[ fig : phase : space](a ) , a small detuning error can largely be corrected with a second pulse whose phase has been shifted by @xmath0 .
we now discuss how to generalize this simple pulse sequence in order to suppress larger errors of this type .
spin - dependent force operation with a small detuning error , @xmath28 .
the solid and dashed curves show the two different trajectories taken by the two different wavepackets associated with spin up and spin down in the @xmath29 basis with only the spin up trajectory being labeled for clarity . after being initialized to a state centered at the origin ,
the spin - up wave - packet begins its clockwise motion near the point labeled @xmath30 .
halfway through the operation , near the point @xmath31 , the phase of the drive field is advanced by @xmath0 , changing the direction of the applied force . at the end of the gate , near point @xmath32
, the wave - packet ends up much closer to the origin than the turning point near point 2 .
therefore , the two wave - packets have more overlap at the end of the gate which is the key to achieving a higher fidelity operation .
( b)the walsh functions @xmath33 , @xmath34 and @xmath35 are shown .
notice that @xmath36 can be constructed as two sequential @xmath37 functions with a phase flip on the second pulse .
likewise , @xmath38 can be constructed as sequential @xmath36 pulses with a phase flip on the second pulse.,width=326 ] is subjected to the standard and composite spin - dependent force operations and then measured in the @xmath15 basis .
the data shown are plotted together with theoretical curves assuming an initial thermal state of motion with the average excitation number @xmath39 .
( a ) the data show the probability of finding the ion in @xmath40 as a function of the symmetric detuning @xmath14 for @xmath41sec . on resonance , @xmath42 ,
the motional wave - packets quickly become entangled with the spin state , resulting in a maximally mixed spin state . for finite @xmath14
, the wavepackets trace out circles in phase space resulting in partial revivals of the initial spin state when @xmath43 is a non - zero integer .
( b ) the spin - dependent force operation is implemented using @xmath44 for the phase @xmath45 .
( c ) @xmath46 is used for the phase @xmath45 .
note the narrow resonance at @xmath47 corresponds to a trajectory where the phase flips occur when the motional wavepackets are not at the origin.,title="fig:",width=153 ] is subjected to the standard and composite spin - dependent force operations and then measured in the @xmath15 basis .
the data shown are plotted together with theoretical curves assuming an initial thermal state of motion with the average excitation number @xmath39 .
( a ) the data show the probability of finding the ion in @xmath40 as a function of the symmetric detuning @xmath14 for @xmath41sec . on resonance , @xmath42 ,
the motional wave - packets quickly become entangled with the spin state , resulting in a maximally mixed spin state . for finite @xmath14
, the wavepackets trace out circles in phase space resulting in partial revivals of the initial spin state when @xmath43 is a non - zero integer .
( b ) the spin - dependent force operation is implemented using @xmath44 for the phase @xmath45 .
( c ) @xmath46 is used for the phase @xmath45 .
note the narrow resonance at @xmath47 corresponds to a trajectory where the phase flips occur when the motional wavepackets are not at the origin.,title="fig:",width=153 ] is subjected to the standard and composite spin - dependent force operations and then measured in the @xmath15 basis .
the data shown are plotted together with theoretical curves assuming an initial thermal state of motion with the average excitation number @xmath39 .
( a ) the data show the probability of finding the ion in @xmath40 as a function of the symmetric detuning @xmath14 for @xmath41sec . on resonance , @xmath42 ,
the motional wave - packets quickly become entangled with the spin state , resulting in a maximally mixed spin state . for finite @xmath14
, the wavepackets trace out circles in phase space resulting in partial revivals of the initial spin state when @xmath43 is a non - zero integer .
( b ) the spin - dependent force operation is implemented using @xmath44 for the phase @xmath45 .
( c ) @xmath46 is used for the phase @xmath45 .
note the narrow resonance at @xmath47 corresponds to a trajectory where the phase flips occur when the motional wavepackets are not at the origin.,title="fig:",width=153 ] suppose there is a symmetric error @xmath48 in the detuning such that @xmath49 that could be the result of a change in the trapping frequency .
the error in the operation results in some residual entanglement between the spin and motion that can be quantified by the magnitude of @xmath50 , ( which goes to zero for @xmath51 at @xmath52 ) .
we will show that by switching either @xmath45 or @xmath53 between @xmath54 and @xmath0 at times prescribed by certain walsh functions , the effect of @xmath48 on the magnitude of @xmath55 can be suppressed to any order .
a walsh function , denoted here as @xmath56 , is a piecewise constant function that alternates between the values @xmath57 at certain values of @xmath58 depending on the dyadic - ordered index @xmath59 @xcite , ( see fig .
[ fig : phase : space](b ) ) .
if @xmath60 and @xmath61 shift together between @xmath54 and @xmath0 , then @xmath45 shifts between @xmath54 and @xmath0 , but @xmath53 remains constant and can be assumed to be @xmath54 without the loss of generality .
note the effect of the phase shift @xmath62 is equivalent to shifting the motional phase @xmath63 while keeping @xmath45 constant .
both of these phase shifts are equivalent to the mapping @xmath64 , which can also be achieved with @xmath0 pulses on the qubit states as done in @xcite .
although the @xmath0 phase shifts and @xmath0 rotations are ideally equivalent , the phase shift switching time and precision is limited by electronics whereas the microwave rotations depend on qubit control that might be subject to the same noise source that generates @xmath48 . if the times at which phase shifts occur is determined by the zero crossing times of @xmath65 , then @xmath66 .
when modulating @xmath45 in this manner , the new displacement operator in ( [ eq : time : evolution ] ) is given by , @xmath67 by choosing a walsh function with index @xmath68 where @xmath69 is an integer and a detuning @xmath70 , phase flips only occur at integer multiples of @xmath71 and the effect of @xmath48 can be suppressed to any order .
this statement rests on the following equality , @xmath72 where @xmath73 is a constant , ( proof in the supplemental material ) .
if the function @xmath74 is expanded in a taylor series , the identity in ( [ eq : walsh : coefficients ] ) ensures that the displacement operator will be given by @xmath75 , where @xmath76 is the identity operator .
is compared for the first three walsh functions being used for @xmath45 . in all three data sets ,
the ions were sideband cooled to the motional ground state before implementing the gate and @xmath77ms . in figures ( a ) , ( b ) and ( c ) the measured state fidelity as a function of the detuning @xmath14 is compared with the theoretical curves . to account for other imperfections in the experiment ,
the theoretical curves have an overall scale factor .
the data shows a maximum fidelity of @xmath78 for the standard pulse , @xmath79 for the @xmath44 sequence and @xmath80 for @xmath46 .
the characteristic width of the high fidelity region is also clearly larger for the higher order pulse sequences as predicted by the theory .
( d ) shows the fidelity curves for @xmath81 and @xmath82 .
the curves are shifted in frequency in order to facilitate comparison and highlight the increasingly large regions of high fidelity for the higher order sequences.,title="fig:",width=153 ] is compared for the first three walsh functions being used for @xmath45 . in all three data sets ,
the ions were sideband cooled to the motional ground state before implementing the gate and @xmath77ms . in figures ( a ) , ( b ) and ( c ) the measured state fidelity as a function of the detuning @xmath14 is compared with the theoretical curves . to account for other imperfections in the experiment ,
the theoretical curves have an overall scale factor .
the data shows a maximum fidelity of @xmath78 for the standard pulse , @xmath79 for the @xmath44 sequence and @xmath80 for @xmath46 .
the characteristic width of the high fidelity region is also clearly larger for the higher order pulse sequences as predicted by the theory .
( d ) shows the fidelity curves for @xmath81 and @xmath82 .
the curves are shifted in frequency in order to facilitate comparison and highlight the increasingly large regions of high fidelity for the higher order sequences.,title="fig:",width=153 ] is compared for the first three walsh functions being used for @xmath45 . in all three data sets ,
the ions were sideband cooled to the motional ground state before implementing the gate and @xmath77ms . in figures
( a ) , ( b ) and ( c ) the measured state fidelity as a function of the detuning @xmath14 is compared with the theoretical curves . to account for other imperfections in the experiment ,
the theoretical curves have an overall scale factor .
the data shows a maximum fidelity of @xmath78 for the standard pulse , @xmath79 for the @xmath44 sequence and @xmath80 for @xmath46 .
the characteristic width of the high fidelity region is also clearly larger for the higher order pulse sequences as predicted by the theory .
( d ) shows the fidelity curves for @xmath81 and @xmath82 .
the curves are shifted in frequency in order to facilitate comparison and highlight the increasingly large regions of high fidelity for the higher order sequences.,title="fig:",width=153 ] is compared for the first three walsh functions being used for @xmath45 . in all three data sets ,
the ions were sideband cooled to the motional ground state before implementing the gate and @xmath77ms . in figures ( a )
, ( b ) and ( c ) the measured state fidelity as a function of the detuning @xmath14 is compared with the theoretical curves . to account for other imperfections in the experiment ,
the theoretical curves have an overall scale factor .
the data shows a maximum fidelity of @xmath78 for the standard pulse , @xmath79 for the @xmath44 sequence and @xmath80 for @xmath46 .
the characteristic width of the high fidelity region is also clearly larger for the higher order pulse sequences as predicted by the theory .
( d ) shows the fidelity curves for @xmath81 and @xmath82 .
the curves are shifted in frequency in order to facilitate comparison and highlight the increasingly large regions of high fidelity for the higher order sequences.,title="fig:",width=153 ] to demonstrate the power of the composite pulse sequence , we use a qubit defined as the clock states in the @xmath83 hyperfine manifold of a @xmath84 ion in an rf paul trap which can be initialized and read out using the techniques described in @xcite .
these states , @xmath85 , have a splitting of @xmath86 ghz and can be coupled to each other directly or through the harmonic oscillator using stimulated raman transitions .
as described in @xcite , the raman transition induced spin - dependent forces are created by the beat notes between two optical frequency combs that are generated by a @xmath87 nm mode - locked pulsed laser .
the uv pulses have a duration of @xmath88 psec at a repetition rate of 80.57 mhz and have a center wavelength that is optimal for minimizing off - resonant scattering from the excited @xmath89 states in @xmath84@xcite . at the position of the ion , the two beams are cross - polarized and mutually orthogonal to a magnetic field of @xmath90 g with a geometry such that the momentum kicks associated with the stimulated raman process only excite the transverse modes of motion which have a resonance frequency of @xmath91 mhz .
the two raman beams are frequency shifted with aoms to set up the appropriate beat notes in the interference field at the location of the ions . as described in @xcite , driving one aom with a single frequency and the other with two frequencies generates the bichromatic beat note that gives rise to the ms interaction .
the red and blue phases @xmath92 are therefore defined by the phases of these two rf drive frequencies . the composite pulse is implemented by splitting the operation into segments , between which the phases @xmath60 and @xmath61 are shifted . in this setup
, symmetric detuning errors might be the result of fluctuating rf trap voltages which manifests itself as noise on the oscillation frequency .
asymmetric detuning errors arise from a change in the qubit splitting which might stem from a noisy local magnetic field .
the effect of walsh modulation on the spin - dependent force can be plainly seen with a single ion . in the case of a single ion ,
the phase @xmath13 is global and the only operation that results in a pure spin state is one that simply restores the initial spin state .
the disentanglement of the spin and motion and consequential revival of the initial spin state should occur when @xmath93 where @xmath94 is the walsh function index .
if the spin is initialized to @xmath95 in the @xmath15 basis and the joint spin - motion state after the operation is @xmath96 , then the state - fidelity is @xmath97 $ ] . ignoring heating effects and assuming an initial thermal state of motion , the fidelity is @xmath98\right)$ ] where @xmath99 is the average excitation number of the harmonic oscillator and @xmath100 . because the lowest order term for the infidelity is @xmath101 , the infidelity of the walsh modulated operation at @xmath93 is @xmath102 .
this effect is clearly seen in fig .
[ fig : one : ion ] where the higher order walsh sequences exhibit spin revivals of high purity over a much larger range of detunings .
the increase in the width of the spin revival regions for the higher order walsh sequences is evidence of the higher tolerance of small detuning errors .
note the same arguments apply to the case of small timing errors and errors resulting from an asymmetric detuning from the red and blue sidebands such as a change in the qubit splitting .
the effect of the walsh modulation on a two - qubit gate is more complicated than that of a single qubit operation since the term in ( [ eq : time : evolution ] ) proportional to @xmath103 must be taken into account .
the walsh modulation of @xmath45 changes the evolution of @xmath13 in general , but not in the case where @xmath70 since the evolution is a series of closed circles in phase space . in this case ,
@xmath104 and a fully entangling operation is achieved when @xmath105 .
this implies that in order to use @xmath106 , the gate time must be at least @xmath107 .
while the exponential nature of this composite gate becomes daunting for large @xmath69 , small errors can easily be corrected with a modest increase in the gate time . in the case of two ions ,
the maximally entangling operation ideally implements the transformation @xmath108 where the phase @xmath109 is determined by the phase of the drive field . with this target state ,
the fidelity is @xmath110 and is measured in the same manner as described in @xcite .
the phase @xmath111-\frac{1}{\delta}\left(t_g-\frac{1}{\delta}\sum_{i=0}^{k}\mathrm{sin}(\delta t_i)\right)$ ] is written here in terms of sums over the different parts of a pulse sequence .
the parameters @xmath112 refer to the duration of the @xmath113 segment of a sequence and the parameters @xmath114 with @xmath115 and @xmath116 .
the data shown in fig .
[ fig : two : ions ] compare the state fidelity of two ions using the composite pulse sequences for the maximally entangling gate and shows that the standard pulse sequence is outperformed by the walsh modulated pulse sequences in terms of both the maximum fidelity and the characteristic width of the high fidelity region .
the value of @xmath2 is used as a fit parameter for the theoretical curves with the same value , @xmath117 khz , used in all three plots .
walsh functions have long been known by the electrical engineering , astronomy and radio communications communities to have useful error correcting properties @xcite . while the walsh functions are not the only option for choosing how to modulate the drive field of the spin - dependent force gate , we hope their introduction in the context of quantum control provides a useful tool for the further development of dynamical decoupling and related areas . in the formalism of dynamical decoupling
, the function @xmath118 can be viewed as an optimized filter function designed to suppress the effects of a noise source centered at @xmath5 @xcite .
walsh modulation is optimal in a different sense than that of the uhrig sequence which can , with an increasing number of pulses , suppresses higher moments of the noise spectrum in the zero frequency limit @xcite . while the uhrig filter is optimal in the number of pulses used for a given order of noise suppression , the walsh filters are optimal in the number of elementary sequences , ( see supplemental material for the definition of an elementary sequence ) .
this minimal number of elementary sequences not only allows for a simple mathematical construction , but also means that these functions are easy to synthesize using simple integrated circuits . by introducing the idea of a walsh modulated spin - dependent force , we have shown theoretically and experimentally that it is possible to suppress errors that are linked to the residual entanglement between the spin and motion , thereby alleviating the required precision of the control fields .
because the detrimental effect of any error increases exponentially with the initial temperature of the harmonic oscillator , the technique may also decrease the amount of resource intensive cooling that must be done in order to achieve a high fidelity operation . as quantum information experiments progress
, this technique of coherent error suppression in quantum bus operations might prove to be an important ingredient in scaling toward larger systems and more complex algorithms .
we acknowledge useful discussions with peter maunz , daniel brennan and michael biercuk .
this work is supported by the army research office ( aro ) with funds from the darpa optical lattice emulator ( ole ) program , iarpa under aro contract , the nsf physics at the information frontier program , and the nsf physics frontier center at jqi . [ 1 ] c. a. sackett et al .
nature * 404 * , 256 ( 2000 ) .
g. kirchmair et al .
new j. phys . * 11 * , 023002 ( 2009 ) .
p.
j. lee et al .
. lett .
* 28 * , 1582 ( 2003 ) .
j. p. home et al .
new j. phys .
* 8 * , 188 ( 2006 ) .
j. benhelm et al .
nature physics * 4 * , 463 ( 2008 ) .
k. mlmer and a. srensen .
phys .
rev .
lett . * 82 * , 1835 ( 1999 ) .
e. solano et al .
phys .
rev .
a * 59 * , r2539 ( 1999 ) .
g. j. milburn et al .
. phys . * 48 * , 801 ( 2000 ) .
r. blatt and d. wineland .
nature * 453 * , 1008 ( 2008 ) .
t. monz et al .
arxiv 1009.6126v1 ( 2010 ) .
m. a. rowe et al .
nature * 409 * , 791 ( 2001 ) .
g. kirchmair et al .
nature * 460 * , 494 ( 2009 ) .
k. kim et al .
nature * 465 * , 590 ( 2010 ) .
a. friedenauer et al .
nature physics * 4 * , 757 ( 2008 ) .
r. gerritsma et al .
nature * 463 * , 68 ( 2009 ) .
h. y. carr and e. m. purcell .
phys . rev . * 94 * , 630 ( 1954 ) .
l. viola and s. lloyd .
phys .
rev .
a * 58 * , 2733 ( 1998 )
. g. s. uhrig .
phys .
rev .
lett . * 98 * , 100504 ( 2007 ) .
m. j. biercuk et al .
phys .
rev .
a * 79 * , 062324 ( 2009 ) .
s .- l .
zhu and z. d. wang .
phys .
rev .
lett . * 91 * , 187902 ( 2003 ) .
j. d. jost et al .
nature * 459 * , 683 ( 2009 ) .
c .- y .
chen and m. feng .
phys .
rev .
a * 77 * , 012325 ( 2008 ) .
h. k. cummins et al .
phys .
rev .
a * 67 * , 042308 ( 2003 ) .
y .- d .
wang , a. kemp , and k. semba .
phys .
rev .
b * 79 * , 024502 ( 2009 )
. k. g. beauchamp .
_ applications of walsh and related functions _ ( 1984 ) .
s. olmschenk et al .
phys .
rev .
a * 76 * , 052314 ( 2007 ) . d. hayes et al .
phys .
rev .
lett . * 104 * , 140501 ( 2010 ) .
l. cywinski et al .
phys .
rev .
b * 77 * , 174509 ( 2008 ) . a. r. thompson , j. m. moran , and g. w. swenson . _
interferometry and synthesis in radio astronomy _ , p. 245
( 2001 ) .
w. c. campbell et al .
phys .
rev .
lett . * 105 * , 090502 ( 2010 ) . |
the concept of quantum coherence and interference plays an important role in resonant nonlinear optics and spectroscopy @xcite .
coherent coupling of laser radiation with multilevel quantum schemes has been used to manipulate energy level populations , nonlinear - optical response , refraction and absorption of a resonance medium @xcite .
much attention has been shown to four - wave mixing ( fwm ) processes with maximal coherence on a raman transition @xcite based on the effect of coherence population trapping ( cpt ) @xcite , which allows a dramatic increase of conversion efficiency . in far - from - degenerate schemes , substantial doppler broadening of the resonant quantum transitions gives rise to inhomogeneous coupling of the driving fields with the atoms in different velocity intervals . therefore , under relatively low intensities , only a small fraction of the atoms can be concurrently coupled with both the fields driving the raman transition and , as a consequence , the cpt conditions can not be fulfilled for the entire thermal velocity interval . this fundamental limitation in resonant nonlinear optics of gases substantially reduces the efficiency of the corresponding processes .
the common methods of doppler - free ( df ) nonlinear spectroscopy entail ladder energy level schemes and equal frequency counter - propagating beams .
this usually leads to large detunings from the intermediate resonance and consequently a decrease of the nonlinear susceptibility . besides that
, the method can not be applied to fwm processes in ladder schemes , due to the phase - matching requirements , and in raman - type schemes due to the difference in frequencies of the coupled fields .
the technique , allowing one to overcome this obstacle by compensating for doppler frequency shifts with light - induced ac - stark shifts and thus enabling one to couple a wide velocity interval concurrently , has been proposed in @xcite .
this was developed in further publications @xcite and most recently in publications @xcite addressing fwm under conditions where cpt is not possible . related effects in the absorption index
were also investigated recently @xcite .
we further mean cpt as the process where two driving fields form nearly equal velocity - averaged populations of the levels at raman transitions , while the velocity - averaged population of the intermediate state is much less .
along with this absorption and amplification indices for the couple fields reach minimum . on the other hand coherence in the raman transition , which is origin of fwm ,
approaches maximum .
that allows one to accomplish a maximum of the conversion efficiency .
the conditions to achieve cpt may differ from those required for the df raman - type resonance .
this also addresses the problem `` nonlinear optics at the level a few photons per atom '' @xcite .
0.5 consider a raman - type scheme of fwm , @xmath0 , depicted in fig .
[ fig1 ] , where @xmath1 and @xmath2 with frequencies @xmath3 and @xmath4 are traveling waves of the driving fields .
the field @xmath5 with the frequency @xmath6 and generated radiation @xmath7 with the frequency @xmath8 are assumed weak ( i.e. , not changing the level populations ) .
all waves are co - propagated .
initially , only the lower level @xmath9 is populated .
the steady - state solution for the off - diagonal elements of the density matrix ( atomic coherence ) can be found in the same form of traveling waves as the the resonant driving radiations .
then the equations for the density matrix amplitudes to the lowest order in @xmath10 can be written as @xmath11,\ , p_{21}r_{21}=-i\left[g_{21}(r_{1}-r_{2})- g_{01}r_{20}\right],&\nonumber\\ & p_{02}r_{02}=-i\left[g_{01}r_{12}-g_{12}r_{01}\right];&\\ & \gamma_{1}r_{1}=2{\hbox{re\,}}[ig_{12}r_{21}+ig_{10}r_{01}],\ , \gamma_{2}r_{2}=-2{\hbox{re\,}}[ig_{12}r_{21}]+\gamma_{2}r_{1},&\nonumber\\ & r_{0}+r_{1}+r_{2}=1;&\label{eq1}\\ & p_{23}r_{23}=-i\left[-g_{23}r_{2}+ g_{21}r_{13}-g_{03}r_{20}\right],&\nonumber\\ & p_{03}r_{03}=-i\left[-g_{03}r_{0}+ g_{01}r_{13}-g_{23}r_{02}\right],\nonumber & \\ & p_{13}r_{13}=-i\left[g_{12}r_{23}-g_{23}r_{12}+ g_{10}r_{03}-g_{03}r_{10}\right]\ , & \end{aligned}\ ] ] here @xmath12 and @xmath13 are the amplitudes of the off - diagonal and diagonal elements of the density matrix , @xmath14 , @xmath15 is the resonance detuning for the corresponding resonant field ( e.g. , @xmath16 ) , @xmath17 are homogeneous half - widths of the transitions ( coherence relaxation rates ) , @xmath18 is the projection of the atom velocity on the direction of the wave vectors @xmath19 , @xmath20 are relaxation rates from level @xmath21 to @xmath9 and from @xmath22 to @xmath23 , accordingly , @xmath24 are relaxation rates of the populations of the levels @xmath22 and @xmath23 accordingly , @xmath25 are the rabi frequencies , and @xmath26 are the electric dipole momenta of the transitions . only the resonant couplings are accounted for .
the solution for the coherence @xmath27 and for those at the allowed transitions , which determine absorption and refraction at the frequencies @xmath3 , @xmath4 , @xmath6 , are found in the form @xmath28 where @xmath29 are intensity - dependent population differences , and @xmath30 and @xmath31 are two - photon denominators `` dressed '' by the driving fields : @xmath32 the coherence @xmath33 consists of two terms .
one , @xmath34 , determines absorption and refraction at the frequency @xmath8 , and the other , @xmath35 , determines fwm at @xmath36 : @xmath37 where @xmath38 and @xmath39 making use of the above equations , the solution for the populations can be found from ( [ eq1 ] ) as @xmath40 where @xmath41 and @xmath42 the appearance of a df resonance and therefore the coupling of molecules from a wide velocity interval can be understood as follows .
@xmath43 and @xmath44 in ( [ den ] ) , in the limiting case , when the detuning from the intermediate resonance is much greater than the doppler hwhm of the allowed optical transitions , the modified two - photon resonance described by the denominator @xmath45 can be presented in the lowest order in @xmath46 as @xmath47 where @xmath48 are the power broadened hwhm and ac - stark shifted two - photon resonance detuning . from ( [ p02 ] )
it follows that the requirements for the induced doppler - free resonance to be achieved are @xmath49 and @xmath50 , where + @xmath51 and @xmath52 + are intensity - dependent doppler shifts .
eventually the equations take form @xmath53 from ( [ eq3])-([eq4 ] ) follows a cubic equation as @xmath54 : @xmath55 which determines the detuning of field @xmath2 , corresponding to the induced resonance , under given values of the rabi frequency and detuning of field @xmath1 . here , the rabi frequency of @xmath2 must be @xmath56 one of the roots of the equation ( [ kub ] ) corresponds to the dressed df two - photon resonance , @xmath57 where @xmath58 @xmath59^{1/2}$ ] , + @xmath60 the conditions for the cpt and df resonances may fit each other or differ , depending on the specific case
. this and corresponding outcomes regarding fwm will be investigated below numerically for the most optimum situations , where the analytical solution can not be obtained .
the graphs presented below are computed numerically based on averages over the velocity equations ( [ chi1 ] ) , ( [ chi2 ] ) and ( [ r ] ) . for the numerical analysis we have used the parameters of the sodium dimer transition with the following wavelengths @xcite : @xmath61 @xmath62 nm )
@xmath63 @xmath64 nm ) @xmath65 nm ) @xmath66 nm ) .
the corresponding homogeneous half - widths of the transition are 20.69 , 23.08 , 18.30 and 15.92 mhz , whereas the doppler half - widths @xmath67 are equal to 0.92 , 0.85 , 0.68 and 0.75 ghz .
the numerical simulations allow us to analyze the velocity - averaged equation for the detunings , where the approximation taken in ( [ p02 ] ) is not valid .
normalized squared module of the fwm polarizations @xmath68 ( a ) and @xmath69 ( b ) vs the detuning of driving field @xmath2 .
the plots correspond to : 1 df resonance , 2 cpt , 3 intermediate case , subplot same resonance , but under weak fields.,title="fig : " ] normalized squared module of the fwm polarizations @xmath68 ( a ) and @xmath69 ( b ) vs the detuning of driving field @xmath2 .
the plots correspond to : 1 df resonance , 2 cpt , 3 intermediate case , subplot same resonance , but under weak fields.,title="fig : " ] figure [ figpom ] , plots 1 @xmath70 [ fig.[figpom ] ( * b * ) ] . the nonlinear susceptibility is reduced by its maximum value in the same frequency range , but for negligibly - weak @xmath1 and @xmath2 fields , @xmath71 , @xmath72 .
, presenting the spectral dependence of the velocity - averaged squared module of the same susceptibility , but in the negligibly weak fields , normalized to unity .
the hwhm of the resonance in the subplot is approximately 70 mhz , which corresponds to the doppler width of the nonperturbated raman transition , whereas the hwhm of the resonance in plots 1 is 15 mhz .
this indicates some power and residual doppler broadening .
the rabi frequencies of the fields @xmath1 and @xmath2 are equal to 157 mhz and 85 mhz .
the detuning of @xmath1 is equal to 413.8 mhz ( @xmath73 ) , and detuning of @xmath5 is equal to -695.5 mhz ( @xmath74 ) .
plots 2 display the same dependences , but related to the cpt conditions ( the half - width of this resonance in @xmath75 is 200 mhz ) . in this case , the required rabi frequencies of @xmath1 and @xmath2 are equal to 351 mhz and 332 mhz , respectively , while the detunings are identical to those in plots 1 .
plots 3 display the same dependence , but under the intensities , that provide narrowing in the intermediate range ( hwhm of resonance is 70 mhz ) . in this case , the rabi frequencies of the fields @xmath1 and @xmath2 are equal to 222 mhz and 235 mhz , respectively , whereas the detunings are the same as in the previous plots . also , we want to stress the substantial difference in magnitudes of the nonlinear susceptibilities @xmath76 and @xmath77 , which follows from the interference of contributions of molecules in different velocity intervals to the macroscopic nonlinear polarizations .
velocity integrated population of the energy levels vs detuning of driving field @xmath2 .
a near cpt and cpt regime , subplot velocity distribution of the coherently trapped populations ; b near df regime , subplot velocity distribution of populations under df regime.,title="fig : " ] velocity integrated population of the energy levels vs detuning of driving field @xmath2 .
a near cpt and cpt regime , subplot velocity distribution of the coherently trapped populations ; b near df regime , subplot velocity distribution of populations under df regime.,title="fig : " ] figure [ figpop](a ) shows dependence of the velocity integrated populations on the detuning of the field @xmath2 , while the rabi frequencies of @xmath1 and @xmath2 are equal to 351 mhz and 332 mhz , respectively so that the cpt conditions can be fulfilled under appropriate detuning .
the subplot displays the distribution of the populations over velocities , whereas the detuning corresponds to the cpt regime for the velocity - integrated populations . here
@xmath78 is the thermal velocity with maxwell envelope removed .
figure [ figpop](b ) is computed for the rabi frequencies of the fields @xmath1 and @xmath2 equal to 157 mhz and 85 mhz , respectively .
the inset displays the distribution of the populations over velocities at the detuning corresponding to df .
the distribution is even more complicated compared to that in the previous case .
figure [ figalp ] displays velocity - averaged absorption indices vs detuning of the field @xmath2 .
velocity averaged absorption indices @xmath79 near df ( a ) and cpt ( b ) regimes vs detuning of the driving field @xmath2 .
, title="fig : " ] velocity averaged absorption indices @xmath79 near df ( a ) and cpt ( b ) regimes vs detuning of the driving field @xmath2 .
, title="fig : " ] the intensities of the fields are the same as in the corresponding graphs given above . the graph [ figalp ] ( a ) shows a substantially decreased absorption index for the first field due to saturation effects , which is almost independent of the detuning of the second field because of its relative weakness .
the second field experiences stokes gain .
the plot for the generated beam ( reduced by 10 ) displays a nonlinear resonance in the absorption index . with growth of the second field intensity , so that cpt becomes possible , both absorption of the first and gain of the second driving field experience substantial decrease in the cpt regime .
absorption of the generated field dramatically changes as well ( figure [ figalp ] ( a ) ) .
a linear approximation in @xmath46 ( see equation ( [ p02 ] ) ) is valid for large detunings from the intermediate resonance .
while for tuning close to the resonance , the concurrently coupled velocity interval decreases , the nonlinear response of the molecules grows . at the same time , absorption of the coupled radiations increases as well .
the output fwm generated radiation is determined by the interplay of the above considered processes . moreover
, resonant fwm may not be viewed as a sequence of the independent elementary acts of absorption , gain and fwm .
interference of these elementary processes plays a crucial role , even for a qualitative prediction of the spatial dynamics of the generated field @xcite .
therefore , appropriate optimization is required in order to achieve the maximum fwm - generation output . in the next section ,
relevant results are presented based on the above analyzed dependencies .
the solution of the maxwell equations for the coupled traveling waves can be found in the form @xmath81\},\ ] ] where @xmath82 are complex wave number at the corresponding frequencies , @xmath83 .
the set of the equations for the amplitudes , relevant to the case under consideration , is @xmath84 where @xmath85 and @xmath86 are cross coupling parameters , @xmath87 and @xmath88 are intensity - dependent nonlinear susceptibilities for the fwm processes @xmath89 and @xmath90 .
the boundary conditions are @xmath91 ( @xmath92 ) and @xmath93 .
we assume that intensities of the weak waves with frequencies @xmath6 and @xmath8 are weak enough so that the change of the strong fields @xmath94 due to fwm conversion can be neglected .
the quantum efficiency of conversion qec of the radiation @xmath5 in @xmath95 is defined by the equation @xmath96 assuming that the condition of phase matching @xmath97 can be ensured ( e.g. , with a buffer gas ) , the solution for qec of the above given equations can be expressed as @xmath98 \bigl|1-\exp(-g_{0}z)\bigr|^{2},&\nonumber \\ & g_{0}=\left[\left [ ( \alpha_{1}+\alpha_{2}+\alpha_{3}-\alpha_{s}/2)\right]^{2}+ 4\sigma_{s}\sigma_{3}\bigl|e_{01}e_{02}\bigl|^2\right]^{1/2 } , \quad \alpha=\alpha_{3}+\alpha_{s}. & \nonumber\end{aligned}\ ] ] these formulas account for nonlinear resonances both in fwm nonlinear polarizations and in absorption indices . in order to compute the absolute magnitude of qec ( under the assumption of @xmath97 ) we have used the data for the frank - condon factors of 0.068 , 0.142 , 0.02 and 0.036 for the transition with wavelengths @xmath99 , @xmath100 , @xmath101 and @xmath102 accordingly .
figure [ figqec ] presents a numerical simulation of the evolution of qec along the vapor cell .
the distance is scaled to the resonance optical density @xmath103 with the driving fields being turned off .
the plot 1 corresponds to the case where the conditions for doppler compensation at the transition 02 are fulfilled on the entrance of the cell
. the doppler - free resonance would be narrower if one of the coupled fields were weak .
however , that would give rise to larger absorption and to decreased nonlinear fwm polarization .
the optimization accounting for the change of the driving fields along the medium also shows that it does not substantially change the optimum input intensity values for the plots presented in fig .
[ figqec ] .
plot 2 displays the same dependence , but for the optimum cpt conditions .
alternatively , they are not optimum for the elimination of doppler broadening .
0.5 plot 3 displays a similar dependence in the intermediate case . all intensities and detunings
are identical to those used for computing the previous figures . for the transitions under consideration , the rabi frequency on the level of 100 mhz corresponds to powers of about 100 mw focused on a spot on the order of @xmath104 @xmath105 , which can be realized with common cw lasers with the confocal parameter of focusing of about 2.5 cm . at a temperature of about 700 k ,
the optimum optical density of the dimers is achievable for vapor lengths of about 2 cm .
this is in accord with the parameters of typical cw fwm experiments ( see , e.g. @xcite ) .
the above indicated intensity corresponds to about @xmath106 photons per cubic centimeter .
accounting for the molecule number density at the indicated temperature , which is on the order on @xmath107 @xmath108 ,
coherent control of populations @xcite and of four - wave mixing @xcite with pulses shorter than the dephasing time @xmath109 has proven to be a powerful tool for manipulating nonlinear - optical and chemical properties of free atoms and molecules . in these cases ,
maximum coherence can be achieved as a result of rabi oscillations of the two - photon transition , and the required driving intensity is much higher than that proposed in this paper .
consequently , doppler broadening of the coupled transitions does not play an important role . on the contrary , this paper considers coherent quantum control of resonant four - wave mixing processes , where doppler - broadening of a double-@xmath80 medium is the factor of crucial importance .
an approach enabling one to enhance the efficiency of nonlinear optical conversion and , at the same time , to decrease the required intensity for the fundamental pump beams is proposed .
this approach is based on the elimination of doppler broadening of the resonant two - photon transition .
the advantages of the proposed method as compared to those based on coherent population trapping , where inhomogeneous broadening may play important role too , are illustrated with numerical simulations for realistic experimental schemes .
the results obtained contribute to the field of nonlinear optics at the level of several photons per atom , which is currently attracting growing interest .
tfg and akp thank u.s .
national research council - national academy of sciences for support of this research through the international collaboration in basic science and engineering ( cobase ) program .
akp and asb acknowledge funding support from the international association of the european community for the promotion of co - operation with scientists from the new independent states of the former soviet union ( intas ) ( grant intas-99 - 19 ) and russian foundations for basic research ( grant 99 - 02 - 39003 ) and from the center on fundamental natural sciences at st .
petersburg university ( grant 97 - 5.2 - 61 ) .
999 s.g . rautian and a.m. shalagin ,
_ kinetic problems of nonlinear spectroscopy _ amsterdam : elsevier 1991 + a.k .
popov , _ introduction in nonlinear spectroscopy _ novosibirsk : nauka 1983 ( in russian ) for review see , e.g. _ papers from atomic coherence and interference _ , crested butte workshop 1993 : journal of the european optical society b * 6 * ( 1994 ) no 4 + a review of early russian works can be found in : a.k . popov and s.g .
rautian : proc .
spie * 2798 * ( 1996 ) 49 ( review ) _ coherent phenomena and amplification without inversion _ ( a.v .
andreev , o. kocharovskaya and p. mandel , editors ) , http://www.spie.org/web/abstracts/2700/2798.html , http://xxx.lanl.gov/abs/quant-ph/0005114 + a.k .
popov : bull .
sci . , physics * 60 * ( 1996 ) 927 ( allerton press .
inc ) [ transl . from : izvestiya ran , ser .
* 60 * ( 1996 ) 92 ] ( review ) , http://xxx.lanl.gov/abs/quant-ph/0005108 + a.k .
popov : proc .
spie * 3485 * ( 1998 ) 252 ( s.g .
rautian , i.m .
beterov , n.m .
rubtsova , editors ) , http://www.spie.org/web/abstracts/3400/3485.html , http://xxx.lanl.gov/abs/quant-ph/0005118 maanesh jain , hui xia , g.y .
yin , a.j .
merriam , and s.e .
harris : phys .
* 77 * ( 1996 ) 4326 , http://ojps.aip.org/prlo/ + y. li and m. xiao : opt .
* 21 * ( 1996 ) 1064 + b. lu , w.h .
burkett , and m. xiao : opt .
* 23 * ( 1998 ) 804 , http://ojps.aip.org/olo/ + a.v .
sokolov , g.y .
yin , and s.e .
harris : proc .
spie * 3485 * ( 1998 ) 26 ( s.g .
rautian , i.m .
beterov , n.m .
rubtsova , editors ) + v.g .
arkhipkin , s.a .
myslivets , d.v .
manushkin , and a.k .
popov : _ ibid _ , 525 , http://www.spie.org/web/abstracts/3400/3485.html + v.g .
arkhipkin , s.a .
myslivets , d.v .
manushkin , and a.k .
popov : quantum electronics * 28 * ( 1998 ) 637 , http://turpion.ioc.ac.ru/tmp/rq2804024592_16.html[http://turpion.ioc.ac.ru/ ] + a.j .
merriam , s.j .
sharpe , h. xia , d. manuszak , g.y .
yin , and s.e .
harris : opt .
* 24 * ( 1999 ) 625 , http://ojps.aip.org/olo/ b.d .
agapev , m.b .
gorniy , b.g .
matisov , yu.v .
rojdestvenskiy : usp .
nauk * 163 * ( 1993 ) 1 + e. arimondo : _ progress in optics _ * xxxv * ( 1996 ) 257 t.ya .
popova , a.k .
popov , s.g .
rautian , a.a .
feoktistov : sov .
jetp , * 30 * , 243 ( 1970 ) [ translated from zh .
57 * , 444 ( 1969 ) ] , http://xxx.lanl.gov/abs/quant-ph/0005081 + c. cohen - tannoudji : metrologia * 13 * ( 1977 ) 161 c. cohen - tannoudji , f. hoffbeck , s. reynaud : opt
* 27 * ( 1978 ) 71 + a.k .
popov and l.n .
talashkevich : optics comm . *
28*(1979 ) 315 + s. reynaud , m. himbert , j. dupon - roc , h.h . stroke and c. cohen - tannoudji : phys .
rev . lett .
* 42 * ( 1979 ) 756 + a.k .
popov and v.m .
shalaev : optics comm .
* 35 * ( 1980 ) 189 + a.k .
popov and v.m .
shalaev : opt . spectrosc . * 49*(1981 ) 336 [ transl . from opt.spektrosk . * 49*(1980 ) 617 ] + a.k .
popov and v.m .
shalaev : sov . j. quant
* 12 * ( 1982 ) 289 [ transl . from kvant
* 9 * ( 1982 ) 488 ] + s. reynaud , m. himbert , j. dalibard , j. dupont - roc and c. cohen - tannoudji : opt
* 42 * ( 1982 ) 39 a.k .
popov and a.s .
bayev : jetp letters * 69 * ( 1999 ) 110 , http://ojps.aip.org/journal_cgi/dbt?key=jtpla2&volume=69&issue=2[http://ojps.aip.org/jetplo/ ] + a.k . popov and a.s .
bayev : phys .
a * 62 * , 025801 ( 2000 ) , http://ojps.aip.org/journal_cgi/dbt?key=plraan&volume=62&issue=2[http://ojps.aip.org/prao/ ] g. vemuri , g.s .
agarwal , b. rao : phys .
rev . a * 53 * ( 1996 ) 2842 + yifu zhu and t.n .
wasserlauf : phys .
a * 54 * ( 1996 ) 3653 + yifu zhu , t.n .
wasserlauf and p. sanchez : phys .
a * 54 * ( 1996 ) 4500 + a.k .
popov and v.m .
shalaev : phys .
a * 59 * ( 1999 ) 946 , http://ojps.aip.org/journal_cgi/dbt?key=plraan&volume=59&issue=2[http://ojps.aip.org/prao/ ] s.e .
harris and l.v .
hau : phys .
rev . lett .
* 82 * ( 1999 ) 4611 , http://ojps.aip.org/prlo/ s. babin , u. hinze , e. tiemann and b. wellegehausen : opt
* 21 * ( 1996 ) 1186 , http://ojps.aip.org/olo/ a.k .
popov , s.a .
myslivets , e. tiemann , b. wellegehausen and g. tartakovsky : jetp letters * 69 * ( 1999 ) 912 , http://ojps.aip.org/journal_cgi/dbt?key=jtpla2&volume=69&issue=12[http://ojps.aip.org/jetplo/ ] .
t.rickes , l.p .
yatsenko , s.steuerwald , t.halfmann , b.w.shore , n.v.vitanov , and k.bergmann : j.chem.phys . 113
( 2000 ) 534 http://ojps.aip.org/journal_cgi/dbt?key=jcpsa6&volume=113&issue=2[http://www.physik.uni-kl.de/w_bergma/publications/1997-2000/ ] o. kittelmann , j. ringling , a. nazarkin , g. korn , and i. v. hertel : phys .
* 76 * ( 1996 ) 2682 , http://ojps.aip.org/prlo/ ; + a. nazarkin , g. korn , o. kittelmann , j. ringling , and i. v. hertel : phys .
rev . a * 56 * ( 1997 ) 671 http://ojps.aip.org/journal_cgi/dbt?key=plraan&volume=56&issue=1[http://ojps.aip.org/prao/ | new feasibity of coherent quantum control of four - wave mixing processes in a resonant doppler - broadened medium are studied .
we propose a technique which enables one to enhance the quantum efficiency of nonlinear optical conversion . at the same time
, it allows one to decrease the required intensities of the fundamental beams compared to those necessary in the approach based on coherent population trapping .
the major outcomes of the analysis are illustrated with numerical simulation addressed within a practical medium . |
a ` flat band ' is a dispersionless energy band composed of entirely degenerate states . as a result of this degneracy ,
any superposition of these states ( even a highly localized one ) is completely static , displaying no evolution dynamics whatsoever .
flat bands are manifested in a variety of condensed matter contexts , including ferromagnetic ground states of hubbard models @xcite , frustrated hopping models @xcite , in the context of localization in the presence of magnetic fields or spin - orbit coupling @xcite , and the fractional quantum hall effect @xcite . with the advent of the recent experimental observations of photonic topological insulators @xcite , topological flat band models are now ripe for study in the optical domain . indeed in optics , flat bands
have direct technological relevance due to their high degeneracy , and thus high photonic density - of - states , leading to enhanced light - matter interaction @xcite .
additionally , for kagome systems single peak solitons have shown to bifurcate from the flat band at zero power threshold @xcite , and a non - diffractive image transmission scheme have been suggested in the linear regime @xcite , showing the chance for concrete applications emerging from the flat band properties . in the lieb
lattice - a line - centered square lattice - a flat band touches two linearly dispersing intersecting bands , i.e. , the flat band intersects a single dirac point . in the low - energy regime ,
the dispersion is described by a quasi - relativistic equation for spin-1 fermions @xcite .
it is manifested in , e.g. , the high - temperature tetragonal phases of the so - called cuprates : layered , ceramic materials that exhibit the phenomenon of high - temperature superconductivity @xcite . in this particular system several studies were conducted only recently , experimentally probing its response to external gauge fields @xcite to nematic ferromagnetism @xcite and nonlinear conical diffraction @xcite .
the flat - band zero - energy edge modes in this lattice exhibit various peculiar features @xcite and , hence , deserve particular attention . here ,
we analyze transport in lieb lattices theoretically and experimentally , in the bulk and at the edges .
we demonstrate that whether or not flat - band modes are excited strongly impacts the transport in the bulk of the lattice . additionally , we demonstrate the presence of tamm - like edge states that arise at a van hove singularity in the band structure @xcite .
finally , we show that our photonic realization holds promise for a thorough analysis of various features of the lieb lattice in terms of transport at and far off the lattice edges .
the structure of the lieb lattice is shown in fig .
[ fig:1](a ) , where it is evident that the unit cell consists of three sites @xmath0 , @xmath1 , and @xmath2 . as in honeycomb lattices ( also known as photonic graphene " ) @xcite , the equation describing the propagation of light in this system is @xmath3 where @xmath4 is the longitudinal propagation distance into the photonic lattice ( playing the role of time in the analogous schrdinger equation ) ; @xmath5 is the envelope of the electric field , defined by @xmath6 ( @xmath7 is the electric field , @xmath8 the wavenumber in free space , @xmath9 and @xmath10 is the speed of light in vacuum ) ; @xmath11 is the refractive index structure defining the photonic lattice , while @xmath12 is the refractive index of the material in which the lattice is embedded ; @xmath13 is the transverse laplacian operator ; @xmath14 as defined in equation ( 1 ) corresponds to the continuum hamiltonian for wave propagation in a photonic lattice . when the refractive index structure is composed of highly confined waveguides , each with a single bound state , we employ a tight - binding approximation .
since within a given unit cell hopping takes place only between @xmath0-@xmath1 and @xmath1-@xmath2 nearest neighbors , the corresponding tight - binding hamiltonian reads @xmath15 with @xmath16 , @xmath17 , @xmath18 as the creation operators in the @xmath19-th unit cell on the @xmath0 , @xmath1 and @xmath2 sites , respectively .
the summation takes place over @xmath20 ( the position of the unit cell ) , and @xmath21 ( the vectors connecting the neighbors via the inter - site coupling @xmath22 in the @xmath23- and @xmath24-direction , respectively ) . in
fourier space , this becomes @xmath25 with @xmath26 ( first brillouin zone ) . from this ,
one readily obtains the dispersion relation , i.e. , the spectrum @xmath27 and @xmath1 is described by @xmath28 ( horizontal direction ) , and between @xmath1 and @xmath2 by @xmath29 ( vertical direction ) .
( b ) the dispersion relation of the isotropic lieb lattice ( @xmath30 ) in the tight - binding approximation ( eq.[tightbindingband ] ) shows three bands : a flat band at @xmath31 and two conically intersecting ones , with all three bands intersecting at the m point ( @xmath32 ) .
( c ) microscope image of a laser - written lieb lattice in a fused silica glass wafer.[fig:1 ] ] the full spectrum ( shown in fig .
[ fig:1](b ) for @xmath33 ) consists of three bands : a flat band ( @xmath34 ) and two conically intersecting bands , with all three bands intersecting at the m point ( @xmath32 ) .
note that for each eigenstate at bloch wavevector @xmath35 in the top band with eigenvalue @xmath36 , there exists an eigenstate also at @xmath35 with eigenvalue @xmath37 ; this is known as particle - hole symmetry @xcite .
the eigenfunctions of the finite lieb lattice with periodic and vanishing boundary conditions are given in ref .
importantly , any superposition of eigenstates belonging to the flat band is also an eigenstate of the system , meaning that it is static and does not contribute to transport .
equivalently , the group velocity in the flat band is zero at all bloch wave vectors , therefore any wavepacket is necessarily diffraction - free .
consequently , to have transport , states in the other bands have to be excited .
we also note that the flat band is not destroyed by any anisotropy @xmath38 , being an intrinsic property of this lattice in the nearest - neighbor coupling limit @xcite . in general ,
linear modes of any periodic structure are completely extended . however
, flat band systems allow the formation of very localized eigenmodes as a destructive linear combination of extended linear wavefunctions @xcite . in the case of the lieb lattice , in any square of the lattice ( formed by four @xmath1 sites at the corners ,
two @xmath2 sites at the upper and lower sides and two @xmath0 sites at the right and left sides of the square ) a stationary localized mode can be created by taking the four @xmath1 sites strictly zero , the two @xmath2 sites with an amplitude @xmath39 , while the last two @xmath0 with an amplitude @xmath40 and the rest of lattice sites are zero . for our experiments ,
we fabricate lieb lattices employing the laser direct - write technology @xcite . to this end , ultrashort laser pulses ( coherent inc .
mira / rega , @xmath41 nm , @xmath42fs at 100khz repetition rate ) are tightly focused through a microscope objective ( 25@xmath43 , na = 0.35 ) into a 10 cm long transparent fused silica wafer and translated by means of a high - precision positioning system ( aerotech inc . ) .
a microscope front view of such a lattice , which consists of @xmath44 waveguides with nearest - neighbor spacing of 20@xmath45 m , is shown in fig . [ fig:1](c ) .
note that due to the slightly elliptical form of our waveguides , we find that the vertical inter - site coupling is slightly stronger than the horizontal one , i.e. the lattice is effectively uniaxially strained and anisotropic with @xmath46 . in our experiments , laser light at @xmath47 nm was launched into a single waveguide at the input facet of the array using a standard microscope objective . at the output facet of the sample ,
intensity patterns were recorded with a ccd camera .
-site , ( b ) the @xmath1-site , and ( c ) the @xmath2-site .
the overlap with the eigenmodes in the three bands is shown in ( d ) for the @xmath0-site , ( e ) for the @xmath1-site , and ( f ) for the @xmath2-site .
it is evident that transport is stronger when the flat band is least populated at the initial plane of excitation .
[ fig:2 ] ] we begin by launching light into the bulk of the lattice , i.e. , far away from the edges .
as there are three elements per unit cell ( see fig .
[ fig:1]a ) , we study the propagation arising from excitations in each of the three positions @xmath0 , @xmath1 , and @xmath2 .
the corresponding diffraction patterns obtained in our experiments are shown in figs .
[ fig:2](a),(c),(e ) .
one can clearly see that the impact of the excitation location on the diffraction pattern is significant , depending on where in the unit cell the lattice is excited .
the reason lies in the unique structure of the lieb lattice , where a single site excitation ( that corresponds to a unit impulse in real space ) does not excite all bands - in contrast to other lattices , like square , hexagonal or honeycomb lattices . to illustrate this effect ,
we plot the projections @xmath48 of the impulse excitation @xmath49 at the @xmath0 , @xmath1 , or @xmath2 sites on the eigenfunctions @xmath50 of the three bands ( @xmath51 is the band index ) in figs .
[ fig:2](b),(d),(f ) , and compare them to the experimental data .
it is evident that , when exciting the @xmath1 site [ fig .
[ fig:2](e ) ] , the projections on the flat band are minimal ( equal to @xmath52 , where @xmath53 is the number of complete unit cells on one side of a finite square lieb lattice , i.e. , @xmath54 in our case ) and , hence , transport is strong . this
is clearly observed in the experiment [ see [ fig:2](b ) ] , where the light spreads across many lattice sites following the 10 cm propagation in the lattice . however , when exciting the @xmath0 site the situation changes drastically [ fig .
[ fig:2](d ) ] . in this case
, the projections on the flat band dominate and , therefore , the transport is minimal .
the experiment shows exactly this : the spreading is very weak , and the light covers only a few lattice sites after propagating through the lattice [ see fig .
[ fig:2](a ) ] .
intermediate spreading is observed when launching light into lattice site @xmath2 , as shown in fig .
[ fig:2](c ) , as a consequence of the partial overlap with the flat band [ see fig .
[ fig:2](f ) ] . at this point
it is interesting to examine the transport through the lieb lattice when strain is applied .
specifically , it has been shown that inhomogeneous strain applied to honeycomb lattices can give rise to landau level splitting @xcite , robust klein tunneling @xcite , topological transitions @xcite , and other unexpected phenomena .
it is therefore instructive to look at the projections on the flat band as a function of the strain when exciting either the @xmath0 , @xmath1 , or @xmath2 sites in the infinite lieb lattice .
our calculations are summarized in fig .
[ fig:3 ] .
interestingly , when exciting a @xmath1 site , the population of the flat band always vanishes , for all values of the anisotropy .
localized flat band modes have @xmath1 sites with zero amplitude , therefore launching light into this site will always result in strong transport , irrespective of the ratio @xmath55 . however , the situation is different for an excitation of the @xmath0 and @xmath2 sites .
whereas for the isotropic case ( @xmath56 ) , the overlap with the flat band for both sites is exactly the same ( due to symmetry reasons ) , for the anisotropic case ( @xmath57 ) the flat band is more populated for an @xmath0-site excitation ( resulting in decreasing transport ) and less populated for a @xmath2-site excitation ( yielding enhanced transport ) .
( anisotropic @xmath31 modes have @xmath58 , being therefore easier to excite them in sites @xmath0 ) .
-site ( red dotted line ) , @xmath1-site ( green solid line ) , and @xmath2-site ( blue dashed line ) .
[ fig:3],width=453 ]
notwithstanding the bulk transport , we pay particular attention to the diffraction of light at the edge of the lattice .
it is important to note that , within tight - binding model ( eq . [ tightbindingmodel ] ) there are no edge states present , in the isotropic as well as in the anisotropic case . however , we surprisingly find that edge states arise in anisotropic lieb lattices in the continuous model ( eq . [ continuousmodel ] ) . in fig .
[ fig:4](a ) we present a plot of the band structure at the edge of the lieb lattice by diagonalizing @xmath59 as given by eq . [ continuousmodel ] , showing clearly the existence of states residing in the band gap that are localized at the lattice edge .
the underlying mechanism for the formation of these edge states is similar to the one recently reported for `` photonic graphene '' @xcite .
note that these calculations are performed by full diagonalizing @xmath59 , as given in eq .
[ continuousmodel ] , as opposed to simply examining the spectrum of the tight - binding model . to explain these edge states in the lieb lattice , one has to take into account the fact that the effective on - site propagation constant of each mode is modified by its neighbouring sites .
this modification differentiates between the @xmath1-sites in the bulk , where they have four neighbours , and the edge , where they have only three .
although this difference is small and is usually neglected in tight - binding theories , it comes into play where the modes are highly degenerate at the van hove singularity . at the degeneracy point , any slight edge perturbation takes the edge mode out of the band and creates a tamm - like edge state as shown in fig .
[ fig:4](b ) .
we emphasize again that this state does not result from any real surface perturbation but rather from the specific surface structure along the edge .
we classify this state as a `` tamm state '' @xcite as opposed to a `` shockley state '' @xcite , because it does not arise due to a band crossing , the criterion for the emergence of shockley states .
it is important to note that these states do not have a topological origin , and only arise from the ` effective defect ' at the edge .
although tamm states are conventionally associated with surface perturbations or defects ( that are inherent in the system , or that arise from modulation @xcite ) , in the present case no defects whatsoever are present . instead , what happens here is that the edge itself acts as a sufficiently strong defect to localize light on the edge .
this effect can not be accounted for in the standard tight - binding model and appears only in the continuum model .
the experimental observation of strong edge diffraction in the anisotropic lieb lattice is presented in fig .
[ fig:4](c ) , the simulations using eq .
[ continuousmodel ] are shown in fig .
[ fig:4](d ) . ) using the continuous model eq .
[ continuousmodel ] , showing the appearance of edge states ( red line ) .
( b ) the intensity distribution of the edge state in the lieb lattice .
( c ) experimental observation of the edge states after an impulse excitation .
( d ) simulation of the light diffraction at the edge .
[ fig:4 ] ]
in conclusion , we presented a theoretical and experimental study of transport of light in photonic lieb lattices .
we find that the transport in the bulk is significantly affected by which site in the unit cell is excited , due to different overlap with the flat band ( since any superposition of flat band states does not diffract and can form localized bulk states ) .
importantly , an excitation of the @xmath1-site in the unit cell leaves the flat band essentially empty , resulting in maximal transport .
we also show that the population of the different bands is a function of the anisotropy of the lattice .
moreover , we demonstrate the existence of new edge states in anisotropic lieb lattices that arise due to a virtual defect at the lattice edges and are not manifested in the standard tight - binding model .
our findings experimentally prove that the lieb lattice exhibits various peculiar and unique features concerning wave transport , which might find their way into new concepts for imaging , routing and switching .
the authors wish to thank the german ministry of education and research ( center for innovation competence programme , grant 03z1hn31 ) , and the thuringian ministry for education , science and culture ( research group spacetime , grant 11027 - 514 ) , fondecyt grant 1110142 , programa icm p10 - 030-f , and programa de financiamiento basal de conicyt ( fb0824/2008 ) . | we analyze the transport of light in the bulk and at the edge of photonic lieb lattices , whose unique feature is the existence of a flat band representing stationary states in the middle of the band structure that can form localized bulk states .
we find that transport in bulk lieb lattices is significantly affected by the particular excitation site within the unit cell , due to overlap with the flat band states .
additionally , we demonstrate the existence of new edge states in anisotropic lieb lattices .
these states arise due to a virtual defect at the lattice edges and are not described by the standard tight - binding model . |
studies of massive galaxy populations show that the most massive galaxies are in place and in possession of the majority of their stellar mass by @xmath7 @xcite . thus , while dark matter halos in a @xmath8cdm cosmology build up hierarchically with the gradual accretion and merging of smaller halos , the most massive galaxies within that context likely have more dramatic origins .
the details of this process are uncertain , and ideally we would like to study sites of ongoing massive galaxy formation .
nebulae or ly@xmath1 `` blobs '' ( labs ) large ( @xmath050 kpc ) clouds of gas emitting strongly in ly@xmath1 ( @xmath9 erg s@xmath10 ) provide such an opportunity .
the large ly@xmath1 equivalent widths and the association with galaxy populations such as lyman break galaxies ( lbgs ) and submillimeter galaxies ( smgs ) strongly suggest that labs are sites of ongoing galaxy formation .
they have been found in small numbers ( only 16 @xmath050 kpc labs are known ) around @xmath11 2 - 3 , a key epoch of black hole and galaxy growth ( @xcite , hereafter s00 ; @xcite ) .
labs span a range in size and surface brightness from the scales of ly@xmath1-emitting galaxies ( laes ) to the largest labs known ( @xmath5150 kpc , @xmath527 mag arcsec@xmath12 ; e.g. , * ? ? ?
labs are similar to the large ly@xmath1 halos observed in the overdensities around higher redshift radio galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
* ) but are found in radio quiet environments .
the dominant power source in labs is difficult to determine ; among the limited sample of large labs , there is evidence for embedded agn , starburst - driven superwinds , gravitational cooling radiation , and spatially - extended star formation , all of which may play a role in powering the ly@xmath1 emission ( prescott et al .
2008 , in preparation ; s00 ; @xcite ) .
most known labs , including two of the largest cases , have been discovered via narrow - band surveys , often by targeting known galaxy overdensities ( s00 ; @xcite ) .
follow - up narrow - band imaging of the ssa22 region surrounding the s00 labs revealed fainter labs associated with the same overdensity traced by the laes @xcite .
this suggests that labs may be confined to overdense regions , as would be expected for sites of massive galaxy formation . a blank field survey by @xcite supported this claim , finding no labs with sizes greater than @xmath530 kpc and an order of magnitude lower number density of labs relative to that found within the ssa22 galaxy overdensity .
however , the association between labs and galaxy overdensities may be misleading , as a truely systematic wide - area search has yet to be completed .
a thorough environmental study has only been done for the s00 labs , which were found by targeting a known galaxy overdensity .
in contrast , one of the largest labs uncovered recently was found by entirely different means . while conducting a study of mid - infrared ( 24@xmath13 m ) sources detected by the
spitzer space telescope _ , @xcite discovered a lab at @xmath14 within the noao deep wide - field survey botes field ( ndwfs ; * ? ? ?
follow - up observations revealed complexity typical of the lab class : the region hosts a buried agn , many young galaxies , an lbg , and diffuse heii and continuum emission suggestive of spatially - extended star formation ( prescott et al .
2008 , in preparation ) .
one of the largest and most luminous labs known , and found without any a priori knowledge of its surroundings , this source represents a unique opportunity to perform an unbiased , complementary test of the association between labs and overdense regions . in this work ,
we present the first results from an ongoing deep intermediate - band ly@xmath1 imaging survey of the environment of the @xcite lab ( hereafter , labd05 ) and report on the spatial distribution of the laes in the immediate vicinity .
a detailed analysis of the multiwavelength properties of labd05 and the properties and clustering of the laes in the region will be presented by prescott et al .
( 2008 , in preparation ) .
we assume the standard @xmath8cdm cosmology ( @xmath15=0.3 , @xmath16=0.7 , @xmath17=0.7 ) ; the angular scale at @xmath14 is 7.96 kpc/. all magnitudes are in the ab system .
we obtained deep imaging of the field around labd05 using the subaru telescope and the suprimecam imager @xcite on u.t .
2007 may 10 - 14 and june 17 .
the survey covers 0.22 deg@xmath18 in an intermediate - band filter , ia445 ( @xmath194458 , @xmath20201 ) , centered on the ly@xmath1 line at the redshift of the nebula ; this corresponds to a comoving volume of 4.27@xmath310@xmath21 h@xmath22 mpc@xmath23 ( 52@xmath345@xmath3180 h@xmath4 mpc ) .
conditions during the may observations were variable ( clouds , variable seeing 0.7 - 1.2 ) and good in june ( clear with 0.7 seeing ) .
we obtained a total of 3 hours of observations .
we reduced the data using the sdfred software @xcite .
the data were overscan - subtracted and corrected for geometric and atmospheric distortions .
we generated the sky flat using object frames that were free of bright stars in combination with other images taken in the same intermediate - band filter ( y. taniguchi , private communication ) .
small portions of the suprimecam field of view are vignetted by the autoguider probe ; the affected areas were masked , as were bad columns , bright star ghosts , and satellite trails .
images were aligned and scaled using common stars and then combined using a clipped mean algorithm , which successfully removed cosmic rays . of the suprimecam field - of - view ( 0.26 deg@xmath18 ) , 73% was usable ( the remainder being of lower signal - to - noise along the edge of the field or in the vicinity of bright stars ) .
the limiting magnitude of the stacked image is 28.3 ab mag ( 1@xmath24 , 2 diameter aperture ) , calculated using 10,000 random apertures .
an approximate magnitude zeropoint was calculated from observations of the standard stars bd+25d4655 and feige 34 . for the ndwfs broad - band imaging , the limiting magnitudes are @xmath25@xmath26=27.9 mag , @xmath27=27.1 mag , @xmath28=26.0 mag ( 1@xmath24 , 2 diameter aperture ) .
a portion of the field was observed in a custom @xmath29-band filter ( @xmath30 ; @xmath313590 , @xmath20116 ) using the mayall 4 m telescope over 6 nights ( u.t .
2007 june 8 - 13 ) .
these data will be described elsewhere ( prescott et al .
2008 , in preparation ) but here provide a useful check on interlopers in our @xmath32 lae sample .
we used source extractor @xcite to select a sample of @xmath3338,600 sources detected in the @xmath34 band down to the 5@xmath24 limit of @xmath34=26.5 mag ( 2 diameter aperture ; @xmath35(@xmath14)=1.5@xmath310@xmath36 erg s@xmath10 ) with the following search parameters : at least 5 contiguous pixels , a threshold of 2@xmath24 per pixel , and a gaussian filter matched to the seeing ( [email protected] arcsec ) .
we measured matched aperture photometry using 2 diameter apertures from the @xmath34 , @xmath25 , @xmath37 , and @xmath38 imaging datasets , which were registered and resampled to match the @xmath25 astrometry and pixel scale ( 0.258 arcsec / pix ) .
aperture corrections were neglected ( @xmath331.08 for an unresolved source ) .
line - emitting sources are strongly detected in @xmath34 relative to @xmath25 , i.e. , they have large negative @xmath34@xmath39@xmath25 colors relative to the normal galaxy locus ( see figure [ fig : ibcolor ] ) .
we removed bright stars ( @xmath34@xmath40 ) using the class_star parameter in source extractor ( @xmath410.91 ) and employed a cut of @xmath34@xmath39@[email protected] mag yielding 1500 candidates . shifting the @xmath34@xmath39@xmath25 cut by @xmath430.2 mag
causes no significant change to the main results presented in section 4 . for a @xmath44 continuum source
, this corresponds to an observed equivalent width ( ew ) cut of @xmath45148 .
laes are known to be young , with estimated ages of 4 - 200 myr @xcite . for the case of a young galaxy ( 25 myr old simple stellar population , solar metallicity , chabrier imf ; * ? ? ?
* ; * ? ? ?
* ) at @xmath46 with standard intergalactic absorption @xcite , this is equivalent to a rest - frame ew cut of @xmath4750 .
we expect our lae candidate sample to be contaminated by high ew , low - redshift [ oii]@xmath483727,3729-emitting galaxies at @xmath49 . since the @xmath34 bandpass lies on the red side of the @xmath25 filter ( figure [ fig : ibcolor ] , see inset ) , we also expect contamination from higher redshift galaxies ( @xmath50 ) for which the lyman limit has entered the @xmath25 filter , thus depressing the @xmath25 flux relative to the @xmath34 flux . using the publicly available ndwfs imaging @xcite , we employ a cut of @[email protected] to remove both contaminant populations ( see figure [ fig : colorcolor ] )
. the final sample of 785 lae candidates corresponds to a mean lae surface density of @xmath334200 deg@xmath12 .
the properties and sizes of the lae sample will be discussed in an upcoming paper ( prescott et al .
2008 , in preparation ) .
there are no other large ( @xmath050 kpc ) labs in the vicinity of labd05 .
sources at these redshifts should show very little flux in the @xmath30-band , which straddles the lyman limit ( @xmath52970 - 1000 at @xmath53 ) . using the 25 myr old model above to represent the typical lae continuum shape , the @xmath54 galaxy luminosity function from @xcite , and standard intergalactic absorption @xcite
, we predict that @xmath51 - 2% of the lae sample should have @xmath30 detected at the 5@xmath24 level ( @xmath30=25.3 ) .
only 1 lae candidate ( @xmath551% ) is detected with @xmath5625.3 , giving us confidence that we have selected a clean sample of laes .
more sophisticated interloper rejection would require deeper @xmath30 imaging over the entire field .
figure [ fig : spatial ] shows the spatial distribution of laes in the vicinity of labd05 .
we find an overabundance of laes in the immediate vicinity of the lab and clear evidence of an elongated overdense structure , with what appears to be a @xmath510 h@xmath4 mpc underdensity at [ 218.36,33.07 ] . in figure
[ fig : density ] , we plot the surface density of laes versus distance from labd05 . this shows a peak overdensity of a factor of 3.0 relative to the edge of the field , or a factor of 2.6 when averaged over the central 10 h@xmath4 mpc radius . for comparison ,
the dashed line shows the expected lae surface density estimated from the @xmath57 luminosity functions of @xcite and @xcite assuming a uniform redshift distribution across our bandpass and no evolution .
the overdensity spans at least @xmath3317@xmath347 h@xmath4 mpc ( comoving ) .
a study of the lae population around the two s00 labs , which are of comparable size and luminosity to labd05 , also found a factor of @xmath53 overdensity relative to the field that was @xmath560 mpc across @xcite .
thus , although it was found without any a priori knowledge of its surroundings , labd05 appears to reside in a similarly overdense environment . without follow - up spectroscopy
, we can not know the true redshift distribution of the sample within the intermediate - band filter . assuming a uniform distribution across the filter ( @xmath58 0.17 , 180 h@xmath4 mpc , comoving , along the line - of - sight )
, our survey yields a mean number density of @xmath592.1@xmath310@xmath60 h@xmath22 mpc@xmath60 for laes with @xmath611.5@xmath310@xmath36 erg s@xmath10 cm@xmath12 .
the inner contour in figure [ fig : spatial ] corresponds to @xmath592.8@xmath310@xmath60 h@xmath22 mpc@xmath60 while at the edge of the field @xmath591.2@xmath310@xmath60
h@xmath22 mpc@xmath60 .
given the presence of an overdensity , we expect the true redshift distribution to be significantly narrower . if we assume the overdensity resembles that hosting the s00 labs , which extends 40 h@xmath4 mpc comoving ( 22% of the filter width ) along the line - of - sight @xcite , and that it is centered within the @xmath34 filter , the inner contour corresponds to a number density of @xmath5912@xmath310@xmath60 h@xmath22 mpc@xmath60 . alternately , if the structure is cylindrical in shape with a radius of 10 h@xmath4 mpc , the corresponding number density at the inner contour is @xmath5925@xmath310@xmath60 h@xmath22 mpc@xmath60 .
we estimate the expected frequency of such overdense structures using the galaxy catalog of @xcite , which is based on the millennium simulation @xcite .
we approximate our observational set - up by sampling the simulation volume randomly with cylinders that are 10 h@xmath4 mpc in radius ( the size of the overdensity peak ) and 180 h@xmath4 mpc in depth ( the full span of the filter ) .
we assumed a ly@xmath1/h@xmath1 ratio of 10 @xcite and scaled from the predicted h@xmath1 luminosity to select model lae galaxies down to the ly@xmath1 limit of our observations ( under the assumption that ly@xmath1 dominates the measured @xmath34 flux ) at a redshift of @xmath32 .
the number density of model laes in this case is 10.8@xmath310@xmath60 h@xmath22 mpc@xmath60 , which is a factor of @xmath55 higher than the mean density we observe .
taken at face value , this could imply that laes have a duty cycle of @xmath520% or that they are a younger , less massive subset of this population . if we restrict the model galaxies to be younger than the sample s median age ( @xmath63172 myr ) and less massive than the median stellar mass ( @xmath639@xmath310@xmath64 @xmath65 ) , the number density of model laes is 3.5@xmath310@xmath60 h@xmath22 mpc@xmath60 . in either case
, we uncover overdensities of greater than a factor of 2 at a rate of @xmath330.3% . therefore , within the large span of our filter , which will tend to average out inhomogeneities and reduce the signal , our imaging survey had only a @xmath330.3% chance of randomly uncovering such an overdense region if labs and overdensities are independent phenomena .
the space density of large ly@xmath1 nebulae is very uncertain , but at the high end is the range quoted for the s00 ly@xmath1 nebulae ; @xmath53 - 400@xmath310@xmath66 mpc@xmath60 @xcite . taking these values ( equivalent to a @xmath517 - 100% chance of finding one large lab within the peak of the overdensity )
, the likelihood of a chance coincidence between a factor of @xmath412 overdensity and a large ly@xmath1 nebulae would be @xmath60.05 - 0.3% .
preliminary results from more recent systematic ly@xmath1 nebulae surveys hint that their true space density may be orders of magnitude lower ( e.g. , prescott et al .
2008 , in preparation ) , making the likelihood of chance coincidence vanishingly small . in terms of the relevant emission mechanisms the ly@xmath1 nebula class appears to be
a highly heterogeneous mix , and this diversity could in principle derive from environmental differences .
the largest ly@xmath1 nebulae ( @xmath0100 kpc ) , including the case studied here , often show evidence for obscured agn and extended star formation ( e.g. , * ? ? ?
* ; * ? ? ?
* prescott et al .
2008 , in preparation ) and have received the most scrutiny in terms of their properties and environments .
presumably , the somewhat smaller `` cooling '' labs ( e.g. , * ? ? ?
* ; * ? ? ?
* ) must also reside in dense regions with sufficient gas supply , but thorough environmental studies of these sources have , to our knowledge , not yet been completed .
the discovery of a large ly@xmath1 nebula at @xmath32 via its strong mid - infrared emission has provided an unbiased test of the association between these rare sources and galaxy overdensities .
using deep ly@xmath1 imaging of the environment surrounding this lab , we identify 785 lae candidates and find evidence for a factor of @xmath53 lae overdensity which spans 20@xmath350 mpc ( comoving ) .
this is comparable to what is found in the vicinity of the well - known s00 ly@xmath1 nebulae .
we rule out a chance coincidence at the @xmath61% level . in conjunction with previous work ,
these results point conclusively to a strong association between the largest ly@xmath1 nebulae and overdense regions of the universe .
we are grateful to masafumi yagi and yoshiaki taniguchi for data reduction advice and assistance .
we thank naveen reddy for the use of data from an ongoing @xmath30 imaging program and the referee for constructive comments on the manuscript .
m. p. was supported by an nsf graduate research fellowship .
support for this work was provided by nasa ( hst - go10591 ) and the national science foundation ( grant no .
0714311 ) .
basu - zych , a. , & scharf , c. 2004 , , 615 , l85 bertin , e. , & arnouts , s. 1996 , , 117 , 393 bower , r. g. , benson , a. j. , malbon , r. , helly , j. c. , frenk , c. s. , baugh , c. m. , cole , s. , & lacey , c. g. 2006 , , 370 , 645 brown , m. j. i. , dey , a. , jannuzi , b. t. , brand , k. , benson , a. j. , brodwin , m. , croton , d. j. , & eisenhardt , p. r. 2007 , , 654 , 858 bruzual , g. , & charlot , s. 2003 , , 344 , 1000 bundy , k. , ellis , r. s. , & conselice , c. j. 2005 , , 625 , 621 chapman , s. c. , scott , d. , windhorst , r. a. , frayer , d. t. , borys , c. , lewis , g. f. , & ivison , r. j. 2004 , , 606 , 85 daddi , e. , et al .
2005 , , 626 , 680 dey , a. , et al .
2005 , , 629 , 654 finkelstein , s. l. , rhoads , j. e. , malhotra , s. , pirzkal , n. , & wang , j. 2007 , , 660 , 1023 francis , p. j. , et al . 2001 , , 554 , 1001 gawiser , e. , et al .
2007 , , 671 , 278 greve , t. r. , stern , d. , ivison , r. j. , de breuck , c. , kovcs , a. , & bertoldi , f. 2007 , , 382 , 48 gronwall , c. , et al .
2007 , , 667 , 79 hayashino , t. , et al .
2004 , , 128 , 2073 jannuzi , b. t. & dey , a. 1999 , in `` photometric redshifts and the detection of high redshift galaxies '' , asp conference series , vol .
191 , edited by r. weymann , l. storrie - lombardi , m. sawicki , and r. brunner .
isbn : 158381 - 017-x , p. 111
lai , k. , et al .
2007 , arxiv e - prints , 710 , arxiv:0710.3384 leitherer et al .
1996 , pasp , 108 , 996 madau 1995 , apj , 441 , 18 matsuda , y. , et al .
2004 , , 128 , 569 matsuda , y. , et al .
2005 , , 634 , 125 matsuda , y. , iono , d. , ohta , k. , yamada , t. , kawabe , r. , hayashino , t. , peck , a. b. , & petitpas , g. r. 2007 , , 667 , 667 mccarthy , p. j. , et al .
2004 , , 614 , l9 miyazaki , s. , et al .
2002 , , 54 , 833 nilsson , k. k. , fynbo , j. p. u. , mller , p. , sommer - larsen , j. , & ledoux , c. 2006 , , 452 , l23 osterbrock , d. e. 1989 , research supported by the university of california , john simon guggenheim memorial foundation , university of minnesota , et al .
mill valley , ca , university science books , 1989 , 422 p. ouchi , m. , et al .
2004 , , 611 , 660 ouchi , m. , et al .
2007 , arxiv e - prints , 707 , arxiv:0707.3161 overzier , r. a. , et al .
2008 , , 673 , 143 palunas , p. , teplitz , h. i. , francis , p. j. , williger , g. m. , & woodgate , b. e. 2004 , , 602 , 545 reddy , n. a. , steidel , c. c. , pettini , m. , adelberger , k. l. , shapley , a. e. , erb , d. k. , & dickinson , m. 2007 , arxiv e - prints , 706 , arxiv:0706.4091 saito , t. , shimasaku , k. , okamura , s. , ouchi , m. , akiyama , m. , & yoshida , m. 2006 , , 648 , 54 smith , d. j. b. , & jarvis , m. j. 2007 , , 378 , l49 springel , v. , et al .
2005 , , 435 , 629 steidel , c. c. , adelberger , k. l. , shapley , a. e. , pettini , m. , dickinson , m. , & giavalisco , m. 2000 , , 532 , 170 taniguchi , y. , & shioya , y. 2000 , , 532 , l13 tremonti , c. a. , et al .
2004 , , 613 , 898 van dokkum , p. g. , et al .
2004 , , 611 , 703 venemans , b. p. , et al .
2007 , , 461 , 823 yagi , m. , kashikawa , n. , sekiguchi , m. , doi , m. , yasuda , n. , shimasaku , k. , & okamura , s. 2002 , , 123 , 66 | large nebulae ( @xmath050 kpc ) emitting strongly in ly@xmath1 ( also known as ly@xmath1 `` blobs '' ) are likely signposts of ongoing massive galaxy formation .
the relative rarity of these sources and their discovery in well - studied galaxy overdensities suggest that they may be associated with regions of high galaxy density .
one of the largest ly@xmath1 nebulae , discovered at a redshift of @xmath2 2.7 via its strong mid - infrared emission , provides an unbiased test of this association .
we have carried out a deep intermediate - band imaging survey for ly@xmath1-emitting galaxies ( laes ) within a 30@xmath326 field of view around this ly@xmath1 nebula .
this is the first study of the environment of a ly@xmath1 nebula found without a priori knowledge of its surroundings .
we find that the nebula is located in an overdense region , at least 20@xmath350 h@xmath4 comoving mpc in size , showing a factor of @xmath53 lae number density enhancement relative to the edge of the field .
given the predicted number of such overdensities , we rule out the possibility of a chance coincidence at the @xmath61% level .
this study , in conjunction with previous work , provides strong confirmation of the association between the largest ly@xmath1 nebulae and overdense regions of the universe . |
constructions of invariants of 3-manifolds associated to chern - simons theory rely on the specialization of a parameter @xmath1 at a root of unity @xcite . in khovanov
s categorification of the jones polynomial @xcite the variable @xmath1 is represented by an extra grading . understanding how to `` set '' this grading to a root of unity has been a difficult problem in the program to derive invariants of 3-manifolds from categorifications of knot polynomials . in this paper
we propose a means by which such categorifications can be evaluated at a root of unity .
this proposal is motivated in section [ handleslide ] by a construction of objects which are invariant under handle slides . in the temperley - lieb algebra the evaluation of @xmath1 at a root of unity
is closely related to taking a suitable quotient by the jones - wenzl projector @xmath2 . in @xcite
the authors constructed chain complexes @xmath3 within the universal categorification of the jones polynomial ( see @xcite ) that become jones - wenzl projectors in the image of the grothendieck group .
these chain complexes are uniquely characterized up to homotopy by the defining properties of the jones - wenzl projectors @xmath2 . in this paper
we consider a quotient of the homotopy category of chain complexes by the projector @xmath3 .
the verdier quotient classically considered in homological algebra turns out to be unsuitable in this context ( see section [ verdier ] ) .
we propose a different notion in section [ cosetcon ] called the coset construction . in this context
we define objects which categorify reshetikhin - turaev invariants at low levels .
we will now outline a few more details of the construction .
the temperley - lieb algebra has coefficients in @xmath4 $ ] .
the evaluation of a polynomial @xmath5 at a root of unity may be implemented algebraically by passing to the quotient @xmath6 where @xmath7 \to \mathbb{z}[q]/(\varphi_p(q))$ ] is the quotient map and @xmath8 is the @xmath9th cyclotomic polynomial .
for example , when @xmath9 is a prime number @xmath10 given an element @xmath11 such that @xmath12 , the ideal generated by @xmath8 contains the ideal generated by @xmath13 . therefore the quotient obtained by setting @xmath1 to be a @xmath9th root of unity factors through the quotient of @xmath14 by @xmath15 : @xmath16 when @xmath17 the jones - wenzl projectors @xmath18 provide a choice of @xmath15 .
in particular , when @xmath9 is a prime number the trace of the @xmath19th projector @xmath20 is given by the quantum integer @xmath21 $ ] and @xmath22 = q^{2(p-1 ) } + \cdots + q^2 + 1 = \varphi_p(q)\varphi_p(-q).\ ] ] since the graded euler characteristic of the categorified projector agrees with the jones - wenzl projector , if a quotient @xmath23 could be defined it would be closely related a desired evaluation of @xmath24 at a root of unity . in this paper
we prove that there is an interesting quotient of this form and motivate the claim that this is sufficient for topological considerations by identifying structures which are familiar from low dimensional topology .
different approaches to categorification of @xmath25-manifold invariants have been proposed in @xcite .
these authors consider @xmath25-manifold invariants in different contexts , not involving evaluation at a root of unity .
sections [ tl spaces ] and [ categorified tl section ] review constructions of @xmath26 tqfts and the relevant material on categorification from @xcite and @xcite .
we introduce a filtration on coefficients corresponding to the level .
the emphasis in both tqft and categorified settings is on the jones - wenzl projectors . in section [ cosetcon ]
we define a suitable version of localization .
the unusual nature of this construction is motivated in section [ verdier ] by examining the structure of verdier localization .
section [ handleslide ] defines objects @xmath27 which are invariant under handle slides .
this section summarizes parts of the constructions of @xmath26 quantum invariants which are relevant in this paper , the turaev - viro theory associated to a surface and the reshetikhin - turaev invariant of @xmath25-manifolds based on a surgery presentation .
the general outline presented here follows the `` picture tqft '' approach of @xcite .
our exposition differs from those available in the literature in the choice of the coefficient ring , and the main point of this section is contained in [ arelcyc ] which shows that a version of the tqft may be defined using the jones - wenzl projectors , without a reference to a root of unity .
a compact oriented surface @xmath28 with boundary is _ labelled _ when paired with a map @xmath29 which identifies each boundary circle with a model circle containing a fixed set of marked points @xmath30 .
[ tl def ] the _ temperley - lieb space _ @xmath31 of a labelled surface @xmath28 is the set of all @xmath4 $ ] linear combinations of isotopy classes of 1-manifolds ( `` multi - curves '' ) @xmath32 intersecting @xmath33 transversely at marked points and subject to the local relation : removing a simple closed curve bounding a disk in @xmath28 from a multi - curve is equivalent to multiplying the resulting element by the quantum integer @xmath34 = q + q^{-1}$ ] .
@xmath31 is a 2-dimensional version of the kauffman skein module of @xmath35 .
the _ temperley - lieb algebra _
@xmath36 is given by @xmath37 where @xmath38 identifies @xmath39 with a circle containing @xmath40 marked points .
if we view @xmath41 as a rectangle with @xmath42 marked points on top and @xmath42 marked points on the bottom then vertical stacking corresponds to the usual algebra structure on @xmath36 .
the construction of @xmath31 is compatible with gluing surfaces along common labels .
in particular , the restriction to genus zero surfaces gives the structure of a planar algebra . for each @xmath43
the temperley - lieb algebra @xmath36 contains special idempotent elements called the _ jones - wenzl projectors_. if we depict elements of @xmath36 graphically with @xmath42 incoming and @xmath42 outgoing chords then @xmath44{pn } \end{minipage } } = { \begin{minipage}{.8 in } \includegraphics[scale=1.1]{pnm1parjw } \end{minipage } } - \frac{[n-1]}{[n]}\ !
{ \begin{minipage}{.8 in } \includegraphics[scale=1.1]{pnm1sad } \end{minipage } } \ ] ] where the first element @xmath45 is given by a single chord .
jones - wenzl projectors will be central to the discussion in section [ arelcyc ] and sections [ cosetcon ] , [ handleslide ] . for more detail , see @xcite .
recall the quantum integers , @xmath46 = \frac{q^n - q^{-n}}{q - q^{-1 } } = q^{n-1 } + q^{n-3 } + \cdots + q^{-n+3 } + q^{-n+1}.\ ] ] [ restricted ring ] consider the power series @xmath47 determined by the expansion of @xmath48 $ ] into positive powers of @xmath1 . for each @xmath49 the _ restricted ring _
@xmath50[s_1,\ldots , s_{n } ] \subset \mathbb{z}[q^{-1}]\llbracket q \rrbracket\ ] ] is obtained by adjoining @xmath51 to the ring of laurent polynomials @xmath4 $ ] .
we have @xmath52 $ ] and inclusions @xmath53\llbracket q \rrbracket.\ ] ] let @xmath28 be a labeled surface ( see definition [ tl def ] ) .
recall that @xmath31 is the @xmath4 $ ] module given by the temperley - lieb space .
define the _
@xmath19th restricted temperley - lieb space _ by extension of coefficients , @xmath54 } r^n.\ ] ] the restricted space @xmath55 is precisely the extension of @xmath14 necessary to define the first @xmath19 jones - wenzl projectors : @xmath56 ( after @xmath48 $ ] is replaced by @xmath57 in equation ( [ projeq ] ) ) .
each restricted space @xmath55 will have a categorified analogue @xmath58 ( see section [ cateofchcx ] ) and these special coefficients will allow us to control the quotient later .
we would like to consider the quotient of @xmath55 by the ideal @xmath59 .
note that @xmath60 \not\in r^n$ ] , this motivates the definition of the restricted ring @xmath61 above .
[ idealtl ] let @xmath62 .
the _ ideal _
@xmath63 generated by @xmath64 in @xmath65 is the smallest submodule containing all elements obtained by gluing @xmath64 to @xmath66 where @xmath67 with @xmath68 , and @xmath66 is any element of @xmath69 .
where @xmath64 is represented by a box , and the surface @xmath28 is a torus . ]
the notion of an ideal @xmath63 allows one to consider an element @xmath64 of the temperley - lieb algebra as a local relation `` @xmath70 '' among multi - curves on a surface @xmath28 , see @xcite for more details .
the term `` ideal '' is motivated by the notion of an ideal in the planar algebra setting ( which corresponds to restricting our discussion to surfaces of genus zero ) .
we start by briefly recalling a construction of the turaev - viro theory @xcite associated to a surface @xmath28 , following @xcite
. we will then present a slight modification of the definition where the evaluation at a root of unity is replaced by taking a suitable quotient by a jones - wenzl projector .
let @xmath17 and consider the temperley - lieb space @xmath71 with coefficients in the ring @xmath72/{\ensuremath{\langle \varphi_{p}(q^2 ) \rangle}}$ ] , where @xmath8 denotes the @xmath9th cyclotomic polynomial .
then the turaev - viro @xmath0 theory associated to @xmath28 , @xmath73 , is the quotient @xmath31 by the ideal @xmath74 , where @xmath75 denotes the @xmath19th jones - wenzl projector in the temperley - lieb algebra .
note that our definition is somewhat different from that in @xcite , specifically rather than setting @xmath1 to be a root of unity in the ground ring ( which may be algebraically implemented by taking a quotient by the cyclytomic polynomial @xmath8 ) we define the coefficient ring @xmath76 as the quotient by @xmath77 .
the relation between the two is @xmath78 .
the reason for our choice of the coefficient ring is explained in section [ arelcyc ] .
the spaces @xmath79 suffice for the study of mapping class group representations .
such spaces are determined in a canonical way by a 3-dimensional tqft @xcite . on the other hand in order to define a 3-dimensional tqft using such spaces one must choose maps associated to the addition of 3-dimensional handles which transform in accordance with handle slides and cancellations .
when such a choice exists it is unique up to multiplication by a scalar @xcite . in our context this information is determined by the special element @xmath80 ( see section [ rkh ] ) . for reasons explained in section [ categorified tl section ] , it will be convenient for us to consider _ cosets _ of the submodule @xmath81 generated by @xmath82 in @xmath83 .
of course , the quotient @xmath84 may be equivalently considered as the set of cosets of @xmath85 in @xmath31 .
we stress this minor detail so it is easier to follow the categorical construction in section [ categorified tl section ] .
set @xmath17 and recall the sequence of steps used in the definition of @xmath86 above : first the coefficients of @xmath83 are set to be @xmath76 ( this is similar to setting the parameter @xmath1 to be a root of unity ) , and then one takes the quotient of @xmath65 by the submodule @xmath59 .
we will next show that in fact the first step may be omitted .
consider the ring @xmath87/{\ensuremath{\langle \varphi_{p}(q^2 ) \rangle}}$ ] as above .
recall the restricted ring @xmath61 ( definition [ restricted ring ] ) and consider the ideal @xmath88 . given a prime @xmath9 ,
let @xmath17 .
then @xmath89 since @xmath90 $ ] is a unit in @xmath91 the quotient map @xmath4\otimes\mathbb{q } \to k^p \otimes \mathbb{q}$ ] extends to a map @xmath92 .
moreover , since any element of @xmath93 evaluates to zero in @xmath94 , this map descends to a map @xmath95 on the other hand , observe that the cyclotomic polynomial @xmath77 is in the ideal @xmath93 , since @xmath96 .
this gives an inverse map @xmath97 setting the variable @xmath1 to a root of unity corresponds to quotienting the ring of coefficients by the ideal @xmath98 .
since @xmath99 the lemma implies @xmath100 the proof of this lemma generalizes to give a `` picture tqft '' construction of @xmath101 turaev - viro theory for a surface @xmath28 by killing @xmath3 in the @xmath0 kauffman skein module of @xmath102 over @xmath61 .
it is a classical theorem of lickorish and wallace that any 3-manifold @xmath103 can be obtained by surgery on a framed link @xmath104 .
kirby calculus is a means of describing when surgery on two different framed links produce homeomorphic 3-manifolds : @xcite two 3-manifolds @xmath103 and @xmath19 obtained by surgery on @xmath105 along framed links @xmath106 and @xmath107 are diffeomorphic if and only if the associated link diagrams @xmath108 and @xmath109 are related by a sequence of kirby moves : @xmath110{blank } \end{minipage } } \cong { \begin{minipage}{1 in } \includegraphics[scale=1.5]{plus1framed2 } \end{minipage } } { \textnormal{\quad \quad and \quad } } { \begin{minipage}{1 in } \includegraphics[scale=1.5]{handleslide1 } \end{minipage } } \cong { \begin{minipage}{1 in } \includegraphics[scale=1.5]{handleslide2 } \end{minipage } } .$ ] the first move describes the addition of a disjoint @xmath111-framed unknot to the diagram .
it is the second , _ handle slide _ , move that will be the focus of this paper .
the invariance under the first kirby move is achieved by a suitable normalization of the quantum invariant ( cf .
this is not addressed in the present paper . in @xcite
( see also @xcite ) a special element @xmath80 is defined which can be represented by a linear combination of projectors : @xmath112 here @xmath113 is the element of the annulus category obtained by placing the projector @xmath114 in the annulus .
this @xmath80 , sometimes referred to as the magic linear combination of temperley - lieb elements , is the element which is invariant under handle slides used in most constructions of quantum 3-manifold invariants .
the element @xmath80 is invariant under handle slides in the module @xmath115 .
a beautiful proof of this lemma was given by lickorish ( see @xcite ) when @xmath1 equals a root of unity .
we observe that his proof remains true in the annulus module above , where the evaluation at a root of unity is replaced by taking a quotient by the ideal generated by the jones - wenzl projector .
the reshetikhin - turaev invariant of a @xmath25-manifold @xmath103 is defined by presenting @xmath103 as a surgery on a framed link @xmath106 in @xmath105 , labeling each link component by the element @xmath80 and evaluating the jones polynomial of the resulting labeled link at the corresponding root of unity , cf .
@xcite . to relate this to the context of the temperley - lieb spaces and the turaev - viro theory @xmath73 associated to surfaces , note that the jones polynomial of a link may be viewed as the evaluation of the link in the skein module associated to @xmath116 .
the turaev - viro theory discussed above is the quantum double of the reshetikhin - turaev tqft , however in the case of the @xmath117-sphere both theories are @xmath118-dimensional .
therefore it makes sense to analyze the handle - slide property in the module @xmath115 which was defined above in the turaev - viro setting .
the reader should note that the omegas introduced in section [ handleslide ] are not precisely the same as the special element @xmath80 defined above .
the two definitions are compared in section [ relationtomagic ] .
a refined version of @xmath0 quantum invariants may be defined for a @xmath25-manifold with a spin structure @xcite .
one considers a decomposition @xmath119 , corresponding to the @xmath120 grading ( parity ) of the summation indexing set in ( [ omega eq ] ) .
given a presentation of a @xmath25-manifold @xmath103 with a spin structure as the surgery on a framed link @xmath121 , one labels the components of the characteristic sublink by @xmath122 and the rest of the components by @xmath123 .
a spin analogue of the kirby calculus is then used @xcite to prove that an appropriate normalization is an invariant of @xmath124 with a given spin structure .
in this section we recall a version of dror bar - natan s graphical formulation @xcite of the khovanov categorification @xcite .
it will be used throughout the remainder of this paper .
the bar - natan formulation extends to planar algebras and surfaces in a transparent way and is essentially universal ( see @xcite ) and so has the advantage of allowing our constructions to apply to a number of variant categorifications which exist in the literature .
let @xmath28 be a labelled surface .
there is an additive category @xmath125 whose objects are isotopy classes of formally @xmath1-graded 1-manifolds @xmath32 intersecting all boundary components transversely at marked points ( compare to definition [ tl def ] ) .
the morphisms are given by @xmath126 $ ] linear combinations of isotopy classes of orientable cobordisms bounded in @xmath127 $ ] between two surfaces @xmath28 containing such diagrams .
each cobordism is constant along the boundary .
the _ degree _ of a cobordism @xmath128 is given by @xmath129 where the topological degree @xmath130 is given by the euler characteristic of @xmath131 and the @xmath1-degree @xmath132 is given by the relative difference in @xmath1-gradings .
homological degree is not part of the definition @xmath133 .
it has become a common notational shorthand to represent a handle by a dot and a saddle by a flattened diagram containing a dark line .
@xmath134{sheetwithhandle } \end{minipage } } \ ! = 2\ ! { \begin{minipage}{.8 in } \includegraphics[scale=1.1]{sheetwithdot } \end{minipage } } = 2 { \begin{minipage}{.8 in } \includegraphics[scale=1.1]{dottedline } \end{minipage } } \textnormal { and } { \begin{minipage}{.8 in } \includegraphics[scale=1.1]{saddle } \end{minipage } } \,\ , = \!\ ! { \begin{minipage}{.8 in } \includegraphics[scale=1.1]{n2 - 1-sad } \end{minipage } }
$ ] ( the topological degrees of the cobordisms above are @xmath135 , @xmath136 respectively . ) when @xmath137 we would like a category @xmath138 such that @xmath139
so we require that the object represented by a closed circle be isomorphic to sum of two empty objects in degrees @xmath111 respectively .
if such maps are to be degree preserving then the most natural choice for these maps is given below .
@xmath140{localcircle } \end{minipage } } \hspace{.1 in } & \pile{\rto^{\left ( { \begin{minipage}{.45 in } \includegraphics[scale=.75]{cap - circ - to - null } \end{minipage } } { \begin{minipage}{.45 in } \includegraphics[scale=.75]{torus - circ - to - null } \end{minipage } } \right ) } \\
\lto_{\left ( { \begin{minipage}{.45 in } \includegraphics[scale=.75]{torus - null - to - circ } \end{minipage } } { \begin{minipage}{.45 in } \includegraphics[scale=.75]{cap - null - to - circ } \end{minipage } } \right ) } } & \hspace{.1 in } q^{-1 } { \begin{minipage}{.45 in } \includegraphics[scale=.75]{localblank } \end{minipage } } \ , \oplus\ , q { \begin{minipage}{.45 in } \includegraphics[scale=.75]{localblank } \end{minipage } } & \hspace{.15 in } : \ ! \psi \end{diagram }
$ ] in order to obtain @xmath141 and @xmath142 we form a new category @xmath143 obtained as a quotient of the category @xmath125 by the local relations given below .
@xmath144{sphere } \end{minipage } } = 0 \hspace{.75 in } { \begin{minipage}{.45 in } \includegraphics[scale=.75]{spheredot } \end{minipage } } = 1 \hspace{.75 in } { \begin{minipage}{.45 in } \includegraphics[scale=.75]{spheredotdot } \end{minipage } } = 0 \hspace{.75 in } { \begin{minipage}{.45 in } \includegraphics[scale=.75]{spheredotdotdot } \end{minipage } } = \alpha$ ] @xmath144{cyl } \end{minipage } } = { \begin{minipage}{.45 in } \includegraphics[scale=.75]{cut1 } \end{minipage } } + { \begin{minipage}{.45 in } \includegraphics[scale=.75]{cut2 } \end{minipage } } $ ] the neck cutting relation can be applied to any compressing disk .
in particular , that this means the isomorphisms @xmath38 and @xmath145 can not be applied to remove a non - trivial circle from the annulus .
the cylinder or neck cutting relation implies that closed surfaces @xmath146 of genus @xmath147 evaluates to @xmath148 .
the relations above imply that a sheet with two dots is equal to @xmath149 times a sheet with no dots .
in what follows we will let @xmath149 be a free variable and absorb it into our base ring ( @xmath150 ) .
one can think of @xmath149 as a deformation parameter , see @xcite for further details .
for any labelled surface @xmath28 there is an isomorphism of @xmath4$]-modules , @xmath151 the proof follows directly from the construction above .
note that when @xmath152 this is precisely the same setting as @xcite and @xcite .
for an extended discussion of grothendieck groups in this context see @xcite .
surfaces with common boundary labels can be glued and the construction of @xmath143 is compatible with this gluing .
[ kom def ] if @xmath28 is a labelled surface let @xmath153 be the category of chain complexes of finite direct sums of objects in @xmath143 . in @xmath154
we allow chain complexes @xmath155 of unbounded positive homological degree and bounded negative homological degree .
let @xmath156 denote the category associated to the disk .
let @xmath157 be the category of bounded complexes .
in previous work the authors showed that the category @xmath158 contains special objects @xmath159 which categorify the jones - wenzl projectors @xmath160 .
@xcite there exists a chain complex @xmath161 called the _ universal projector _ which satisfies 1 .
@xmath159 is positively graded with degree zero differential .
2 . the identity diagram appears only in homological degree zero and only once .
the chain complex @xmath162 is contractible under turnbacks .
these three properties guarantee that @xmath159 is unique up to homotopy , @xmath163 and @xmath164 .
other authors @xcite constructed projectors using different techniques . given @xmath165 and a labeled surface @xmath28 ,
let @xmath166 be the full subcategory of @xmath167 consisting of those chain complexes whose image under @xmath168 is contained in @xmath169 . by construction ,
@xmath170 where @xmath157 is the category of bounded complexes and @xmath167 is as in definition [ kom def ] .
the idea of localization is central to this paper . we seek to carry out a categorified analogue of the `` picture tqft '' construction in sections
[ su2 invariants ] , [ arelcyc ] , and a central step in this program is to kill @xmath3 in @xmath166 for all @xmath28 . in order to accomplish this
we find a subcategory @xmath171 of @xmath166 the image of which corresponds to the ideal @xmath81 in the grothendieck group @xmath172 .
a natural family of morphisms is then inverted ( definition [ localset ] ) in which we identify any object @xmath173 in @xmath166 with any cone of @xmath173 on any object @xmath174 .
this localization can be understood concretely ( definition [ coset definition ] ) in terms of iterated cones over the ideal @xmath171 . in section [ verdier ]
we explain why the usual notion of verdier localization is not suitable in our context , leading us to a different version of localization described next . in what follows ,
we fix @xmath175 and use @xmath176 to denote @xmath3 and @xmath177 to denote @xmath166 for some @xmath28 .
also @xmath24 is the homotopy category of @xmath177 .
[ ideal ] given @xmath64 in @xmath178 , the _ ideal _ @xmath179 generated by @xmath64 is the smallest full subcategory of @xmath180 which contains all objects obtained by gluing @xmath64 to @xmath66 where @xmath181 with @xmath68 , @xmath66 is any object of @xmath182 , and which is closed under cones and grading shifts .
the reader should compare this to definition [ idealtl ] .
@xmath134{omega3eq1a } \end{minipage } } , \quad { \begin{minipage}{.8 in } \includegraphics[scale=1.1]{omega3eq2a } \end{minipage } } .$ ] notice that @xmath183 ; the subcategory @xmath184 is not thick , compare with section [ verdier ] .
the definition is chosen so that the following proposition holds .
if @xmath24 and @xmath185 are as above , then @xmath186 where @xmath187 is the ideal generated by @xmath176 ( definition [ idealtl ] ) .
the proof follows from definition [ ideal ] . for more on @xmath168 of infinite complexes
see @xcite . for a discussion of @xmath168 of triangulated categories
see @xcite . ( localization of a category)[loccat ] given a collection @xmath188 of morphisms of a category @xmath189
, there exists a category @xmath190 $ ] and a functor @xmath191 $ ] satisfying the universal property : * if @xmath192 is a functor and @xmath193 is an isomorphism for all @xmath194 then there exists a functor @xmath195 \to { \ensuremath{\mathcal{d}}}$ ] such that @xmath196 .
the category @xmath190 $ ] is constructed by realizing @xmath189 as a directed graph , adding edges corresponding to inverses of elements of @xmath188 and taking morphisms to be paths in this graph subject to the most obvious relations .
for details see @xcite .
( `` brusque quotient'')[localset ] given @xmath197 , if @xmath198 or @xmath199 there are maps @xmath200 and @xmath201 $ ] .
we wish to invert all such maps . set @xmath202 and define @xmath203.\ ] ] note that this version of localization has some unusual properties .
for instance , @xmath204 in the quotient , and the familiar triangulated structure is now gone .
section [ verdier ] explains how this definition is different from the usual verdier localization .
we will now show that the above construction can be reformulated in a more concrete way .
( the reader may note that the theme of characterizing localization in terms of iterated extensions appears in algebraic topology literature , cf .
we begin with an equivalence relation on objects in @xmath177 .
( @xmath205equivalence ) let @xmath206 be as above then for any @xmath207 we say that @xmath208 is @xmath184-_equivalent _ to @xmath66 if there exists a sequence of maps @xmath209 and objects @xmath210 such that 1 .
@xmath211 or @xmath212 for some @xmath213 2 .
@xmath214 or @xmath215 and @xmath216 this equivalence relation amounts to isomorphism in the category @xmath217 defined above : @xmath218 in @xmath217 if and only if @xmath208 is @xmath184-equivalent to @xmath66 in @xmath24 .
the first direction is obvious .
suppose @xmath218 in @xmath217 then by definition an isomorphism @xmath219 is a map of the form @xmath220 where @xmath113 is an isomorphism in @xmath24 and @xmath221 is an isomorphism of the form @xmath222 or @xmath223 as in definition [ localset ] above .
a map @xmath224 tells us that either @xmath225 is equivalent to @xmath226 coned on @xmath227 or vice versa .
given an object @xmath228 we can combine all of the objects which are @xmath184-equivalent to it to form a category which is a natural analogue of the coset @xmath229 in the image of the grothendieck group .
( coset categories ) [ coset definition ] given @xmath230 , the _ coset _ associated to @xmath208 is the full subcategory , denoted @xmath231 , consisting of objects @xmath232 which are @xmath184-equivalent to @xmath208 in @xmath24 .
isomorphism classes of coset subcategories are in bijection with @xmath184-equivalence classes : @xmath208 is @xmath184-equivalent to @xmath66 if and only if @xmath233 notice that the grothendieck group of a coset category @xmath234 can be identified with the coset @xmath235 . in particular , the non - triviality of @xmath236 implies that the quotient @xmath217 is highly non - trivial .
more precisely , in light of remark at the end of section [ arelcyc ] , the machinery developed in this section may be viewed as a categorification of the @xmath117-dimensional part of the turaev - viro theory . in @xcite khovanov
considered additive monoidal categories @xmath237 of graded objects in which @xmath238 such an isomorphism implies that the polynomial @xmath239 is equal to @xmath148 in the grothendieck group @xmath240 . in this paper
, @xmath241 is isomorphic to a chain complex consisting only of empty diagrams @xcite and the relation @xmath242 implies that the polynomial @xmath243 is zero in the grothendieck group .
there are two other more notable differences between these ideas . in our context
, the quotient must be homotopy invariant because @xmath3 is only determined up to homotopy .
also inasmuch as @xmath3 is not determined by its graded euler characteristic , the effect of the quotient on the category has less to do with the polynomial @xmath244 and more to do with the structure of @xmath3 itself .
in section [ cosetcon ] we quotiented @xmath166 by @xmath3 in an unusual way . in this section
we recall elements of the most common quotient construction for triangulated categories . in particular , we explain why it is not useful in the context of this paper .
readers interested in a detailed discussion of the verdier quotient should consult @xcite .
given a triangulated category @xmath189 we would like to invert some family of maps @xmath245 so that the objects of a full triangulated subcategory @xmath246 become isomorphic to zero in a triangulated quotient category @xmath247 = { \ensuremath{\mathcal{c}}}/{\ensuremath{\mathcal{d}}}$ ] . in order to produce a triangulated quotient category @xmath247 $ ] the maps in @xmath245
are required to admit a calculus of left and right fractions .
the collection @xmath245 is additionally required to be closed under homological grading shifts and if @xmath248 is a map between triangles with @xmath249 then there exists an @xmath250 so that @xmath251 is a map between triangles .
most of these conditions are not imposed by the brusque quotient of section [ cosetcon ] .
if @xmath252 is a full triangulated subcategory of a triangulated category @xmath189 , then @xmath245 is defined to be the collection of maps @xmath253 which fit into an exact triangle : @xmath254\ ] ] where @xmath255 .
if the collection @xmath245 is chosen in this manner then it satisfies all of the properties described above .
the quotient functor @xmath256 $ ] is a triangulated functor and universal in the category of small triangulated categories .
the quotient functor @xmath257 in proposition [ loccat ] does not satisfy these properties .
a _ thick _ ( or _ paisse _ ) subcategory @xmath258 is a subcategory which is closed under retracts : @xmath259 .
the verdier quotient only sees the smallest thick subcategory containing the subcategory under consideration . more specifically ,
if @xmath258 is a subcategory and @xmath260 is the smallest thick subcategory containing @xmath252 then @xmath261 the following proposition explains our problem .
the smallest thick subcategory containing the ideal @xmath262 is @xmath263 . in @xcite
it was shown that @xmath264 .
it follows that @xmath265 .
the following general statement also holds . if @xmath106 is any link or @xmath266tangle then the smallest thick subcategory containing the ideal @xmath267 is @xmath268 the chain complex associated to @xmath106 is homotopy equivalent to a chain complex in which the rasmussen invariant can be computed from the grading of a trivial direct summand ( see @xcite ) .
a conceptual explanation could be formulated by instead considering the category @xmath269 ( compare @xcite ) . here
thick subcategories are in correspondence with closed subschemes @xcite and @xmath270/(x^2))$ ] has no interesting subschemes .
other quotients such as those of bousfield and drinfeld also respect thickness in the manner described above .
it seems that the intricacies of these homotopy theories are not seen by the usual quotients .
in this section we use the framework discussed above to define objects @xmath27 which are invariant under handle slides .
these objects categorify the magic elements @xmath271 in the annular temperley - lieb space for low levels @xmath19 . recall that @xmath271 is a crucial ingredient in the construction of the @xmath0 tqft at level @xmath19 , see section [ rkh ] .
the construction of these objects relies on a detailed analysis of the universal projector complexes from @xcite .
the second projector is defined @xcite to be the chain complex @xmath272{p2box } \end{minipage } } \ ! & = & \ ! { \begin{minipage}{.45 in } \includegraphics[scale=.75]{n2 - 1 } \end{minipage } } & \rto^ { { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n2 - 1-sad } \end{minipage } } } & q { \begin{minipage}{.45 in } \includegraphics[scale=.75]{n2-s } \end{minipage } } & \rto^ { { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n2-tops } \end{minipage } } \!\!\!- { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n2-bots } \end{minipage } } } & q^{3 } { \begin{minipage}{.45 in } \includegraphics[scale=.75]{n2-s } \end{minipage } } & \rto^ { { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n2-tops } \end{minipage } } \!\!\ ! + { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n2-bots } \end{minipage } } } & q^{5 } { \begin{minipage}{.45 in } \includegraphics[scale=.75]{n2-s } \end{minipage } } & \cdots \end{diagram}\ ] ] @xmath275{n2 - 1-sad } \end{minipage } } & : { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n2 - 1 } \end{minipage } } \!\ ! \to
q { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n2-s } \end{minipage } } & n = 0\\ { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n2-tops } \end{minipage } } \!\!\ ! + { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n2-bots } \end{minipage } } & : q^{4k-1 } { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n2-s } \end{minipage } } \!\ ! \to q^{4k + 1 } { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n2-s } \end{minipage } } & n \ne 0 , n = 2k\\ { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n2-tops } \end{minipage } } \!\!\ ! - { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n2-bots } \end{minipage } } & : q^{4k + 1 } { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n2-s } \end{minipage } } \!\
! \to q^{4k+3 } { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n2-s } \end{minipage } } & n = 2k+1 \end{array } \right.\ ] ] @xmath278{n2-s } \end{minipage } } \hspace{-.04in}\bigg ] } & = & q { \begin{minipage}{.45 in } \includegraphics[scale=.75]{n2-s } \end{minipage } } & \rto^ { { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n2-tops } \end{minipage } } \!\!\!- { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n2-bots } \end{minipage } } } & q^{3 } { \begin{minipage}{.45 in } \includegraphics[scale=.75]{n2-s } \end{minipage } } & \rto^ { { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n2-tops } \end{minipage } } \!\!\ ! + { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n2-bots } \end{minipage } } } & q^{5 } { \begin{minipage}{.45 in } \includegraphics[scale=.75]{n2-s } \end{minipage } } & \cdots \end{diagram}\ ] ] @xmath279{p2box } \end{minipage } } = \operatorname{cone}\left(\hspace{-.02 in } { \begin{minipage}{.45 in } \includegraphics[scale=.75]{n2 - 1 } \end{minipage } } \hspace{-.05 in } \to { \bigg [ { \begin{minipage}{.45 in } \includegraphics[scale=.75]{n2-s } \end{minipage } } \hspace{-.04in}\bigg]}\right ) { \textnormal{\hspace{.08 in } and \hspace{.08 in } } } { \begin{minipage}{.45 in } \includegraphics[scale=.75]{n2 - 1 } \end{minipage } } \hspace{-.09 in } \simeq \operatorname{cone}\left({\bigg [ { \begin{minipage}{.45 in } \includegraphics[scale=.75]{n2-s } \end{minipage } } \hspace{-.04in}\bigg ] } \to { \begin{minipage}{.45 in } \includegraphics[scale=.75]{p2box } \end{minipage } } \right)\ ] ] ( @xmath281 ) [ omega2 ] in the category @xmath282)$ ] let @xmath283 be the object consisting of a single curve about the origin and @xmath284 be the object consisting of a single nullhomotopic simple closed curve .
let @xmath285 , we represent @xmath27 by a curve decorated with a dot .
graphically , @xmath286{omega2p } \end{minipage } } , \quad \omega_{- } = { \begin{minipage}{.8 in } \includegraphics[scale=1.1]{omega2 m } \end{minipage } } , \quad { \begin{minipage}{.8 in } \includegraphics[scale=1.1]{omega } \end{minipage } } = { \begin{minipage}{.8 in } \includegraphics[scale=1.1]{omega2p } \end{minipage } } \oplus { \begin{minipage}{.8 in } \includegraphics[scale=1.1]{omega2 m } \end{minipage } } .\ ] ] consider first the partial trace of @xmath277 in @xmath282)$ ] , @xmath288{omega2eq2a } \end{minipage } } = \operatorname{cone}\left(\hspace{-.05 in } { \begin{minipage}{.8 in } \includegraphics[scale=1.1]{omega2eq2b } \end{minipage } } \hspace{-.05 in } \to { \bigg[\hspace{-.0 in } { \begin{minipage}{.8 in } \includegraphics[scale=1.1]{omega2eq2c } \end{minipage } } \hspace{-.04in}\bigg]}\right).\ ] ] as above , the square brackets denote the tail of the second projector .
consider also the following diagram : @xmath288{omega2eq1a } \end{minipage } } = \operatorname{cone}\left(\hspace{-.05 in } { \begin{minipage}{.8 in } \includegraphics[scale=1.1]{omega2eq1b } \end{minipage } } \hspace{-.05 in } \to { \bigg[\hspace{-.0 in } { \begin{minipage}{.8 in } \includegraphics[scale=1.1]{omega2eq1c } \end{minipage } } \hspace{-.04in}\bigg]}\right).\ ] ] note that the tail of the first equation and the tail of the second equation above are equal chain complexes .
also the left - hand sides of both equations are in the ideal @xmath262 ( definition [ ideal ] ) , since they contain the second projector .
these two observations taken together imply that the equation @xmath288{omega2eq1b } \end{minipage } } \cong { \bigg[\hspace{-.0 in } { \begin{minipage}{.8 in } \includegraphics[scale=1.1]{omega2eq1c } \end{minipage } } \hspace{-.04in}\bigg ] } = { \bigg[\hspace{-.0 in } { \begin{minipage}{.8 in } \includegraphics[scale=1.1]{omega2eq2c } \end{minipage } } \hspace{-.04in}\bigg ] } \cong { \begin{minipage}{.8 in } \includegraphics[scale=1.1]{omega2eq2b } \end{minipage } } \ ] ] holds among @xmath262 cosets .
for example , the first isomorphism holds since the diagram on the left is a cone of the inclusion of the tail into an element of @xmath262 .
this says that a single strand placed on the lefthand side of @xmath283 is isomorphic to a single strand placed on the righthand side of @xmath284 as @xmath262 cosets and vice versa .
this equation together with its vertical reflection imply that @xmath288{omega2l } \end{minipage } } \cong { \begin{minipage}{.8 in } \includegraphics[scale=1.1]{omega2r } \end{minipage } } \ ] ] we start this section by recalling the minimal ( in the sense that it is not homotopic to a chain complex with fewer temperley - lieb diagrams ) chain complex for @xmath289 , introduced in @xcite .
then we show how it can be used to define an element @xmath27 in the annular category @xmath290)$ ] which is invariant under handle slides in the associated coset category .
@xmath291{n3 - 1 } \end{minipage } } \ , & \rto^a & \ , q^{1}\left(\ ! { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-e1 } \end{minipage } } \,\ , \oplus { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-e2 } \end{minipage } } \!\right ) & \rto^b & \ , q^{2}\left(\ ! { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-e1e2 } \end{minipage } } \,\ , \oplus { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-e2e1 } \end{minipage } } \!\right ) & \rto^c & \ , q^{4}\left(\ ! { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-e1e2 } \end{minipage } } \,\ , \oplus { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-e2e1 } \end{minipage } } \!\right ) \\ & & & & & & \dto^d\\ \cdots & \lto & \ , q^{8}\left(\ ! { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-e1e2 } \end{minipage } } \,\ ,
\oplus { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-e2e1 } \end{minipage } } \!\right ) & \lto^b & \ , q^{7}\left(\ ! { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-e1 } \end{minipage } } \,\ , \oplus { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-e2 } \end{minipage } } \!\right ) & \lto^e & \ , q^{5}\left ( { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-e1 } \end{minipage } } \,\ , \oplus { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-e2 } \end{minipage } } \!\right ) \\ \end{diagram}\ ] ] @xmath293{n3-b-00 } \end{minipage } } & - { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-b-01 } \end{minipage } } \\ - { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-b-10 } \end{minipage } } & { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-b-11 } \end{minipage } } \\
\end{array } \right ) & c & = \left(\begin{array}{cc } { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-e1e2-dot - u } \end{minipage } } + { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-e1e2-dot - d } \end{minipage } } & { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-c-01 } \end{minipage } } \\ { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-c-10 } \end{minipage } } & { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-e2e1-dot - u } \end{minipage } } + { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-e2e1-dot - d } \end{minipage } } \\
\end{array } \right ) & \\ d & = \left(\begin{array}{cc } { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-d-00 } \end{minipage } } & - { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-d-01 } \end{minipage } } \\ - { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-d-10 } \end{minipage } } & { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-d-11 } \end{minipage } } \\
\end{array } \right ) & e & = \left(\begin{array}{cc } { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-e2-dot - u } \end{minipage } } + { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-e2-dot - d } \end{minipage } } & { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-e-01 } \end{minipage } } \\ { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-e-10 } \end{minipage } } & { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-e1-dot - u } \end{minipage } } + { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-e1-dot - d } \end{minipage } } \\
\end{array } \right ) & \\\end{aligned}\ ] ] @xmath295{n3-e1 } \end{minipage } } \hspace{-.04 in } \oplus\!\ ! { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-e2 } \end{minipage } } \ ! \rightarrow \ ! { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-e1e2 } \end{minipage } } \hspace{-.04 in } \oplus\!\ ! { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-e2e1 } \end{minipage } } \!\bigg]}}= \operatorname{cone}\left ( { \begin{minipage}{.35 in } \includegraphics[scale=.45]{p3box } \end{minipage } } \to { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3 - 1 } \end{minipage } } \right)\ ] ] @xmath295{n3-e1 } \end{minipage } } \hspace{-.04 in } \oplus\!\ ! { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-e2 } \end{minipage } } \
! \rightarrow \ !
{ \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-e1e2 } \end{minipage } } \hspace{-.04 in } \oplus\!\ ! { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-e2e1 } \end{minipage } } \!\bigg]}}\simeq q^{1}\left(\ ! { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-e1 } \end{minipage } } \,\ ,
\oplus { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-e2 } \end{minipage } } \!\right ) \to q^{2}\left(\ ! { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-e1e2 } \end{minipage } } \,\ , \oplus { \begin{minipage}{.35 in } \includegraphics[scale=.45]{n3-e2e1 } \end{minipage } } \!\right ) \to \cdots\ ] ] ( @xmath296 ) [ omega3 ] in the category @xmath290)$ ] let @xmath283 be the object consisting of two curves about the origin and @xmath284 be the object consisting of a single curve about the origin and a single nullhomotopic simple closed curve .
let @xmath285 , we represent @xmath27 by a curve decorated with a dot then in terms of pictures . graphically , @xmath286{omega3p } \end{minipage } } , \quad \omega_{- } = { \begin{minipage}{.8 in } \includegraphics[scale=1.1]{omega3 m } \end{minipage } } , \quad { \begin{minipage}{.8 in } \includegraphics[scale=1.1]{omega } \end{minipage } } = { \begin{minipage}{.8 in } \includegraphics[scale=1.1]{omega3p } \end{minipage } } \oplus { \begin{minipage}{.8 in } \includegraphics[scale=1.1]{omega3 m } \end{minipage } } .\ ] ] @xmath288{omega3eq1a } \end{minipage } } = { \begin{minipage}{.8 in } \includegraphics[scale=1.1]{omega3eq1b } \end{minipage } } \hspace{-.03in}\to { \ensuremath{\bigg[\ ! { \begin{minipage}{.8 in } \includegraphics[scale=1.1]{omega3eq1tailbox1 } \end{minipage } } \hspace{-.04 in } \oplus\!\ ! { \begin{minipage}{.8 in } \includegraphics[scale=1.1]{omega3eq1tailbox2 } \end{minipage } } \
! \rightarrow \ ! { \begin{minipage}{.8 in } \includegraphics[scale=1.1]{omega3eq1tailbox3 } \end{minipage } } \hspace{-.04 in } \oplus\!\ ! { \begin{minipage}{.8 in } \includegraphics[scale=1.1]{omega3eq1tailbox4 } \end{minipage } } \!\bigg]}}\ ] ] @xmath288{omega3eq2a } \end{minipage } } = { \begin{minipage}{.8 in } \includegraphics[scale=1.1]{omega3eq2b } \end{minipage } } \hspace{-.03in}\to { \ensuremath{\bigg[\ ! { \begin{minipage}{.8 in } \includegraphics[scale=1.1]{omega3eq2tailbox1 } \end{minipage } } \hspace{-.04 in } \oplus\!\ ! { \begin{minipage}{.8 in } \includegraphics[scale=1.1]{omega3eq2tailbox2 } \end{minipage } } \ ! \rightarrow \ ! { \begin{minipage}{.8 in } \includegraphics[scale=1.1]{omega3eq2tailbox3 } \end{minipage } } \hspace{-.04 in } \oplus\!\ ! { \begin{minipage}{.8 in } \includegraphics[scale=1.1]{omega3eq2tailbox4 } \end{minipage } } \!\bigg]}}.\ ] ] @xmath299{omega3eq1tailbox1 } \end{minipage } } \hspace{.05 in } \oplus\!\ ! { \begin{minipage}{.45 in } \includegraphics[scale=.75]{omega3eq1tailbox2 } \end{minipage } } \
! \rightarrow \ !
{ \begin{minipage}{.45 in } \includegraphics[scale=.75]{omega3eq1tailbox3 } \end{minipage } } \hspace{.05 in } \oplus\!\ ! { \begin{minipage}{.45 in } \includegraphics[scale=.75]{omega3eq1tailbox4 } \end{minipage } } \!\bigg]}}\cong { \ensuremath{\bigg[\ ! { \begin{minipage}{.45 in } \includegraphics[scale=.75]{omega3eq2tailbox1 } \end{minipage } } \hspace{.05 in } \oplus\!\ ! { \begin{minipage}{.45 in } \includegraphics[scale=.75]{omega3eq2tailbox2 } \end{minipage } } \
! \rightarrow \ !
{ \begin{minipage}{.45 in } \includegraphics[scale=.75]{omega3eq2tailbox3 } \end{minipage } } \hspace{.05 in } \oplus\!\ ! { \begin{minipage}{.45 in } \includegraphics[scale=.75]{omega3eq2tailbox4 } \end{minipage } } \!\bigg]}}\ ] ] since the left hand side of both equations above include the third projector they are both elements of the ideal @xmath300 .
it follows that @xmath288{omega3eq1b } \end{minipage } } \cong { \begin{minipage}{.8 in } \includegraphics[scale=1.1]{omega3eq2b } \end{minipage } } , \ ] ] but this equation says that a free strand can be slid over @xmath283 at the cost of turning the @xmath283 into an @xmath284 .
this equation together with its vertical reflection imply that the @xmath300 coset associated to the @xmath27 defined above is invariant under handle slides .
here we compare the @xmath301 elements categorified in [ second projector section ] and [ third projector section ] to the special @xmath80 of section [ rkh ] .
we use @xmath173 to represent a single essential circle in the annulus and the relation @xmath302 below . in section [ second projector section ]
when @xmath303 we set @xmath304 $ ] .
it follows that @xmath305 x = \phi_2 + \phi_0 + [ 2 ] \phi_1 = \omega_2 $ ] .
so , @xmath306 .
the handle slide property implies that @xmath307\omega$ ] . in section [ third projector section ] when @xmath308 we set @xmath309 x$ ] . if @xmath310 \phi_1 + [ 3 ] \phi_2 $ ] then @xmath311 + 1 ) x + [ 2 ] x^2 $ ] since @xmath312 when @xmath308 and @xmath313 + 1=[2]^2 \rightarrow x\omega_3 = [ 2 ] \omega$ ] .
since @xmath314 satisfies the handle slide property , @xmath34 \omega_3 = x \omega_3 = [ 2]\omega \rightarrow \omega_3 = \omega$ ] since @xmath34 $ ] is a unit in @xmath315 by construction .
one may use definitions [ omega2 ] , [ omega3 ] in the context of @xmath25-manifolds with a spin structure . specifically , one labels the components of the characteristic sublink by @xmath316 and the rest of the components by @xmath317 , compare with section [ spin ]
. theorems [ omega2 thm ] , [ omega3 thm ] then state that sliding a strand interchanges the cosets associated to @xmath316 and @xmath317 , as one expects in the context of spin kirby calculus .
this section introduced a categorification of the elements @xmath318 , where @xmath271 is defined in ( [ omega eq ] ) .
it is shown that these objects are invariant under handle slides , understood in the context of the coset construction of section [ cosetcon ] .
the authors conjecture that there is also a categorification of @xmath271 for all @xmath19 , however it is likely that a construction would require a further non - trivial extension of ideas presented here .
one can construct an object in the brusque quotient category ( section [ cosetcon ] ) by labelling each component of a framed link @xmath121 with @xmath27 .
theorems [ omega2 thm ] and [ omega3 thm ] imply that the isomorphism type of this object is unchanged by handle slides .
this reflects information about the 3-manifold obtained by surgery on @xmath106 ( see section [ kirby ] ) .
we expect that there are maps between such objects associated to @xmath294-dimensional cobordisms .
i. frenkel , c. stroppel and j. sussan , _ categorifying fractional euler characteristics , jones - wenzl projector and @xmath322-symbols with applications to exts of harish - chandra bimodules _
, arxiv:1007.4680 . | we propose a means by which some categorifications can be evaluated at a root of unity .
this is implemented using a suitable localization in the context of prior work by the authors on categorification of the jones - wenzl projectors . within this construction
we define objects , invariant under handle slides , which decategorify to the @xmath0 quantum invariants at low levels . |
wir untersuchen , inwiefern man auf einer carnot - mannigfaltigkeit @xmath0 ein spektrales tripel konstruieren kann , welches die carnot - carathodory metrik und die gradierte dimension von @xmath0 erkennen soll .
zu diesem zweck konstruieren wir selbst - adjungierte horizontale dirac operatoren @xmath1 und zeigen , dass zwar jeder horizontale dirac operator ber connes formel die metrik erkennt , allerdings in keinem fall hypoelliptisch ist und somit keine kompakte resolvente besitzen kann
. zunchst betrachten wir ein beispiel auf kompakten carnot nilmannigfaltigkeit detailliert , wobei wir eine konstruktion fr einen horizontalen dirac operator ber den pull - back des dirac operators auf dem torus durchfhren .
einer methode von christian br folgend knnen wir das horizontale clifford bndel dieses operators zerlegen und erkennen , dass der operator einen unendlich dimensionalen kern besitzt .
dennoch knnen wir i m fall von heisenberg nilmannigfaltigkeiten die gradierte dimension aus dem asymptotischen verhalten der eigenwerte dieses horizontalen dirac - operators erkennen .
anschlieend wenden wir uns dem allgemeinen fall zu , indem wir zeigen dass ein beliebiger horizontaler dirac operator nicht hypoelliptisch ist .
dazu entwickeln wir ein kriterium mit dem man die hypoelliptizitt von bestimmten gradierten differentialoperatoren ausschlieen kann indem man die situation auf einer heisenberg mannigfaltigkeit betrachtet , fr welche eine vollstndige charakterisierung der hypoelliptizitt durch die rockland - bedingung gegeben ist .
schlielich zeigen wir wie man spektrale tripel aus horizontalen laplace operatoren mit hilfe des heisenberg pseudodifferentialkalkls , das von richard beals und peter greiner entwickelt wurde , konstruieren kann .
wir stellen ein paar explizite konstruktionen vor und diskutieren , unter welchen voraussetzungen es mglich sein kann aus diesen operatoren eine zu der carnot - carathodory metrik quivalente metrik zu erhalten .
zustzlich erwhnen wir eine formel , mit der die carnot - carathodory metrik aus beliebigen horizontalen laplace operatoren erkannt werden kann .
* schlsselwrter : * spektrales tripel , carnot - carathodory metrik , hypoelliptizitt .
_ du sollst dich zu einer stadt begeben , die den namen tsnips - egn - rih trgt ! _ ( walter moers )
in the 1980s , alain connes presented the concept of non - commutative geometry as an extension of the usual notion of a topological space ( see e.g. @xcite , @xcite , @xcite ) .
the idea goes back to the 1940s , when israel gelfand and mark naimark showed that every commutative @xmath2-algebra is isomorphic to the @xmath2-algebra @xmath3 of continuous functions on a locally compact hausdorff space @xmath4 vanishing at infinity ( see @xcite ) . from this starting point
, one sees that a lot of properties of the space @xmath4 can be translated into properties of its @xmath2-algebra @xmath3 .
motivated by this , in non - commutative geometry one considers a general @xmath2-algebra as a non - commutative space . to describe geometry on a non - commutative space , connes introduced so - called _ spectral triples_. the definition of a spectral triple is suggested by the fact that many geometric properties of a compact connected riemannian spin manifold @xmath0 without boundary can be obtained from the dirac operator @xmath5 acting on a clifford bundle @xmath6 over @xmath0 .
for example , one can reproduce the dimension of @xmath0 by the asymptotic growth of the eigenvalues of @xmath5 via @xmath7 and one can detect the geodesic distance on @xmath0 by the formula @xmath8\right\|_{l^2(\sigma m ) } \leq 1 \right\}.\ ] ] these properties can be transported to the picture of @xmath2-algebras : a spectral triple is a triple @xmath9 where @xmath10 is a @xmath2-algebra , @xmath11 is a hilbert space carrying a faithful action of @xmath10 on @xmath12 and @xmath13 is a self - adjoint operator on @xmath11 such that @xmath14 $ ] is bounded for @xmath15 belonging to a dense sub - algebra of @xmath10 and such that the resolvent @xmath16 of @xmath13 is compact . for a general spectral triple ,
one can define notions of dimension and metric in analogy to the formulas above .
hence on a closed riemannian spin manifold with dirac operator @xmath5 , a spectral triple which detects the dimension and the geodesic metric on @xmath0 is given by the triple @xmath17 . during the last decades
there have been several approaches to construct spectral triples for more examples than the one of a closed riemannian spin manifold .
in particular there are some constructions for certain fractals which have a non - integer hausdorff dimension , done for example by erik christensen , christina ivan and their collaborators ( see e.g. @xcite , @xcite ) : for some fractals , it is possible to detect the hausdorff dimension as well as the geodesic distance of the space from a spectral triple .
a more general approach has been suggested by ian palmer ( @xcite ) and by john pearson and jean bellissard @xcite ) .
palmer shows that under mild conditions the hausdorff dimension of every compact metric space can be discovered by a spectral triple . for these triples
one can also find an estimate from above for the metric .
but the constructions palmer is considering lead away from the geometry of the space , since these constructions make use of an approximation of the space by a discrete subset . from the aspect of dimension
, it is also interesting to note that there has been a construction by erik christensen and christina ivan showing that to any given positive real number @xmath18 , one can construct a spectral triple of dimension @xmath18 , where @xmath10 is a limit of finite - dimensional @xmath2-algebras and @xmath13 is a limit of finite - dimensional operators ( @xcite ) .
hence at least theoretically it is possible to define spectral triples of arbitrary dimension on certain spaces .
the question we are dealing with in this thesis is whether it is possible to define two different spectral triples on one space , which both give reasonable geometries ( in terms of dimension and metric ) .
therefore we consider so - called _ sub - riemannian manifolds _ ( or in a more specialized setting _ carnot manifolds _ ) , which are riemannian manifolds @xmath0 equipped with a bracket - generating horizontal sub - bundle @xmath19 of their tangent bundle @xmath20 .
it has been detected by wei - liang chow ( @xcite ) in 1939 that in this case and if @xmath0 is connected , any two points on @xmath0 can be connected by a curve which is tangent to the horizontal distribution @xmath19 .
this means that we obtain a metric on @xmath0 via the formula @xmath21 which differs from the metric induced by the geodesic distance on @xmath0 .
in addition , the hausdorff dimension of the metric space @xmath22 turns out to be strictly greater than the hausdorff dimension of the metric space @xmath23 .
this result is due to john mitchell ( @xcite ) and is also known as _
mitchell s measure theorem_. the most important example for such a sub - riemannian manifold is the @xmath24-dimensional heisenberg group @xmath25 . it can be represented as the matrix group consisting of matrices of the form @xmath26 where the group composition is given by matrix multiplication .
note that as a point set @xmath25 is isomorphic to @xmath27 , and that @xmath25 has the structure of a graded nilpotent lie group .
the grading is induced by the lie algebra @xmath28 of @xmath25 , which is of the form @xmath29 such that @xmath30 , @xmath31 and @xmath32 = v_2 $ ] .
more generally , we will consider carnot groups .
a carnot group is a nilpotent lie group @xmath33 whose lie algebra is carrying a grading such that @xmath34 with @xmath35 = v_{s+1}$ ] for @xmath36 and @xmath37 = 0 $ ] for all @xmath38 . for our work , we will consider carnot manifolds which will be defined to be riemannian manifolds which carry such a grading structure on their tangent bundle . in this thesis
, we will use the geometric approach by dirac operators to discuss whether one can construct spectral triples on compact carnot manifolds , which furnish the hausdorff dimension of @xmath22 and the carnot - carathodory metric of @xmath39 .
we will indeed be able to construct so - called _ horizontal dirac operators _ @xmath1 acting on a horizontal clifford bundle @xmath40 over a carnot manifold @xmath0 in analogy to classical dirac operators in a quite general setting . in a frame @xmath41 of the horizontal distribution of a carnot manifold
such an operator will have the form @xmath42 where @xmath43 denotes the clifford action of @xmath19 and @xmath44 is an endomorphism on the bundle .
then indeed the carnot - carathodory metric can be detected from @xmath1 , as we will show in detail , and it does not matter if one uses the lipschitz functions with respect to the carnot - carathodory metric or the smooth functions for the dense sub - algebra of @xmath45 appearing in connes metric formula .
but we will see that these operators are not hypoelliptic , which means that they do not have a compact resolvent ( arising from pseudodifferential calculus ) .
hence they do not furnish a spectral triple , providing us the unexpected result that the theory of spectral triples does not apply to the carnot manifolds @xmath22 in the way one would expect . after introducing horizontal dirac operators and proving the metric formula
, we will construct a class of examples .
to this end we consider local homogeneous spaces of carnot groups , which arise from the left - action of a lattice sub - group @xmath46 of a carnot group @xmath33 .
we call these closed carnot manifold @xmath47 compact carnot nilmanifolds .
then the idea is to consider the submersion @xmath48 and to define a horizontal clifford module and a horizontal dirac operator @xmath1 by pulling back the spinor bundle and the dirac operator from a spin structure on @xmath49 .
for the case where @xmath50 , we will be able to calculate the spectrum of this horizontal dirac operator completely . to do this
, we adapt an argument by christian br to our situation(@xcite ) : br decomposes the spinor bundle belonging to a classical dirac operator on a heisenberg nilmanifold @xmath51 into its irreducible components under which the operator is invariant . from these irreducible components ,
he is able to calculate the spectrum of the dirac operators by means of the representation theory of the heisenberg group . in our case
, the horizontal clifford bundle decomposes in the same way .
we will present these calculations in detail , and afterwards we will deduce that our horizontal pull - back dirac operator @xmath1 on a compact carnot nilmanifold possesses at least one infinite dimensional eigenspace , hence it can not have a compact resolvent . in spite of
that we will be able to detect the hausdorff dimension of @xmath22 in the heisenberg group case from the asymptotic behavior of the non - degenerate eigenvalues of @xmath1 .
we will also extend this approach to the setting of the compact nilmanifold of a general carnot group @xmath33 , where we will show that in the spectral decomposition of such a group there is at least one infinite dimensional eigenspace .
the reason for this is that in the spectral decomposition of @xmath1 there is at least one subspace isomorphic to the space of @xmath52-sections on a horizontal clifford bundle belonging to a group of nilpotency step @xmath53 . on this subspace , we will be able to detect an infinite dimensional eigenspace of the horizontal pull - back dirac operator from the heisenberg case .
further we will see that the problem we detected in this concrete example is not due to a bad choice of the horizontal dirac operator .
in fact it is a general phenomenon : we will argue for that using techniques from pseudodifferential calculus .
there is a calculus invented by richard beals and peter greiner ( see @xcite ) on heisenberg manifolds , from which it can be derived that hypoellipticity of a ( self - adjoint ) operator of positive order implies that it has a compact resolvent and that furthermore the hausdorff dimension of the heisenberg manifold can be detected by the eigenvalue asymptotics of this operator .
we will prove in this thesis that any horizontal dirac operator on an arbitrary carnot manifold can not be hypoelliptic .
this is a big difference to the classical case , where any dirac operator is elliptic . on the way
, we will develop a criterion for the non - hypoellipticity of an arbitrary graded differential operator of the from @xmath54 where @xmath55 is a frame of @xmath56 which is forming a graded nilpotent lie algebra @xmath57 .
as soon as there is a graded differential operator @xmath58 induced by a projection @xmath59 of @xmath57 onto a lower dimensional lie algebra @xmath60 , which is not hypoelliptic , then @xmath5 can not be hypoelliptic .
this criterion will serve us well , since we will be able to reduce the problem of showing general non - hypoellipticity of a horizontal dirac operator to the heisenberg case . in the heisenberg case
we have a complete characterization of hypoellipticity of horizontal laplacians arising from the representation theory of the heisenberg group , from which we will be able to exclude that a horizontal dirac operator is hypoelliptic . finally , once having introduced the heisenberg pseudodifferential calculus , we make use of this calculus and show how hypoelliptic heisenberg pseudodifferential operators furnishing a spectral triple and detecting in addition the hausdorff dimension of the heisenberg manifold can be constructed .
we will suggest a few concrete operators , but it remains unclear whether one can detect or at least estimate the carnot - carathodory metric from them .
but we will show that the carnot - carathodory metric can be detected by horizontal laplacians instead of horizontal dirac operators via the formula @xmath61,f\right ] \right\| \leq 1 \right\}.\ ] ] hence , maybe in this case the second order operators are the right operators to look at when we want to do non - commutative geometry on carnot manifolds .
we conclude the thesis with some approaches from which one may be able to estimate or to approximate the connes metric by first order heisenberg pseudodifferential operators , and we present some criteria from which such an estimate would follow . the structure of this thesis will be the following . * in chapter 1 ,
we introduce the notion of spectral triples , their metric dimension and the connes metric , and state a few well - known examples .
then we will turn to the more general approach by mark rieffel of compact quantum order - unit spaces and prove a few criteria to detect convergence of a family of spectral triples to such a compact quantum metric space ( which does not necessarily have to be a spectral triple ) .
* in chapter 2 , we give an overview of sub - riemannian geometry and carnot manifolds .
we introduce the carnot - carathodory metric and state some important theorems in this context . the sections 2.3 and 2.4 serve to introduce the concept of a levi form and of certain submersions between carnot groups which will be of importance later .
* in chapter 3 , we construct the horizontal dirac operator .
we start by analyzing horizontal connections , which will be the connections we want a horizontal clifford bundle to be compatible with .
then we construct a self - adjoint horizontal dirac operator on these bundles . in preparation of what we need
later we will calculate its square locally and we will state a proposition about the eigenvalues of certain sums of clifford matrices . finally , in section 3.3 , we will show that any horizontal dirac operator on a carnot manifold @xmath0 detects the carnot - carathodory metric via connes metric formula , since the norm of the commutator @xmath62 $ ] coincides with the lip - norm of @xmath63 with respect to the carnot - carathodory metric .
we will also show that the metric is already detected by the smooth functions on @xmath0 . * in chapter 4
, we treat in detail the example of nilmanifolds @xmath47 from carnot groups , arising from the left action of the standard lattice .
first we construct a horizontal dirac operator on @xmath0 by pulling back the spinor bundle of the horizontal torus .
then it is our aim to show that this horizontal dirac operator possesses infinite dimensional eigenspaces .
therefore we use an approach which was presented by christian br and bernd amman ( @xcite , @xcite ) for the case of the classical dirac operator on heisenberg nilmanifolds to find a spectral decomposition of the horizontal clifford bundle @xmath40 which is invariant under @xmath1 . for the case of a general carnot group
, we find that one part of this decomposition is isomorphic to the horizontal clifford bundle belonging to a horizontal pull - back operator of a carnot nilmanifold of lower commutator step .
for the case of heisenberg nilmanifold ( where the horizontal distribution is of co - dimension 1 ) we will be able to calculate all eigenvalues of @xmath1 from the approach by br and ammann , and we will detect that @xmath1 has an infinite - dimensional kernel .
but in spite of that we will show that the hausdorff dimension of @xmath0 can still be detected by these eigenvalues for the case @xmath64 . in section 4.4
we will put all these results together to show that on any carnot nilmanifold arising from the left - action of the standard lattice sub - group of @xmath33 the horizontal dirac operator we constructed has an infinite dimensional kernel . * in chapter 5 , we introduce the heisenberg pseudodifferential calculus while referring to richard beals and peter greiner ( @xcite ) and to raphal ponge .
we will discuss symbol classes , composition of operators , parametrices and the role of hypoellipticity .
finally we mention the well - known results that hypoellipticity in the heisenberg calculus implies the existence of complex powers and certain eigenvalue asymptotics . * in chapter 6
, we show that the absence of a compact resolvent detected in chapter 4 for a specific class of examples is of general nature .
in particular we show that any horizontal dirac operator is not hypoelliptic .
therefore we first review some classical hypoellipticity theorems with special attention to the case of heisenberg manifolds .
then in section 6.2 we prove a theorem from which hypoellipticity can be excluded for certain graded differential operators by going back to a graded differential operator acting on a lower dimensional carnot group .
we will present some consequences which arise from going back to the heisenberg case via this reduction : this reduction criterion provides us with the possibility to exclude hypoellipticity of an arbitrary graded differential operator by looking at the levi form of its underlying graded lie algebra .
finally , we will apply the reduction criterion to show that an arbitrary horizontal dirac operator can not be hypoelliptic . * in chapter 7
, we discuss the possibility of constructing spectral triples by taking square roots of horizontal laplacians and the question if one can get any metric information from these spectral triples .
first we show how the application of heisenberg calculus furnishes spectral triples , and we suggest a few operators being not too far away from the horizontal dirac operator which produce spectral triples . in section 7.2 , we show that the carnot - carathodory metric can be detected by any horizontal laplacian , while the business is much more difficult if we want to detect the metric from an arbitrary first - order hypoelliptic and self - adjoint heisenberg operator .
unfortunately , all we can present concerning the last question are some criteria and some ideas which might approximate the carnot - carathodory metric by metrics arising from spectral triples , but so far we have not been able to prove such an approximation or estimate completely .
before starting with the text of the thesis , there must be some time to say thank you to all the people who supported me during the last three years .
first of all , i want to express my gratitude to my advisor elmar schrohe who was always interested in my work and found time for discussion , no matter how busy he was .
he supported me all the time and motivated me , especially during the times when i got stuck with my work .
i am deeply grateful to the members of the institute of analysis and the graduiertenkolleg 1643 for a warm and friendly atmosphere at work and many interesting discussions , not only on mathematics .
i found some wonderful friends there !
in particular i have to mention magnus goffeng in this context , a great mathematician , who supported my work with many discussions and hints .
he found the time to proofread this thesis and made some really helpful remarks .
further more i say thank you to erik christensen for hosting me in copenhagen for two months , and to christian br for accepting the task of being the second referee for this thesis .
and it is an important matter to me to send some special thanks to natascha krienen and susanne rudolph , not only for taking care of all the administrative stuff but also for being the `` good souls '' of the graduiertenkolleg and the institute .
thanks for financial supports are adressed to the dfg , since for most of the time of my phd studies i was a member of the graduiertenkolleg 1643 - analysis , geometry and string theory .
but i think i would never have finished this thesis without all the support from my friends and my parents .
finally my gratitude towards my dear wife susanne can not be expressend in words : she really had a hard time with me during the last months listening to all of my lamentation , but nonetheless she always kept on encouraging me and giving me amounts of love .
the intention of the first chapter of this thesis is to introduce the concepts of non - commutative geometry we want to consider . in section 1.1
we will present the concept of a spectral triple introduced by alain connes , which describes the geometry of a space in non - commutative geometry .
this will only be a rough introduction covering only the notion of dimension and metric arising from a spectral triple , since these are the objects we are interested in . after defining these objects
, we will give a few classical examples for commutative spectral triples to motivate the work of this thesis .
the carnot - carathodory spaces which we want to examine from the point of view of non - commutative geometry will be treated in chapter 2 .
since it will turn out that we do not get spectral triples in the sense of connes from our constructions , we will state a modified definition of a degenerate spectral triple , to denote an object being a spectral triple except for a degeneration of certain eigenspaces of its dirac operator . later in chapter 4
we will see that this definition fits in our situation .
besides that , in section 1.2 we will present the more general concept of compact quantum metric spaces , which was introduced by mark rieffel and which only provides metric information in the abstract setting of lip - norms . a lip - norm can arise from an operator which does not provide a spectral triple , as it will be the case for our situation .
we will close section 1.2 by some simple observations we made leading to a sufficient criterion for the convergence of a sequence of compact quantum metric spaces to a certain given space . in detail , we show that the metrics @xmath65 arising from a family of lip - norms @xmath66 converge to a given metric @xmath67 , if the corresponding lip - norms converge to the lip - norm @xmath68 corresponding to @xmath67 in a uniform way
. we will refer to these observations in the final chapter of this thesis , where we will be in the situation that we have a family of spectral triples providing metrics close to the metric we want to discover . but the desired metric itself will be detected by an operator which does not furnish a spectral triple
. we will give a rough overview about the theory of spectral triples now , mentioning only the concepts we will use in this thesis .
most of the definitions and examples we give are well known and can be found in any textbook on non - commutative geometry , see e.g. @xcite , @xcite or @xcite .
first of all we define what a spectral triple is .
[ def spectral triple ] let @xmath10 be a unital @xmath2-algebra and let @xmath11 be a hilbert space which carries an injective unitary representation @xmath69 .
furthermore let @xmath13 be an unbounded self - adjoint operator on @xmath11 such that a. the algebra @xmath70 \text { is densely defined and bounded } \right\}\ ] ] is a dense sub - algebra of @xmath10 .
b. for any number @xmath71 the operator @xmath72 is compact .
then the triple @xmath73 is called a _ spectral triple_.
@xmath74 * remark : * considering condition ( ii ) , it follows from the hilbert identity that the property of @xmath75 being compact for one @xmath71 implies that @xmath72 is compact for every @xmath71 .
hence one often reformulates condition ( ii ) in the sense that @xmath76 or by simply demanding that _ @xmath13 has a compact resolvent_. @xmath74 for a spectral triple we can define a notion of dimension . [ def metric dimension ] a spectral triple @xmath9 is called @xmath18-summable for some @xmath77 , if we have @xmath78 where @xmath79 denotes the ideal of trace - class operators on @xmath11 . the number @xmath80 is called the _ metric dimension _ of @xmath9 . @xmath74
* remark : * like in definition [ def spectral triple ] , one can replace the operator @xmath81 in and by any operator of the form @xmath82 for @xmath71 .
@xmath74 one can also define a metric from a spectral triple @xmath9 on the state space @xmath83 of its @xmath2-algebra @xmath10 , equipped with the weak @xmath84-topology .
it is not hard to check that the expression @xmath85
\right\| \leq 1 \right\},\ ] ] where @xmath86 is the dense sub - algebra from condition ( i ) of the definition of a spectral triple , gives a metric on @xmath83 .
[ def metric on state space ] let @xmath9 be a spectral triple , @xmath83 be the state space of its @xmath2-algebra @xmath10 .
then the metric @xmath87 on @xmath83 from is called the _ connes metric _ arising from @xmath9 .
@xmath74 it turns out that it is sufficient to consider the expression only for the positive elements of @xmath10 .
we cite this proposition here , for a proof we refer to @xcite , section 2 , lemma 1 .
[ prop metric from positive elements ] let @xmath9 be a spectral triple , and let @xmath88 denote the subset of positive elements of the @xmath2-algebra @xmath10
. then the connes metric on the state space @xmath83 is given by @xmath89 \right\| \leq 1 \right\}\ ] ] for all @xmath90 .
@xmath91 spectral triples can be defined on arbitrary @xmath2-algebras . in non - commutative geometry , the notion of a @xmath2-algebra replaces in a way the notion of a space .
this is motivated by the gelfand - naimark theory : the famous theorem of israel gelfand and mark naimark ( see @xcite ) states that every unital commutative @xmath2-algebra @xmath10 is isometrically isomorphic to the @xmath2-algebra @xmath92 of the continuous functions on some compact hausdorff space @xmath4 . in this case
the state space of @xmath10 is exactly the space @xmath4 , which turns into the expression @xmath93
\right\| \leq 1 \right\},\ ] ] defining a metric on @xmath4 . in this thesis
, we will only consider this commutative situation . we will now state the classical ( and most important ) example of a commutative spectral triple , where @xmath4 is a compact riemannian manifold .
the central point of this thesis is to construct another commutative spectral triple on this manifold @xmath4 which describes a different geometry , meaning that one gets a different dimension and a different metric from it .
[ ex canonical spectral triple on manifold ] let @xmath0 be an @xmath94-dimensional closed riemannian manifold equipped with a clifford bundle @xmath6 and a dirac operator @xmath5 .
we further consider the algebra @xmath45 of continuous functions on @xmath0 and the hilbert space @xmath95 of @xmath52-sections in the spinor bundle , on which @xmath45 has a representation by left multiplication .
then we have the following ( see e.g. @xcite , @xcite or @xcite ) : * the triple @xmath17 is a spectral triple . a dense sub - algebra of @xmath45 , which furnishes a bounded commutator with @xmath5 , is given by @xmath96 or also by the lipschitz functions @xmath97 on @xmath0 . and
since @xmath5 is an elliptic and self - adjoint differential operator of order @xmath98 with discrete spectrum , it has a compact resolvent . *
the metric dimension of this spectral triple is exactly the dimension @xmath94 of the manifold : by weyl asymptotics we have @xmath99 for the eigenvalues of @xmath5
( since it is an elliptic and self - adjoint differential operator of order @xmath98 ) , which implies that for @xmath100 the operator @xmath101 is trace class if and only if @xmath102 .
* the geodesic distance @xmath103 on @xmath0 can be detected by the formula @xmath104
\right\| \leq 1 \right\}.\ ] ] this is the case , since @xmath105 \right\|$ ] is exactly the essential supremum of the gradient of @xmath63 : one can show @xmath106 for every geodesic curve connecting @xmath107 and @xmath108 , which gives @xmath109 . on the other hand we have @xmath110 since for a fixed @xmath111 the essential supremum norm of the gradient of the function @xmath112
is bounded by @xmath98 .
@xmath74 in the commutative world , one can detect further examples for spectral triples which give a meaningful geometry by considering the hausdorff dimension of a metric space .
we briefly sketch its construction .
[ def hausdorff dimension ] let @xmath113 be a metric space , @xmath114 .
a. let @xmath115 .
for @xmath116 we set @xmath117 then the _ @xmath18-dimensional hausdorff measure _ of @xmath118 is given by the number @xmath119 b. the unique number @xmath120 such that @xmath121 for all @xmath122 and @xmath123 for all @xmath124 is called the _ hausdorff dimension _ of @xmath118 .
@xmath74 the hausdorff dimension is in particular used in fractal geometry , where one gets metric spaces of non - integer hausdorff dimension .
one classical example for a spectral triple by alain connes is a spectral triple for the cantor set ( see @xcite , chapter 3.4.@xmath125 ) , whose metric dimension coincides with the hausdorff dimension of this set . during the last years
, spectral triples were constructed for fractals , detecting the hausdorff dimension and the geodesic distance on this fractal .
we refer for example to the work of erik christensen , christina ivan and their collaborators for that ( see e.g. @xcite or @xcite ) . under mild conditions
, it is indeed possible to construct a spectral triple for an arbitrary compact metric space @xmath4 .
the idea is to approximate the space by finite sets of points .
since @xmath4 is assumed to be compact , one can find a sequence of open covers of @xmath4 whose diameter decreases to zero .
then one chooses two points in any open set belonging to one of the covers . using this sequence
, it is possible to define a dirac operator of a spectral triple on the hilbert space of @xmath126-sequences indexed by the sets belonging to the covers with values in @xmath127 .
this construction can be found in detail in the phd thesis of ian christian palmer ( see @xcite ) and has also been used by john pearson and jean bellissard ( see e.g. @xcite ) .
it is shown that there exists such a sequence such that the corresponding spectral triple gives back the hausdorff dimension of the space ( @xcite , theorem 4.2.2 ) , and that the connes metric provided by this spectral triple dominates the original metric on @xmath4 ( @xcite , proposition 5.2.1 ) . in this thesis
, we will consider compact riemannian manifolds which carry a second metric besides the geodesic one : these so - called _ carnot manifolds _ will be introduced in chapter 2 .
indeed we will be able to construct a differential operator detecting the second metric , but it will turn out that this operator does not provide a spectral triple .
but we can modify the definition of a spectral triple a little bit , such that it will fit to our situation .
[ def degenerate spectral tripel ] let @xmath10 be a unital @xmath2-algebra and let @xmath11 be a hilbert space which carries an injective unitary representation @xmath128 . furthermore let @xmath13 be an unbounded self - adjoint operator on @xmath11 such that a. the algebra @xmath129 \text { is densely defined and bounded } \right\}\ ] ] is a dense sub - algebra of @xmath10 .
b. the spectrum of the operator @xmath13 is discrete .
if in addition @xmath130 denotes the set of all eigenvalues @xmath131 of @xmath13 which have an infinite dimensional eigenspace @xmath132 , then for any number @xmath133 the operator @xmath134 is compact .
then we call the triple @xmath9 a _ degenerate spectral triple_. @xmath74 in this section we consider a more general version of a non - commutative metric space .
the idea is that for any compact metric space @xmath113 , on can associate a lipschitz semi - norm to the space @xmath92 of continuous functions on @xmath4 via @xmath135 now since a non - commutative space can be viewed as a generalization of the ( commutative ) @xmath2-algebra @xmath92 , the idea is to introduce the notion of a lip - norm on a general unital @xmath2-algebra , from which one gets a metric on the state space of @xmath10 .
this approach has been introduced by mark rieffel in @xcite and @xcite . in his approach , instead of @xmath2-algebras , rieffel works in the more abstract setting of _ order - unit spaces _ ( see @xcite ) .
[ def order - unit space ] an _ order - unit space _ is a real partially ordered vector space @xmath136 with a distinguished element @xmath137 ( the _ order unit _ ) , which satisfies a. for each @xmath138 there is an @xmath139 such that @xmath140 .
b. if @xmath138 and if @xmath140 for all @xmath139 with @xmath141 , then @xmath142 .
furthermore , the norm of an order - unit space is given by @xmath143 @xmath74 note that the self - adjoint elements of a unital @xmath2-algebra @xmath10 form an order unit space , such that the non - commutative spaces in the sense of alain connes are included in this setting .
now we introduce a semi - norm on these spaces .
[ def compact quantum metric space ] let @xmath144 be an order - unit space .
then a _ lip - norm _ on @xmath136 is a semi - norm @xmath68 on @xmath136 with the following properties .
a. we have @xmath145 @xmath146 @xmath147 .
b. the topology on the state space @xmath148 of @xmath136 from the metric @xmath149 is the @xmath150-topology .
we call a pair @xmath151 , consisting of an order - unit space @xmath136 and a lip - norm @xmath68 on @xmath136 a _ compact quantum metric space_. @xmath74 we recognize that the metric looks similar to the metric arising from a spectral triple . indeed , if @xmath152 is a spectral triple , where the dense sub - algebra of @xmath10 for which @xmath14 $ ] is bounded is denoted by @xmath86 , then the term @xmath153 \right\|\ ] ] gives a lip - norm on the order unit space @xmath136 consisting of self - adjoint elements of @xmath86 . by proposition [ prop metric from positive elements ] , the connes metric does only depend on the positive elements of @xmath86 , and hence it is in particular determined by the self - adjoint elements of @xmath86 forming the order - unit space @xmath136 .
it is shown in @xcite , section 11 , that for any essentially self - adjoint operator @xmath5 on a hilbert space @xmath11 , which carries a faithful representation @xmath154 of an order - unit space @xmath136 such that @xmath155 $ ] is bounded for a dense subspace @xmath156 of @xmath136 , furnishes a lip - norm via @xmath157\|.\ ] ] on the other hand , in @xcite , appendix 2 , it is shown that for any lower semi - continuous lip - norm @xmath68 on an order - unit space @xmath136 one can define such an operator @xmath5 describing @xmath68 via @xmath158\|$ ] on the space @xmath159 of continuous functions on the state space of @xmath136 .
note that in both cases the operator @xmath5 does not need to have a compact resolvent .
in particular , the operator @xmath5 constructed to a given semi - norm @xmath68 is in general far away from having a compact resolvent .
we will see that for the compact spaces we study in this thesis , the metric will be detected by a dirac operator which does not have a compact resolvent .
but it will fit into the setting of a compact quantum metric space . and
in addition , we will detect that we can repair the lack of not being a spectral triple by disturbing the operator a little bit .
then the question arises whether the metrics from the disturbed spectral triples are equivalent to the original metric or whether they even converge towards this metric as the perturbation goes to zero .
sadly we have not been able to give a satisfying answer to this in the later chapters of this thesis .
but nonetheless we have made some simple observations , which give criteria for the convergence of a series of compact quantum metric spaces , which we will present now .
we will see that the metric convergence of a family of compact quantum metric spaces will follow from the condition that the corresponding family of lip - norms is uniformly continuous .
our version of uniform continuity of lip - norms is characterized by the following definition : [ def uniformly continuous families of cqms ] let @xmath136 be a commutative order unit space ( which means that it can be realized as the set of continuous real - valued functions on a compact metric space ) , and for @xmath160 $ ] let @xmath66 be a family of lip - norms on @xmath136 .
we set @xmath161 then we call the family @xmath162 of compact quantum metric spaces _ uniformly continuous _ towards @xmath163 ,
if @xmath164 @xmath74 we will see now that the property of uniform continuity of a family of lip - norms implies the fact that the metrics @xmath65 corresponding to @xmath165 converge to @xmath166 , and that they are equivalent in case @xmath167 is small enough .
remember that @xmath65 is defined on the state space @xmath148 of @xmath136 , which is isometrically isomorphic to a compact metric space @xmath4 with @xmath168 . in this case
the metric is given via the formula @xmath169 then we have the following proposition .
[ prop uniformly continuity implies convergence of metrices ] let @xmath136 be a commutative order unit space and for @xmath160 $ ] let @xmath66 be a family of lip - norms on @xmath136 , such that the family @xmath162 of compact quantum metric spaces is uniformly continuous in the sense of . then for every @xmath116 with @xmath170 there is a @xmath171 such that @xmath172 for every @xmath173 .
* proof : * since @xmath136 is commutative , we consider @xmath174 for a compact metric space @xmath4 such that @xmath175 .
note that we can forget about the constant functions when calculating @xmath65 since in this case we have @xmath176 for any two points @xmath107 and @xmath108 .
now assume @xmath162 is uniformly continuous , and let @xmath116 .
then we find a @xmath177 such that @xmath178 for every @xmath179 with @xmath180 .
for an arbitrary function @xmath63 which is not constant on @xmath4 we have @xmath181 .
therefore we can set @xmath182 such that @xmath183 and the above estimate becomes @xmath184 uniformly for all @xmath185 which are not constant .
now for given points @xmath186 , leads to the estimates @xmath187 since for every constant @xmath188 one has @xmath189 this shows , and therefore the proposition is proved .
we note a similar observation , which gives equivalence of the metrics in case one can estimate the difference of two lip - norms against one of them .
[ prop equivalence condition for metricies ] let @xmath136 be a commutative order unit space and let @xmath190 , @xmath191 be two lip - norms on @xmath136 .
assume that there is a @xmath192 such that we have an estimate @xmath193 for all @xmath185 .
then the metrics @xmath166 and @xmath194 , arising from @xmath191 and @xmath190 , are equivalent .
* proof : * on the one hand , we have the estimate @xmath195 and on the other hand we have @xmath196 now the equivalence of the metrics follows by an argument analogous to the argument given in the proof of proposition [ prop uniformly continuity implies convergence of metrices ] , using @xmath197 instead of @xmath125 in .@xmath91 now one can keep on playing the game and show that the statement of proposition [ prop uniformly continuity implies convergence of metrices ] ( and therefore the uniform continuity of the family of lip - norms ) implies that the sequence of compact quantum metric spaces @xmath198 converges to the compact quantum metric space @xmath199 in the so - called _ quantum gromov - hausdorff convergence_. the concept of quantum gromov - hausdorff convergence was introduced by mark rieffel ( see @xcite ) as an quantum analogy to the gromov - hausdorff convergence of classical metric spaces . to prove this
, one can adopt the arguments given by rieffel in @xcite , section 11 , where this convergence is proved for a field of lip - norms on finite - dimensional vector spaces .
in our situation , we have to assume the uniform continuity condition from definition [ def uniformly continuous families of cqms ] , which can be shown to be fulfilled in the finite dimensional case .
but since this does not affect the questions we are considering in this thesis , we will not write down the argumentation here .
this chapter is devoted to the introduction of the spaces for which we want to construct spectral triples . on some riemannian manifolds
@xmath0 it is possible to establish another metric besides the geodesic one .
one can define the so - called carnot - carathodory distance , determined by shortest paths tangent to a certain sub - bundle of the tangent bundle . whenever this distance gives a metric on @xmath0 ( meaning that any two points of @xmath0 can be connected by such a path ) , the hausdorff dimension of this metric space differs from the usual dimension of @xmath0 . in the first section we will introduce these sub - riemannian spaces and review the properties mentioned above .
then , we will turn to the example of carnot groups , which are nilpotent lie groups carrying a certain grading inside their lie algebras , and define carnot manifolds .
these will be the central objects to construct spectral triples on in this thesis ; and they are fundamental within sub - riemannian geometry since the tangent space of riemannian geometry is generalized to a carnot group in the sub - riemannian case .
the last two sections of this chapter are meant to provide some techniques we will take advantage of later : section 2.3 deals with the case of heisenberg manifolds , which have a horizontal distribution of co - dimension @xmath98 and occur at different points in mathematics and physics , and introduces the tool of the levi form to describe the lie group structure . finally , in section 2.4
, we will construct submersions between carnot groups which will allow us to reduce certain problems for general carnot groups to lower dimensional cases . throughout this section
, @xmath200 will denote a riemannian manifold with tangent bundle @xmath20 .
since most parts of this chapter are intended to sum up things which are already known , we will refer to the basic literature , which includes in our case the books and monographs by richard montgomery @xcite , by luca capogna , donatella danielli , scott d. pauls and jeremy t. tyson @xcite and by mikhael gromov @xcite .
in addition , we refer to books by ovidiu calin and der - chen chang @xcite and by a. bonfiglioli , e. lanconelli , f. uguzzoni @xcite .
we start this chapter by giving the definition of a horizontal distribution , on which a sub - riemannian geometry is based on .
[ def horizontal distribution ] let @xmath0 be a riemannian manifold with a riemannian metric @xmath201 .
a. a _ horizontal distribution _ of the tangent bundle @xmath20 is given by a sub - bundle @xmath202 together with a fiber wise inner product @xmath203 , such that at each point @xmath204 we have @xmath205 for all @xmath206 . +
a vector field @xmath207 is called _ horizontal _ if @xmath208 .
b. a ( smooth ) path @xmath209 \rightarrow m$ ] is called _ horizontal _ if at any point @xmath210 $ ] we have @xmath211 .
c. if , for @xmath212 , @xmath213 and @xmath214 are linearly independent at each point @xmath215 , the integer @xmath216 is called the _ rank _ of @xmath19 .
in this case we call @xmath41 a _ frame _ for the horizontal distribution @xmath19 .
d. let @xmath41 be a frame for @xmath19 .
then for @xmath217 , the vector field @xmath218 is called the _ horizontal gradient _ of @xmath63 .
e. a horizontal distribution @xmath19 is called _ involutive _ if @xmath219 \in hm$ ] for any @xmath206 . f. a horizontal distribution @xmath19 is called _ bracket - generating _ if the lie hull of @xmath19 , which is the collection of all vector fields of @xmath19 and their ( multi - step ) commutators , generates the tangent bundle @xmath20 .
the smallest number @xmath220 , such that @xmath19 together with all its @xmath221-step commutators generates @xmath20 , is called the _ step _ of a bracket generating distribution .
* remark : * a horizontal distribution can also be defined by a set of @xmath98-forms : if @xmath55 is a frame of @xmath20 such that for @xmath222 @xmath41 spans @xmath19 , we denote by @xmath223 the corresponding dual frame of @xmath224 .
then we have @xmath225 @xmath74 the frobenius theorem ( see for example @xcite , theorem 1.3.1 ) asserts that @xmath19 is involutive if and only if it is integrable , which means that the set of all horizontal paths through a fixed point @xmath215 sweeps out a smooth immersed sub - manifold @xmath226 of @xmath0 with @xmath227 . in this thesis
we are interested in non - involutive , but bracket generating horizontal distributions : one can define a distance between two points @xmath186 by considering the length of the shortest horizontal part connecting these points .
[ def carnot - caratheodory distance ] let @xmath200 be a riemannian manifold which is equipped with a horizontal distribution @xmath213 . a. the _ horizontal length _ of a smooth horizontal path @xmath209 \rightarrow m$ ] is given by the number @xmath228 here , @xmath203 denotes the fiber wise inner product of @xmath19 induced by the riemannian metric of @xmath0 , see definition [ def horizontal distribution ] .
b. the _ carnot - carathodory distance _ between to points @xmath186 is given by the ( not necessarily finite ) number @xmath229 c. if @xmath230 is finite for any @xmath186 , we call @xmath231 a _ sub - riemannian manifold_. in general , this distance needs of course not to be finite , but a famous theorem by chow says that this distance is indeed finite for arbitrary @xmath186 if @xmath19 is bracket - generating ( see also @xcite , theorem 1.17 ) : [ thm chows theorem ] if @xmath19 is a bracket - generating horizontal distribution on a connected manifold @xmath0 , then any two points @xmath186 can be joined by a horizontal path .
@xmath91 in particular
this means that if the horizontal distribution @xmath19 is bracket - generating , then @xmath22 is a metric space .
this leads us to the question whether one can compare the metric spaces @xmath22 and @xmath23 , where @xmath103 denotes the usual geodesic distance .
first of all , it is obvious that for any @xmath186 we have @xmath232 , but in general the metrics are not equivalent .
but they do indeed induce the same topologies on @xmath0 , as the following theorem shows ( see @xcite , theorem 2.3 ) .
[ thm theorem on topologies ] if @xmath19 is a bracket - generating horizontal distribution on @xmath0 , then the topology on @xmath0 induced by the carnot - carathodory distance @xmath39 is the same as the usual manifold topology induced by @xmath103 . @xmath91
an unexpected observation is that the metric spaces @xmath22 and @xmath23 have a different hausdorff dimension ( see definition [ def hausdorff dimension ] for the construction of this measure theoretic dimension ) . while the hausdorff dimension of @xmath23 coincides with the ( usual ) topological dimension @xmath94 of @xmath0 , the hausdorff dimension of @xmath22 is in general strictly greater than @xmath94
: we will see that the hausdorff dimension of @xmath22 is exactly the so - called _ graded dimension _ of the horizontal distribution . to define the graded dimension of a sub - riemannian manifold ,
we need a little bit of preparation .
since the horizontal distribution @xmath19 is bracket generating of step @xmath221 , we have a sequence of vector bundles @xmath233 where @xmath234\ ] ] with @xmath235 : = { \mathrm{span}}\left\{[x , y ] : x \in hm , y \in h^sm \right\}$ ] for @xmath236 ( using the convention that @xmath237 ) .
now we assume that for a given @xmath38 the dimension of the space @xmath238 is the same for every point @xmath215 ( this will be an assumption throughout this thesis ) .
in this situation , we are able to make the following definition .
[ def graded dimension ] let @xmath0 be a sub - riemannian manifold with horizontal distribution @xmath19 of rank @xmath216 and step @xmath221 as above . for @xmath236
we denote by @xmath239 and by @xmath240 the ranks of the spaces @xmath241 .
then the _ graded dimension _ ( or _ homogeneous dimension _ ) of @xmath0 is the number @xmath242 * remark : * note that the topological dimension of @xmath0 is given by the number @xmath243 in this context .
@xmath74 now we are ready to formulate the so - called _ mitchells measure theorem _ , from which it follows that the hausdorff dimension of @xmath22 differs from its topological dimension ( see also @xcite , theorem 2.17 ) .
[ thm mitchells measure theorem ] the hausdorff dimension of a sub - riemannian manifold @xmath0 is equal to its graded dimension , i.e. under the notations of definition [ def graded dimension ] we have @xmath244 this means in particular , that in general the hausdorff dimension of the metric space @xmath22 is strictly greater than the hausdorff dimension of the metric space @xmath23 .
@xmath91 we now finish this basic section by giving the most important example for a sub - riemannian manifold : the @xmath24-dimensional heisenberg group . in dimension
@xmath245 this is the easiest example where we have one space on which two different ( with respect to metric and dimension ) geometries can be established , and it will also be an example for a carnot group , which we will introduce in the next section .
[ ex heisenberg group ] for @xmath246 , we consider the space @xmath27 equipped with a riemannian metric @xmath247 such that the vector fields @xmath248 for @xmath249 form an orthonormal frame of @xmath250 .
note that @xmath251 = x_{2m+1}$ ] , such that @xmath252 forms a non - involutive , @xmath53-step bracket generating horizontal sub - bundle of @xmath250 .
therefore @xmath253 is a metric space of hausdorff dimension @xmath254 .
since at each point , the tangent space generated by @xmath255 has the structure of a @xmath53-step nilpotent lie algebra @xmath28 the exponential mapping furnishes a simply - connected nilpotent lie group @xmath256 .
@xmath25 is called the _
@xmath257-dimensional heisenberg group_. it is a
well known fact that @xmath25 can be realized as the space of upper triangle @xmath258-matrices which have the form @xmath259 where the group rule is given by matrix multiplication .
on @xmath27 , this multiplication can be written as @xmath260 the lie algebra @xmath28 of @xmath25 can be realized by strictly upper triangle matrices @xmath261 where the algebra multiplication is once again the matrix multiplication .
@xmath74 * remark : * the composition rule of the heisenberg group is written down in the so - called _ polarized coordinates_. these are exactly the coordinates derived from the matrix model of this lie group .
on the other hand , one can describe points of @xmath25 by their _ exponential coordinates _ , which are the coordinates of the corresponding lie algebra having a one - to - one correspondence with the coordinates of @xmath25 by the exponential mapping .
we will introduce both kinds of coordinates in the next section in the context of carnot groups .
@xmath74 the heisenberg group @xmath25 introduced in example [ ex heisenberg group ] is an example for a greater class of nilpotent lie groups , which are called carnot groups .
[ def carnot group ] a _ carnot group _ of step @xmath220 is a simply connected lie group @xmath33 whose lie algebra @xmath57 has a stratification ( or grading ) @xmath262 such that @xmath263 are vector spaces satisfying the conditions a. @xmath264 = v_{s+1}$ ] for @xmath265 b. @xmath37 = 0 $ ] for @xmath266 .
we call the number @xmath267 the _ bracket - generating dimension _ of @xmath33 and the number @xmath268 the _ bracket - generating co - dimension _ of @xmath33 .
@xmath74 carnot groups are the canonical generalization of the euclidean space @xmath56 in sub - riemannian geometry , since the tangent space ( or better the tangent cone ) of a carnot manifold has the structure of a carnot group .
this can be seen as a generalization of the riemannian case , where the tangent space at any point is isomorphic to the ( @xmath98-step nilpotent ) carnot group @xmath56 . without going into detail ,
the situation is as follows : let @xmath4 be any metric space , @xmath269 . then we define the _ tangent cone _ as the pointed gromov - hausdorff limit of the family @xmath270 of pointed metric spaces for @xmath271 , if it exists .
it has been proved by john mitchell @xcite that in the case where @xmath22 is a carnot manifold , this limit exists and has the structure of a carnot group ( see also @xcite , theorem 8.8 ) : [ thm mitchells theorem on tangent cone ] let @xmath107 be a regular point of a carnot manifold @xmath22 .
then , at every @xmath215 , the tangent cone exists and is a carnot group , which is arising from the nilpotentization of the horizontal distribution at @xmath107 . @xmath91
* remark : * we will not explain the details of the process of nilpotentization here and refer to @xcite or @xcite instead , since the technical details will not affect this thesis .
@xmath74 now getting back to the objects we are considering , we finally define a carnot manifold as a sub - riemannian manifold , which has globally a carnot group structure on its tangent bundle .
[ def carnot manifold ] let @xmath200 be an @xmath94-dimensional riemannian manifold , whose tangent bundle @xmath20 carries a grading @xmath272 such that @xmath273 are vector bundles satisfying the conditions a. @xmath274 = v_{s+1}m$ ] for @xmath265 b. @xmath275 = 0 $ ] for @xmath266 .
then we call @xmath0 a _ carnot manifold _ of step @xmath220 .
we further call the number @xmath276 the _ bracket - generating dimension _ of @xmath0 and the number @xmath277 the _ bracket - generating co - dimension _ of @xmath0 . @xmath74 * remark : * it is clear that every carnot group is a carnot manifold .
it is also clear by the conditions ( i ) and ( ii ) on the brackets of the vector fields that every carnot manifold contains the structure of a sub - riemannian manifold , where the horizontal distribution is given by @xmath278 .
it should be mentioned that every carnot manifold has locally the structure of a carnot group , while for a general sub - riemannian manifold this is a bit more involved ( see theorem [ thm mitchells theorem on tangent cone ] ) .
note that we have not fixed the geometry we are considering in the above definition .
so by section 2.1 we can establish at least two different geometries on a carnot manifold @xmath0 : the riemannian one ( equipped with the geodesic distance @xmath103 ) and the sub - riemannian one ( equipped with the carnot - carathodory distance @xmath39 ) .
@xmath74 when working on a carnot manifold , we will always assume that we have a riemannian metric @xmath247 on @xmath0 such that the spaces @xmath279 appearing in the grading are mutually orthogonal .
for @xmath280 we set @xmath281 and denote by @xmath282 an orthonormal frame of @xmath279 .
hence , a frame of @xmath20 is given by @xmath283 we will sometimes use the abbreviation @xmath284 to denote the frame of @xmath279 . from the grading of its lie algebra @xmath57 , one can introduce coordinates on a carnot group @xmath33 . thereby we make use of the fact that because of the nilpotency the exponential mapping @xmath285 is a diffeomorphism .
we present two different types of coordinates here , where it will be depending on the situation which type is more comfortable to work with . [ def coordinates of carnot groups ] let @xmath33 be a carnot group with grading @xmath34 of its lie algebra , such that for @xmath38 , @xmath286 denotes a basis of @xmath287 .
a. the coordinates @xmath288 given by @xmath289 are called _ exponential coordinates _ or _
canonical coordinates of the first kind _ of @xmath33 .
b. the coordinates @xmath290 given by @xmath291 are called _ polarized coordinates _ or _
canonical coordinates of the second kind _ of @xmath33 .
thereby , @xmath292 denotes the group composition on @xmath33 , and the products are taken with respect to this group composition . @xmath74 * remark : * it is known that there is an isomorphism between the exponential and the polarized coordinates of a carnot group .
for example , in the case of the @xmath24-dimensional heisenberg group @xmath25 this isomorphism is given via @xmath293 where @xmath294 denote the exponential coordinates ( see @xcite , section 1.2 ) .
@xmath74 when calculating a group composition on @xmath33 , one uses the baker - campbell - hausdorff formula .
this formula is given on @xmath57 by @xmath295 where @xmath296 is a sum of multi - step commutators of order @xmath297 , see @xcite .
therefore , @xmath296 is a polynomial of degree smaller or equal to the step of @xmath33 , which does not depend on vectors belonging to @xmath298 since @xmath298 commutes with every @xmath299 . in the case of a @xmath53-step carnot group
, we have @xmath300\right).\ ] ] using this expression on the lie algebra , one can derive the composition rule in exponential or polarized coordinates on the carnot group @xmath33 .
maybe the most important property of a carnot group is its homogeneity , which is expressed by a weighted dilation of it .
once again there is a one - to - one relation between the dilation on a carnot group and the corresponding weighted dilation on its graded lie algebra , given by the exponential map .
[ def delation on carnot group ] a.
let @xmath33 be a carnot group with coordinates @xmath301 ( exponential or polarized ) , @xmath302 .
then we define by @xmath303 the _ ( weighted ) dilation _ on @xmath33 by @xmath131 . +
a function @xmath304 is called _ homogeneous _ of degree @xmath305 with respect to @xmath306 if we have @xmath307 for all @xmath308 , @xmath302 .
b. let @xmath57 be a graded nilpotent lie algebra with grading @xmath309 , @xmath302 .
then we define by @xmath310 the _ ( weighted ) dilation _ on @xmath57 by @xmath131 . +
a function @xmath311 is called _ homogeneous _ of degree @xmath305 with respect to @xmath312 of degree @xmath313 if we have @xmath314 for all @xmath299 , @xmath302 . c. let @xmath0 be a carnot manifold , @xmath306 and @xmath312 the weighted dilations by @xmath302 defined point - wise on @xmath20 .
a vector field @xmath315 is called _ homogeneous _ of degree @xmath305 if we have @xmath316 for every @xmath302 .
* remark : * note that if the grading of @xmath20 is given by @xmath317 , we have that @xmath207 is homogeneous of degree @xmath318 if and only if @xmath319 .
@xmath74 we remark that the weight @xmath318 of @xmath306 ( or @xmath312 ) belonging to certain coordinates reflects exactly the commutator step of the space @xmath287 of @xmath57 which belongs to @xmath320 ( or @xmath286 ) .
these weighted dilations play an important role when one wants to describe a symbol calculus with respect to a carnot group structure : asymptotic expansions are given in terms of homogeneous functions with respect to weighted dilations .
furthermore one can use weighted dilations to define a carnot group , see for example @xcite , definition 2.2.1 , where a ( homogeneous ) carnot group is defined to be a lie group structure on @xmath56 which respects these dilations and has a bracket generating structure .
in @xcite it is also shown that these two definitions are equivalent ( up to isomorphisms ) .
let us mention now how every carnot group can be realized as a certain lie group structure on @xmath56 , using certain vector fields to represent the basis of its lie algebra .
we have already seen this in example [ ex heisenberg group ] for the case of the heisenberg group .
the idea is to find a frame of vector fields @xmath321 which satisfies the grading conditions of definition [ def carnot group ] and which has the property that @xmath322 for each @xmath323 .
this is indeed possible , and the requested vector fields @xmath324 have polynomial coefficients , as the following proposition shows ( see @xcite , remark 1.4.6 ) .
[ prop carnot lie algebra as vector fields ] let @xmath33 be a carnot group with lie algebra @xmath57 , where the grading of @xmath57 is given by @xmath325 with @xmath326 .
then for each @xmath287 there exists a frame of vector fields @xmath327 such that @xmath328 here , @xmath329 is a polynomial which is homogeneous with respect to the dilations @xmath306 from of degree @xmath330 , which means @xmath331 @xmath91 now one can consider the homogeneous vector fields from as homogeneous differential operators .
thus any polynomial of these vector fields is a differential operator .
these so - called _ graded differential operators _ will play a big role in this thesis : the horizontal dirac operators we will construct in chapter @xmath245 fall into this category , and the horizontal laplacians of ( homogeneous ) degree @xmath53 will also turn out to be very important . [ def graded differential operator ] let @xmath309 be a graded lie algebra with @xmath332 , which is represented by vector fields on @xmath333 as in proposition [ prop carnot lie algebra as vector fields ] . a frame for @xmath287 shall consist of the vector fields @xmath334 , where we write @xmath41 for the frame of @xmath335 ( with @xmath336 ) . then
a _ graded differential operator _ is a differential operator of the from @xmath337 where @xmath338 is a polynomial with matrix - valued @xmath339 coefficients . if @xmath338 is homogeneous of degree @xmath305 we call @xmath68 _ homogeneous _ of degree @xmath313 .
@xmath74 in particular , a _
horizontal laplacian _ is a graded differential operator of order @xmath53 , which means it is an operator of the form @xmath340 where all the @xmath341 and @xmath342 are smooth ( matrix valued ) functions on @xmath343 . @xmath74 the above definition suggests to consider a graded differential operator @xmath68 as an element of the universal enveloping algebra @xmath344 of the graded lie algebra @xmath57 generated by @xmath41 .
this interpretation will be used in chapter @xmath345 when we analyze graded differential operators using the representation theory of their underlying lie algebras @xmath57 .
we return once again to the dilations from definition [ def delation on carnot group ] : for our purposes we should mention an additional another property of @xmath306 : for arbitrary points @xmath346 we have @xmath347 so the carnot - carathodory metric on @xmath33 respects these dilations .
now we will introduce a quasi - norm ( or gauge norm ) @xmath348 on a carnot group @xmath33 , which is homogeneous with respect to these dilations and provides us with a quasi - metric on @xmath33 which is equivalent to @xmath39 .
it will rather be a quasi - norm than a norm since the triangle inequality on @xmath33 must be replaced by the condition @xmath349 this quasi - norm will play an important role when we define a symbol calculus according to carnot groups .
[ def koranyi gauge ] let @xmath33 be a carnot group of step @xmath221 .
the quasi - norm @xmath350 on @xmath33 , defined by @xmath351 is called the _ koranyi gauge _ of @xmath33 .
the quasi - metric @xmath352 arising from the koranyi gauge will be called the _ koranyi ( quasi-)metric _ on @xmath33 .
@xmath74 * remark : * note that in the case @xmath353 the koranyi gauge is given by the formula @xmath354 we also note that there are several definitions of the koranyi gauge on a carnot group which are equivalent , for example in @xcite the koranyi gauge on @xmath25 is defined via @xmath355 where @xmath356 denotes the euclidean norm of the corresponding vector .
@xmath74 as noted in @xcite , the koranyi quasi - metric on a carnot group @xmath33 is equivalent to the carnot - carathodory metric on @xmath33 . [
prop equivalence koranyi metric carnot - caratheodory metric ] let @xmath33 be a carnot group , and let @xmath357 denote the koranyi quasi - metric and @xmath39 the carnot - carathodory metric on @xmath33
. then there are constants @xmath358 and @xmath359 such that for all @xmath346 we have @xmath360 @xmath91 now having finished a rough review about the theory on carnot groups and carnot manifolds , we want to finish this section by giving some simple examples for compact carnot manifolds .
we consider the local homogeneous space of a discrete lattice subgroup of a carnot group @xmath33 . in a way , this is the analogous object to the torus @xmath361 , which arises as the quotient by @xmath362 on @xmath56 .
since it is quite comfortable to do calculations on these objects , they will serve as the main example for the considerations of this thesis .
[ ex compact carnot nilmanifolds ] let @xmath33 be a carnot group , equipped with a ( left - invariant ) riemannian metric , with grading @xmath34 , where @xmath41 is a frame for @xmath335 .
we consider the discrete subgroup @xmath46 of @xmath33 generated by the basis vectors of @xmath335 , that is @xmath363 @xmath46 acts on @xmath33 via the left multiplication @xmath364 for all @xmath308 .
it follows easily that @xmath46 is a lattice in @xmath33 ( see @xcite , section 9.3 ) , which we will call the _ standard lattice _ of @xmath33 .
since @xmath46 is a lattice , the local homogeneous space @xmath365 , consisting of the orbits of this group action , is a compact riemannian manifold which is locally isometric to @xmath33 .
hence @xmath366 is a compact carnot manifold , where the grading of the tangent bundle @xmath20 comes from the grading of @xmath57 .
we will call @xmath366 the ( standard ) _ compact carnot nilmanifold _ of @xmath33.@xmath74 the group @xmath46 defined via can also be viewed as a discrete nilpotent group . from the grading of @xmath33 one
can detect a so - called _ central descending series _ @xmath367 which means we have @xmath368 $ ] , where @xmath369 is the standard lattice of the carnot group @xmath370 .
one can check that for the graded dimension ( and therefore by theorem [ thm mitchells measure theorem ] also for the hausdorff dimension ) of @xmath33 we have the identity @xmath371 now there is a famous theorem by bass , milnor and wolf ( see @xcite , theorem 9.3 ) which states that the right hand side of is equal to the polynomial growth of @xmath46 .
so , altogether , the hausdorff dimension of @xmath33 coincides with the polynomial growth of its standard lattice subgroup .
see @xcite , sections 9.2 and 9.3 , for details on this .
@xmath74 we will now pay attention to the most important class of carnot manifolds : it is the case where the graded co - dimension is equal to @xmath98 .
those manifolds are also known as heisenberg manifolds . in the context of non - commutative geometry
, they were treated in details in the work of raphal ponge ( see for example @xcite and the references there ) , and recently they have also been a big research area in index theory , for which we refer to erik van erp and his collaborators ( see e.g. @xcite ) . for most of the following definitions and
proposition , we refer to ponge ( @xcite ) .
[ def heisenberg manifold ] a heisenberg manifold is a smooth riemannian manifold @xmath200 of dimension @xmath372 equipped with a hyperplane bundle @xmath202 of rank @xmath216 , which is bracket - generating .
* remark : * one can formulate the definition of a heisenberg manifold a little bit more general by dropping the assumption that @xmath19 has to be bracket generating , see e.g. @xcite . in this case , objects like contact manifolds or cr manifolds are included in the class of heisenberg manifolds .
but since we are following more or less theoretical aspects and want to consider manifolds on which we have the carnot - carathodory geometry , we make this further assumption .
@xmath74 an important tool to handle heisenberg manifolds is the so called _
levi form_. in case @xmath0 is a carnot manifold of graded co - dimension @xmath98 , this is the @xmath98-form describing the lie algebra structure of the ( graded ) tangent bundle .
[ defin levi form ] let @xmath0 be a carnot manifold of co - dimension @xmath98 , where the grading of its tangent bundle is given by @xmath373 , with @xmath374 and @xmath375 $ ] .
let @xmath41 be a frame of @xmath376 and let @xmath377 be a frame of @xmath378 . then the _ levi form _
@xmath379 of @xmath0 is the ( antisymmetric ) bilinear form @xmath380 \mod v_1m.\ ] ] for @xmath381 with @xmath382 for @xmath383 , we denote by @xmath384 the antisymmetric @xmath385-matrix describing @xmath379 according to the basis @xmath41 .
now one can define a levi form on any heisenberg manifold @xmath0 the same way : on can see rather easily that from the lie bracket of vector fields on @xmath19 one gets a @xmath53-form @xmath386 such that for any section @xmath206 we have @xmath387(x_0 ) \mod h_{x_0}m\ ] ] near a point @xmath204 ( see @xcite ) .
but this allows us to define a bundle of graded @xmath53-step nilpotent lie algebras @xmath388 , which gives rise to a bundle of carnot groups @xmath389 of step @xmath53 and graded co - dimension @xmath98 over @xmath0 via the exponential mapping .
this bundle @xmath389 is also called the _ tangent lie group bundle _ of @xmath0 .
we can even say how this tangent lie group bundle looks like : [ prop tangent lie group bundle and rank of levi form ] let @xmath200 be a heisenberg manifold .
then for any point @xmath204 , the point - wise levi form @xmath390 has rank @xmath391 if and only if @xmath392 . in
especially , this means @xmath379 has constant rank @xmath391 if and only if @xmath389 is a fiber bundle with typical fiber @xmath393 .
in addition , if @xmath394 , it is always possible to find an orthonormal basis @xmath41 of @xmath376 such that the matrix representation @xmath395 of @xmath379 becomes @xmath396 where @xmath397 is a diagonal matrix carrying the absolute values @xmath398 of the non - zero eigenvalues on its diagonal .
* proof : * the first statement is exactly the statement of @xcite , proposition 2.1.6 .
the second statement follows by linear algebra , since @xmath68 is a skew - symmetric matrix , so it is known to have the non - zero eigenvalues @xmath399
. then for any orthonormal frame of @xmath19 the form can be reached by an orthonormal basis transformation at every point of @xmath19 .
@xmath91 * remark : * by our assumption of @xmath19 being bracket - generating , we always have @xmath400 and therefore @xmath401 .
of course the case @xmath402 also fits into the theory : in this case the hyperplane bundle @xmath19 induces a foliation on @xmath0 .
@xmath74 we can further introduce levi forms on arbitrary carnot manifolds : since @xmath403 = v_2m$ ] for the first two summands appearing in the grading of the tangent space of any carnot manifold , the following definition is well - defined .
[ def generalized levi form ] let @xmath0 be a carnot manifold with grading @xmath404 of its tangent bundle .
we denote by @xmath405 an orthonormal frame of @xmath376 and by @xmath406 an orthonormal frame of @xmath378 .
then for @xmath407 , the _ @xmath408-levi form _ of @xmath0 is given by the bilinear form @xmath409 \mod \left ( { \mathrm{span}}\{x_{2,\nu}\ } \right)^\perp.\ ] ] for @xmath410 with @xmath411 , we denote by @xmath412 the antisymmetric matrix describing @xmath413 .
@xmath74 in other words , the collection of @xmath408-levi forms @xmath413 describes the structure of the first step commutators of a carnot manifold . we will use this notation later when we look for structures of heisenberg manifolds inside a carnot manifold . in the final section of this chapter
we will introduce a technique which will play an important role in our later considerations : we will show how one gets a submersion from a given carnot group @xmath414 onto a lower dimensional carnot group @xmath415 .
this will provide us the possibility to pull back objects defined on @xmath415 to objects on @xmath414 , such that statements about @xmath415 can be transported to @xmath414 .
once again we consider a carnot group @xmath33 whose lie algebra has the grading @xmath325 .
let @xmath416 be a linear subspace of ( the vector space ) @xmath57 which has the structure @xmath417 for some @xmath418 .
we take a look at the canonical orthogonal projection @xmath419 we show now that @xmath420 induces a homomorphism of lie algebras , which gives rise to a submersion ( and also to a homomorphism ) of lie groups . [ prop properties projection lie algebra ] let @xmath33 , @xmath57 , @xmath421 and @xmath422 be as above .
then the vector space @xmath423 can be equipped with a lie algebra structure via @xmath424_\mathrm{pr } : = \mathrm{pr } \left([x , y]\right ) \ \ \ \forall x , y \in \tilde{v}\ ] ] such that @xmath425 ) \rightarrow ( \mathfrak{\tilde{g } } , [ \cdot,\cdot]_\mathrm{pr})$ ] is an homomorphism of graded lie algebras .
further , if we set @xmath426 , the spaces @xmath427 and @xmath428 are carnot groups such that @xmath429 and hence @xmath430 .
the resulting map @xmath431 is a homomorphism of carnot groups and a submersion of riemannian manifolds .
* proof : * first of all , one checks that the lie brackets defined via are indeed lie brackets : the bilinearity of @xmath432_\mathrm{pr}$ ] is obvious because of the bilinearity of @xmath432 $ ] and the linearity of @xmath420 ; and the jacobi identity also follows from these properties in connection with the jacobi identity of @xmath432 $ ] . to check that @xmath420 is a homomorphism of lie algebras
, one has to check that @xmath433_\mathrm{pr } = \mathrm{pr } \left([x , y]\right ) \ \ \ \forall \ x , y \in \mathfrak{g}.\ ] ] this is by definition true for all @xmath434 .
now let ( without loss of generality ) @xmath435 .
in this case , @xmath436 and therefore the left hand side of is zero .
but the right hand side is also zero , since @xmath437 and therefore @xmath219 \in \bigoplus_{s = m+1}^r v_s$ ] because of the graded structure of @xmath57 .
therefore is true , which shows that the linear map @xmath420 is a lie algebra homomorphism . but
this means that @xmath438 is an ideal in @xmath57 and therefore also a graded lie algebra .
the grading structures of @xmath439 follows immediately from the grading structure of @xmath57 since @xmath440 thus @xmath60 is obviously a graded lie algebra of step @xmath0 .
now we consider the map @xmath441 from . because of the nilpotency the exponential maps from the lie algebras @xmath60 and @xmath57 to their lie groups @xmath442 and @xmath33 are diffeomorphisms , and since @xmath420 is a surjective linear map ( which is smooth because of its linearity ) we have the result that @xmath441 is a submersion . to check that it is a group homomorphism , we calculate , using the baker - campbell - hausdorff formula : @xmath443 in this calculation we have used the fact that @xmath422 is a lie algebra homomorphism . in particular , the third equation is true since @xmath444 is a sum of ( multi - step )
commutators of the vector fields @xmath445 , such that applying @xmath420 to @xmath444 furnishes the polynomial @xmath446 in the baker - campbell - hausdorff formula on @xmath442 .
finally we have @xmath447 such that @xmath226 is a normal lie subgroup of @xmath33 . but
this shows in addition that @xmath448 and that @xmath430 .
now every statement of the proposition is proved .
we now show briefly that such a submersion can be lifted to the compact nilmanifolds arising from @xmath33 from example [ ex compact carnot nilmanifolds ] .
let @xmath46 be generated by the images of the basis vector fields of @xmath335 under the exponential mapping , and let @xmath47 be the local homogeneous space of the left action of @xmath46 on @xmath33 .
since the submersion @xmath441 from is a lie group homomorphism , the image @xmath449 under @xmath441 is a discrete subgroup of @xmath450 , and by the definition of @xmath441 it is clear that @xmath449 is generated by the image of the basis vectors of @xmath335 under the corresponding projection @xmath451 of the lie algebras .
therefore , we get a compact carnot nilmanifold @xmath452 . for the action of @xmath449 on @xmath450 we have @xmath453 for any @xmath454 and @xmath308 , therefore any orbit of the action of @xmath46 on @xmath33 is mapped to a orbit of the action of @xmath449 on @xmath450 .
let @xmath455_\gamma$ ] denote the orbit belonging to an element @xmath308 under the action of @xmath46 . because @xmath441 is a submersion by proposition [ prop properties projection lie algebra ]
, this means that the mapping @xmath456_\gamma \mapsto [ \psi(x)]_{\psi(\gamma)}\ ] ] is also a submersion .
we summarize the above argumentations in the following corollary : [ cor projection carnot nilmanifolds ] in the above situation , the map @xmath154 from is a submersion of riemannian manifolds .
locally , @xmath154 coincides with the submersion @xmath457 from of the corresponding carnot groups .
* proof : * the fact that @xmath154 is a submersion has already been deduced in the discussion ahead of this corollary . since the nilmanifolds @xmath365 and @xmath458 are locally isometric to the carnot groups @xmath33 and @xmath442 ( see example [ ex compact carnot nilmanifolds ] ) , the second statement is obvious by the construction of @xmath154 from @xmath441 .
the intention of this chapter is to find a first order differential operator which detects the carnot - carathodory metric via connes formula .
therefore we follow an approach analogous to the standard example for a spectral triple on a riemannian manifold : we construct a so - called _ horizontal dirac operator _ , arising from the clifford action of a horizontal distribution of a carnot manifold . to really cover the horizontal geometry , we want our horizontal dirac operator to be compatible with the horizontal part of the levi - civita connection on a carnot manifold @xmath0 .
we thus define and discuss a ( partial ) connection @xmath459 according to the sub - bundle @xmath19 of @xmath20 in section 7.1 from the levi - civita connection on @xmath0 .
then in section 7.2 , we introduce horizontal clifford bundles arising from the clifford action of the bundle @xmath19 , which carry a connection compatible with @xmath460 . on these bundles
, we will be able to define horizontal dirac operators @xmath1 in a general sense ( in analogy to classical dirac operators on clifford bundles ) , and we will be able to modify these operators such that they are self - adjoint . in the end of section 7.2
, we calculate the square of @xmath1 and proof a technical proposition concerning the eigenvalues of a sum of certain clifford matrices in preparation for future arguments . finally in section 3.3 , we will show that the horizontal dirac operators we constructed are indeed the right operators to detect the carnot - carathodory metric on a carnot manifold @xmath0 : we show that the norm of @xmath62 $ ] coincides with the lipschitz norm of @xmath63 with respect to the carnot - carathodory metric , such that we can apply connes metric formula to the sub - algebra of @xmath45 consisting of these lipschitz functions .
in addition we show that this metric is already detected by the sub - algebra of @xmath339-functions . in the following chapters , we will seen that in spite of all these good properties this operator does not define a spectral triple . throughout this section , @xmath0 will be a compact carnot of step @xmath221 with carnot - carathodory metric @xmath39 of dimension @xmath94 .
we will use the notation @xmath461 for a frame of the horizontal distribution @xmath19 of @xmath0 , and if not stated otherwise the riemannian metric @xmath247 on @xmath0 will be chosen such that this frame is orthonormal .
the horizontal dirac operator will be defined on closed ( meaning compact without boundary ) carnot manifolds . to define a horizontal dirac operator using the clifford module arising from the horizontal distribution @xmath19 we need to introduce a connection on @xmath19 .
this happens straight forward and can be found at various parts of the literature ( see e.g. @xcite , but a horizontal connection is also mentioned in the textbooks @xcite and @xcite ) .
the idea is simply to start with the levi - civita connection on @xmath0 and to project it onto the horizontal bundle .
this approach is justified by the following proposition ( see @xcite , proposition 4.2 ; not that the case of carnot manifolds follows directly from the case of carnot groups formulated there ) .
[ prop extension of horizontal metric ] let @xmath0 be a carnot manifold with horizontal distribution @xmath202 which carries a smoothly varying inner product @xmath203 .
let @xmath462 be a sub - bundle which is complementary to @xmath19 .
if @xmath463 and @xmath464 are riemannian metrics which make @xmath465 and @xmath19 orthogonal such that for @xmath466 @xmath467 then the associated levi - civita connections @xmath468 and @xmath469 coincide when projected to @xmath19 : we have @xmath470 for all sections @xmath471 of @xmath19 .
@xmath91 this proposition allows us to define a connection only depending on the horizontal bundle , and it shows that it is in fact well defined when constructing it from any levi - civita connection of the extension of the horizontal inner product .
we follow @xcite , where this is done for the heisenberg group .
[ def horizontal connection ] let @xmath0 be a carnot manifold with horizontal distribution @xmath202 , which carries a smoothly varying inner product @xmath203 .
let @xmath472 denote the projection of a tangent vector field onto its horizontal component .
we then define @xmath473 where @xmath474 denotes the levi - civita connection of any extension of the horizontal inner product .
@xmath460 is called the _ horizontal ( levi - civita-)connection _ of @xmath0 .
@xmath74 [ prop properties horizontal connection ] the map @xmath460 defined in definition [ def horizontal connection ] is indeed a ( partial ) connection defined on the horizontal bundle , which means we have a. for all @xmath475 and for all @xmath476 : @xmath477 b. for all @xmath478 and for all @xmath476 : @xmath479 @xmath460 is _ metric _ with respect to the point - wise inner product @xmath203 of @xmath19 , which means that we have a. for all @xmath478 : @xmath480 @xmath460 is _ torsion free _ in the horizontal direction , which means that we have a. for all @xmath481 : @xmath482).\ ] ] * proof : * let @xmath474 be the levi - civita connection corresponding to any riemannian metric @xmath247 on @xmath20 which is an extension of @xmath203 . the conditions ( i ) and ( ii ) which show that @xmath460 is indeed a connection follow immediately from the corresponding properties of the levi - civita connection after projection onto the horizontal distribution .
the metric property ( iii ) follows immediately from the metric property of @xmath474 since @xmath203 is just a restriction of the riemannian metric @xmath247 .
finally , the torsion freeness into the horizontal direction follows because @xmath483\right)\end{aligned}\ ] ] for all @xmath484 , since @xmath485 because @xmath486 is a projection and since the levi - civita connection @xmath474 is torsion free .
hence the proposition is proved .
* remark : * we note that we can not expect to have torsion freeness in the sense that @xmath487 $ ] for all @xmath484 , as it is the case for the levi - civita connection : since @xmath19 is not involutive , we have @xmath219 \notin hm$ ] for some @xmath206 , but by definition of the horizontal connection the vector field @xmath488 must be horizontal again .
@xmath74 we have only defined our horizontal connection to be a partial connection , which means we only allow differentiation into horizontal directions .
there are several possibilities to extend @xmath460 to a connection @xmath489 for instance , the expression makes sense for any @xmath490 , and proposition [ prop properties horizontal connection ] would also hold for this case with the same proof .
another possibility of defining a connection is given by setting @xmath491 for all @xmath492 .
@xmath74 we can not get rid of the vertical directions completely as soon as we want to apply horizontal covariant derivatives more than one time : by the tensorial property ( ii ) of proposition [ prop properties horizontal connection ] we have for any @xmath493 and for any @xmath494 : @xmath495f \right ) \cdot y,\end{aligned}\ ] ] and since @xmath19 is not involutive we have @xmath496 \notin \gamma^\infty(hm)$ ] .
this has consequences when one wants to define some kind of curvature belonging to the horizontal connection : it turns out that this curvature shows into the transversal direction of @xmath20 .
but we will not discuss this aspect any further and refer to @xmath497 and @xmath498 instead .
@xmath74 we close this short section with a proposition which shows us how the calculate a horizontal covariant derivative in local coordinates respecting the grading structure of our carnot manifold @xmath0 .
this is more or less trivial , since after calculating the covariant derivative with respect to the levi - civita connection we simply project onto the horizontal distribution .
[ prop horizontal connection in local coordinates ] let @xmath0 be a carnot manifold with a horizontal distribution @xmath19 of rank @xmath216 .
let @xmath55 denote an orthonormal frame for @xmath20 such that for @xmath499 @xmath41 is an orthonormal frame for @xmath19 .
then we have locally @xmath500 where the @xmath501 are the christoffel symbols of the levi - civita connection @xmath474 of @xmath20 with respect to the frame @xmath55 .
@xmath91 we want to construct a horizontal dirac operator in analogy to the ( classical ) dirac operator . in the classical case
this construction is outlined for example in @xcite or @xcite .
we start with a review about the definition of a clifford algebra and clifford action .
remember that for each vector space @xmath502 equipped with a symmetric bilinear form @xmath503 there exists an ( up to isomorphisms ) unique _ clifford algebra _
@xmath504 which is a unital algebra , equipped with a map @xmath505 such that @xmath506 in addition , @xmath507 is supposed to fulfill the universal property in the sense that for any other unital algebra @xmath156 equipped with map @xmath508 satisfying , there is a unique algebra homomorphism @xmath509 such that @xmath510 ( see @xcite , definition 3.1 and proposition 3.2 ) .
multiplication in this unique algebra is determined by the rule @xmath511 and we know that for @xmath512 we have @xmath513 . using this map @xmath507 to define multiplication , we can construct a left module over the complex algebra @xmath514 which will be called a _
clifford module_. in other words , a _ clifford module _ @xmath318 for a real inner product space
@xmath502 is a complex vector space @xmath318 equipped with an @xmath515-linear map @xmath516 such that @xmath517 for all @xmath518 .
the map @xmath519 is called the _
action_. now the question arises what the minimal possible dimension of such a clifford module @xmath318 is , depending on the dimension of @xmath502 ; in other words , we are looking for irreducible representations of @xmath136 .
such an irreducible representation is given by the so - called _ spin representation _ @xmath520 , and it is known that for @xmath521 we have @xmath522},\ ] ] where @xmath523 $ ] denotes the gaussian bracket which gives the greatest integer smaller or equal then @xmath524 ( see @xcite , chapter 4 , or @xcite ) .
the idea for our situation is now to take a look at the clifford algebras generated by the fibers of the horizontal distribution @xmath213 .
since the riemannian metric on @xmath0 is chosen in a way that @xmath41 forms an orthonormal frame at each point , we immediately get the fundamental properties of this horizontal clifford action , which we will denote by @xmath43 , from the above discussion .
[ prop clifford module over hm ] for @xmath215 , let @xmath525 be a clifford module for @xmath526 with ( horizontal ) clifford action @xmath527 .
then we have : a. @xmath528 for all @xmath529 .
b. @xmath530 for all @xmath531 , @xmath532 .
@xmath91 now let @xmath533 be a bundle of clifford modules for the horizontal distribution @xmath19 .
we need to equip @xmath533 with an point - wise ( hermitian ) inner product and with a connection , for which we claim certain compatibility conditions . for a classical dirac operator ,
we ask the connection on @xmath534 to be compatible with the levi - civita connection @xmath474 on @xmath20 , see e.g. @xcite , definition 3.4 . hence to define a suitable bundle and connection where a horizontal dirac operator can act on
, we would like to have compatibility with the horizontal connection @xmath460 on @xmath19 defined in section 3.1 via @xmath535 where @xmath486 is the orthogonal projection onto the horizontal distribution .
for this horizontal connection we formulate the compatibility conditions in the following way : [ def horizontal clifford bundle ] let @xmath536 be a bundle of horizontal clifford moduls over @xmath0 which is equipped with a fiber - wise hermitian metric @xmath537 and a metric connection @xmath538 .
@xmath536 is called a _ horizontal clifford bundle _ if a. for each @xmath215 we have @xmath539 for all @xmath540 , @xmath541
. b. if @xmath460 is the horizontal levi - civita connection on @xmath19 , we have @xmath542 for all @xmath484 and for all sections @xmath543 .
@xmath74 now , still in analogy to the classical dirac operator , we can define a horizontal dirac operator .
later we will see that this operator has to be modified by adding an endomorphism in order to be symmetric .
[ def formal horizontal dirac operator ] the _ formal horizontal dirac operator _ @xmath544 of @xmath533 is the first order differential operator on @xmath545 defined by the composition @xmath546 here the first arrow is given by the connection @xmath538 , the second arrow is given by the identification of @xmath547 and @xmath19 via the horizontal metric , and the third arrow is given by the clifford action .
if we choose a local orthonormal frame @xmath41 of sections of @xmath19 , we can write @xmath548 @xmath74 before we go on with our construction to get a self - adjoint operator from @xmath544 , we have a look at the most natural example .
[ ex exterior bundle ] consider the exterior bundle @xmath549 on a compact riemannian manifold @xmath0 without boundary equipped with its natural metric and connection .
it is well known ( see for example @xcite , 49ff . )
that ( in the classical sense ) this is a clifford bundle with multiplication given by the wedge product and clifford action of @xmath550 given by @xmath551 the dirac operator of this clifford bundle is given by @xmath552 .
remember that the interior product is defined by @xmath553 via the hodge star operator @xmath554 , which ( for a @xmath555-form @xmath556 ) is defined to be the unique @xmath557-form @xmath558 such that @xmath559 using this hodge star operator , the operator @xmath560 is defined by @xmath561 now a horizontal clifford bundle can be constructed analogously .
we assume @xmath0 to be a carnot manifold with horizontal distribution @xmath213 such that the frame @xmath562 is orthonormal , hence we can identify this frame with its dual frame @xmath223 .
we further identify @xmath547 with the sub - bundle of @xmath224 annihilating @xmath563 , which means @xmath564 .
then @xmath565 is a horizontal clifford bundle : the clifford action of @xmath566 is again given by @xmath567 note that in case @xmath568 and @xmath569 the right side of this equation lies still in @xmath533 .
if we use the horizontal exterior derivative @xmath570 where @xmath486 is the orthonormal projection onto @xmath547 , we get a connection which is compatible with the horizontal levi - civita connection is defined on @xmath547 . using this , we can define the formal horizontal dirac operator on @xmath536 and see that it is given by @xmath571 @xmath74 our approach of constructing a horizontal clifford bundle and a horizontal dirac operator can be found in the literature in a greater generality .
igor prokhorenkov and ken richardson recently introduced a class of so - called _ transversally dirac operators _ in @xcite by considering a distribution @xmath572 of the tangent bundle ( which may be integrable or not ) , which furnishes a @xmath573-module structure on a complex hermitian vector bundle @xmath574 over @xmath0 with clifford action @xmath575 and a connection @xmath576 on @xmath574 fulfilling the requirements of definition [ def horizontal clifford bundle ] .
hence our construction of a horizontal clifford bundle fits into this setting .
it turns out that the gap of this construction so far is that the operator @xmath544 is not symmetric : when one wants to calculate the @xmath52-adjoint of @xmath544 , stokes theorem causes an additional term of mean curvature into the vertical direction , projected to @xmath19 via the horizontal connection . hence to get a symmetric operator
, we have to add the ( horizontal ) clifford action by the mean curvature of the orthogonal complement of the horizontal distribution .
this was shown by prokhorenkov and richardson for the general case of a distribution of the tangent bundle ( @xcite ) , such that we refer to their work formulating the following theorem .
it shows in addition that the resulting operator is essential self - adjoint , which follows by a theorem of paul chernoff about the self - adjointness of certain differential operators ( @xcite ) .
altogether , we have the following theorem ( see @xcite , theorem 3.1 ) , which will finally provide us with the possibility to define ( essential self - adjoint ) horizontal dirac operators .
[ thm d^h selfadjoint ] let @xmath544 be a formal horizontal dirac operator , acting on the smooth sections of a horizontal clifford bundle @xmath533 over a compact carnot manifold @xmath0 without boundary .
then we have a. the formal @xmath52-adjoint of @xmath544 is given by @xmath577 where @xmath578 is the orthogonal projection onto the horizontal distribution and @xmath474 is the levi - civita connection of @xmath20 .
b. the operator @xmath579 is essentially self adjoint on @xmath580 .
@xmath91 [ def horizontal dirac operator ] let @xmath536 be a horizontal clifford bundle over a closed carnot manifold @xmath0 .
then the operator defined by , acting on @xmath581 , is called the _ horizontal dirac operator _ of @xmath533 . @xmath74 at this point , we want to state another example which is due to @xcite .
it states that on any ( classical ) clifford bundle @xmath574 over a closed carnot manifold @xmath0 one can implement the structure of a horizontal clifford bundle by adjusting the bundle connection .
we refer to @xcite , section 2 , for the following proposition .
[ prop horizontal clifford bundle from spinor bundle ] let @xmath0 be a carnot manifold with horizontal distribution @xmath19 , such that @xmath41 is an orthonormal frame for @xmath19 and @xmath582 is an orthonormal frame for @xmath583 .
let in addition @xmath574 be a clifford bundle over @xmath0 with clifford action @xmath519 and bundle connection @xmath576 , and let @xmath43 denote the restriction of @xmath519 to the horizontal distribution @xmath19 .
then there is a connection @xmath584 on @xmath585 such that @xmath585 equipped with @xmath43 and @xmath584 is a horizontal clifford module .
this connection @xmath584 is given by @xmath586 where @xmath587 denotes the orthogonal projection of @xmath20 onto @xmath19 and @xmath474 denotes the levi - civita connection on @xmath20 .
@xmath91 in chapter 4 , we will consider another example in detail : on the local homogeneous space of a carnot group @xmath33 of rank @xmath216 under the action of a lattice subgroup , there is a submersion @xmath588 onto the @xmath216-dimensional torus by section 2.4 .
then one can define a horizontal clifford bundle and a horizontal dirac operator by pulling back the objects from @xmath49 , which does not depend on the non - horizontal directions of @xmath20 .
we will discuss this example extensively in chapter 4 , since it serves well as a toy model to the question whether one can define spectral triples from horizontal dirac operators .
for further considerations it will be important to work with the square of the horizontal dirac operator .
especially it will follow in section 6.3 that this operator , considered as an operator on @xmath589 , is not hypoelliptic , and we will conclude from that that a horizontal dirac operator in not hypoelliptic . hence we calculate @xmath590 , whose principal term will be a horizontal laplacian , in a local expression .
[ prop d^h squared ] let @xmath0 be a carnot manifold , and let @xmath41 be an orthonormal frame for its horizontal distribution @xmath19 .
let @xmath1 be a horizontal dirac operator acting on a horizontal clifford bundle @xmath536 over @xmath0 .
then we have locally @xmath591 + o_h(1),\ ] ] where @xmath324 is to be understood as a component - wise directional derivative in a local chart and @xmath592 denotes a graded differential operator of order smaller or equal to @xmath98 ( which only depends on the differential operators @xmath214 and endomorphisms on the bundle ) .
* proof : * we work with the local expression of @xmath1 , given by . from this
we get for @xmath568 @xmath593 where @xmath594 is the mean curvature of the bundle @xmath583 like in theorem [ thm d^h selfadjoint ] .
note that we have @xmath595 .
we observe that the second summand on the right hand side of consists only of differential operators of graded order @xmath98 or @xmath596 applied to @xmath556 , so it is contained in the expression @xmath592 .
further we have @xmath597 by definition [ def horizontal clifford bundle ] , since @xmath533 is a horizontal clifford bundle , so this expression is also contained in @xmath592 .
now we can not assume @xmath598 as one does when calculating the square of an ordinary dirac operator in this case , since our frame is fixed and therefore can not be chosen to be synchronous at a point @xmath215 .
we calculate for @xmath531 , using once again the properties of a horizontal clifford bundle : @xmath599 since we are working in local charts , @xmath600 is locally given by the expression @xmath601 with @xmath602 . applying this to the last expression
, we find @xmath603 and therefore @xmath604 since @xmath605 and @xmath606 by the characterization of the clifford algebra .
finally we plug this into and find together with the discussion above @xmath607 \left(\sigma\right ) + o_h(1)\sigma.\end{aligned}\ ] ] this proves the proposition .
@xmath91 * remark : * the proof of this proposition shows that the leading term @xmath608\ ] ] in the local expression of @xmath609 will not change if we modify @xmath1 by an term of graded order zero . for an operator which is locally of the type @xmath610 with @xmath611 we still have @xmath612 + o_h(1).\ ] ] in particular this shows that for a modified connection @xmath613 with @xmath614 the leading term of the local expression of @xmath609 according to this connection will not change .
we will refer to this remark when we check @xmath1 for hypoellipticity in section 6.3 , since the property of being hypoelliptic only depends of the leading term of a horizontal laplacian .
before we conclude this section , we want to prove a rather technical lemma about the eigenvalues a certain combination of clifford matrices can have .
this will be needed in the following chapters , and since it will be needed more than one time we decided to put it here .
its statement is quite general , since it can be applied to the clifford action arising from any vector bundle , horizontal or not .
[ prop eigenvalues sum c(x_j)c(x_k ) ] let @xmath502 be a vector space of dimension @xmath216 , and let @xmath615 denote its ( complexified ) clifford algebra , which is represented by matrices on a complex vector space @xmath318 via .
we assume that we have an @xmath246 such that @xmath616 and such that @xmath617 is a basis of @xmath502
. then we have the following : a. for any @xmath618 with @xmath619 , all the eigenvalues of the matrix @xmath620 are given by the numbers @xmath621 , where both eigenvalues have the same multiplicity .
b. the eigenvalues of the matrix @xmath622 are exactly the numbers @xmath623 in the case where @xmath624 and the clifford action of @xmath502 on @xmath318 is irreducible , which means we have @xmath625 , each eigenvalue @xmath626 has the multiplicity @xmath627 . c. for @xmath628 with @xmath629 ( with @xmath630 ) ,
all the eigenvalues of the matrix @xmath631 are included in the interval @xmath632 \subset i { \mathbb{r}}\ ] ] on the imaginary line .
thereby , the numbers @xmath633 and @xmath634 are eigenvalues of the matrix . * remark : * although this proposition has been formulated for vector spaces , it can be transferred to clifford bundles immediately , since point - wise we have the eigenvalues noted above . in particular
, the situation of the proposition occurs for the horizontal clifford action on a clifford bundle on a carnot manifold @xmath0 , where @xmath0 is locally diffeomorph to a carnot group of the type @xmath393 . in this case
we will identify the horizontal frame @xmath41 at each point @xmath215 with the vector space spanned by @xmath635 , for which the proposition is formulated .
* proof : * first of all , since @xmath636 each of the products @xmath620 has eigenvalues @xmath637 and @xmath638 . furthermore , the matrices @xmath620 and @xmath639 commute since @xmath640 and are thus simultaneously diagonalizable .
because of the relation @xmath641 this means that we have the same number of @xmath637- and @xmath638-eigenvalues , and therefore statement ( i ) is proved .
the next observation is that for @xmath532 and @xmath642 , the matrices @xmath643 and @xmath644 commute : we have @xmath645\\ & = & c(e_j)c(e_{m+j})c(e_k)c(e_{m+k } ) - c(e_k)c(e_{m+k})c(e_j)c(e_{m+j})\\ & = & c(e_k)c(e_{m+k})c(e_j)c(e_{m+j } ) - c(e_k)c(e_{m+k})c(e_j)c(e_{m+j})\\ & = & 0\end{aligned}\ ] ] by the commutation rules of the horizontal clifford action , see proposition [ prop clifford module over hm ] . but
this means that all the summands of the matrix @xmath646 are simultaneously diagonalizable .
let @xmath318 be a matrix diagonalizing these summands simultaneously , this means its diagonal matrix is given by @xmath647 using this , we can prove the rest of the statements by induction over @xmath648 . from this point on , we will assume for a moment that the representation of our clifford algebra is irreducible .
this means that we have a map @xmath649 for a vector space @xmath318 with @xmath650}$ ] , @xmath651 $ ] denoting the gaussian bracket , such that for every @xmath529 we have @xmath652 } \times 2^{[d/2]}}({\mathbb{c}})\ ] ] ( see the discussion in the beginning of this section for this ) .
we further denote the ( unique ) complexified clifford algebra arising from a ( complex ) vector space of dimension @xmath94 by @xmath653 . for the proof by induction
, we will use the isomorphism @xmath654 see @xcite , theorem i.4.3 .
now we start with the induction argument : for the case @xmath655 the statements ( ii ) and ( iii ) already follow from statement ( i ) .
now we assume the statements to be true for @xmath656 . in detail , this means the following : for a vector space @xmath657 with basis @xmath658 , we have a clifford algebra @xmath659 .
since we assumed irreducibility of the clifford action , the elements @xmath660 are given by @xmath661 - 1 } \times 2^{[d/2 ] - 1})$]-matrices , such that the statements ( ii ) and ( iii ) are true for sums of the form @xmath662 considering @xmath656 instead of @xmath648 .
now for a vector space @xmath663 we have @xmath664 and consider a basis @xmath665 .
hence after using the isomorphism from , we can work with the following representations of the elements @xmath666 : @xmath667 - 1 } } \otimes \begin{pmatrix } 0&i \\
i&0 \end{pmatrix } & \text{for } \ j = m \\ c_{m-1}(\tilde{e}_{j-1 } ) \otimes \begin{pmatrix } i&0 \\
0&-i \end{pmatrix } & \text{for } \
m+1 \leq j \leq 2m-1 \\ { \text{id}}_{2^{[d/2 ] - 1 } } \otimes \begin{pmatrix } 0&-1
\\ 1&0 \end{pmatrix } & \text{for } \
j=2 m \\
c_{m-1}(\tilde{e}_{j-2 } ) \otimes \begin{pmatrix } i&0 \\
0&-i \end{pmatrix } & \text{for } \
2m+1 \leq j \leq d,\\ \end{cases}\ ] ] where @xmath668 denotes the clifford action by @xmath669 and @xmath670 denotes the clifford action by @xmath671 on the corresponding vector spaces .
note that the matrices @xmath672 together with the @xmath673-unit matrix , form a basis for the clifford algebra @xmath674 .
for the representations from we get by the rules of the tensor product for matrices : @xmath675 - 1 } } \otimes \begin{pmatrix } 0&i \\
i&0 \end{pmatrix } \right ) \cdot \left ( { \text{id}}_{2^{[d/2 ] - 1 } } \otimes \begin{pmatrix } 0&-1 \\ 1&0 \end{pmatrix } \right ) \\ & = & \sum_{j=1}^{m-1 } \lambda_j \left(c_{m-1}(\tilde{e}_j ) c_{m-1}(\tilde{e}_{m-1+j } ) \otimes \begin{pmatrix } -1&0 \\
0&-1 \end{pmatrix } \right ) + \lambda_m \cdot { \text{id}}_{2^{[d/2 ] - 1 } } \otimes \begin{pmatrix } i&0 \\
0&-i \end{pmatrix}\\ & = & \left ( \sum_{j=1}^{m-1 } \lambda_j c_{m-1}(\tilde{e}_j ) c_{m-1}(\tilde{e}_{m-1+j } ) \right ) \otimes \begin{pmatrix } -1&0 \\
0&-1 \end{pmatrix } + \lambda_m \cdot { \text{id}}_{2^{[d/2 ] - 1 } } \otimes \begin{pmatrix } i&0 \\ 0&-i \end{pmatrix}.\end{aligned}\ ] ] now let @xmath676 be the matrix diagonalizing the matrices @xmath677 simultaneously , such that we have in this situation .
but then we see from the above calculation that the matrix @xmath678 diagonalizes @xmath679 simultaneously , which means we have because of the above calculation @xmath680 - 1 } } \otimes \begin{pmatrix } i&0 \\ 0&-i \end{pmatrix}. \end{split}\ ] ] since all these matrices are diagonal matrices , we immediately see that for each eigenvalue @xmath681 , @xmath682 , of @xmath683 with multiplicity @xmath684 the numbers @xmath685 are eigenvalues of @xmath686 . but from we can prove the statements ( ii ) and ( iii ) by induction : * statement ( iii ) can be seen immediately from .
if @xmath687 is the greatest and @xmath688 is the lowest eigenvalue of @xmath689 , then the greatest and lowest eigenvalue of are given by the numbers @xmath690 which proves both statements of ( iii ) .
* statement ( ii ) follows after choosing @xmath691 in . by the assumption that ( ii ) is true for @xmath656
, we have @xmath692 and @xmath693 for @xmath694 , where each @xmath695 has multiplicity @xmath696 .
we observe for @xmath697 that @xmath698 hence each of these eigenvalues has multiplicity @xmath699 .
in addition we have the eigenvalues @xmath700 and @xmath701 , which are both of multiplicity @xmath98 .
we finally drop the restriction that the representation of the clifford algebra is irreducible : we do not get any new eigenvalues , because every ( reducible ) representation is a direct sum of irreducible ones , and hence ( ii ) and ( iii ) are also true for this case .
but we can not make a general statement about the multiplicity of the eigenvalues in the reducible case .
altogether , every statement of this proposition is proved.@xmath91 throughout this section we will assume that @xmath0 is a compact carnot manifold without boundary , such that the algebra @xmath45 is a unital @xmath2-algebra which can be represented on @xmath702 via left multiplication . our intention is to show that the operator @xmath1 from the previous section detects the carnot - carathodory metric via the connes metric formula @xmath703\right\| \leq 1\right\},\ ] ] where @xmath86 is a dense sub - algebra of @xmath45 .
this means that although @xmath1 does not furnish a spectral triple ( which we will see in general in chapter 6 ) , we can consider the triple @xmath704 as a compact quantum metric space in the sense of mark rieffel , see definition [ def compact quantum metric space ] , with the corresponding lip - norm @xmath705\right\|$ ] , whose metric is exactly the carnot - carathodory metric on @xmath0 .
the key observation is that , in analogy to the classical case from the standard example for a spectral triple , the commutator @xmath62 $ ] acts as clifford action by the horizontal gradient of a function @xmath63 .
remember that on a carnot manifold @xmath0 the horizontal gradient of a function @xmath217 is given by the vector field @xmath706 where , like before , @xmath707 is a horizontal distribution of @xmath0 .
[ prop d^h , f = c grad^hf ] let @xmath1 be a horizontal dirac operator acting on a horizontal clifford bundle @xmath533 with horizontal clifford action @xmath43 over a closed carnot manifold @xmath0
. then , for any function @xmath217 , we have @xmath708 = c^h({\mathrm{grad}}^h f).\ ] ] * proof : * let us first show that the horizontal dirac operator fulfills the leibniz rule .
this is just an easy calculation using the properties of the connection and the clifford action : for any @xmath543 we have @xmath709 now the statement follows immediately : @xmath708 \sigma = d^h(f \cdot \sigma ) - f
\cdot d^h \sigma = c^h\left({\mathrm{grad}}^h f\right ) \sigma,\ ] ] and the proposition is proved .
@xmath91 using this proposition , we have to show that the lip - norm defined by the connes metric coming from @xmath1 coincides with the supremum of the horizontal gradient of a function @xmath63 .
but yet it is not clear how the sub - algebra @xmath86 of @xmath45 has to look like such that the connes metric formula is true . in the classical case
, this is exactly the algebra of lipschitz functions , and it will turn out that in the carnot case this will be the algebra of functions which are lipschitz with respect to the carnot - carathodory metric .
[ def lip_cc ] let @xmath22 be a carnot manifold , @xmath710 .
then we call the number @xmath711 the _ carnot - carathodory - lipschitz constant _ of @xmath63 .
if @xmath712 is finite , we call @xmath63 a _ carnot - carathodory - lipschitz function_. we denote the algebra of all carnot - carathodory - lipschitz functions on @xmath0 , equipped with the semi - norm @xmath712 , by @xmath713 . @xmath74 the proof that the horizontal dirac - operator @xmath1 detects the carnot - carathodory metric works similar to the classical one that a dirac operator detects the geodesic metric on a riemannian spin manifold ( see e.g. @xcite , @xcite or @xcite ) :
if we assume that for @xmath714 the horizontal gradient @xmath715 exists almost everywhere , we show that the number @xmath712 coincides with the essential supremum of @xmath716 .
after that we will show that the carnot - carathodory metric can be described via the carnot - carathodory - lipschitz constant , such that we can apply proposition [ prop d^h , f = c grad^hf ] to get the result .
we need to show that the assumption that @xmath715 exists almost everywhere for @xmath714 .
note that the analogous statement for classical lipschitz functions is well known ( see e.g. @xcite ) .
the horizontal case is shown in @xcite for the case of the heisenberg group ( see @xcite , proposition 6.12 ) , and this result can be generalized easily to arbitrary carnot manifolds as we show in the following proposition . in a greater generality
, this is also a consequence of the pansu - rademacher theorem ( @xcite ; see also for example @xcite , theorem 6.4 ) , which states that a lipschitz map between two carnot groups , the so - called _ pansu differential _ ( see @xcite ) , exists .
[ prop lip_cc functions are differentialble almost everywhere ] let @xmath0 be a closed carnot manifold with horizontal distribution @xmath213 and let @xmath717 be a carnot - carathodory - lipschitz function .
then the horizontal gradient @xmath718 exists almost everywhere on @xmath0 .
* proof : * we consider the case where @xmath719 is an open subset of a carnot group @xmath33 ; then the general case follows after restricting ourselves to local coordinates ( on which we have a carnot group structure ) since the coordinate changes are smooth and therefore do not affect the regularity of @xmath63 .
thus we choose an arbitrary point @xmath720 .
we fix a @xmath529 . after an affine change of coordinates on @xmath118 ( such that @xmath721 ) we can assume @xmath722 where the coordinates on @xmath723 are meant to be either exponential or polarized coordinates , see section 2.2 .
the idea is to consider the integral curves arising from @xmath724 into the direction of the vector field @xmath324 via the exponential map , which means we have a curve @xmath725 \rightarrow { \mathbb{g } } , \ \ \ t \mapsto x_0 . \exp{t
x_j}\ ] ] for any @xmath726 .
we denote by @xmath727 ) \cap \omega$ ] the path of @xmath728 .
since this path is horizontal , the map @xmath729 is a ( classical ) lipschitz function on @xmath515 because of the carnot - carathodory lipschitz property of @xmath63 .
since @xmath730 is lipschitz , it is differentiable almost everywhere ( see e.g. @xcite ) , and its derivative is given by @xmath731 but this shows that @xmath732 exists almost everywhere on @xmath733 .
this is true for any starting point @xmath734 , such that we can conclude that @xmath735 exists almost everywhere on @xmath118 because any point of a carnot group @xmath33 can be reached from the hyperplane @xmath736 by a horizontal curve of the type . now
the above argument is true for any @xmath529 , and hence any horizontal partial derivative @xmath735 and therefore also the horizontal gradient exists almost everywhere on @xmath118 .
@xmath91 to show that the carnot - carathodory - lipschitz constant coincides with the supremum norm of the horizontal gradient , we work with the corresponding object from the cotangent bundle of @xmath0 .
we assume we have a riemannian metric @xmath247 on @xmath0 such that @xmath55 is an orthonormal frame with respect to @xmath247 , which respects the grading structure of @xmath20 .
remember from section 2.1 that we can choose a basis @xmath223 of @xmath224 such that @xmath737 according to this basis , the horizontal differential of @xmath63 is given by @xmath738 obviously we have @xmath739 for any function @xmath217 , where @xmath356 denotes the ( euclidean ) norm of a horizontal tangent ( or cotangent ) vector on @xmath0 , coming from the riemannian metric @xmath247 .
[ lemma lip_cc = grad^h ] let @xmath0 be a closed carnot manifold with riemannian metric @xmath247 as above , and let @xmath714 such that , by proposition [ prop lip_cc functions are differentialble almost everywhere ] , the horizontal gradient @xmath716 of @xmath63 exists almost everywhere .
then we have @xmath740 * proof : * for @xmath714 we show that @xmath741 and that @xmath742 .
let @xmath186 , and let @xmath209 \rightarrow m$ ] be a smooth horizontal curve connecting @xmath107 and @xmath108 , which means @xmath743 , @xmath744 and @xmath211 for all @xmath210 $ ] .
note that because of and the characterization of the horizontal differential this means @xmath745 for all @xmath746 , and we have for @xmath217 @xmath747 where we have used the cauchy - schwarz inequality . taking the infimum over all horizontal curves connecting @xmath107 and @xmath108 , we find @xmath748 now the above calculation holds not only for @xmath749-functions @xmath63 , but for all functions which have a horizontal gradient almost everywhere , such that @xmath750 is defined as an essentially bounded vector field on @xmath0 .
hence for every @xmath63 which fulfills the assumptions of the lemma , becomes @xmath751 and since @xmath107 and @xmath108 can be chosen arbitrarily we see @xmath752 on the other hand , we choose an @xmath204 such that @xmath753 exists and consider the integral curve along the vector field @xmath754 arising from the exponential map @xmath755 from riemannian geometry , that is @xmath756 for a small @xmath116 we set @xmath757 and denote the length of a horizontal curve @xmath758)$ ] which is connecting @xmath724 and @xmath107 by @xmath759 .
we observe that we have @xmath760 now , since @xmath728 is a horizontal curve connecting @xmath107 and @xmath724 we have @xmath761 , and hence we get from @xmath762 by definition of the carnot - carathodory - lipschitz constant .
but since this works for every @xmath204 where @xmath763 exists ( which is almost every @xmath204 ) , we have @xmath764 altogether and prove the statement of the lemma .
@xmath91 to make use of the above lemma , one has to show that the carnot - carathodory distance can be expressed using the carnot - carathodory - lipschitz constant .
this should be obvious , since it is just the lipschitz semi - norm belonging to the compact metric space @xmath231 , but for completeness we write down the proof . from our point of view
it is important that the function describing the carnot - carathodory distance from a fixed point @xmath204 is a carnot - carathodory - lipschitz function , which has lip - norm bounded by @xmath98 and is differentiable into horizontal directions almost everywhere .
but note that this is not a @xmath749-function .
[ lemma d_cc from lip_cc ] on a closed carnot manifold @xmath0 the carnot - carathodory distance between two points @xmath186 is given by @xmath765 * proof : * for every @xmath710 such that @xmath766 we have for all @xmath186 : @xmath767 on the other hand , we define a function @xmath768 via @xmath769 for a given @xmath215 .
obviously @xmath770 is continuous , and since @xmath39 is a metric on @xmath0 we have for any @xmath771 @xmath772 this shows @xmath773 , and since @xmath774 we get @xmath775 from and the statement of the lemma follows .
@xmath91 * remark : * the fact that the function @xmath770 appearing in the proof is differentiable almost everywhere can also be deduced from the fact that any two points @xmath186 can be joint by a so - called minimizing geodesic , which is a part from @xmath107 to @xmath108 realizing the carnot - carathodory distance ; fulfilling the additional property that it has a derivative for almost all @xmath746 whose components are measurable functions ( see @xcite , theorem 1.19 ) .
we do not need this argument , since the proof above shows that @xmath770 is a carnot - carathodory - lipschitz function which has the property mentioned above by proposition [ prop lip_cc functions are differentialble almost everywhere ] .
@xmath74 now we simply have to put everything together to get the identity of the metrics .
[ thm d_cc from d^h ] let @xmath0 be a closed carnot manifold and @xmath1 the horizontal dirac operator acting on a horizontal clifford bundle @xmath533 over @xmath0 . then the carnot - carathodory metric of @xmath0 can be detected via the formula @xmath776
\right\| \leq 1\right\}.\ ] ] * proof : * by lemma [ lemma lip_cc = grad^h ] and lemma [ lemma d_cc from lip_cc ] we have @xmath777 now proposition [ prop d^h , f = c grad^hf ] tells us that @xmath778 = c^h({\mathrm{grad}}^h f)\ ] ] for every @xmath217 , where @xmath779 denotes the horizontal clifford action , and the norm of the operator @xmath780 is given by @xmath781 but since because of @xmath782 the map @xmath783 is an isometry for any @xmath215 , this shows @xmath784 hence together with we see that the identity @xmath785 \right\|\ ] ] is true for any @xmath217 . since for @xmath714
the horizontal gradient @xmath750 exists almost everywhere by proposition [ prop lip_cc functions are differentialble almost everywhere ] , implies @xmath786 \right\| \ \ \ \forall f \in \mathrm{lip}_{cc}(m),\ ] ] and therefore the theorem is proved .
@xmath91 in theorem [ thm d_cc from d^h ]
we have seen that any horizontal dirac operator detects the carnot - carathodory metric via connes metric formula , where the supremum is taken over the carnot - carathodory - lipschitz functions on @xmath0 .
we will see now that it suffices to take the supremum over all @xmath339-functions , since each @xmath714 can be approximated by functions @xmath787 with smaller @xmath788-norm by a standard approximation argument .
[ cor d_cc from d^h via smooth functions ] let @xmath0 be a closed carnot manifold and @xmath1 the horizontal dirac operator acting on a horizontal clifford bundle @xmath533 over @xmath0 .
then the carnot - carathodory metric of @xmath0 can be detected via the formula @xmath789
\right\| \leq 1\right\}.\ ] ] * proof : * let @xmath33 be the tangent carnot group @xmath33 of the carnot manifold @xmath0 , where like before @xmath221 is the nilpotency step of @xmath33 and for @xmath38 the number @xmath790 denotes the dimension of the vector space @xmath287 belonging to the grading @xmath34 of @xmath33 . we show that each function @xmath791 can be approximated uniformly by a sequence of @xmath339-functions @xmath792 such that @xmath793 for all @xmath125 .
then the statement for @xmath788-functions on the manifold @xmath0 follows , since @xmath0 is compact , by considering local charts , and hence the corollary follows immediately from theorem [ thm d_cc from d^h ] .
let @xmath791 .
we use the koranyi gauge @xmath794 see definition [ def koranyi gauge ] , and consider the unit ball @xmath795 with respect to this semi - norm .
then we can consider the smooth function @xmath796 and set @xmath797 where we choose the constant @xmath798 such that @xmath799 .
but this means that we have for all @xmath116 @xmath800 where @xmath801 denotes the weighted dilation on @xmath33 by @xmath802 ( see definition [ def delation on carnot group ] ) , and @xmath803 is the hausdorff dimension of @xmath33 ( see theorem [ thm mitchells measure theorem ] ) .
note that because of the smoothness of @xmath804 the function @xmath507 is a smooth function on @xmath33 , which is compactly supported in the unit ball with respect to the koranyi gauge on @xmath33 .
we finally define @xmath805 which provides us a family of functions @xmath806 with @xmath807 such that @xmath808 for all @xmath116 ( because of ) and @xmath809 in the sense of distributions . because of the equivalence of the koranyi gauge and the carnot - carathodory metric ( see proposition [ prop equivalence koranyi metric carnot - caratheodory metric ] ) there is a @xmath359 such that each of these functions @xmath810 is supported in @xmath811 , where @xmath812 denotes the carnot - carathodory ball with radius @xmath125 . using the compactly supported smooth functions @xmath810 from [ eq smooth family varphi_epsilon ] , we consider @xmath813 where @xmath814 denotes the composition on @xmath33 .
it is clear from the rules of convolution that we have @xmath815 .
further , since @xmath63 in continuous ( with respect to the carnot - carathodory metric on @xmath33 ) , we know that for any compact subset @xmath816 and any @xmath171 there is an @xmath817 such that @xmath818 for all @xmath819 and for all @xmath820 , where @xmath821 and the constant @xmath822 is chosen such that @xmath823 is supported in @xmath824 . but
this shows that for all @xmath825 we have because of the translation invariance of the carnot - carathodory metric with respect to the composition on @xmath33 ( see @xcite ) and the properties of @xmath810 @xmath826 for all @xmath819 . since @xmath0 is a compact manifold
( hence we can restrict ourselves to compact subsets in any chart ) , this shows that any @xmath788-function @xmath63 on @xmath0 can be approximated by these @xmath339-functions @xmath792 in the supremum norm .
finally , we show that for any @xmath116 we have @xmath793 .
this follows once again because the carnot - carathodory metric is translation invariant with respect to the composition on the carnot group @xmath33 ( see e.g. @xcite ) . from the definition of the carnot - carathodory - lipschitz constant this invariance in connection with the properties of @xmath810 leads to the estimate @xmath827
altogether we have proved that any function @xmath828 can be approximated uniformly by a sequence of @xmath339-functions @xmath792 such that @xmath793 for all @xmath125 , and the statement of the corollary follows .
@xmath91 since @xmath713 ( or @xmath96 ) is a dense sub - algebra of @xmath45 , proposition [ prop d^h , f = c grad^hf ] shows that @xmath62 $ ] is bounded for a dense sub - algebra of @xmath45 and therefore @xmath1 fulfills condition ( i ) for a spectral triple ( see definition [ def spectral triple ] .
but as we have already mentioned ( and will see in the next chapter for a concrete class of examples and in chapter 6 in general ) , @xmath1 fails to have a compact resolvent , and therefore @xmath829 is not a spectral triple .
but on the other hand , theorem [ thm d_cc from d^h ] and the preceding lemmas suggest that @xmath1 seems to be the logical candidate for a first order differential operator to detect the carnot - carathodory metric .
we will now show that at least @xmath830 gives rise to a compact quantum metric space in the sense of mark rieffel ( see definition [ def compact quantum metric space ] ) if we consider the real - valued functions of @xmath45 as an order - unit space .
[ cor d^h provides compact quantum metric space ] the pair @xmath151 , where @xmath831 \right\|,\ ] ] is a compact quantum metric space which detects the carnot - carathodory metric on a closed carnot manifold @xmath0 .
* proof : * it is obvious that @xmath136 is an order - unit space with norm @xmath832 now for calculating the connes metric from @xmath68 it suffices to consider only the self - adjoint elements of @xmath45 ( see proposition [ prop metric from positive elements ] , which is exactly the space @xmath136 .
we have to show that @xmath68 is a lip - norm on @xmath136 , i.e. a. for every @xmath185 we have @xmath833 @xmath146 @xmath834 .
b. the topology on the state space @xmath148 of @xmath136 from the connes metric defined by @xmath68 is the @xmath150-topology .
the non - trivial part of condition ( i ) follows from proposition [ prop d^h , f = c grad^hf ] : since @xmath835 = c^h(\left ( { \mathrm{grad}}^h f \right),\ ] ] @xmath62 = 0 $ ] implies @xmath836 almost everywhere , which implies @xmath837 for all @xmath529 almost everywhere .
but this also means @xmath838 almost everywhere for every @xmath839 because every vector field of the @xmath840 s can be written as a commutator consisting of the @xmath324 s .
since @xmath55 spans the tangent space of @xmath0 and @xmath63 is continuous by assumption , this implies that @xmath63 must have been a constant .
therefore ( i ) is proved . for condition ( ii ) , note that @xmath841 by gelfand - naimark theory , where the @xmath150-topology on @xmath0 is exactly the usual manifold topology .
now , by the sub - riemannian theorem on topologies ( see theorem [ thm theorem on topologies ] ) , this topology coincides with the topology induced by the carnot - carathodory metric @xmath39 .
but by theorem [ thm d_cc from d^h ] , @xmath842 which shows condition ( ii ) .
hence @xmath151 is a compact quantum metric space , and the statement about the metric is just the statement of theorem [ thm d_cc from d^h ] . @xmath91 * remark : * note that the carnot - carathodory - lipschitz constant @xmath68 also provides a lip - norm on @xmath136 . as we have seen in lemma [ lemma d_cc from lip_cc ] , this is exactly the lip - norm which belongs to the carnot - carathodory metric @xmath39 . in this context ,
corollary [ cor d^h provides compact quantum metric space ] shows that the compact quantum metric spaces @xmath843 and @xmath844\right\| \right)$ ] are identical .
in the previous chapter we constructed horizontal dirac operators more or less analogous to classical dirac operators and we saw that they detect the carnot - carathodory metric . therefore they seem to be a natural candidate to construct a spectral triple which covers the horizontal geometry of a carnot manifold . in this chapter
we will do a concrete construction of horizontal dirac operators @xmath1 on local homogeneous spaces of carnot groups which arise from the action of a lattice subgroup , namely compact carnot nilmanifolds @xmath47 .
this can be seen as a generalization of the torus in the non - abelian case .
we will make use of the spin structures of the horizontal torus , arising as the image of a submersion from @xmath0 , and observe that we obtain a horizontal clifford structure via pullback where the representation of the horizontal clifford algebra is irreducible .
afterwards we will use an approach developed by christian br and bernd ammann ( see @xcite and @xcite ) to get information about the spectrum of @xmath1 .
it will follow that our horizontal dirac operator has an infinite dimensional eigenspace .
in particular its resolvent in not compact and thus it does not furnish a spectral triple .
the strategy is to decompose the @xmath52-space of horizontal clifford sections for the case where the horizontal distribution has co - dimension @xmath98 : in this case we can use well - known results from the representation theory of heisenberg groups to calculate the eigenvalues of @xmath1 . using this , we will be able to deduce the statement for the general case .
these calculations on carnot nilmanifolds can be seen as an example of a more general statement : in later chapters , we will use this idea of transferring the problem to the co - dimension @xmath98 case to show that on any carnot manifold a horizontal dirac operator is not hypoelliptic . despite this lack , as an additional result we are still able to extract the hausdorff dimension of @xmath22 from the asymptotic behavior of the non - degenerate eigenvalues of @xmath1 in the case where @xmath353 is a heisenberg group .
we will use the notion of a compact carnot nilmanifold introduced in example [ ex compact carnot nilmanifolds ] . throughout this chapter ,
we assume that @xmath46 is the standard lattice of a carnot group @xmath33 .
we denote the resulting local homogeneous space by @xmath47 .
let @xmath33 be a carnot group with horizontal distribution of rank @xmath216 and nilpotency step @xmath221 , and let @xmath845 be its standard lattice given by @xmath846 see example [ ex compact carnot nilmanifolds ] .
we consider the local homogeneous space @xmath47 , where the action of @xmath46 on @xmath33 is given by the group composition from the left .
we equip @xmath0 with a riemannian metric @xmath247 such that the vector bundles @xmath847 , forming the grading of @xmath20 , are pairwise orthogonal at each point .
to construct an example for a horizontal clifford connection on @xmath0 , we start by calculating the christoffel symbols for the horizontal connection on @xmath0 arising from the levi - civita connection ( see section 3.1 ) .
[ prop christoffel symbols homogeneous space ] let @xmath460 be the horizontal connection arising from the levi - civita connection @xmath474 on the compact carnot nilmanifold @xmath47 .
if @xmath41 is an orthonormal frame for the ( bracket generating ) horizontal distribution @xmath278 , then we have @xmath848 for all christoffel symbols
@xmath501 of @xmath460 , @xmath849 , corresponding to this basis .
in addition , if we extend the horizontal frame @xmath41 to an orthonormal tangent frame @xmath55 of @xmath0 which respects the grading of @xmath20 , then all christoffel symbols of @xmath474 belonging to this frame satisfy @xmath850 for @xmath851 and @xmath852 .
* proof : * by the construction of the horizontal connection the horizontal christoffel symbols are exactly the horizontal christoffel symbols of the levi - civita connection on @xmath0 , see proposition [ prop horizontal connection in local coordinates ] .
if @xmath247 is a riemannian metric on @xmath0 such that @xmath41 forms an orthonormal frame at every point , they can be calculated using the properties of the levi - civita connection .
we have locally @xmath853 ( with @xmath854 , and hence we get by the koszul formula for the levi - civita connection because of the orthonormality of the frame @xmath41 @xmath855 ) - g(x_l,[x_k , x_j ] ) + g(x_j,[x_k , x_l])\right)\\ = & \frac{1}{2 } \left(-g(x_k,[x_j , x_l ] ) - g(x_l,[x_k , x_j ] ) + g(x_j,[x_k , x_l])\right ) \end{split}\ ] ] for all @xmath856 . since for any choice of vector fields @xmath857 the vector fields @xmath858 and @xmath859 \in v_2m$ ] are orthogonal with respect to @xmath247 , the right hand side of vanishes
. therefore all the horizontal christoffel symbols are @xmath596 .
the fact that @xmath860 for all @xmath861 and for all @xmath862 can also be seen immediately from : if @xmath863 for @xmath864 , we have @xmath865 = 0 $ ] and @xmath866 \in v_{s+1}m$ ] for @xmath852 ( which means @xmath867 ) by the grading structure of @xmath20 . but
this means that we also have @xmath868 ) = 0 $ ] since @xmath869 .
hence every term on the right hand side of is @xmath596 , and the additional statement of the proposition follows .
@xmath91 * remark : * using formula , we can also calculate all the other christoffel symbols belonging to the levi - civita connection of @xmath0 as soon as we know the commutator relations of the lie algebra generated by the vector fields @xmath41 .
@xmath74 we now construct an irreducible horizontal clifford bundle over @xmath0 only involving the horizontal distribution of @xmath0 .
the idea is to consider the submersion @xmath870 from section 2.4 , which is given in exponential coordinates via @xmath871 we have seen in corollary [ cor projection carnot nilmanifolds ] that @xmath441 can be lifted to the nilmanifold given by the action of @xmath46 on @xmath33 .
this means we have a submersion @xmath872 of @xmath0 onto the @xmath216-dimensional torus , which ( by corollary [ cor projection carnot nilmanifolds ] ) coincides locally with the submersion @xmath441 of the carnot groups .
if @xmath873 is a local frame for @xmath874 such that we have @xmath875 locally on the carnot groups , then the differential of the submersion @xmath154 applied to our orthonormal frame @xmath876 is given by @xmath877 we choose the riemannian metric @xmath878 on @xmath49 such that @xmath879 forms an orthonormal frame .
the idea for our construction of a horizontal dirac operator is now to exploit the fact that @xmath49 is a spin manifold .
the starting point for our constructions is the following theorem , which summarizes the well known facts about the spin structures on @xmath49 and the realization of their corresponding spinor bundles .
[ thm spin structures on t^d ] there are @xmath880 different spin structures @xmath881 on @xmath49 , which are in one - to - one correspondence to the group homomorphisms @xmath882 they are indexed by @xmath883 where @xmath884 is the image of the generator @xmath885 of @xmath886 under @xmath125 . furthermore , for each spin structure the spinor bundle @xmath887 is a ( complex ) vector bundle of rank @xmath888}$ ] , where @xmath651 $ ] denotes the gaussian bracket , and the space of sections of @xmath889 can be identified with functions @xmath890}})$ ] such that @xmath891 * proof : * the spin structures of a connected riemannian manifold @xmath0 ( if they exist ) are characterized by group homomorphisms from the fundamental group of @xmath0 to @xmath892 ( see e.g. @xcite , theorem ii.2.1 ) . in the case of the @xmath216-dimensional torus @xmath49 ,
the fundamental group is @xmath886 , and since a group homomorphism @xmath893 is uniquely determined by the images @xmath894 of the @xmath216 generators @xmath885 of @xmath886 , there are @xmath880 possibilities for such a homomorphism . the second statement of the theorem follows from the construction of the spinor
bundle corresponding to a spin structure ( see e.g. @xcite or @xcite).@xmath91 * remark : * note that is also true for @xmath895 to be an @xmath52-spinor , since @xmath889 is dense in @xmath896 . for a given spin structure @xmath881 on the torus @xmath49
, we have a corresponding spinor bundle @xmath887 , equipped with a spinor connection @xmath897 .
since @xmath154 is a submersion , there exist unique pullbacks of these objects on @xmath0 ( see e.g. @xcite ) .
in detail , we have the following : * the sections of the pullback @xmath898 of the spinor bundle have the form @xmath899 } } f_j \pi^\ast\varphi_j : \ f_j \in c^\infty(m ) , \
\varphi_1 , \ldots , \varphi_{2^{[d/2 ] } } \ \text{basis sections of } \
\sigma_\delta { \mathbb{t}}^d\right\}.\ ] ] * if @xmath900 is the bundle metric on @xmath887 , a metric on @xmath898 is given via the pullback @xmath901 * the pullback @xmath902 of the spinor connection to @xmath898 has the form @xmath903 on the pull - backs of sections of @xmath887 .
for an arbitrary element of @xmath904 described by @xmath902 is defined by using the linear and tensorial behavior of a connection .
considering the clifford action @xmath905 on @xmath887 , we can use the identification @xmath906 given by the differential @xmath907 from to define a horizontal clifford action on @xmath898 .
this is done by the pull - back of the endomorphism bundle : for @xmath208 we define @xmath908 which is an endomorphism on @xmath898 since @xmath909 is an endomorphism on @xmath887 . in detail ,
point - wise we have by the definition of the pull - back of an endomorphism @xmath910 for any basis section @xmath507 of @xmath887 , which extends to the whole bundle via linearity . from the identification @xmath906 via @xmath907 we can conclude that the restriction of the riemannian metric @xmath247 on @xmath0 to @xmath19 is exactly the pull - back of the riemannian metric @xmath878 on @xmath49 , which was chosen such that @xmath911 forms an orthonormal frame for @xmath874 . by writing @xmath912
we will now show that this structure indeed gives a horizontal clifford bundle .
furthermore , we can write down the horizontal dirac operator for this bundle .
[ thm horizontal cilfford bundle from submersion ] let @xmath47 be the nilmanifold of a carnot group @xmath33 , and let @xmath913 , @xmath914 and @xmath43 be as above
. then @xmath913 equipped with the connection @xmath914 and the horizontal clifford multiplication @xmath43 is a horizontal clifford bundle over @xmath0 .
@xmath52-sections @xmath915 can be identified with @xmath916}}$]-valued functions @xmath556 on @xmath33 such that @xmath917 for all @xmath918 and @xmath308 , where @xmath125 is the group homomorphism describing the spin structure of @xmath49 and @xmath919 denotes the group operation on @xmath33 .
the horizontal dirac operator @xmath1 acting on @xmath920 is given in local coordinates by @xmath921 where @xmath922 denotes the partial derivative belonging to a local coordinate chart of @xmath0 . *
proof : * first of all , it is clear that the action of @xmath19 on @xmath923 via @xmath43 furnishes a clifford module over @xmath0 : since @xmath909 furnishes a clifford module structure on @xmath924 , we have for any @xmath208 @xmath925 for the basis sections @xmath926 of @xmath927 by .
note thereby that the restriction of the riemannian metric @xmath247 on @xmath0 to the horizontal distribution @xmath19 is exactly the pull - back of the riemannian metric @xmath878 we chose above .
analogously , the condition ( i ) and ( ii ) of definition [ def horizontal clifford bundle ] follow because the ( classical ) clifford bundle @xmath887 on the torus fulfills these conditions : for the metric on @xmath927 we have point - wise for any basis sections @xmath928 , @xmath929 and for any horizontal vector field @xmath208 @xmath930 by and , and hence the metric compatibility ( i ) follows .
to show the compatibility of @xmath914 with the horizontal connection @xmath460 , we first calculate this for any basis section @xmath926 of @xmath923 . for any @xmath206 we have @xmath931 by , and since the compatibility condition is true on the spinor bundle @xmath887 , using the levi - civita connection @xmath932 over @xmath49 .
thereby the fourth equation needs a bit more explanation .
the crucial point is that for all @xmath206 we have @xmath933 under the identification @xmath906 via @xmath907 from : if we write both sides of in local coordinates , we see that they coincide since the christoffel symbols of @xmath49 with respect to frame @xmath911 vanish as well as the christoffel symbols of the horizontal levi - civita connection on @xmath0 with respect to the frame @xmath41 do . but
this means by @xmath934 for all @xmath935 by linearity of @xmath909 and @xmath43 and by . we still need to show the compatibility with the connection for an arbitrary element @xmath936 . by
, such a @xmath556 can be written in the form @xmath937 where @xmath938 is a local spinor basis for @xmath887 and @xmath939 for @xmath940 .
but then the general compatibility follows from the above calculation and by the linearity and the tensorial behavior of the connection @xmath941 , since for each @xmath323 we have @xmath942 altogether we have shown that @xmath923 is a horizontal clifford bundle over @xmath0 . turning to the representation of our horizontal spinors , we use to describe a general section @xmath943 .
since @xmath47 is a local homogeneous space , each @xmath939 can be considered as a function @xmath944 such that @xmath945 for all @xmath918 and @xmath308 . and
since the submersion @xmath154 is a lie group homomorphism according to the group operations on @xmath33 and @xmath946 , we have for each @xmath947 @xmath948 by equation from theorem [ thm spin structures on t^d ] . from this , follows for sections @xmath936 , and since @xmath949 is dense in the hilbert space @xmath950 of @xmath52-sections of this vector bundle , the second statement of the theorem is proved .
finally we prove the local description of the horizontal dirac operator from this clifford module .
let @xmath951 denote the christoffel symbols of @xmath874 according to the frame @xmath952 , which are all zero .
then the formula follows , because it is known that we have locally @xmath953 for the spinor connection in the torus ( see e.g. @xcite , lemma 4.1 ) : using once again the representation for elements of @xmath954 , we get from the definition of @xmath955 for every @xmath529 : @xmath956 now by definition [ def horizontal dirac operator ] , the horizontal dirac operator is given locally via @xmath957 where @xmath474 is the levi - civita connection on @xmath20 and @xmath587 denotes the orthonormal projection of @xmath20 onto the horizontal distribution @xmath19 .
for the first term on the right hand side of we can plug in the result from the above calculation , and for the second term we can use the additional statement of proposition [ prop christoffel symbols homogeneous space ] which states that all the christoffel symbols @xmath958 for @xmath851 and @xmath852 are zero : by definition of the levi - civita connection this implies that we have locally @xmath959 for any @xmath960 .
altogether , the local expression for the horizontal dirac operator we constructed follows .
@xmath91 let @xmath47 be a compact carnot nilmanifold , equipped with the objects @xmath913 , @xmath914 , @xmath43 and @xmath1 from theorem [ thm horizontal cilfford bundle from submersion ] .
then we call @xmath961 the _ horizontal pull - back dirac operator _ on @xmath0 .
further , we call @xmath913 the _ horizontal spinor bundle _ and @xmath914 the _ horizontal spinor connection _ on @xmath0
. @xmath74 let us summarize what we have done in this section : we have constructed a horizontal clifford bundle together with a horizontal dirac operator on an arbitrary compact carnot nilmanifold @xmath47 only depend on the horizontal distribution of this manifold . for dimensional reasons
this representation of the ( bundle of ) clifford algebras over @xmath19 is irreducible , since our clifford bundle @xmath962 has rank @xmath888}$ ] whenever @xmath216 is the rank of @xmath19 .
therefore we can claim we have constructed a natural candidate for a horizontal dirac operator on @xmath0 .
we remark that one can also construct horizontal dirac operators from already existing ( classical ) clifford or spinor bundles of @xmath0 . for this
we take a look at proposition [ prop horizontal clifford bundle from spinor bundle ] and note that the horizontal connection in this case depends on the clifford action of the whole tangent bundle @xmath20 . but nonetheless , whenever we choose an orthonormal frame @xmath55 of @xmath20 we will get a local expression similar to for the resulting horizontal dirac operator , which only differs by a matrix term arising from the clifford action .
therefore , methods similar to the ones described in the following sections can also be applied to this situation , and we expect similar results . our aim is to show that the horizontal pull - back dirac operator constructed in the preceding section , which detects the carnot - carathodory metric by section 3.3 , does not have a compact resolvent
. this will be the case if we find an eigenvalue of @xmath1 which possesses an infinite dimensional eigenspace .
therefore we are interested in getting information about the spectrum of this operator .
we intend to use the local expression of the horizontal pull - back operator @xmath1 we constructed in the last section . from this expression
, we will be able to use techniques involving the representation theory of its underlying carnot group : the following proposition shows that @xmath1 can be expressed using the right regular representation of @xmath33 ( see also @xcite for the case of the classical dirac operator ) .
[ prop expression of horizontal dirac via representation ] let @xmath47 be a compact carnot nilmanifold , and let @xmath1 be the horizontal pull - back dirac operator defined on the horizontal spinor bundle @xmath923 , arising from a spin structure @xmath881 on @xmath49 .
we denote by @xmath963 the right regular representation of the carnot group @xmath33 on the hilbert space @xmath964 , which is defined by @xmath965 for all @xmath966 .
then @xmath1 can be expressed locally using @xmath221 via @xmath967 for any @xmath936 . as a consequence ,
a closed subspace @xmath968 is invariant under @xmath1 if it is invariant under @xmath221 and under the clifford action of @xmath19 .
* remark : * the expression is well - defined since elements of @xmath964 can be viewed as periodic functions on @xmath33 by theorem [ thm horizontal cilfford bundle from submersion ] . @xmath74 * remark : * for an element @xmath299 , where @xmath57 is the lie algebra of the carnot group @xmath33 , we can define @xmath969 this is the so - called _ right - regular representation _ of the lie algebra @xmath57 on @xmath950 adopted from @xmath221 . using this representation @xmath970 , we can rewrite in the form @xmath971
@xmath74 * proof : * the local expression of the horizontal pull - back dirac operator follows immediately from the expression of the directional differentiation along a vector field @xmath4 on a lie group , which is @xmath972 the statement about the invariance is straight forward .
@xmath91 the idea , which has been used by christian br and bernd amman for the classical dirac operator on compact nilmanifolds of heisenberg groups and which we will adopt to our situation , is now to find a direct sum decomposition of the horizontal clifford bundle @xmath964 , which is invariant under the horizontal dirac operator .
thus the determination of the spectrum of @xmath1 splits into parts which are easier to handle .
remember that by from theorem [ thm horizontal cilfford bundle from submersion ] we can consider elements of @xmath964 as functions on @xmath33 which have certain periodicity properties .
we intend to use these properties to decompose @xmath964 . in the language of representation theory and in view of proposition [ prop expression of horizontal dirac via representation ]
this is just the decomposition of the unitary right - regular representation @xmath221 , acting on the hilbert space @xmath964 , into its irreducible components .
let @xmath34 be the grading of the lie algebra belonging to @xmath33 .
we start with the decomposition of @xmath964 using the periodicities arising from the subgroup @xmath973 of @xmath33 .
[ def carnot center ] for the grading @xmath34 of the carnot group @xmath974 , we call @xmath975 the _ carnot center _ of @xmath33
. @xmath74 * remark : * note that the center of a general lie group @xmath976 is the maximal ( normal ) subgroup of @xmath976 commuting with every element of @xmath976 . in the case
@xmath33 is a carnot group , the center of @xmath33 needs not to coincide with the carnot center of @xmath33 defined above .
consider for example the carnot group @xmath977 for @xmath978 , where the components belonging to @xmath979 belong to the center but not to the carnot center of @xmath33 .
@xmath74 now the decomposition of @xmath950 from the carnot center works as follows .
in addition , we will detect that the horizontal spinor space of a compact carnot nilmanifold of lower step can be found in this decomposition .
[ thm decomposition from center ] let @xmath47 be the compact nilmanifold of a carnot group @xmath33 with center @xmath980 , and let @xmath923 be the horizontal clifford bundle arising via pull - back from a spin structure @xmath927 on @xmath49 .
then there is a decomposition @xmath981 into hilbert spaces @xmath982 , which are invariant under the pull - back horizontal dirac operator @xmath1 acting on @xmath923 .
the elements @xmath983 are exactly those elements @xmath915 fulfilling @xmath984 in addition , the space @xmath985 is isomorphic to the space @xmath986 , where @xmath987 is the compact nilmanifold of the carnot group @xmath988 of step @xmath989 . under this isomorphism
, the restriction of the pull - back horizontal dirac operator @xmath1 to @xmath985 can be identified with the pull - back horizontal dirac operator @xmath544 acting on @xmath990 .
* proof : * throughout this proof we will use the characterization of a horizontal spinor @xmath991 as a map @xmath992}}$ ] such that @xmath993 for every @xmath994 given by theorem [ thm horizontal cilfford bundle from submersion ] .
note that this @xmath556 , considered as a periodic function on @xmath33 , is an @xmath52-function on any fundamental domain for the action of @xmath46 on @xmath33 by left translation .
we will work with the projection @xmath995 .
the carnot center of @xmath33 can be described via ( exponential or polarized ) coordinates by @xmath996 where @xmath309 is the grading of the lie algebra @xmath57 of @xmath33 with @xmath997 .
the image @xmath998 of @xmath999 has the structure of a carnot group of step @xmath989 : this follows from proposition [ prop properties projection lie algebra ] , using the inherited group composition from @xmath33
. we can realize @xmath442 the subset @xmath1000 of @xmath33 , which leads to the coordinate expression @xmath1001 of @xmath999 . in this description , the group composition on @xmath442 is given by executing the group composition on @xmath33 first and projecting the result to @xmath442 afterwards , i.e. @xmath1002 for the rest of the proof we will work with this coordinate description .
in addition , the carnot center of the standard lattice @xmath845 of @xmath33 is given by @xmath1003 hence , and since the generators of the lie algebra of @xmath33 can be identified with the generators of the lie algebra of @xmath442 via @xmath999 , the discrete carnot group @xmath1004 is the standard lattice of @xmath988 .
this means that we can form the local homogeneous space @xmath1005 , which is a compact carnot nilmanifold of @xmath442 . from the projection
@xmath995 we see that @xmath0 has the structure of a principle @xmath1006-bundle over @xmath1007 , since @xmath1006 is the local homogeneous space of @xmath1008 under the action of @xmath1009 . after this preparation
, we can finally prove the statements of the theorem .
we fix a point @xmath1010 . then for any @xmath1011
, we can define a map @xmath1012 via @xmath1013 } } , \ \ \ z \mapsto \sigma\left(y ._{\mathbb{g}}(0 , \ldots , 0 , z ) \right).\ ] ] as one can calculate using the baker - campbell - hausdorff formula , we have for each @xmath1014 @xmath1015 since @xmath1016 lies in the carnot center of @xmath33 , which means it commutes with any other element of @xmath33 .
thus , since @xmath1017 for any homomorphism @xmath125 characterizing a spin structure @xmath927 on @xmath49 ( see theorem [ thm spin structures on t^d ] ) , provides us the periodicity @xmath1018 for each @xmath1019 .
furthermore , @xmath1012 is an @xmath52-section on @xmath1020 ( or on a corresponding fundamental domain of @xmath1021 on @xmath1022 , which is any cube of edge length @xmath98 ) : this follows from the fact that @xmath0 is a @xmath1006-principle bundle over @xmath1007 and hence integration over @xmath0 is integration over the fiber @xmath1006 followed by integration over the base space @xmath1007 .
but since @xmath556 is an @xmath52 section on @xmath0 , it must also be an @xmath52-section on @xmath1023 .
together with the periodicity this means that we can develop @xmath1012 into a fourier series @xmath1024 with fourier coefficients @xmath1025 .
if we choose @xmath1026 in and , we find @xmath1027 for any @xmath1028 , which proves the decomposition .
to find an explicit description of the elements of the spaces @xmath982 , we use the expression of the fourier coefficients from via the integral @xmath1029^{d_r } } \sigma\left(y.(0 , \ldots , 0 , t)\right ) e^{-2\pi i \langle \tau , t \rangle } { \ensuremath{\ , \mathchar'26\mkern-12mu d}}^{d_r}t.\ ] ] assume @xmath983 , which means @xmath1030 by .
let the coordinates of @xmath33 belonging to the carnot center be denoted by @xmath1014 , which means we set @xmath1031 .
since we have @xmath1032 from the group rule in @xmath33 , we have the following calculation using the substitution @xmath1033 in the integral and the periodicity of @xmath556 : @xmath1034^{d_r } } \sigma\left((y^{(1 ) } , \ldots , y^{(r-1 ) } , 0 ) .
( 0 , \ldots , 0 , z+t)\right ) e^{-2\pi i \langle \tau , t \rangle } { \ensuremath{\ , \mathchar'26\mkern-12mu d}}^{d_r}t\\ & = & e^{2\pi i \langle \tau , z \rangle } \int_{[z_1,z_1 + 1 ] \times \ldots \times [ z_{d_r } , z_{d_r}+1 ] } \sigma \left((y^{(1 ) } , \ldots , y^{(r-1 ) } , 0).(0 , \ldots , 0 , u ) \right)e^{-2 \pi i \langle \tau , u \rangle } { \ensuremath{\ , \mathchar'26\mkern-12mu d}}^{d_r}u \\ & = & e^{2\pi i \langle \tau , z \rangle } \int_{[0,1]^{d_r } } \sigma \left((y^{(1 ) } , \ldots , y^{(r-1 ) } , 0).(0 , \ldots , 0 , u ) \right)e^{-2 \pi i \langle \tau , u \rangle } { \ensuremath{\ , \mathchar'26\mkern-12mu d}}^{d_r}u\\ & = & e^{2\pi i \langle \tau , z \rangle } \cdot \sigma \left(y^{(1 ) } , \ldots , y^{(r-1 ) } , 0\right ) , \end{aligned}\ ] ] where we used the identity once again in the last line . but from this we get the description for any element @xmath983 . on the other hand , for any @xmath991 fulfilling @xmath1035 we fix @xmath1028 and write down the fourier series of the corresponding function @xmath1036 from .
for any @xmath1037 we calculate the fourier coefficients @xmath1038 from : @xmath1039^{d_r } } \sigma\left(y.(0 , \ldots , 0 , t)\right ) e^{-2\pi i \langle \tau',t \rangle } { \ensuremath{\ , \mathchar'26\mkern-12mu d}}^{d_r}t \\ & = & \int_{[0,1]^{d_r } } e^{2\pi i \left\langle \tau , y^{(r ) } + t \right\rangle } \cdot \sigma \left(y^{(1 ) } , \ldots , y^{(r-1 ) } , 0\right ) e^{-2\pi i \langle \tau',t \rangle } { \ensuremath{\ , \mathchar'26\mkern-12mu d}}^{d_r}t \\ & = & e^{2\pi i \left\langle \tau , y^{(r ) } \right\rangle } \cdot \sigma \left(y^{(1 ) } , \ldots , y^{(r-1 ) } , 0\right ) \int_{[0,1]^{d_r } } e^{2\pi i \left\langle \tau - \tau ' , t \right\rangle } { \ensuremath{\ , \mathchar'26\mkern-12mu d}}^{d_r}t \\ & = & \begin{cases } \sigma(y ) & \text{for } \ \tau = \tau ' \\ 0 & \text{otherwise } \end{cases}.\end{aligned}\ ] ] this shows immediately that we must have @xmath983 .
using this description , we also see the invariance under @xmath1 : by proposition [ prop expression of horizontal dirac via representation ] we have to check the invariance of @xmath983 under the clifford action of @xmath19 and the right - regular representation @xmath221 .
while the first invariance is trivial , the invariance under @xmath221 follows from a small calculation : let @xmath983 .
for an @xmath966 we set @xmath1040 because of the baker - campbell - hausdorff formula we have in ( exponential or polarized ) coordinates @xmath1041 where @xmath1042 is a polynomial in the coordinates of @xmath1043 and @xmath1044 ( see section 2.2 )
. then we have because of and @xmath1045 since @xmath1042 does not depend on the components @xmath1046 and @xmath1047 .
this shows @xmath1048 and therefore the invariance of @xmath982 under @xmath221 . the next step is to show that @xmath1049 : first of all , the horizontal distributions of the compact carnot nilmanifolds @xmath47 and @xmath1050 can be identified using the carnot group homomorphism @xmath995 from , which means that for a given spin structure @xmath881 on the horizontal torus @xmath49 we find horizontal spinor bundles @xmath923 and @xmath990 on both manifolds . now for any @xmath1051 we find can define @xmath1052 which is an element of @xmath985 : if we consider @xmath1053 as the subset of @xmath33 , we have @xmath1054 , which is exactly the characterization of the space @xmath985 from . and since elements of @xmath985 are uniquely determined by its values on @xmath1053 , the map @xmath1055 is an isomorphism .
finally , we show that the restriction of the pull - back horizontal dirac operator @xmath1 on @xmath950 to @xmath985 can be identified with the pull - back horizontal dirac operator @xmath544 on @xmath1056 , coming from the same spin structure @xmath881 on the horizontal torus @xmath49 .
note that by the characterization of @xmath442 as a subset of @xmath33 , a horizontal frame @xmath41 of @xmath19 is also a horizontal frame of @xmath1057 , and since the horizontal spinor bundles on @xmath0 and @xmath1007 are constructed from the same horizontal torus @xmath49 the horizontal clifford action from elements of this frame coincides in both cases .
then the identification of the operators follows immediately from the isomorphism when we use the local expression of the pull - back horizontal dirac operator from proposition [ prop expression of horizontal dirac via representation ] : we have for every @xmath1058 @xmath1059 and since @xmath985 is invariant under @xmath1 this shows @xmath1060 for the isomorphism @xmath507 from .
altogether every statement of the theorem is proved .
* remark : * note that the coordinate expressions we used in the proof do not depend on whether we choose exponential or polarized coordinates on the carnot group @xmath33 .
this is the case since only actions by the carnot center are involved . in the next section
, it will turn our to be more comfortable to work with polarized coordinates . @xmath74
the above theorem reduces the problem of showing that the horizontal dirac operator has an infinite dimensional eigenspace to the corresponding horizontal dirac operator on a carnot group of a lower step .
we are interested in using this theorem inductively until @xmath1007 is the compact nilmanifold of a carnot group of step @xmath53 .
as we will see in the next section , for a compact heisenberg nilmanifold @xmath1007 we will be able to do a complete spectral decomposition of @xmath1061 and detect infinite dimensional eigenspaces from this . then in section 4.4
, we will use theorem [ thm decomposition from center ] to lift these infinite dimensional eigenspace to compact nilmanifolds of general carnot groups .
we will now do the complete spectral decomposition of the horizontal pull - back dirac operator on compact nilmanifolds @xmath47 , where @xmath1062 is a carnot group of step @xmath53 and rank @xmath216 for some integer @xmath1063 .
hence the horizontal distribution is of co - dimension @xmath98 . for our spectral decomposition
we will follow an argument of christian br ( see section ii.2 of his phd thesis @xcite ) , which was used to determine the spectrum of dirac operators on nilmanifolds from heisenberg groups .
this method was also used by christian br and bernd ammann in @xcite for the case of the dirac operator on the @xmath245-dimensional heisenberg group .
although the situation in case of our horizontal dirac operator differs a bit , the space @xmath950 can be decomposed in exactly the same way since our local expression of @xmath1 from is quite similar to the local expression of the dirac operator br is considering . in this thesis , we have to consider a slightly more general case where @xmath1062 , but the commutative part of the group will not effect the general strategy too much . after doing the decomposition
, we will be able to detect infinite - dimensional eigenspaces of @xmath1 ; and as an additional result we will show that the asymptotic behaviors of the non - degenerate eigenvalues gives back the homogeneous dimension of @xmath0 .
we start by introducing some notation .
let the grading of @xmath20 be given by @xmath373 , where @xmath41 is an orthonormal frame of @xmath376 such that the levi form according to this frame is given by @xmath1064 note that this can always be achieved by proposition [ prop tangent lie group bundle and rank of levi form ] and that the numbers @xmath398 are exactly the absolute values of the non - zero eigenvalues @xmath1065 of the levi form of @xmath0 . if @xmath378 is spanned by @xmath1066 , this means that we have the commutator relations @xmath1067
= \begin{cases } \lambda_j x_{d+1 } & \text{for } \ 1 \leq j \leq m , \
k = m+j\\ -\lambda_j x_{d+1 } & \text{for } \
1 \leq k \leq m , \
j = k+m \\ 0 & \text{otherwise } \end{cases}.\ ] ] in what follows , it will be more comfortable to use the polarized coordinates instead of the exponential coordinates of @xmath33 ( see definition [ def coordinates of carnot groups ] ) .
we will use polarized coordinates with respect to the frame @xmath1068 of @xmath20 , where @xmath1069 and @xmath1070 otherwise .
we denote these polarized coordinates on @xmath33 by @xmath1071 with @xmath1072 , @xmath1073 , @xmath1074 and @xmath1075 , where @xmath1076 obviously , @xmath1077 denotes the product according to the group composition @xmath814 of @xmath33 . in other words ,
the coordinates @xmath1078 describe the @xmath25-part of @xmath33 and @xmath1079 describes the @xmath979-part of @xmath33 . now let us calculate the composition rule on @xmath33 in these coordinates . in exponential coordinates corresponding to the frame @xmath1068 we set for abbreviation @xmath1080 and @xmath1081 such that we have because of @xmath1082 & = & \sum_{j=1}^m \left ( a_jy_j \cdot \frac{1}{\lambda_j } [ x_j , x_{m+j } ] + b_jx_j \cdot \frac{1}{\lambda_j } [ x_{m+j},x_j ] \right ) \\ & = & \left(\sum_{j=1}^m a_jy_j - b_jx_j \right ) \tilde{x}_{d+1}.\end{aligned}\ ] ] by the baker - campbell - hausdorff formula ( see section 2.2 ) ,
we calculate @xmath1083 \right ) \\ & = & \exp \left ( \sum_{j=1}^m ( a_j+x_j ) \tilde{x}_j + \sum_{j=1}^{m } ( b_j+y_j ) \tilde{x}_{m+j } + \sum_{j=1}^{d-2 m } ( c_j+z_j ) \tilde{x}_{2m+j } \right .
\times \\ & & \ \ \ \ \ \ \ \times \left .
+ \frac{1}{2}\left(s+t + \sum_{k=1}^m ( a_ky_k - b_kx_k ) \right ) { x}_{d+1 } \right)\\ & = & \left ( a+x , \
c+z , \ s+t + \frac{1}{2}\sum_{k=1}^m ( a_ky_k - b_kx_k ) \right)\end{aligned}\ ] ] for the composition rule on @xmath33 in exponential coordinates .
we can now use the isomorphism between the exponential and the polarized coordinates on a heisenberg group ( see the remark after definition [ def coordinates of carnot groups ] ) and conclude that for the composition rule in coordinates according to the frame @xmath1068 we have @xmath1084 using this composition rule , we can start with the spectral decomposition of @xmath950 .
the first decomposition of @xmath964 is already given by theorem [ thm decomposition from center ] .
in the situation @xmath1062 we have @xmath1085 which is invariant under @xmath1 according to the theorem , where elements of @xmath982 are identified by the relation @xmath1086 we set @xmath1087 considered as a function on @xmath946 .
note that @xmath1088 must be an @xmath52-function on the torus @xmath49 since @xmath556 is an @xmath52-function on @xmath0 , and that any @xmath983 is fully determined by @xmath1088 .
the strategy is to look for periodicities of @xmath1088 in order to find a further decomposition of the spaces @xmath982 .
these periodicities can be detected from the lemma below .
[ lemma periodicity f_sigma ] let @xmath893 be the group homomorphism characterizing the spin structure @xmath881 on @xmath49 from theorem [ thm spin structures on t^d ] , which induces the horizontal spinor bundle @xmath923 on the compact heisenberg nilmanifold @xmath0 . for @xmath1089
, we assume @xmath983 according to the decomposition of @xmath950 .
then for any @xmath1090 we have @xmath1091 * proof : * we use the periodicities of @xmath556 from of theorem [ thm horizontal cilfford bundle from submersion ] and the composition rule to calculate for every @xmath1090 : @xmath1092 thereby , in the third equation we have used the characterization of @xmath1093 .
@xmath91 from this lemma , we can detect further decompositions of the spaces @xmath1094 .
we start with the case @xmath1095 : [ lemma spectral decomposition h_0 ] for the space @xmath985 from the decomposition of @xmath950 we have the further decomposition @xmath1096 with @xmath1097}}$ ] for every @xmath558 , where each of the spaces @xmath1098 is invariant under the horizontal pull - back dirac operator @xmath1 on @xmath0 .
* proof : * let @xmath1058 , which means we have by and @xmath1099 with @xmath1088 as in lemma [ lemma periodicity f_sigma ] .
if we set @xmath1095 in , we immediately see that @xmath1088 is @xmath1100-periodic , and therefore can be developed into a fourier series @xmath1101 with @xmath1102}}$ ] . now
the homomorphism @xmath893 from is given by @xmath1103 see theorem [ thm spin structures on t^d ] , such that lemma [ lemma periodicity f_sigma ] provides us further @xmath1104 for every @xmath1090 .
but this leads to the constraint that for every @xmath529 we must have @xmath1105 and therefore @xmath1106 .
this means that we have @xmath1107 and so the fourier series of @xmath1088 becomes @xmath1108 with @xmath1109}$ ] .
therefore we have the decomposition @xmath1110 where each @xmath1098 is spanned by @xmath1111 and hence isomorphic to @xmath916}}$ ] .
now we immediately see that each @xmath1098 is invariant under the clifford action of any @xmath1112 ( which only acts on the coefficient @xmath1113 ) and also under the right regular representation of @xmath33 : since @xmath1114 , we have for every given @xmath1115 @xmath1116 with @xmath1109}$ ] . but
this shows that @xmath1117 , and altogether every statement of the lemma is proved .
* remark : * since @xmath1118 , we already know by theorem [ thm decomposition from center ] that the space @xmath985 is isomorphic to the space @xmath1119 , which is by construction exactly the space @xmath896 , and the horizontal pull - back dirac operator acting on @xmath896 is exactly the classical dirac operator @xmath1120 on the torus . since the spectral decomposition of the dirac operator on the torus is well - known , we could have deduced the statement of this lemma directly from theorem [ thm decomposition from center ] .
@xmath74 for the cases @xmath1121 the spectral decomposition is a bit more involved .
[ lemma spectral decomposition h_tau , tau neq 0 ] let @xmath1121 .
then for the space @xmath982 from the decomposition of @xmath950 we have the decomposition @xmath1122 with @xmath1123}})$ ] , where each of the spaces @xmath1124 is invariant under the horizontal pull - back dirac operator @xmath1 on @xmath0 .
* proof : * once again we use the identity for @xmath1088 from lemma [ lemma periodicity f_sigma ] , looking for periodicities in the case @xmath1121 .
we write @xmath1125 with @xmath1126 , @xmath1127 and @xmath1128 to distinguish the periodicities which belong to @xmath107 , @xmath108 and @xmath1079
. remember that we have @xmath1062 for our carnot group @xmath33 , such that the coordinates @xmath107 and @xmath108 ( together with the coordinate from the center ) belong to the heisenberg part and @xmath1079 forms the ( commutative ) @xmath979-part of @xmath33 . from we see immediately that for any @xmath1128 we have @xmath1129 with @xmath1130 describing the spin structure of @xmath49 from which @xmath1 is constructed .
therefore @xmath1088 is @xmath1131-periodic in @xmath1079 and after fixing @xmath107 and @xmath108 we can develop @xmath1088 in a fourier series @xmath1132 where @xmath1133 is a @xmath916}}$]-valued function .
we note that @xmath1134 must be an @xmath52-function on @xmath1135 , since @xmath1088 is an @xmath52-function on @xmath1136 . as we did before we can put further conditions on the indices @xmath44
over which we sum up in : for any @xmath1137 we have @xmath1138 and thus from lemma [ lemma periodicity f_sigma ] we can deduce the restriction @xmath1139 for every @xmath1137 by , which leads to the constraint that for all @xmath1140 we must have @xmath1141 and therefore @xmath1142 .
but this means that we have @xmath1143 and so the fourier series of @xmath1088 becomes @xmath1144 in equation we also see that we have the same periodicity properties for @xmath1088 in the variable @xmath108 ( since the @xmath1145-components of @xmath1146 do not appear in the exponent of @xmath137 ) , so we can repeat the above argument to show that @xmath1147 which gives us @xmath1148 together with .
but therefore we can develop @xmath1134 into a fourier series with respect to @xmath108 , and after taking care of the restrictions arising from the components @xmath1149 of the spin structure similarly to the way we did above we find that @xmath1150 for a @xmath916}}$]-valued function @xmath1151 on @xmath1152 .
plugging this into , we can write @xmath1153 we continue to find periodicities for @xmath1088 by considering dilations by @xmath1126 of the @xmath107 variable .
this will lead to a further characterization of the functions @xmath1151 appearing in .
let @xmath1093 .
we use the the composition rule on @xmath33 , the characterization of elements belonging to @xmath1094 and the fourier series development to calculate for any @xmath1126 @xmath1154 where the summation over @xmath44 and @xmath1145 is given as in .
now , since @xmath1155 and therefore @xmath1156 , we can use the substitution @xmath1157 in the last equation of the above calculation and find @xmath1158 because of @xmath1159 for every element @xmath915 this means @xmath1160 on the other hand we have because of and @xmath1161 and together with this leads to the identity @xmath1162 or , equivalently , @xmath1163 for all @xmath1126 .
but this means that for a given @xmath44 we have altogether @xmath1164 independent functions @xmath1151 which characterize @xmath1094 .
now we can choose @xmath1164 independent functions which determine @xmath983 completely .
after fixing a @xmath1165 , we consider the set @xmath1166 where @xmath1167 are the generators of @xmath1168 .
obviously we have @xmath1169 , and after enumerating the elements of @xmath444 by @xmath1170 we have the identity @xmath1171 for the index set over which the @xmath1145 s are summated .
note that the union on the right hand side of this identity is disjoint , and after using the property we see that for every @xmath1126 we have @xmath1172 plugging this into the expression for elements @xmath983 , we see that every @xmath983 is uniquely determined by the functions @xmath1173 via @xmath1174 we further make use of the fact that for any @xmath44 @xmath1133 is an @xmath52-function on @xmath1135 to derive that @xmath1175}})$ ] .
we can argue using the expression as follows : for any @xmath44 we have @xmath1176^{2 m } } \left| a_\gamma^\sigma ( x , y ) \right|^2 dy dx & = & \int_{[0,1]^m } \int_{[0,1]^m } \left| \sum_{j=1}^{|\tau|^m } \sum_{\xi \in { \mathbb{z}}^m } \varepsilon(\xi,0,0 ) e^{2\pi
i \langle \beta_j + \tau \xi , y \rangle } \right|^2 dy dx\\ & = & \int_{[0,1]^m } \int_{[0,1]^m } \sum_{j=1}^{|\tau|^m } \sum_{\xi \in { \mathbb{z}}^m } \left| b_{\gamma,\beta_j}^\sigma(x + \xi ) \right|^2 dy dx\\ & = & \sum_{j=1}^{|\tau|^m } \int_{[0,1]^m } \sum_{\xi \in { \mathbb{z}}^m } \left| b_{\gamma,\beta_j}^\sigma(x + \xi ) \right|^2 dx\\ & = & \sum_{j=1}^{|\tau|^m } \int_{{\mathbb{r}}^m } \left| b_{\gamma,\beta_j}^\sigma(x ) \right|^2 dx,\end{aligned}\ ] ] and from the last line we can conclude @xmath1175}})$ ] . since elements of @xmath982 are fully characterized by the values of @xmath1177 from the above calculation it furthermore follows that there is an isometry @xmath1178}}).\ ] ] but this shows the decomposition stated in the theorem .
to finish the proof , we have to show that this decomposition is invariant under @xmath1 . while once again the invariance under the clifford action is trivial , the invariance under the right regular representation @xmath221 follows after a simple calculation .
remember that in the polarized coordinates we are using on @xmath33 , for a given point @xmath1179 the right regular action on @xmath950 is given via @xmath1180 see .
now since @xmath983 , we can use the characterization of @xmath556 and find that @xmath1181 but since for @xmath1182}})$ ] we also have @xmath1183}})$ ] , this proves the invariance of this decomposition under @xmath221 , which means that it is also invariant under @xmath1 .
hence the lemma is proved.@xmath91 the arguments for the decomposition of @xmath950 used so far ( except for the small modification of including the commutative part @xmath979 ) are exactly the arguments br uses to decompose the ( not horizontal ) spinor bundle on @xmath0 in @xcite .
but now we are ready to make some interesting observation for our situation by looking at the eigenvalues resulting from this decomposition .
we can calculate all eigenvalues of @xmath1 , and first of all we will see that @xmath1 does not have a compact resolvent , and therefore does not provide a spectral triple .
[ thm spectral decomposition d^h on heisenberg group ] let @xmath1062 be a carnot group of step @xmath53 , rank @xmath216 and co - rank @xmath98 of its horizontal distribution with compact nilmanifold @xmath47 .
the non - zero eigenvalues of the levi form of @xmath33 shall be given by the numbers @xmath1184 with @xmath629 for @xmath249 .
let @xmath1 be a horizontal pull - back dirac operator on @xmath0 arising via pullback from a spin structure @xmath881 with @xmath1185 on the torus @xmath49 .
then we have the following statements about the spectrum of @xmath1 : a. in any case , the spectrum of @xmath1 is discrete .
the absolute values of the eigenvalues of @xmath1 are of the types a. @xmath1186 of multiplicity @xmath888}$ ] for @xmath1187 such that @xmath1188 .
b. @xmath1189 of multiplicity @xmath888 } |\tau|^m$ ] for @xmath1190 , @xmath1191 and @xmath1192 such that @xmath1142 .
b. in the case @xmath1193 , @xmath1 has an infinite dimensional kernel , and any eigenspace belonging to an eigenvalue @xmath1194 of @xmath1 is finite dimensional . c. in the case @xmath1195 , there are infinitely many eigenvalues of @xmath1 which have an infinite dimensional eigenspace . in particular , since @xmath1 has in any case an infinite - dimensional eigenspace , it does not have a compact resolvent . *
proof : * the strategy is to use the decomposition of @xmath964 we get from theorem [ thm decomposition from center ] , lemma [ lemma spectral decomposition h_0 ] and lemma [ lemma spectral decomposition h_tau , tau neq 0 ] , which is invariant under @xmath1 .
then we can use proposition [ prop expression of horizontal dirac via representation ] to write @xmath1 locally in the form @xmath1196 to shorten notation in what follows , we set @xmath1197 we will keep on working in the polarized coordinates according to the frame @xmath1068 from , such that the group composition on @xmath33 is given by .
by we have @xmath1198 for all @xmath1199 , and the representation of @xmath1 with respect to the frame @xmath1068 becomes @xmath1200 since each of the spaces @xmath1098 from lemma [ lemma spectral decomposition h_0 ] and @xmath1201 from lemma [ lemma spectral decomposition h_tau , tau neq 0 ] is invariant under the right regular representation @xmath221 , this decomposition also provides a decomposition of @xmath221 into its irreducible parts . for @xmath1095
these will be the trivial representations , which can be calculated directly , and for @xmath1121 we can make use of the fact that the representation theory for the heisenberg group is fully known to detect the corresponding infinite dimensional representations . for @xmath1114 , which means by @xmath1202 with @xmath1102}}$ ]
, we calculate from ( using the notation @xmath1203 instead of @xmath1071 for the coordinates of @xmath33 ) by @xmath1204 this shows that the spectrum of @xmath1 restricted to each @xmath1098 consists of the eigenvalues of the matrices @xmath1205 . to calculate the absolute values of these eigenvalues
, we calculate the eigenvalues of @xmath609 restricted to @xmath1098 : @xmath1206 since @xmath1207 for @xmath532 and @xmath1208 by the rules of the clifford action . but
this shows that on each @xmath1098 the absolute values of the eigenvalues of @xmath1 are given by @xmath1209 each one with multiplicity @xmath888}$ ] because @xmath923 is a vector bundle of rank @xmath888}$ ] .
note that this also shows that the part of the spectrum belonging to @xmath985 coincides with the spectrum of the dirac operator @xmath1120 on the torus @xmath49 ( equipped with a riemannian metric such that @xmath879 is orthonormal ) , from which @xmath1 was constructed .
this can also be deduced from theorem [ thm decomposition from center ] . to determinate the spectrum on the spaces @xmath1123}})$ ] , we use results from the representation theory of the heisenberg group .
it is known that @xmath1210 is the representation space of irreducible unitary representations of @xmath25 , and what we have done so far is to decompose the right regular representation @xmath1211 into its irreducible components .
thereby , the frame @xmath1068 exactly corresponds to the carnot group @xmath393 , while the frame @xmath1212 only corresponds to a carnot group isomorphic to @xmath393 .
now it is well known by the theorem of stone and von neumann how they have to look like on @xmath1210 ( see e.g. @xcite , theorem ( 1.50 ) ) : in the polarized coordinates we are using , the infinite dimensional unitary irreducible ( schrdinger ) representations of @xmath25 are given by @xmath1213 for any @xmath1214 . in our case , for @xmath1215}}$ ] , we have @xmath1216 .
from now on , we write @xmath63 instead of @xmath556 whenever we consider an element of @xmath1217}})$ ] .
hence using the expression of the horizontal dirac operator via the right - regular representation @xmath221 we have , after plugging in , @xmath1218 we still need to express the action of @xmath1219 for @xmath1220 on @xmath1201 .
but this is just multiplication by @xmath1221 , since @xmath983 is characterized via @xmath1222 ( see in the proof of lemma [ lemma spectral decomposition h_tau , tau neq 0 ] ) , and therefore we get @xmath1223 plugging everything into the expression we see that the action of @xmath1 on every @xmath1224}}$ ] is given by @xmath1225 to calculate the absolute values of the eigenvalues of @xmath1 restricted to @xmath1201 , we once again consider the square of the operator .
from we get @xmath1226 and because of the clifford relation @xmath1227 for any @xmath1228 , the sum of the third and the forth term on the right hand side of vanish . for the same reason , and since @xmath1229 for any @xmath1230 , we get @xmath1231 for the second term .
finally , using again the rules of clifford action and the leibniz rule for differentiation , we calculate @xmath1232 and @xmath1233 because of the identity @xmath1234 for any @xmath1235 , we have @xmath1236 and thus get for the first term of from these two calculations @xmath1237 now after plugging everything into the expression for the restriction of @xmath609 onto the space @xmath1123}})$ ] we see that @xmath1238 on these spaces . from
we observe that we can calculate the eigenvalues of @xmath1 from the eigenvalues of the harmonic oscillator .
it is well known that the operator @xmath1239 possesses the eigenvalues @xmath1240 , where @xmath1241 ( see e.g. @xcite , section 1.7 ) .
since this operator is acting on the @xmath52-space of @xmath1242}$]-valued functions , each of these eigenvalues has the multiplicity @xmath888}$ ] .
in addition , for each @xmath1190 we have to add the eigenvalues of the operator @xmath1243 by proposition [ prop eigenvalues sum c(x_j)c(x_k ) ] from section 3.2 , every eigenvalue @xmath1244 , @xmath1245}\}$ ] of this @xmath661 } \times 2^{[d/2]})$]-matrix is included in the set @xmath1246 the third term of is just an additive constant ( since @xmath44 is fixed by choosing a space @xmath1124 ) , and hence on each @xmath1201 the operator @xmath609 from has the eigenvalues @xmath1247 with @xmath1190 and @xmath1248 for @xmath1249}\}$ ] .
but since each number @xmath1250 is bounded by constants depending only on the constants @xmath398 which are given by the levi form on @xmath0 , they do not change the asymptotic behavior of the eigenvalues from . therefore we can say that @xmath1251 where each @xmath1252 has the multiplicity @xmath888}$ ] , for the eigenvalues of @xmath609 restricted to @xmath1201 . from
we get for the absolute values of the eigenvalues of @xmath1 @xmath1253 for @xmath1191 , @xmath1254 such that @xmath1255 and @xmath1256 .
each of these ( asymptotic ) eigenvalues has the multiplicity @xmath888 } |\tau|^m$ ] , since there are @xmath1164 copies of the spaces @xmath1257}})$ ] .
thus we have proved statement ( a ) .
we still have to show the statements ( b ) and ( c ) about the degeneracy of @xmath1 .
this will be done by showing that for each @xmath1191 , the first term of the operator @xmath609 from has at least one @xmath596-eigenvalue on @xmath1201 : from that it will follow that for every @xmath1128 the number @xmath1258 appears as an eigenvalue on each space @xmath1201 , and is thus an eigenvalue of infinite multiplicity .
furthermore all the degenerate eigenvalues of @xmath1 are of this form , since every other eigenvalue on @xmath1259 depends on @xmath1260 and is of finite multiplicity on this space .
this shows statement ( c ) , and since for @xmath64 the term @xmath1261 does not appear in the expression of @xmath1 on @xmath1259 the statement ( b ) also follows .
indeed the first term @xmath1262 of has at least one @xmath596-eigenvalue .
as we already noted , the eigenvalues of the harmonic oscillator are the numbers @xmath1263 where @xmath1241 , and the eigenvalues of the matrix @xmath1264 belong to the set @xmath1265 hence , the only chance for the operator to have a @xmath596-eigenvalue is that @xmath1266 and that the number @xmath1267 is indeed an eigenvalue of @xmath1268 .
but the last statement is true by proposition [ prop eigenvalues sum c(x_j)c(x_k ) ] from section 3.2 , and hence the above argumentation proves the statements ( b ) and ( c ) of the theorem.@xmath91 as mentioned before , the results of theorem [ thm spectral decomposition d^h on heisenberg group ] imply that @xmath1 does not furnish a spectral triple on @xmath47 .
but it fits into our definition of a degenerate spectral triple and it detects the carnot - carathodory metric of our carnot manifold @xmath0 , which we write down in the following corollary .
[ cor d^h gives degenerate spectral triple ] let @xmath47 be the compact nilmanifold of a carnot group @xmath1062 .
let @xmath1 be a horizontal dirac operator on @xmath0 arising via pull - back from a spin structure @xmath881 on the torus @xmath49 .
then the triple @xmath1269 is a degenerate spectral triple , which detects the carnot - carathodory metric via connes metric formula .
* proof : * since @xmath1 is a horizontal dirac operator in the sense of section 3.2 , it has been shown in section 3.3 that it detects the carnot - carathodory metric on @xmath0 via connes metric formula .
in particular it has also been shown that @xmath1270\| = \mathrm{ess } \sup_{x \in m } \left\|{\mathrm{grad}}^h f \right\|\ ] ] for every @xmath714 , and since the number on the right hand side is bounded and @xmath713 is a dense sub - algebra of @xmath45 , condition ( i ) for a spectral triple ( see definition [ def spectral triple ] ) is fulfilled .
now from theorem [ thm spectral decomposition d^h on heisenberg group ] we know that the spectrum of @xmath1 is discrete . if we exclude the eigenvalues of infinite multiplicity , then for a given number @xmath1271 there are only finitely many eigenvalues in the spectrum of @xmath1 which are smaller than @xmath130 .
this shows us that @xmath1 has a compact resolvent if we restrict it to the orthonormal complement of the degenerate eigenspaces , which means we have a degenerate spectral triple according to definition [ def degenerate spectral tripel ] .
@xmath91 as a further consequence of theorem [ thm spectral decomposition d^h on heisenberg group ] , we show that besides the carnot - carathodory metric we can also detect the hausdorff dimension of the metric space @xmath231 from the degenerate spectral triple @xmath1272 .
therefore we consider the asymptotic behavior of the non - degenerate eigenvalues of @xmath1 .
to detect the asymptotic behavior in our case , we will make use of the following proposition , which shows the equivalence about the growth function of the number of eigenvalues within a certain radius and their asymptotic growth .
we write down the statement in a very general matter , since for its proof it is not important what exactly the elements of the sets whose growth rate we analyze describe . for a proof
, we refer to @xcite , proposition 13.1 .
[ prop asymptotic behaviour versus growth rate ] let @xmath1273 be a discrete ordered set of positive real numbers , which means we have @xmath1274 . for @xmath1275 , we denote by @xmath1276 the number of all elements of @xmath130 which are bounded by @xmath746 .
in addition , we consider numbers @xmath1277 and @xmath1278
. then the following statements are equivalent : a. @xmath1279 as @xmath1280 . b. @xmath1281 as @xmath1282 .
@xmath91 using this proposition , we can show that the hausdorff dimension on @xmath22 coincides with the metric dimension of the degenerate spectral triple @xmath1283 .
[ cor hausdorff dimension from d^h ] let @xmath64 be isomorphic to the @xmath257-dimensional heisenberg group , and let @xmath47 be its compact nilmanifold .
let @xmath1 be a horizontal dirac operator on @xmath0 arising via pull - back from a spin structure @xmath881 on the torus @xmath49 with @xmath624 .
then the metric dimension of the degenerate spectral triple @xmath1284 is @xmath1285 , i.e. it coincides with the hausdorff dimension of the metric space @xmath22 .
* proof : * by theorem [ thm spectral decomposition d^h on heisenberg group ] , the only degenerate eigenvalue of @xmath1 on @xmath1286 is @xmath596 .
hence let @xmath130 denote the set of all eigenvalues of @xmath1 which are not @xmath596 .
we have to show by definition [ def metric dimension ] : @xmath1287 now we can decompose @xmath130 into two disjoint sets @xmath1288 where @xmath1289 contains all the eigenvalues of @xmath1 listed under ( i ) and @xmath1290 contains all the eigenvalues of @xmath1 listed under ( ii ) in theorem [ thm spectral decomposition d^h on heisenberg group ] which are not @xmath596 .
the set @xmath1289 contains exactly the eigenvalues of the classical dirac operator , acting on the spinor bundle @xmath887 of the torus @xmath49 ( with respect to a riemannian metric on @xmath49 such that the vector fields @xmath1291 from used in the proof of the theorem form an orthonormal frame for @xmath49 ) , and since this classical dirac operator is elliptic it is known that they grow proportional to the function @xmath1292 by weyl asymptotic . but this shows @xmath1293 therefore the crucial point of the asymptotic behavior of our eigenvalues lies in the set @xmath1290 .
applying theorem [ thm spectral decomposition d^h on heisenberg group ] to our situation where @xmath64 , we see that the eigenvalues belonging to @xmath1290 are given asymptotically by @xmath1294 for @xmath1191 and @xmath1295 .
we make a further ( disjoint ) decomposition @xmath1296 with @xmath1297 containing exactly the eigenvalues @xmath1298 with @xmath1190 .
note that each of these eigenvalues occurs with the multiplicity @xmath888 } |\tau|^m$ ] . for a fixed @xmath1191
we denote the absolute values of these elements belonging to @xmath1297 by @xmath1299 such that we can work with the number @xmath1300 from proposition [ prop asymptotic behaviour versus growth rate ] .
it is well known that for any @xmath1301 we have @xmath1302 ( one can consider these numbers as the eigenvalues of the elliptic dirac operator on the @xmath648-dimensional torus , and then this statement follows from weyl asymptotic ) , which means @xmath1303 and hence for every @xmath1191 it follows from that @xmath1304 at this point we can apply proposition [ prop asymptotic behaviour versus growth rate ] , which tells us that is equivalent to the fact that @xmath1305 for the elements of @xmath1297 denoted by , where we do not care about the multiplicity @xmath888}|\tau|^m$ ] of these numbers as eigenvalues of @xmath1 for a moment . with these results , we can check for the set @xmath1290 which will prove the corollary : because of our decomposition of @xmath1290 and we have @xmath1306 } \left| \tau \right|^m \cdot \frac{1}{\left(|\tau|^{\frac{1}{2 } } k^{\frac{1}{2m}}\right)^p}\\ & = & \sum_{\tau \in { \mathbb{z}}\setminus \{0\ } } 2^{\left[\frac{d}{2}\right ] } \frac{1}{\left|\tau\right|^{\frac{p}{2}-m } } \cdot \sum_{k \in { \mathbb{n } } } \frac{1}{k^{\frac{p}{2m}}},\\\end{aligned}\ ] ] where we have taken care of the multiplicity @xmath888 } |\tau|^m$ ] of every eigenvalue @xmath1307 in the second equation .
now the second geometric series in the last line of this calculation converges if and only if @xmath1308 , and the first geometric series in this line converges if and only if @xmath1309 .
but this , together with tells us that the series @xmath1310 converges if and only if @xmath1309 , which shows that is true and hence the corollary is proved .
* remark : * the same statement should also be true for the case @xmath1062 , but in this case the argument is a bit more involved
. @xmath74 from the argumentation of theorem [ thm spectral decomposition d^h on heisenberg group ] it is also possible to calculate the eigenvalues of the horizontal pull - back dirac operator exactly for given numbers @xmath246 and @xmath212 ( after determining the matrices @xmath1311 describing the clifford action ) . the idea is to use the orthonormal basis of @xmath1210 consisting of the hermite functions @xmath1312 which allows us to decompose @xmath1201 into finite dimensional subspaces which are invariant under @xmath1 . on these subspaces
the determination of the spectrum is done by calculating eigenvalues of matrices . in the case @xmath655 where @xmath1313 is the @xmath245-dimensional compact heisenberg nilmanifold ,
this has been done in @xcite and @xcite for an ordinary dirac operator .
the argument used there can be transfered to our case .
after these calculations , we find that the eigenvalues of @xmath1 are exactly the numbers a. @xmath1314 for @xmath1315 such that @xmath1316 , @xmath1317 of multiplicity @xmath98 , b. @xmath1318 for @xmath1319 , @xmath1191 of multiplicity @xmath1320 , c. @xmath1321 for @xmath1191 of multiplicity @xmath1320 . considering the inverses of the absolute values of these eigenvalues which are not zero , we find ( for the set @xmath1290 from in the proof of theorem [ thm spectral decomposition d^h on heisenberg group ] ) that @xmath1322 which shows the statement of corollary [ cor hausdorff dimension from d^h ] for this example . using the hermite functions from ,
it is possible to describe the kernel of @xmath1 on the sub - spaces @xmath1217}})$ ] of the decomposition of each space @xmath982 from lemma [ lemma spectral decomposition h_tau , tau neq 0 ] . in the case @xmath655 and @xmath1313 , one finds that on each space @xmath1323 belonging to @xmath982 the kernel of @xmath1 is spanned by the @xmath127-valued functions @xmath1324 for @xmath1325 and @xmath1326 for @xmath1327 . in particular we observe that these functions are smooth .
this observation leads to the idea that @xmath1 may be an example for a so - called _ weakly hypoelliptic operator _ , in the sense that @xmath1328 implies that @xmath507 is smooth .
recently , these weakly hypoelliptic operators have been considered by christian br in @xcite .
@xmath74 to close this section , let us once again summarize the results we achieved so far : we have seen that for the case @xmath1062 , the horizontal pull - back dirac operator on @xmath365 does not have a compact resolvent .
hence it only furnishes a degenerate spectral triple , but from this degenerate spectral triple the most important ingredients of the carnot - carathodory geometry on @xmath0 , which are the carnot - carathodory metric and the hausdorff dimension , can be detected . in the next section
, we will deduce from this co - dimension @xmath98 case that the absence of a compact resolvent of @xmath1 occurs on any compact carnot nilmanifold .
this section will provide the most general result of this chapter concerning compact carnot nilmanifolds : on an arbitrary compact carnot nilmanifold the horizontal pull - back dirac operator has ( at least one ) infinite dimensional eigenspace and does therefore not have a compact resolvent .
later we will prove that any horizontal dirac operator on an arbitrary carnot manifold fails to be hypoelliptic , which can be seen as a generalization of this statement .
the starting point is theorem [ thm decomposition from center ] , which allows us to reduce the problem from an arbitrary carnot group to a @xmath53-step nilpotent one . in the last section we have made a detailed treatment of compact heisenberg nilmanifolds , which are exactly those @xmath53-step nilmanifolds whose horizontal distribution has co - dimension @xmath98 , and seen that we have a degenerate eigenspace in this case .
so the only step missing is to get from an arbitrary space of step @xmath53 to a compact heisenberg nilmanifold . to do this
, we consider a submersion of the type introduced in section 2.4 .
let @xmath1329 be the compact nilmanifold of a carnot group @xmath415 of rank @xmath53 .
we assume that the lie algebra @xmath1330 of @xmath33 has the grading @xmath1331 , where @xmath1332 is an orthonormal frame for @xmath335 and @xmath406 is an orthonormal frame for @xmath1333 . for a @xmath407
we consider the orthonormal projection @xmath1334 we have seen in section 2.4 that the vector space @xmath1335 can be canonically equipped with a lie bracket , such that it is a graded nilpotent lie algebra of rank @xmath53 , where @xmath335 is bracket generating of step @xmath53 and co - dimension @xmath98 for @xmath1335 .
we denote the carnot group arising from @xmath1335 by @xmath1336 , and the lie group homomorphism arising from @xmath1337 by @xmath1338 for @xmath1339 , we further define the compact carnot nilmanifold @xmath1340 over @xmath1336 .
assume that a horizontal clifford bundle @xmath1341 and a horizontal pull - back dirac operator @xmath1342 on @xmath1343 , arising from a spin structure @xmath881 on @xmath49 , are given .
now @xmath881 also defines a horizontal clifford bundle @xmath1344 on @xmath1345 , which is a vector bundle of the same rank as @xmath1346 ; and since the horizontal distributions of @xmath1347 and @xmath1345 can both be identified with @xmath874 we have a horizontal clifford action on @xmath1345 which coincides with the horizontal clifford action on @xmath1347 , meaning @xmath1348 for all @xmath1349 as endomorphisms on the vector bundles of rank @xmath888}$ ] .
hence we have a horizontal pull - back dirac operator @xmath1350 on @xmath1345 . but this operator allows us to prove the following lemma .
[ lemma reduction to codimension 1 case on homogeneous spaces ] let @xmath1342 be the horizontal pull - back dirac operator on the compact carnot nilmanifold @xmath1329 , where @xmath415 is a carnot group of nilpotency step @xmath53 , and for a @xmath407 let @xmath1351 be the horizontal pull - back dirac operator from on the compact heisenberg nilmanifold @xmath1352 constructed above .
then if there is a @xmath407 such that the section @xmath1353 lies in the kernel of @xmath1351 , the section @xmath1354 is an element of the kernel of @xmath1342 .
* proof : * the idea is to use the expression from theorem [ prop expression of horizontal dirac via representation ] of the horizontal dirac operators involving the right - regular representation @xmath221 of the carnot groups @xmath415 and @xmath1336 on the spaces @xmath1355 and @xmath1356 .
this means in the current situation that the operators @xmath1357 and @xmath1358 , applied to @xmath556 and @xmath1359 , are given by @xmath1360 and @xmath1361 we want to show that for any @xmath1359 such that @xmath1362 we have @xmath1363 , where @xmath1364 following the discussion preceding this lemma , we have @xmath1348 as endomorphisms on the vector bundles of rank @xmath888}$ ] .
hence , from the expressions and the desired statement will follow if we have @xmath1365 for any @xmath1366 .
will check this via a small calculation .
let @xmath1075 .
then for the left hand side of we get for every @xmath1367 , using exponential coordinates of @xmath415 and the baker - campbell - hausdorff formula ( see equation in section 2.2 ) , @xmath1368 \right ) \right ) .
\end{aligned}\ ] ] now we can calculate the commutators occurring in the last line using for @xmath1369 the @xmath313-levi forms ( see definition [ def generalized levi form ] ) @xmath1370 of @xmath415 , which gives us the identity @xmath1371 = \sum_{\mu=1}^{d_2 } l_{jk}^{(\mu ) } x_{2,\mu}\ ] ] for all @xmath1372 . plugging this into the above calculation we get @xmath1373 now we set @xmath1374 in and get from the definition of @xmath1375 @xmath1376 since @xmath1377 for all @xmath1194 . since @xmath1375 is a lie algebra homomorphism , we have @xmath1378 = { \mathrm{pr}}_\nu ( [ x_{1,j } , x_{1,k}])\ ] ] for all @xmath1379 .
thus we see immediately ( after using the baker - campbell - hausdorff formula on @xmath1336 in the same way we did above on @xmath415 ) that is exactly the right hand side of , and therefore we have @xmath1380 this shows the statement of the lemma .
@xmath91 now we can put the things together to prove that the horizontal pull - back dirac operator we constructed has an infinite dimensional eigenspace on any carnot group one can choose .
[ thm d^h has infinite dimensional kernel on any homogeneous carnot space ] let @xmath47 be the compact nilmanifold of a carnot group @xmath33 of rank @xmath216 and step @xmath221 .
let @xmath1 be the horizontal pull - back dirac operator acting on the horizontal clifford bundle @xmath913 which is arising from a spin structure @xmath881 of the torus @xmath49 .
then the kernel of @xmath1 is infinite - dimensional .
this means in particular that @xmath1 does not have a compact resolvent .
* proof : * let @xmath325 be the grading of the lie algebra of @xmath33 .
we consider the decomposition @xmath1381 of @xmath964 from theorem [ thm decomposition from center ] , where all the spaces @xmath982 are invariant under @xmath1 . by the second statement of theorem [ thm decomposition from center ] ,
we have @xmath1382 where @xmath1383 is the compact nilmanifold of the carnot group @xmath1384 and the restriction of @xmath1 to @xmath985 can be identified with a horizontal pull - back dirac operator @xmath544 acting on @xmath1385 . but
this means that we can apply the same decomposition to the space @xmath985 , which contains a hilbert space isomorphic to @xmath1386 , with @xmath1387 the compact nilmanifold of the step @xmath1388 carnot group @xmath1389 , and so on .
inductively , we find a hilbert space @xmath1390 , where @xmath1347 is the compact nilmanifold of a carnot group @xmath415 of step @xmath53 , which is invariant under @xmath1 and on which @xmath1 can be identified with a horizontal pull - back dirac operator @xmath1342 , acting on @xmath1346 .
but for this operator , we find an infinite dimensional kernel by lemma [ lemma reduction to codimension 1 case on homogeneous spaces ] : since @xmath33 is not abelian , we find a @xmath407 , where @xmath1391 is the dimension of the space @xmath1333 from the grading @xmath1331 of the lie algebra of @xmath415 , such that @xmath1392 for some @xmath401 .
we can define a horizontal clifford bundle @xmath1393 and a horizontal pull - back operator @xmath1351 from the corresponding objects on @xmath1347 like we did in the discussion preceding lemma [ lemma reduction to codimension 1 case on homogeneous spaces ] .
now we know by theorem [ thm spectral decomposition d^h on heisenberg group ] that @xmath1351 has an infinite dimensional kernel , and that we can choose a basis @xmath1394 of this kernel . using lemma [ lemma reduction to codimension 1 case on homogeneous spaces ] , we can lift this basis to an orthonormal system of infinitely many independent sections of @xmath1395 , given by @xmath1396 , which all lie in the kernel of @xmath1 .
( note that these sections are indeed linear independent since this is the case for the @xmath1397 s and @xmath1398 is a submersion . ) hence we have shown that @xmath1 has an infinite - dimensional kernel on @xmath1395 , and from the above argumentation this is also the case on @xmath964 .
the statement that @xmath1 can not have a compact resolvent follows trivially from the fact that @xmath1399 is infinite - dimensional .
@xmath91 we have shown by theorem [ thm d^h has infinite dimensional kernel on any homogeneous carnot space ] that the horizontal pull - back dirac operator @xmath1 does not furnish a spectral triple on arbitrary compact carnot nilmanifolds .
theoretically it is possible to do spectral decompositions like in section 4.3 for any given carnot group and thus get statements about the asymptotic behavior of the non - degenerate eigenvalues : we have to know about the representation theory of @xmath33 .
now there is an algorithm to determine all the irreducible unitary representation of a carnot group @xmath33 ( up to equivalence ) developed by alexander kirillov ( @xcite ) , which is also referred to as the orbit method ( see e.g. @xcite or @xcite ) .
but since this algorithm makes use of the concrete structure of @xmath33 it is hard to get general results in our context . and
even for a given carnot group which is not isomorphic to @xmath393 we expect the calculations to be very long and complicated .
but anyway we will show in the following chapters that the phenomenon of the degeneracy of the horizontal dirac operator occurs in general , such that the description of the carnot - carathodory geometry via spectral triples does not work as one would expect .
in the previous chapter we presented an explicit construction for horizontal dirac operators on compact carnot nilmanifolds and we saw that these operators do not have a compact resolvent . to put these observations into a greater generality , we want to adopt tools from pseudodifferential calculus . in the classical case ,
the dirac operator @xmath5 on a compact manifold is an elliptic operator of order @xmath98 , and therefore it follows from pseudodifferential theory that it admits a parametrix of order @xmath1400 , which is compact because of the sobolev embedding theorem .
the existence of a parametrix leads to the possibility to construct complex powers within the calculus , which shows that the resolvent @xmath1401 of @xmath5 is compact .
now we intend to present something analogous for operators on carnot manifolds , respecting the grading of a graded nilpotent lie algebra .
indeed there is a pseudodifferential symbol calculus for heisenberg manifolds : it has been developed simultaneously by richard beals and peter greiner ( see @xcite ) and michael taylor ( see @xcite ) in the 1980s . in the last decade ,
some properties which are important for our work have been presented by raphal ponge ( see @xcite ) .
we will see that in this calculus hypoellipticity takes the place of ellipticity , since hypoellipticity implies the existence of complex powers .
in addition , we will see that on a compact manifold operators of negative order are compact , such that one can get a compact resolvent for a given operator of positive heisenberg order . for hypoelliptic self - adjoint horizontal laplacians on compact heisenberg manifolds
it is also known that their eigenvalues grow polynomial with a rate which gives back the graded dimension of the manifold .
this can be seen as an analogy to the weyl asymptotics in the elliptic case . in this chapter
we give an overview over the heisenberg calculus developed by richard beals and peter greiner , explain the composition of symbols and explore the role of hypoellipticity .
for details we refer to the books by beals and greiner ( @xcite ) and by raphal ponge ( @xcite ) .
afterwards we present results concerning the asymptotic growth of the eigenvalues and the existence of complex powers of hypoelliptic operators , which can be derived by an expansion of the heat kernel of the operator similarly to the classical case . here
we refer to @xcite and also to @xcite .
it is expected that there are analogous results for arbitrary carnot manifolds , see @xcite , but we did not find this generalization worked out in the literature and hence we restrict ourselves to heisenberg manifolds .
let @xmath0 be a heisenberg manifold of dimension @xmath1402 , which means by section 2.3 that we have a grading @xmath1403 of the tangent bundle such that @xmath19 is a bracket generating horizontal distribution of rank @xmath216 and @xmath1404 $ ] is of rank @xmath98 .
an orthonormal frame for @xmath19 shall be given by the vector fields @xmath41 , while @xmath377 shall span @xmath465 .
as we know by section 2.2 , the graded ( and therefore the hausdorff ) dimension of the metric space @xmath22 , where @xmath39 is the carnot - carathodory metric , is equal to @xmath1285 .
in section 2.3 we have seen how one can identify the tangent space of @xmath0 with a bundle @xmath1405 of graded nilpotent lie algebras , and hence we also have this structure on the cotangent bundle @xmath224 which we denote by @xmath1406 . at a point @xmath204 , we have @xmath1407 ( as a vector space ) , and we have the dilations @xmath1408 for coordinates @xmath1409 ( with @xmath1410 ) .
we will further use the koranyi gauge @xmath1411 see definition [ def koranyi gauge ] .
first we will consider @xmath1412 to be an open subset of @xmath56 ; the generalization to vector bundles and manifolds will be the content of a theorem we mention at a later point of this section .
let @xmath1413 denote the ( classical ) symbol of the vector fields @xmath1414 , and set @xmath1415 .
note that by proposition [ prop carnot lie algebra as vector fields ] we can detect @xmath1416 from the symbol @xmath556 of the homogeneous operator . using the koranyi gauge @xmath1417 and the notation @xmath1418 for a multi - index @xmath1419 we take care of the homogeneity of the @xmath324 considered as differential operators .
we can define the following symbol classes on which the heisenberg calculus will be based ( see @xcite , ( 10.5)-(10.18 ) , and @xcite , section 3.1.2 ) .
[ def symbol classes heisenberg calculus ] let @xmath1420 be an open subset . then : a. for @xmath1421 , we set @xmath1422 we will further denote the class of these homogeneous functions , which do not depend on @xmath107 , by @xmath1423 . b. for @xmath1421 , we set @xmath1424 where the asymptotic expansion @xmath1425 is meant in the sense that for all multi - indices @xmath1426 and all @xmath1427 , we have @xmath1428 for a locally bounded function @xmath1429 on @xmath1430
. + we will further denote the class of these functions , which do not depend on @xmath107 , by @xmath1431 : @xmath1432 c. for @xmath1421 , we set @xmath1433 d. for @xmath1421 , we set @xmath1434 for @xmath1435 , the asymptotic expansion of @xmath1436 is given by @xmath1437 we call elements belonging to the class @xmath1438 _ heisenberg symbols _ of order @xmath648 . e. the symbol class @xmath1439 induces the class of _ heisenberg pseudodifferential operators _ on @xmath1430 , which we denote by @xmath1440 .
+ in detail , for @xmath1441 , the corresponding operator @xmath1442 is given by @xmath1443 for any function @xmath1444 .
on the other hand , if an operator @xmath1445 can be written in the form for a function @xmath1446 , we call @xmath1447 the heisenberg symbol of @xmath1445
. f. the class @xmath1448 is called the class of _ smoothing operators _ in the heisenberg calculus . @xmath74 before we go on with the theory , we will state an example for heisenberg pseudodifferential operators we already know . [ ex graded differential operators as heisenberg pseudos ] let @xmath1449 equipped with a frame @xmath1212 such that @xmath41 forms a bracket generating horizontal distribution .
then any graded differential operator @xmath1450 of order @xmath1451 with respect to this frame ( see definition [ def graded differential operator ] ) is a heisenberg pseudodifferential operator of heisenberg order @xmath313 .
its heisenberg symbol is a polynomial @xmath1452 of ( homogeneous ) degree @xmath313 , and its asymptotic expansion is given by the homogeneous terms of this polynomial . in case the coefficients of @xmath5 are constant , we have a heisenberg symbol belonging to the class @xmath1453 .
in particular , a horizontal laplacian is a heisenberg pseudodifferential operator of heisenberg order @xmath53 .
@xmath74 we do not know yet if the expression from the last item of definition [ def symbol classes heisenberg calculus ] makes sense . but
this will be the case , since a symbol of @xmath1438 belongs to a hrmander symbol class @xmath1454 where @xmath1455 is again a locally bounded function on @xmath1430 .
these symbol classes were established by lars hrmander in @xcite , where it was also shown that the corresponding operators can be extended to bounded operators on certain sobolev spaces . for the following theorem ,
we refer to @xcite , proposition ( 10.22 ) .
[ thm embedding of s_hei^m ] for every @xmath1421 , we have @xmath1456 @xmath91 from this embedding , we see that the operator from equation is well defined for every test function @xmath1444 .
moreover , we immediately get some regularity properties for heisenberg pseudodifferential operators , which follow immediately from the corresponding regularity properties of operators belonging to the class @xmath1457 .
[ cor heisenberg pseudos regularity ] let @xmath1421 , @xmath1458 .
we can define an operator @xmath1459 which has the following properties : a. @xmath1460 is a continuous linear operator .
b. for every @xmath77 , @xmath1445 can be extended to a bounded linear operator @xmath1461 where @xmath1462 and @xmath1463 denote the @xmath52-sobolev spaces .
+ in particular , this means @xmath1464 if @xmath1465 .
@xmath91 in @xcite , it is also shown that the class @xmath1466 does not depend on the choice of the frame @xmath1467 of @xmath20 , as long as it respects the grading structure ( see @xcite , proposition ( 10.46 ) ) .
after defining our symbol classes , we show that they induce a meaningful calculus .
this is in general not the case for symbols of the class @xmath1468 , but it turns out that the composition of two heisenberg pseudodifferential operators gives a @xmath1469 again .
we briefly sketch how this composition is defined , referring to @xcite , chapters 12 - 14 , and to @xcite .
section 3.1.3 , for the details .
assume first that the symbols of the operators which shall be composed are given by @xmath1470 and @xmath1471 , where @xmath1472 for homogeneous functions @xmath1473 .
* we fix an @xmath1474 and choose coordinates on @xmath1430 which are centered at @xmath107 .
this provides us with symbols @xmath1475 with @xmath1476 for @xmath1477 , defined on the dual @xmath1478 of the tangent graded lie algebra @xmath1479 at @xmath107 , which corresponds to a carnot group @xmath1480 via the exponential mapping .
then @xmath1480 can be identified with the tangent carnot group @xmath1481 of @xmath1430 .
* after these identifications , we can use the lie group composition @xmath814 of the tangent carnot group @xmath1482 to define a convolution product @xmath1483 in the following way : for @xmath1484 , @xmath1485 , the operator associated to the symbol is simply the convolution operator @xmath1486 acting on the carnot group @xmath1481 .
the composition of two convolution operators of this type gives the bilinear mapping ( see @xcite , proposition ( 12.14 ) ) .
* finally , one finds that the product @xmath1487 from depends smoothly on @xmath107 ( see @xcite , proposition ( 13.3 ) ) , which gives us a continuous bilinear product @xmath1488 the above construction shows how homogeneous symbols can be composed .
now every heisenberg symbol has an asymptotic expansion into homogeneous symbols , and therefore one can find an asymptotic expansion for the composition of two arbitrary heisenberg symbols , which is again a heisenberg symbol and furnishes a @xmath1469 . the details for this expansion
are formulated within the next theorem ( see @xcite , theorems ( 14.1 ) and ( 14.7 ) , and @xcite , proposition 3.1.9 ) .
[ thm composition of heisenberg pseudos ] for @xmath466 let @xmath1489 have the symbol @xmath1490 in the sense of and assume that one of these operators is properly supported .
then for the operator @xmath1491 we have @xmath1492 , and its symbol @xmath338 is given by @xmath1493 in the sense of , where @xmath1494 in , @xmath1495 denotes the sum over all the indices such that @xmath1496 and the functions @xmath1497 are polynomials in the derivatives of the coefficients of the vector fields @xmath1498 .
@xmath91 so far , we have developed the heisenberg calculus on open subsets of @xmath56 . to define this calculus for operators acting on vector bundles over arbitrary heisenberg manifolds
, one has to show that the class of @xmath1469s is invariant under the change of charts respecting the heisenberg structure . in detail
, we have the following theorem which shows us that we are able to extend the theory to the manifold case ( see @xcite , theorem ( 10.67 ) , and @xcite , proposition 3.1.18 ) .
[ thm invariance theorem for heisenberg calculus ]
let @xmath1499 and @xmath1500 be open subsets on @xmath1501 together with hyperplane bundles @xmath1502 and @xmath1503 , and let @xmath1504 be a heisenberg diffeomorphism , which means we have @xmath1505 for the differential of @xmath1506 .
then if @xmath1507 is a heisenberg pseudodifferential operator of order @xmath648 on @xmath1500 , the pullback @xmath1508 of this operator to @xmath1499 is a heisenberg pseudodifferential operator of order @xmath648 on @xmath1499 .
@xmath91 because of this theorem , we can consider heisenberg pseudodifferential operators on heisenberg manifolds @xmath0 , acting on vector bundles @xmath585 from now on , and write @xmath1509 for this class of operators .
in particular , we can derive the following consequence from corollary [ cor heisenberg pseudos regularity ] which states that @xmath1469s of negative order defined on a compact manifold are compact .
it can be used if we want to show that a certain heisenberg pseudodifferential operator has a compact resolvent .
[ cor heisenberg pseudos of negative order are compact ] for @xmath1510 , let @xmath1511 be a heisenberg pseudodifferential operator acting on a vector bundle @xmath585 over a heisenberg manifold @xmath0 . then @xmath1445 is a compact operator on the hilbert space @xmath1512 . *
proof : * by corollary [ cor heisenberg pseudos regularity ] and theorem [ thm invariance theorem for heisenberg calculus ] , for every @xmath77 the operator @xmath1513 is bounded on the @xmath52-sobolev spaces @xmath1514 .
we can now choose @xmath1515 and use the fact that the embedding @xmath1516 is compact by the sobolev embedding theorem to derive that that operator @xmath1517 is compact as the composition of a bounded and a compact operator .
@xmath91 after having established a composition rule inside the class of @xmath1518s in theorem [ thm composition of heisenberg pseudos ] , the next thing to examine is the existence and the regularity of parametrices inside this class .
first of all let us recall the definition of a parametrix .
[ def parametrix ] let @xmath585 be a vector bundle over a heisenberg manifold @xmath0 , and let @xmath1519 be a @xmath1469 .
then an operator @xmath1520 with @xmath1521 is called a _ heisenberg parametrix _ or a _ heisenberg pseudodifferential inverse _ of @xmath1522 , if we have @xmath1523 which means that the operators @xmath1524 and @xmath1525 are smoothing .
@xmath74 in the classical calculus for symbols of the type @xmath1526 , a parametrix for a pseudodifferential operator @xmath1522 exists if the operator is elliptic .
we will see that a necessary condition for the existence of a parametrix in the heisenberg calculus is the hypoellipticity of the operator .
the following classical definition of hypoellipticity is due to lars hrmander and can also be found in @xcite and @xcite .
[ def hypoelliptic ] let @xmath1522 be a @xmath1469 of order @xmath1421 , acting on a vector bundle @xmath585 over a heisenberg manifold @xmath0 .
then @xmath1522 is called _ hypoelliptic _ , if for any distribution @xmath1527 we have @xmath1528 in more detail , we call @xmath1522 _ hypoelliptic with the loss of @xmath555 derivatives _ , if we have for any @xmath77 : @xmath1529 @xmath74 in the classical case , a pseudodifferential operator is elliptic if its principal symbol is invertible .
this generalizes to the case of heisenberg pseudodifferential operators in terms of hypoellipticity : the main result will be that a @xmath1469 is hypoelliptic if the @xmath1469 associated to the principal part in the asymptotic ( heisenberg ) expansion of its symbol is invertible in the heisenberg calculus .
we first introduce the notion of the principal symbol and the model operator , as it is done in @xcite and @xcite .
[ def principal symbol and model operator ] let @xmath1522 be a @xmath1469 of order @xmath1421 with heisenberg symbol @xmath1530 , acting on a vector bundle @xmath585 over a heisenberg manifold @xmath0 . for a point @xmath1531 , we choose an open subset @xmath1532 of @xmath0 and local coordinates of @xmath1430 centered at @xmath15 to consider the asymptotic expansion @xmath1533 in the sense of definition [ def symbol classes heisenberg calculus ] . for @xmath1534 with @xmath1413 the classical symbol of @xmath1414
, we find a @xmath1535 such that in @xmath1536 .
then the symbol @xmath1537is called the _ ( homogeneous ) principal symbol _ of @xmath1522 at @xmath15 . the corresponding operator @xmath1538 where @xmath1539 is the tangent carnot group at @xmath15 , is called the _ ( homogeneous ) model operator _ of @xmath1522 at @xmath15 . @xmath74 * remark : * it is also possible to define a global principal symbol @xmath1540 on @xmath0 using the kernel representation of heisenberg pseudodifferential operators , for which we refer to @xcite , theorem 3.2.2 .
@xmath74 [ ex principal symbol graded differential operator ] for a graded differential operator @xmath1541 from example [ ex graded differential operators as heisenberg pseudos ] , the principal symbol of @xmath5 at a point @xmath1531 is given by the leading homogeneous term of the polynomial @xmath338 after freezing the coefficients of @xmath338 in @xmath15 .
hence the model operator of @xmath5 is given by the homogeneous graded differential operator with constant coefficients belonging to this homogeneous term .
in particular , for the case of a horizontal laplacian of the form @xmath1542 the model operator of @xmath1543 at @xmath1531 is the operator @xmath1544 @xmath74 one can show that the convolution of two principal symbols gives the principal symbol of the composition of the corresponding operators , and hence the composition of two model operators gives the model operator of the composition of the original operators .
see for example @xcite , proposition 3.2.9 . in particular
, the existence of a parametrix for a heisenberg pseudodifferential operator implies the existence of a parametrix of its model operator at each point .
we even have equivalence for these two statements , and both statements imply the hypoellipticity of each model operator , which is formulated in the next theorem ( see @xcite , proposition 3.3.1 , theorem 3.3.18 and proposition 3.3.20 ) .
[ thm hypoellipticity and existence of parametricies ] let @xmath585 be a vector bundle over a heisenberg manifold @xmath0 , and let @xmath1545 be a @xmath1469 of order @xmath246 .
then the following statements are equivalent .
a. @xmath1522 has a heisenberg parametrix @xmath1445 with symbol @xmath1546 .
b. at each point @xmath1531 , the model operator @xmath1547 from of @xmath1522 has a heisenberg parametrix @xmath1548 with symbol @xmath1549 . c. the global principal symbol @xmath1550 of @xmath1522 is invertible with respect to the convolution product for homogeneous symbols .
if any of these conditions is fulfilled , @xmath1522 is hypoelliptic with loss of @xmath1551 derivatives .
we note that in @xcite it is only shown that hypoellipticity is implied by the conditions ( i ) - ( iii ) of theorem [ thm hypoellipticity and existence of parametricies ] . but
in the case of a graded differential operator @xmath5 with constant coefficients , the hypoellipticity of @xmath5 is equivalent to the invertibility of @xmath5 in the heisenberg calculus .
this can be seen by the so - called _ rockland condition _ , which we will introduce in chapter 6 :
it is shown in @xcite , that this rockland condition ( in a more general version compared to the one we will introduce ) for a heisenberg pseudodifferential operator @xmath1522 is equivalent to the statements ( i ) - ( iii ) of theorem [ thm hypoellipticity and existence of parametricies ] ( see @xcite , theorem 3.3.18 ) . but for graded differential operators with constant coefficients , the rockland condition is equivalent to the hypoellipticity of @xmath5 , as we will see in section 6.1 . in case of a horizontal laplacian of the form @xmath1552
it was proved in @xcite that the hypoellipticity of this operator is equivalent to its invertibility in the heisenberg calculus .
since we will work in particular with horizontal laplacians , we state this theorem here ( see @xcite , theorem ( 18.4 ) ) .
[ thm hypoellipticity and existence of parametricies horizontal laplacian ] for the horizontal laplacian @xmath1543 from , the following statements are equivalent : a. @xmath1543 has a heisenberg parametrix @xmath1553 b. at each @xmath1554 , the model operator of @xmath1543 has a heisenberg pseudodifferential inverse . c. @xmath1543 is hypoelliptic with loss of one derivative .
@xmath91 we will see in the next section that from the point of view of constructing and analyzing spectral triples using the techniques of heisenberg calculus indeed hypoellipticity is the central property . right now , we close this section with a remark concerning a possible generalization to arbitrary carnot manifolds .
[ rem calculus on general carnot manifold ] let @xmath0 be an arbitrary carnot manifold of step @xmath221 with grading @xmath317 such that for any @xmath38 @xmath1555 forms an orthonormal frame for @xmath279 ( with @xmath1556 ) .
then we assume that we can generalize definition [ def symbol classes heisenberg calculus ] as follows : for the carnot group @xmath33 corresponding to @xmath0 we consider the exponential coordinates @xmath1557 with @xmath1558 , where @xmath1559 , the dilations @xmath1560 and the koranyi gauge @xmath1561 see definition [ def koranyi gauge ] .
in addition , for multi - indices @xmath1562 we set @xmath1563 then we can define symbol classes in analogy to definition [ def symbol classes heisenberg calculus ] , and it seems natural that it is possible to generalize the further definitions and theorems of this section to the general carnot case .
the proofs should work more or less analogous to the presentation in @xcite and @xcite , but it would by very laborious to write everything down in detail.@xmath74 in this short section we present some results in heisenberg calculus which open the door to constructing meaningful spectral triples from this calculus .
as we will see , the crucial assumption in all these results is the hypoellipticity of the operator .
first of all , we provide a theorem from which it can be shown that the operator @xmath1564 is compact for a hypoelliptic self - adjoint operator @xmath5 of heisenberg order @xmath98 .
we formulate a combinations of the theorems 5.3.1 and 5.4.10 from @xcite .
[ thm complex powers of heisenberg pseudos ] let @xmath0 be a heisenberg manifold .
suppose that @xmath1522 is a hypoelliptic self - adjoint heisenberg pseudodifferential operator of order @xmath1565 which is bounded from below and which is satisfying @xmath1566 .
then , for any @xmath1567 , the operator @xmath1568 defined via functional calculus is a @xmath1469 of order @xmath1569 .
@xmath91 * remark : * note that we have not introduced @xmath1469s of non - integer order properly in this thesis . for this
, we refer to @xcite .
since in our case every operator appearing in this context will be of integer order , the theorems formulated in section 5.1 are sufficient .
note also that the formulation in @xcite of this theorem is more general in the sense that he does not assume that the kernel of @xmath1522 is only assumed to be finite dimensional , which is the case because of the hypoellipticity of @xmath1522 . in this general case
, one can also construct complex powers using projections onto the orthonormal complement of the kernel .
@xmath74 the next thing is to discover the asymptotic behavior of the eigenvalues of a self - adjoint heisenberg pseudodifferential operator , from which one can detect the metric dimension of a spectral triple . in classical pseudodifferential calculus
, one can consider the heat kernel of a positive self - adjoint elliptic operator @xmath1522 to get asymptotics for the growth of its eigenvalues .
this is done by expanding the trace of the operator @xmath1570 .
now something similar works for hypoelliptic self - adjoint horizontal laplacians which are bounded from below : mostly , this is the content of the paper @xcite by richard beals , peter greiner and nancy stanton .
some further considerations have been carried out by raphal ponge ( see e.g. @xcite ) . without going into the details , we just state the results here .
the following theorem shows how the trace expansion of the heat kernel of such a horizontal laplacian looks like ( see @xcite , theorem ( 5.6 ) , or @xcite , proposition 6.1.1 ) .
[ thm heat trace expansion horizontal laplacian ] let @xmath1571 be a hypoelliptic and self - adjoint horizontal laplacian which is bounded from below , acting on a vector bundle @xmath585 over a heisenberg manifold @xmath0 of dimension @xmath1572 ( which means that the horizontal distribution of @xmath0 has rank @xmath216 ) .
then for @xmath1573 we have the expansion @xmath1574 with @xmath1575 , where @xmath1576 can be computed from the term of degree @xmath1577 in the asymptotic expansion of the symbol of the parametrix of @xmath1543 in local coordinates .
now we denote by @xmath1578 the eigenvalues of @xmath1571 , counted with multiplicity .
a consequence of , in connection with karamatas tauberian theorem , is that the following asymptotic behavior of these eigenvalues holds ( see @xcite , proposition 6.1.2 ) .
[ thm asymptotic behaviour of eigenvalues horizontal laplacian ] let @xmath1571 be a hypoelliptic and self - adjoint horizontal laplacian which is bounded from below , acting on a vector bundle @xmath585 over a heisenberg manifold @xmath0 of dimension @xmath1572 .
then for @xmath1579 we have @xmath1580 for the eigenvalues @xmath1581 of @xmath1571 , where @xmath1582 is a constant depending on the dimension @xmath216 of the horizontal distribution and the term @xmath1583 in the heat trace expansion of @xmath1571 .
@xmath91 note that the above theorem just gives a qualitative statement about the growth of the eigenvalues , which suffices to detect the metric dimension of a spectral triple constructed from a first order hypoelliptic and self - adjoint operator on @xmath0 .
we will carry out this construction in section 7.1 . but
first of all , we want to return to the horizontal dirac operator and show that this theory can not be applied to @xmath1 , since @xmath1 is not hypoelliptic .
as we saw in the last chapter , the condition of hypoellipticity allows us to define complex powers of a self - adjoint heisenberg pseudodifferential operator within the heisenberg calculus .
since operators of negative heisenberg order on a compact carnot manifold are compact , one can argue that a hypoelliptic graded differential operator has a compact resolvent .
this means that if a horizontal dirac operator @xmath1 acting on a compact carnot manifold @xmath0 is hypoelliptic , condition ( ii ) for a spectral triple will be fulfilled on @xmath1584 .
but in chapter 4 we already constructed a horizontal dirac operator on an arbitrary compact carnot nilmanifold which does not have a compact resolvent .
the aim of this chapter is to show a generalization of the results of chapter 4 in the setting of pseudodifferential calculus : we show that any horizontal dirac operator on a carnot manifold is not hypoelliptic . from this
we will draw the conclusion that a horizontal dirac operator does not have a compact resolvent , and hence does not furnish a spectral triple .
we start this chapter by reviewing some well - known hypoellipticity criteria , starting with hrmanders _ sum - of - squares theorem _ and leading to the _ rockland condition _ , which states an equivalence between hypoellipticity of a graded differential operator and the non - degeneracy in the irreducible representations of its associated lie algebra
. then we will draw special attention to the situation where this lie algebra is a heisenberg algebra , since in this situation one has a good classification for the hypoellipticity of horizontal laplacians .
this will pay off , because in the second section we develop a criterion to exclude hypoellipticity of a graded differential operator by reducing the case to the co - dimension @xmath98 case .
similar to the argument given in chapter 4 for a specific example on compact carnot nilmanifolds , we will make use of the submersions between graded nilpotent lie algebras introduced in section 2.4 . finally , in section 6.3 we prove that any horizontal dirac operator is not hypoelliptic .
this follows quickly from the previous work by considering its square .
the development of hypoellipticity criteria for certain differential operators was a great matter in the 1970s and 1980s , and there are some celebrated results .
the origin of all these criteria is the famous _ sum - of - squares theorem _ by lars hrmander ( see @xcite ) .
[ thm hormander sum of squares ] let @xmath1585 be homogeneous vector fields on @xmath56 with real @xmath339-coefficients on an open set @xmath1586 and @xmath1587 real valued .
then the operator @xmath1588 is hypoelliptic , if among the operators @xmath324 and all their commutators there exist @xmath94 which are linearly independent at any given point in @xmath118 .
@xmath91 the problem is that this theorem only works for vector fields with real coefficients , so it will not apply to our case of the square of a horizontal dirac operator where the clifford action causes complex coefficients . during the following years
there were several generalizations of hrmander s theorem , for example by kohn ( @xcite and @xcite ) or rothschild and stein ( @xcite ) .
rothschild and stein developed a close - to - complete characterization for the hypoellipticity of horizontal laplacians of the form @xmath1589,\ ] ] where the @xmath324 are homogeneous vector fields on @xmath56 and @xmath1590 is assumed to be a real skew - symmetric matrix .
then rothschild and stein proved the following theorem in a slightly more general version ( see @xcite , theorem 1 and theorem 2 ) : [ thm hypoellipticity rothschild stein ] consider the space @xmath1591 = 0 \right\},\ ] ] and its orthonormal complement @xmath1592 with respect to the inner product @xmath1593 on @xmath1594 .
a. if @xmath1595 then @xmath1596 from is hypoelliptic
. b. assume that the lie algebra @xmath57 spanned by @xmath41 is graded , which means @xmath1597 such that @xmath1598 \subset v_{s+t}$ ] if @xmath1599 and @xmath1598 = 0 $ ] if @xmath1600 ( see also section 2.2 ) .
then if @xmath1601 and if the algebra @xmath1602 is not the lie algebra of a heisenberg group @xmath25 , @xmath1596 from is not hypoelliptic . @xmath91 * remark : * for @xmath1603 , the situation is more involved : it can be shown that there are situations where @xmath1330 is the lie algebra of a heisenberg group , in which the operator @xmath1596 is hypoelliptic , even though @xmath1604 for a more detailed treatment of the co - dimension @xmath98 case , we refer to theorem [ thm hypoellipticity of laplacian codimension 1 ] below . @xmath74
the idea for the proof of the second statement of this theorem is to reduce the situation to the case @xmath1605 , and to describe the operator @xmath1543 on @xmath1330 via irreducible unitary representation of @xmath1330 .
doing this , it is possible to write down a function which is not @xmath339 but belongs to the kernel of @xmath1543 , which is a contradiction to the hypoellipticity of the operator .
remember that we already used techniques from representation theory for the case of horizontal dirac operators on compact carnot nilmanifolds in chapter 4 , and it turned out that there is indeed a close connection between the hypoellipticity of graded differential operators and the representation theory of the underlying carnot group .
we will describe this connection now . as we noted in section 2.2 ,
a graded differential operator is an operator of the form @xmath1606 where @xmath338 is a polynomial with matrix - valued @xmath339 coefficients and where @xmath334 is a frame for the vector space @xmath287 appearing in the grading @xmath34 of a graded lie algebra @xmath57 .
if the coefficients of @xmath5 are constant , this suggests to consider @xmath5 as an element of the universal enveloping algebra @xmath344 of @xmath57 .
since @xmath57 is nilpotent , the exponential mapping is an isomorphism from @xmath57 onto its lie group @xmath33 which is a carnot group .
now let @xmath154 be an irreducible unitary representation of @xmath33 on a hilbert space @xmath11 , consisting of @xmath52-functions .
since @xmath154 gives rise to an irreducible , unitary representation of @xmath57 on @xmath11 via @xmath1607 we obtain a representation @xmath1608 of the graded differential operator @xmath5 .
let @xmath1609 denote the unitary dual of @xmath33 , which is the space of all irreducible unitary representations of @xmath33 .
using the above concept , there is a representation theoretic criterion which characterizes the hypoellipticity of graded differential operators completely .
we will now state this criterion , which is called the _ rockland condition _ since it was developed by charles rockland @xcite for heisenberg algebras .
b. helffer and j. nourrigat extended the rockland condition to the case of arbitrary graded nilpotent lie algebras ( see @xcite or also @xcite ) .
[ thm rockland condition ] let @xmath57 be a graded nilpotent lie algebra , @xmath1610 , and let @xmath1611 be a graded differential operator which is homogeneous of degree @xmath648
. then @xmath68 is hypoelliptic if and only if @xmath1612 is injective for all nontrivial @xmath1613 .
@xmath91 although there is a way to determine the irreducible unitary representations of an arbitrary nilpotent lie group ( see e.g. @xcite , we also mentioned this at the end of section 4.4 ) , this is a quite difficult task for specific examples .
hence we follow the same approach as in chapter 4 : for the co - dimension @xmath98 case , it is not too difficult to formulate hypoellipticity criteria deduced from the rockland condition
. then in the next section , we will see how this case enables us to make more general statements . for the rest of this section ,
we assume @xmath372 and @xmath1614 with @xmath1615 .
let @xmath1212 be a frame for @xmath1616 such that @xmath1617 and @xmath1618 for the representation of this lie algebra as vector fields on @xmath56 from proposition [ prop carnot lie algebra as vector fields ] .
we want to check horizontal laplacians of the form @xmath1619 acting on a vector bundle @xmath585 of rank @xmath1620 over an open subset @xmath1621 with @xmath1622 for hypoellipticity . here , the term @xmath592 denotes a graded differential operator of order smaller or equal to @xmath98 ( which means it is a differential operator of order @xmath98 only depending on the horizontal vector fields @xmath214 ) .
to formulate the criterion we recall the notion of the levi form from chapter 2.3 ( see definition [ defin levi form ] ) , which is the bilinear form @xmath1623 \mod v_1.\ ] ] for @xmath381 with @xmath382 , we denote by @xmath384 the antisymmetric matrix describing @xmath379 .
the following result states that the hypoellipticity of @xmath1571 from only depends on how the eigenvalues of @xmath136 behave in comparison with the eigenvalues of @xmath379 .
it is well known and can be found at various places in the literature ( see e.g. @xcite or @xcite ) , but because of its importance for our future arguments we give a proof , orientated towards the one given in @xcite .
[ thm hypoellipticity of laplacian codimension 1 ] let @xmath1571 be the horizontal laplacian given by with corresponding graded nilpotent lie algebra @xmath1624 , carnot group @xmath1625 and levi form @xmath379 , which is described by the levi matrix @xmath1626 corresponding to this basis .
the non - zero eigenvalues of @xmath379 are denoted by @xmath1627 ( including multiplicity ) with @xmath629 for all @xmath630 , where @xmath616 is the rank of @xmath379 .. then the hypoellipticity of @xmath1571 can be characterized as follows : a. if @xmath1628 with @xmath1629 the @xmath24-dimensional heisenberg group ( with @xmath1630 ) , then @xmath1571 is hypoelliptic if and only if no eigenvalue of @xmath136 belongs to the set @xmath1631 b. if @xmath1062 with @xmath1195 , then @xmath1571 is hypoelliptic if and only if no eigenvalue of @xmath136 belongs to the set @xmath1632 \cup \left [ \frac{1}{2 } \left\|l\right\|_1 , \infty \right).\ ] ] in both cases , @xmath1633 denotes the trace norm of @xmath68 .
* proof : * since @xmath1543 is a horizontal laplacian , by theorem [ thm hypoellipticity and existence of parametricies horizontal laplacian ] its hypoellipticity is equivalent to the hypoellipticity of its model operator , which is in this case the homogeneous horizontal laplacian @xmath1634 hence we only have to check the operator for hypoellipticity to prove the theorem , which can be done by the rockland condition ( see theorem [ thm rockland condition ] ) .
first of all , for every carnot group @xmath33 of co - dimension @xmath98 we have @xmath1635 , where @xmath1636 , by proposition [ prop tangent lie group bundle and rank of levi form ] . by the same proposition , there exists an orthonormal basis transformation of @xmath1637 such that after this transformation we have @xmath1638 for the levi matrix of @xmath379 , where @xmath5 is a diagonal matrix with diagonal entries @xmath1639 .
since an orthonormal change of the frame of @xmath335 does not change the form of the horizontal laplacian , as one can see after a small calculation , we can assume to be the matrix of the levi form @xmath379 of @xmath33 .
but this gives the commutator relations @xmath1640
= \lambda_j x_{d+1 } \ \ \ \text{and } \ \ \ [ x_{j+m},x_j ] = -\lambda_j x_{d+1}\ ] ] if @xmath249 , while all the other commutators are zero . for @xmath691
these are exactly the commutator relations for the lie algebra of @xmath1641 .
thus , the isomorphism @xmath1642 is coordinate - wise defined by @xmath1643 where we work on exponential coordinates of the carnot groups .
the next step is to show that the general case of @xmath1644 acting on a vector bundle @xmath585 of rank @xmath338 can be restricted to the scalar case : we can choose ( point - wise ) a basis of @xmath585 such that in this basis the matrix @xmath136 in is given by @xmath1645 where @xmath1646 are the eigenvalues of @xmath136 .
with respect to this basis the operator @xmath1644 takes the form @xmath1647 where for each @xmath323 with @xmath1648 we have a horizontal laplacian @xmath1649 acting on scalar valued functions .
now it is obvious that the operator fulfills the rockland condition from theorem [ thm rockland condition ] if and only if each of the scalar operators @xmath1650 does ( because otherwise the matrix would not be invertible which would be a contradiction to the injectivity of the corresponding irreducible representations ) . therefore we can restrict our considerations to the scalar case , working with the eigenvalues @xmath1646 of @xmath136 . from now on
, we will assume @xmath136 is scalar .
after the above simplifications , our task is to check the rockland condition for the scalar case with levi form of @xmath33 given by .
we start with the situation where @xmath1193 ( which means @xmath1651 ) . for the lie algebra of the heisenberg group
the representation theory is well known : up to equivalence , the nontrivial irreducible representations of the operators @xmath1652 , considered as basis vectors of the heisenberg algebra @xmath28 are of two types ( see e.g. @xcite , ( 3.3.5 ) - ( 3.3.7 ) , or @xcite ; compare also to the proof of theorem [ thm spectral decomposition d^h on heisenberg group ] ) : a. the infinite dimensional representations @xmath1653 of @xmath25 , parametrized by @xmath1654 , give rise to the representations @xmath1655 b. the one dimensional representations @xmath1656 of @xmath25 , parametrized by @xmath1657 , give rise to the representations @xmath1658 under the lie group isomorphism @xmath1659 given by , the representations of the basis vectors of the lie algebra @xmath57 we are interested in are given by the above representations applied to @xmath1660 and @xmath1661 for @xmath249 . hence all the irreducible representation for @xmath1543 are @xmath1662 and @xmath1663 obviously is injective for any @xmath1657 , so we only have to check the operators given by for the rockland condition . for these operators ,
their injectivity is equivalent to the injectivity of the operators @xmath1664 it is known that the spectrum of the harmonic oscillator @xmath1665 is exactly the set @xmath1666 ( see e.g. @xcite , section 3.4 , or @xcite ; compare also to the proof of theorem [ thm spectral decomposition d^h on heisenberg group ] ) , and therefore the invertibility of is equivalent to the condition @xmath1667 but this shows statement ( i ) of the theorem for a scalar @xmath1668 , and because of the above simplifications statement ( i ) is also proved for @xmath1669 . the second statement where @xmath1062 with @xmath1670 can be proved via the rockland condition in a similar way
: one can use that the irreducible representations of the abelian group @xmath979 are simply the trivial ones . note
that for the case @xmath1671 the second statement can also be proved using theorem [ thm hypoellipticity rothschild stein ] , if we set @xmath1672 in .
this works for both directions of the equivalence we want to prove since in this situation @xmath33 is not a heisenberg group . by a more general version of the theorem by rothschild and stein ( see @xcite ) , the case @xmath1673 is also covered . @xmath91
* remark : * if we look at the proof of this theorem , we note that the strategy is very similar to the strategy of proving theorem [ thm spectral decomposition d^h on heisenberg group ] : in both cases , we make use of the fact that we can describe our operator using the irreducible representations of @xmath33 . while in theorem [ thm spectral decomposition d^h on heisenberg group ] we deduced that the horizontal pull - back dirac operator on a compact carnot nilmanifold has infinite dimensional eigenspaces , we will deduce from this theorem that the square of any horizontal dirac operator ( and hence the operator @xmath1 itself ) on a heisenberg manifold is not hypoelliptic in section 6.3 .
@xmath74 before we close this section , we remark that of course one can combine theorem [ thm hypoellipticity of laplacian codimension 1 ] with theorem [ thm hypoellipticity rothschild stein ] to deduce a better characterization of hypoellipticity for horizontal laplacians of the form , acting on scalar - valued functions .
since this is straight - forward , we will not write it down here . in this section , we want to develop a criterion which we can use to show that any horizontal dirac operator on a carnot manifold is not hypoelliptic .
note that theorem [ thm hypoellipticity rothschild stein ] is not very practical for this situation : first of all , it has to be extended to the case of vector bundles , and even after doing this is seems to be a quite complicated task to work with the spaces @xmath1674 from , which depends heavily on the concrete structure of the carnot group , in a general setting .
but on the other hand theorem [ thm hypoellipticity of laplacian codimension 1 ] can be applied easily to the square of such a horizontal dirac operator , as long as we are in the case of heisenberg manifolds .
thus the idea we want to follow is to reduce the general problem of excluding hypoellipticity on an arbitrary carnot manifold to the co - dimension @xmath98 case .
we have already seen that this approach works for the case of horizontal pull - back dirac operators on compact carnot nilmanifolds in chapter 4 : it was possible to lift the infinite dimensional eigenspaces of @xmath1 to a higher dimensional nilmanifold . in chapter 4
we used the submersions between the corresponding carnot group introduced in section 2.4 , and for our current ( and more general ) situation this strategy will also work .
like in section 6.1 , we consider a graded differential operator with constant coefficients @xmath1675 as an element of the universal enveloping algebra of a graded lie algebra @xmath325 belonging to a carnot group @xmath33 of step @xmath221 .
the ( vector - valued ) functions @xmath1 is acting on are supposed to be defined on a carnot group @xmath33 , realized as a non - abelian group structure on @xmath56 .
we remember the orthogonal projection @xmath1676 from section 2.4 , where @xmath1677 for a linear subspace @xmath1678 for some @xmath418 . by proposition
[ prop properties projection lie algebra ] , @xmath1679 has the structure of a nilpotent graded lie algebra which is induced by @xmath57 . after applying the map @xmath420 to the elements of @xmath57
, we get a new differential operator @xmath1680 considered as an element of the universal enveloping algebra of @xmath60 .
note that of course the operator @xmath1681 is supposed to act on the ( compared to @xmath33 lower dimensional ) carnot group @xmath1682 , realized on the euclidean space @xmath1683 . by proposition
[ prop properties projection lie algebra ] the projection @xmath422 gives rise to a submersive lie group homomorphism @xmath1684 between the carnot groups @xmath1625 and @xmath1685 . the operator @xmath1681 is supposed to act on vector - valued functions living on the carnot group @xmath442 . since @xmath441 is a submersion , a function or distribution on @xmath442 can be extended to a function or distribution on @xmath33 via pullback along @xmath441 .
this observation leads us to the idea to deduce that @xmath5 is not hypoelliptic if @xmath1681 is not hypoelliptic for some projection of the type , as we will do via the following theorem .
[ thm non - hypoellipticity on vector bundles over carnot groups ] let @xmath33 be a carnot group with grading @xmath325 of its lie algebra , and let @xmath1686 be a graded differential operator with constant coefficients .
assume that for some @xmath1687 there is a linear space @xmath1678 such that the corresponding orthonormal projection @xmath1688 defined by furnishes a graded differential operator @xmath1689 which is not hypoelliptic .
then @xmath5 is not hypoelliptic .
* proof : * we write down the argument for the case @xmath1690 where @xmath5 is acting on complex - valued functions functions ; the vector - valued case works analogously .
assume the operator @xmath1691 constructed above is not hypoelliptic , which means that there exists a distribution @xmath1692 such that @xmath1693 , but @xmath1694 for the carnot group @xmath1695 .
we have to find a @xmath1696 which is not a @xmath339-function , such that @xmath1697 .
we consider the submersive lie group homomorphism @xmath1698 arising from @xmath420 , see .
in addition we use the projection @xmath1699 , which is given by @xmath1700 , of @xmath57 onto the kernel of @xmath422 . then from proposition [ prop properties projection lie algebra ] we get an isomorphism @xmath1701 from @xmath441 , where @xmath1702 and @xmath1703 is the projection onto the kernel of @xmath441 , arising from @xmath1704 . from now on , we will use the coordinates @xmath1705 on @xmath430 which are induced by the isomorphism @xmath558 from .
if @xmath1706 are exponential coordinates ( corresponding to a vector field @xmath299 ) on @xmath33 , then we have @xmath1707 as one sees immediately from the definition of @xmath558 via @xmath422 .
but from these exponential coordinates we see that for the differential of @xmath558 we have @xmath1708 and hence we see how the application of a vector field @xmath299 to a function @xmath944 carries over to the push - forward @xmath1709 of @xmath63 on @xmath1710 : we have @xmath1711 where @xmath1712 . from now on , we will consider functions and distributions on @xmath33 as functions and distributions on @xmath1710 , where the identification is given via the isomorphism @xmath558 .
after these preparations , we are ready to prove the theorem .
let @xmath1692 be given such that @xmath1693 , but @xmath1713 for our graded differential operator @xmath5 .
for a test function @xmath1714 we can define the push - forward @xmath1715 of @xmath63 along @xmath441 via @xmath1716 where @xmath1717 is the haar measure on the nilpotent lie group @xmath226 .
note that the expression is well - defined since @xmath226 is a normal subgroup of @xmath33 as the kernel of the lie group homomorphism @xmath441 , and that @xmath1715 is a @xmath339-function with compact support on @xmath442 since this is the case for @xmath63 on @xmath1710 .
but this means we get a distribution @xmath1718 via @xmath1719 obviously , since we assumed @xmath1693 , we have @xmath1720 .
we will show now that @xmath1721 , from which the theorem will be proved . if we apply a vector field @xmath299 to the distribution , which means by to apply @xmath1722 to @xmath1718 , we get from @xmath1723 \right\rangle \end{split}\ ] ] for any test function @xmath1714 .
now for the push - forward of @xmath1724 to @xmath442 via @xmath441 we calculate , using , @xmath1725(\tilde{x } ) & = & \int_n \left({\mathrm{pr}}(x ) , { \mathrm{pr}}^\perp(x)\right)f(\tilde{x},n ) d\mu(n)\\ & = & \int_n \left .
\frac{d}{dt } f\left(\tilde{x } \ , ._{\tilde{{\mathbb{g } } } } \ , \exp_{\tilde{{\mathbb{g } } } } t { \mathrm{pr}}(x ) , \
n \ , ._n \ , \exp_n t { \mathrm{pr}}^\perp(x ) \right ) \right|_{t=0 } d\mu(n ) \\ & = & \left
. \frac{d}{dt } \int_n f\left(\tilde{x } \ , ._{\tilde{{\mathbb{g } } } } \ , \exp_{\tilde{{\mathbb{g } } } } t { \mathrm{pr}}(x ) , \
n \ , ._n \ , \exp_n t { \mathrm{pr}}^\perp(x )
\right ) d\mu(n ) \right|_{t=0 } \\ & = & \left .
\frac{d}{dt } \int_n f\left(\tilde{x } \ , ._{\tilde{{\mathbb{g } } } } \ , \exp_{\tilde{{\mathbb{g } } } } t { \mathrm{pr}}(x ) , \
n \right ) d\mu(n ) \right|_{t=0 } \\
& = & \left.\frac{d}{dt } \psi_\ast f \left(\tilde{x } \ , ._{\tilde{{\mathbb{g } } } } \ , \exp_{\tilde{{\mathbb{g } } } } t { \mathrm{pr}}(x)\right ) \right|_{t=0}\\ & = & { \mathrm{pr}}(x ) \psi_\ast f(\tilde{x } ) .
\end{aligned}\ ] ] here we are allowed to interchange differentiation and integration since @xmath63 is a test function , and the forth equation is true since the haar measure is translation invariant .
the same calculation works if we apply another vector field @xmath1726 to @xmath1727 , which provides @xmath1728(\tilde{x } ) = { \mathrm{pr}}(y ) { \mathrm{pr}}(x ) \psi_\ast f(\tilde{x}),\ ] ] and we can go on inductively for higher order applications of vector fields .
now we can apply these identities to to find that @xmath1729 and because of the same identity is true if we apply further vector fields to @xmath1727 .
but this means that for any graded differential operator @xmath5 we get @xmath1730 now we assumed @xmath1731 to be a smooth function . but
from this it follows immediately from that @xmath1732 is a smooth function on @xmath1710 : we have by this equation and by the definition of @xmath1715 @xmath1733 where @xmath1734 and @xmath1735 are the corresponding haar measures on @xmath226 and @xmath442 .
but this shows that we must have @xmath1736 and since @xmath1737 is smooth this must also be the case for @xmath1732 .
altogether we have found a distribution @xmath507 on @xmath430 , for which we have @xmath1697 , but @xmath1738 , whenever there is a projection @xmath422 of the type such that @xmath1681 is not hypoelliptic . but
this shows that the graded differential operator @xmath5 is not hypoelliptic in this situation , such that the theorem is proved .
@xmath91 we did the proof of theorem [ thm non - hypoellipticity on vector bundles over carnot groups ] in the very general setting of distributions .
if in a more specific sense the operator @xmath1681 is assumed to be hypoelliptic in the sense that there is a function @xmath1739 on @xmath442 , which lies in the domain of @xmath5 but which is not @xmath339 , such that @xmath1731 , then the hypoellipticity of @xmath5 is neglected by the function @xmath1740 on @xmath33 .
the calculation is straight forward , making use of the isomorphism @xmath430 like in the proof of the above theorem . via the same calculation
one can show that @xmath1741 implies @xmath1742 .
but from this , we immediately get theorem [ thm d^h has infinite dimensional kernel on any homogeneous carnot space ] , which states the degeneracy of the horizontal pull - back operator on any compact carnot nilmanifold ( if we use the degeneracy in the heisenberg case ) .
this is no surprise , since the techniques of lifting an operator from a low - dimensional carnot manifold to a higher dimensional one work very similar in both cases . in this way , the results from this chapter are the more general ones , while we get some extra features ( like the detection of the metric dimension ) from the specific examples in chapter 4 which we were not able to derive in the general case .
@xmath74 in the next section ,
we want to use theorem [ thm non - hypoellipticity on vector bundles over carnot groups ] to show that any horizontal dirac operator constructed in section 3.2 is not hypoelliptic .
the idea is to use theorem [ thm non - hypoellipticity on vector bundles over carnot groups ] after reducing the operator to an operator acting on a carnot group of co - dimension @xmath98 , which will be assumed to be of the type @xmath393 . for this operator ,
the hypoellipticity can be neglected by theorem [ thm hypoellipticity of laplacian codimension 1 ] .
like before , let @xmath1743 be the dimension of @xmath287 and let @xmath327 be an orthonormal basis of @xmath287 for the grading @xmath325 of @xmath57 .
then a projection of @xmath57 onto a graded nilpotent lie algebra of co - dimension @xmath98 can be constructed as follows : for any @xmath407 , consider the @xmath98-dimensional linear subspace @xmath1744 of @xmath1333 .
we then define @xmath1745 , such that our orthogonal projection becomes @xmath1746 we can apply proposition [ prop properties projection lie algebra ] to @xmath1375 , which gives us graded nilpotent lie algebras @xmath1747 and @xmath1748 with corresponding carnot groups @xmath1336 and @xmath1749 .
now let @xmath1750 + o_h(1 ) \ \ \ \in \mathcal{u}(\mathfrak{g } ) \otimes { \mathbb{c}}^q\ ] ] be a horizontal laplacian acting on a vector bundle @xmath585 of rank @xmath338 , which is considered as an element of the universal enveloping algebra of @xmath57 tensored with @xmath1751 .
the term @xmath592 denotes a graded differential operator of order smaller or equal to @xmath98 ( which is a first order differential operator depending only on the vector fields @xmath214 ) .
the @xmath1752 are @xmath1753-matrices with complex valued entries .
using the lie algebra homomorphism @xmath1375 from we can define a horizontal laplacian @xmath1754 , acting on a carnot group with a horizontal distribution @xmath1755 of co - dimension @xmath98 : for @xmath1756 we get @xmath1757 \right ) + o_h(1 ) \ \ \
\in \mathcal{u}(\mathfrak{g}_{2,\nu } ) \otimes { \mathbb{c}}^p.\ ] ] after calculating the commutators we have an @xmath1758 such that ( for @xmath336 ) @xmath1759 with @xmath1760 , which is a horizontal laplacian for which its hypoellipticity can be determined by theorem [ thm hypoellipticity of laplacian codimension 1 ] .
this argument enables us to formulate non - hypoellipticity criteria involving the @xmath408-levi form introduced in section 2.3 : in the above situation , for @xmath407 this is the bilinear form @xmath1761 \mod \left ( { \mathrm{span}}\{x_{2,\nu}\ } \right)^\perp.\ ] ] if we fix a basis for @xmath335 , @xmath413 is described by an antisymmetric matrix @xmath1762 , such that @xmath1763 .
now we can use theorem [ thm non - hypoellipticity on vector bundles over carnot groups ] together with theorem [ thm hypoellipticity of laplacian codimension 1 ] to formulate the following criterion .
[ cor non - hypoellipticity from reduced group of codimension 1 ] let @xmath33 be a carnot group with corresponding lie algebra @xmath325 . for any number @xmath1764
, we consider the projection @xmath1337 from of @xmath57 onto the lie algebra @xmath1765 together with its corresponding carnot group @xmath1336 .
let @xmath1766 be a horizontal laplacian of the type , such that the horizontal laplacian @xmath1767 on @xmath1336 is given via @xmath1768 like in , where @xmath1758 and @xmath592 denotes a graded differential operator of order smaller or equal to @xmath98 .
assume there is a @xmath1764 such that there is an eigenvector of @xmath1769 which is contained in the singular set @xmath1770 defined as follows : a. if @xmath1771 is isomorphic to the @xmath24-dimensional heisenberg group ( with @xmath1193 ) , we define @xmath1772 where @xmath399 denote the non - zero eigenvalues and @xmath1773 denotes the trace norm of the @xmath408-levi form @xmath1774 .
b. if @xmath1062 with @xmath1195 is not isomorphic to a heisenberg group , then we define @xmath1775 \cup \left [ \frac{1}{2 } \left\|l_\nu\right\|_1 , \infty \right),\ ] ] where @xmath1776 denotes the trace norm of the @xmath408-levi form @xmath1774 .
then @xmath1543 is not hypoelliptic .
* remark : * note that for an arbitrary @xmath1777 , the cases @xmath1778 and @xmath1779 , with @xmath401 , are indeed the only possible cases since @xmath1336 is a carnot group of step @xmath53 and horizontal co - dimension @xmath98 .
* proof : * by theorem [ thm non - hypoellipticity on vector bundles over carnot groups ] , @xmath1543 is not hypoelliptic if @xmath1780 is not hypoelliptic for some @xmath1764 . but to the operator @xmath1780 we can apply theorem [ thm hypoellipticity of laplacian codimension 1 ] , which states that the operator @xmath1780 is not hypoelliptic if and only if there is an eigenvalue of @xmath1769 belonging to one of the singular sets @xmath1770 from ( i ) and ( ii ) , depending on how the structure of @xmath1336 concretely looks like .
@xmath91 from corollary [ cor non - hypoellipticity from reduced group of codimension 1 ] , we can immediately derive a simple criterion which ensures us that a horizontal laplacian is not hypoelliptic if one of the matrices from has a certain eigenvalue .
in detail we can use the matrices @xmath1781 of the @xmath408-levi forms @xmath413 from according to the frame @xmath41 of @xmath335 , such that for all @xmath531 the commutators have the form @xmath1782 = \sum_{\nu_1}^{d_2 } l_{jk}^{(\nu ) } x_{2,\nu}.\ ] ] thus the horizontal laplacian can be rewritten in the form @xmath1783 such that the matrices @xmath1769 are given by @xmath1784 but from these matrices one can check immediately that a given horizontal laplacian is not hypoelliptic : [ cor non - hypoellipticity by eigenvalue of matrix ] if there is a @xmath1764 such that there is an eigenvalue @xmath313 of the @xmath1753-matrix @xmath1769 from with @xmath1785 then the operator @xmath1543 is not hypoelliptic . here , for @xmath630
the numbers @xmath1786 are supposed to be the non - zero eigenvalues of the @xmath408-levi form @xmath413 of @xmath33 .
* proof : * using the projection @xmath1375 for any @xmath407 from , we get from @xmath1787 now we can apply corollary [ cor non - hypoellipticity from reduced group of codimension 1 ] to this operator and see that it is not hypoelliptic if one of the eigenvalues of @xmath1769 belongs to one the sets @xmath1770 from corollary [ cor non - hypoellipticity from reduced group of codimension 1 ] . since the numbers @xmath1788 are included in both sets , the operator @xmath1543 fails to be hypoelliptic if one of them is an eigenvalue of one of the matrices @xmath1769 , no matter how the carnot groups @xmath1789 looks like . @xmath91
using the preparing work done in the last section we are finally ready to show that the horizontal dirac operators we constructed in chapter 3 can not be hypoelliptic .
we have constructed our horizontal dirac operators @xmath1 more or less analogously to classical dirac operators , acting on a horizontal clifford bundle which is arising from the natural horizontal connection on a carnot manifold @xmath0 .
we have seen that these operators detect the carnot - carathodory metric and that they are therefore also a generalization of the classical case from the connes metric point of view .
if @xmath1 would be hypoelliptic , it would follow that it has a compact resolvent by heisenberg calculus ( which is developed for the case of heisenberg manifolds in chapter 5 ) , and hence it would give a spectral triple .
but this is not the case : we will prove that @xmath1 is not hypoelliptic , and therefore the machinery of graded pseudodifferential calculus does not work .
we have already seen this in detail for a specific example in chapter 4 , and from the following theorems it will follow that these results fit into the general situation . to prove the non - hypoellipticity of @xmath1 , we will consider its square calculated locally in proposition [ prop d^h squared ] from chapter 3 .
this strategy is justified by the following simple observation .
[ prop hypoellipticity d and d square ] let @xmath5 be a ( pseudo-)differential operator acting on a vector bundle @xmath585 over a manifold @xmath0 .
then @xmath5 is hypoelliptic of and only if @xmath1790 is hypoelliptic .
* proof : * assume @xmath5 is hypoelliptic , which means for any distribution @xmath507 we have that @xmath1791 implies @xmath1792 .
this gives the implication @xmath1793 and hence @xmath1790 is hypoelliptic . on the other hand ,
let @xmath1794 imply that @xmath1792 .
we have to show that @xmath5 is hypoelliptic . but
this follows immediately since for @xmath1791 , we also have @xmath1795 , which shows @xmath1792 by assumption . altogether the proposition is proved .
@xmath91 since @xmath609 is a horizontal laplacian , we can use theorem [ thm hypoellipticity of laplacian codimension 1 ] to decide about its hypoellipticity for the case it is acting on a heisenberg manifold . for the case of a general carnot manifold
, we can use the criterion developed in section 6.2 to get a corresponding horizontal laplacian on a heisenberg manifold to which we can apply theorem [ thm hypoellipticity of laplacian codimension 1 ] .
we have already fulfilled this reduction step in section 6.2 , such that we can simply use the corollaries [ cor non - hypoellipticity from reduced group of codimension 1 ] and [ cor non - hypoellipticity by eigenvalue of matrix ] to prove the following theorem .
[ thm horizontal dirac is not hypoelliptic ] let @xmath0 be a carnot manifold with horizontal distribution @xmath19 , equipped with a horizontal clifford bundle @xmath1796 arising from the horizontal connection on @xmath0 , and let @xmath1 be any horizontal dirac operator acting on @xmath533 . then @xmath1 is not hypoelliptic .
* proof : * assume the grading of @xmath20 is given by @xmath1797 and that we have a riemannian metric on @xmath0 such that these sub - bundles are point - wise orthogonal to each other .
further we assume @xmath41 ( with @xmath1798 ) to be an orthonormal frame for @xmath19 and @xmath406 to be an orthonormal frame for @xmath378 ( with @xmath1799 ) . after fixing a @xmath1800
we can assume the frame of @xmath19 to have the additional property that the matrix @xmath1801 describing the @xmath1802-levi form @xmath1803 with respect to this frame is given by @xmath1804 where @xmath5 is a diagonal matrix carrying the absolute values @xmath398 of the non - zero eigenvalues of @xmath1803 on its diagonal .
this can always be achieved by on orthonormal transformation of the horizontal frame because any @xmath408-levi matrix is skew symmetric . for the proof of this theorem ,
we work with the expression of @xmath1 according to this horizontal frame @xmath41 , see e.g. equation from theorem [ thm d^h selfadjoint ] .
we show that @xmath609 is not hypoelliptic , then the non - hypoellipticity of @xmath1 follows from proposition [ prop hypoellipticity d and d square ] .
note that it suffices to show the non - hypoellipticity locally in an environment of any point @xmath215 . by proposition
[ prop d^h squared ] we have locally @xmath1805 + o_h(1)\ ] ] with horizontal clifford action @xmath1806 on the horizontal clifford bundle @xmath533 .
as before @xmath324 is to be understood as a component wise directional derivative in a local chart and @xmath592 denotes a graded differential operator of order smaller than or equal to @xmath98 . for any @xmath1807 let @xmath1781 denote the matrix of the @xmath408-levi form corresponding to the frame @xmath41 of @xmath376 .
hence for any pair @xmath531 we have @xmath1782 = \sum_{\nu=1}^{d_2 } l_{j , k}^{(\nu ) } x_{2,\nu},\ ] ] and plugging this into we find that @xmath1808 now we can use corollary [ cor non - hypoellipticity by eigenvalue of matrix ] applied to the matrix @xmath1809 where @xmath1800 is the number we chose in the beginning of the proof such that the @xmath1802-levi matrix @xmath1810 has the form . but
this means that we have @xmath1811 where @xmath398 denote the absolute values of the non - zero eigenvalues of @xmath413 .
according to corollary [ cor non - hypoellipticity by eigenvalue of matrix ] , we have to show that there is an eigenvalue of the matrix @xmath1769 from with absolute value @xmath1812 .
but this has already been done by proposition [ prop eigenvalues sum c(x_j)c(x_k ) ] from section 3.2 : this proposition states that @xmath633 and @xmath634 are eigenvalues of the matrix @xmath1813 .
hence the theorem is proved because of corollary [ cor non - hypoellipticity by eigenvalue of matrix ] .
@xmath91 * remark : * the property of being hypoelliptic only depends on the leading term of the local expression for the square of @xmath1 . as we already noted in the remark after proposition [ prop d^h squared ] , this term does not change if we modify @xmath1 by adding a section of the endomorphism bundle of @xmath533 , meaning an order zero term in the language of differential operators .
hence there is no chance of getting a hypoelliptic first order horizontal differential operator on @xmath536 via a modification of the connection on @xmath533 by an endomorphism .
@xmath74 in chapter 5 , theorem [ thm complex powers of heisenberg pseudos ]
we have seen that on a compact heisenberg manifold @xmath0 we need the condition of being hypoelliptic to get a resolvent for a heisenberg pseudodifferential operator of positive order which is compact . on the other hand ,
it is clear that from the existence of a compact resolvent of a differential operator @xmath5 of order @xmath648 one can expect this operator to be hypoelliptic : for every @xmath77 , @xmath5 can be extended to a bounded operator @xmath1814 between the @xmath52-sobolev spaces @xmath1815 and @xmath1816 . hence we expect a ( compact ) resolvent of @xmath5 to be a mapping from @xmath1816 to @xmath1815 for any @xmath77 .
now assume @xmath5 is not hypoelliptic , i.e. that there is an element @xmath1817 such that @xmath1818 . but
this means that the resolvent of @xmath5 maps the @xmath339-function @xmath1732 ( which is an element of @xmath1815 for every @xmath77 ) to a distribution which does not belong to @xmath1816 for one @xmath77 .
this is a contradiction . from the above discussion
, we can formulate the following corollary .
[ cor horizontal dirac has no compact resolvent ] let @xmath0 be a carnot manifold with horizontal distribution @xmath19 , equipped with a horizontal clifford bundle @xmath1796 arising from the horizontal connection on @xmath0 , and let @xmath1 be any horizontal dirac operator acting on @xmath533 .
then @xmath1 does not have a compact resolvent .
we emphasize once again that theorem [ thm horizontal dirac is not hypoelliptic ] is true for any example of a horizontal dirac operator according to the horizontal levi - civita connection on @xmath19 one can imagine .
it is true for the horizontal pull - back dirac operators we discussed in chapter 4 , and it is also true for horizontal dirac operators defined on classical clifford or spinor bundles ( by skipping the non - horizontal derivatives ) , see proposition [ prop horizontal clifford bundle from spinor bundle ] . in particular , it is true for the operator @xmath1819 , where @xmath1820 is the horizontal exterior derivative ( see example [ ex exterior bundle ] ) . note that all these constructions are based on the horizontal connection on @xmath0 , which is induced by the levi - civita connection .
this shows that the degeneracy we detected in chapter 4 for the horizontal pull - back dirac operator on compact carnot nilmanifolds is not because of a bad choice for @xmath1 .
rather it reflects a general phenomenon : the natural differential operator , which detects the carnot - carathodory metric on a carnot manifold via connes formula , does not give a spectral triple , and the classical construction of a spectral triple on a compact spin manifold can not be transported to the carnot case .
in the last chapter we saw that the canonical candidate for a spectral triple over a carnot manifold detecting the horizontal geometry , the horizontal dirac operator @xmath1 , reproduces the carnot - carathodory metric but is not hypoelliptic , and therefore does not have a compact resolvent in the heisenberg calculus .
but on the other hand the classical hypoellipticity criteria imply that there are a lot of horizontal laplacians which are hypoelliptic , and in addition ( at least in the heisenberg case ) give back the hausdorff dimension of @xmath22 via their eigenvalue asymptotics .
now this chapter is devoted to studying horizontal laplacians and to discussing how they furnish the geometry in the sense of connes . in the first section
we show explicitly how one can construct a spectral triple from a positive hypoelliptic horizontal laplacian using the heisenberg calculus .
we present some operators which are induced by a small perturbation of the square of a horizontal dirac operator .
then in section 7.2 we will show that any horizontal laplacian detects the carnot - carathodory metric of a carnot manifold via a formula similar to connes one .
we also discuss what this means for a first order operator whose square is a horizontal laplacian .
the last section of this chapter is rather speculative : we intend to give some ideas how one can find estimates for the connes metric of spectral triples arising from horizontal laplacians towards the carnot - carathodory metric ( which is exactly detected by a horizontal dirac operator @xmath1 ) .
therefore we use some of the observations we made in section 1.2 concerning the convergence of the metrics belonging to compact quantum order unit spaces .
but sadly we have not been able to prove such an estimate , so there are a few open problems formulated in that section .
we now point out how we can construct spectral triples on compact heisenberg manifolds using the heisenberg calculus we introduced in chapter 5 .
the idea is to start with a hypoelliptic , positive horizontal laplacian @xmath1543 and to consider the operator @xmath1821 , defined via functional calculus . to show that such a construction indeed furnishes a spectral triple
, we check that the commutator of any heisenberg pseudodifferential operator of order @xmath98 with a smooth function @xmath63 is bounded .
[ prop boundedness of commutator ] let @xmath0 be a heisenberg manifold and @xmath1822 a heisenberg pseudodifferential operator of order @xmath98 acting on a vector bundle @xmath585 over @xmath0 . for any function @xmath494 we denote by @xmath1823 the operator of multiplication by @xmath63 on the hilbert space @xmath1512
. then the commutator @xmath1824 $ ] is bounded .
* proof : * we assume that @xmath585 is the trivial line bundle over @xmath0 , the general case works analogously .
our strategy is to show that the symbol of the commutator @xmath1824 $ ] lies in @xmath1825 , and hence provides a heisenberg pseudodifferential operator of order @xmath596 .
since @xmath1826 for the hrmander class @xmath1827 by theorem [ thm embedding of s_hei^m ] , this implies by classical pseudodifferential calculus that @xmath1824 $ ] is bounded ( see also corollary [ cor heisenberg pseudos regularity ] ) .
first of all we note that the multiplication operator @xmath1823 is a heisenberg pseudodifferential operator of order @xmath596 , and its symbol is exactly the function @xmath63 by the definition of the symbol classes ( see definition [ def symbol classes heisenberg calculus ] ) .
let @xmath1828 denote the symbol of @xmath1522 .
then we can calculate the symbol of the commutator @xmath1824 $ ] using theorem [ thm composition of heisenberg pseudos ] : for general heisenberg symbols @xmath1829 of order @xmath1830 and @xmath1831 of order @xmath1832 , the asymptotic expansion of the symbol @xmath1833 denoting the symbol of the composition of @xmath1834 , is given by the terms @xmath1835 where @xmath1495 denotes the sum over all the indices such that @xmath1836 and the functions @xmath1497 are polynomials in the derivatives of the coefficients of the vector fields @xmath1498 forming a graded frame for @xmath20 .
the operation @xmath1837 denotes the operation of point - wise convolution , varying smoothly over @xmath107 ( see equations and from section 5.1 ) . see the discussion before theorem [ thm composition of heisenberg pseudos ] for a more detailed explanation . to get the boundedness of the commutator we only have to show that the order @xmath98 part of this asymptotic expansion vanishes , since all the lower order terms have maximal heisenberg order @xmath596 which gives rise to a bounded operator by corollary [ cor heisenberg pseudos regularity ] . since @xmath1838 in our situation , these are exactly the terms in such that @xmath1839 . but
this means @xmath1840 and therefore also @xmath1841 by .
now the heisenberg symbol of the commutator @xmath1824 = pf - fp$ ] is given by @xmath1842 ) = p \ # f - f \ # p,\ ] ] and for its leading ( heisenberg order @xmath98 ) symbol we have because of the above argumentation @xmath1843 where @xmath1844 denotes the leading term of the heisenberg symbol of @xmath1522 .
note that the polynomial @xmath1845 only depends on the coefficients of the vector fields @xmath1652 , but not on the symbols of the operators to be composed .
hence all we have to show is that the convolution of @xmath1844 and @xmath63 commutes .
but this is clear by the definition of the convolution : whenever we fix an @xmath215 , the terms commute point - wise in @xmath107 since in this case @xmath1846 is a constant .
now since @xmath1837 varies smoothly in @xmath107 , this shows that @xmath1847 and from we get @xmath1848 .
we have seen that the leading term in the asymptotic expansion of the symbol belonging to the commutator @xmath1824 $ ] has mostly heisenberg order zero , and we can conclude that this commutator is bounded by corollary [ cor heisenberg pseudos regularity ] . therefore the proposition is proved .
@xmath91 to check the compactness of the resolvent , we refer to theorem [ thm complex powers of heisenberg pseudos ] , which states that for a hypoelliptic self - adjoint @xmath1469 @xmath1522 of order @xmath408 the operator @xmath1568 is a @xmath1469 of order @xmath1569 .
then it is clear that @xmath1469s of order one can provide spectral triples .
[ thm spectral triple from operator heisenberg order 1 ] let @xmath0 be a compact heisenberg manifold , and let @xmath1522 be a hypoelliptic self - adjoint @xmath1469 of order @xmath98 acting on a vector bundle @xmath585 over @xmath0 , which is bounded from below . statement ( i ) is immediately clear from proposition [ prop boundedness of commutator ] , since @xmath96 is a dense sub - algebra of @xmath45 . and
statement ( ii ) follows immediately from theorem [ thm complex powers of heisenberg pseudos ] : by this theorem , the operator @xmath1854 exists in the heisenberg calculus and is a @xmath1469 of order @xmath1400 .
since @xmath0 is compact , this is a compact operator by corollary [ cor heisenberg pseudos of negative order are compact ] , and therefore @xmath1522 possesses a compact resolvent .
@xmath91 from the above considerations we find that it is in general possible to construct spectral triples via the heisenberg calculus .
we will see that concrete examples can be comfortably derived from horizontal laplacians , and by the results stated in section 5.2 concerning the eigenvalue asymptotics we will also see that the metric dimension of these spectral triples provides the hausdorff dimension of @xmath22 .
[ thm spectral triple from horizontal laplacian ] let @xmath0 be a compact heisenberg manifold equipped with a horizontal distribution of rank @xmath216 , and let @xmath1543 be a hypoelliptic self - adjoint horizontal laplacian which is bounded from below , acting on a vector bundle @xmath585 over @xmath0 . then the triple @xmath1855 is a spectral triple of metric dimension @xmath1285 . *
proof : * without loss of generality we assume @xmath1543 to be invertible ( since it is bounded from below , this can always be achieved by adding a constant ) .
thus , by theorem [ thm complex powers of heisenberg pseudos ] , @xmath1856 exists and is a @xmath1469 of order one , which is hypoelliptic , self - adjoint and positive .
then the statement that is a spectral triple follows from theorem [ thm spectral triple from operator heisenberg order 1 ] . to get the additional statement about the metric dimension of this spectral triple
, we can apply theorem [ thm asymptotic behaviour of eigenvalues horizontal laplacian ] , which says that the eigenvalues of @xmath1543 have the asymptotic behavior @xmath1857 which gives us @xmath1858 this asymptotic behavior shows that @xmath1859 is trace class if and only if @xmath1860 .
thus the metric dimension of the spectral triple is @xmath1285 , which is also the hausdorff dimension of @xmath22 by theorem [ thm mitchells measure theorem ] .
@xmath91 from now on , we consider once again the orthonormal frame @xmath1467 of the tangent bundle of our heisenberg manifold @xmath0 , such that @xmath41 span the horizontal distribution @xmath19 .
we assume that the levi form of @xmath0 according to this frame is given by the matrix @xmath1861 for an @xmath1862 .
note that such a levi - form can always be achieved by an orthonormal change of the frame ( see proposition [ prop tangent lie group bundle and rank of levi form ] ) .
but this means we have @xmath1062 for the underlying carnot group , and the commutator relation of the vector fields belonging to the frame of @xmath0 become @xmath1782 = \begin{cases } \lambda_j , & \text{for } \ , 1 \leq j \leq m , \ , k = m+j \\
-\lambda_k & \text{for } \ , 1 \leq k \leq m , \ , j = k+m\\ 0 & \text{otherwise } \end{cases}.\ ] ] now an obvious candidate for a horizontal laplacian which provides a spectral triple is the sum - of - squares operator @xmath1863 where @xmath474 is any connection acting on a vector bundle @xmath585 over @xmath0 .
this operator is obviously self - adjoint , and after applying hrmander s sum - of - squares theorem ( see theorem [ thm hormander sum of squares ] ) in any local chart we see that is is hypoelliptic .
it is also known that this operator is bounded from below , as it is the case for any self - adjoint horizontal laplacian , whose principal symbol is invertible in the heisenberg calculus . see @xcite , remark 5.2.10 for that .
as we have seen in the previous chapters , the horizontal dirac operator @xmath1 detecting the carnot - carathodory metric on @xmath0 is not hypoelliptic and therefore does not provide a spectral triple . to see why the hypoellipticity fails , we take a look at the proof of theorem [
thm horizontal dirac is not hypoelliptic ] adapted to this situation : the square of @xmath1 is ( locally ) given by @xmath1864 and this operator fails to be hypoelliptic since @xmath1865 and @xmath1866 are eigenvalues of the matrix @xmath1867 .
this is the content of theorem [ thm hypoellipticity of laplacian codimension 1 ] , but the same theorem also states that the operator from would be hypoelliptic if the absolute value of every eigenvalue of the matrix coefficient of @xmath1066 would be smaller than @xmath1812 .
since this is fulfilled for any other eigenvalue of @xmath1867 ( see proposition [ prop eigenvalues sum c(x_j)c(x_k ) ] ) , one can disturb this operator by a small number @xmath1868 and conclude that the operator @xmath1869 is hypoelliptic for any @xmath1870 .
we thus construct a spectral triple which shall be a small perturbation of the horizontal dirac operator , using the sum - of - squares operator from .
the connection we are using is the horizontal clifford connection @xmath538 from section 3.2 , acting on a clifford bundle @xmath533 over @xmath0 on which the horizontal dirac operator @xmath1 is defined .
[ cor disturbed spectral triple on heisenberg manifold ] let @xmath0 be a closed heisenberg manifold , and let @xmath533 be a horizontal clifford bundle over @xmath0 on which a horizontal dirac operator @xmath1 is acting .
we further define the horizontal laplacian @xmath1871 then for any @xmath1870 the operator @xmath1872 provides a spectral triple @xmath1873 of metric dimension @xmath1285 . *
proof : * we have to show that for any @xmath1874 $ ] the operator @xmath1875 is positive ( which also means that is well - defined ) and hypoelliptic .
in this case we can apply theorem [ thm spectral triple from horizontal laplacian ] to get that @xmath1876 provides a spectral triple of metric dimension @xmath1285 and the corollary is proved .
we assume that @xmath1877 , which can always be achieved by adding a constant because it is bounded from below .
since the horizontal dirac operator @xmath1 is self - adjoint by theorem [ thm d^h selfadjoint ] , its square @xmath590 is positive , and therefore @xmath1878 from is positive as a convex combination of positive operators . to show that @xmath1878 is hypoelliptic , we note that @xmath1879 we consider this operator in the local frame @xmath1212 introduced above , such that @xmath213 and the matrix of the levi form corresponding to this basis is given by .
but in these local coordinates , the argument we sketched in the discussion preceding this corollary applies : we have locally @xmath1880 and therefore @xmath1881 by the local expression of @xmath609 from proposition [ prop d^h squared ] .
hence has locally the form @xmath1882 where like in the previous chapters @xmath592 denotes a graded differential operator of order smaller than or equal to @xmath98 .
now by proposition [ prop eigenvalues sum c(x_j)c(x_k ) ] , the eigenvalues of the matrix @xmath1883 are included in the interval @xmath1884 \ \subset { \mathbb{r}}i\ ] ] on the imaginary line , and since @xmath1885 the absolute value of any of these eigenvalues is smaller than @xmath1886 . but this means by theorem [ thm hypoellipticity of laplacian codimension 1 ] that @xmath1878 is ( locally ) hypoelliptic .
after we have constructed spectral triples from horizontal laplacians , we ask ourselves whether one can detect the carnot - carathodory metric from these spectral triples . while we postpone the discussion of estimates for the connes metric formula of these operators to section 7.3 , we want to show now that there is a formula to detect the carnot - carathodory metric directly from any horizontal laplacian . in the case where we have an ordinary laplacian @xmath1887 ,
it is known that for @xmath494 we have @xmath1888,f \right ] = -2 \left\|df\right\|^2,\ ] ] see for example @xcite , proposition 2.3 .
now an analogous result is true for horizontal laplacians : the key observation is the following lemma .
[ lemma double commutator from horizontal laplacian ] let @xmath1543 be a horizontal laplacian , acting on a vector bundle @xmath585 over a carnot manifold @xmath0
. then we have for any function @xmath494 @xmath1889,f\right ] \sigma = - \left\| { \mathrm{grad}}^h f \right\|^2 \sigma\ ] ] for @xmath1890 . here
, @xmath356 denotes the ( point - wise ) norm of a vector in @xmath1891 induced by the riemannian metric @xmath247 on @xmath0 .
* proof : * let @xmath1892 be the grading of @xmath0 , such that @xmath41 is an orthonormal frame of @xmath376 and @xmath1893 is an orthonormal frame of @xmath378 .
we prove the statement locally , meaning that @xmath1543 is given in the form @xmath1894 where @xmath1895 denotes the partial derivative along @xmath840 in any chart , @xmath1896 is any differential operator of ( classical ) order @xmath98 and @xmath1897 is a ( matrix - valued ) function .
we now plug the expression term by term into the double commutator of and use the linearity of the commutator . since @xmath1897 commutes with any smooth function @xmath63 ,
this is zero for the last summand .
it is also zero for the second summand of , since for any first order differential operator @xmath1898 we have because of the leibniz rule @xmath1899,f\right]\sigma & = & \left[bf - fb , f\right ] \sigma\\ & = & b(f^2 \sigma ) - 2f \cdot b(f\sigma ) + f^2 \cdot b\sigma\\ & = & 2(bf)\cdot f\sigma + f^2 b\sigma - 2f \cdot ( bf ) \cdot
\sigma -2f^2 \cdot b\sigma + f^2 \cdot b\sigma\\ & = & 0\end{aligned}\ ] ] locally for any section @xmath1890 .
calculating the first term of , we find that for the double commutator applied to any @xmath1900 we have @xmath1901,f\right ] \sigma \\ & = & \partial_{x_j}^2 ( f^2\sigma ) - 2f \cdot \partial_{x_j}^2(f\sigma ) + f^2 \cdot \partial_{x_j}^2 \sigma\\ & = & \partial_{x_j } \left(2(\partial_{x_j}f ) \cdot f\sigma + f^2 \cdot \partial_{x_j}\sigma\right ) -2f
\cdot \partial_{x_j } \left((\partial_{x_j}f ) \cdot \sigma + f \cdot \partial_{x_j } \sigma\right ) + f^2 \cdot \partial_{x_j}^2 \sigma \\
& = & 2(\partial_{x_j}^2f ) \cdot f\sigma + 2(\partial_{x_j } f)^2 \cdot \sigma + 4 ( \partial_{x_j } f ) \cdot f \cdot \partial_{x_j } \sigma + f^2 \cdot \partial_{x_j}^2 \sigma\\ & & -2f \cdot ( \partial_{x_j}^2 f ) \cdot \sigma -4f \cdot ( \partial_{x_j } f ) \cdot ( \partial_{x_j } \sigma ) -2f^2 \cdot \partial_{x_j}^2 \sigma + f^2 \cdot \partial_{x_j}^2 \sigma\\ & = & 2(\partial_{x_j } f)^2 \cdot \sigma.\end{aligned}\ ] ] finally , we plug everything together into the double commutator expression , which shows us that @xmath1902,f\right ] \sigma & = & -\frac{1}{2}\sum_{j=1}^d \left[\left[\partial_{x_j}^2,f\right],f\right ] \sigma \\ & = & -\sum_{j=1}^d ( \partial_{x_j } f)^2 \cdot \sigma\\ & = & - \left\| { \mathrm{grad}}^h f \right\|^2 \sigma.\end{aligned}\ ] ] hence the statement of the lemma is proved .
@xmath91 from the identity
we get an expression depending on a horizontal laplacian instead of a horizontal dirac operator which can detect the carnot - carathodory metric , following the theory from section 3.3 . by corollary
[ cor d_cc from d^h via smooth functions ] , we have for the carnot - carathodory metric @xmath1903 using this expression , we can prove the following theorem .
[ thm metric detection from horizontal laplacian ] let @xmath1543 be a horizontal laplacian , acting on a vector bundle @xmath585 over a closed carnot manifold @xmath0
. then we have for any @xmath186 @xmath1904,f\right ] \right\| \leq 1 \right\},\ ] ] where @xmath39 is the carnot - carathodory metric on @xmath0 .
* proof : * by lemma [ lemma double commutator from horizontal laplacian ] we have @xmath1905,f\right ] \right\| = \sup_{x \in m } \left\| { \mathrm{grad}}^h f(x ) \right\|^2.\ ] ] in addition , we have shown in section 3.3 that @xmath1906 \right\| = \sup_{x \in m } \left\| \mathrm{grad}^h f ( x ) \right\|\ ] ] for any smooth function @xmath494 .
thus provides us @xmath1907 since @xmath1908 if and only if @xmath1909 , the statement of the theorem follows from lemma [ lemma double commutator from horizontal laplacian].@xmath91 theorem [ thm metric detection from horizontal laplacian ] states that it is possible to detect the carnot - carathodory metric by arbitrary positive horizontal laplacians . in the case
@xmath1 is a horizontal dirac operator constructed in chapter 3 , both @xmath1 and @xmath609 detect the carnot - carathodory metric via the corresponding formulas .
now it is a natural question whether one can use the metric detection by a positive horizontal laplacians to find estimates for the connes metric given by its square root operator . for the rest of this section , we take the algebra @xmath96 as the dense sub - algebra of @xmath45 over which the connes metric is defined , i.e. we consider the metric @xmath1910 \right\| \leq 1\right\}\ ] ] for an operator @xmath5 defining a spectral triple
. then we can note the following criterion .
[ prop metric estimate square root of horizontal laplacian from below ] let @xmath1691 be a self - adjoint operator acting on a closed carnot manifold @xmath0 and defining a spectral triple such that @xmath1911 is a horizontal laplacian .
let @xmath1912 denote the connes metric defined by @xmath1691 via .
* proof : * this follows immediately from the formulas for the metrics . by
we see that @xmath1915,f\right ] \right\| = \left\| \left[\tilde{d},f\right]^2 \right\| \leq \left\| \left[\tilde{d},f\right]\right\|^2.\ ] ] but this shows that @xmath1916\| \leq 1 $ ] implies @xmath1917,f]\| \leq 1 $ ] , which gives together with theorem [ thm metric detection from horizontal laplacian ] the desired estimates for the corresponding metrics : @xmath1918\right\| \leq 1 \right\}\\ & \leq & \sup \left\ { \left|f(x)-f(y)\right| : \left\| \frac{1}{2}\left[[\delta^{\mathrm{hor}},f],f\right ] \right\| \leq 1 \right\}\\ & = & d_{cc}(x , y).\end{aligned}\ ] ] @xmath91 now in case @xmath1 is a horizontal dirac operator , condition is fulfilled as we will see in a minute . in this situation
we even have equality of the metrics , since estimate in the proof of the proposition is an equality .
the reason for this is that @xmath62 $ ] is exactly the horizontal clifford action @xmath43 of the horizontal gradient of @xmath63 , and that at each point @xmath1919 @xmath1920 is an isometry .
[ prop condition [ a , b]^2 = 1/2[[a^2,b],b ] ] let @xmath1921 be linear operators on a hilbert space
. then we have @xmath1922 ^ 2 = 1/2[[a^2,b],b ] } \left[a , b\right]^2 = \frac{1}{2 } \left[\left[a^2,b\right]b\right]\ ] ] if and only if @xmath1923,b]+[[a , b],b]a=0 } a\left[[a , b],b\right ] + \left[[a , b],b\right]a = 0\]]*proof : * this is just a simple calculation involving commutator rules : @xmath1924b\right ] & = & \left[a[a , b]+[a , b]a , b\right]\\ & = & \left[a[a , b],b\right ] + \left[[a , b]a , b\right]\\ & = & a \left[[a , b],b\right ] + [ a , b][a , b ] + [ a , b][a , b ] + \left[[a , b],b\right]a.\end{aligned}\ ] ] from this equation , we see the equivalence of the two statements .
@xmath91 from the aspect of spectral triples , we choose in proposition [ prop condition [ a , b]^2 = 1/2[[a^2,b],b ] ] @xmath1925 ( the dirac operator of the triple ) and @xmath1926 ( the representation of the corresponding algebra on the hilbert space , which is multiplication by functions in the commutative case ) .
hence it is obvious that is fulfilled if @xmath5 is a first - order differential operator , since in this case we have @xmath1927 = 0 $ ] .
more generally , is fulfilled if the operator @xmath5 fulfills the order - one condition for a spectral triple by alain connes ( see e.g. @xcite ) , or @xcite , section 10.5 ) , which states in the commutative situation that @xmath1928,g ] = 0 $ ] for all @xmath476 .
therefore we can write down the following corollary .
let @xmath0 be a closed carnot manifold .
if @xmath1929 is a spectral triple such that @xmath1930 is a horizontal laplacian , which fulfills the order one condition , then we have for any points @xmath186 @xmath1931\right\| \leq 1 \right\}.\ ] ] * proof : * this follows immediately by proposition [ prop condition [ a , b]^2 = 1/2[[a^2,b],b ] ] and proposition [ prop metric estimate square root of horizontal laplacian from below ] together with the above discussion .
@xmath91 but despite this corollary , we note that the condition or , equivalently by proposition [ prop condition [ a , b]^2 = 1/2[[a^2,b],b ] ] , @xmath1932,f\right ] + \left[[\tilde{d},f],f\right]\tilde{d } = 0\ ] ] seem to be rather strong conditions for the case where @xmath1691 is not a differential operator .
although the horizontal dirac operator , which detects the carnot - carathodory metric , does not furnish a spectral triple , we have seen in section 7.1 that there are quite a lot of spectral triples arising from heisenberg calculus which give at least the right dimension . in particular the spectral triples constructed in corollary [ cor disturbed spectral triple on heisenberg manifold ] only differ by a small parameter from the horizontal dirac operator .
but sadly we do not know how their connes metric behaves with respect to the carnot - carathodory metric .
now in this final section we want to present a few ideas how one can approximate the carnot - carathodory metric by a family of spectral triples .
therefore we use the observations made in section 1.2 concerning arbitrary compact quantum metric spaces to develop criteria for this approximation .
first of all , we give a reformulation of proposition [ prop uniformly continuity implies convergence of metrices ] and proposition [ prop equivalence condition for metricies ] for our setting .
[ prop conditions for metric convergence to d_cc ] let @xmath0 be a closed carnot manifold , where @xmath39 is the carnot - carathodory metric on @xmath0 .
let further @xmath1 denote a horizontal dirac operator on @xmath0 , acting on a horizontal clifford bundle @xmath533 .
a. let @xmath1691 be an operator on @xmath580 such that @xmath1933 is a spectral triple . if there exists a constant @xmath1934 such that @xmath1935 \right\| \leq c \left\| [ d^h , f ] \right\|\ ] ] for any @xmath63 belonging to a suitable sub - algebra of @xmath45 , then the connes metric @xmath1912 arising from @xmath1691 is equivalent to @xmath39 .
b. let for @xmath1870 the family @xmath1936 be a family of operators on @xmath533 with the property that @xmath1937 \right\| < \varepsilon \ \forall f \in \sigma_0,\ ] ] where @xmath1938 for a suitable sub - algebra @xmath86 of @xmath45 . then for every @xmath116 with @xmath170 there is a @xmath171 such that @xmath1939 for every @xmath173 .
* proof : * we set @xmath1940 \|$ ] , which is a lip - norm on @xmath45 coinciding with the carnot - carathodory - lipschitz constant @xmath712 of @xmath63 by corollary [ cor d^h provides compact quantum metric space ]
. further @xmath1941 \|$ ] in ( i ) and @xmath1942\|$ ] in ( ii ) are also lip - norms on @xmath45 , since they arise from spectral triples .
but in this context ( i ) is just a reformulation of proposition [ prop equivalence condition for metricies ] and ( ii ) is just a reformulation of proposition [ prop uniformly continuity implies convergence of metrices ] . @xmath91
[ ass estimate for general first - order operator ] let @xmath1522 be a first order heisenberg pseudodifferential operator , acting on a heisenberg manifold @xmath0 , and let @xmath494 or ( more generally ) @xmath714 . then there exists a @xmath359 such that @xmath1943 \right\| \leq c \cdot \mathrm{lip}_{cc}(f),\ ] ] where @xmath1944 denotes the carnot - carathodory - lipschitz constant of @xmath63 from definition [ def lip_cc ] .
@xmath74 this assumption is motivated by the fact that something analogous is true for the classical lipschitz functions @xmath97 and for classical pseudodifferential operators of order @xmath98 belonging to the class @xmath1945 . for this classical situation , the statement can be found in @xcite , section 3.6 .
but the proof presented in the book by taylor is quite involved , such that it is far away from being trivial to transfer this proof to the carnot - carathodory situation .
unfortunately , we have not found a way to prove the estimates or , and the proof of the above assumption ( if possible ) seems to be an even harder and much involved problem .
but we assume that the connes metrics arising from the spectral triples constructed in corollary [ cor disturbed spectral triple on heisenberg manifold ] are a good approximation of the carnot - carathodory metric ( note that we have in the formulation of the corollary @xmath1946 with @xmath1 the horizontal dirac operator ) . as an open problem
, we present an approximation of the degenerate spectral triple @xmath1947 for the case where @xmath0 is the compact nilmanifold of a heisenberg group @xmath25 , and @xmath1 is the horizontal pull - back dirac operator arising from a spin structure @xmath1948 of the @xmath391-dimensional torus constructed in chapter 4 .
let @xmath41 be an orthonormal frame of the horizontal distribution @xmath19 of @xmath0 ( with @xmath624 ) , such that the horizontal pull - back dirac operator is given by @xmath1949 see section 4.1 for the explicit construction .
we have seen in section 4.3 that the only space where this horizontal dirac operator degenerates is its infinite dimensional kernel , but otherwise it has all the properties we ask for ( see corollary [ cor d^h gives degenerate spectral triple ] and corollary [ cor hausdorff dimension from d^h ] ) .
the idea is now to fix the gap of degeneracy on the kernel of @xmath1 by using the sum - of - squares operator @xmath1950 it is known that @xmath1951 is hypoelliptic and bounded from below ( see also section 7.1 for this ) .
now let @xmath1522 denote the orthogonal projection onto the kernel of @xmath1 .
since the spectrum of @xmath1 is discrete , this projection operator is a pseudodifferential operator of order @xmath596 , and it is also an order - zero operator in the heisenberg calculus .
then for a small @xmath1868 we define an operator @xmath1876 via @xmath1952 by the above argumentation , this is a hypoelliptic heisenberg pseudodifferential operator , and hence it does define a spectral triple .
in addition it detects the metric dimension of @xmath22 via its eigenvalue asymptotics , since we have the right eigenvalues asymptotics on @xmath1953 by corollary [ cor hausdorff dimension from d^h ] and on @xmath1954 because of the eigenvalue asymptotics of @xmath1951 ( see theorem [ thm asymptotic behaviour of eigenvalues horizontal laplacian ] ) .
now if assumption [ ass estimate for general first - order operator ] is true , we have a constant @xmath822 such that @xmath1955\right\| \leq c \cdot \mathrm{lip}_{cc}(f).\ ] ] note that in this situation it would be enough to have the estimate on the kernel of @xmath1 , which may be much easier to prove than the general case of .
this would lead to the estimate @xmath1956 \right\| \leq \left\| \theta [ p \left(\delta^h\right)^\frac{1}{2 } p , f]\right\| \leq \theta c \cdot \mathrm{lip}_{cc}(f),\ ] ] because of the definition of @xmath1876 , and we could use proposition [ prop conditions for metric convergence to d_cc ] to show that @xmath1957 for @xmath1958 for all points @xmath186 .
cccccc b. ammann , c. br , the dirac operator on nilmanifolds and collapsing circle bundles .
global anal .
( 1998 ) , no . 3 _ , 221 - 253 .
b. andrews , c. hopper : _ the ricci flow in riemannian geometry_. springer , 2011 .
p. baum , e. van erp : _ k - homology and index theory on contact manifolds_. arxiv:1107.1741 , 2011 . c. br : _ das spektrum von dirac - operatoren_. bonner math .
schr . 217 , 1991 .
c. br : _ some properties of solution to weakly hypoelliptic equations_. arxiv:1207.4036 , 2012 .
r. beals , p.c .
greiner : _ calculus on heisenberg manifolds_. princeton university press , 1988 .
r. beals , p.c .
greiner , n.k .
stanton , the heat equation on a cr manifold , _ j. differential geom . 20
( 1984 ) , no .
2_. 343 - 387 .
a. bellaiche , the tangent space in sub - riemannian geometry .
_ sub - riemannian geometry , progress in mathematics 144 , pp .
1 - 78 _ , birkhuser , 1996 .
n. berline , e. getzler , m. vergne : _ heat kernels and dirac operators_. springer , 2004 .
a. bonfiglioli , e. lanconelli , f. uguzzoni : _ stratified lie groups and potential theory for their sub - laplacians_. springer , 2007 .
o. calin , d .-
chang : _ sub - riemannian geometry : general theory and examles_. cambridge university press , 2009
. l. capogna , d. danielli , s.d .
pauls , j.t .
tyson : _ an introduction to the heisenberg group and the sub - riemannian isoperimetric problem_. birkhuser , 2007 .
chernoff , essential self - adjointness of powers of generators of hyperbolic equations .
_ j. funct .
12 ( 1973 ) _ , 401 - 414 .
chow , ber systeme von linearen partiellen differentialgleichungen erster ordnung .
_ mathematische annalen 117 ( 1939 ) _ , 98 - 105 .
e. christensen , c. ivan , spectral triples on af @xmath2-algebras and metrics on the cantor set .
_ j. operator theory 56 ( 2006 ) , no . 1 _ 17 - 46 .
e. christensen , c. ivan , m.l .
lapidus , dirac operators and spectral triples for some fractal sets built on curves .
( 2008 ) , no . 1 _ , 42 - 78 .
e. christensen , c. ivan , e. schrohe , spectral triples and the geometry of fractals .
_ j. noncommut .
( 2012 ) , no . 2 _ , 249 - 274 .
a. connes , noncommutative differential geometry . _
publications mathmatiques de l i.h..s .
( 1985 ) , tome 62 , no .
2 _ , 41 - 144 . a. connes , compact metric spaces , fredholm modules , and hyperfiniteness .
_ ergodic theory dynamical systems 9 ( 1989 ) , no . 2 _ , 207 - 220 .
a.connes : _ noncommutative geometry_. academic press , 1994 .
a. connes : gravity coupled with matter and the foundations of noncommutative geometry .
182 , ( 1996 ) , no . 1 _ , 155 - 176 .
a. connes , h. moscovici , the local index formula in noncommutative geometry .
anal . 5 ( 1995 ) , no . 2 _ , 174 - 243 .
l. corwin , f.p .
greenleaf : _ representations of nilpotent lie groups and their applications , part 1 : basic theory and examples_. cambridge university press , 1989 .
d. danielli , n. garofalo , d.m .
nhieu , sub - riemannian calculus on hypersurfaces in carnot groups .
( 2007 ) , no .
1_. 292378 .
h. federer : _ geometric measure theory _ , springer , 1969 .
folland : _ harmonic analysis in phase space _ , princeton university press , 1989 .
gelfand , m.a .
naimark , on the embedding of normed rings into the ring of operators in hilbert space .
sbornik ( 1943 ) , no . 2 _ 197 - 217 .
gracia - bonda , j.c .
vrilly , h. figueroa : _ elements of noncommutative geometry _ , birkhuser , 2001 .
m. gromov , carnot - carathodory spaces seen from within .
_ sub - riemannian geometry , progress in mathematics 144 , pp .
79 - 323 _ , birkhuser , 1996 . b. helffer , r. nourrigat , characterisation des operateurs hypoelliptiques homogenes invariants a gauche sur un groupe de lie nilpotent gradue . _ com . in partial differential equations , 4 ( 1979 ) , no . 8 _ , 899 - 958 .
l. hrmander , pseudo - differential operators and hypoelliptic equations .
_ singular integrals ( proc
pure math .
x , chicago ill .
1966 ) , pp .
138 - 183 _ , american mathematical society , providence , r.i . ,
1967 . l. hrmander , hypoelliptic second order differential equations . _ acta math . 119
( 1967 ) _ , 147 - 171 .
b. iochum , t. krajewski , p. martinetti , distances in finite spaces from noncommutative geometry .
_ j. geom .
37 ( 2001 ) , no . 1 - 2 _ , 100 - 125 .
kirillov , unitary representations of nilpotent lie groups .
survey 17 ( 1962 ) _ , 53 - 104 .
kirillov : _ lectures on the orbit method_. american mathematical society , 2004 .
knapp : _ lie groups beyond an introduction _ , 2nd edition .
, birkhuser , 2005 .
kohn : complex hypoelliptic operators .
mathematica , vol .
vii ( convegno sulle equazioni ipoellittiche e spazi funzionali , indam , rome , 1970 ) , pp .
459 - 468 _ , acamedic press , london , 1971 .
kohn , pseudo - differential operators and hypoellipticity .
_ partial differential equations ( proc .
pure math .
xxiii , univ .
california , berkeley , calif .
1971 ) , pp .
61 - 69 _ , american mathematical society , providence , r.i .
g. landi , _ an introduction to noncommutative spaces and their geometries _ , springer , 1997 .
lawson , m .-
michelsohn : _ spin geometry _ , princeton university press , 1989 .
j. mitchell , on carnot - carathodory metrics .
_ j. differential geom . 21
( 1985 ) , no . 1 _ , 35 - 45 .
r. montgomery : _ a tour of sub - riemannian geometries , their geodesics and applications _ , american mathematical society , 2002 .
palmer , _ riemannian geometry of compact metric spaces _ , phd - thesis , georgia institute of technology , 2010 .
p. pansu , mtriques de carnot - carathodory et quasiisomtries des espaces symtriques de rang un .
_ ann . of math .
( 2 ) 129 ( 1989 ) , no .
1 _ , 1 - 60 .
j. pearson , j. bellissard , noncommutative riemannian geometry and diffusion on ultrametric cantor sets .
_ j. noncommut .
( 2009 ) , no . 3 _ , 447 - 480 .
r. ponge : _ heisenberg calculus and spectral theory of hypoelliptic operators on heisenberg manifolds _ , american mathematical society , 2008 .
i. prokhonrenkov , k. richardson , natural equivariant transversally elliptic dirac operators .
dedicata 151 ( 2011 ) _ , 411 - 429 .
rieffel , metrics on states from actions of compact groups .
math . 3 ( 1998 )
_ , 215 - 229 .
rieffel , metrics on state spaces .
math . 4 ( 1999 )
_ , 559 - 600 .
rieffel : _ gromov - hausdorff distance for compact quantum metric spaces_. american mathematical society , 2004 .
c. rockland , hypoellipticity on the heisenberg group - representation - theoretic criteria . _ trans .
( 1978 ) _ , 1 - 52 .
j. roe : _ elliptic operators , topology and asymtotic methods _ ,
2nd edition .
addison wesley longman limited , 1998 .
rothschild , a remark of hypoellipticity of homogeneous invariant differential operators on nilpotent lie groups .
_ comm . in part .
( 1983 ) , no
. 15 _ , 1679 - 1682 .
rothschild , e.m .
stein , hypoelliptic differential operators and nilpotent groups .
_ acta math . 137
( 1977 ) , no . 3 - 4 _ , 247 - 320 .
shubin : _ pseudodifferential operators and spectral theory _ ,
2nd edition .
springer , 2001 .
stein : _ harmonic analysis_. princeton university press , 1993 .
taylor : _ noncommutative microlocal analysis i_. american mathematical society , 1984 .
taylor : _ pseudodifferential operators and nonlinear pde_. birkhuser , 1991 . | we analyze whether one can construct a spectral triple for a carnot manifold @xmath0 , which detects its carnot - carathodory metric and its graded dimension .
therefore we construct self - adjoint horizontal dirac operators @xmath1 and show that each horizontal dirac operator detects the metric via connes formula , but we also find that in no case these operators are hypoelliptic , which means they fail to have a compact resolvent .
first we consider an example on compact carnot nilmanifolds in detail , where we present a construction for a horizontal dirac operator arising via pullback from the dirac operator on the torus . following an approach by christian br to decompose the horizontal clifford bundle
, we detect that this operator has an infinite dimensional kernel .
but in spite of this , in the case of heisenberg nilmanifolds we will be able to discover the graded dimension from the asymptotic behavior of the eigenvalues of this horizontal dirac operator .
afterwards we turn to the general case , showing that any horizontal dirac operator fails to be hypoelliptic . doing this
, we develop a criterion from which hypoellipticity of certain graded differential operators can be excluded by considering the situation on a heisenberg manifold , for which a complete characterization of hypoellipticity in known by the rockland condition .
finally , we show how spectral triples can be constructed from horizontal laplacians via the heisenberg pseudodifferential calculus developed by richard beals and peter greiner .
we suggest a few of these constructions , and discuss under which assumptions it may be possible to get an equivalent metric to the carnot - carathodory metric from these operators . in addition , we mention a formula by which the carnot - carathodory metric can be detected from arbitrary horizontal laplacians .
* keywords : * spectral triple , carnot - carathodory metric , hypoellipticity . |
the mechanism of quark confinement within hadrons is still an open problem .
numerical simulations of qcd on space - time lattices demonstrate the confinement phenomenon but could not yet clarify completely the mechanism for the formation of a gluonic flux tube .
effective theories have been created to improve our understanding of this mechanism .
an intuitive physical picture is the dual superconductor model @xcite : dynamically generated colour magnetic monopole currents form a solenoid which squeezes the chromoelectric flux between quarks into a narrow flux tube .
the energy of this flux tube increases linearly with its length and confinement is achieved .
another approach is the effective string theory @xcite which should describe flux tubes at large quark distances .
it is an interesting goal of lattice gauge theories to test the validity of such effective models and therefore deepen our understanding of the confinement mechanism .
@xmath0 lattice gauge theory in four dimensions undergoes a phase transition at strong coupling and has been widely used as a prototype of a confining theory . in @xmath0 theory the condensation of magnetic monopoles is made responsible for the confinement of electric charges .
the results of lattice simulations have been interpreted in terms of the dual superconductor picture of confinement , e.g. the validity of a dual london relation has been checked @xcite , and the @xmath0 results for electric fields and monopole currents were found to agree with the predictions of a classical effective model considering also string fluctuations @xcite .
most of the obtained results are expected to hold also in non - abelian gauge theories , at least qualitatively @xcite .
the data accumulated so far , however , for larger flux tube lengths are not of sufficient accuracy to perform a quantitative analysis in terms of effective models for the formation of a gluonic flux tube .
furthermore , it was realized many years ago @xcite that one can perform a duality transformation of the path integral in compact abelian gauge theories . in this way a new partition function is obtained which can be regarded as a limit of the dual non - compact abelian higgs model @xcite .
we will use the duality relation for expectation values in the presence of external charges to clarify the connection between dually transformed @xmath0 theory and the dual superconductor model . besides this support for the interpretation of the results , the dual theory provides a very efficient tool for simulating flux tubes .
this fact has already been realized in refs .
@xcite , but to our knowledge no dual simulations of four - dimensional @xmath0 lattice gauge theory with external sources have been performed yet .
the organisation of this article is as follows : in section 2 we review the duality transformation of @xmath0 lattice gauge theory for the wilson action with special consideration of fields and their squares in the presence of external charges .
the consequences for the interpretation in the dual superconductor picture are discussed by connecting the results to a dual higgs model in section 3 . after pointing out the advantages of dual simulations of flux tubes in section 4
, we present our results for electric field strength and magnetic current in the presence of a charge anticharge
pair in section 5 , analyzing the flux tube as a function of spatial lattice size and temperature .
in section 6 we return to the dual superconductor picture , subjecting our data to a quantitative comparison with the london model . in section 7
the dependence of the flux tube width on its length is investigated and compared with the prediction of an effective string theory . finally in section 8
we calculate the total electromagnetic energy and analyze the various components contributing to it .
we begin with the usual prescription for simulating @xmath0 lattice gauge theory .
expectation values of physical observables @xmath1 in the presence of a static charge pair at distance @xmath2 are determined by the correlation function @xmath3 where @xmath4 is the polyakov loop @xmath5 and the angle brackets denote the evaluation of the path integral using the standard wilson action @xmath6.\ ] ] @xmath7 is the inverse coupling , and @xmath8 is the discretized exterior derivative of the phases @xmath9 of the link variables @xmath10 and is assigned to a plaquette of size @xmath11 . below we will use the notation of lattice differential forms and therefore suppress indices . the observables @xmath12 we are interested in are the electric field strength , the magnetic current and the electric and magnetic energy densities . for the field strength @xmath13
we use the identification @xmath14 which was shown to fulfil the electric gauss law for the wilson action @xcite .
the magnetic current can be constructed via the dual maxwell equations .
the squared fields are calculated in the usual way by means of the operator @xmath15 .
let us briefly review the most important steps of the duality transformation .
we start from the @xmath0 partition function using the wilson action which reads @xcite @xmath16 where the product is over all links on the lattice , and the sum in the exponent is over the plaquettes .
expanding the boltzmann factor for each plaquette in a fourier series one can write @xmath17 where we have used the plaquette variables @xmath18 and the property of the inner product @xmath19 , and then performed the integration over the links @xmath9 .
@xmath20 is the number of lattice sites , and the product has to be performed over all plaquettes .
@xmath18 is an integer valued 2-form corresponding to the field strength @xmath13 .
the constraint @xmath21 describes the absence of electric charges , @xmath22 .
we solve this constraint by introducing an integer valued 3-form @xmath23 with @xmath24 @xmath25 where we have finally switched to the dual link variables @xmath26 , representing an integer valued dual potential .
the product in the partition function extends over the links of the dual lattice .
the next step is the inclusion of external charges into the @xmath0 path integral . @xmath27 and @xmath28 denote the world - lines of a static charge pair .
starting from @xmath29 one arrives at an expression equivalent to ( [ ztp ] ) , with a modified constraint @xmath30 for links belonging to @xmath31 ( equivalent to @xmath32 ) .
this constraint is solved by an additional integer 2-form @xmath33 : @xmath34 external electric currents are represented by the boundaries of a dual dirac sheet in the dual theory .
the dually transformed path integral therefore reads @xmath35 where the product is performed over all links of the dual lattice .
the dual dirac sheet @xmath36 can be fixed as the minimal straight area between the two polyakov loops , while the integer gauge field @xmath26 is kept as dynamical variable and updated in simulations .
fluctuations in @xmath26 result in fluctuations of the flux string @xmath37 .
further , we will perform the duality transformation for the correlation function ( [ correlation ] ) . for the determination of the flux @xmath38 through a given plaquette @xmath39
, we have to insert the operator @xmath40 into the path integral in addition to the polyakov loops .
so we can write @xmath41 taking the imaginary part reflects the change of sign of the field strength under an exchange of charge and anticharge , i. e. under complex conjugation . our constraint for @xmath42
is further modified : @xmath42 has to be non - zero along the electric current and the boundary of the considered plaquette @xmath43 . solving this constraint by a redefinition of @xmath18 on this plaquette ( @xmath44 ) and using the identity for modified bessel functions @xmath45,\ ] ] one obtains @xmath46 where @xmath47 as before .
note that in this dual expression the validity of gauss law for electric charges is checked easily because of @xmath48 .
performing the duality transformation for the expectation value of @xmath49 in an analogous way leads to @xmath50 we would like to mention that an identity equivalent to ( [ cos ] ) was already derived in ref .
@xcite for calculating the electromagnetic energy density in three - dimensional @xmath0 lattice gauge theory , using the polymer formulation of the path integral .
the results of @xmath0 simulations in the confinement phase verify the dual superconductor picture @xcite : the electric flux between a charge pair is squeezed into a flux tube , encircled by monopole currents acting like a coil .
we want to discuss how this scenario can be described by the dual non - compact abelian higgs model , in order to establish an exact connection to the dually transformed lattice @xmath0 theory presented above .
the dual higgs model is written in terms of dual potentials @xmath51 assigned to links on the dual lattice ( the @xmath52-notation is again used for all fields which are located on the dual lattice ) and of a dual higgs field @xmath53 which is located on the dual sites .
electric charges in this dual model correspond to monopoles ( assigned to three - dimensional cubes on the dual lattice ) and can be introduced by the dual of a dirac string in the same way as in the last section . on the original lattice the world lines of electric charges
define the boundary of a continous surface ( the dirac sheet ) : @xmath54 the field strength @xmath55 is given by @xmath56 , where @xmath57 is the exterior derivative of the 1-form @xmath51 , and @xmath58 is the magnetic coupling . in the london limit the higgs field @xmath53 is constrained to @xmath59 and therefore the coherence length vanishes . including electric sources the corresponding action reads @xmath60 where @xmath61 .
the function @xmath62 determines the non - compact gauge field action and approaches @xmath63 in the continuum limit .
the second part of the action is the compact lattice version of the gauge invariant kinetic term of the higgs field @xmath53 .
the ratio @xmath64 has the meaning of the bare squared london penetration length @xmath65 .
the higgs current @xmath66 , given by @xmath67 - g\ , { } ^*a \right),\ ] ] is constructed in such a way that the fluxoid quantisation @xmath68 @xcite is valid for every single field configuration .
the integer two - form @xmath69 characterizes the physical fluxoid string which is allowed to fluctuate in simulations due to the compactness of the phase @xmath70 of the higgs field . @xmath71
$ ] in eq .
( [ jh ] ) indicates the reduction of @xmath72 to the interval @xmath73 $ ]
. therefore @xmath74 $ ] may be non - zero for topologically non - trivial field configurations , and @xmath75 $ ] .
if the value of @xmath76 $ ] on one dual link flips , i. e. the phase difference between neighbouring points exceeds the range @xmath73 $ ] , this effects six dual plaquettes corresponding to a closed cube on the original lattice .
if such a cube has one plaquette in common with the dirac sheet ( @xmath77 ) , @xmath78 and @xmath74 $ ] may cancel each other , changing the shape of the physical fluxoid string .
this higgs model can be regarded as four - dimensional generalisation of the static classical effective model investigated in ref .
it can also serve to extend numerically the predictions from the dual qcd ansatz @xcite or from similar analytical calculations in effective models @xcite , especially in investigating fluctuations of the fluxoid string .
for the @xmath0 gauge theory , it is even possible to establish an exact connection at the level of the path integral .
let us assume @xmath79 which imposes a new constraint on the fields : choosing a gauge where the phase @xmath70 of the higgs field in eq .
( [ higgs ] ) is zero , we realize that the integral over @xmath51 for each dual link reduces to a sum over an integer @xmath26 , according to the constraint @xmath80 .
up to a @xmath81-dependent factor the @xmath79 limit of the partition function of this higgs model equals the dually transformed path integral ( [ wje ] ) of @xmath0 gauge theory , including two polyakov loops .
this duality property is well - known for the villain form of the action @xcite . in the case of the wilson action
the function @xmath62 describing the gauge field action in the corresponding higgs model is determined by the modified boltzmann factor in eq .
( [ ztp ] ) @xmath82 = e^{-\beta } i_{\parallel g\,{}^*f \parallel}(\beta),\ ] ] and @xmath83 .
especially interesting for us is the interpretation of expectation values in @xmath0 in terms of the dual higgs model .
let us first regard the field strength in @xmath0 derived in eq .
( [ sin ] ) : it also agrees with the field stength of the dual higgs model in the @xmath79 limit which reads @xmath84 this means that the field strength expectation values and due to the validity of the extended maxwell equations also the magnetic currents in @xmath0 gauge theory agree with the corresponding fields and currents in the dual higgs model in the limit of both zero coherence length and zero penetration length . from this point of view
a @xmath0 flux tube looks rather like an infinitely thin fluctuating string than like a flux tube in a classical dual superconductor .
this correspondence to the higgs model does not hold for energy densities .
the dually transformed expectation value of the squared field strength , eq .
( [ cos ] ) , obviously disagrees with the corresponding result in the @xmath79 limit of the dual higgs model .
this is not astonishing , since the expectation values of the squared fields are connected to the vacuum fluctuations which show the opposite behaviour in the dual theory , where the confinement phase is the weakly coupled phase and fluctuations become more important with increasing @xmath81 .
hence , a direct interpretation of the @xmath0 energy density in the dual higgs model is not possible , which makes it difficult to distinguish by analytical considerations between dual type - i and type - ii superconductivity .
as has been pointed out above , one has to be careful when comparing @xmath0 lattice gauge theory to a dual higgs model on a `` microscopic '' level .
nevertheless , simulating the dual @xmath0 gauge theory gives us the possibility to calculate any expectation value in the presence of static charges with significantly higher precision .
this has several reasons : the confinement phase is the weakly coupled one in the dual theory , therefore we have less quantum fluctuations in the region of interest . simulating the dually transformed path integral ( [ sin ] ) resp .
( [ cos ] ) is also much faster because of the integer valued gauge fields . most important ,
however , is the fact that it is not necessary to project the charge anticharge state out of the vacuum : charge pairs with arbitrary distance can be simulated with equal accuracy , as well as multiply charged systems @xcite .
the updating procedure itself is very simple : as already discussed in the previous section , the dual dirac string sheet @xmath36 connecting the two polyakov lines is held fixed , while the integer gauge field @xmath26 , located on the dual links , is changed by @xmath85 during the updates .
we have to mention that this is essentially the same procedure as the `` change - a - cube '' algorithm used in ref .
the only difference for the three - dimensional model used there is that the dual potential is located on the dual sites instead of the dual links . in ref .
@xcite the electromagnetic energy distribution was calculated to extract the string tension in a range of weaker couplings as total field energy per length . however
, this quantity must not be confused with the string tension calculated from usual correlation functions .
recently , a detailed analysis of simulating dual @xmath0 lattice gauge theory in three dimensions was performed in ref .
the authors found that at weak coupling the advantages are outweighed by large autocorrelation times as the dual model becomes disordered .
we also observed the increase of the autocorrelation time and correspondingly of the error for increasing values of @xmath81 with a peak at the phase transition of the four - dimensional theory . in the deconfinement phase
there are large fluctuations , but the autocorrelation time is smaller than around the phase transition . since our goal is the investigation of the confinement mechanism , we take the main advantage from simulations in the confinement phase , anyhow . in this context
we want to mention that we used a duality transformation exactly valid for the whole @xmath81 range , not only in the weak coupling limit as in ref .
further , as stated above , the power of dual simulations lies in the inclusion of the sources in the action , which already has been realized in ref .
@xcite . for each charge distance
@xmath2 a separate simulation has to be performed , and of course for the subtraction of the vacuum expectation value in ( [ correlation ] ) a distinct run is needed .
therefore one gets a good estimate of the reliability of the errorbars . for smaller distances on a @xmath86 lattice
an accurate comparison was performed between expectation values from ordinary and from dual simulations @xcite .
the only difference in the observed results is the effect of the periodic boundary conditions which is absent in the dual model : there can be no electric flux over the period per construction . for @xmath87
the dual simulation is clearly more efficient , because in ordinary simulations the signal decreases exponentially with the charge distance @xmath2 , while in the dual simulation it stays constant ( the necessary amount of computer time still depends of course on the lattice size used ) . for the updating procedure we use a standard metropolis algorithm , which has proven sufficient for the investigated range of couplings , although a nonlocal cluster algorithm would certainly lead to a further increase in performance .
our typical measurements are taken from @xmath88 configurations , where each 10th configuration is evaluated .
for determining the errorbars we took blocks of 100 evaluated configurations .
for the results shown in the next sections we used the coupling @xmath89 , where the system is clearly in the confinement phase . on the other hand the fluctuations in the dual theory are already large enough to produce a flux tube of the desired transverse extent .
if one likes to compare with the qcd flux tube , we can require that our string tension equals the physical value , which gives a lattice spacing @xmath90 of roughly @xmath91 .
we present now our numerical results for the electric field and magnetic current distribution in the presence of a static charge pair . the great advantage compared to the results reported in ref .
@xcite is that we are now able to investigate much larger charge distances and much lower temperatures . before discussing the temperature dependence of the flux tube profile ,
let us comment on the necessary spatial size of the lattice .
there is the simple condition that the spatial extent of the lattice has to be greater than the extent of the considered flux tube both in longitudinal and in transverse direction in order to prevent strong finite size effects . for our simulations we used a spatial extent of @xmath92 for charge distances @xmath93 , @xmath94 for simulating @xmath95 , and @xmath96 for @xmath97 .
further we performed some simulations on larger lattices ; the largest investigated charge distance was @xmath98 ( on a @xmath99 lattice ) .
results for this distance are displayed in a 3d - plot in fig .
[ plot3d ] .
we focus on the longitudinal component of the electric field strength and on the azimuthal component of the magnetic current which encircles the flux tube like a coil .
it can already be seen in these plots that a flux tube of constant thickness is not observed even for this large charge distance .
the increase of the flux tube width with the distance between charges may be explained by roughening effects and will be discussed in section [ wsq ] .
the right figure demonstrates that the signal for the magnetic current drops in the middle between charges , its radial dependence in the symmetry plane exhibits a maximum at finite @xmath100 .
this behaviour is interesting for the dual superconductor interpretation as will be discussed in the next section . in all our simulations
we represent static charges by polyakov loops and therefore have to deal with finite temperature effects due to the finite time extent of the lattice .
it is therefore important to separate the occuring thermal fluctuations from quantum fluctuations which are essential for the extended flux tube .
this is achieved by analyzing the dependence of the results on the time extent nt , as is shown in fig .
[ e_nt ] for the electric field profile and in fig .
[ j_nt ] for the magnetic current profile in the symmetry plane between charges for various charge distances @xmath2 . from the results
we conclude that if the time extent nt of the lattice exceeds the charge distance @xmath2 finite temperature effects can safely be neglected . for smaller
nt , however , we realize a significant widening of the flux tube , connected with a decreasing magnetic current . from the point of view of the dual simulation , this widening can easily be explained .
quantum fluctuations of the world sheet of the string correspond to cube - like excitations on the original lattice . at finite temperature
there also appear time - independent excitations which wrap around the lattice in time direction and therefore increase the effective flux tube width .
these excitations vanish if nt is large compared to the length of the flux tube .
after the discussion of the techniques used for the simulation of long flux tubes and the dependence of the results on the temperature , we will now turn again to the dual superconductor interpretation .
in the presence of a pair of static charges , the classical solution of the higgs model ( [ higgs ] ) yields an electric field and a magnetic current distribution according to @xmath101where the ( straight ) string @xmath102 connects the point charges , carrying a quantum of electric flux @xmath103 . integrating eq .
( [ fluxoid ] ) , a generalisation of the london equation , over a surface intersecting the string gives the fluxoid quantisation .
for an infinitely long string eq .
( [ ampere ] ) and eq .
( [ fluxoid ] ) yield an electric field profile of the form @xmath104 .
these classical equations were used in ref .
@xcite for comparison with the results of @xmath0 lattice simulations .
the free parameter @xmath105 ( the london penetration length ) could be fitted by comparing @xmath106 and @xmath107 off axis , and also the fluxoid quantisation was verified .
an extension of this analysis was performed in ref .
@xcite by including fluctuations of the string @xmath102 in the above model .
further the effective model was solved numerically on a lattice of same size , to avoid being misled by lattice artifacts . due to the exponential falloff of the signal , however , the @xmath0 data analyzed so far contain only charge distances up to @xmath108 .
we will now consider the data obtained in the dual simulations , in order to see if the above interpretation holds for longer flux tubes . in the classical model with a straight string @xmath102 one would expect that * the flux tube width as a function of @xmath2 soon approaches a constant , this means that the fitted parameter @xmath105 should not depend on @xmath2 , * @xmath65 is constant everywhere off axis , * the azimuthal magnetic current encircling the axis should exhibit a maximum in the symmetry plane between the charges , and not near the charges . the radial dependence should exhibit an exponential falloff .
it was shown that these requirements are not even fulfilled for a charge distance @xmath108 @xcite , nevertheless the flux tube profile in the symmetry plane may be described well by the effective model with a static string @xcite . considering larger charge distances
confirms the discrepancy to the items mentioned above . in a comparison of electric field profiles for @xmath108 and @xmath109 ( see fig .
[ e_mod ] ) it can be seen that the fitted values for @xmath105 not only differ very much , but it is already impossible to describe the field distribution for larger charge distances by the numerical solution of coupled maxwell and london equation . for the magnetic current distribution we already see in fig .
[ plot3d ] and in fig .
[ j_nt ] that its behaviour can not be explained in the above effective model .
[ jm_mod ] shows the corresponding classical solutions , again for @xmath108 and @xmath109 . for larger distances ,
there is not even qualitative agreement : the azimuthal component of the magnetic current as a function of the transverse distance first increases and takes its maximum value at @xmath110 .
this phenomenon together with the other described observations may be explained within the dual superconductor picture by two different approaches : first , it can be regarded as evidence for the importance of string fluctuations .
this way was chosen in ref .
@xcite , where the intrinsic thickness of the flux tube @xmath105 turned out to be much smaller than the effective flux tube width which increases with charge distance .
the extreme scenario ( @xmath111 ) leads us again to the exact correspondence of @xmath0 to the @xmath79 limit of a dual higgs model .
the second possibility for arguing is to stay in the classical model of a dual superconductor , but leave the london ( extreme type ii ) limit described so far .
a strong suppression of the higgs condensate in the region of electric fields may also lead to the behaviour of the current observed in fig .
[ jm_mod ] .
a similar current distribution measured in abelian projection of @xmath112 was interpreted in this way in ref .
hence , one may conclude that @xmath0 lattice gauge theory in this respect behaves like a dual type - i superconductor .
stronger evidence for such a statement is expected of course from an investigation of the interaction between flux tubes ( which also can be performed very efficiently in a dual simulation ) .
indeed preliminary results @xcite show an attraction between @xmath0 flux tubes for the regarded value of @xmath81 .
we may conclude that interpreting the flux tube results in terms of the dual superconductor picture still makes sense , but @xmath0 lattice gauge theory certainly can not be regarded as a classical dual superconductor in the london limit if one examines larger charge distances . on the other hand
the effective string description of flux tubes represents an independent feature of the lattice gauge system , and from the point of view of dual simulations this looks somehow easier to argue .
we will use this approach in the next section for a closer look at the flux tube width in dependence of the charge distance .
in ref . @xcite the string model approximation to wilson loop correlation functions was used to study the behaviour of the flux tube width for large charge distances .
this model assumes a very thin bare flux tube subjected to quantum mechanical fluctuations . a convenient measure for the transverse extent of the flux tube is the squared width @xmath113 where @xmath114 represents a quantity characteristic of the flux tube , for example the ( chromo-)electric field energy density .
it was found in ref .
@xcite that the squared width @xmath115 diverges logarithmically for @xmath116 , @xmath117 with some constants @xmath118 and @xmath119 . a numerical test of this hypothesis could not be performed up to now for @xmath0 lattice gauge theory . in the three - dimensional @xmath120 gauge model
this prediction was verified in ref .
@xcite , where the authors also exploited the duality relation to a three - dimensional ising model and combined their data with high precision data on the interface physics of the three - dimensional ising model .
they also achieved agreement with the prediction @xmath121 ( where @xmath122 is the string tension ) for the two - dimensional free gaussian model .
there is the conjecture that this behaviour of the flux tube width should be observed in all gauge theories .
it is therefore interesting to test this prediction in dual @xmath0 simulations where large flux tubes become accessible .
we determine the flux tube width by means of the parallel component of the electric field strength @xmath123 .
our defintion on the lattice reads : @xmath124 this slightly reduces the estimated error compared to the definition via the electric field energy , in all other aspects the results roughly agree .
the sum extends over a transverse plane of the lattice with @xmath125 : for @xmath126 we evaluate the sum until a transverse distance @xmath127 , for larger distances we use @xmath128 . in fig .
[ w_nt ] one can see again the influence of finite temperature effects on the flux tube width .
the calculation with time extent 16 can be taken as acceptable approximation to the zero temperature case for the regarded charge distances and is analyzed according to the prediction ( [ wlog ] ) in fig .
[ cas ] in analogy to ref .
the logarithmic increase of the flux tube width is realized very well for @xmath0 lattice gauge theory .
the constant @xmath129 , however , disagrees with expr .
( [ w0sq ] ) by roughly a factor 2 .
the above data demonstrate that the thickness of @xmath0 flux tubes does not saturate for large charge distances , as expected in a naive formulation of the dual superconductor picture .
it seems to diverge logarithmically , as predicted by the effective string description . as can be seen from fig .
[ cas ] this behaviour sets in already at @xmath130 , which corresponds to a physical length of around @xmath131 in agreement with ref . @xcite .
in section 2 we have discussed the duality transformation for the expectation value of @xmath49 in the presence of two polyakov loops
( [ cos ] ) which gives the squared field strength on a specified plaquette @xmath43 .
this allows for the determination of the euclidean action density @xmath132 and energy density @xmath133 .
as has been pointed out , the squared field strength calculated in the dual simulation differs from the squared field strength of the @xmath79 higgs model .
the contributions of all plaquettes can be summed up and give the total action respective energy of the system . in ordinary lattice simulations this is a very difficult task , since the difference between the squared field strength of the pure @xmath0 vacuum and the charge
anticharge state is small compared to the vacuum expectation value itself .
therefore it can not be resolved for larger charge distances .
further , concerning the energy density there are large cancellations between electric and magnetic field contributions . in the dual simulation of @xmath0 lattice gauge theory
it is possible to determine the total energy of the electromagnetic fields as a function of the charge distance .
the influence of finite temperature effects is analyzed in fig .
[ en_16_4 ] .
the thermal fluctuations increase the total energy .
this is again due to the periodic excitations in time direction which contribute to the electric part of the energy only .
the total field energy shown in fig .
[ en_16_4 ] qualitatively behaves like the free energy for a charge pair ( as determined in usual lattice simulations ) , i.e. up to a constant like a coulomb and a linear term .
however , as already realized in ref .
@xcite , a quantitative comparison shows that the integral over the energy density differs from the potential .
we mention here that it is also possible to determine the free energy via dual simulations ; this opens the opportunity for the application of lattice sum rules @xcite .
such calculations are in progress and will be reported elsewhere .
it is also interesting to look at the individual contributions to the total field energy . for @xmath134
there are no string fluctuations , and therefore only the parallel component of the electric field contributes . if we consider single cube - like excitations in transverse directions , we get also contributions from transverse electric and transverse magnetic fields . in this case
they cancel each other exactly and therefore do not change the total energy . for stronger fluctuations , i. e. increasing roughening of the string , these components differ , and also the parallel magnetic component gets nonzero .
the results for @xmath89 are plotted in fig .
one can argue that for charge distances @xmath135 the string fluctuations are not fully developped .
the `` coulomb '' behaviour of the total field energy observed in fig .
[ en_16_4 ] is mainly due to the parallel component of the magnetic field which gives the smallest contribution to the linear increase of the energy .
this demonstrates that the behaviour of the flux tube can not be explained easily within a classical static model .
our motivation was a detailed quantitative analysis of the dual superconductor picture in @xmath0 lattice gauge theory . by considering the duality transformation of correlation functions we improved the understanding of the behaviour of expectation values of fields and currents in the presence of a charge pair , as well as the significance of string fluctuations .
moreover , simulating the dually transformed theory provided numerical flux tube data for low temperatures and large charge distances not obtained before in four - dimensional @xmath0 lattice gauge theory .
we found that finite temperature effects may be neglected for time extents exceeding the charge distance , while for smaller time extents the flux tube widens significantly .
further we examined the electric field and magnetic current distribution and found that its description by the classical solution of maxwell and london equation fails for large charge distances .
the observed effects may be explained by fluctuations of the flux tube , on the other hand the current profile resembles that of a dual type - i superconductor .
roughening effects which become important for large distances are described well by the effective string picture of confinement which predicts a logarithmic increase of the flux tube width .
finally , the analysis of the total field energy revealed the responsibility of the parallel magnetic field for the coulomb behaviour at small distances . at this time , no final statement should be made which effective model is really able to describe strongly coupled @xmath0 gauge theory .
certain aspects of the flux tube may be explained by a `` classical '' dual superconductor picture , while others can not .
further studies considering periodically closed flux tubes ( torelons ) are in progress and promise to answer some of the remaining questions .
we want to thank richard w. haymaker for interesting discussions on the description of confinement in @xmath0 lattice gauge theory by effective models .
further we thank jeff greensite for making us aware of an article ( formerly not known to us ) , where the energy density in three - dimensional @xmath0 was calculated in an analogous way .
finally we acknowledge critical comments from andrei ivanov and tefan olejnk .
99 g. t hooft , in _ high energy physics _ , eps international conference , palermo 1975 , ed .
a. zichichi , s. mandelstam , phys . rep .
23c ( 1976 ) 245 . y. nambu , phys .
d10 ( 1974 ) 4262 .
m. lscher , k. symanzik and p. weisz , nucl .
b173 ( 1980 ) 365 . v. singh , r. w. haymaker , d. a. browne , phys . rev .
d47 ( 1993 ) 1715 . m. zach , m. faber , w. kainz , p. skala , phys .
b358 ( 1995 ) 325 , hep - lat 9508017 .
p. skala , m. faber , m. zach , nucl .
b , in press , npb preprint 4397 . t. banks , r. myerson , j. kogut , nucl .
b129 ( 1977 ) 493 .
j. frhlich , t. spencer , commun .
phys . 83 ( 1982 ) 411 , + j. frhlich , p. a. marchetti , europhys . lett . 2 ( 1986 ) 933 . t. sterling , j. greensite , nucl . phys . b220 ( 1983 ) 327 .
m. caselle , f. giozzi , u. magnea , s. vinti , nucl .
b460 ( 1996 ) 397 . for
a review see e.g. m. baker , j. s. ball , f. zachariasen , phys .
d41 ( 1990 ) 2612 .
m. faber , a. n. ivanov , w. kainz , n. i. troitskaya , nucl .
b475 ( 1996 ) 73 , + m. faber , a. n. ivanov , w. kainz , n. i. troitskaya , phys .
b386 ( 1996 ) 198 . for the villain form of the action the same procedure
is reviewed in : r. w. haymaker , lecture notes of the int .
school of physics , `` enrico fermi '' , varenna , italy , 1995 ; hep - lat 9510035 .
m. zach , m. faber , p. skala , nucl .
b ( proc . suppl . ) 53 ( 1997 ) 521 , hep - lat 9608009 .
m. zach , m. faber , p. skala , to be published .
k. coyle , i. g. halliday , p. suranyi , preprint swat/102 , hep - lat 9610031 .
k. bernstein , g. di cecio , r. w. haymaker , nucl .
phys.b ( proc
. suppl . ) 53 ( 1997 ) 481 . c. michael , nucl .
b280[fs18 ] ( 1987 ) 13 , + h. g. dosch , o. nachtmann , m. rueter , heidelberg preprint hd - thep-95 - 12 , + c. michael , phys .
d53 ( 1996 ) 4102 . | the dually transformed path integral of four - dimensional @xmath0 lattice gauge theory is used for the calculation of expectation values in the presence of external charges . applying the dual simulation to flux tubes for charge distances up to around 20 lattice spacings
, we find a deviation from the behaviour of a dual type - ii superconductor in the london limit at large distances .
the roughening of the flux tube agrees with the effective string picture of confinement .
further we show that finite temperature effects are negligible for time extents larger than the charge distance .
finally we analyze the different contributions to the total energy of the electromagnetic field .
the parallel component of the magnetic field turns out to have the strongest influence on the deviation from the linear behaviour at small charge distances . |
we will denote the set of density operators on @xmath5 by @xmath6 .
we will denote the set of quantum channels @xmath7 by @xmath8 .
we will put @xmath9 . for any linear map @xmath10
, we define it _
choi - jamiokowski matrix _ as @xmath11 this isomorphism was first studied by choi @xcite and jamiokowski @xcite .
note that some authors prefer to add a normalization factor of @xmath12 if front of the expression for @xmath13 .
other authors use the other order for the tensor product factors , a choice resulting in an awkward order for the space in which @xmath13 lives .
the rank of the matrix @xmath13 is called the _ choi rank _ of @xmath14 ; it is the minimum number @xmath15 such that the map @xmath14 can be written as @xmath16 for some operators @xmath17
. given two pairs of unitary operators , @xmath18 , respectively @xmath19 , define the `` rotated map '' @xmath20 it is an easy exercise to check that the choi - jamiokowski matrix of the rotated map is given by @xmath21 the transposition appearing in the equation above is due to the following key fact ( here @xmath22 is an arbitrary unitary operator ) : @xmath23 the _ diamond norm _ was introduced in quantum information theory by kitaev ( * ? ? ?
* section 3.3 ) as a counterpart to the @xmath24-norm in the task of distinguishing quantum channels .
first , define the @xmath25 norm of a linear map @xmath26 as @xmath27 kitaev noticed that the @xmath25 norm is not stable under tensor products ( as it can easily be seen by looking at the transposition map ) , and considered the following `` regularization '' : @xmath28 in operator theory , the diamond norm was known before as the _ completely bounded trace norm _
; indeed , the @xmath25 norm of an operator is the @xmath29 norm of its dual , hence the diamond norm of @xmath14 is equal to the completely bounded ( operator ) norm of @xmath30 ( see ( * ? ? ?
* chapter 3 ) ) .
we shall need two simple properties of the diamond norm .
first , note that the supremum in the definition can be replaced by taking the value @xmath31 ( recall that @xmath32 is the dimension of the input hilbert space of the linear map @xmath14 ) ; actually , one could also take @xmath33 equal to the choi rank of the map @xmath14 , see ( * ? ? ?
* theorem 3.3 ) or ( * ? ? ? * theorem 3.66 ) .
second , using the fact that the extremal points of the unit ball of the @xmath24-norm are unit rank matrices , we always have @xmath34 moreover , if the map @xmath14 is hermiticity - preserving ( e.g. @xmath14 is the difference of two quantum channels ) , one can optimize over @xmath35 in the formula above , see ( * ? ? ?
* theorem 3.53 ) . given a map @xmath14 , it is in general difficult to compute its diamond norm .
computationally , there is a semidefinite program for the diamond norm , @xcite , which has a simple form and which has been implemented in various places ( see , e.g. @xcite ) .
we bound next the diamond norm in terms of the partial trace of the absolute value of the choi - jamiokowski matrix .
the diamond norm has application in the problem of quantum channel discrimination .
suppose we have an experiment in which our goal is to distinguish between two quantum channels @xmath14 and @xmath36 .
each of the channels may appear with probability @xmath37 .
then , celebrated result by helstrom @xcite gives us an upper bound on the probability of correct discrimination @xmath38 from the discussion in this section we readily arrive at @xmath39 the main goal of this work is to show the asymptotic behavior of the diamond norm . to achieve this , in section [ sec : sdp ] we find a new upper bound of on the diamond norm of a general map . in the case of a hermiticity preserving map it has a nice form @xmath40 next , in section [ sec : lower - bound ]
we prove that the well known lower bound on the diamond norm @xmath41 converges to a finite value for random independent quantum channels @xmath14 and @xmath36 in the limit @xmath42 .
we obtain that for channel sampled from the flat hilbert - schmidt distribution the value of the lower bound is @xmath43 the general case is discussed in - depth in the aforementioned section .
finally , in section [ sec : upper - bound ] we show that the upper bound also converges to the same value as the lower bound . from these results we have for channels sampled from the hilbert - schmidt distribution @xmath44
we discuss in this section some bounds for the diamond norm . for a matrix @xmath45 ,
we denote by @xmath46 and @xmath47 its right and left absolute values , i.e. @xmath48 when @xmath49 is the svd of @xmath45 . in the case where @xmath45 is self - adjoint
, we obviously have @xmath50 . in the result below
, the lower bound is well - known , while the upper bound appear in a weaker and less general form in ( * ? ? ?
* theorem 2 ) .
[ prop : bound - diamond ] for any linear map @xmath10 , we have @xmath51 consider the semidefinite programs for the diamond norm given in ( * ? ? ?
* section 3.2 ) : @xmath52 \text{subject to:}\quad & \begin{bmatrix } \rho_0 \otimes i_{d_2 } & x\\ x^ * & \rho_1 \otimes i_{d_2 } \end{bmatrix } \geq 0\\ & \rho_0,\rho_1\in m_{d_1}^{1,+}(\mathbb c)\\ & x \in m_{d_1d_2}(\mathbb c ) \end{aligned}\ ] ] @xmath53 \text{subject to:}\quad & \begin{bmatrix } y_0 & -j(\phi)\\ -j(\phi)^ * & y_1 \end{bmatrix } \geq 0\\ & y_0 , y_1 \in m_{d_1d_2}^+(\mathbb c ) \end{aligned}\ ] ] the lower and upper bounds will follow from very simple feasible points for the primal , resp . the dual problems .
let @xmath54 be a svd of the choi - jamiokowski state of the linear map .
for the primal problem , consider the feasible point @xmath55 and @xmath56 .
the value of the primal problem at this point is @xmath57 showing the lower bound . for the upper bound , set @xmath58 and @xmath59 , both psd matrices .
the condition in the dual problem is satisfied : @xmath60 and the proof is complete . if the map @xmath14 is hermiticity - preserving ( i.e. the matrix @xmath13 is self - adjoint ) , the inequality in the statement reads simply @xmath61 the two bounds in are equal iff the psd matrices @xmath62 and @xmath63 are both scalar .
indeed , the lower bound in can be rewritten as @xmath64 and the two bounds are equal exactly when the spectra of @xmath65 and @xmath66 ae flat .
the upper bound in can be seen as a strengthening of the following inequality @xmath67 , which already appeared in the literature ( e.g. ( * ? ? ?
* section 3.4 ) ) . indeed , again in terms of @xmath65 and @xmath66 , we have @xmath68 and @xmath69 .
there are several ways to endow the convex body of quantum channels with probability distributions . in this section
, we discuss several possibilities and the relations between them .
recall that the choi - jamiokowski isomorphism puts into correspondence a quantum channel @xmath70 with a bipartite matrix @xmath71 having the following two properties * @xmath13 is positive semidefinite * @xmath72 .
the above two properties correspond , respectively , to the fact that @xmath14 is complete positive and trace preserving .
hence , it is natural to consider probability measures on quantum channels obtained as the image measures of probabilities on the set of bipartite matrices with the above properties .
given some fixed dimensions @xmath73 and a parameter @xmath74 , let @xmath75 be a random matrix having i.i.d .
standard complex gaussian entries ; such a matrix is called a _
ginibre random matrix_. define then @xmath76 the random matrices @xmath77 and @xmath78 are called , respectively , _ wishart _ and _ partially normalized wishart_. the inverse square root in the definition of @xmath78 uses the moore - penrose convention if @xmath77 is not invertible ; note however that this is almost never the case , since the wishart matrices with parameter @xmath79 larger than its size is invertible with unit probability .
it is for this reason we do not consider here smaller integer parameters @xmath79 .
note that the matrix @xmath78 satisfies the two conditions discussed above : it is positive semidefinite and its partial trace over the second tensor factor is the identity : @xmath80 \\ & = ( { { \mathrm{tr}}}_2 w)^{-1/2 } \left ( \operatorname{tr}_2 w \right ) ( { { \mathrm{tr}}}_2 w)^{-1/2 } = i_{d_1}.\end{aligned}\ ] ] hence , there exists a quantum channel @xmath81 , such that @xmath82 ( note that @xmath78 , and thus @xmath14 are functions of the original ginibre random matrix @xmath83 ) .
[ def : measure - partially - normalized - wishart ] the image measure of the gaussian standard measure through the map @xmath84 defined in , and the equation @xmath82 is called the _ partially normalized wishart measure _ and is denoted by @xmath85 .
another way of introducing a probability distribution on the set of quantum channels is via the stinespring dilation theorem @xcite : for any channel @xmath86 , there exists , for some given @xmath87 , an isometry @xmath88 such that @xmath89 [ def : measure - isometries ] for any integer parameter @xmath79 , let @xmath90 be the image measure of the haar distribution on isometries @xmath91 through the map in . finally , one can consider the lebesgue measure on the convex body of quantum channels , @xmath92 . in this work
, we shall however be concerned only with the measure @xmath93 coming from normalized wishart matrices . in this section
we introduce and study the basic properties of a two - parameter family of probability measures which will appear later in the paper .
this family generalizes the symmetrized marcenko - pastur distributions from @xcite , see also @xcite for other occurrences of some special cases .
before we start , recall that the marcenko - pastur ( of free poisson ) distribution of parameter @xmath94 has density given by ( * ? ? ?
* proposition 12.11 ) @xmath95}(u)\ , du,\ ] ] where @xmath96 and @xmath97 .
[ def : smp - xy ] let @xmath98 be two free random variables having marcenko - pastur distributions with respective parameters @xmath99 and @xmath100 . the distribution of the random variable @xmath101 is called the _ subtracted marcenko - pastur distribution _ with parameters @xmath102 and is denoted by @xmath103 .
in other words , @xmath104 we have the following result . [
prop : smp - wishart ] let @xmath105 ( resp .
@xmath106 ) be two wishart matrices of parameters @xmath107 ( resp @xmath108 ) . assuming that @xmath109 and @xmath110 for some constants @xmath111 , then , almost surely as @xmath112 , we have @xmath113 the proof follows from standard arguments in random matrix theory , and from the fact that the schatten @xmath24-norm is the sum of the singular values , which are the absolute values of the eigenvalues in the case of self - adjoint matrices
. we gather next some properties of the probability measure @xmath103
. examples of this distribution are shown in fig .
[ fig : smp - xy ] .
let @xmath114 .
then , 1 .
if @xmath115 , then the probability measure @xmath103 has exactly one atom , located at 0 , of mass @xmath116 .
if @xmath117 , then @xmath103 is absolutely continuous with respect to the lebesgue measure on @xmath118 .
2 . define @xmath119 ^ 2 - 4 \left [ t_{x , y}(u ) \right]^3}. \end{split}\ ] ] the support of the absolutely continuous part of @xmath103 is the set @xmath120 ^ 2 - 4 \left [ t_{x , y}(u ) \right]^3 \geq 0\}.\ ] ] this set is the union of two intervals if @xmath121 and it is connected when @xmath122 , with @xmath123 3 . on its support , the density of @xmath103 is given by @xmath124^{\frac23}-2^{\frac23 } t_{x , y}(u)}{2^{\frac43 } \sqrt{3 } \pi u \left [ y_{x , y}(u ) \right]^{\frac13 } } \right|.\ ] ] the statement regarding the atoms follows from ( * ? ? ?
* theorem 7.4 ) .
the formula for the density and equation comes from stieltjes inversion , see e.g. ( * ? ? ?
* lecture 12 ) .
indeed , since the @xmath125-transform of the marcenko - pastur distribution @xmath126 reads @xmath127 , the @xmath125-transform of the subtracted measure reads @xmath128 the cauchy transform @xmath83 of @xmath103 is the functional inverse of @xmath129 . to write down the explicit formula for @xmath83
, one has to solve a degree 3 polynomial equation , and we omit here the details . the statement regarding the number of intervals of the support follows from .
the inequality is given by a polynomial of degree 6 which factorizes by @xmath130 , hence an effective degree 4 polynomial .
the nature of roots of this polynomial is given by the sign of its discriminant , which , after some algebra , is the same as the sign of @xmath131 , see @xcite .
in the case where @xmath35 , some of the formulas from the result above become simpler ( see also @xcite ) . the distribution @xmath132 is supported between @xmath133 . finally , in the case when @xmath134 , which corresponds to a flat hilbert - schmidt measure on the set of quantum channels , we get that @xmath135 .
+ we state here the main result of the paper . for the proof , see the following two sections , each providing one of the bounds needed to conclude .
[ thm : main ] let @xmath14 , resp .
@xmath36 , be two _ independent _ random quantum channels from @xmath136 having @xmath93distribution with parameters @xmath137 , resp .
. then , almost surely as @xmath42 in such a way that @xmath139 , @xmath140 ( for some positive constants @xmath102 ) , and @xmath141 , @xmath142 the proof follows from theorems [ thm : lower ] and [ thm : upper ] , which give the same asymptotic value . combining theorem [ thm : main ] with hellstrom s theorem for quantum channels , we get that the probability @xmath143 of distinguishing two quantum channels is equal to : @xmath144 additionally , any maximally entangled state may be used to achieve this value .
in this section we compute the asymptotic value of the lower bound in theorem [ thm : main ] . given two random quantum channels @xmath145 , we are interested in the asymptotic value of the quantity @xmath146 .
[ thm : lower ] let @xmath14 , resp .
@xmath36 , be two _ independent _ random quantum channels from @xmath136 having @xmath93distribution with parameters @xmath137 , resp .
@xmath138 . then , almost surely as @xmath147 in such a way that @xmath139 and @xmath140 for some positive constants @xmath102 , @xmath148 the proof of this result ( as well as the proof of theorem [ thm : lower ] ) uses in a crucial manner the approximation result for partially normalized wishart matrices .
[ prop : approximation - partial - normalization ] let @xmath149 a random wishart matrix of parameters @xmath150 , and consider its `` partial normalization '' @xmath78 as in . then , almost surely as @xmath42 in such a way that @xmath151 for a fixed parameter @xmath152 , @xmath153 note that in the statement above , the matrix @xmath77 is not normalized ; we have @xmath154 the marchenko - pastur distribution of parameter @xmath155 . in other words , @xmath156 , where @xmath83 is random matrix of size @xmath157 , having i.i.d .
standard complex gaussian entries .
let us introduce the random matrices @xmath158 the first observation we make is that the random matrix @xmath45 is also a ( rescaled ) wishart matrix .
indeed , the partial trace operation can be seen , via duality , as a matrix product , so we can write @xmath159 where @xmath160 is a complex gaussian matrix of size @xmath161 ; remember that @xmath79 scales like @xmath162 . since , in our model , both @xmath32 , @xmath163 grow to infinity , the behavior of the random matrix @xmath45 follows from @xcite . as @xmath42
, the random matrix @xmath164 converges in moments toward a standard semicircular distribution .
moreover , almost surely , the limiting eigenvalues converge to the edges of the support of the limiting distribution : @xmath165 the proof is a direct application of ( * ? ? ?
* corollary 2.5 and theorem 2.7 ) ; we just need to check the normalization factors . in the setting of (
* section 2 ) , the wishart matrices are not normalized , so the convergence result deals with the random matrices ( here @xmath166 and @xmath167 ) @xmath168 we look now for a similar result for the matrix @xmath169 ; the result follows by functional calculus .
[ lem : convergence - y ] almost surely as @xmath42 , the limiting eigenvalues of the random matrix @xmath170 converge respectively to @xmath171 : @xmath172 by functional calculus , we have @xmath173^{-1/2}$ ] , so , using the previous lemma , we get @xmath174^{-1/2 } = 1 + \frac{1}{2 } \frac{2}{\sqrt{t } d_2 } + o(d_2^{-1}),\ ] ] and the conclusion follows .
the case of @xmath175 is similar .
we have now all the ingredients to prove proposition [ prop : approximation - partial - normalization ] .
we have @xmath176 note that , almost surely , the three random matrix norms in the last equation above converge respectively to the following finite quantities @xmath177 the first and the third limit above follow from lemma [ lem : convergence - y ] , while the second one is the bai - yin theorem ( * ? ? ? *
theorem 2 ) or ( * ? ? ?
* theorem 5.11 ) .
let us now prove theorem [ thm : lower ] .
the result follows easily by approximating the partially normalized wishart matrices with scalar normalizations . by the triangle inequality , with @xmath178 and @xmath179
, we have @xmath180 the conclusion follows from propositions [ prop : smp - wishart ] and [ prop : approximation - partial - normalization ] .
the core technical result of this work consists of deriving the asymptotic value of the upper bound in theorem [ thm : main ] . given two random quantum channels @xmath145 , we are interested in the asymptotic value of the quantity @xmath181 .
[ thm : upper ] let @xmath14 , resp .
@xmath36 , be two _ independent _ random quantum channels from @xmath136 having @xmath93distribution with parameters @xmath137 , resp .
@xmath138 . then , almost surely as @xmath42 in such a way that @xmath139 , @xmath140 ( for some positive constants @xmath102 ) , and @xmath182 , @xmath183 the proof of theorem [ thm : upper ] is presented at the end of this section .
it is based on the following lemma which appears in @xcite , the proof being left to the reader ; see also ( * ? ? ?
* eq . ( 5.10 ) ) or ( * ? ? ?
* chapter x ) .
[ lem : absolute - value - bound ] for any matrices @xmath184 of size @xmath185 , the following holds : @xmath186 for a universal constant @xmath187 which does not depend on the dimension @xmath185 . for the sake of completeness
, we give here a proof , relying on a similar estimate for the schatten classes proved in @xcite .
using ( * ? ? ?
* theorem 8) , we have , for any @xmath188 : @xmath189 for some universal constant @xmath190 .
choosing @xmath191 gives the desired bound , for @xmath185 large enough .
the case of small values of @xmath185 is obtained by a standard embedding argument .
using the triangle inequality and lemma [ lem : absolute - value - bound ] , we first prove an approximation result ( as before , we write @xmath192 and @xmath193 ) : @xmath194 where we have used proposition [ prop : approximation - partial - normalization ] and the fact that @xmath195 .
this proves the approximation result , and we focus now on the simpler case of wishart matrices .
let us define @xmath196 it follows from ( * ? ? ?
* proposition 4.4.9 ) that the random matrix @xmath197 converges almost surely ( see appendix [ app : partial - trace - unitarily - invariant ] for the definition of almost sure convergence for a sequence of random matrices ) to a non - commutative random variable having distribution @xmath103 , see . moreover , using a standard strong convergence argument @xcite , the extremal eigenvalues of @xmath197 converge almost surely to the extremal points of the support of the limiting probability measure @xmath103 . hence , the almost sure convergence extends from the traces of the powers of @xmath197 to any continuous bounded function ( on the support of @xmath103 ) , in particular to the absolute value , i.e. to @xmath198 . from proposition
[ prop : partial - trace - flat ] , the asymptotic spectrum of the random matrix @xmath199 is flat , with all the eigenvalues being equal to @xmath200 which , by proposition [ prop : smp - wishart ] , is equal to @xmath201 , finishing the proof . we think that the condition @xmath202 in the statement can be replaced by a much weaker condition .
in this work we analyzed properties of generic quantum channels concentrating on the case of large system size . using tools provided by the theory of random matrices and the free probability calculus we showed that the diamond norm of the difference between two random channels asymptotically tends to a constant specified in theorem [ thm : main ] . in the case of channels corresponding to the simplest case @xmath134 , the limit value of the diamond norm of the difference is @xmath203 . in fig .
[ fig : convergence ] we illustrate the convergence of the upper and lower bound to this value .
this statement allows us to quantify the mean distinguishability between two random channels . ) .
the results were obtained via monte carlo simulation with 100 samples for each data point . ] to arrive at this result we considered an ensemble of normalized random density matrices , acting on a bipartite hilbert space @xmath204 , and distributed according to the flat ( hilbert - schmidt ) measure .
such matrices , can be generated with help of a complex ginibre matrix @xmath83 as @xmath205 . in the simplest case of square matrices
@xmath83 of order @xmath206 the average trace distance of a random state @xmath1 from the maximally mixed state @xmath207 behaves asymptotically as @xmath208 @xcite .
however , analyzing both reduced matrices @xmath209 and @xmath210 we can show that they become @xmath211 close to the maximally mixed state in sense of the operator norm , so that their smallest and largest eigenvalues do coincide .
this observation implies that the state @xmath1 can be directly interpreted as a jamiokowski state @xmath212 representing a stochastic map @xmath14 , as its partial trace @xmath213 is proportional to identity .
furthermore , as it becomes asymptotically equal to the other partial trace @xmath214 , it follows that a generic quantum channel ( stochastic map ) becomes unital and thus bistochastic .
the partial trace of a random bipartite state is shown to be close to identity provided the support of the limiting measure characterizing the bipartite state is bounded .
in particular , this holds for a family of subtract marchenko
pastur distributions defined in eq . as a free additive convolution of two rescaled marchenko
pastur distributions with different parameters and determining the density of a difference of two random density matrices . in this way
we could establish the upper bound for the average diamond norm between two channels and show that it asymptotically converges to the lower bound @xmath201 given in theorem [ thm : lower ] .
the results obtained can be understood as an application of the measure concentration paradigm @xcite to the space of quantum channels .
_ acknowledgments . _
i.n . would like to thank anna jencov and david reeb for very insightful discussion regarding the diamond norm , which led to several improvements and simplifications of the proof of proposition [ prop : bound - diamond ] .
i.n.s research has been supported by the anr projects rmtqit anr-12-is01 - 0001 - 01 and stoq anr-14-ce25 - 0003 - 01 , as well as by a von humboldt fellowship .
in this section we show a general result about unitarily invariant random matrices : under some technical convergence assumptions , the partial trace of a unitarily invariant random matrix is `` flat '' , i.e. it is close in norm to its average .
recall that the normalized trace functional can be extended to arbitrary permutations as follows : for a matrix @xmath215 , write @xmath216 recall the following definition from ( * ? ? ?
* section 4.3 ) . a sequence of random matrices @xmath217 is said to have _ almost surely limit distribution _
@xmath218 if @xmath219 [ prop : partial - trace - flat ] consider a sequence of hermitian random matrices @xmath220 and assume that 1 . both functions @xmath221 grow to infinity , in such a way that @xmath222 .
the matrices @xmath223 are unitarily invariant .
the family @xmath224 has almost surely limit distribution @xmath218 , for some compactly supported probability measure @xmath218 .
then , the normalized partial traces @xmath225(a_d)$ ] converge almost surely to multiple of the identity matrix : @xmath226 where @xmath227 is the average of @xmath218 : @xmath228 in the proof , we shall drop the parameter @xmath112 , but the reader should remember that the matrix dimensions @xmath229 are functions of @xmath185 and that all the matrices appearing are indexed by @xmath185 . to conclude , it is enough to show that @xmath230 since the statement for the smallest eigenvalue follows in a similar manner .
let us denote by @xmath231 ^ 2 = \operatorname{tr}_{(12)}(b ) - [ \operatorname{tr}_{(1)}(b)]^2\end{aligned}\ ] ] the average eigenvalue and , respectively , the variance of the eigenvalues of @xmath232 ; these are real random variables ( actually , sequences of random variables indexed by @xmath185 ) . by chebyshev s inequality
, we have @xmath233 note that one could replace the @xmath234 factor in the inequality above by @xmath235 by using samuelson s inequality @xcite , but the weaker version is enough for us .
we shall prove now that @xmath236 almost surely and later that @xmath237 almost surely , which is what we need to conclude .
to do so , we shall use the weingarten formula @xcite . in the graphical formalism for the weingarten calculus introduced in @xcite ,
the expectation value of an expression involving a random haar unitary matrix can be computed as a sum over diagrams indexed by permutation matrices ; we refer the reader to @xcite or @xcite for the details . using the unitary invariance of @xmath238
, we write @xmath239 , for a haar - distributed random unitary matrix @xmath240 , and some ( random ) eigenvalue vector @xmath241 . note that traces of powers of @xmath238
depend only on @xmath241 , so we shall write @xmath242 .
we apply the weingarten formula to a general moment of @xmath232 , given by a permutation @xmath243 : @xmath244 where @xmath245 are the cycles of @xmath246 , and @xmath247 denotes the conditional expectation with respect to the haar random unitary matrix @xmath22 . from the graphical representation of the weingarten formula (
* theorem 4.1 ) , we can compute the conditional expectation over @xmath22 ( note that below , the vector of eigenvalues @xmath241 is still random ) : @xmath248 above , @xmath249 is the weingarten function @xcite and @xmath250 is the moment of the diagonal matrix @xmath251 corresponding to the permutation @xmath252 .
the combinatorial factors @xmath253 and @xmath254 come from the initial wirings of the boxes respective to the vector spaces of dimensions @xmath32 ( initial wiring given by @xmath243 ) and @xmath163 ( initial wiring given by the identity permutation ) , see figure [ fig : graphical - wg ] .
the pre - factors @xmath255 contain the normalization from the ( partial ) traces . finally , the ( random ) factors @xmath250 are the normalized power sums of @xmath241 : @xmath256 where @xmath257 are the cycles of @xmath252 .
recall that we have assumed almost sure convergence for the sequence @xmath224 ( and , thus , for @xmath258 ) : @xmath259 -th group in the diagram corresponding to @xmath260.,scaledwidth=50.0% ] as a first application of the weingarten formula , let us find the distribution of the random variable @xmath261 .
obviously , @xmath262 actually , @xmath263 does not depend on the random unitary matrix @xmath22 , since @xmath264 from the hypothesis ( with @xmath265 ) , we have that , almost surely as @xmath112 , the random variable @xmath263 converges to the scalar @xmath266 .
let us now move on to the variance @xmath267 of the eigenvalues .
first , we compute its expectation @xmath268 . we apply now the weingarten formula for @xmath269 ; the sum has @xmath270 terms , which we compute below : * @xmath271 : @xmath272 * @xmath273 , @xmath274 : @xmath275 * @xmath276 , @xmath277 : @xmath278 * @xmath279 : @xmath280 . combining the expressions above with
, we get @xmath281 using the hypothesis , we have thus , as @xmath42 , @xmath282 let us now proceed and estimate the variance of @xmath267 ; more precisely , let us compute @xmath283 . as
before , we shall compute the expectation in two steps : first with respect to the random haar unitary matrix @xmath22 , and then , using our assumption , with respect to @xmath241 , in the asymptotic limit . to perform the unitary integration , note that the weingarten sum is indexed by a couple @xmath284 ,
so it contains @xmath285 terms , see @xcite . in appendix
[ sec : variance ] we have computed the variance of @xmath267 with the usage of symmetry arguments .
the result , to the first order reads @xmath286 ^ 2.\ ] ]
taking the expectation over @xmath241 and the limit ( we are allowed to , by dominated convergence ) , we get @xmath287 ^ 2.\ ] ] we put now all the ingredients together : @xmath288 ^ 2 } \sim \frac{cd_1^{-2}d_2^{-4}}{[\varepsilon^2 d_1^{-1 } - ( 1+o(1))c'd_2^{-2}]^2},\ ] ] where @xmath289 non - negative constants depending on the limiting measure @xmath218 . using @xmath141 ,
the dominating term in the denominator above is @xmath290 , and thus we have : @xmath291 since the series @xmath292 is summable , we obtain the announced almost sure convergence by the borel - cantelli lemma , finishing the proof .
we remind here , that @xmath293 and @xmath294 ( a)$ ] . because we assume that @xmath238 has unitarly invariant distribution , we can write @xmath295 where @xmath296 is @xmath297-th column of matrix @xmath22 and @xmath298 we denote @xmath299 and consider mixed moments computed in appendix [ sec : mixed - moments ] . @xmath300 and symmetric mixed moments @xmath301 let @xmath302 we have @xmath303 as @xmath304 , in the above @xmath305 . direct computations with the usage of symmetric moments @xmath306 give us @xmath307 \\ & = \frac{2 \left(d_1 ^ 2 - 1\right ) \left(d_2 ^ 2 - 1\right ) } { d_2 ^ 2 \left(d_1 ^ 2 d_2 ^ 2 - 1\right){}^2 \left(d_1 ^ 4 d_2 ^ 4 - 13 d_1 ^ 2 d_2 ^ 2 + 36\right ) } \big(d_1 ^ 4 d_2 ^ 4 \left(\mu _ 1 ^ 2-\mu _ 2\right){}^2 \\ & \ \ \ \ \ \ \ + d_1 ^ 2 d_2 ^ 2 \left(11 \mu _ 1 ^ 4 - 22 \mu _ 2 \mu _ 1 ^ 2 + 20 \mu _ 3 \mu _ 1 - 4 \mu _ 2 ^ 2 - 5 \mu _ 4\right)+5 \left(3 \mu _ 2 ^ 2 - 4 \mu _ 1 \mu _ 3+\mu _
4\right)\big ) \\ & = \frac{2 ( \mu_1 ^ 2 - \mu_2)^2}{d_1 ^ 2 d_2 ^ 4}(1 + o(1 ) ) \end{split}\ ] ]
we have the following formulas for mixed moments , which covers all possible cases ( because of the symmetry ) @xmath308 here we will consider expectations of the following kind @xmath312 note , that if we multiply matrix @xmath22 by a unitary matrix which does not change the first column we will not change the expectation , in fact we can integrate over the subgroup of matrices which does not change the first column of @xmath22 . now for a moment we fix matrix @xmath22 and consider @xmath313 where matrices @xmath91 are in the form @xmath314 the @xmath315 is an expectation with respect to the haar measure on @xmath316 embedded in @xmath317 , in the above way .
note , that the vector @xmath318 represents a random orthogonal vector to the @xmath319 .
first we calculate @xmath320 now , using standard integrals we obtain @xmath321 where @xmath322 and incorporates the condition that first element of vector @xmath323 is zero .
now we obtain , after elementary calculations , using the fact , that @xmath22 is unitary @xmath324 where @xmath325 is a swap operation on two systems of dimensions @xmath326 , i.e. @xmath327 . so we get @xmath328 next we consider @xmath329 so we have obtained @xmath330 in the above formulas we used @xmath331 and fact , that @xmath332 . | in this work we analyze properties of generic quantum channels in the case of large system size .
we use the random matrix theory and free probability to show that the distance between two independent random channels tends to a constant value as the dimension of the system grows larger . as a measure of the distance we use the diamond norm . in the case of a flat hilbert - schmidt distribution on quantum channels
, we obtain that the distance converges to @xmath0 .
furthermore , we show that for a random state @xmath1 acting on a bipartite hilbert space @xmath2 , sampled from the hilbert - schmidt distribution , the reduced states @xmath3 and @xmath4 are arbitrarily close to the maximally mixed state .
this implies that , for large dimensions , the state @xmath1 may be interpreted as a jamiokowski state of a unital map .
= 1 |
solar coronal magnetic field is a major source of dynamics and thermodynamics of global coronal corona and transient coronal events including the coronal heating , solar flares , coronal mass ejections and the solar wind .
its dynamics affects space weather processes that may impact earth s magnetosphere and atmosphere and affect life on our planet .
thus , the knowledge of coronal magnetic field strength and topology represents one of major goals of the solar physics .
currently , coronal magnetic fields can not directly measured .
the major techniques that are currently used to derive the global magnetic structures of the solar corona represent indirect methods including potential field source surface ( pfss ) models , non - linear force - free field ( nlfff ) models and multi - dimensional magnetohydrodynamic ( mhd ) models of the global solar corona .
these methods use solar photospheric scalar or vector magnetic field as input directly derived from photospheric magnetograms .
however , as we discussed in @xcite , all these methods can not adequately describe the dynamics of magnetic fields observed in the solar corona driven by current carrying structures .
moreover , they do not provide accurate information about the coronal thermodynamics , and thus can not be used for modeling the coronal emission measures derived from extreme ultraviolet ( euv ) observations .
the mhd modeling of the solar corona provides a self - consistent time - dependent treatment of the plasma pressure , gravitational and magnetic forces , but are limited by approximations used for describing the coronal heating , and the uncertainties in the boundary conditions that are deduced from synoptic data .
thus , direct measurements of the coronal magnetic field remains one of the most reliable and challenging ways for characterization of solar coronal processes .
direct measurement of the coronal magnetic field is the most challenging problem in observational solar physics .
a major progress here was reached with the deployment of the coronal multichannel polarimeter ( comp ) @xcite . in order to interpret such type of data
, a vector tomography method has been developed for 3d reconstruction of the global coronal magnetic field @xcite , and recently , first 3d reconstruction of the global coronal magnetic field has been performed based on the comp data @xcite .
another , thought implicit but also based on coronal observations , way to reconstruct some global coronal magnetic field structures was investigated in @xcite .
there we applied the tomography method that employs stereo data to reconstruct 3d density of global solar corona and related it to the coronal magnetic field structure .
this method was applied to characterize the solar corona over the carrington rotation 2066 occurred during deep solar minimum for the range of coronal heights from @xmath0 to @xmath3 and provided description of the open / close magnetic field boundaries in the solar corona . in this paper
, we apply this methodology to study global solar coronal structures for the period of over the half of carrington rotation 2131 , which represents the case of solar maximum .
specifically , we use the stereo / cor1 coronagraph data for half a solar rotation period during cr 2131 as input for the tomographic reconstruction of the 3d coronal electron density and the configuation of magnetic field associated with them .
our results are complemented by the 3d emissivity obtained by tomography method applied for the the stereo _ extreme ultraviolet imager _
( euvi ) data in the 195 band .
we also compared the reconstructed 3d coronal structures with those implied by the pfss model . in section 4 , we present the comparison of the 3d coronal structure near the solar maximum to those at deep solar minimum ( cr 2066 ) .
the corona is optically thin at visible and euv wavelengths dominated by lines of ionized iron and calcium .
thus , the observed visible and euv coronal flux represents an integral of the emissivity along any line of sight ( los ) .
this makes it impossible to reconstruct the spatial distribution of the emissivity from a single ( in a geometric sense ) measurement or projection .
the solution space is reduced if we have measurements from many different viewpoints .
the three - dimensional ( 3d ) reconstruction of optically thin emission over depths can be performed by using tomography methods that employ the 2d observations of an object from different view angles .
the possibility of reconstructing a function from its projections was first studied by @xcite .
several decades later , this purely mathematical method formed the basis for the x - ray computer tomography that produces cross - sectional slices of human bodies @xcite and @xcite .
today , tomography is used in many fields of science including medical research , material structure testing , geophysics , heliophysics and astrophysics @xcite . in astrophysical applications
the input data can suffer from noise and data incompleteness .
while the application of tomography to astrophysical objects is limited by the noisy input data and its incompleteness , the regularization method allows solar coronal tomography to produce reliable 3d reconstructions of emissivity and plasma density @xcite .
the key input for tomographic reconstruction of the plasma density of extended solar coronal structures requires observational data from more than two vantage points .
tomography applications for coronal studies typically assume a rigid rotation of the coronal density structures . for reliable density reconstruction ,
the algorithm utilizes coronagraph data for half a solar rotation as input if observed from a single spacecraft and requires coronal structures to remain stable over their observation periods @xcite .
however , depending on the positions of a coronal structure relative to the spacecraft during the observation period , the stationarity assumption for that structure can be reduced to about a week @xcite . for our density reconstructions we used the polarized brightness ( pb ) intensity images from the cor1 instrument onboard
the stereo - b spacecraft taken 28 images per half a solar rotation as input for the tomographic inversion .
we limited here the data input for the tomography based on cor1 data to the stereo - b spacecraft because cor1-b had lower levels of stray light during cr 2066 than cor1-a . in the stereo / cor1-b field of view ( below @xmath4 ) , the white - light pb coronal emission
is dominated by scattering sunlight on the free electron in the corona @xcite .
the intensity of the pb - signal as a fraction of the mean solar brightness is given as @xmath5 where @xmath6 is the electron density , @xmath7 is a vector in plane - of - sky ( pos ) from the sun center to the los and perpendicular to los , @xmath8 is length along the los , @xmath9 is the unit vector along the los , and @xmath10 is the radius - vector .
the kernel function [ @xmath11 is defined by the thompson scattering effect @xcite : @xmath12 \frac{\rho^2}{r^2 } \label{kernelfunction_pb}\ ] ] where the expressions for @xmath13 and @xmath14 are the same as those given by @xcite , @xmath15 is the thompson scattering cross - section for a single electron , and the linear limb - darkening coefficient [ @xmath16 is set to @xmath17 in the present calculations . because cor1 views the corona close to the limb , the instrument has a significant amount of scattered light , which must be subtracted from the image prior to be applied in the reconstruction method .
proper removal of instrumental scattered light is essential for coronal reconstruction .
one way is to subtract a monthly minimum ( mm ) background .
the monthly minimum approximates the instrumental scatter by bf calculating the lowest value of each pixel in all images during a period of about one month .
however , this method tends to overestimate the scattered light in the streamer belt ( equatorial region ) .
the lowest value of these pixels during a month usually contains both the scattered light and the steady - intensity value from the corona .
hence , the subtraction of mm background would output an electron density in streamer regions that is lower than the actual density .
another way to account for the scattered light is to subtract a roll minimum ( rm ) background .
the roll minimum background is the lowest value of each pixel obtained during a roll maneuver of the spacecraft ( instrument ) around its optical axis . because the coronal polar regions are much darker than the equatorial ones , the lowest pixel values in the equatorial region during the roll maneuver are nearer to the value of the scattered - light intensity than the mm .
the sensitivity of the cor1-b instrument decreases at a rate of about @xmath18 % per month @xcite .
moreover , variations in the spacecraft s distance from the sun cause changes of the amount of scattered light in the coronagraph images .
but the roll manoeuvres are done rarely .
therefore , it is impossible to use an rm background obtained in one month for data for the following month when the highest possible photometric accuracy is needed .
a background image for the period between the roll maneuveres can be derived from interpolation of the rm backgrounds over time in a such a way that this temporal dependence follows the temporal dependence of the mm backgrounds , because the mm background images are available for every month . in our study
we used the approach implemented by w. thompson in the solarsoft idl routine secchi_prep with the keyword parameter calroll .
the photometric calibration is based on jupiter s passage through the cor1 fov @xcite . after subtracting the scattered light ,
a median filter with a width of three pixels was applied to reduce anomalously bright pixels caused by cosmic rays .
then , for the cr 2066 reconstruction , every third image pixel was selected ( resulting in a @xmath19 pixel image from original @xmath20 ) to reduce the computer memory size . in case of cr 2112 and 2131 reconstructions
, we used original @xmath21 input data resolution .
we used the reconstruction domain as a spherical grid with a size of @xmath22 covering heliocentric distances from @xmath0 to @xmath1 , carrington latitudes from @xmath23 to @xmath24 , and carrington longitudes from @xmath25 to @xmath26 , respectively .
the inversion was performed for the function @xmath27 here , the elements @xmath28 of the column matrix @xmath29 contain the values of electron density [ @xmath6 ] in the grid cells with index @xmath31 , and @xmath32 is the data value for the @xmath33-th ray , where index @xmath34 accounts for both the viewing direction and pixel position in the image . the element @xmath35 of the matrix @xmath36 represents the intersection of volume element @xmath37 with los related to pixel @xmath33 , multiplied by the kernel function that is defined by the thompson scattering effect for the pb - intensity signal
( see equation ( [ thompsonscat_pb ] ) ) .
the second term on the right - hand side of equation ( [ minfun ] ) is the regularization term that minimizes the effects of noise and data gaps @xcite .
the matrix @xmath38 is a diagonal - like matrix such that the regularization is the first - order smoothing term , _
i.e. _ operation @xmath39 produces the square difference in value between two neighboring grid cells , summed over all cells .
the regularization parameter [ @xmath40 regulates balance between the smoothness of the solution on one hand and the noise and reconstruction artifacts on the other .
the result of the inversion depends on a number of factors , including the number of iterations and the value of @xmath41 .
the value of @xmath41 was chosen using the cross - validation method @xcite .
the iterations are performed until the first term in equation ( [ minfun ] ) became slightly lower than the data noise level , which is essentially the poisson noise in the data .
the coronal electron density drops very rapidly with distance from the sun , introducing a wide dynamic range in the data , which causes linear artifacts in the reconstruction . to increase the contribution of signals from those los that pass through the low density regions , and to reduce the artifacts in the numerical reconstruction at larger distances from the sun , we applied a set of weighting coefficients ( or preconditioning ) @xmath42 were applied for the first term in equation ( [ minfun ] ) in such way that @xmath43 . here ,
@xmath44 is the inverse fourier transform of the function @xmath45 on @xmath46 with harmonics taken up to first order , where @xmath45 is the data value at the position @xmath47 in the polar coordinate system for some particular image .
the value of @xmath48 is fixed for a given pixel and set equal to the radial distance from the center of the sun s disk to the pixel .
a more detailed description of used tomography method is given in @xcite .
the error estimation of the tomographic method has been investigated in @xcite .
the reconstruction results are discussed in section [ sect_cr2131 ] .
the stereo / euvi instrument observes the solar corona at heights up to about @xmath49 in four spectral channels ( 171 , 195 , 284 , and 304 ) that span the 0.1 to 20 mk temperature range @xcite .
the measured coronal emission in the 171 , 195 , and 284 channels can be represented as the result of emission integrated over the los as @xmath50 where @xmath51 is the emissivity at the position @xmath10 in the selected channel , _
i.e. _ light intensity ( in photons per second for example ) emitted per unit volume , per unit solid angle .
the coefficient @xmath52 accounts for pixel size , aperture , and distance to the sun . as input , we used euvi 195 images calibrated by applying idl solarsoft routines . to reduce anomalously bright pixels caused by cosmic rays ,
the idl solarsoft routine despike_gen was applied .
three images taken with about two hours difference were averaged into one .
three averaged images per day were taken during a period of half a solar rotation .
then , every fourth pixel was taken , resulting in @xmath53 input image .
we inverted @xmath51 in the same manner as for the electron density in the white - light tomography with @xmath54 and @xmath6 in equation ( [ thompsonscat_pb ] ) substituted by @xmath52 and @xmath55 , respectively , according to equation ( [ euvi_emiss_los ] ) .
the inversion result is the 3d emissivity distribution for the euvi 195 channel in the coronal range from @xmath56 to @xmath57 .
figure [ fig_cr2131_euvi_rec_sph ] shows a spherical cross - section of the reconstructed euvi 195 emissivity at a heliocentric distance of @xmath58 for cr 2131 .
the reconstruction results are discussed in section [ sect_cr2131 ] .
cr 2131 represents the period near maximum of solar activity cycle . in this study , we performed two types of tomographic reconstructions : 3d reconstruction of the electron density based on stereo / cor1 data and 3d reconstruction for the euvi 195 emissivity [ units of photons s@xmath59 sr@xmath59 @xmath60 based on stereo / euvi data . in order to demonstrate a general structure of the coronal streamer belt for cr 2131 , figure [ fig_cr2131_rec_sph ] shows a spherical cross - section of the electron density at the heliocentric distance of @xmath61 and figure [ fig_cr2131_euvi_rec_sph ] shows a spherical cross - section of the euvi 195 emissivity at the heliocentric distance of @xmath58 .
figure [ fig_rss_cor1_phi_cr2131 ] shows several meridional cross - sections of the electron density ( range from @xmath0 to @xmath3 ) and euvi 195 emissivity ( range from @xmath56 to @xmath62 ) .
the figure with a set of all cross - sections is available in the electronic suplemental material .
the black contour lines in figure [ fig_rss_cor1_phi_cr2131 ] show boundaries between open and closed magnetic - field structures in two pfss models with source surface heliocentric distances [ @xmath63 at @xmath0 and @xmath64
. the pfss model with @xmath65 does not coincide with the derived positions of the streamer and pseudo - streamer as well as with the coronal hole positions indicated by the stereo / euvi 195 emissivity 3d reconstruction .
the pfss model with @xmath66 appears to fit the latter structures better .
it was previously demonstrated that the position of maximal gradient of the density and 195 emissivity can be an indicator for the position of boundary between closed and open magnetic structures @xcite .
therefore , in order to exhibit the possible real boundary positions , figure [ fig_rss_cor1_phi_cr2131_grad ] shows gradient of electron density and euvi 195 emissivity , @xmath67 and @xmath68 , for cr 2131 and for the same longitudinal cross - sections as in figure [ fig_rss_cor1_phi_cr2131 ] . both figures [ fig_rss_cor1_phi_cr2131 ] and [ fig_rss_cor1_phi_cr2131_grad ]
demonstrate that the magnetic field configuration during cr 2131 has a tendency to become radially open at heliocentric distances below @xmath2 .
this is lower than corresponding distance during the deep solar minimum near cr 2066 @xcite .
figure [ fig_rec_comp ] shows the radial profiles of the maximum ( left panel ) and average ( right panel ) electron density values at different heliocentric distances for cr 2066 ( dashed green ) , 2112 ( solid red ) , and 2131 ( solid black ) .
cr 2066 corresponds to deep solar minimum , while cr 2131 corresponds to period near solar maximum .
the maximum values of coronal density mostly reflect the closed magnetic field structures represented by solar active regions , while average values include contributions from both open and closed field regions .
the plot suggests that the maximum coronal density at the solar maximum is larger than the coronal density at two carrington rotations representing the periods of lower magnetic activity .
specifically , at the lower boundary ( at 1.8 @xmath69 ) of the reconstruction region the maximum density at solar maximum is more that twice greater than the maximum coronal density at solar minimum represented by cr 2066 .
this reflects the fact that the total unsigned magnetic flux that is structured in the form of coronal active regions during solar maximum is a factor of 2 - 3 greater than the emerging magnetic flux during solar minimum @xcite .
the average coronal density is a factor of 1.6 greater during solar maximum as compared to solar minimum , because it reflects the contribution of plasma emission formed in the diffuse corona as well as coronal active regions .
figure [ fig_rec_comp_3d ] shows the snapshot of the isosurface of coronal electron density value of @xmath70 ( in blue ) for cr 2066 occured at solar minimum ( left ) and cr 2131 at solar maxium ( right ) .
the corresponding movie showing the density isosurface over these two carrington rotations are available provided in online version of the paper . the figure and the movie
clearly demonstrate that the filling factor of the plasma with the density of @xmath70 increases at solar maximum , which is consistent with figure [ fig_rec_comp ] .
we applied the tomography method to stereo - b / cor1 data to derive the reconstructions of the 3d global coronal electron density and euvi 195 emissivity based on stereo/ cor1 and stereo / euvi observations near solar maximum represented by carrington rotation cr 2131 ( december 2012 ) .
we complemented the tomography with mhd simulations to obtain the open / closed magnetic field boundaries in the solar corona .
because the 3d reconstructions are based entirely on _ coronal _ data , these results could serve as an independent test and/or as an additional constraint for the models of the solar corona .
specifically , we used the pfss model for different source surface distances as a test case for the reconstructed 3d electron density and euvi 195 emissivity structures .
we have shown that magnetic field structures deviate significantly from the pfss approximation as evident from the derived boundaries of open / closed field for both solar minimum @xcite and solar maximum .
this suggests that potential field approximation for the low corona is not valid approximation when describing its large - scale structures .
it is important to note that the magnetic field becomes radial at heights greater than @xmath71 during solar minimum , while its open up at heights below @xmath72 during solar maximum .
our studies also show that the average and the maximum electron densities in the low solar corona at heights @xmath73 are about 2 times greater than that obtained during solar minimum .
this is an important result that can be extrapolated to coronal densities of active and young solar - like stars .
these stars show much greater levels of stellar activity in terms of at @xmath74 times greater surface magnetic flux , up to 1000 times more luminous in x - rays and at least 10 times denser and hotter corona .
these results open up an interesting opportunity to provide scaling of the electron density of the solar corona with different levels of magnetic activity traced by its x - ray luminosity and average surface magnetic flux .
these scaling laws can be then used for characterization of stellar corona of active stars , which may provide insights on the nature of stellar coronal heating at various phases of their evolution . ,
p . , lemen , j. r. , tarbell , t. d. , et al .
2004 , in society of photo - optical instrumentation engineers ( spie ) conference series , vol . 5171 , telescopes and instrumentation for solar astrophysics , ed . s. fineschi & m. a. gummin , 111122 .
the reconstruction is obtained by tomography based on cor-1 data obtained during december 4 - 17 , 2012 ( cr 2131 ) .
solid black and dashed red lines indicate the magnetic neutral line in pfss model with source surface located at @xmath0 and @xmath61 , respectively . ] .
the reconstruction is obtained by tomography based on cor-1 data obtained during december 4 - 17 , 2012 ( cr 2131 ) .
solid black and dashed red lines indicate the magnetic neutral line in pfss model with source surface located at @xmath0 and @xmath61 , respectively . ] to @xmath3 ) and euvi 195 data ( emissivity in the range from @xmath56 to @xmath62 ) .
carrington longitudes for cross - sections are shown at upper right corners .
the figure with a set of all cross - sections is available in the electronic supplementary material .
the contour black lines show boundaries between open and closed magnetic - field structures for the pfss models with source surface located at @xmath0 and @xmath64.,title="fig : " ] to @xmath3 ) and euvi 195 data ( emissivity in the range from @xmath56 to @xmath62 ) .
carrington longitudes for cross - sections are shown at upper right corners .
the figure with a set of all cross - sections is available in the electronic supplementary material .
the contour black lines show boundaries between open and closed magnetic - field structures for the pfss models with source surface located at @xmath0 and @xmath64.,title="fig : " ] + to @xmath3 ) and euvi 195 data ( emissivity in the range from @xmath56 to @xmath62 ) .
carrington longitudes for cross - sections are shown at upper right corners .
the figure with a set of all cross - sections is available in the electronic supplementary material .
the contour black lines show boundaries between open and closed magnetic - field structures for the pfss models with source surface located at @xmath0 and @xmath64.,title="fig : " ] to @xmath3 ) and euvi 195 data ( emissivity in the range from @xmath56 to @xmath62 ) .
carrington longitudes for cross - sections are shown at upper right corners .
the figure with a set of all cross - sections is available in the electronic supplementary material .
the contour black lines show boundaries between open and closed magnetic - field structures for the pfss models with source surface located at @xmath0 and @xmath64.,title="fig : " ] + to @xmath3 ) and euvi 195 data ( emissivity in the range from @xmath56 to @xmath62 ) .
carrington longitudes for cross - sections are shown at upper right corners .
the figure with a set of all cross - sections is available in the electronic supplementary material .
the contour black lines show boundaries between open and closed magnetic - field structures for the pfss models with source surface located at @xmath0 and @xmath64.,title="fig : " ] to @xmath3 ) and euvi 195 data ( emissivity in the range from @xmath56 to @xmath62 ) .
carrington longitudes for cross - sections are shown at upper right corners .
the figure with a set of all cross - sections is available in the electronic supplementary material .
the contour black lines show boundaries between open and closed magnetic - field structures for the pfss models with source surface located at @xmath0 and @xmath64.,title="fig : " ] + to @xmath3 ) and euvi 195 data ( emissivity in the range from @xmath56 to @xmath62 ) .
carrington longitudes for cross - sections are shown at upper right corners .
the figure with a set of all cross - sections is available in the electronic supplementary material .
the contour black lines show boundaries between open and closed magnetic - field structures for the pfss models with source surface located at @xmath0 and @xmath64.,title="fig : " ] to @xmath3 ) and euvi 195 data ( emissivity in the range from @xmath56 to @xmath62 ) .
carrington longitudes for cross - sections are shown at upper right corners .
the figure with a set of all cross - sections is available in the electronic supplementary material .
the contour black lines show boundaries between open and closed magnetic - field structures for the pfss models with source surface located at @xmath0 and @xmath64.,title="fig : " ] and @xmath68 , for cr 2131 .
carrington longitudes for cross - sections are shown at upper right corners .
color scales are different for different cross - sections to enhance the images .
the contour black lines show boundaries between open and closed magnetic - field structures for the pfss models with source surface located at @xmath0 and @xmath64.,title="fig : " ] and @xmath68 , for cr 2131 .
carrington longitudes for cross - sections are shown at upper right corners .
color scales are different for different cross - sections to enhance the images .
the contour black lines show boundaries between open and closed magnetic - field structures for the pfss models with source surface located at @xmath0 and @xmath64.,title="fig : " ] + and @xmath68 , for cr 2131 .
carrington longitudes for cross - sections are shown at upper right corners .
color scales are different for different cross - sections to enhance the images .
the contour black lines show boundaries between open and closed magnetic - field structures for the pfss models with source surface located at @xmath0 and @xmath64.,title="fig : " ] and @xmath68 , for cr 2131 .
carrington longitudes for cross - sections are shown at upper right corners .
color scales are different for different cross - sections to enhance the images .
the contour black lines show boundaries between open and closed magnetic - field structures for the pfss models with source surface located at @xmath0 and @xmath64.,title="fig : " ] + and @xmath68 , for cr 2131 .
carrington longitudes for cross - sections are shown at upper right corners .
color scales are different for different cross - sections to enhance the images .
the contour black lines show boundaries between open and closed magnetic - field structures for the pfss models with source surface located at @xmath0 and @xmath64.,title="fig : " ] and @xmath68 , for cr 2131 .
carrington longitudes for cross - sections are shown at upper right corners .
color scales are different for different cross - sections to enhance the images .
the contour black lines show boundaries between open and closed magnetic - field structures for the pfss models with source surface located at @xmath0 and @xmath64.,title="fig : " ] + ( in blue ) for cr 2066 ( left ) and 2131 ( right ) .
the orange sphere is located at @xmath75 just for scaling purpose .
the movie demonstrating the isosurfaces for the whole longitudinal projection range is provided in online version of the paper . ] | measurement of the coronal magnetic field is a crucial ingredient in understanding the nature of solar coronal dynamic phenomena at all scales .
we employ stereo / cor1 data obtained near maximum of solar activity in december 2012 ( carrington rotation , cr 2131 ) to retrieve and analyze the three - dimensional ( 3d ) coronal electron density in the range of heights from @xmath0 to @xmath1 using a tomography method and qualitatively deduce structures of the coronal magnetic field .
the 3d electron density analysis is complemented by the 3d stereo / euvi emissivity in 195 band obtained by tomography for the same cr period .
we find that the magnetic field configuration during cr 2131 has a tendency to become radially open at heliocentric distances below @xmath2 .
we compared the reconstructed 3d coronal structures over the cr near the solar maximum to the one at deep solar minimum .
results of our 3d density reconstruction will help to constrain solar coronal field models and test the accuracy of the magnetic field approximations for coronal modeling . |
@xcite performed a uniform survey over a @xmath3 12@xmath8 strip of the southern milky way extending from _
l _ @xmath9
250@xmath8 and _ l _ @xmath9 360@xmath8 .
they employed the curtis - schimidt telescope of the cerro tololo interamerican observatory and a pair of blue and red filters . from that survey
the authors recognised 262 star clusters , 63 of which had not been previously catalogued . for each idenfied object , they assessed the richness of stars on both plates as well as the possible reality of being a genuine star cluster . up to date , less than 25 per cent of the van den bergh - hagen ( vdbh ) objects have some estimation of their fundamental properties ( reddening , distance , age , etc ) . in general terms , according to the most updated version of the open cluster catalogue compiled by ( * ? ? ?
* as of january 2016 ) , vdbh clusters are mostly of relatively small size , with diameters smaller than @xmath3 5 pc , although some few ones have diameters twice as big this value . on the other hand , although @xmath3 60 per cent of them are located inside a circle of 2 kpc in radius from the sun , the remaining ones reach distances as large as @xmath3 12 kpc . indeed , nearly 15 per cent of the sample is located at distances larger than 5 kpc . as for their ages , the vdbh clusters expand over an interesting age regime , from those with some few myr up to the older ones with more than 3 gyr . at this point , it appears interesting to estimate fundamental parameters of those overlooked vdbh objects , particularly those located far away from the sun , in order to improve our knowledge of the galactic open cluster system beyond the solar neighbourhood . in this paper
, we present a comprehensive photometric study of vdbh1 , 10 , 31 , 72 , 87 , 92 and 118 ; the last four clusters were discovered by @xcite .
as far as we are aware , previous photometric studies were performed for vdbh 1 ( = haffner7 ) , 10 (= ruprecht35 ) and 31 (= ruprecht60 ) @xcite . in section 2
we describe the collection and reduction of the available photometric data and their thorough treatment in order to build extensive and reliable data sets .
the cluster structural and fundamental parameters are derived from star counts and colour - magnitude and colour - colour diagrams as described in section 3 .
the analysis of the results of the different astrophysical parameters obtained is carried out in section 4 , where implications about the stage of their dynamical evolution are suggested .
finally , section 5 summarizes the main conclusion of this work .
we make use of images obtained with the johnson @xmath0 , kron - cousins @xmath1 and washington @xmath10 filters , using a 4k@xmath114k ccd detector array ( scale of 0.289@xmath12/pixel ) attached to the 1.0-m telescope at the cerro tololo inter - american observatory ( ctio ) , chile , in 2011 january 31february 4 ( ctio program # 2011a-0114 , image header information : pi : clari , observers : clari - palma ) .
the nights were of photometric quality with a typical seeing of 1.1@xmath12 .
the data sets used in this work were downloaded from the public website of the national optical astronomy observatory ( noao ) science data management ( sdm ) archives.the log of the observations is presented in table [ tab : table1 ] , where the main astrometric and observational information is summarized . [ cols="<,^,^,^,^,^,^,^ " , ] note : to convert 1 arcsec to pc , we use the following expression,10@xmath1110@xmath13sin(1/3600 ) , where @xmath14 is the true distance modulus .
we used the mean cluster radii to extract the cluster colour - magnitude diagrams ( cmds ) .
they account for the luminosity function , colour distribution and stellar density of the stars distributed along the cluster line of sights , so that we statistically cleaned the cmds before using them to estimate the cluster fundamental parameters .
we employed the cleaning procedure developed by ( * ? ? ? *
see their fig .
the method compares the extracted cluster cmd to distinct cmds composed of stars located reasonably far from the object , but not too far so as to risk losing the local field - star signature in terms of stellar density , luminosity function and/or colour distribution .
here we chose four field regions , each one designed to cover an equal area as that of the cluster , and placed around the cluster .
note that the four selected fields could not adequately represent the fore / background of the cluster if the extinction varies significantly accross the field of view .
comparisons of field and cluster cmds have long been done by comparing the numbers of stars counted in boxes distributed in a similar manner throughout both cmds . however , since some parts of the cmd are more densely populated than others , counting the numbers of stars within boxes of a fixed size is not universally efficient .
for instance , to deal with stochastic effects at relatively bright magnitudes ( e.g. , fluctuations in the numbers of bright stars ) , larger boxes are required , while populous cmd regions can be characterized using smaller boxes .
thus , the use of boxes of different sizes distributed in the same manner throughout both cmds leads to a more meaningful comparison of the numbers of stars in different cmd regions .
precisely , the procedure of @xcite carries out the comparison between field - star and cluster cmds by using boxes which vary their sizes from one place to another throughout the cmd and are centred on the positions of every star found in the field - star cmd . by starting with reasonably large boxes typically ( @xmath15(magnitude),@xmath15(colour ) ) = ( 1.00 , 0.50 ) mag centred on each star in the four field cmds and by subsequently reducing their sizes until they reach the stars closest to the boxes centres in magnitude and colour , separately , we defined boxes which result in use of larger areas in field cmd regions containing a small number of stars , and vice versa .
note that the definition of the position and size of each box involves two field stars , one at the centre of the box and another -the closest one to box centre - placed on the boundary of that box .
next , we plotted all these boxes for each field cmd on the cluster cmd and subtracted the star located closest to each box centre .
since we repeated this task for each of the four field cmd box samples , we could assign a membership probability to each star in the cluster cmd .
this was done by counting the number of times a star remained unsubtracted in the four cleaned cluster cmds and by subsequently dividing this number by four .
thus , we distinguished field populations projected on to the cluster area , i.e. , those stars with a probability @xmath16 @xmath17 25% ; stars that could equally likely be associated with either the field or the object of interest ( @xmath16 = 50% ) ; and stars that are predominantly found in the cleaned cluster cmds ( @xmath18 75% ) rather than in the field - star cmds . statistically speaking ,
a certain amount of cleaning residuals is expected , which depends on the degree of variability of the stellar density , luminosity function and colour distribution of the field stars .
[ fig : fig4 ] illustrates the performance of the cleaning procedure in the field of vdbh72 , where we plotted three different ( @xmath19,@xmath20 ) cmds : a single field - star cmd ( top left - hand panel ) for a circular region with an area equal to that of the cluster ; that for the stars located within the cluster radius ( bottom left - hand panel ) ; and the cleaned cluster cmd ( bottom right - hand panel ) for stars with @xmath21 25% ( pink ) , @xmath22 50% ( light blue ) and @xmath21 75% ( dark blue ) . in the field - star cmd we overplotted the boxes generated by the cleaning procedure . a schematic finding chart with a circle of radius equal to the cluster radius is shown in the top right - hand panel . figures [ fig : fig5 ] to [ fig : fig11 ] show the whole set of cmds and colour - colour ( cc ) diagrams for the cluster sample that can be exploited from the present extensive multi - band photometry .
they include every magnitude and colour measurements of stars located within the respective cluster radii ( see table [ tab : table9 ] ) .
we have also incorporated to the figures the statistical photometric memberships obtained in sect .
3.2.1 by distinguishing stars with different colour symbols as in fig .
[ fig : fig4 ] . at first glance ,
the cleaned cluster cmds ( stars with @xmath18 75% ) resemble those moderately young to intermediate - age , projected on to star fields with different levels of crowdness .
@xcite introduced a new age - metallicity diagnostic diagram for the washington photometric system , @xmath23@xmath24 versus @xmath23@xmath10 - @xmath23@xmath24 , which has shown the ability of unambiguously providing age and metallicity estimates , simultaneously . @xmath23@xmath10 and @xmath23@xmath24 are the respective magnitude differences between the giant branch clump and the main sequence turnoff ( msto ) .
the new procedure allows to derive ages from 1 up to 13 gyr and metallicities [ fe / h ] from -2.0 up to + 0.5 dex , and is independent of the cluster reddening and distance modulus .
we used here that procedure to estimate the age and metallicity of three clusters ( vdbh1 , 31 and 118 ) whose cleaned cmds show a handful of red clump ( rc ) stars ( see figs .
[ fig : fig5 ] , [ fig : fig7 ] and [ fig : fig11 ] ) , besides their mstos .
we used the cleaned cluster cmds to measure @xmath10 and @xmath24 magnitudes at the msto and rc , then computed @xmath23@xmath10 and @xmath23@xmath24 and entered into the age - metallicity diagnostic diagram to estimate cluster ages and metallicities .
the resulting @xmath23@xmath24 and @xmath23@xmath10 - @xmath23@xmath24 values with their uncertainties are drawn in fig .
[ fig : fig12 ] , where we have traced iso - age lines and marked iso - abundance positions using colour - coded lines and filled circles , respectively . from this figure
we estimated by interpolation ages of [email protected] gyr , [email protected] gyr and [email protected] gyr for vdbh1 , 31 and 118 , respectively .
as for the mean metallicities , although rather more uncertain than the ages ( @xmath26[fe / h]= 0.25 dex ) , the clusters appear to be of solar or slightly subsolar metal content . for the remaining clusters in our sample ( vdbh10 , 72 , 87 and 92 ) whose cmds and cc diagrams resemble those of young clusters ( see figs . [
fig : fig6 ] , [ fig : fig8 ] , [ fig : fig9 ] and [ fig : fig10 ] ) we also adopted a solar metal content ( see last column of table [ tab : table9 ] ) .
note that , by considering the whole metallicity range of the milky way open clusters ( see , e.g. * ? ? ?
* ; * ? ? ?
* ) and by using the theoretical isochrones of @xcite , the differences at the zero age main sequence ( zams ) in @xmath20 and @xmath27 colours is smaller than @xmath3 0.08 and 0.04 mag , respectively .
this result implies that negligible differences between the zamss for the cluster metallicity and that of solar metal content would appear , keeping in mind the intrinsic spread of the stars in the @xmath19 vs @xmath20 and @xmath24 vs @xmath27 cmds . the availability of six cmds and three different cc diagrams covering wavelengths from the blue up to the near - infrarred allowed us to derive reliable ages , reddenings and distances for the studied clusters .
particularly noticeable in the @xmath19 vs @xmath28 cmd , but applicable to every cmd , the shape of the main sequence ( ms ) , its curvatures ( those less and more pronounced ) , the relative distance between the rc and the msto in magnitude and colour separately , among others , are features tightly related to the cluster age , regardless their reddenings and distances . for this reason , we started by selecting theoretical isochrones @xcite with the adopted cluster metallicities in order to chose those which best match the clusters features in the cmds . from our first choices , we derived the cluster reddenings by shifting those isochrones in the three cc diagrams following the reddening vectors until their bluest points coincided with the observed ones .
note that this requirement allowed us to use the @xmath29 vs @xmath30 cc diagram as well , even though the reddening vector runs almost parallell to the cluster sequence .
finally , the mean @xmath31 colour excesses were used to properly shift the chosen isochrones in the cmds in order to derive the cluster true distance modulii by shifting the isochrones along the magnitude axes .
we iterated this procedure whenever refinements in the cluster ages were necessary .
nevertheless , we found that isochrones bracketing the initial age choice by @xmath15 log(@xmath32 yr@xmath33 ) = @xmath250.10 represent the overall age uncertainties owing to the observed dispersion in the cluster cmds and cc diagrams .
although in some cases the age dispersion is smaller than @xmath15 log(@xmath32 yr@xmath33 ) = 0.10 , we prefer to keep the former value as an upper limit to our error budget . in order to enter the isochrones into the cmds and cc diagrams we used the following ratios : @xmath34/@xmath31 = 0.72 + 0.05@xmath11@xmath31 @xcite ; @xmath35/@xmath31 = 0.65 , @xmath36/@xmath31 = 1.25 , @xmath37/@xmath31 = 3.1 @xcite ; @xmath38/@xmath31 = 1.97 , @xmath39/@xmath31 = 0.692 , @xmath40/@xmath31 = 2.62 @xcite .
the adopted best matched isochrones are overplotted on figs .
[ fig : fig5 ] to [ fig : fig11 ] , while the resulting values with their errors for the cluster reddenings , true distance modulii and ages are listed in table [ tab : table9 ] .
the masses of the clusters in our sample were derived by summing the individual masses of stars with membership probabilities @xmath18 75% .
the latter were obtained by interpolation in the theoretical isochrones traced in figs .
[ fig : fig5 ] to [ fig : fig11 ] from the observed @xmath19 magnitude of each star , properly corrected by reddening and distance modulus .
we estimate the uncertainty in the mass to be @xmath41 @xmath3 0.2 dex . note
that this error comes from propagation of the @xmath19 magnitude errors in the mass distribution along the theoretical isochrones .
it does not reflect the deviation of the cluster mass computed from stars with @xmath18 75% from the actual total cluster mass . nevertheless , at first glance , the appearance of the cluster cmds and cc diagrams ( @xmath18 75% ) do not seem to significantly differ from those including any other observed stars placed along the adopted isochrones with @xmath42 75% , thought to be cluster stars . at the same time
, the computed cluster masses include some unavoidable interlopers , which mitigate the loss of some cluster stars . in the case of vdbh118 ,
the derived mass should be considered as a very lower limit , since its cmds barely reach the cluster msto . for the remaining clusters ,
our photometry reach well below the fainter ms cluster stars . using the resulting masses and the half - mass radii @xmath43 of table [ tab : table9 ] , we computed the half - mass relaxation times using the equation @xcite
: @xmath44 where @xmath45 is the cluster mass and @xmath46 is the mean mass of the cluster stars . the derived masses and relaxation times are listed in table [ tab : table9 ] . if we considered non - oberved stars with masses between 1 and 0.5 @xmath6 and the salpeter s mass function , the relaxation times would increase in @xmath3 10 per cent .
@xmath24 versus @xmath23@xmath10 - @xmath23@xmath24 diagram with iso - age lines and iso - metallicity locii .
metallicity and age labels are given in dex and gyr , respectively . ]
vdbh10 (= ruprecht35 ) and 31 (= ruprecht60 ) have age estimates of 400@xmath25100 myr derived from two - micron all - sky survey ( 2mass ) photometry @xcite , which clearly differs from our values ( see table [ tab : table9 ] ) . by inspecting their 2mass @xmath47 vs @xmath48 cmds ( figures 4 and 5 in @xcite ) , and considering the relationship @xmath49 vs @xmath50 for the cluster ages computed by @xcite and our cluster distances
, we found that the faintest reached @xmath47 magnitude ( @xmath9 15.5 mag ) corresponds to @xmath19 @xmath9 16.3 mag and 16.6 mag for vdbh10 and 31 , respectively .
this suggests that the used 2mass photometry is shallower than the present one . on the other hand
, we speculate with the possibility that the field - star cleaning procedure applied by @xcite have left significant residuals ( see * ? ? ?
* ; * ? ? ?
* ) ; thus leading them to derive ages which reflect the composite stellar population along the cluster line - of - sights .
curiously , both clusters have nearly similar angular galactic coordinates .
additionally , the 2mass cmds were extracted using sky areas much larger than those embracing the clusters , so that field - stars can quantitatively prevail over cluster stars . for these reasons ,
we are confidence of the present multi - band photometry and resulting cluster parameters . the distance modulus and age of vdbh10 (= ruprecht35 )
were previously estimated by @xcite and soon after used by @xcite in an updated study of the spiral structure of the outer galactic disc .
@xcite estimated a distance from the sun of 5.32 kpc and an age of 70 myr . while the cluster age is in excellent agreement with the present value ,
the cluster distance differs significantly .
this could be due either to a photometric zero point offset or to the use of field star contaminated cmd diagrams in the analysis carried out by @xcite .
both sources of error have been checked within our photometric data set . on the one hand , we used six different cmds involving @xmath19 and @xmath24 magnitudes , as well as @xmath28 , @xmath51 , @xmath29 , @xmath30 , @xmath52 and @xmath27 colours . on the other hand , we have had particular care in cleaning the cluster cmds from field star contamination ( sect .
as fas as we are aware , we are confident of the realibility of the present cluster distance .
vdbh31 (= ruprecht60 ) was more recently studied by @xcite using @xmath53 photometry .
the authors mentioned that their photometry does not cover the object completely and that they did not perform any cleaning of field stars in the cluster cmd .
they estimated an age of @xmath31.5 gyr , a metallicity of [ fe / h ] @xmath3 -0.5 dex , an @xmath31 colour excess of [email protected] mag and a distance from the sun of 4.5 kpc .
their age , reddening and distance are in very good agreement with our values . as for the remarkable low metal abundance ,
our metallicity sensitive washington @xmath10 photometry do not support such a low value .
@xcite could have included field stars in their cmd which , alongside the fewer cluster stars measured , could lead to such a low value . finally , @xcite studied vdbh31 (= ruprecht60 ) also from @xmath53 photometry .
they derived @xmath31 = [email protected] mag , a distance from the sun of [email protected] kpc and an age of 0.8 - 1.0 gyr . while the estimated reddening is in faily good agreeement with our value
, they derived a slightly younger age which in turn led to estimate a larger distance .
we are confident of our age value which comes from a cmd analysis as well as from the washington age - metallicity diagnostic diagram analysis .
we computed galactic coordinates using the derived cluster heliocentric distances , their angular galactic coordinates and a galactocentric distance of the sun of r@xmath54 = 8.3 kpc ( * ? ? ? * and references therein ) .
the resulting spatial distribution is depicted in fig .
[ fig : fig13 ] , where we added for comparison purposes the 2167 open clusters catalogued by ( * ? ? ?
* as of january 2016 ) and the schematic positions of the spiral arms @xcite .
the studied clusters are mostly located between the carina and perseus arms , with the sole exception of vdbh118 , which lies between the carina and crux spiral arms and is one of the most distant open clusters from the sun in that direction as well . in general terms , the seven vdbh clusters are distributed outside the circle around the sun ( d @xmath3 2.0 kpc ) where the catalogued clusters are mostly concentrated . the structural parameters estimated from the stellar density radial profiles in combination with the fundamental properties derived from the analysis of cmds and cc diagrams allow us to investigate the internal dynamical state of the clusters . here
we consider the effect of the internal dynamical evolution ( two - body relaxation , mass segregation , etc ) . as for the galactic tidal field , bearing in mind the galactocentric distances of the studied clusters ( r@xmath55 @xmath9 [email protected] kpc , with a mean value of 9@xmath251 kpc ) and their ages ( see table [ tab : table9 ] ) , the differences in the potential will lead to @xmath3 10 per cent variations in the half mass radius ( * ? ?
* see , e.g. , their figures 1 ) .
the age/@xmath5 ratio is a good indicator of the internal dynamical evolution , since it gives the number of times the characteristic time - scale to reach some level of energy equipartition @xcite has been surpassed .
star clusters with large age/@xmath5 ratios have reached a higher degree of relaxation and hence are dynamically more evolved . as fig .
[ fig : fig14 ] shows , our clusters appear to cover a wide range of age/@xmath5 ratios , from @xmath3 20 up to 320 , all of them suggesting that the clusters have had enough time to evolve dynamically . in the figure we included in grey colour 236 open clusters analysed by @xcite , who derived from them homogeneous scales of radii and masses .
they derived core and tidal radii for their cluster sample , from which we calculated the half - mass radii and , with their clusters masses and eq .
12 , relaxation times , by assuming that the cluster stellar density profiles can be indistinguishably reproduced by king and plummer models .
their cluster sample are mostly distributed inside a circle of @xmath3 1 kpc from the sun . as can be seen ,
our clusters cover the most evolved limit of the age/@xmath5 distribution ( right - hand panels ) . ) , half - mass ( @xmath43 ) and tidal ( @xmath56 ) radii , concentration parameter ( @xmath7 ) , mass , age and relaxation time ( @xmath5 ) . ]
grey dots correspond to 236 star clusters with homogeneous estimations of masses and radii derived by @xcite .
[ fig : fig14 ] since dynamical evolution implies the loss of stars ( mass loss ) , we expect some trend of the present - day cluster mass with the age/@xmath5 ratio .
this is confirmed in the top - right panel of fig .
[ fig : fig14 ] , where the larger the present - day mass the less the dynamical evolution of a cluster in the solar neighbourhood , with a noticeable scatter .
the studied clusters appear to have relatively large masses for their particular internal dynamical states .
curiously , selection against poor and old clusters could suggest the beggining of cluster dissolution , with some exceptions .
@xcite presented a unified picture for the evolution of star clusters on the two - body relaxation timescale from direct n - body simulations of star clusters in a tidal field .
their treatment of the stellar evolution is based on the approximation that most of the relevant stellar evolution occurs on a timescale shorter than a relaxation time , when the most massive stars lose a significant fraction of mass and consequently contribute to a global expansion of the system . later in the life of a star cluster , two - body relaxation tends to erase the memory of the initial density profile and concentration .
they found that the structure of the system , as measured by the core to half mass radius ratio , the concentration parameter @xmath7= log(@xmath57 ) , among others , evolve toward a universal state , which is set by the efficiency of heating on the visible population of stars induced by dynamical interactions in the core of the system . in the bottom panels of fig .
[ fig : fig14 ] we plotted the dependence of the concentration parameter @xmath7 with the cluster mass and the age/@xmath5 ratio , respectively .
they show that our dynamically evolved clusters are within those with relatively high @xmath7 values , and that star clusters tend to initially start their dynamical evolution with relatively small concentration parameters .
likewise , star clusters in an advanced dynamical state can also have relatively lower @xmath7 values due to their smaller masses . according to ( * ? ? ?
* see , e.g. , their figure 33.2 ) a star cluster dynamically evolving with its tidal radius filled , moves in the @xmath58 vs @xmath59 plane parallel to the @xmath58 axis ( @xmath59 @xmath3 0.21 ) toward low values due to violent relaxation in the cluster core region followed by two - body relaxation , mass segregation , and finally core - collapse .
top - left panel in fig .
[ fig : fig14 ] shows that the studied clusters have @xmath59 ratios in fairly agreement , within the uncertainties , with that of tidally filled clusters , and @xmath58 ratios suggesting mass segregation in their core regions .
we finally built the cluster mass functions ( mfs ) employing the masses of stars with photometric memberships @xmath18 75% .
the resulting mfs are shown in fig .
[ fig : fig15 ] where the errorbars come from applying poisson statistics .
we did not include vdbh118 because of incompleteness in our photometry for its ms . for comparison porpuses we superimposed the relationship given by (
* slope = -2.35 ) for the stars in the solar neighbourhood .
as can be seen , despite their advanced state of dynamical evolution , vdbh1 , 72 and 87 still keep their mfs close to that of salpeter s law , while the remaining clusters ( vdbh10 , 31 and 92 ) show mfs which depart from it for smaller masses toward a relation with a smaller slope ( see , e.g. * ? ? ?
* ; * ? ? ?
their total masses , structural parameters and age/@xmath5 ratios do not suggest any hint to explain such differences , and prevent us to draw any conclusion from such a small cluster sample .
we present results of seven star clusters identified by @xcite , namely , vdbh1 , 10 , 31 , 72 , 87 , 92 , and 118 , for which particular attention was not given until now . the clusters were observed through the johnson @xmath0 , kron - cousins @xmath1 and washington @xmath10 filters ; four of them ( vdbh72 , 87 , 92 and 118 ) are photometrically study for the first time .
the multi - band photometric data sets were used to trace the cluster stellar density radial profiles and to build cmds and cc diagrams , from which we estimate their structural parameters and fundamental astrophysical properties .
their radial profiles were built from star counts carried out throughout the observed fields using the final photometric catalogues .
we derived the cluster radii from a careful placement of the background levels and fitted king and plummer models to derive cluster core , half - mass and tidal radii .
we then applied a subtraction procedure developed by @xcite to statistically clean the star cluster cmds and cc diagrams from field star contamination in order to disentangle star cluster features from those belonging to their surrounding fields .
the employed technique makes use of variable cells in order to reproduce the field cmd as closely as possible . the availability of three cc diagrams and six cmds covering wavelengths from the blue up to the near - infrarred allowed us to derive reliable ages , reddenings and distances for the studied clusters .
we exploited such a wealth in combination with a new age - metallicity diagnostic diagram for the washington system and theoretical isochrones computed by @xcite to find out that the clusters in our sample cover a wide age range , from @xmath3 60 myr up to 2.8 gyr , are of relatively small size ( @xmath3 1 @xmath4 6 pc ) and are placed at distances from the sun which vary between 1.8 and 6.3 kpc , respectively .
they are located between the carina and perseus spiral arms , with the exception of vdbh118 , which is closer to the galactic centre than the carina arm .
it also belongs to the sample of the farthest known clusters placed between the carina and crux spiral arms .
we estimated lower limits for the cluster present - day masses as well as half - mass relataxion times .
their resulting values suggest that the studied clusters have advanced states of their internal dynamical evolution , i.e. , their ages are many times the relaxation times ( @xmath3 20 @xmath4 320 ) .
when combined with the obtained structural parameters , we found that the clusters are possible in the phase typical of those tidally filled with mass segregation in their core regions .
we compared the cluster masses , concentration parameters and age/@xmath5 ratios with those for 236 clusters located in the solar neighbourhood .
the seven vdbh clusters are within more massive ( @xmath3 80 @xmath4 380@xmath6 ) , with higher @xmath7 values , and dynamically evolved ones .
there is also a broad correlation between the cluster massses and the dynamical state , in the sense that the more the advanced the internal dynamical evolution , the less the cluster mass .
the apparent decrease of clusters with masses smaller than @xmath3 200@xmath6 and ages @xmath3 100 times older than their relaxation times could suggest the beggining of cluster dissolution , with some exceptions .
finally , we found that the mfs of vdbh1 , 72 and 87 follow approximately the @xcite s relationship , while vdbh10 , 31 and 92 seem to show shallower mfs for the lower mass regime .
we thank the anonymous referee whose thorough comments and suggestions allowed us to improve the manuscript . | we present results from johnson @xmath0 , kron - cousins @xmath1 and washington @xmath2 photometries for seven van den bergh - hagen ( vdbh ) open clusters , namely , vdbh1 , 10 , 31 , 72 , 87 , 92 , and 118 . the high - quality , multi - band photometric data sets were used to trace the cluster stellar density radial profiles and to build colour - magnitude diagrams ( cmds ) and colour - colour ( cc ) diagrams from which we estimated their structural parameters and fundamental astrophysical properties .
the clusters in our sample cover a wide age range , from @xmath3 60 myr up to 2.8 gyr , are of relatively small size ( @xmath3 1 @xmath4 6 pc ) and are placed at distances from the sun which vary between 1.8 and 6.3 kpc , respectively .
we also estimated lower limits for the cluster present - day masses as well as half - mass relaxation times ( @xmath5 ) .
the resulting values in combination with the structural parameter values suggest that the studied clusters are in advanced stages of their internal dynamical evolution ( age/@xmath5 @xmath3 20 @xmath4 320 ) , possibly in the typical phase of those tidally filled with mass segregation in their core regions .
compared to open clusters in the solar neighbourhood , the seven vdbh clusters are within more massive ( @xmath3 80 @xmath4 380@xmath6 ) , with higher concentration parameter values ( @xmath7 @xmath3 [email protected] ) and dynamically evolved ones .
[ firstpage ] techniques : photometric galaxy : open clusters and associations : general . |
the observed accelerated expansion of the universe @xcite is most easily explained by a constant vacuum energy density , the characteristic feature of de sitter space . this has led to an intense search for de sitter vacua in supergravity and superstring theories , which represent attractive extensions of the standard model of particle physics .
although much progress has been made , challenging questions still remain @xcite . a particularly attractive proposal by kachru , kallosh , linde and trivedi @xcite is based on quantized fluxes in type iib string theory and additional nonperturbative effects stabilizing the overall volume modulus .
these features are captured in the kklt superpotential that yields an anti - de sitter minimum . to achieve the required ` uplift ' to de sitter space
additional ingredients are needed such as anti - d3 branes @xcite , @xmath1-terms induced by magnetic flux @xcite , a gauged r - symmetry @xcite or t - branes @xcite .
alternatively , additional matter fields have to be added allowing for an f - term uplift ( see e.g. @xcite ) . also extensions of the kklt proposal , such as the large volume scenario @xcite or
khler uplift @xcite need additional ingredients to achieve an uplift to de sitter vacua .
conditions for de sitter vacua in four - dimensional supergravity theories derived from string theory compactifications have previously been analyzed in @xcite .
in particular , constraints on the khler potential and , in case of gauged shift symmetries , on the associated fayet - iliopoulos ( fi ) terms have been derived @xcite . for
a related discussion of de sitter and minkowski vacua in string theory see , for instance , @xcite . in this letter
, we study de sitter vacua in six - dimensional supergravity models , compactified on orbifolds with flux .
such models have been suggested as an intermediate step in string compactifications to four dimensions @xcite .
our results represent an application of previous work @xcite .
the crucial ingredients are the flux and the green - schwarz term of an anomalous local @xmath0 symmetry .
they lead to two fayet - iliopoulos terms of opposite sign , and they modify the gauge kinetic function . together with a nonperturbative kklt - type superpotential one naturally obtains de sitter vacua , without any additional degrees of freedom below the flux induced mass scale . in the following
we derive relations between the superpotential parameters and de sitter vacua in moduli space and give an explicit example .
in this work we consider a simple @xmath2 supergravity model that has been derived as the effective four - dimensional lagrangian for flux compactifications of six - dimensional supergravity @xcite .
it contains three chiral multiplets , @xmath3 , @xmath4 , @xmath5 and the real vector multiplet @xmath6 of a @xmath0 gauge symmetry .
the khler potential is given by @xmath7 where @xmath8 are purely imaginary and constant killing vectors parameterizing a gauged shift symmetry .
note that the khler potential is of no - scale type , i.e. @xmath9 , where @xmath10 ( @xmath11 ) denotes the derivative with respect to the @xmath10-th chiral ( @xmath11-th anti - chiral ) multiplet and @xmath12 is the inverse khler metric . to completely define the gauge sector one further has to specify the gauge kinetic function @xmath13 , whose real part @xmath14 corresponds to the effective gauge coupling . in the following
it will be crucial that , in addition to the classical term linear in @xmath3 , the gauge kinetic function contains a second part linear in @xmath4 , @xmath15 a @xmath4-dependence is known to arise due to quantum corrections @xcite . following the convention @xmath16 , @xmath17 , one has @xmath18 .
finally , we have to specify the superpotential .
it has to be gauge invariant and can therefore only depend on the linear combination @xmath19 where in a first step , we ignore the shape modulus @xmath5 . for reasons discussed above
we consider a kklt - type potential , @xmath20 where , w.l.o.g .
we choose @xmath21 , @xmath22 and @xmath23 to be real .
the third modulus @xmath5 can be stabilized by inclusion of a term @xmath24 in the superpotential
. however , the resulting equations are considerably more involved than in the two moduli case .
they will be given in @xcite . in the following
it will be important that the parameter @xmath25 in the gauge kinetic function is negative while the classical contribution @xmath26 is positive .
furthermore , the two killing vectors have to be of opposite sign
. as we shall see in the next section , both conditions are satisfied in flux compactifications of 6d supergravity . the line along which @xmath27 divides the moduli space into a physical ( @xmath28 ) and an unphysical region ( @xmath29 ) .
the physical region thus satisfies the condition @xmath30 the scalar potential is a sum of @xmath31- and @xmath1-term contributions , @xmath32 where @xmath33 is the khler covariant derivative of the superpotential and @xmath34 is a gauge coupling , resulting in @xmath35 as the effective gauge coupling .
the @xmath1-term is given in terms of the killing vectors , @xmath36 due to their opposite sign @xmath37 defines a second line in @xmath38 space given by @xmath39 if this line is not part of the physical region in moduli space , i.e.@xmath40 , the @xmath1-term potential is positive definite .
this is the situation we want to study in the following .
solutions of the equations of motion leading to minkowski or de sitter vacua have to fulfill , @xmath41 which yield three relations between the @xmath31-term and @xmath1-term contributions to the potential .
it is convenient to use the linear combinations @xmath42 instead of @xmath43 .
derivatives of the superpotential then produce the factors @xmath44 and @xmath45 , where @xmath46 in terms of @xmath1 and @xmath47 the scalar potential can be written in the compact form @xmath48 with @xmath49 note that the term @xmath50 in the scalar potential has disappeared due to the no - scale structure of the khler potential .
choosing opposite signs for @xmath22 and @xmath23 , the axion @xmath51 is stabilized at zero , and in the following we set the cosine term to one .
the orthogonal axion gives a mass to the @xmath0 vector field via the stueckelberg mechanism @xcite . in order to find minima
we invert the problem and solve for the superpotential parameters in terms of @xmath52 and @xmath53 . in this way we can see which superpotential parameters lead to minima in the @xmath38 plane for realistic parameter ranges .
the conditions for extrema yield three relations between the superpotential parameters and the position @xmath38 of the extrema in moduli space .
a straightforward calculation leads to @xmath54 where we have introduced the ratio @xmath55 .
the quantities on the r.h.s . of these equations , @xmath14 , @xmath47 and @xmath56 , only depend on the values of the moduli fields @xmath52 and @xmath53 , whereas @xmath57 , @xmath58 and @xmath59 are related to the superpotential by the definitions and .
eqs . can now be used to identify consistency conditions for the existence of de sitter vacua and to construct explicit solutions .
a similar discussion for pure @xmath31-term breaking has recently been carried out in @xcite . from eq .
we see that @xmath57 has to be positive , and by definition @xmath60 .
furthermore , @xmath61 is needed for a cancellation of the positive contributions in the potential , thus allowing for a vanishing or small cosmological constant . from eqs .
one reads off that these conditions can be satisfied if @xmath62 .
the reason is that the terms proportional to @xmath14 , which would give the unwanted sign , can be made small due to the opposite sign of the two contributions @xmath63 and @xmath64 .
given a solution for a chosen extremum at @xmath38 , one then has to examine whether the obtained values for @xmath14 , @xmath59 , @xmath22 and @xmath23 are physically meaningful .
the de sitter solutions obtained this way correspond to metastable vacua since the familiar runaway solution at infinity has zero energy density . due to
the similarity with the kklt model one expects no further vacua . we have checked this numerically for various sets of superpotential parameters and we have indeed found no other minima .
let us finally emphasize again that the obtained minkowski and de sitter minima are a consequence of the opposite signs of the moduli contributions to both the gauge kinetic function and the @xmath1-term in combination with the nonperturbative superpotential
. in the de sitter vacua constructed this way supersymmetry is broken by a @xmath1-term since , as discussed above , @xmath65 in the physical region of the moduli space .
vanishing f - terms , i.e. @xmath66 , would imply @xmath67 , which is not possible for @xmath65 and a kklt - type superpotential .
hence , as expected , supersymmetry is also broken by @xmath31-terms and the goldstino is a mixture of the gaugino and modulini , with @xmath68 .
it is well known that no - scale models require carefully chosen superpotentials in order to allow for metastable de sitter vacua @xcite . in the above analysis simple solutions
are found due to the gauged shift symmetry , a possibility which has already been discussed in @xcite .
in @xcite we derive the low - energy effective action for a six - dimensional supergravity model with a @xmath0 gauge field and a charged bulk matter field , compactified on an orbifold @xmath69 with magnetic flux .
the antisymmetric tensor field couples to the @xmath0 gauge field via a chern - simons term .
the four - dimensional effective action involves the khler potential with a gauged shift symmetry in @xmath3 and @xmath4 .
the killing vectors are given by @xmath70 where @xmath71 is the quantized flux , @xmath72 , @xmath73 , and @xmath74 is a length scale in planck units .
@xmath75 is a consequence of the green - schwarz term which is needed to cancel bulk and fixed point anomalies .
it represents a quantum correction which , to our knowledge , has been neglected in previous analyses of moduli stabilization .
note that the two killing vectors contribute to the @xmath1-term with opposite signs : @xmath76 , @xmath77 .
the orbifold compactification with @xmath78 flux quanta generates @xmath79 massless weyl fermions , as discussed in @xcite .
their scalar superpartners @xmath80 obtain masses from the @xmath1-term potential @xmath81 with @xmath1 given by eq . .
the stabilization of the charged matter fields @xmath80 at the origin requires @xmath82 , as assumed in the previous section .
note that all matter fields @xmath80 have the same @xmath0 charge since they originate from the same bulk field .
the size of @xmath1 determines the mass splitting within matter multiplets and therefore the scale of supersymmetry breaking @xcite .
furthermore , we want to point out that the quantity @xmath47 introduced above is always positive , @xmath83 , and @xmath84 .
as discussed in the previous section , the existence of de sitter minima crucially depends on the gauge kinetic function for which one obtains @xcite @xmath85 at the classical level , the gauge kinetic function only depends on @xmath3 .
the contribution proportional to @xmath4 is a direct consequence of the green - schwarz term , and it is remarkable that the sign of @xmath25 is indeed negative . note that for @xmath86 , @xmath75 , @xmath26 and @xmath25 as given in eqs . and , @xmath28 always implies @xmath65 for @xmath87 .
hence , for nonzero flux , the matter fields are always stabilized at the origin . in the expression for the scalar potential only the term @xmath88 can become negative . for the 6d model considered here one has @xmath89 ( see eqs . , ) .
hence , @xmath58 is bounded from below and from above , and @xmath88 can approach @xmath90 only for @xmath91 ( see eq . ) .
however , in this limit the positive third term in eq .
is more singular . the scalar potential is therefore bounded from below .
we are now ready to apply the general analysis of the previous section to our example . for simplicity , we set @xmath92 in eqs . and look for minkowski vacua . obviously , there exist very similar solutions for de sitter vacua with @xmath93 .
having gut scale extra dimensions and three generations in mind @xcite , we choose @xmath94 , @xmath95 and @xmath96 . using eqs . and writing @xmath97 , we find for a minkowski minimum at @xmath98 typical superpotential parameters , @xmath99 for the effective gauge coupling we obtain @xmath100 , and the size of the compact dimension is @xmath101 , implying gut scale supersymmetry breaking .
note that also larger extra dimensions can be obtained without fine tuning for larger values of @xmath14 , i.e. for smaller gauge couplings and smaller superpotential parameters @xmath22 , @xmath23 .
the masses of all moduli are of order @xmath102 . in fig . [ minimum ] the scalar potential is shown close to the minkowski minimum .
it diverges as @xmath52 and @xmath53 approach the critical line @xmath103 .
vanishes along the red dashed line and @xmath1 vanishes along the blue , dashed - dotted line .
the minimum is at @xmath104 as indicated by the orange star . ]
we have presented a new class of metastable de sitter vacua .
they arise in no - scale supergravity models with a gauged shift symmetry involving two moduli fields , @xmath3 and @xmath4 .
this leads to two fayet - iliopoulos terms with different moduli dependence .
also the gauge kinetic function has contributions linear in @xmath3 and @xmath4 .
the superpotential of kklt - type depends on the gauge invariant linear combination of @xmath3 and @xmath4 .
as we have shown , de sitter or minkowski vacua are naturally obtained if the @xmath4-dependent contribution to the gauge kinetic function is negative .
it is remarkable that the low - energy effective action of six - dimensional supergravity with flux , compactified on an orbifold , fulfills the sufficient conditions for de sitter vacua .
the @xmath4-dependent fi term is generated by magnetic flux , whereas the @xmath3-dependent fi term and the @xmath4-dependent contribution to the gauge kinetic function are due to the green - schwarz term needed to cancel the @xmath0 gauge anomaly .
the green schwarz - term leads to the negative sign of the @xmath4-dependent part in the gauge kinetic function .
flux and anomaly cancellation by the green - schwarz mechanism are generic ingredients of string compactifications .
we therefore believe that our discussion of de sitter vacua is relevant far beyond the six - dimensional supergravity example discussed in this letter
. it will be interesting to study applications for the heterotic string as well as for type iib string theory and f - theory .
furthermore , it is intriguing that typical values for the superpotential parameters and the gauge coupling lead to gut scale extra dimensions and a related large supersymmetry breaking scale .
this is a challenge for the electroweak hierarchy problem , analogous to the cosmological constant problem which is encoded in the fine tuning of the superpotential parameters .
we thank michele cicoli , emilian dudas , stefan groot nibbelink , zygmunt lalak , jan louis , hans - peter nilles , alexander westphal and clemens wieck for valuable discussions .
this work was supported by the german science foundation ( dfg ) within the collaborative research center ( sfb ) 676 `` particles , strings and the early universe '' .
m.d . acknowledges support from the studienstiftung des deutschen volkes .
s. perlmutter _ et al .
_ [ supernova cosmology project collaboration ] , `` measurements of omega and lambda from 42 high redshift supernovae , '' _ astrophys .
j. _ * 517 * ( 1999 ) 565 http://arxiv.org/abs/astro-ph/9812133[[astro-ph/9812133 ] ] .
+ a. g. riess _ et al .
_ [ supernova search team collaboration ] , `` observational evidence from supernovae for an accelerating universe and a cosmological constant , '' _ astron .
j. _ * 116 * ( 1998 ) 1009 http://arxiv.org/abs/astro-ph/9805201[[astro-ph/9805201 ] ] .
m. r. douglas and s. kachru , `` flux compactification , '' _ rev .
* 79 * ( 2007 ) 733 http://arxiv.org/abs/hep-th/0610102[[hep-th/0610102 ] ] .
s. kachru , r. kallosh , a. d. linde and s. p. trivedi , `` de sitter vacua in string theory , '' _ phys .
d _ * 68 * , 046005 ( 2003 ) http://arxiv.org/abs/hep-th/0301240[[hep-th/0301240 ] ] .
k. choi , a. falkowski , h. p. nilles and m. olechowski , `` soft supersymmetry breaking in kklt flux compactification , '' _ nucl .
b _ * 718 * ( 2005 ) 113 http://www.arxiv.org/abs/hep-th/0503216[[hep-th/0503216 ] ] .
v. balasubramanian , p. berglund , j. p. conlon and f. quevedo , `` systematics of moduli stabilisation in calabi - yau flux compactifications , '' _ jhep _ * 0503 * ( 2005 ) 007 http://www.arxiv.org/abs/hep-th/0502058[[hep-th/0502058 ] ] .
v. balasubramanian and p. berglund , `` stringy corrections to kahler potentials , susy breaking , and the cosmological constant problem , '' _ jhep _ * 0411 * ( 2004 ) 085 http://www.arxiv.org/abs/hep-th/0408054[[hep-th/0408054 ] ] .
l. covi , m. gomez - reino , c. gross , j. louis , g. a. palma and c. a. scrucca , `` de sitter vacua in no - scale supergravities and calabi - yau string models , '' _ jhep _ * 0806 * ( 2008 ) 057 http://www.arxiv.org/abs/0804.1073[[0804.1073 ] ] .
l. covi , m. gomez - reino , c. gross , g. a. palma and c. a. scrucca , `` constructing de sitter vacua in no - scale string models without uplifting , '' _ jhep _ * 0903 * , 146 ( 2009 ) http://www.arxiv.org/abs/0812.3864[[0812.3864 ] ] .
r. blumenhagen , d. lust and t. r. taylor , `` moduli stabilization in chiral type iib orientifold models with fluxes , '' _ nucl .
b _ * 663 * ( 2003 ) 319 http://www.arxiv.org/abs/hep-th/0303016[[hep-th/0303016 ] ] .
g. villadoro and f. zwirner , `` d terms from d - branes , gauge invariance and moduli stabilization in flux compactifications , '' _ jhep _ * 0603 * ( 2006 ) 087 http://www.arxiv.org/abs/hep-th/0602120[[hep-th/0602120 ] ] .
w. buchmuller , m. dierigl , f. ruehle and j. schweizer , `` chiral fermions and anomaly cancellation on orbifolds with wilson lines and flux , '' _ phys .
d _ * 92 * ( 2015 ) 10 , 105031 http://www.arxiv.org/abs/1506.05771[[1506.05771 ] ] . | we find a new class of metastable de sitter solutions in compactifications of six - dimensional supergravity motivated by type iib or heterotic string vacua .
two fayet - iliopoulos terms of a local @xmath0 symmetry are generated by magnetic flux and by the green - schwarz term canceling the gauge anomalies , respectively .
the interplay between the induced d - term , the moduli dependence of the effective gauge coupling , and a nonperturbative superpotential stabilizes the moduli and determines the size of the extra dimensions .
desy-16 - 030 + .8 cm + + + |
the associated production of vector boson ( v ) and jets is a major background to single and pair - produced top quarks , as @xmath3 and @xmath4 heavy flavor jets share much of the final - state signature of top events .
the contribution of @xmath3 + heavy flavor quark ( @xmath5 ) for instance leads to one of the main systematic uncertainties to the identification and measurements of single top at the tevatron , but also searches for the higgs boson in the @xmath6 channel . in recent measurements of top quarks at cdf @xcite , the estimates of the @xmath7 , @xmath8 and @xmath9 backgrounds carry a precision of only @xmath10% . calculations of v+jets at next - to - leading - order ( nlo ) in qcd are available mostly for lower jet multiplicities .
moreover , a number of monte carlo ( mc ) simulation tools are also available : these include simulations based on leading - order ( lo ) matrix elements followed by parton shower modeling ( e.g. pythia , herwig ) , or higher - order matrix elements for the hard - scattering process ( e.g. alpgen , sherpa ) , followed by modeling of the parton shower .
the production of v+jet , in addition to be crucial for the understanding of top quark candidate samples , gives important information on parton distribution functions ( pdfs ) inside the proton and allows for stringent qcd and standard model tests .
we present preliminary measurements of @xmath0+jets , @xmath1 and @xmath2 and review the latest measurements of v + inclusive and heavy flavor jets at the tevatron .
the production cross section of @xmath11 jets in @xmath12 collisions at @xmath13 tev has been measured @xcite using a data sample corresponding to an integrated luminosity of @xmath14320 pb@xmath15 , collected with the cdf ii detector .
the cdf ii detector is described in detail elsewhere @xcite .
@xmath3 bosons are identified in their electron decay channel , and electrons from the @xmath3 decay are required to have transverse energy @xmath16 gev and pseudo - rapidity @xmath17 @xcite .
jets are reconstructed using a cone algorithm @xcite with radius 0.4 and are required to have @xmath18 gev and @xmath19 , where @xmath20 is calibrated such that it measures on average the summed @xmath21 of the particles within the jet cone that are the result of the @xmath22 interaction .
the measurement s results include the determination of the total cross section : @xmath23 and the differential cross section @xmath24 for ( @xmath25 ) .
the data is seen to agree to better than @xmath26% with nlo calculations done with the mcfm @xcite program ( @xmath27 ) .
the analysis also presents comparisons of the data with mc modeling based on lo perturbative qcd ( pqcd ) calculation plus parton shower and hadronization ( madgraph+ @xcite and alpgen+herwig @xcite ) .
the production cross section measured in the data ( @xmath28 ) , as a function of @xmath29-jets , results higher than that predicted by these models by about 40% , although the ratio @xmath30 is well reproduced .
the production of @xmath0+jets at the tevatron has been measured in the @xmath31 decay channel @xcite by the cdf collaboration and in @xmath32 and @xmath33 decay channels @xcite by the d0 collaboration , using @xmath142.5 fb@xmath15 and @xmath141.0 fb@xmath15 of collision data respectively .
we present here a preliminary measurement in the @xmath32 decay channel using @xmath142.4 fb@xmath15 of data collected with the cdf ii detector .
jets are reconstructed using the midpoint algorithm @xcite with cone radius @xmath34 and are required to have transverse momentum @xmath35
gev/@xmath36 , rapidity @xmath37 @xcite and a minimum distance between the jets and the muons @xmath38 .
the value of @xmath39 is corrected back to the particles that enter the jet cone , and includes also corrections for the multiple @xmath12 interactions per crossing at high instantaneous luminosity . in fig
[ fig : zjets ] the data is compared to lo and nlo predictions computed using the mcfm program . fig .
[ fig : zjets](a ) shows the cross section as a function of the number ( @xmath40 ) of jets in the event , while fig .
[ fig : zjets](b ) shows the inclusive @xmath39 differential cross section for @xmath41 jet .
the theoretical predictions include parton - to - particle level correction factors that account for underlying event and fragmentation processes .
the measurements are well described by nlo pqcd predictions ( for @xmath42 only lo predictions are available ) , and the @xmath43 differential cross section measurement suggests a common lo - to - nlo multiplicative factor of @xmath44 , approximately independent of @xmath43 .
[ cols="^,^ " , ] the extensive @xmath45-boson+jets sample collected at the tevatron has also allowed detailed comparisons of the data with lo and nlo predictions in a variety of phase space regions . in the most recent of these analyses @xcite , differential
@xmath0+jet+x cross sections are measured , binned in the azimuthal angle between the @xmath0 and leading jet , @xmath46 , absolute value of the rapidity difference between the @xmath0 and leading jet , @xmath47 , and the absolute value of the average rapidity of the @xmath0 and leading jet , @xmath48 .
the @xmath49 distribution is sensitive to qcd radiation .
the rapidity variables , @xmath50 and @xmath51 have primary contributions from the relative momenta of the incoming partons in the hard scatter but are modified by any additional qcd radiation .
[ fig : d0_mm_2 ] shows the normalized differential cross section as a function of @xmath52 .
the data corresponds to an integrated luminosity of [email protected] fb@xmath15 collected with the d0 detector .
a full description of the d0 detector is available elsewhere @xcite , jets are reconstructed using an iterative midpoint algorithm @xcite with cone radius 0.5 and their transverse momentum is corrected for the calorimeter response , instrumental noise and pile - up from multiple @xmath12 interactions and bunch crossings .
the analysis selects events with @xmath54 gev/@xmath36 and @xmath55 .
the data are compared with lo and nlo pqcd calculations obtained using mcfm , after correcting the predictions for hadronization and underlying event using pythia .
the data are also compared in fig .
[ fig : d0_mm_2](c)(d ) with matrix element calculations with matched parton shower implementations such as sherpa and alpgen .
the nlo pqcd calculation provides a good modeling of the data and is a significant improvement in both shape and uncertainty over lo .
production with @xmath56 gev/@xmath36 , in bins of @xmath52 @xcite .
all ratios in ( b ) , ( c ) and ( d ) are shown relative to the prediction from sherpa.,width=472 ] the production of @xmath58 at a center - of - mass energy @xmath13 tev has been measured @xcite using @xmath141.0 fb@xmath15 of data collected by the d0 detector .
photons are reconstructed in the central rapidity region @xmath59 with transverse momenta in the range @xmath60 gev/@xmath36 while jets are reconstructed in either the central @xmath61 or forward @xmath62 rapidity intervals with @xmath63 gev/@xmath36 .
different angular configurations between the photon and the jets can be used to simultaneously test the dynamics of qcd hard - scattering subprocesses in different regions of parton momentum fraction @xmath64 and large hard - scattering scales @xmath65 .
the data are compared to nlo qcd predictions obtained using jetphox @xcite with cteq6.5 m pdf @xcite and bfg fragmentation functions of partons to photons @xcite .
while the nlo predictions are seen to give a good overall description of the @xmath66 differential cross section , discrepancies of order 3040% appear when attempting the describe simultaneously the data in four separate kinematic regions defined by the rapidity of the photon and that of the leading jet ( central jets with @xmath67 and @xmath68 , and forward jets with @xmath67 and @xmath68 ) .
the measurements are compared to the corresponding theoretical predictions , which are seen to overestimate slightly the production measured for photons with @xmath69 gev/@xmath36 .
the measurement of the production of @xmath71-jets provides and important test of qcd , and the understanding and modeling of this process is crucial e.g. in searches for standard model higgs boson via @xmath6 and measurements of single top . theoretical predictions for @xmath71-jets production are based on nlo calculations in the 1-jet and 2-jet multiplicities @xcite and show an enhancement over lo of up to a factor of two .
the cross section for jets from @xmath72 quarks produced with a @xmath3 boson has been measured @xcite in @xmath12 collision data at @xmath13 tev using @xmath141.9 fb@xmath15 of integrated luminosity recorded with the cdf ii detector .
the measurement is based on the selection of events consistent with the electronic or muonic decay of a @xmath3 boson and containing one or two jets .
the identification of heavy flavor - originated jets is based on the detection of a secondary vertex ( vertex @xmath72-tagging ) , reconstructed from the charged particles within each jet , and well displaced from the primary @xmath12 interaction vertex location .
the invariant mass of the charged particles forming the secondary decay vertex is used as a discriminant to separate @xmath72-jets from charm and certain light flavor hadrons which are accidentally @xmath72-tagged .
the measurement s result : @xmath73 is in agreement , although about two standard deviations higher , with the nlo calculation for the corresponding phase space of [email protected] pb .
modelings based on lo from pythia and summed fixed - order from alpgen of the jet @xmath21 and jet @xmath74 differential cross sections are seen to agree with the data , although the overall production cross section predicted by these programs is lower than the measured value by factors of 2.53.5 .
calculations of the production of @xmath11charm quark at the tevatron carry a precision of only @xmath1430% , due to uncertainties in the strange - quark density in the proton and on the renormalization and factorization scales in the nlo calculation @xcite . therefore , experimental constraints are crucial to help in searches and measurement of signatures with the @xmath3+heavy flavor final state .
the experimental determination of @xmath3+charm is based on the correlation between the charge of the @xmath3 boson and the charge of the lepton ( @xmath75 or @xmath76 ) from the semileptonic decay of the charm hadron .
charge conservation in the process @xmath77 ( @xmath78 ) allows as final states only the pairings @xmath79 and @xmath80 and therefore the charge of the lepton from the semileptonic decay of the @xmath36 and the charge of the lepton from the @xmath3 decay are always of opposite sign . by subtracting the number of same - sign combinations from the opposite - sign sample it is possible to highlight a @xmath3+charm signal from which the production cross section is measured
we report a preliminary measurement using the semileptonic electron decay of the charm , performed with @xmath144.3 fb@xmath15 of collision data recorded by the cdf ii detector .
the analysis s result : @xmath81 is in agreement with the nlo theoretical calculation of @xmath82 .
more precise measurements of @xmath1 at the tevatron have been derived using the semileptonic muon decay of the charm , as the identification of soft muons in jets leads to purer charm samples than those based on jets with soft electrons .
the two most accurate measurements to - date of @xmath3 boson production with a single charm quark in @xmath12 collisions at @xmath13 tev , have been made using @xmath141.8 fb@xmath15 and @xmath141.0 fb@xmath15 of data collected with the cdf ii and d0 detectors respectively @xcite .
the cdf anaysis measures @xmath83 .
the d0 analysis measures the ratio @xmath84 , in agreement with a lo prediction of [email protected] .
both these measurements are based on the identification of the charm - jet via the semileptonic muon decay of the charm .
a precision of @xmath85% could be reached in the near future by analyzing the larger datasets of several fb@xmath15 already available at the tevatron .
similarly to @xmath71 or @xmath36 quarks , the associated production of @xmath45 with one or more @xmath72 jets provides an important test of qcd and its accurate modeling is crucial in particle searches , e.g. for the standard model higgs boson in the @xmath86 decay mode @xcite .
we present a preliminary measurement of @xmath87 jets performed using @xmath144.2 fb@xmath15 of data collected with the d0 detector .
jets are reconstructed using the iterative midpoint algorithm @xcite with a cone of radius @xmath88 .
the energy of jets is corrected for detector response , noise , the presence of additional interactions , and energy deposited outside the reconstructed jet cone .
figure [ fig : zbbpt ] shows the @xmath89 distribution of the candidate @xmath72 jets in data and mc , where the mc expectations for various jet flavors have been weighted by the fractions determined by a maximum likelihood fit of the distribution of probability that a jet originates from the primary vertex .
this analysis supersedes an earlier measurement @xcite and yields : @xmath90 for jets with @xmath91 and @xmath92 .
the result is compatible with the nlo qcd prediction of [email protected]% @xcite .
a measurement of @xmath2 had also been performed using @xmath142 fb@xmath15 of collision data collected by the cdf ii detector @xcite . in the analysis ,
jets are selected with @xmath93 and @xmath94 and are identified as @xmath72 jets by the detection of a secondary vertex .
the measurement determines the ratio of the integrated @xmath87 jet cross section to the inclusive @xmath45 production cross section : ( [email protected](stat)@xmath530.42(syst))@xmath95 .
the production ratio @xmath96 for jets and @xmath72 jets with @xmath16 gev and @xmath94 is measured as @xmath97% , both values consistent with predictions from lo mc generators ( alpgen and pythia ) and nlo qcd calculations using mcfm . in @xmath87-jet candidate events , with contributions from @xmath72 , @xmath36 and mis - identification of light - flavor jets.,width=245 ]
the production of @xmath99+heavy flavor jets has been measured @xcite using an integrated luminosity of @xmath141 fb@xmath15 of data collected by the d0 detector .
photons produced in association with heavy quarks ( @xmath36 or @xmath72 ) provide valuable information about the parton ( @xmath100 and gluon ) distributions in the proton and anti - proton .
measurements of the dynamics of @xmath101 also provides testing of perturbative qcd and its modeling helps in searches for new physics that shares the @xmath99+heavy flavor final states .
the analysis provides the first measurements of the differential cross sections @xmath102 in @xmath103 candidate events , for photons with transverse momenta @xmath104 gev/@xmath36 , photon rapidities @xmath59 , jet rapidities @xmath61 , and jet transverse momenta @xmath63 gev/@xmath36 .
the @xmath105 production is seen to agree with the prediction based on nlo pqcd using cteq6.6 m pdfs over the entire range of @xmath106 .
the @xmath107 production is seen to agree with the nlo pqcd prediction for @xmath108 gev/@xmath36 whereas the data becomes larger by about 2 standard deviations than the calculated cross section for harder photons ( @xmath109 gev/@xmath36 ) .
we have presented preliminary measurements of @xmath0+jets , @xmath1 and @xmath2 , and reviewed the latest measurements of vector boson plus inclusive and heavy flavor jets at the tevatron .
all measurements of the production and kinematics of v+jets are in good agreement with nlo pqcd calculations .
more precise measurements of @xmath70 and @xmath1 are needed to constrain the theoretical uncertainties , and help reducing uncertainties on measurements of single top quark and @xmath110 searches .
significant improvements on measurements of @xmath111 production are expected in the near future based on the latest tevatron datasets and developments of the analyses technique .
we use a ( @xmath112 , @xmath113 , @xmath114 ) coordinate system where the @xmath112 axis is in the direction of the proton beam , and @xmath113 and @xmath114 are the azimuthal and polar angles , respectively .
the pseudorapidity is @xmath115 .
transverse energy and momentum are @xmath116 and @xmath117 , respectively , where @xmath118 and @xmath119 are energy and momentum . | we present preliminary measurements of @xmath0+jets , @xmath1 and @xmath2 at the tevatron , and review recent measurements of vector boson plus inclusive and heavy - flavor jets production .
all measurements are in agreement with next - to - leading - order qcd calculations within the experimental and theoretical uncertainties .
we also point to comparisons of the production rate and kinematics of the data with several monte carlo simulation programs of vector boson + jets processes .
[ 1999/12/01 v1.4c il nuovo cimento ] |
in this section we derive the free energy of an ordered lattice at given macroscopic deformations . from the resulting free energy we get fluctuation corrected elastic moduli of the effective medium .
we write the lattice deformation as @xmath91 where @xmath25 labels lattice sites , @xmath92 is the deformation gradient tensor of the ( homogeneous ) macroscopic deformation , and @xmath37 is the fluctuation deviating from the macroscopic deformation . to capture bulk and shear moduli
we take @xmath41 the lattice free energy at given @xmath92 can be obtained by integrating out fluctuations @xmath93 . to do this we first expand the hamiltonian to second order in @xmath93 @xmath94 where @xmath95 is the energy of the uniformly deformed state , @xmath96 is the fourier transform of @xmath37 , and @xmath97 is the @xmath98 dimensional ( @xmath4 being the spatial dimension ) dynamical matrix . in this dynamical matrix
the scalar @xmath99 and second rank tensor @xmath100 are determined by the potentials as follows @xmath101 where @xmath102 is a sum over bonds in one unit cell ( here we only write down the simple case of one site in each unit cell ) , @xmath103 is the vector of this bond connecting site @xmath25 and @xmath104 in the undeformed state , and @xmath105 represent the bond potential in the uniformly deformed state ( of deformation gradient @xmath92 ) and its first and second derivatives .
there is no term linear in @xmath106 in eq .
( [ eq : hexpa ] ) because of the periodicity constraint .
a similar derivation has been used in ref .
@xcite .
we then integrate out fluctuations @xmath93 to obtain lattice free energy , @xmath107 it was shown in ref .
@xcite that @xmath108 only depend on the uniform deformation through the combination of strain tensor @xmath109 so it keeps the same symmetry as @xmath110 and thus the ward identity is satisfied .
in this section we derive finite temperature elastic moduli of the effective medium triangular lattice , and obtain fluctuation corrections to the em spring constant @xmath32 .
the triangular lattice unit cell contains @xmath111 nn bonds with @xmath112 where @xmath53 is the lattice constant . at finite temperature the rigidity of the em triangular lattice in response to shear deformation
is given by @xmath113 where the constant @xmath114 where @xmath115 denote a trace of the whole @xmath116 dimensional matrix @xmath117 ( where @xmath5 is the total number of particles ) and @xmath118 denote a trace of the @xmath98 dimensional matrix @xmath117 at a given @xmath119 .
this finite temperature shear modulus includes contributions from both potential energy in @xmath95 and fluctuation correction term which is proportional to @xmath15 .
because at @xmath0 the shear modulus is solely determined by spring constant @xmath32 , at finite @xmath15 , by identifying @xmath120 we can get from eq . the renormalized value of @xmath32
, @xmath121 this equation characterizes fluctuation stabilization of the lattice . in the cpa , the em spring constant @xmath32
is determined through the following method .
we first consider picking a random bond between sites @xmath104 and @xmath18 from the lattice , and replace the spring constant of the bond from @xmath32 to @xmath122 .
the scattering potential of such a replacement can be written as @xmath123 which acts as a perturbation on the dynamical matrix @xmath124 the resulting green s function can be written in terms of a perturbative expansion @xmath125 where @xmath126 is the ( unperturbed ) phonon green s function of the em , and @xmath127 is the @xmath49-matrix of this perturbation .
the cpa self - consistency condition that the disorder average of the perturbed green s function @xmath46 to be equal to @xmath128 can then be written in the form @xmath129 this equation determines the em spring constant @xmath130 . in the case of replacing one cf bond this sum for the @xmath49-matrix can be done exactly , because @xmath131 where @xmath132 is the unit vector along bond @xmath133 , and in the last line we write the integral as @xmath134 where @xmath135 is a dimensionless number . using this
the @xmath49-matrix can be written as @xmath136 plugging this into the self - consistency equation we have @xmath137 at @xmath0 we use the bare value @xmath32 in the em green s function @xmath128 , and the resulting integral is @xmath138 which give the meanfield @xmath0 rigidity percolation point for the triangular lattice @xmath139 . at @xmath140 the renormalized value as in eq .
should be used in the green s function and we have @xmath141 and the self - consistency equation becomes @xmath142 which leads to different regimes of entropic elasticity as discussed in the main text .
in this model the nn bonds are all present so the nn spring constant keeps the same value @xmath143 , and the cpa refers to mapping randomly placed nnn bonds to uniform nnn bonds on every site with an em spring constant @xmath65 . each unit cell in the em has @xmath144 nn bonds and @xmath144 nnn bonds , with vectors @xmath145 in order to extract the renormalized @xmath65 we need to consider a simple shear of the square lattice , which to leading order only deforms the nnn bonds .
similar to the calculation for the triangular lattice we have @xmath146 in contrast to the fluctuation correction in the triangular lattice case as in eq .
, here the non - analytic dependence on @xmath18 and @xmath65 comes from the sub - extensive number of floppy modes along the @xmath70 and @xmath71 directions in momentum space .
these floppy modes provide a divergent fluctuation correction to elastic moduli as the lattice approaches instability @xmath147 .
the integral involving these floppy modes has been discussed in detail in ref .
@xcite . as a result
the renormalized nnn spring constant is @xmath148 rigorously speaking , the nn spring constant @xmath18 also get renormalized by fluctuation , but because in this problem @xmath18 is always large its renormalization does not have any significant effect on stability . in the cpa of the randomly braced square lattice
we replace one arbitrary nnn bond .
the calculation for the perturbed green s function follows a similar procedure as we did for the triangular lattice .
the self - consistency equation takes a similar form @xmath149 with @xmath150 where @xmath133 represent an arbitrary nnn bond .
the function @xmath151 for nnn bonds in the square lattice behaves rather differently from @xmath151 in the triangular lattice as we discussed above . at @xmath0 , instead of approaching a constant we have @xmath152 with the non - analytic dependence coming from the sub - extensive number of floppy modes .
this @xmath0 solution has been discussed in ref .
@xcite . at @xmath140
, we replace @xmath153 in the phonon green s function @xmath128 ( not in @xmath151 which has an extra prefactor of @xmath65 ) and expand at low temperature .
this lead to the finite temperature cpa self - consistency equation @xmath72
in this section we discuss the simple picture of one particle tethered between two walls by two collinear springs . to calculate shear modulus , we introduce a transverse displacement of the two walls , @xmath154 on the top and @xmath155 on the bottom .
( this is equivalent to fixing one wall and only displace the other one by @xmath43 but that way one need to expand around new equilibrium position of the particle . )
the hamiltonian is then @xmath156 one can then calculate the canonical partition function at fixed @xmath15 by integrating over particle position @xmath157 , @xmath158 where @xmath159 represent modified bessel function of the second kind .
taking derivatives of the resulting free energy @xmath160 with respect to @xmath43 gives @xmath161 so the system exhibit a shear modulus that is proportional to @xmath162 .
48ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop * * , ( ) \doibase doi : 10.1016/0020 - 7683(78)90052 - 5 [ * * , ( ) ] @noop * * , ( ) \doibase doi : 10.1016/0038 - 1098(85)90381 - 3 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop ( ) link:\doibase 10.1103/physrevlett.104.085504 [ * * , ( ) ] @noop * * ( ) link:\doibase 10.1103/physreve.83.011111 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.75.4051 [ * * , ( ) ] link:\doibase 10.1103/physreve.53.3682 [ * * , ( ) ] link:\doibase 10.1103/physreve.76.041135 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.74.4472 [ * * , ( ) ] @noop * * , ( ) in http://dblp.uni-trier.de/db/conf/soda/soda2011.html#kasiviswanathanmt11[__ ] , ( , ) pp .
link:\doibase 10.1103/physrevlett.97.148701 [ * * , ( ) ] link:\doibase 10.1103/physreve.87.042602 [ * * , ( ) ] @noop ( ) @noop ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) in @noop _ _ , ( , ) chap .
@noop _ _ ( , ) @noop _ _ ( , ) link:\doibase 10.1103/physrevlett.93.225701 [ * * , ( ) ] link:\doibase 10.1103/physreve.80.031140 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.68.405 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physreve.67.011903 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.80.4907 [ * * , ( ) ] @noop ( ) link:\doibase 10.1103/physrevlett.109.018301 [ * * , ( ) ] @noop * * , ( ) @noop ( ) link:\doibase 10.1103/physrev.178.1136 [ * * , ( ) ] link:\doibase 10.1103/physrevb.31.7276 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.52.216 [ * * , ( ) ] link:\doibase 10.1103/physrevb.31.276 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.54.2107 [ * * , ( ) ] link:\doibase 10.1103/physrevb.37.5539 [ * * , ( ) ] link:\doibase 10.1103/physreve.87.042601 [ * * , ( ) ] @noop * * , ( ) | mechanical instability takes different forms in various ordered and disordered systems .
we study the effect of thermal fluctuations in two disordered central - force lattice models near mechanical instability : randomly diluted triangular lattice and randomly braced square lattice . these two lattices exhibit different scalings for the emergence of rigidity at @xmath0 due to their different patterns of self stress at the transition . using analytic theory
we show that thermal fluctuations stabilize both lattices .
in particular , the triangular lattice displays a critical regime in which the shear modulus scales as @xmath1 , whereas the square lattice shows @xmath2 .
mechanical instability takes a rich variety of forms in systems ranging from granular matter to biological tissue and man - made structures @xcite . even in lattice models where there is a remarkably simple principle
the maxwell s rule which compares discrete numbers of degrees of freedom and constraints @xcite multiple classes of distinct behaviors arise .
mechanical instabilities in lattice models can take the form of either mean - field @xcite or non mean - field @xcite transitions , with continuous @xcite or discontinuous @xcite changes of various elastic moduli , and in some case even with a phonon structure of nontrivial topology @xcite .
one central concept in understanding these different classes of behaviors is _ self stress _ , which describes distributions of stress on bonds in the lattice with which the net force on every site vanishes .
this is characterized by the generalized maxwell s rule @xcite @xmath3 where @xmath4 is the dimension , @xmath5 is the number of sites , @xmath6 is the number of central force ( cf ) bonds , @xmath7 is the number of floppy modes , and @xmath8 is the number of states of self stress . according to this rule , each bond added to a lattice either removes one floppy mode or increases one state of self stress , and in the latter the bond can be considered as `` redundant . ''
moreover , states of self stress are intimately related to elastic moduli because it describes the nature of stress the lattice can sustain @xcite . whether a bond is redundant or not strongly depend on the architecture of the lattice @xcite .
thus a small difference in the construction of the lattice can lead to very different behaviors near mechanical instability .
different classes of lattice models have been found to relate to different experimental systems , such as jamming of spheres @xcite , glasses @xcite , and semi - flexible polymer networks @xcite .
little is known about how thermal fluctuations affect mechanical instability in these different classes of models . in general , fluctuations have been shown to change conclusions from mean - field theories about transitions dramatically , e.g. , fluctuations can alter the order and scaling exponents of the transition , or even ruling out the ordered phase completely @xcite .
it has been pointed out that thermal fluctuations may stabilize under - coordinated networks in some special cases including crosslinked flexible polymer @xcite , tethered networks @xcite , diluted lattices @xcite , and jammed packings @xcite , but a generic theoretical framework for finite - temperature mechanical instability , especially in disordered systems , is still lacking
. such generic theory will be important in understanding experimental observations of mechanical instability which are made at finite temperature . in this letter
, we construct an analytic theory for disordered lattices near mechanical instability at finite temperature .
our theory treats thermal fluctuations through renormalization of elastic moduli and treats lattice disorder through the coherent potential approximation ( cpa ) @xcite .
we use this theory to discuss two cf lattice models , the randomly diluted triangular lattice @xcite and the randomly braced square lattice @xcite , both of which display mechanical instabilities when they pass through the cf isostatic point @xmath9 where @xmath10 is the coordination number .
our main findings are summarized in the phase diagrams in fig .
[ fig : pd ]
. the triangular lattice is stabilized by thermal fluctuations near the isostatic point , and it displays three regimes of elasticity , @xmath11 ( mechanical ) , @xmath12 ( critical ) , and @xmath13 ( entropic ) , consistent with numerical results in ref .
@xcite .
interestingly the square lattice shows very different exponents , e.g. , @xmath14 in the critical regime , as well as a nonaffine - affine crossover , originating from the sub - extensive number of states of self stress at the transition .
our theory is readily generalizable to other models to map out finite - temperature phase diagrams for different classes of systems near mechanical instability . for having each bond ( nnn bonds in the case of square lattice ) and temperature @xmath15 measured in units of
@xmath16 shown in ( c ) and ( d ) respectively . in
( d ) nnn spring constant @xmath17 .
red dots mark the @xmath0 mechanical instability point . ]
_ model _ we consider two types of lattices near mechanical instability .
the first one is the randomly diluted cf triangular lattice [ fig .
[ fig : pd](a ) ] , in which each nearest - neighbor ( nn ) bond is of spring constant @xmath18 and is present with probability @xmath19 .
previous numerical studies @xcite have shown that the @xmath0 rigidity transition in the triangular lattice is continuous and occur at @xmath20 with non - mean - field exponents .
the second one is the randomly braced cf square lattice [ fig .
[ fig : pd](b ) ] , in which all nn bonds are present with spring constant @xmath18 and each next - nearest - neighbor ( nnn ) bond of spring constant @xmath21 is present with probability @xmath19 .
the @xmath0 rigidity percolation of this model has been studied both numerically and analytically , revealing a transition at @xmath22 with mean - field exponents and features of both first- and second - order transitions .
the hamiltonian of both lattices can be written as @xmath23 for a deformation that moves positions @xmath24 for each lattice site @xmath25 . here
the sum is over all bonds @xmath26 , @xmath27 if the bond is present and @xmath28 if not , and @xmath29 is the vector of the bond in the reference space ( assuming all bonds to be at rest length in the reference space ) . in the triangular lattice @xmath30 , and in the square lattice @xmath30 for the nn bonds and @xmath31 for the nnn bonds .
we choose to use the coherent potential approximation ( cpa ) for the disorder in these models .
this method , also called the effective medium theory ( emt ) , has been shown to work very well for mechanical properties of disordered lattices @xcite .
the spirit of the original zero temperature cpa is that the disordered lattice can be mapped into an effective medium ( em ) regular lattice in which all bonds are present with an effective medium spring constant @xmath32 as a function of dilution @xmath19 .
this mapping is determined by the self - consistency condition that if one randomly chosen bond in the em is replaced according to the rules of the original disordered lattice ( change into spring constant @xmath18 with probability @xmath19 and removed with probability @xmath33 ) , the average of the perturbed phonon green s function equals to the unperturbed em green s function . in this letter
we develop a version of the cpa that incorporates thermal fluctuation corrections in order to study finite-@xmath15 mechanical instability .
assuming that the mapping to an em regular lattice still holds at finite temperature , but with a temperature dependent em spring constant @xmath34 , we characterize thermal fluctuation effects by calculating renormalized elastic moduli of the em as follows .
consider a deformation @xmath35 where @xmath25 labels lattice sites , @xmath36 denote a macroscopic deformation gradient , and @xmath37 denote finite wavelength fluctuations around @xmath38 .
free energy of the lattice at a given macroscopic deformation @xmath36 in finite temperature can be obtained by integrating out fluctuations @xmath37 , which yield a vibrational entropy term .
although the hamiltonian contains only harmonic springs , expanding @xmath39 in terms of displacements generate higher order terms in displacements , and thus this entropic term depends on @xmath36 .
it was shown in ref .
@xcite that ward identity holds for the fluctuation corrected free energy @xmath40 .
details of this calculation is shown in the supplementary information ( si ) .
we consider generic deformations of the form @xmath41 and taking derivatives of @xmath42 with respect to @xmath43 and @xmath44 yields the fluctuation corrected elastic moduli .
we then obtain renormalized spring constants from these corrected elastic moduli .
next we use the cpa procedure to calculate the perturbed green s function as one bond is replaced @xmath45 where in the phonon green s function @xmath46 renormalized spring constants instead of the @xmath0 ones are used . here
@xmath47 is the perturbative potential describing changing the spring constant of a randomly chosen bond in the em , and @xmath48 is the @xmath49-matrix of this perturbation . in the case of replacing one cf bond this sum for the @xmath49-matrix
can be done exactly .
the cpa self - consistency condition can then be written in the form @xmath50 we use this equation to determine the em spring constants .
details are included in the si .
_ results _ we first apply this finite temperature cpa to the diluted triangular lattices .
the renormalized elastic moduli @xmath51 is determined by calculating shear rigidity of the em lattice @xmath52 where @xmath53 is the lattice constant and @xmath54 .
we then extract the renormalized spring constant @xmath51 through the relation @xmath55 and find @xmath56 where the last term indicates an increase of em spring constant with @xmath15 , characterizing fluctuation stabilization of this lattice . using this renormalized @xmath51
we find the cpa self - consistency equation @xmath57 where @xmath58 with @xmath59 being the meanfield rigidity transition point for the triangular lattice .
taking @xmath60 recovers the zero temperature cpa result @xcite .
in contrast , at finite @xmath15 , with the addition of this fluctuation correction term , we find the solution @xmath61 which has 3 distinct regimes @xmath62 as shown in fig .
[ fig : g](a ) . because elastic moduli of the lattice at finite temperature are both proportional to @xmath32
, these 3 regimes correspond to mechanical , critical , and entropic regimes of elasticity respectively , as shown in fig .
[ fig : pd](c ) . for diluted triangular lattice as a function of @xmath63 .
the solutions given by eq .
are shown in solid curves , and the asymptotic forms for the mechanical / critical / entropic regimes are shown in dotted curve / markers / dashed curves , for @xmath64 from top to bottom .
( b ) cpa solution of @xmath65 for randomly braced square lattice as a function of @xmath19 .
numerical solutions of eq . are shown in solid curves , and the asymptotic solution for the nonaffine mechanical / affine mechanical / critical regimes are shown in dotted / dot - dashed / dashed lines , for @xmath66 from top to bottom . ]
we apply a similar calculation to randomly braced square lattices . in this case
we determine renormalized nnn spring constant @xmath67 by again calculating finite temperature shear modulus , which yields @xmath68 comparing to eq .
we find that the renormalization of the nnn bond spring constant contains a non - analytic dependence on the ratio @xmath69 .
this comes from the unique phonon structure of the square lattice , which exhibits lines of states of self stress and floppy phonon modes along @xmath70 and @xmath71 axis in the first brillouin zone , leading to non - analyticity in phonon green s functions .
the resulting finite-@xmath15 cpa self - consistency equation then reads @xmath72 its solution can be expanded in 3 regimes too @xmath73 corresponding to nonaffine mechanical , affine mechanical , and critical regimes , as shown in fig .
[ fig : g](b ) and the phase diagram fig .
[ fig : pd](d ) .
it is worth noting that the @xmath74 scaling agrees with the ordered lattice results in ref .
@xcite with @xmath22 .
_ discussions _ in both lattice models we found that thermal fluctuations provide stabilizing effect to elasticity , even when the lattice is below the isostatic point .
moreover , the square lattice finite-@xmath15 phase diagram near mechanical instability differs from that of the triangular lattice , because of the different structure of states of self stress in the square lattice .
there is an intuitive picture to understand the different regimes of the triangular lattice .
the mechanical regime for @xmath75 at low @xmath15 is straightforward the lattice is stable at @xmath0 and thermal fluctuation simple provide small correction to elastic moduli . in the critical regime ,
the system starts to have some modes that are of very low stiffness .
finite temperature behavior of these modes can be understood by considering the following simple example [ fig .
[ fig : onep](a ) ] . here a particle is connected by 2 springs in the @xmath76 direction to 2 walls and the springs are at rest length when the particle is at @xmath77 . assuming the 2 springs to be harmonic with force constant @xmath18 , the hamiltonian of the particle is to leading order @xmath78 where @xmath79 is the rest length of the springs
it is clear that vibrations in @xmath80 is a floppy mode of this simple system because its @xmath0 stiffness @xmath81 .
equivalently , the `` shear modulus '' of the system ( resistance to transverse displacement @xmath43 of the top wall ) is also @xmath28 . at finite @xmath15
, the shear modulus can be calculated using canonical ensemble ( see si ) @xmath82 where @xmath83 .
this @xmath84 thermal stiffness for floppy modes is an intuitive way to understand the critical regime in the diluted triangular lattice , because in this regime , there is only a small number of floppy modes @xmath85 with positive coefficients for @xmath86 terms in @xmath39 , giving them entropic rigidity of @xmath87 . resistance to transverse deformation .
one example of instantaneous fluctuation configuration is shown in gray .
( b ) a schematic picture of floppy modes further below the isostatic point , in which large fluctuations ( gray ) cost no energy .
this leads to @xmath88 rigidity . ]
when the diluted lattice is deeper below @xmath89 there are more floppy modes which allow large fluctuations ( flat energy landscape , no quartic term ) , and the corresponding physics , as shown in fig.3b , is more like a traditional `` entropic elasticity '' picture for flexible polymers .
thus the corresponding entropic elasticity is proportional to @xmath15 as in the case of polymers .
in addition , in these simple pictures and our lattice model , at finite @xmath15 a _ negative stress _ is generated by thermal fluctuations , and we take fixed volume ensemble to avoid collapsing .
however , fluctuation stabilization of elasticity is not limited to the case with fixed volume .
it was shown in ref .
@xcite that with zero pressure but self - avoidance between bonds , similar stabilization and scalings are observed . in our theory , this correspond to allow a negative thermal expansion at finite @xmath15 , and add additional anharmonic terms in the hamiltonian to provide stability .
an analytic theory of this effect in ordered lattices has been discussed in ref .
@xcite .
our analytic calculations on lattice models provide a simple framework to explore fluctuation effects in more complicated systems near mechanical instability .
for example , recent studies on jamming at finite temperature revealed significant fluctuation stabilization effect below point j ( @xmath90 ) @xcite .
it would be interesting to consider variations of our lattice models to include effects such as one - sided potential and bond - breaking to capture regimes of elasticity near point j. |
the problem as to whether meteor showers associated with near earth asteroids ( neas ) exist or not has been discussed by many authors ( e.g. , olsson - steel 1988 ; hasegawa et al .
1992 ; babadzhanov 1994 ) .
the most prominent case is the association between nea ( 3200 ) phaethon and the geminid stream complex .
however , it should be an exceptional case since phaethon is considered to be a dormant or extinct cometary nucleus ; hence the geminid stream complex meteoroids were presumably released from phaethon in its active cometary phase ( gustafson 1989 ) .
meanwhile , meteoroids of asteroidal origin certainly exist and they are usually observed and recognized as meteorites , meteoritic fireballs , or high bulk - density meteors .
they are probably produced as impact ejecta by asteroidal collisions .
although there is no asteroid
meteor association as strong as the phaethon
geminid complex yet , most likely the asteroidal meteor ( or meteorite ) streams indeed exist ( e.g. , halliday 1987 ; terentjeva & barabanov 2002 ; spurn et al .
2003 ) .
the hayabusa ( muses - c ) sample return mission target , nea ( 25143 ) itokawa = 1998 @xmath0 , is one of the apollo - type neas , whose orbital parameters at present epoch are perihelion distance @xmath3 au and small eccentricity @xmath4 along with an inclination near the ecliptic plane at @xmath5 .
the taxonomic class of itokawa is less the cometary nature than a typical s(iv)-type with rather high albedo ( e.g. , binzel et al . 2001
; sekiguchi et al . 2003 ; ishiguro et al . 2003 ) .
a meteoritic analog would correspond to a ll chondrite ( binzel et al .
since itokawa is a small asteroid of @xmath6 m in mean axis , the escape speed from itokawa will be only @xmath7 cm s@xmath8 ( ostro et al .
consequently , impact ejecta should easily escape from itokawa toward interplanetary space and some may become meteoroids
. the earth approaches itokawa s orbit to within 0.05 au from late march to early july , over a span of three months , in which we would be able to observe some meteors originated from itokawa ( hereafter , we designate such a meteor as itokawid " ) every year . if we can obtain the orbital and physical data of the itokawid meteors , they must give us very important information when compared with the forthcoming analytical results of the hayabusa onboard scientific instruments and the sample of itokawa .
on the basis of the above assumption , we surveyed whether there are itokawid meteors or not from among the optical meteor orbit database , considering orbital similarity with itokawa . as a result
, we could locate itokawid meteor candidates .
we also analyzed the corresponding mineral materials of the itokawid candidates through their trajectory and atmospheric data .
we called japanese visual - meteor observers to watch the itokawid meteor activity through the observing campaign from the 2001 season .
however , no detection has been reported so far .
this is supported by the negative results of the japanese fireball network ( jn ) monitoring , a total of @xmath9 hr , during the 20002004 seasons .
one of the reasons is that a predicted geocentric velocity for the itokawids is very slow at only @xmath10 km s@xmath8 , which should scatter their apparent radiants in a large celestial area around leonis .
this makes it difficult for visual meteor observers to distinguish a member of the itokawids or not .
another reason is that the activity of asteroidal meteors is substantially weak .
indeed , we could not find any past itokawid records in the several visual meteor catalogues .
thus , the itokawid activity has never been reported by single - station visual meteor observers . unlike the visual meteor observations
, we might trace itokawid activity by retrieving data from the meteor orbit database recorded by multiple - station observations . considering the orbital similarity with itokawa
, we surveyed whether there are itokawid meteors among the multi - station optical ( photographic and tv ) meteor orbit data of @xmath1 orbits .
our database includes : i ) the iau mdc optical meteor orbit data of @xmath11 orbits as of august , 2001 ( lindblad 2001 ; lindblad et al . 2001 ) , based upon many astronomical papers that have been published ( steel 1996 ) ; ii ) 6500 more optical meteor orbits from released catalogues on the internet websites ( e.g. dms additional photo and tv ; mss - wg tv ; ondrejov tv ; etc . ) and unpublished tmn photographic and jn fireball reductions ; and iii ) super - schmidt graphical reductions ( mccrosky & posen 1961 ) of @xmath12 orbits , although they were not included in the iau mdc photographic meteor data ( lindblad 2001 ) by reason of their lower precision than other photographic reductions ; they are comparable in quality to the tv reductions and hence they deserve to be retrieved here . in investigating the orbital similarity between meteors or a comet / asteroid and meteors ,
we often use the @xmath2-criteria , e.g. @xmath13 ( southworth & hawkins 1963 ) ; @xmath14 ( drummond 1979 ) ; @xmath15 and @xmath16 ( valsecchi et al .
we applied these @xmath2-criteria as a retrieving engine , mainly taking account of discrimination @xmath17 .
the traditional @xmath13 is defined by a distance of two points of two objects ( @xmath18 and @xmath19 ) in the five - dimensional coordinate space systemized by the orbital elements , @xmath20 : eccentricity , @xmath21 : perihelion distance , @xmath22 : argument of perihelion , @xmath23 : longitude of the ascending node , and @xmath24 : inclination , as follows ( southworth & hawkins 1963 ) : @xmath25 ^ 2 = ( e_a - e_b)^2+(q_a - q_b)^2+[2 \sin ( i_{ab } / 2)]^2 \\ + \{[(e_a+e_b ) / 2][2 \sin ( \mathit{\pi}_{ab } / 2)]\}^2 \nonumber \end{split}\ ] ] where @xmath26 is the angle between both orbital planes and @xmath27 is the difference between the longitudes of perihelion measured from the intersection of the orbits .
therefore , you can easily comprehend that the smaller @xmath13 is , the stronger the association between itokawa and meteor will be .
we found five itokawid meteor candidates , as listed in table 1 , where the column heads from left to right are as follows : object name , observed date of the meteors in ut , @xmath20 , @xmath21 in au , @xmath28 : semimajor axis in au , the angular elements , @xmath22 , @xmath23 , and @xmath24 in degree in j2000 , @xmath29 and @xmath30 : right ascension and declination of the geocentric radiant in degree in j2000 , and @xmath31 : geocentric velocity in km s@xmath8 .
the orbital parameters of itokawa at epoch 2004 july 14.0 tt (= jdt 2453200.5 ) taken from the nasa - jpl neo program website are also presented in table 1 , when mean anomaly was @xmath32 .
the equinox of angular data , except for those of the meteor 00504004 , was transformed from b1950 to j2000 .
the results of the @xmath2-criteria are also given in table 2 , where the column heads are the following : object name , @xmath13 between itokawa and each meteor , @xmath33 and @xmath34 : longitude and latitude of the perihelion in degree in j2000 ( main terms of @xmath14 ) , and @xmath35 and @xmath36 : unperturbed geocentric encounter velocity to the earth and cosine of the angle between @xmath35 and the direction of earth s heliocentric motion ( main terms of @xmath15 and @xmath16 ) .
@xmath35 equals @xmath37 , where @xmath38 is the tisserand s invariant of each object with respect to jupiter .
the meteors , h7220 and h10330 , were picked out from the super - schmidt graphical reductions ( mccrosky & posen 1961 ) .
@xmath13 between itokawa
h7220 is 0.05 , the smallest among all the itokawa
meteor pairs and comparable to the confirmed comet meteor associations ( e.g. , 55p leonids , 103p perseids , phaethon geminids ; and others ) ; thus both objects are probably associated with one another .
although outside the predicted months mentioned above , the meteor , ha10000 , came from the super - schmidt automatic reductions ( mccrosky & shao 1967 ) .
the fireball , morp172 , was multiply photographed by the canadian fireball network , meteorite observation and recovery project ( morp ) , on april 11 , 1975 ( halliday et al . 1996 ) ; @xmath13 of 0.09 between itokawa morp172 is somewhat larger than that of itokawa h7220 .
however it is still within a probable association range .
it is also notable that the secular near - invariants of morp172 in the @xmath35-@xmath39 coordinates match well with those of itokawa .
this suggests that both objects are dynamically in strong relationship to each other ( valsecchi et al .
1999 ) .
[ table1 ] [ table2 ] the tv meteor , 00504004 , was recorded by the ondrejov observatory team ( koten et al .
2003 ) on may 4 , 2000 ; the @xmath13-criterion with itokawa is just outside the 0.15-limit , but its @xmath33 and @xmath34 are very close to those of itokawa and hence in a possible association range . true anomaly of these meteors , at observed epoch , equals @xmath40 for the super - schmidt meteors and morp172 and @xmath41 for the meteor , 00504004 .
furthermore , we approached the above subject in a physical analysis , in order to ascertain the association between itokawa and the itokawid candidates .
unfortunately , most of the physical information ( e.g. , bulk density , ablation coefficient , spectral data , etc . ) , are not given in the published data , except for the mass data .
however , we could analyze corresponding mineral matters of the itokawid candidates through their trajectory and atmospheric data instead , i.e. , atmospheric density at meteor s luminous height , velocity , entry angle to earth atmosphere , and mass , using ceplecha s ( 1988 ) classification .
the itokawid candidates are classified on the basis of evaluating the coefficients @xmath42 only for optical ( super - schmidt and tv ) meteors and @xmath43 and @xmath44 only for fireballs . in these evaluations
, we adopted the atmospheric data from cira1965 .
the results of classification are summarized in table 3 , where the column heads from left to right are as follows : object name , @xmath45 : initial velocity in km s@xmath8 , @xmath46 : cosine of zenith distance of the apparent radiant , @xmath47 and @xmath48 : beginning and end height of meteor in km , @xmath49 : initial ( photometric ) mass in gram , @xmath42 , @xmath43 and @xmath44 , classified group , and classified mineral matter . it should be also noted that the @xmath42 coefficient for the tv meteors is adjusted adding 0.15 and the logarithm of the total luminosity of morp172 for @xmath44 is 3.31 ( in 0 mag sec ) .
[ tab3 ] these evaluations indicate all the itokawid candidates evidently belong to asteroidal meteors , as shown in table 3 : all the super - schmidt and tv meteors are classified in the range of group a as carbonaceous chondrite ( @xmath50 ) , while the fireball morp172 is in group i as ordinary chondrite by @xmath43 classification , @xmath51 , but in group ii as carbonaceous chondrite by @xmath44 classification , @xmath52 .
hence , the classified mineral matter of morp172 shows more similarity to itokawa s surface composition of a ll chondrite analogue suggested by binzel et al .
( 2001 ) than the four other meteors .
interestingly , according to halliday et al ( 1989 ) , morp172 was observed as a likely meteorite dropping fireball with an estimated terminal mass of @xmath53 kg ; this meteorite might have fallen , and still remains , somewhere near edam , canada .
as a result of a survey of the itokawid meteors from among our multi - station optical meteor orbit data of @xmath1 orbits , and applying the orbital similarity criteria , @xmath2-criteria , we could find five itokawid meteor candidates .
two out of them , the meteor h7220 and the fireball morp172 , are probably associated with itokawa , judging from their @xmath13 values .
it is also notable that the secular near - invariants of morp172 in the @xmath35-@xmath39 coordinates match well with those of itokawa : this suggests both objects have a strong dynamical relationship to each other .
the @xmath42 , @xmath43 and @xmath44 evaluations indicate all the itokawid candidates evidently belong to asteroidal meteors .
morp172 is the only one classified at ordinary / carbonaceous chondrite , the others being in carbonaceous chondrites .
hence , the classified mineral matter of morp172 shows more similarity to itokawa s surface composition of a ll chondrite analogue than the other candidates .
we conclude , on the basis of the investigations above , that the fireball , morp172 , is the strongest itokawid candidate . however , if you ask whether there are itokawid meteors or not , we will answer , possible " , since only a small sample like this is available so far
. therefore , more optical detections and analyses for the itokawid meteors are desirable in the future work .
we thank messrs . c. shimoda and s. okumura for providing the jn monitoring data and dr .
p. koten for offering unpublished trajectory data of the meteor , 00504004 .
babadzhanov , p. b. 1994 , in seventy - five years of hirayama asteroid families : the role of collisions in the solar system history , y. kozai , r. p. binzel , & t. hirayama ( eds ) , asp conf .
63 , 168 binzel , r. p. , rivkin , a. s. , bus , s. j. , sunshine , j. m. , & burbine , t. h. 2001 , m&ps , 36 , 1167 ceplecha , z. 1988 , bull .
czsl . , 39 , 221 drummond , j. d. 1979 , proc . southwest reg
astrophys . , 5 , 83 gustafson , b. a. s. 1989 , a&a , 225 , 533 halliday , i. 1987 , icarus , 69 , 550 halliday , i. , blackwell , a. t. , & griffin , a. a. 1989 , jrasc , 83 , 49 halliday , i. , griffin , a. a. , & blackwell , a. t. 1996 , m&ps , 31 , 185 hasegawa , i. , ueyama , y. , & ohtsuka , k. 1992 , pasj , 44 , 45 ishiguro , m. , abe , m. , ohba , y. , fujiwara , a. , fuse , t. , terada , h. , goto , m. , kobayashi , n. , tokunaga , a. t. , & hasegawa , s. 2003 , pasj , 55 , 691 koten , p. , spurn , p. , borovicka , j. , & stork , r. 2003 , pub .
czech rep .
, 91 , 1 lindblad , b. a. 2001 , in proceedings of the meteoroids 2001 conference , b. warmbein ( ed ) , esa sp-495 , 71 lindblad , b. a. , neslusan , l. , svoren , j. , & porubcan , v. 2001 , in proceedings of the meteoroids 2001 conference , b. warmbein ( ed ) , esa sp-495 , 73 mccrosky , r. e. & posen , a. 1961 , smithson .
astrophys . , 4 , 15 mccrosky , r. e. & shao , c. -y .
1967 , harvard meteor res .
prog . , semiannual tech
3 olsson - steel , d. 1988 , icarus , 75 , 64 ostro , s. j. , benner , l. a. m. , magri , c. , giorgini , j. d. , rose , r. , jurgens , r. f. , yeomans , d. k. , hine , a. a. , nolan , m. c. , scheeres , d. j. , kaasalainen , m. , vokrouhlick , d. , chesley , s. r. , & margot , j. l. 2004 , in the 1st hayabusa symposium sekiguchi , t. , abe , m. , boehnhardt , h. , dermawan , b. , hainaut , o. r. , & hasegawa , s. 2003 , a&a , 397 , 325 spurn , p. , oberst , j. , & heinlein , d. 2003 , nature , 423 , 151 steel , d. 1996 , space sci .
, 78 , 507 southworth , r. b. & hawkins , g. 1963 , smithson .
, 7 , 261 terentjeva , a. k. & barabanov , s. i. 2002 , sol .
res . , 36 , 431 valsecchi , g. b. , jopek , t. j. , & froeschle , cl .
1999 , mnras , 304 , 743 | as a result of a survey of itokawid meteors ( i.e. , meteors originated from near earth asteroid ( 25143 ) itokawa = 1998 @xmath0 ) , from among the multi - station optical meteor orbit data of @xmath1 orbits , and applying the @xmath2-criteria , we could find five itokawid meteor candidates . we also analyzed corresponding mineral materials of the itokawid candidates through their trajectory and atmospheric data .
we conclude , on the basis of our investigation , that the fireball , morp172 , is the strongest itokawid candidate . |
in spectral graph theory one uses spectral analysis to study the interplay between the topology of a graph and the dynamical processes modelled through the graph @xcite .
in fact , various dynamical processes in disciplines ranging from physics , biology , information theory , and chemistry to technological and social sciences are modelled with graph theory .
therefore , graph theory forms a unified framework for their study @xcite .
in particular , one associates a certain matrix to a graph ( e.g. the adjacency matrix , the laplacian , the google matrix ) and studies the connection between the spectral properties of this matrix and the properties of the dynamical processes governed through them .
we mention some studies in this context : the stability of synchronization processes @xcite , the robustness and the effective resistance of networks @xcite , error - correcting codes @xcite , etc .
these examples illustrate how spectral graph theory relies to a large extent on the capability of determining spectra of sparse graphs .
it is thus important to develop mathematical methods which allow to derive in a systematic way exact analytical as well as numerical results on spectra of large graphs . for an overview of analytical results on the spectra of infinite graphs we refer to the paper of mohar and woess @xcite .
recently , the development of exact results for large sparse graphs has been reconsidered using ideas from statistical physics of disordered systems . in this approach one
formulates the spectral analysis of graphs in a statistical - mechanics language using a complex valued measure @xcite .
the spectrum is given as the free energy density of this measure , which can be calculated using methods from disordered systems such as the replica method @xcite , the cavity method @xcite or the super - symmetric method @xcite .
this approach is exact for infinitely large graphs that do not contain cycles of finite length . in recent
works the cavity method has been generalized to the study of spectra of graphs with a community structure @xcite and spectra of small - world graphs @xcite .
the replica and cavity methods have also been used to derive the largest eigenvalue of sparse random matrix ensembles @xcite .
although the cavity method is heuristic , it has been considered in a rigorous setting for undirected sparse graphs using the theory of local weak convergence @xcite . however , for directed sparse graphs the asymptotic convergence of the spectrum to the resultant cavity expressions has not been shown @xcite .
nevertheless , recent studies have shown the asymptotic convergence of the spectrum of highly connected sparse matrices to the circular law @xcite and have proven the asymptotic convergence of the spectrum of markov generators on higly connected random graphs @xcite .
these studies , however , do not concern finitely connected graphs . in this work
we extend the statistical - mechanics formulation , in particular the cavity method for the spectra of large sparse graphs which are locally tree - like @xcite , to large sparse graphs with many short cycles .
such an extension is relevant because cycles do appear in many real - world systems such as the internet @xcite .
we derive a set of resolvent equations for graph ensembles that contain many cycles of finite length .
these equations are exact for infinitely large ( un)directed husimi graphs @xcite and solving them constitutes an algorithm for determining the spectral density of these graphs .
first , we show how this algorithm determines the spectra of irregular husimi graphs up to a high accuracy , well corroborated by numerical simulations .
then we derive novel analytical results not only for the spectra of undirected regular husimi graphs , but also for the spectra of directed regular husimi graphs .
in particular , we show that the boundary of the spectrum of directed husimi graphs composed of cycles of length @xmath0 , is determined by a hypotrochoid with radii @xmath1 ( @xmath2 being the radius of the fixed circle and @xmath3 of the moving circle ) .
not many analytical expressions for the spectra of directed sparse random graphs and non - hermitian random matrices are known ( besides some exceptions , see @xcite ) .
a short account of some of our results on regular graphs has appeared in @xcite .
networks appearing in nature are usually modelled with theoretical ensembles of graphs .
these ensembles consist of randomly constructed graphs with certain topological constraints on their connectivity .
model systems allow for a better understanding of the properties of the more complex real - world systems @xcite . in this work we consider simple graphs @xmath4 of size @xmath5 consisting of a discrete set of vertices @xmath6 and a set of edges @xmath7 .
simple graphs are uniquely defined by their adjacency matrix @xmath8 , with elements @xmath9_{ij } = a_{ij}\in \left\{0,1\right\}$ ] for @xmath10 and @xmath9_{ii } = 0 $ ] for @xmath11 .
we have @xmath12 when @xmath13 and zero otherwise .
an undirected edge is present between @xmath14 and @xmath15 when @xmath16 , a directed edge is present from @xmath14 to @xmath15 when @xmath12 and @xmath17 , while @xmath18 indicates that there is no edge present between @xmath14 and @xmath15 .
we define graph ensembles through a normalized distribution @xmath19 for the adjacency matrices @xmath8 . selecting a graph from the ensemble corresponds with drawing randomly an adjacency matrix from the distribution @xmath19 .
we always consider ensembles of infinite graphs , for which @xmath20 .
this is implicitly assumed throughout the whole paper .
below , we first define random graphs which are locally tree - like .
their spectral properties have been considered in several studies @xcite .
second , we define ensembles of graphs with many cycles : the cacti or husimi graphs @xcite . ensembles of random graphs with only certain constraints on the vertex degrees are locally tree - like .
some well - studied examples with these characteristics are the ensemble of @xmath21-regular graphs ( also called bethe lattices ) and the ensemble of irregular erds - rnyi graphs @xcite . for undirected graphs ,
these ensembles can be formally defined as follows : * @xmath21-regular graphs with fixed connectivity @xmath21 : @xmath22 with @xmath23 the degree of the @xmath14-th vertex .
figure [ cavities1a ] presents a sketch of a bethe lattice . * erds - rnyi graphs with mean connectivity @xmath21 : @xmath24 in this ensemble
the degrees fluctuate from vertex to vertex within the graph .
the distribution of vertex degrees converges to a poissonian distribution with mean @xmath21 in its asymptotic limit @xmath25 .
ensembles of directed graphs can be defined similarly by leaving out the symmetry constraints in the definitions eqs ( [ eq : ensemble1 ] ) and ( [ eq : ensemble2 ] ) and by taking into consideration the difference between @xmath26 and @xmath27 .
for instance , we define a @xmath21-regular ensemble of directed graphs by the constraint @xmath28 for any @xmath14 , with the indegree defined as @xmath29 and the outdegree as @xmath30 .
the erds - rnyi and the @xmath21-regular ensemble have the local tree property , i.e. the probability to encounter a cycle of finite length in the local neighbourhood of a vertex becomes arbitrary small for @xmath31 @xcite .
the local tree property of random graphs is an important characteristic which has allowed to determine the spectra of sparse graphs in previous works @xcite .
this property is illustrated in figure [ cavities1a ] for the bethe lattice : cycles only appear at the leaves of the tree , where the vertex constraints have to be satisfied such that their typical length is of the order @xmath32 .
next , we define graph ensembles which contain many short cycles and are , therefore , not locally tree - like .
a cycle of length @xmath0 is defined by a set of nodes @xmath33 with an edge connecting each pair @xmath34 , for @xmath35 .
the set of all @xmath0-tuples is denoted by @xmath36 .
we define the following ensembles which are very similar to the husimi graphs in @xcite : * @xmath37-regular husimi graphs with fixed connectivity @xmath21 and fixed cycle length @xmath0 : @xmath38 where @xmath39 denotes now the number of @xmath0-cycles of @xmath8 incident to the @xmath14-th vertex . in figure [ cavities1 ] we show the typical local neighbourhood of an undirected ( 4,2)-regular husimi graph .
* irregular husimi graphs with fixed cycle length @xmath0 and mean loop connectivity @xmath21 : @xmath40 \times \prod_{i < j}\delta(a_{ij } ; a_{ji } ) \nonumber \\ & & \times \prod_{(i , j)}\delta\left(a_{ij}\sum_{(i , k_1,k_2,\cdots , k_{\ell-2 } , j)\in v^{\ell}_{ij}}a_{ik_1}\left(\prod^{\ell-3}_{n=1 } a_{k_n k_{n+1}}\right ) a_{k_{\ell-2}j } ; a_{ij } \right).\nonumber \label{eq : poiss3}\end{aligned}\ ] ] the set @xmath41 consists of all @xmath0-tuples @xmath42 .
the first factor gives a weight @xmath43 to each randomly drawn cycle of length @xmath0 , the second factor is the constraint which makes the graph undirected and the last factor is a constraint which takes into consideration that each edge must belong to exactly one cycle of length @xmath0 .
we remark that in our notation of husimi graphs the mean cycle connectivity @xmath21 is the average number of cycles of length @xmath0 connected to a certain vertex , while @xmath44 denotes the mean degree of each vertex .
note that for @xmath45 the ensembles of husimi graphs introduced here reduce to the corresponding ensembles of random graphs which are locally tree - like ( discussed in the previous subsection ) .
as before , we can define as well directed ensembles of husimi graphs by removing the symmetry constraint @xmath46 and taking into consideration the possible difference beween @xmath26 and @xmath27 .
for instance , for a regular husimi graph we set the mean vertex indegree and the mean vertex outdegree equal to @xmath44 . in figure [ cavities1 ] we present the neighbourhood of a vertex in a directed ( 3,2)-regular husimi graph .
husimi graphs are not locally tree - like since they are composed of cycles , but they do have an infinite - dimensional character . indeed , the local neighbourhood of a vertex contains different branches which are only connected by cycles of order @xmath32 , see figures [ cavities1a ] and [ cavities1 ] .
this property allows us to present an exact spectral analysis .
in this section we show how to determine the spectral properties of undirected graphs with cycles . for regular graphs , our approach , briefly presented in @xcite , consists of four steps .
first , we write the spectrum in statistical mechanics terms using a complex measure
. then we use the infinite - dimensional nature of the graphs to derive a closed expression in the submatrices of the resolvent .
next , we simplify the resolvent equations to the specific case of regular undirected husimi graphs . finally , we solve the resultant algebraic expressions to find analytical expressions for the spectra of regular husimi graphs .
we define the resolvent of a real symmetric matrix @xmath8 as @xmath47 where @xmath48 is the @xmath49 identity matrix and @xmath50 , with @xmath51 and @xmath52 . from the resolvent
we can determine the asymptotic density @xmath53 of the real eigenvalues @xmath54 of an ensemble of matrices @xmath8 @xcite : @xmath55 .
\label{eq : specres}\end{aligned}\ ] ] we have left out ensemble averages since we assume that the spectrum self - averages in the limit @xmath31 to a deterministic value .
a common method in statistical physics is to write the resolvent elements and the spectrum as averages over a certain complex - valued measure @xmath56 @xcite . indeed , when defining the normalized gaussian function @xmath57 the diagonal elements @xmath58_{ii } = g_{i}$ ] are given by @xcite @xmath59 with @xmath60 the average with respect to @xmath61 . in this formalism
the spectrum ( [ eq : specres ] ) is written as @xcite @xmath62 \ , , \label{eq : rhoenergy}\end{aligned}\ ] ] with @xmath63 such that @xmath64\rightarrow 0^+ } \frac{1}{\pi n } \sum^n_{i=1}{\rm{im}}g_i\left(z\right ) , \
\ \lambda = \rm{re}[z ] .
\label{eq : spectrumrec}\end{aligned}\ ] ] hence , in the language of statistical mechanics the spectrum is the imaginary part of the derivative of the free energy density with respect to @xmath54 , as is seen explicitly in eq.([eq : rhoenergy ] ) .
we employ the cavity method to determine the spectral properties of locally tree - like graphs @xcite .
we construct a matrix @xmath65 ( and its corresponding resolvent @xmath66 ) through deletion of the @xmath14-th column and the @xmath14-th row in @xmath8 .
the graph @xmath67 induced by the matrix @xmath65 is usually referred to as the cavity graph @xcite . on this graph
we use the bethe - peierls tree approximation @xcite to arrive at the following set of resolvent equations : @xmath68 where @xmath69 is the set of vertices neighbouring @xmath14 , and @xmath70 denotes the set of all neighbours of @xmath14 except for node @xmath15 . in appendix
[ app : treestruct ] we report the essential steps in deriving these resolvent equations .
the resolvent equations ( [ eq : resolv2 ] ) have appeared in the literature for the first time , as far as we know , in the study of electron localization @xcite .
similar equations have been derived in the context of lvy matrices @xcite and in the more general context of sparse random graphs @xcite .
the resolvent equations ( [ eq : resolv2 ] ) form a set of @xmath71 equations in @xmath71-unknowns .
we interpret the set of equations for @xmath72 as a message - passing algorithm along the induced graph @xmath73 , similar to belief propagation which is widely used in statistical inference and information theory @xcite . here
the messages are the diagonal elements of the resolvent of the cavity graph @xcite .
next , we extend the resolvent equations ( [ eq : resolv2 ] ) to husimi graphs with many short cycles .
although the bethe - peierls approximation is no longer valid on these graphs , they still contain an infinite - dimensional character allowing for an analogous approximation . in statistical mechanics
this forms the basis for the kikuchi approximation @xcite . for husimi graphs the kikuchi approximation gives us the resolvent equations : @xmath74 with @xmath75 and with @xmath76 the set of all @xmath77-tuples @xmath78 which form a cycle of length @xmath0 with the node @xmath14 .
the elements of the @xmath77-dimensional matrix @xmath79 are @xmath80_{nm } = \left[\bl_{(j_1 , j_2 , \cdots , j_{\ell-1})}\right]_{nm } = \delta_{n(m+1)}a_{j_nj_m}$ ] , with @xmath81 .
the quantity @xmath82 is a diagonal matrix with elements @xmath83_{kk } = \sum_{\beta\in \partial^{(\ell-1)}_{j_k}\setminus ( i , j_1,\ldots , j_{k-1 } , j_{k+1 } , \ldots , j_{\ell-1 } ) } \ba^t_{j_k , \beta } g^{(j_k)}_{\beta } \ba_{j_k , \beta } \ , , \nonumber \end{aligned}\ ] ] with @xmath84 .
we have also defined the @xmath77-dimensional vector @xmath85 .
the diagonal elements of the resolvent are given by @xmath86 in appendix [ app : cactusstruct ] we elaborate on the precise derivation of these equations .
the resolvent equations ( [ eq : reshus1 ] ) determine the spectra of graphs with many short cycles of fixed length @xmath0 . note that they can be straightforwardly generalized to graphs with variable cycle lengths and , even more general , to graphs composed from arbitrary figures , by replacing @xmath79 by the corresponding adjacency matrix of the figure @xmath87 in the absence of a node @xmath14 , and @xmath88 by the submatrix of @xmath8 connecting the figure @xmath89 and the node @xmath90 .
in our case , the figures of the graph are the @xmath77-tuples @xmath87 .
the set of resolvent equations ( [ eq : reshus1])-([eq : reshus2 ] ) can also be seen as a message - passing algorithm between regions of the graph , which is similar to the generalized belief propagation algorithm in information theory @xcite . here
the messages of the algorithm are the @xmath91 matrices @xmath92 sent from the @xmath77-tuples @xmath87 to a vertex @xmath14 , with whom the tuple @xmath87 forms a cycle of length @xmath0 in the graph .
we have verified the exactness of the resolvent equations ( [ eq : reshus1])-([eq : reshus2 ] ) on irregular husimi graphs with @xmath93 through a comparison with direct diagonalization results .
this is presented in figure [ fig : c6 ] , where we also compare the spectrum of irregular husimi graphs with mean cycle connectivity @xmath94 with the spectrum of erds - rnyi random graphs with the same mean vertex degree @xmath95 . from these results
it appears that the spectrum of erds - rnyi graphs converges faster to the wigner semicircle for @xmath96 .
finally , we point out that the ensemble definitions in section 2 are global by specifying @xmath19 . in the derivation of the resolvent equations ( [ eq : reshus1])-([eq : reshus2 ] )
these global definitions are not explicitely used .
instead , our analysis uses the typical local neighbourhoods of the vertices in the graph .
therefore , our results are valid for all graphs which have a local neighbourhood similar to the one given by the resolvent equations .
the connection between the global definitions and the resolvent equations can be made explicit through the distribution of local vertex neighbourhoods , see for instance @xcite .
this is how we have derived the results in figure [ fig : c6 ] . in the next two sections we solve the resolvent equations for the case of undirected regular husimi graphs for which they simplify considerably . in this section
we consider graphs with a simple topology that allows to extract exact analytical solutions from the resolvent equations ( [ eq : reshus1])-([eq : reshus2 ] ) : regular husimi graphs with cycles .
for this ensemble of random graphs @xmath97 and @xmath98 for all connected pairs @xmath99 . in figure [ cavities1 ] the square undirected husimi graph is illustrated .
since there is no local disorder every node has the same local neighbourhood .
for @xmath20 the graph becomes transitive and we can set @xmath100 in the resolvent equations ( [ eq : reshus1 ] ) .
we get the following closed equations in the @xmath77-dimensional matrices @xmath101 : @xmath102 and the resolvent follows from @xmath103 where @xmath104 .
it is convenient to introduce the scalar variable @xmath105 and rewrite ( [ reg ] ) as @xmath106 the matrix in ( [ reg1 ] ) has a simple structure .
we have calculated the inverse analytically using specific methods for tridiagonal matrices @xcite .
this leads to our final formula for @xmath107 when @xmath108 @xmath109 the coefficients @xmath110 are complex numbers given by the recurrence relation @xmath111 with initial values @xmath112 and @xmath113 . equations
( [ gs ] ) and ( [ alpha ] ) are a set of very simple equations that determine the resolvent elements of undirected regular husimi graphs .
it is one of the main results of our previous work @xcite . from the expression of @xmath107 we find straightforwardly the green function @xmath114 and the spectrum , according to eq .
( [ eq : spectrumrec ] ) and ( [ greensimp ] ) , @xmath115 we solve the algebraic equations ( [ gs])-([alpha ] ) to find the spectrum of @xmath37-regular undirected husimi graphs . due to their simple linear structure , these recurrence relations can be solved excactly and allow to obtain the analytical form of the polynomials in the variable @xmath107 for any value of @xmath116 : * @xmath93 corresponds with a quadratic equation : @xmath117 - g_s = 2 .
\nonumber \end{aligned}\ ] ] * @xmath118 gives a cubic equation : @xmath119 } { \left [ z - ( c-1 ) g_s \right]^2 -2}. \nonumber \end{aligned}\ ] ] * @xmath120 is determined by a quintic equation : @xmath121^{3 } - 4\left [ z - ( c-1 ) g_s \right ] + 2 } { \left [ z - ( c-1 ) g_s \right]^{4 } - 3 \left [ z - ( c-1 ) g_s \right]^{2 } + 1}. \nonumber \end{aligned}\ ] ] * @xmath122 follows from a quartic equation : @xmath121^{2 } - 4 } { \left [ z - ( c-1 ) g_s \right]^{3 } - 3\left [ z - ( c-1 ) g_s \right]}. \nonumber \end{aligned}\ ] ] * for @xmath123 , the polynomials have a degree larger than four . the root of these polynomial equations which gives the expression for the spectrum is a stable fixed point of eq .
( [ reg1 ] ) . in general , one finds the roots of polynomials in terms of radicals up to degree four . for larger degrees , algebraic solutions in terms of radicals no longer exist , apart from some particular situations @xcite .
the roots of general polynomials with degree larger than four are given in terms of elliptic functions @xcite .
we have solved the above equations for @xmath124 and have found the analytical expression for @xmath125 , generalizing the kesten - mckay law @xcite to regular graphs with cycles . for other values of @xmath0 , eqs .
( [ gs])-([alpha ] ) are solved numerically in a straightforward way , allowing to study the spectrum as a function of @xmath0 with high accuracy . from the solution of the resolvent equations ,
we have found the expressions for the spectra @xmath126 of @xmath127-regular undirected husimi graphs .
we point out that for @xmath45 we recover a sparse regular graph without short cycles . in this case
we have the kesten - mckay eigenvalue - density distribution @xcite @xmath128 for @xmath129 $ ] , and @xmath130 otherwise . the expression ( [ eq : km ] )
follows from solving eq .
( [ reg1 ] ) . for @xmath93
we recover the spectra @xmath131 of triangular husimi graphs @xcite
@xmath132 ^ 2 } + \frac{1}{3}\delta_{c,2}\:\delta(\lambda ; -2 ) , \nonumber\end{aligned}\ ] ] for @xmath133 , and @xmath134 otherwise .
therefore , the support of @xmath131 is @xmath135 $ ] . for @xmath118
we find the spectrum of a square husimi graph @xcite @xmath136^{2 } + 3 \ , \pi \ , c^{2 } \ , q_{-}^{2}(\lambda ) , } \nonumber\ ] ] for @xmath137 and @xmath138 otherwise , where @xmath139 and @xmath140 the edges of @xmath141 are determined by finding the roots of @xmath142 .
for @xmath143 , @xmath141 converges to the wigner semicircle law . for @xmath144 ,
the spectrum contains a power - law divergence @xmath145 as @xmath146 .
we have also found the analytical expression for @xmath122 : @xmath147 ^ 2 + \pi c^2 f(\lambda ) } , \nonumber \end{aligned}\ ] ] for @xmath148 and @xmath149 otherwise , with @xmath150}{3(c-1)^2 } , \nonumber \\
r(\lambda ) & = & \sqrt{u_{r}(\lambda ) - \frac{3 \lambda^2}{4 ( c-1)^2 } - \frac{(2c-5)}{(c-1)^2 } } , \nonumber \\
f(\lambda ) & = & -\frac{|\lambda|^3}{(c-1)^3 } + \frac{4(2c+1)|\lambda|}{(c-1)^3 } + 4 r(\lambda ) \left [ \frac{(2c-5)}{(c-1)^2 } - \frac{3 \lambda^2}{2 ( c-1)^2 } + u_r(\lambda ) \right],\nonumber \end{aligned}\ ] ] and @xmath151 \lambda^2 \nonumber \\ & & + \frac{16(2c-5)}{3}+ \frac{(2c-5)^3}{27(c-1 ) }
. \nonumber\end{aligned}\ ] ] the function @xmath152 is the discriminant of the cubic polynomial associated to the quartic polynomial for @xmath107 .
the edges of @xmath153 are obtained from the roots of @xmath154 .
the spectrum converges to the wigner semicircle law when @xmath155 . for @xmath120 and @xmath156
we obtain the spectrum by solving numerically the set of eqs .
( [ gs])-([alpha ] ) .
one can show that @xmath125 converges to the wigner law in the limit @xmath155 , and converges to the kesten - mckay law in the limit @xmath157 ( this limit corresponds to a graph without short cycles ) .
in figure [ fig:4ell ] we present the evolution of the spectrum as a function of the connectivity @xmath21 for the case @xmath118 .
we notice indeed the fast convergence to the wigner semi - circle law for @xmath96 . in figure [ oddell ]
we present the evolution of the spectrum as a function of the cycle length @xmath0 , for fixed @xmath144 .
we see how the spectrum converges rapidly to the kesten - mckay law for @xmath158 .
in our study of spectral properties of regular directed graphs with cycles we follow again four steps .
first , we write the spectrum in statistical mechanics terms @xcite .
this is done by mapping the resolvent calculation of a directed graph on a resolvent calculation of a related undirected graph , using the hermitization procedure @xcite .
the resolvent of this undirected graph is then analyzed using the methodology presented in section 3 .
thereby , the infinite - dimensional nature of the resultant graphs again allows to derive a closed expression in submatrices of the resolvent matrix .
thirdly , we show how this closed set of equations determines the spectrum of directed regular husimi graphs and we derive their explicit form in this case .
finally , the resultant algebraic expressions are solved to arrive at explicit analytical expressions in the support of the spectra .
the eigenvalues @xmath159 of the @xmath49 adjacency matrix @xmath8 of directed husimi graphs are distributed in the complex plane , contrary to the real eigenvalues for undirected husimi graphs . by introducing @xmath160 and using the relation @xmath161 , the density of states @xmath162 at a certain point @xmath163
can be written formally as @xmath164 , where @xmath165 is the resolvent @xmath166 .
the operation @xmath167 denotes complex conjugation .
the @xmath49 resolvent @xmath165 is the central object of interest and its non - analytic behavior at the eigenvalues of @xmath8 poses difficulties in applying various techniques well - developed for hermitian matrices @xcite .
an elegant way to avoid this problem is the hermitization method .
we define a @xmath168 block matrix @xmath169 where @xmath170 is a regularizer and @xmath171 is the @xmath172-dimensional identity matrix
. the lower - left block of @xmath173 is precisely the matrix @xmath165 .
thus , the problem reduces to calculating the diagonal matrix elements @xmath174_{j+n , j}$ ] ( @xmath175 ) , from which the spectrum is determined according to @xmath176 the form of the enlarged matrix @xmath177 depends on the problem at hand and different proposals have appeared in the literature @xcite .
the form ( [ defh ] ) is particularly convenient here , since its hermitian part is a positive - definite matrix and one can represent @xmath178 as a gaussian integral . for this purpose
we introduce a set of two - dimensional column vectors @xmath179 with complex elements and the `` hamiltonian '' function @xmath180 where @xmath181 and @xmath182 $ ] , with ( @xmath183 ) denoting pauli matrices .
a graphical representation using an induced graph is again useful in these calculations . for a non - hermitian matrix with real entries
the induced graph is directed , contrary to an undirected graph for real symmetric matrices . graphically the matrix elements @xmath26 correspond then with a directed edge from node @xmath14 to node @xmath15 . combining the representation of @xmath184_{ij}$ ] as a gaussian integral with eq .
( [ hamilt ] ) , @xmath185 reads @xmath186 \left [ \bphi^{\dagger}_{k}.(\sigma_{x } + i \sigma_{y})\bphi_{k } \right ] \mathcal{p } ( \ { \bphi_{j},\bphi^{\dagger}_{j } ; \lambda \ } ) \ , , \label{matrh}\end{aligned}\ ] ] where we have introduced the complex - valued measure @xmath187 @xmath188 } } { \int \left [ \prod_{j=1}^{n } d \bphi_{j } d
\bphi^{\dagger}_{j } \right ] \exp{\left
[ -\mathcal{h}(\ { \bphi_{i},\bphi^{\dagger}_{i } \ } ; \lambda)\right ] } } \,.\end{aligned}\ ] ] equation ( [ matrh ] ) shows that the local marginals @xmath189 determine the functions @xmath190 , from which the spectrum follows through eq .
( [ spectrgj ] ) .
we use again the infinite dimensional nature of husimi graphs to derive an exact equation in the resolvent elements . due to the sparse structure of @xmath8 ,
the average number of nodes in a path connecting two randomly chosen cycles scales as @xmath191 ( see figure [ fig : cavities ] ) .
this fundamental property allows to compute @xmath192 using the cavity method , as demonstrated in appendix a.2 .
the main difference with the undirected case resides in the `` state variables '' describing the nodes of the graph and the corresponding `` hamiltonian '' .
while in the undirected case they are scalar real variables , here there is a two - dimensional complex - vector @xmath193 associated to each vertex and they mutually interact through @xmath194 , see eq .
( [ hamilt ] ) .
hence the resolvent equations involve two - dimensional matrices . extending the cavity formulation @xcite to directed graphs with cycles , see appendix a.2 , we have derived the following equation @xmath195 where @xmath76 is the set of all @xmath77 tuples @xmath78 which form a cycle of length @xmath0 with node @xmath14 .
the @xmath196 block matrix @xmath197 encodes the interaction between @xmath14 and a given tuplet @xmath198 .
the matrix @xmath199 is a two - dimensional matrix filled with zeros .
the @xmath200 block matrices @xmath201 fulfill the cavity equations @xmath202 where @xmath203 and @xmath204 .
the matrix @xmath205 is composed of @xmath206 block elements defined by @xmath207_{nm } = [ \mathbb{l}_{(j_1 , j_2 , \cdots , j_{\ell-1})}]_{nm } = \delta_{n , m+1 } \ , \mathcal{j}_{j_{n},j_{m}}$ ] , where @xmath208 .
the matrix @xmath209 is a diagonal matrix formed by the following @xmath206 block elements @xmath210_{kk } = \sum_{\beta \in \partial^{(\ell-1)}_{j_k}\setminus ( i , j_1,\ldots , j_{k-1 } , j_{k+1 } , \ldots , j_{\ell-1 } ) } \mathbb { a}^{t}_{j_{k } \beta } \mathcal { d}_{\beta}^{(j_k ) } \mathbb{a}_{j_k \beta } , \label{matrb}\ ] ] with @xmath211 and @xmath212 .
once eqs .
( [ cavd ] ) have been solved , the spectrum follows from eqs .
( [ gj ] ) and ( [ spectrgj ] ) .
the cavity equations have an interpretation in terms of a message - passing algorithm : the matrix @xmath213 is seen as the message sent by the @xmath77 nodes of cycle @xmath87 to node @xmath14 of the same cycle @xcite .
this completes the general solution of the problem .
we determine now the resolvent equations for the spectrum @xmath214 of infinitely large @xmath37-regular directed husimi graphs .
these graphs have @xmath215 for @xmath203 , i.e. , each vertex is incident to @xmath21 cycles of length @xmath0 .
we set @xmath12 and @xmath216 when there is a directed edge from node @xmath14 to @xmath15 , such that the corresponding matrix @xmath194 assumes the form @xmath217 . as a consequence , the matrices @xmath218
become independent of the indices ( @xmath219 ) .
it is convenient to define the two - dimensional matrix @xmath220 , where @xmath221 .
we write @xmath53 in terms of @xmath222 as follows @xmath223_{21}^{-1}\ , .
\label{spectrcav}\ ] ] from eqs .
( [ cavd ] ) and ( [ matrb ] ) one obtains that , for @xmath108 , the two - dimensional matrix @xmath222 solves the equation @xmath224^{-1 } \mathbb{a } \ , , \label{eqd } \ ] ] where @xmath225 is a @xmath77-dimensional matrix with elements @xmath226_{i j } = \delta_{i , j+1}$ ] .
the derivative of eq .
( [ eqd ] ) yields an equation in @xmath227 , which has to be solved together with ( [ eqd ] ) to find @xmath125 through eq .
( [ spectrcav ] ) . equation ( [ eqd ] ) allows to derive accurate numerical results for the spectrum of directed husimi graphs as a function of @xmath0 . as an illustration , we present in figure [ cuts ] some cuts of @xmath131 and @xmath141 along the real direction for fixed values of @xmath228 .
these results correspond very well with direct diagonalization , confirming the exactness of ( [ eqd ] ) .
for a three - dimensional graph of @xmath229 we refer the reader to @xcite .
analogously to undirected husimi graphs , @xmath125 converges to the spectrum of a directed regular graph without short cycles for @xmath230 @xcite : @xmath231 for @xmath232 , and @xmath233 otherwise .
this convergence is shown numerically in @xcite . by rescaling @xmath234
, the solution of ( [ eqd ] ) leads to girko s law in the limit @xmath155 . in order to derive analytical equations for the support of @xmath125
, we determine the inverse of the @xmath200 matrix present in eq .
( [ eqd ] ) . since this matrix has a tridiagonal block structure its inverse can be computed analytically using the method in @xcite . applying this scheme to the matrix in eq .
( [ eqd ] ) , we have simplified ( [ eqd ] ) into an equation involving sums and products of only two - dimensional matrices .
this equation forms the equivalent for directed regular graphs of the equation ( [ gs ] ) for undirected regular graphs .
the resultant equations in @xmath235 can be solved using the following _ ansatz _
@xcite @xmath236 where @xmath237 is complex and @xmath238 and @xmath239 are both real variables . the hermitian part of this matrix is positive - definite provided that @xmath238 and @xmath239 are positive .
this condition ensures that the gaussian integrals in the cavity method are convergent @xcite . solving numerically eq .
( [ eqd ] ) for a finite regularizer @xmath170 , we find that @xmath240 and @xmath241 . in the limit
@xmath242 , @xmath238 and @xmath239 vanish at the boundary of @xmath53 .
therefore , setting @xmath243 and @xmath244 in eq .
( [ eqd ] ) , and solving the resulting equations for @xmath237 leads to an analytical expression for the support of @xmath245 . in this way
we obtain the following equations for @xmath237 and @xmath54 at the boundaries of the support : * @xmath246 : @xmath247^{2}}{|\lambda - ( c-1 ) b & |\lambda & - ( c-1 ) b |^{4 } = ( c-1 ) \left [ 1 + |\lambda - ( c-1 ) b |^{2 } \right ] .
\label{ell3polyn}\end{aligned}\ ] ] * @xmath248 : @xmath247^{2}}{|\lambda - ( c-1 ) b & |\lambda & - ( c-1 ) b |^{6 } = ( c-1 ) \left [ 1 + |\lambda - ( c-1 ) b |^{2 } + \label{ell4polyn}\end{aligned}\ ] ] the solutions of the polynomials ( [ ell3polyn ] ) and ( [ ell4polyn ] ) in the variable @xmath249 are given by * @xmath246 : @xmath250^{\frac{1}{2 } } \ , , \label{rell3}\ ] ] * @xmath248 : @xmath251^{\frac{1}{3 } } + [ r_{-}(c)]^{\frac{1}{3 } } + \frac{1}{3 } ( c-1 ) \right]^{\frac{1}{2 } } \ , , \label{rell4}\end{aligned}\ ] ] where we have defined @xmath252^{\frac{1}{2 } } \ , , \nonumber\end{aligned}\ ] ] with @xmath253 . by parametrizing @xmath254 , with @xmath255
$ ] , and substituting this form in eqs .
( [ ell3b ] ) and ( [ ell4b ] ) , we find the following expressions for the boundary of the support @xmath256 of triangular @xmath257 and square @xmath258 regular directed husimi graphs : * @xmath246 : @xmath259 * @xmath248 : @xmath260 these are the parametric equations which describe , for each corresponding @xmath0 , an hypotrochoid in the complex plane .
the parameter @xmath261 as a function of the cycle degree @xmath21 for @xmath93 and @xmath118 is given by , respectively , eqs .
( [ rell3 ] ) and ( [ rell4 ] ) .
a hypotrochoid is a cyclic function in the complex plane which is drawn by rotating a small circle of radius @xmath3 in a larger circle of radius @xmath2 @xcite .
the support of triangular and square husimi graphs is therefore given by hypotrochoids with , respectively , @xmath262 and @xmath263 .
these analytical results for the support of the spectra of directed husimi graphs for @xmath93 and @xmath118 are shown in the lower graphs of figure [ mozaic ] .
the agreement with direct diagonalization results for @xmath144 is excellent , confirming the exactness of our analytical results .
-regular husimi graphs for several values of the cycle length @xmath0 and the following values @xmath21 of the number of cycles incident to each vertex : @xmath144 ( solid line ) , @xmath94 ( dashed line ) , @xmath264 ( dotted line ) and @xmath265 ( dot - dashed line ) .
the hypotrochoids have a rotational symmetry by the angle @xmath266 , from which one obtains the value of the cycle length @xmath0 .
direct diagonalization results ( dots ) for @xmath49 matrices with @xmath267 and @xmath144 are shown . ]
based on the form of eqs .
( [ ell3b]-[ell4polyn ] ) , we conjecture that , for a given @xmath0 and @xmath21 , the following equations are fulfilled at the boundary of the support of @xmath245 @xmath268^{\ell-1}}{s^{2(\ell-1 ) } } \ , , \label{eqbanyell } \\
s^{2(\ell-1 ) } & = & ( c-1 ) \sum_{n=0}^{\ell-2 } s^{2n } \ , . \label{eqranyell}\end{aligned}\ ] ] substituting @xmath269 ( @xmath255 $ ] ) in eq .
( [ eqbanyell ] ) reads @xmath270 } \ , \label{hypogen}\ ] ] for the boundary of the support of a @xmath127 directed regular husimi graph .
remarkably , eq . ( [ hypogen ] ) is a hypotrochoid with a fraction @xmath1 .
the parameter @xmath261 is determined from the roots of a polynomial of degree @xmath0 in the variable @xmath271 , see eq .
( [ eqranyell ] ) .
equation ( [ eqranyell ] ) can also be written as @xmath272 we have found an analytical expression for the roots for @xmath93 and @xmath118 . for larger values of @xmath0
, we have solved eq .
( [ simple ] ) numerically and , by choosing the stable solution , we have derived accurate values for the parameters of the hypotrochoids .
we present explicit results for @xmath120 and @xmath122 in the upper graphs of figure [ mozaic ] .
direct diagonalization results exhibit once more an excellent agreement with the theoretical results , strongly supporting our conjecture that the support of the spectrum of directed regular husimi graphs for general @xmath0 is given by eqs .
( [ hypogen]-[simple ] ) .
the support of regular husimi graphs converges to the circle @xmath273 in the limit @xmath274 , corresponding with the expression ( [ spectruml2 ] ) for a graph without short cycles .
in this work we have obtained the spectra of ( un)directed husimi graphs .
the main result is a set of exact equations which determines a belief - propagation like algorithm in the resolvent elements of the matrix . for irregular graphs
we have shown a very good correspondence between direct diagonalization results and our approach . for regular graphs
we have derived several novel analytical expressions for the spectrum of undirected husimi graphs and the boundaries of the spectrum of directed husimi graphs . remarkably , the boundaries of directed regular husimi graphs consist of hypotrochoid functions in the complex plane .
our results indicate that , at high connectivities , the spectrum of undirected random graphs converges to the wigner semicircle law , while the spectrum of directed random graphs converges to girko s circular law .
this convergence seems to be rather universal and independent of the specific graph topology .
it would be interesting to better understand the conditions under which finitely connected graphs converge to these limiting laws @xcite .
finally , we point out that the eigenvalues of random unistochastic matrices are distributed over hypocycloids in the complex plane @xcite .
this close similarity with our results suggests an interesting connection between the spectra of @xmath275 unistochastic matrices and regular directed husimi graphs with cycles of length @xmath0 .
this paper is dedicated to fritz gesztesy , on the occasion of his 60th birthday .
db wants to thank fritz not only for many years of stimulating and fruitful collaborations but especially for a lifetime friendship !
flm is indebted to karol yczkowski for illuminating discussions .
we present the essential steps to determine the resolvent equations using the cavity method @xcite .
this method is based on the introduction of cavities in a graph @xmath73 forming subgraphs @xmath67 , where the node @xmath14 and all of its incident edges have been removed , see figure [ fig : cavities ] . in analogy
one can also remove the i - th column and the i - th row from a matrix @xmath8 to obtain the submatrix @xmath65 .
the cavity method is based on the consideration that a probability distribution @xmath61 defined on a locally tree - like graph has the factorization property : @xmath276 the quantity @xmath277 is the @xmath14-th marginal of @xmath278 on the cavity subgraph @xmath279 of @xmath73 , and @xmath280 is the marginal of @xmath281 with respect to the set of variables @xmath282 . in the language of spin models , the factorization property follows from the locally tree - like structure of a typical neighbourhood in the graph , see figure [ fig : cavities ] . following the derivation
as presented in @xcite , we find a set of closed equations in the marginals @xmath277 @xmath283 from which the marginals @xmath284 follow : @xmath285 finally , we use the fact that the @xmath277 are gaussian functions @xmath286 to recover the resolvent equations ( [ eq : resolv2 ] ) , after substitution of ( [ eq : ansatz ] ) in ( [ eq : marg ] ) and ( [ eq : marg2 ] ) .
we apply now a similar logic to graphs with many cycles which have an infinite - dimensional structure . in this case
we use the factorization property @xmath287 on the marginals of the distribution @xmath61 .
the factorization property follows from the fact that the average distance between different branches connected to @xmath14 is of the order @xmath191 after removal of @xmath14 , see figure [ fig : cavities ] .
note that @xmath288 is the marginal of @xmath61 for a @xmath77 tuple @xmath289 forming a cycle with the @xmath14-th vertex .
we have generalized the derivation for regular husimi graphs presented in @xcite to the case of arbitrary husimi graphs .
we find a set of closed equations in the marginals @xmath288 : @xmath290 \nonumber \\ & & \times \prod_{j\in \alpha } \left[\prod_{\beta \in \partial^{(\ell-1)}_j\setminus \alpha } \exp\left(i\sum_{j\in \alpha}\sum_{\beta\in \partial^{(\ell)}_j\setminus \alpha}\left ( a_{j j^{(\beta)}_1}\:\ : x_{j^{(\beta)}_1}x_{j } + a_{j^{(\beta)}_{\ell-1}j}\:\ : x_{j^{(\beta)}_{\ell-1}}x_{j } \right)\right)\right ] , \nonumber \end{aligned}\ ] ] with the @xmath77-tuple @xmath291 .
the marginals @xmath284 are given as a function of the marginals @xmath288 : @xmath292 we now use the gaussian ansatz for the marginals @xmath293 : @xmath294 with @xmath295 the @xmath91 submatrix of the resolvent @xmath66 of the cavity matrix @xmath65 : @xmath296 we also use a gaussian ansatz of the type ( [ eq : ansatz ] ) for the marginal @xmath284 .
substitution of these anstze in ( [ eq : marginalst ] ) and ( [ eq : marginalstt ] ) gives the resolvent equations ( [ eq : reshus1 ] ) and ( [ eq : reshus2 ] ) .
a b. mohar , _ some applications of laplace eigenvalues of graphs _ , graph symmetry : algebraic methods and applications , vol .
497 kluwer ( 1997 ) , 227 - 275 .
f. r. k. chung , _ spectral graph theory _
, american mathematical society ( 1997 ) a. e. brouwer , w. h. haemers , spectra of graphs , springer ( 2010 ) .
b. mohar , w. woess , _ a survey on spectra of infinite graphs _
london math .
* 21 * ( 1989 ) , 209 - 234 .
a. barrat , m. barthlemy , a. vespignani , _ dynamical processes on networks _ , cambridge university press ( 2008 ) m. e. j. newman , _ networks , an introduction _ , oxford university press ( 2010 ) .
m. baharona , l. m. pecora , _ synchronization in small - world systems _ , phys . rev
* 89 * ( 2002 ) , 054101 .
s. f. edwards , d. r. wilkinson , _ the surface statistics of granular aggregate _
london , ser .
a * 381 * , 17 ( 1982 ) ; b. kozma , m. b. hastings , g. korniss , _
roughness scaling for edwards - anderson relaxation in small - world networks _ , phys .
lett . * 92 * , 108701 ( 2004 ) .
d. j. klein , r. randi , _ resistance distance _ , j. math .
* 12 * ( 1993 ) , 81 - 95 .
s. hoory , n. linial , a. wigderson , _ expander graphs and their applications _ , bull .
* 43 * ( 2006 ) , 439 - 561 s. f. edwards , r. c. jones , _ the eigenvalue spectrum of a large symmetric random matrix _ ,
a * 9 * ( 1976 ) , 1595 - 1603 .
r. khn , _ spectra of sparse random matrices _ ,
a : math . theor .
* 41 * ( 2008 ) , 295002 .
t. rogers , k. takeda , i. p. castillo and r. khn , _ cavity approach to the spectral density of sparse symmetric random matrices _ , phys
e. * 78 * ( 2008 ) , 031116 .
t. rogers , i. p. castillo , _ cavity approach to the spectral density of non - hermitian sparse matrices _ ,
e * 79 * ( 2009 ) , 012101 .
y. v. fyodorov , a. d. mirlin , _ on the density of states of sparse random matrices _ ,
a * 24 * ( 1991 ) , 2219 - 2223 .
t. rogers , c. p. vicente , k. takeda , i. p. castillo , _ spectral density of random graphs with topological constraints _ , j. phys .
a : math . theor .
* 43 * ( 2010 ) , 195002 .
r. khn , j. m. van mourik , _ spectra of modular and small - world matrices _ ,
j. phys . a : math .
, * 44 * ( 2011 ) , 165205 y. kabashima , h. takahashi , o. watanabe , _
cavity approach to the first eigenvalue problem in a family of symmetric random sparse matrices _
, j.phys . :
* 233 * ( 2010 ) , 012001 ; y. kabashima , h. takahashi , _ first eigenvalue / eigenvector in sparse random symmetric matrices : influences of degree fluctuation _ ,
j. phys . a : math
* 45 * ( 2012 ) , 325001 c. bordenave , m. lelarge , _ resolvent of large random graphs _ , random structures and algorithms * 37 * ( 2010 ) , 332 .
c. bordernave , d. chafai , _ around the circular law , probability surveys _ , * 9 * ( 2012 ) , 1 - 89 .
t. tao , v. vu .
, _ random matrices : the circular law _ , commun . contemp .
math.,*10 * , 261 - 307 ( 2008 ) .
t. tao , v. vu , m. krishnapur , _ universality of esds and the circular law _ , ann .
probab . * 38 * , 2023- 2065 ( 2010 ) .
f. gtze , a. tikhomirov , _ the circular law for random matrices _ ,
probab . , * 38 * , 1444- 1491 ( 2010 ) .
p. m. wood , _ universality and the circular law for sparse random matrices _ ,
* 22 * , 1266 ( 2012 ) . c. bordenave , p. caputo , d. chafai , _ spectrum of markov generators on sparse random graphs _ , arxiv:1202.0644 ( 2012 ) .
p. m. gleiss , p. f. stadler , a. wagner , d. a. fell , _ small cycles in small worlds _ , arxiv : cond - mat/0009124 ( 2000 ) . g. bianconi , g. caldarelli , a. capocci , _ loops structure of the internet at the autonomous system level _ ,
e * 71 * ( 2005 ) , 066116 .
g. bianconi , n. gulbahce , a. e. motter , _ local structure of directed networks _ ,
* 100 * ( 2008 ) , 118701 .
k. husimi , _ note on mayers theory of cluster integrals _ , j. chem .
* 18 * ( 1950 ) , 682 - 684 .
f. harary , g. uhlenbeck , _ on the number of husimi trees i _ ,
* 39 * ( 1953 ) , 315 - 322 .
i. neri , f. l. metz , _ spectra of sparse non - hermitian random matrices : an analytical solution _
rev . lett .
* 109 * ( 2012 ) , 030602 f. l. metz , i. neri , d. boll , _ spectra of sparse regular graphs with loops _
, phys . rev .
e * 84 * ( 2011 ) , 055101(r ) .
b. d. mckay , _ the expected eigenvalue distribution of a large regular graph _
, linear algebra appl . * 40 * ( 1981 ) , 203 - 216 .
o. khorunzhy , m. shcherbina , v. vengerovsky , _ eigenvalue distribution of large weighted random graphs _ , j. math
* 45 * ( 2004 ) , 1648 - 1672 .
b. bollobs , _ random graphs _ , cambridge university press ( 2001 ) . f. l. metz , i. neri , d. boll , _ localization transition in symmetric random matrices _ ,
e * 82 * ( 2010 ) , 03115 .
m. mzard , a. montanari , _ information , physics and computation _ , oxford university press ( 2009 ) .
h. a. bethe , _ statistical theory of superlattices _ , proc .
london ser a * 150 * ( 1935 ) , 552 - 575 .
r. abou - chacra , d. j. thouless , p. w. anderson , _ a selfconsistent theory of localization _ , j. phys .
c : solid state phys .
* 6 * ( 1973 ) , 1734 p. cizeau , j. p. bouchaud,_theory of lvy matrices _ , phys .
e * 50 * ( 1994 ) , 1810 - 1822 .
p. judea , _ probabilistic reasoning in intelligent systems : networks of plausible inference _ , morgan kaufmann ( 1988 ) , san francisco , ca .
r. kikuchi , _ a theory of cooperative phenomena _ , phys .
* 81 * ( 1951 ) , 988 - 1003 w. t. yedidia , j. s. freeman , y. weiss , _ bethe free energy , kikuchi approximations , and belief propagation algorithms _ , technical report tr-2001 - 16 , mitsubishi electric reseach , 2001 . y. huang , w. f. mccoll , _ analytic inversion of general tridiagonal matrices _ ,
a : math . gen .
* 30 * ( 1997 ) , 7917 .
r. b. king , _ beyond the quartic equation _
, birkhuser boston , 1996 . h. kesten , _ symmetric random walks on groups _ , trans .
soc . * 92 * , ( 1959 ) 336354 .
m. eckstein , m. kollar , k. byczuk , d. vollhardt , _ hopping on the bethe lattice : exact results for density of states and dynamical mean - field theory _ ,
b * 71 * , 235119 ( 2005 ) ; m. galiceanu , a. blumen , _ spectra of husimi cacti : exact results and applications _ , j. chem .
phys.*127 * , 134904 ( 2007 ) .
j. feinberg , a. zee , _ non - hermitian random matrix theory : method of hermitian reduction _
b * 504 * , 579 ( 1997 ) .
j. t. chalker , z. j. wang , _ diffusion in a random velocity field : spectral properties of a non - hermitian fokker - planck operator _
* 79 * , 1797 ( 1997 ) .
j. t. chalker , b. mehlig , _ eigenvector statistics in non - hermitian random matrix ensembles _ , phys .
81 * , 3367 ( 1998 ) .
r. a. janik , w. nrenberg , m. a. nowak , g.papp , i. zahed , _ correlation of eigenvectors for non - hermitian random matrix models _ , phys .
e * 60 * , 2699 ( 1999 ) .
t.rogers , _ new results on the spectral density of random matrices _ , thesis ( 2010 )
. j. d. lawrence , _ a catalog of special plane curves _
, new york : dover , pp .
165 - 168 , ( 1972 ) .
i. dumitriu , s. pal , _ sparse regular random graphs : spectral density and eigenvectors _ , arxiv:0910.5306 .
l. tran , v. vu , k. wang , _ sparse random graphs : eigenvalues and eigenvectors _ , arxiv:1011.6646 k. yczkowski , m. kus , w. somczynski and h .- j .
sommers , _ random unistochastic matrices _
, j. phys a : math .
gen . * 36 * , 3425 ( 2003 ) .
m. mzard , g. parisi , m.virasoro , _ spin glass theory and beyond _ , world scientific ( 1986 ) . | we present a general method for obtaining the spectra of large graphs with short cycles using ideas from statistical mechanics of disordered systems .
this approach leads to an algorithm that determines the spectra of graphs up to a high accuracy .
in particular , for ( un)directed regular graphs with cycles of arbitrary length we derive exact and simple equations for the resolvent of the associated adjacency matrix . solving these equations
we obtain analytical formulas for the spectra and the boundaries of their support . |
the thermal - like surface radiation detected recently from a number of isolated neutron stars ( nss ) ( see , e. g. , gelman 1995 for a review ) provides valuable information about the ns temperatures and radii .
this enables one to trace the thermal history of nss and to put constraints on the properties of the superdense matter in ns interiors ( e. g. , pethick 1992 ) .
however , the observed spectra are substantially modified by reprocessing of the thermal radiation in ns atmospheres ( romani 1987 ) .
pavlov et al .
( 1995 ; and references therein ) have shown that the strong magnetic fields of nss , @xmath6 g , affect the emergent spectra significantly .
in particular , photoabsorption by the strongly magnetized atomic hydrogen gas which presumably covers the ns surface should be taken into account .
the magnetic field strongly affects the structure and radiative transitions of the hydrogen atom when the electron cyclotron energy @xmath7 is much greater than the rydberg energy @xmath8 ev , i. e. , @xmath9 .
in particular , the atom becomes stretched along the magnetic field , the ionization potential is increased by a factor of @xmath10 , and radiative transitions become strongly dependent on the photon polarization .
photoionization of strongly magnetized hydrogen atoms _ at rest _ has been considered by a number of authors . the most thorough consideration in the _ adiabatic approximation _
, applicable for very large @xmath11 and @xmath12 ( @xmath13 is the photon frequency ) , was presented by potekhin & pavlov ( 1993 ; hereafter pp93 ) .
this approximation neglects coupling of states with different values of the landau quantum number @xmath14 which specifies the energy of the electron motion transverse to the magnetic field ( the other quantum number associated with the transverse motion , the projection @xmath15 of the angular momentum onto the field , is exact for atoms at rest ) .
potekhin , pavlov , & ventura ( 1996 ; hereafter ppv97 ) went beyond the adiabatic approximation and resolved , in particular , a longstanding contradiction : some authors ( e. g. , schmitt et al .
1981 ; wunner et al . 1983
; ventura et al .
1992 ) concluded that the photoionization cross section is exactly zero for photons polarized transversely to the magnetic field , whereas other authors ( e. g. , gnedin , pavlov , & tsygan 1973 ; miller & neuhauser 1991 ; pp93 ) obtained finite cross sections .
pp93 showed that the discrepancy rooted in different forms of the interaction potential ( so - called `` velocity form '' and `` length form '' ) employed by the two groups of authors , and ppv97 proved that the velocity form used by the former group resulted in missing the main contribution to the cross section .
the nonadiabatic approach enabled ppv97 to study correlation of different ionization channels and manifestation of autoionization ( beutler fano ) resonances in the photoabsorption spectra .
the above - cited papers did not take into account the atomic motion inevitable in any real situation .
if there were no strong magnetic field , the motion would only lead to the trivial doppler broadening of the spectral lines and photoionization edges . on the contrary
, the finiteness of the atomic velocities in strong magnetic fields leads to many qualitatively new effects ( pavlov & mszros 1993 ; hereafter pm93 ) associated with strong distortion of the atomic structure due the lorentz electric field generated by the motion . since the moving atom loses its cylindrical symmetry ,
the projection @xmath15 of the angular momentum is no longer a good quantum number ( in other words , the @xmath15-adiabatic approach which neglects coupling of states with different @xmath15 becomes invalid ) .
hence , the selection rules for the radiative transitions are changed so that new types of transitions become allowed . since the atomic structure depends on the velocity of the atom , and different atomic states
are affected by the motion differently , a new type of broadening of spectral features , _
magnetic broadening _
, becomes the most important .
the energies and wave functions of the moving atom depend on the transverse component @xmath2 of a _
generalized momentum _
@xmath3 ( see below ) , which , contrary to the canonical and kinetic momenta , is conserved for a neutral atom moving in a magnetic field .
pm93 analyzed various effects of the motion of the hydrogen atom in the perturbative regime , when @xmath2 is small .
when @xmath2 exceeds some critical value , the magnetic and ( motion - generated ) electric fields form a potential well at a distance @xmath16 from the proton , which gives rise to so - called _ decentered _ ( or shifted ) states ( burkova et al . 1976 ) .
the decentered states recover the cylindrical symmetry ( @xmath15 becomes a good quantum number again at large @xmath2 ) , but the axis of symmetry , with respect to which @xmath15 is defined , is shifted by @xmath17 . in other words , these states can be described in the _ shifted _
@xmath15-adiabatic approximation .
binding energies , wavefunctions and oscillator strengths of radiative transitions between the discrete levels of an arbitrarily moving h atom ( beyond the adiabatic approximation ) have been studied by potekhin ( 1994 ; hereafter p94 ) .
our recent study of the hydrogen bound - bound opacity in ns atmospheres ( pavlov & potekhin 1995 ; hereafter pp95 ) was based on these results .
less investigated are effects of the motion across the field on the bound - free radiative transitions of atoms .
pm93 have shown , in the perturbation approach , that the motion strongly affects photoionization cross sections through opening and enhancement of additional photoionization channels forbidden for the non - moving atom due to its cylindrical symmetry , and through the strong magnetic broadening of the photoionization edges .
bezchastnov & potekhin ( 1994 ; hereafter bp94 ) suggested a modification of the adiabatic approach which allows for different displacements of the initial and final states and presented a convenient formalism for calculating the transition rates between such states .
they have also performed calculations with the initial state treated non - adiabatically ; the two approaches exhibited a fair agreement with each other .
kopidakis , ventura & herold ( 1996 ; hereafter kvh96 ) also went beyond the perturbation approximation and calculated the frequency dependences of the cross sections for various values of @xmath2 , assuming these values small enough to neglect the decentered bound states .
the results of the two latter papers differ from each other by orders of magnitude for radiation polarized _ transversely _ to the field .
the origin of the discrepancy is the same as for atoms at rest kvh96 employed the velocity form of the interaction potential which led to wrong results when used with the @xmath14-adiabatic approximation which treats the landau number @xmath14 as the exact quantum number .
both bp94 and kvh96 treated final ( continuum ) states of electron in the ( shifted ) adiabatic approximation , neglecting coupling of final states with different @xmath15 and contribution of quasi - bound ( autoionizing ) states to the photoionization cross section .
the general non - adiabatic numerical approach developed by ppv97 for the continuum wave functions of the strongly magnetized hydrogen has been applied there to the particular case of non - moving atoms . in the present work
, we employ that approach , together with the theoretical formalism of p94 and bp94 , for an accurate treatment of both discrete and continuum states of the moving atom and the radiative transitions between them .
unlike bp94 and kvh96 , we do not restrict ourselves to the case of relatively small generalized momenta of atoms , but consider also the decentered states which arise at large @xmath2 .
this enables us to calculate the cross sections of photoionization by arbitrarily polarized photons of the hydrogen atom arbitrarily moving in a magnetic field @xmath6 g. averaging these cross sections with thermal distributions of atoms yields the bound - free opacities needed for modeling ns atmospheres .
in section 2 we outline basic techniques for numerical treatment of a strongly magnetized h atom and its interactions with radiation ( the general form of the interaction matrix elements is given in the appendix ) . in section 3
we present the photoionization cross sections of h atoms in various states of motion across the magnetic field and cross sections averaged over thermal motion for conditions typical for atmospheres of isolated nss .
the center - of - mass motion of an atom in a magnetic field can be conveniently described by the generalized momentum @xmath3 ( gorkov & dzyaloshinskii 1968 ; see also johnson , hirschfelder , & yang 1983 ) , @xmath18 where @xmath19 is the canonical momentum , @xmath20 is the relative coordinate , the subscript `` e '' or `` p '' indicates electron or proton , respectively , and the cylindric gauge of the vector potential , @xmath21 , is used .
we shall consider the representation in which all three components of @xmath3 have definite values .
then the two - body wavefunction can be factorized into a phase factor depending on the center - of - mass coordinate and the wavefunction @xmath22 depending on @xmath20 .
it is useful to define a basic electron - proton separation @xmath23 and to regard the deviation from it , @xmath24 , as an independent variable .
here @xmath25 is the position of the relative guiding center , and @xmath26 is the _ shift parameter_. it is also convenient to use a special axial gauge of the vector potential ( vincke , le dourneuf , & baye 1992 ; p94 ) .
then we arrive at a set of hamiltonians @xmath27 and wavefunctions @xmath28 depending on @xmath26 . in the particular cases of @xmath29 and @xmath30 ,
the `` conventional '' and `` shifted '' representations of gorkov & dzyaloshinskii ( 1968 ) are recovered .
wavefunctions with different @xmath26 are related to each other by the phase transformation ( bp94 ) @xmath31\ , \psi^{(0)}_{{\mbox{\scriptsize\boldmath $ k$}}}({\mbox{\boldmath $ r$}}+\eta{\mbox{\boldmath $ r$}}_c ) , \label{transform}\ ] ] where @xmath32 is the coordinate transverse to the field @xmath33 , and @xmath34 is the total mass of the atom . if there were no coulomb attraction , then the electron landau number @xmath35 and the @xmath36-projection of the angular momentum of the relative motion @xmath37 would be exact quantum numbers . in this case the energy of the transverse motion ( with the constant minimum energy subtracted ) is @xmath38 \hbar\omega_c~,\ ] ] and the transverse part of the wave function can be described ( for a given @xmath1 ) by a landau function @xmath39 which depends on @xmath40 ( @xmath41 ) .
here @xmath42 is the azimuthal angle of @xmath43 , @xmath44 is the magnetic length , @xmath45 is a laguerre function ( sokolov & ternov 1968 ) , and the @xmath36-projection @xmath15 of the angular momentum of the relative motion is defined in the shifted reference frame , @xmath30 ( since @xmath1 is definite , the electron and proton do not possess definite @xmath36-projections of the angular momenta separately from each other see johnson et al . 1983 ) .
equation ( [ landau ] ) implies that after photoionization of a moving hydrogen atom , when the photoelectron moves away to large distances at which the coulomb interaction can be neglected , it acquires the displacement @xmath46 from the proton in the transverse plane .
a wavefunction @xmath47 of an atomic state @xmath48 can be expanded over the complete set of the landau functions @xmath49 ( note that @xmath50 depends on @xmath26 ) .
the adiabatic approximation would correspond to retaining only one term in this expansion .
a _ bound state _ of the atom can be numbered as @xmath51 , where @xmath52 and @xmath53 relate to the leading term of the expansion ( [ expan ] ) , and @xmath54 enumerates longitudinal energy levels @xmath55 and controls the @xmath36-parity : @xmath56 . for a moving atom
, this way of numbering is unambiguous at @xmath57 ( p94 ) . since the transverse factors @xmath58 in equation ( [ expan ] )
are known analytically , only the one - dimensional longitudinal functions @xmath59 are to be found numerically . for the bound states , numerical calculations ( vincke et al .
1992 ; p94 ) yield the following . at small transverse generalized momenta @xmath60 , the states @xmath61 remain tightly bound and _ centered _ , with the electron cloud concentrated around the proton . for such states , the conventional representation ( @xmath29 )
is appropriate . at large @xmath60 ,
the states are _ decentered _ , with the electron localized in a potential well shifted apart from the proton .
then @xmath30 is the apt choice . with growing @xmath2
, the transition from the centered tightly - bound states to the decentered states occurs in a narrow range of @xmath2 near @xmath62 , and the shifted ( @xmath30 ) adiabatic approximation becomes fairly good at @xmath63 . for the hydrogen - like states
@xmath64 , however , the mean electron - proton separation grows steadily , being close to @xmath17 at arbitrarily small @xmath60 , so that these states can be described by the shifted adiabatic approximattion at any @xmath60 .
finally , at very large @xmath60 ( @xmath65 , where @xmath66 is the bohr radius ) the longitudinal functions become oscillator - like for any states , corresponding to a wide , shallow parabolic potential well of a depth @xmath67 below the continuum boundary ( burkova et al . 1976 ; p94 ) .
a _ continuum state _ can be numbered as @xmath68 , where @xmath69 and @xmath70 correspond to an open channel for the energy @xmath71 , and @xmath72 corresponds to a type of the final state e. g. , to a given @xmath36-parity or to a direction of electron escape , depending on asymptotic conditions . in the photoionization process
the value of @xmath73 is constrained by the momentum conservation law : @xmath74 , where @xmath75 is a photon wavevector . for the continuum without motion ( @xmath76 ) ,
a technique for calculating the longitudinal wavefunctions has been developed by ppv97 .
it can be easily extended to the general case , @xmath77 , by numerating the orbitals @xmath78 in equation ( [ expan ] ) for the state @xmath79 with single integers @xmath80 such that larger values of @xmath80 correspond to higher energies @xmath81 .
since the coulomb interaction ceases to distort the transverse wavefunctions at infinity , the orbitals in equation ( [ expan ] ) become uncoupled at large @xmath36 provided that the full - shift representation ( @xmath82 ) of the continuum state is chosen .
we shall adopt this choice hereafter .
the cross section for ionizing an atomic state @xmath83 into a continuum state @xmath79 by a photon with frequency @xmath13 , wavevector @xmath75 , and polarization vector @xmath84 is @xmath85 where @xmath86 is the reduced mass , @xmath87 is the @xmath36-extension of the periodicity volume of the final state , and @xmath88 is the dimensionless interaction operator . for an arbitrarily moving atom of finite mass , the expression for @xmath88
is given in the appendix .
using equation ( [ expan ] ) for both the initial and final atomic states , we obtain @xmath89 where @xmath90 corresponds to the longitudinal wavefunction @xmath91 , and @xmath92 corresponds to @xmath58 .
the interaction operator @xmath88 in this equation depends on @xmath93 , because the common phase factor ( eq . [ [ transform ] ] ) has been included in it ( see the appendix ) .
the landau functions for the initial and final states in equation ( [ m_expan ] ) depend on differently displaced arguments if @xmath94 .
this is the case for the tightly bound initial states at @xmath95 , for which the apt choice of the shift parameter is @xmath29 ( while @xmath82 ) .
the different displacements strongly complicate explicit integration in the inner ( transverse ) matrix elements . in order to reduce the initial and final wavefunctions in the same basis
, kvh96 expressed the displaced landau state as a series over the undisplaced states .
this reexpansion involves an additional summation index which runs over a huge number of values if the relative displacement is large enough , which renders the reexpansion method impractical at high @xmath2 .
the difficulty can be circumvented using the creation - annihilation operator formalism of bp94 .
analytical calculation of @xmath96 is performed in the appendix . as a result
, the interaction matrix element @xmath97 reduces to a sum of one - dimensional quadratures ( eq.[[m_final ] ] ) feasible for numerical evaluation .
many properties of the photoionization cross section can be understood from the @xmath2-dependence of the level energies shown in figure [ fig1 ] .
the monotonic increase of the energies shifts the photoionization thresholds redward .
admixture of many orbitals near `` anticrossings '' of levels belonging to different @xmath15-manifolds ( e. g. , at @xmath99 a.u . ) causes peculiarities in the radiative transition rates ( p94 ) .
the levels which lie above zero in figure [ fig1 ] correspond to quasi - bound states in the continuum .
they give rise to sharp resonances in the cross sections at the photon energies corresponding to transitions between the initial bound state and the quasi - bound states .
figure [ fig2 ] demonstrates a comparison of our numerical results ( solid lines ) with those obtained using approximations encountered in literature .
the short - dash - dot curves are obtained using the velocity form of the interaction operator with the @xmath100-coupling `` switched - off '' ( @xmath100-adiabatic approximation ) .
we see that this coupling is very important for evaluating the cross section @xmath101 ( for the left circular polarization transverse to the magnetic field ) and less important for the other two basic cross sections , @xmath102 ( right circular polarization transverse to the field ) and @xmath103 ( linear polarization along the field ) .
the dotted lines are obtained using the shifted adiabatic approximation for the final state and the length representation of the interaction operator .
note that they are close to the short - dash - dot lines in the upper panels .
this is in accordance with approximate equivalence of using the @xmath14-adiabatic approximation ( @xmath104 ) with the length form and using the velocity form with allowance for @xmath105 .
this equivalence has been suggested by pp93 and proved by ppv97 under the condition that the non - adiabatic corrections were small .
it fails for @xmath106 if the @xmath100-coupling is neglected ( compare dotted and dot - short - dash curves in the lower panels ) , but we have checked numerically that it holds if the @xmath107-coupling is included .
the dot - long - dash curves correspond to the approach of bp94 .
it differs from the plain shifted @xmath108-adiabatic approximation ( dotted curves ) by employing the orthogonalization of the final state with respect to the initial one ( see eq . [ 32 ] of bp94 ) . at relatively small @xmath2 ( upper panels ) , this approximation yields qualitatively correct results . with increasing @xmath2 ,
the approximate @xmath102 becomes progressively less accurate due to violation of the cylindrical symmetry of the continuum states of the atom ; for instance , it exceeds our cross section by a factor of @xmath109 at @xmath110 a.u .
however , the approximate @xmath101 and @xmath103 remain surprisingly accurate .
short dashes correspond to the approach of kvh96 , who use the @xmath14-adiabatic approximation for the initial state , and the shifted adiabatic and born approximations for the final state .
if to abandon the born approximation ( retaining the adiabatic one ) , then @xmath111 would become considerably lower , by 0.5 to 3 orders of magnitude for the parameters chosen in fig .
[ fig2 ] , and even more different from the true cross sections
. this lowering is due to the approximate orthogonality of the initial and final _ longitudinal _ wave functions , which is not provided by the born approximation .
all the discussed approximations miss the beutler
fano resonances ( narrow peaks and dips near the thresholds ) , which occur at the energies of the quasi - bound states in the continuum .
these resonances are analogous to those discussed by ppv97 , except that now these quasi - bound states all belong to the ( @xmath112)-manifold .
autoionization in this case becomes possible due to the motion - induced coupling of different @xmath100-channels . except for the missing resonances , the discussed approximations appear to be sufficiently accurate for the longitudinal ( @xmath103 ) polarization .
figure [ fig3 ] demonstrates a representative sample of our numerical results for @xmath113 . in accordance with figure 1 ,
the photoionization thresholds shift redward with increasing @xmath2 . since @xmath114 a.u . at this field strength
, the left and right panels correspond to the centered and decentered atoms , respectively . for the centered atoms ,
the slopes of the curves for each given polarization remain practically coincident with those at @xmath115 .
the values of the cross sections for the parallel and right polarizations , corresponding to the radiative transitions with @xmath116 and @xmath117 allowed at @xmath115 ( when @xmath15 is a `` good '' quantum number ) , decrease only slightly with increasing @xmath2 from 0 to @xmath118 . for the left polarization ,
radiative transitions from the ground state are forbidden ( @xmath119 ) at @xmath115 ( pp93 ) . with growing @xmath2 , the coupling between different @xmath15-channels arises , and @xmath101 grows ( approximately as @xmath120
see pm93 ) until it becomes comparable to @xmath102 at @xmath106 . for the _ decentered _ atoms ( right panels of fig .
[ fig3 ] ) , the energy dependences steepen with growing @xmath2 .
this is explained by the change of the effective potential from coulomb - like at @xmath121 to oscillator - like as the atom becomes decentered .
the effective potentials corresponding to different @xmath15 become nearly identical , that results in a significant decrease of @xmath102 .
still faster is the decrease of @xmath101 caused by restoring applicability of the @xmath15-adiabatic approximation and corresponding selection rules at large @xmath2 . at the same time , the threshold value of @xmath103 remains almost independent of @xmath2 .
since the moving atom loses its cylindrical symmetry , the cross sections depend on the angle @xmath122 between @xmath1 and the transverse component of the wavevector .
the @xmath122-dependence is most pronounced when both the wavevector @xmath75 and polarization direction @xmath84 are perpendicular to @xmath33 .
the cross sections @xmath103 and @xmath123 are drawn in figure [ fig3 ] by dot - long - dash and dashed lines , respectively .
the former cross section does not depend on @xmath122 in the dipole approximation , whereas the small non - dipole corrections ( eq . [ [ mz ] ] ) cause a weak dependence with an amplitude @xmath124% .
therefore , we show only one curve for @xmath103 on each panel .
for the transverse polarization , the @xmath122-dependence becomes pronounced at @xmath125 a.u .
the upper and lower dashed curves correspond to the largest and smallest possible values of @xmath126 , at @xmath127 and @xmath128 , respectively .
the reason is that the transversely polarized photons `` see '' the largest transverse size of the atom ( stretched perpendicular to @xmath1 ) when they propagate along @xmath1 , and the smallest one when propagating transversely to @xmath1 .
the @xmath122-dependence between these two extremes , illustrated in figure [ fig4 ] , is described by the formula @xmath129 note that in the absence of coupling of different @xmath15-channels the amplitude @xmath130 in equation ( [ phi ] ) would equal @xmath131 , which is always greater than the actual value of @xmath130 , calculated numerically .
the curve corresponding to @xmath132 ry in figure [ fig4 ] goes higher than that corresponding to @xmath133 ry because the ionization channel @xmath134 opens at @xmath135 ry , where @xmath136 is the proton cyclotron frequency .
the curve corresponding to @xmath137 ry goes much higher than the other curves because it falls onto a slope of a beutler fano resonance .
such resonances are most prominent at @xmath106 ( cf .
[ fig3 ] ) because of the violation of the @xmath15-adiabatic approximation in this range of the transverse generalized momenta .
nevertheless , these peaks remain very narrow .
figure [ fig5 ] presents some autoionization widths @xmath138 calculated according to equations ( 14 ) and ( 15 ) of ppv97 ( with obvious generalization to the case of moving atoms ) .
the left panel shows the widths of the tightly - bound states @xmath139 through 4 which enter the continuum at @xmath140 , 560 , 340 , and 237 a.u . ,
respectively ( see fig .
[ fig1 ] ) . owing to the similarity of the effective potentials related to different @xmath15 values ,
the jumps of the curves corresponding to opening of additional autoionization channels occur at nearly the same values of @xmath2 .
the hydrogen - like levels of the ( @xmath139)-manifold ( right panel of fig . [ fig5 ] ) belong to the continuum at any @xmath2 .
the autoionization widths vanish at @xmath141 and at @xmath142 because of restoring cylindrical symmetry of the wavefunctions .
the bound - free absorption coefficient for a hydrogen gas of given temperature and density is proportional to the photoionization cross section @xmath143 folded with @xmath144 , the distribution of the number density of atoms in the initial state . since the level energy @xmath145 tends to a constant value at @xmath146 , the normalization integral for the boltzmann distribution @xmath147 $ ] diverges .
the divergence is eliminated if nonideality of the gas is taken into account . at high @xmath2 ,
the binding energies become small , and atomic sizes become large .
this means that the atom can be easily destroyed by surrounding particles , so that very high @xmath2 should not contribute to the corresponding integrals . in order to allow for this effect
, we use the occupation probability formalism , described in detail by hummer & mihalas ( 1988 ) and generalized to the case of the @xmath2-dependent atomic structure by pp95 .
following this approach , we write @xmath148,\ ] ] where @xmath149 is the occupation probability , equal to the fraction of atoms ( in the specified state @xmath150 ) which essentially preserve quantum - mechanical ( particularly , optical ) properties of an isolated atom .
we estimate @xmath149 according to equation ( 14 ) of pp95 , based on the quantum - mechanical atomic sizes calculated by p94 . a rapid decrease of @xmath149 at @xmath151 ( see fig
. 3 of pp95 ) limits contribution from atoms with large generalized momenta and provides convergence of integrals over @xmath2 .
since the threshold energies decrease with increasing @xmath2 , folding the photoionization cross sections with @xmath152 leads to the redward broadening of the photoionization edges , with a typical width @xmath153 ( pm93 ) .
it has been shown by pm93 and pp95 that the `` magnetic width '' is much greater than the doppler and ( impact ) collisional widths under typical conditions of ns atmospheres . in this paper
we take full account of the magnetic broadening and neglect other broadening mechanisms .
the averaged photoionization cross sections for atoms on the ground level are shown in fig .
[ fig6 ] for two values of temperature @xmath154 . for comparison ,
we show also cross sections for the bound - bound ( b - b ) transitions from the ground level ( pp95 ) .
both the b - b and bound - free ( b - f ) averaged cross sections are corrected for the stimulated emission , i. e. , multiplied by @xmath155)$ ] . for the circular polarizations ,
the b - b and b - f transitions dominate in different ( adjacent ) spectral ranges the low - energy b - f component , arising from the decentered states ( cf .
[ fig3 ] ) , turns out to be much weaker than the b - b component in the same energy range . for @xmath103 ,
this is also the case at @xmath156 , but not at @xmath157 . in the latter case , the low - energy b - f component dominates the absorption spectrum , except at very low energies @xmath158 ev .
the peak of @xmath102 in the right panel of figure [ fig6 ] at @xmath159 ev and the corresponding peaks of the ` b - b ' curves arise near the proton cyclotron frequency @xmath160 , around which the transitions of strongly decentered atoms are concentrated ( cf .
pp95 ) .
figure [ fig7 ] demonstrates the effect of density on the averaged photoionization cross sections . with increasing @xmath161
, the low - energy component rapidly decreases because the decentered atoms , due to their large sizes , are easier destroyed by the plasma microfields .
the growth of the high - energy component noticeable at higher @xmath154 is explained by the relative increase of the number of centered atoms when the decentered atoms are destroyed .
it should be noted that at high densities one more effect , broadening of the photoionization edges by quasi - static interactions with surrounding particles , may become important .
as the density grows , highly excited bound states become randomly disturbed by the surrounding and form a quasi - continuum .
this results in continuous photon absorption below the undisturbed photoionization threshold ( e. g. , dppen , anderson , & mihalas 1987 ) .
these `` bound - quasi - free '' transitions require a separate treatment and are not included in the present paper .
figure [ fig8 ] shows the effects of @xmath154 and @xmath162 on the photoabsorption spectral shapes . with growing @xmath154
, the low - energy components significantly increase , and the high - energy components of @xmath103 and @xmath102 decrease slightly , because the decentered states become more populated . at lower @xmath154
, the high - energy component of @xmath101 grows with @xmath154 linearly , @xmath163 in accordance with the result of pm93 obtained in the perturbation regime ( @xmath164 is the energy of the main b - b transition for the right polarization e. g. , @xmath165 ev for @xmath166 ) . at higher @xmath154
, the growth of @xmath101 decelerates because of decreasing fraction of the centered states and nonperturbative effects . at higher @xmath162 ,
the above - discussed proton cyclotron resonance for @xmath102 becomes stronger .
owing to the approximate similarity of the energy spectra related to different @xmath15-manifolds ( see fig .
[ fig1 ] ) , the individual beutler fano resonances ,
although smoothed by the thermal averaging , are collected together into regularly placed peaks .
such peaks corresponding to four proton cyclotron harmonics are indicated by arrows for @xmath101 at @xmath167 in fig .
[ fig8 ] . at high densities ,
the peaks can be further smoothed by interaction with surrounding particles .
the results presented here provide , together with the results of pp95 on the bound - bound transitions , a basis for calculations of realistic opacities of ns atmospheres . to find the opacities , one should add transitions from excited atomic levels , which is straightforward .
more difficult are the problems of ionization equilibrium in the dense , non - ideal plasmas and closely related problem of the non - ideality effects on the bound - free transitions ( including transitions to the quasi - free states below the unperturbed photoionization threshold ) .
finally , polarizations of the normal modes for any frequency and direction of propagation should be found for solving the radiative transfer equations .
bulik & pavlov ( 1996 ) have described how these polarizations can be calculated if the basic cross sections @xmath111 and @xmath103 are known .
our results show that atomic motion in strong magnetic fields drastically changes both the bound - bound and bound - free opacities .
in particular , it considerably broadens the spectral lines and photoionization edges ( with a typical magnetic width @xmath168 ) , gives rise to additional absorption at low frequencies due to decentered atoms , and opens new transition channels forbidden for atoms at rest .
these unusual properties of the spectral opacities alter the structure of the ns atmospheres and manifest themselves in the spectra and light curves of radiation emergent from the ns surface layers .
the authors are grateful to victor bezchastnov and joseph ventura for fruitful discussions , and to cole miller and the referee , tomek bulik , for useful remarks .
this work was partially supported through nasa grant nag5 - 2807 , intas grant 94 - 3834 , rbrf grant 96 - 02 - 16870a , and dfg rbrf grant 96 - 02 - 00177 g .
ayp gratefully acknowledges the hospitality of joseph ventura at the university of crete , gilles chabrier at ens - lyon , and chris pethick at nordita , where a part of this work has been done .
velocity and length representations of the operator of interaction of a moving hydrogen atom with radiation have been derived by bp94 .
here we employ the velocity representation as the simpler one .
kvh96 have noted that the interaction of the radiation field with the electron spin should be taken into account .
it has been found by ppv97 that the spin term almost does not change the cross sections at energies below 10 kev .
nevertheless , we include this term for the sake of generality , neglecting the terms responsible for the spin - flip transitions ( schmitt et al . 1981 ) which are strictly forbidden at the considered photon energies below @xmath169 .
in the conventional representation of the wavefunctions , equation ( a6 ) of bp94 yields for the dimensionless interaction operator @xmath88 : @xmath170 + \left({{m_{\rm e}}\over{m_{\rm p}}}{{\mbox{\boldmath $ \pi$}}\over\hbar } - { { m_{\rm e}}\over{m_{\rm h}}}{{\mbox{\boldmath $ k$}}\over\hbar } \right ) { \mbox{\boldmath $ e$}},\ ] ] where @xmath75 is the photon wavevector , @xmath84 is its polarization vector , and negligibly small terms @xmath171 are omitted .
here @xmath172 is the relative coordinate , and @xmath173 and @xmath174 are the electron- and proton - type kinetic momentum operators : where @xmath177 and @xmath178 for the right ( + ) and left ( @xmath179 ) components , respectively .
if the full - shift representation , @xmath82 , is used for the final state , and an arbitrary shift parameter @xmath26 for the initial one , then , according to equation ( a15 ) of bp94 , @xmath180 \nonumber\\ & \times & { \mbox{\boldmath $ e$}}\left ( \exp\left[\rule{0mm}{2ex}{\rm i}({\mbox{\boldmath $ q$}}{\mbox{\boldmath $ r$}}_\perp/2 + q_z z ) \right ] { \mbox{\boldmath $ f$}}_{\rm e } + \exp\left [ -{\rm i}\eta{\mbox{\boldmath $ r$}}_c { \mbox{\boldmath $ q$ } } - { \rm i}{\mbox{\boldmath $ q$}}{\mbox{\boldmath $ r$}}_\perp/2 \rule{0mm}{2ex } \right ] { \mbox{\boldmath $ f$}}_{\rm p } \right ) , \label{m_shift}\end{aligned}\ ] ] where @xmath24 . once again , the terms @xmath181 are omitted .
the components of the vector operators @xmath182 and @xmath183 are in order to proceed with equation ( [ m_expan ] ) , we need to calculate the transverse matrix elements @xmath185 . in the particular case of the dipole approximation ( @xmath186 ) , such calculation has been performed by bp94 . in the general case , we use the following result obtained with the same operator technique as in bp94 : @xmath187 | ns\eta,{{\mbox{\boldmath $ k$}}_\perp}\rangle_\perp = j_{nn'}(\xi_{\rm e } ) j_{n+s , n'+s'}(\xi_{\rm p } ) , \label{jj}\ ] ] where @xmath188 is arbitrary , using equations ( [ m_shift ] ) , ( [ pi_pm ] ) , ( [ pi_pm ] ) , and ( [ jj ] ) for @xmath82 , and neglecting an insignificant common phase , we finally arrive at the expression for the interaction matrix element ( [ m_expan ] ) : @xmath191 , \label{m_final}\ ] ] where @xmath192 j_{n+s , n'+s'}(\xi_{{\rm p},+})\ , { \cal{z}}^{\rm ( e)}_{n's'ns } \nonumber\\ & - & { { m_{\rm e}}\over{m_{\rm p}}}\exp(-{\rm i}\eta{\mbox{\boldmath $ r$}}_c { \mbox{\boldmath $ q$ } } ) \left [ \sqrt{n+s+1}\ , j_{n+s+1,n'+s'}(\xi_{{\rm p},- } ) - \zeta_{{\rm p},- } j_{n+s , n'+s'}(\xi_{{\rm p},- } ) \right ] \nonumber\\ & \times & j_{nn'}(\xi_{{\rm e},-})\ , { \cal{z}}^{\rm ( p)}_{n's'ns } , \label{m-}\\ m^{(+)}_{n's'ns } & = & \left [ \sqrt{n+1}\ , j_{n+1,n'}(\xi_{{\rm e},+ } ) - \zeta_{{\rm e},+ } j_{nn'}(\xi_{{\rm e},+ } ) \right ] j_{n+s , n'+s'}(\xi_{{\rm p},+})\ , { \cal{z}}^{\rm ( e)}_{n's'ns } \nonumber\\ & + & { { m_{\rm e}}\over{m_{\rm p}}}\exp(-{\rm i}\eta{\mbox{\boldmath $ r$}}_c { \mbox{\boldmath $ q$ } } ) \left [ \sqrt{n+s}\ , j_{n+s+1,n'+s'}(\xi_{{\rm p},- } ) + \zeta_{{\rm p},+ } j_{n+s , n'+s'}(\xi_{{\rm p},- } ) \right ] \nonumber\\ & \times & j_{nn'}(\xi_{{\rm e},-})\ , { \cal{z}}^{\rm ( p)}_{n's'ns } , \label{m+}\\ m^{(z)}_{n's'ns } & = & \left [ \widetilde{\cal{z}}^{\rm ( e)}_{n's'ns } + { { \rm i}\over 2 } { a_{\rm m}}q_z { \cal{z}}^{\rm ( e)}_{n's'ns } \right ] j_{nn'}(\xi_{{\rm e},+ } ) j_{n+s , n'+s'}(\xi_{{\rm p},+ } ) \nonumber\\ & + & { { m_{\rm
e}}\over{m_{\rm p}}}\exp(-{\rm i}\eta{\mbox{\boldmath $ r$}}_c { \mbox{\boldmath $ q$ } } ) \left [ \widetilde{\cal{z}}^{\rm ( p)}_{n's'ns } - { { \rm i}\over 2 } { a_{\rm m}}q_z { \cal{z}}^{\rm ( p)}_{n's'ns } \right ] j_{nn'}(\xi_{{\rm e},- } ) j_{n+s , n'+s'}(\xi_{{\rm p},- } ) \label{mz}\end{aligned}\ ] ] we have presented the result for @xmath82 . for arbitrary @xmath198 , the matrix elements are obtained in analytic form using equation ( a15 ) of bp94 and our equations ( a6 ) , ( a7 ) with @xmath199{\mbox{\boldmath $ q$}}/{m_{\rm h}}$ ] and @xmath200{\mbox{\boldmath $ q$}}/{m_{\rm h}}$ ] .
the result is a straightforward generalization of equations ( a10 ) .
bezchastnov , v. g. , & potekhin , a. y. 1994 , j. phys .
b , 27 , 3349 ( bp94 ) bulik , t. , & pavlov , g. g. 1996 , , 469 , 373 burkova , l. a. , dzyaloshinskii , i. e. , drukarev , g. f. , & monozon , b. s. 1976 , sov .
phys.jetp , 44 , 276 dppen , w. , anderson , l. s. , & mihalas , d. 1987 , , 319 , 195 gnedin , yu .
n. , pavlov , g. g. , & tsygan , a. i. 1974 , sov .
phys.jetp , 39 , 301 gorkov , l. p. , & dzyaloshinskii , i. e. 1968 , sov .
phys.jetp , 26 , 449 hummer , d. g. , & mihalas , d. 1988 , ap.j . , 331 , 794 johnson , b. r. , hirschfelder , j. o. , & yang , k .- h .
1983 , rev .
phys . , 55 , 109 kopidakis , n. , ventura , j. , & herold , h. 1996 , , 308 , 747 ( kvh96 ) miller , m. c. , & neuhauser , d. 1991 , mnras , 253 , 107 gelman , h. 1995 , in the lives of the neutron stars , eds .
m. a. alpar , . kizilolu , & j. van paradijs ( dordrecht : kluwer ) , 101 pethick , c. j. 1992 , rev . mod .
, 64 , 1133 pavlov , g. g. , & mszros , p. 1993
, , 416 , 752 ( pm93 ) pavlov , g. g. , shibanov , yu .
a. , zavlin , v. e. , & meyer , r. d. 1995 , in the lives of the neutron stars , eds .
m. a. alpar , . kizilolu , & j. van paradijs ( dordrecht : kluwer ) , 71 pavlov , g. g. , & potekhin , a. y. 1995 , , 450 , 883 ( pp95 ) potekhin , a. y. 1994 , j. phys .
b , 27 , 1073 ( p94 ) potekhin , a. y. , pavlov , g. g. 1993 , apj , 407 , 330 ( pp93 ) potekhin , a. y. , pavlov , g. g. , & ventura , j. 1997 , , 317 , 618 ( ppv97 ) romani , r.w .
1987 , , 313 , 718 schmitt , w. , herold , h. , ruder , h. , wunner , g. 1981 , , 94 , 194 sokolov , a. a. , & ternov , i. m. 1968 , synchrotron radiation ( berlin : academic ) ventura , j. , herold , h. , ruder , h. , & geyer , f. 1992 , , 261 , 235 vincke , m. , le dourneuf , m. , & baye , d. 1992 , j. phys .
b , 25 , 2787 wunner , g. , ruder , h. , herold , h. , & schmitt , w. 1983 , , 117 , 156 | the strong magnetic fields ( @xmath0 g ) characteristic of neutron stars make all the properties of an atom strongly dependent on the transverse component @xmath1 of its generalized momentum . in particular ,
the photoionization process is modified substantially : ( i ) threshold energies are decreased as compared with those for an atom at rest , ( ii ) cross section values are changed significantly , and ( iii ) selection rules valid for atoms at rest are violated by the motion so that new photoionization channels become allowed . to calculate the photoionization cross sections , we ,
for the first time , employ exact numerical treatment of both initial and final atomic states .
this enables us to take into account the quasi - bound ( autoionizing ) atomic states as well as coupling of different ionization channels .
we extend the previous consideration , restricted to the so - called centered states corresponding to relatively small values of @xmath2 , to arbitrary states of atomic motion .
we fold the cross sections with the thermal distribution of atoms over @xmath3 . for typical temperatures of neutron star atmospheres ,
the averaged cross sections differ substantially from those of atoms at rest .
in particular , the photoionization edges are strongly broadened by the thermal motion of atoms ; this `` magnetic broadening '' exceeds the usual doppler broadening by orders of magnitude .
the decentered states of the atoms give rise to the low - energy component of the photoionization cross section .
this new component grows significantly with increasing temperature above @xmath4 k and decreasing density below 1 g @xmath5 , i. e. , for the conditions expected in atmospheres of middle - aged neutron stars .
accepted for publication in _ the astrophysical journal _ |
the leading model for long gamma - ray bursts ( grbs ) and hypernovae assumes that they are driven by energy released from a disk accreting onto a black hole resulting from a massive stellar collapse .
this accretion energy is converted to the explosion energy in the grb / hypernova jet either through some magnetic field mechanism or through the annihilation of neutrinos emitted by the hot disk . in this paper , we study the physics behind the neutrino driven explosion mechanism , deriving critical densities at which an explosion is launched . for long - duration ( @xmath0 ) grbs and all hypernovae ,
the progenitor is a massive star that either does nt or weakly explodes via the normal supernova mechanism : a.k.a .
a collapsar ( macfadyen & woosley 2000 , macfadyen , heger , & woosley 2001 ) . by studying the structure of these massive stars at collapse and comparing to our critical densities
, we can estimate the delay between stellar collapse and the launch of the jet as well as the ultimate black hole remnant mass .
future observations of these delays and remnant masses provide constraints on the progenitor and/or jet mechanism for gamma - ray bursts .
neutrinos emitted from black hole accretion disks push out against the infalling star both through the direct absorption and scattering of neutrinos on infalling matter and through the energy deposited in it as electron / positron pairs and photons from neutrino annihilation .
neutrino annihilation , which provides most of the explosion energy , occurs primarily above the black hole along the rotation axis of the accretion disk .
popham et al .
( 1999 ) studied a range of likely accretion disk structures and calculated their resultant energy deposition through neutrino annihilation . using this study
, we can derive the value of the matter density above the black hole for which the energy deposited by neutrinos is able to drive an explosion , and obtain an estimate of the mass of the black hole remnant left behind by this explosion .
we will define the onset of the explosion as the moment when neutrino momentum deposition exceeds the pull of gravity on matter flowing onto the accretion disk along the rotation axis .
the scattering opacity ( @xmath1 ) for neutrinos is roughly ( janka 2001 ) : @xmath2 where @xmath3 g is atomic mass unit , @xmath4mev is electron rest - mass energy , @xmath5@xmath6 , @xmath7 is the neutrino energy , @xmath8 is the density above the rotation axis , @xmath9 and @xmath10 are the number fractions of free neutrons and protons .
the corresponding absorption opacity ( @xmath11 ) for neutrinos is ( janka 2001 ) @xmath12 for the conditions in our disks , electron capture produces over half of the total neutrinos and electron neutrinos are the most abundant neutrino species .
assuming this species dominates the absorption and scattering , the total neutrino opacity is @xmath13 here we have used the following assumptions : axial - vector couplings set to the charged - current axial - vector coupling constant in a vacuum , @xmath14 , @xmath15 , the neutrino temperature ( @xmath16 ) is related to neutrino energy by @xmath17 and that the entire neutrino flux is in the electron neutrino species .
had we assumed an equal mix of electron and anti - electron neutrinos , the total opacity would not be different by more than 20% . as a shell of matter falls towards the black hole ,
it is supported by the momentum of the scattered and absorbed neutrinos : @xmath18 where @xmath19 is the neutrino luminosity , @xmath20 is the mass , @xmath21 the thickness , @xmath22 the radius of the shell , @xmath23 is neutrino velocity @xmath24 the speed of light .
the corresponding acceleration from neutrino annihilation requires detailed models of the black hole accretion disk system .
the disk models and annihilation calculations from popham et al .
( 1999 ) led to a value for the energy deposited along a surface per unit path length ( @xmath25= { \rm ergs \
, s^{-1 } \ , cm^{-1}}$ ] ) . with this deposition rate @xmath26 , which is a function of height above the disk
, we can calculate the acceleration due to neutrino annihilation inside a @xmath27 cone along the rotation axis : @xmath28 the onset of the explosion occurs when @xmath29 where @xmath30 is the gravitational acceleration with gravitational constant @xmath31 , and black hole mass @xmath32 . even for this high disk mass ,
the change in the gravitational acceleration is less than 20% . ] . using equations 4 and 5
, the acceleration condition for the onset of the explosion can be translated into a threshold condition on the density : @xmath33 for our estimates , we take the radii where the energy deposition from neutrino annihilation peaks ( @xmath3420 km ) .
figure 1 gives critical densities for set values of @xmath35 versus an effective luminosity [ @xmath36 .
the critical density rises sharply as the `` eddington '' luminosity for neutrinos is reached . the data ( @xmath35 at 20 km and @xmath19 ) for a series of disks from popham et al .
( 1999 ) are also shown for comparison .
these critical densities correspond to an accretion rate inside the @xmath27 cones along the rotation axes of @xmath37 where @xmath38 and @xmath39 .
the accretion rates for massive stars are initially much higher than these critical values , but the accretion rate decreases as material from increasingly outward layers of the star falls onto the black hole . in figure 2
we plot the critical accretion rates for disks accreting @xmath40 and @xmath41 @xmath42 .
note that the accretion rate through the disk need not be related to ( and at late times , is much higher than ) the accretion along the rotation axis . for comparison ,
we plot the accretion rates of stars with masses ranging from 20 - 60@xmath43 ( rauscher et al .
2002 ) versus mass coordinate of the accreting mass layer .
when the accretion rate of stellar material falling along the polar axis drops below the critical accretion rate for a given disk , the neutrinos from the disk will drive an explosion and , very likely , disrupt the entire star .
for a given disk structure , then , this crossing point gives a rough estimate of the remnant mass ( e.g. from fig .
2 we see that for a disk accretion rate of @xmath44 onto a black hole rotating near breakup : a=0.95 , the remnant black hole mass for a @xmath45 star is roughly @xmath46 . )
note also , that if the neutrino emission was as high as that given by the disks accreting at @xmath47 from popham et al .
( 1999 ) , then the explosion would occur immediately .
however , di matteo , perna , & narayan ( 2002 ) have found that these disks actually produce a much lower neutrino flux .
even if such high neutrino luminosities could be constructed , they would disrupt their star immediately .
we note that these results are only rough approximations , since a number of caveats limit the quantitative accuracy of these calculations . besides assuming a fairly simple model for the momentum deposition , relativistic effects on the neutrino energy and momentum , the effects of rotation on the infalling matter and the evolution of the black hole mass and spin were all neglected .
if magnetic fields drive the explosion , the explosion can occur at much earlier times , and similarly , the disk winds seen by macfadyen & woosley ( 2000 ) also would eject the star at earlier times . most of these effects would lead to higher critical densities and lower remnant black hole masses . however , if we restrict ourselves to neutrino - driven explosions only , the major uncertainty is the structure of the progenitor star .
figure 3 shows the infall accretion rates for rotating and non - rotating stars of 40 and 60 @xmath43 ( heger 2002 ) .
note that the remnant mass differs dramatically depending upon the amount of rotation . comparing
the non - rotating 60 @xmath43 case of fig.3 with that of fig.2 shows the variations arising from stars produced with different versions of the same stellar code : rauscher et al .
( 2002 ) and heger ( 2002 ) use different versions of the kepler code ( weaver , zimmerman , & woosley 1978 ) .
in addition to uncertainties in single star evolution , it is likely that collapsar grbs arise from binary systems , and binary progenitors of grbs have not yet been constructed .
for most values of the disk accretion rates ( within roughly @xmath48 ) the critical density lies in a fairly narrow range between @xmath49 .
these densities agree well with the results from macfadyen & woosley ( 1999 ) .
for lower disk accretion rates , the energy deposition from neutrino annihilation decreases dramatically and the critical density for explosions drops below @xmath50 . the material along the poles does not reach this density until nearly all of the star has accreted onto the black hole , and this will not produce a grb .
similarly , higher neutrino luminosities would disrupt the star immediately and probably never form in nature .
with such a narrow range of explosion conditions at the black hole source , one would expect some similarities in the grb outbursts produced by neutrinos . however , bear in mind that the propogation of the jet through the star ( which can differ in different progenitors ) can significantly alter the observable outburst .
what does differ dramatically is the time after the collapse at which the explosion occurs . from figures 2 and 3 , we know the mass zone of the star that is driven to explosion by the neutrinos . using the @xmath51 of heger ( 2002 ) and assuming that , along the rotation axis , the critical mass zone collapses at free - fall , we find that disks accreting at @xmath52 drive explosions at much different times .
for the 40@xmath43 star , infalling material along the pole is turned around at 50,60 or @xmath53s after the initial collapse for black hole spin rates of a=0.95,0.75 and 0 , respectively . for the 60@xmath43 star ,
the corresponding explosion times are : 35 , 300 , @xmath54s .
hence , we should expect a considerable delay ( at least @xmath34 30s ) between the collapse ( as signalled by the initial neutrino and gravitational wave signal ) and the launch of the grb explosion .
there are , however , upper limits to the delay .
the explosion must occur , after all , before the disk accretes . using the convection dominated accretion flow ( cdaf ) solutions from narayan , piran , & kumar ( 2001 ) , we find that for accretion rates above @xmath55 and a total mass accreted through the disk less than @xmath46 , the total time including the delay since the collapse plus the burst duration can not be higher than 200s .
those disks which do not explode before this time will not explode before the disk dissipates . requiring that the accretion rate is at least @xmath55 places an upper limit on the disk formation radius ( the radius at which the angular momentum in the infalling stellar material supports it against the gravitational pull of the black hole ) .
similarly , requiring that the disk maintains its accretion rate long enough to survive the delay times given above places a lower limit on the disk formation radius . using the cdaf solutions from narayan et al .
( 2001 ) and assuming that the disk @xmath56 viscosity @xmath57 , @xmath58 , @xmath59 , a total mass accreted through the disk @xmath60 with at least @xmath61 in the disk , yields a range of disk formation radii from 28 - 7000 km .
this corresponds to specific stellar angular momenta in the range : @xmath62 .
although this is a narrow range of stellar angular momenta , it does lie within many of the current rotating models ( e.g. heger 2002 ) . for neutrino - driven explosions from black hole accretion disks
, we can derive the remnant mass of black holes .
stars between @xmath63 are likely to have weak supernova explosions which ultimately lead to considerable fallback and the formation of a black hole ( fryer 1999 ) . if these stars have insufficient angular momentum , they produce a range of black hole masses between @xmath64 ( fryer & kalogera 2001 ) .
if instead , they are rapidly rotating , they can form hypernovae and lower mass black holes ( nakamura et al .
simulations by macfadyen , woosley , & heger ( 2001 ) show that @xmath65s after their weak explosions , the infall rates of these stars will drop below our critical densities ( assuming disk accretion rates in the range : @xmath66 ) and will drive explosions .
these `` collapsar type ii '' objects will range from @xmath67 . for direct collapse black holes which are the more likely grb candidate ,
the currently most reliable progenitors ( rotating 40,60 @xmath43 stars ) , yield black hole remnants masses to range within 14 - 23@xmath43 with disk accretion rates in the range : @xmath68 .
if other effects ( e.g. disk winds , magnetic fields ) are important , these masses could be lower .
but not all stars will form grbs . for stars with insufficient angular momentum to produce these disk accretion rates , weak or
no explosions are produced , and the remnant can be much more massive ( up to the mass of the star ) . | we study the physics behind the neutrino - driven mechanism for gamma - ray bursts and hypernovae , deriving the critical density at which these outbursts occur in the collapsar model .
the agreement between this derivation and results from past collapsar simulations ( macfadyen & woosley 2000 ) is excellent , implying that we have captured the essential physics .
we then use this derivation to study a range of progenitors for collapsar gamma - ray bursts .
we derive how much of the star will accrete onto the black hole core before the infall density drops below this critical density , leading to an estimate of the remnant black hole mass for grbs and hypernovae .
we also estimate the time delays between gravity wave or neutrino signals and the onset of the explosion or burst event .
this derivation , combined with future observational constraints , provides a physical insight into the structure of the grb progenitor . |
as present grids of theoretical stellar spectra suffer from intrinsic inhomogeneities and incompleteness and show large systematic discrepancies with empirical calibrations due to unavailable molecular opacity , we have undertaken the construction of a comprehensive combined library of _ realistic _ stellar flux distributions .
empirical @xmath2-color relations in @xmath1 photometry are used to calibrate the spectra using an algorithm developed by cuisinier et al .
+ in a first step , the calibration was carried out for solar metallicity , and the differential properties with metallicity were preserved ( lejeune et al . 1997 ) . as expected , some discrepancies at low @xmath3}$ ] remained , so the next step was to collect photometry of galactic globular clusters in order to do the calibration in a metallicity - dependent way .
+ although much work still has to be done , first results are presented here .
in order to cover the necessary range of parameters , we combined the kurucz ( 1995 ) models ( @xmath4 } \leq + 1.0 , 0.0 \leq \log g \leq 5.0 $ ] ) with the scholz ( 1997 ) models ( @xmath5 } \leq + 0.5 , -1.02 \leq \log g \leq 2.0 $ ] ) and the allard & hauschildt ( 1995 ) models ( @xmath6 } \leq + 0.5 , 3.5 \leq \log g \leq 5.5 $ ] ) . + this grid of models was then calibrated using @xmath0 photometric data of solar neighbourhood stars ( dwarfs and giants ) and an algorithm developed by cuisinier et al .
( 1995 , see lejeune et al .
1998 ) , which modifies the model spectra to match the calibration colors for solar abundances and then preserves the differential properties with metallicity and surface gravity of the original models . + as expected , this so - called semi - empirical library still has its limitations at low metallicities .
although the tendencies with decreasing @xmath3}$ ] seem to be well reproduced for visible to near infrared colors ( @xmath7-@xmath8 , @xmath8-@xmath9 , @xmath8-@xmath10 ) , the effect of metallicity in the far - infrared ( @xmath11-@xmath12 , @xmath12-@xmath10 , @xmath11-@xmath10 , @xmath10-@xmath13 ) and the ultraviolet ( @xmath14-@xmath7 ) are less well reproduced .
this was expected from the fact that some important molecular opacity data are still missing .
+ the semi - empirical ( basel 3 ) sed library is currently available by ftp from the astronomical institute of the university of basel , switzerland .
in order to overcome these shortcomings , we performed a metallicity - dependent calibration of the grid , using empirical photometry from the literature ( see the reference list ) of the best - studied galactic globular clusters with metallicities down to -2.16 ( @xmath3}$]-values from carretta & gratton 1997 ) , as globulars are abundant in stars of known metallicity and all evolutionary stages .
+ combined multi - color ( @xmath1 ) - @xmath3}$ ] - @xmath2 - @xmath15 relations were synthesised using @xmath16 and @xmath17-@xmath18 values from the harris online catalog of globular cluster parameters , the @xmath2 - @xmath8-@xmath10 relation from ridgway et al .
( 1980 ) and empirical @xmath2 - @xmath15 relations for red giants from frogel , persson and cohen ( 1978 , 1981 and 1983 , @xmath3}$]-dependent ) and dwarfs from angelov ( 1996 , @xmath3}$]-independent ) , where the large gaps were filled using the differential properties ( in @xmath2 ) of the semi - empirical grid .
at present , there has only been time to produce temperature - color relations for `` rgb ' 's ( the models evaluated for the parameter values of empirical rgbs ) .
these first results are presented in the following figure , wherein the new ( metallicity - calibrated ) models ( dashed ) are compared with the basel 3 ( semi - empirical ) models ( dotted ) and the empirical relations from the clusters that were used for this calibration and solar metallicity ( solid ) .
as expected , the changes in most colors are relatively minor , but the new grid appears to show smoother relations , and improvements can be seen for @xmath14-@xmath7 , @xmath11-@xmath12 , @xmath12-@xmath10 and @xmath11-@xmath10 , especially at low temperatures .
although more relevant tests still have to be performed , some small improvements seem to have been achieved with this first metallicity - calibration of the basel library of stellar seds . the real breakthrough however , can only be achieved after the huge gaps in the empirical data are filled . in particular , @xmath19 photometry needs to be performed for globular clusters of low @xmath3}$ ] down to the main sequence .
+ we expect to obtain new observations by using the eso ntt / 3.6 m telescopes soon . in a first run , we plan to collect @xmath20 photometry data for galactic globular cluster stars covering luminosities from the tip of the rgb down to the main sequence at @xmath21 , and metallicities in the range @xmath22 } < -1.2 $ ] .
alcaino , g. , & liller , w. 1986 , , 91 , 87 alcaino , g. , liller , w. , alvarado , f. , kravtsov , v. , ipatov , a. , samus , n. , & smirnov , o. 1997 , , 114 , 1067 allard , f. , & hauschildt , p. h. 1995 , , 445 , 433 angelov , t. 1996 , bobeo , 153 , 19 arribas , s. , & martinez roger , c. 1987 , , 178 , 106 bessell , m. s. , castelli , f. , & plez , b. 1998 , , 333 , 231 bica , e. , ortolani , s. , & barbuy , b. 1994 , , 106 , 161 von braun , k. , chiboucas , k. , minske , j. c. , salgado , j. f. , & worthey , g. 1998 , , 110 , 810v carretta , e. , & gratton , r. g. 1997 , , 121 , 95 cohen , j. g. , frogel , j. a. , & persson , s. e. 1978 , , 222 , 165 cuisinier , f. , lejeune , t. , & buser , r. 1995 , proc . of the iau symp . , 171 , 355 dacosta , g. s. , & armandroff , t. e. 1990 , , 100 , 162 drissen , l. , & shara , m. m. 1998 , , 115 , 725 ferraro , f. r. , carretta , e. , corsi , c. e. , fusi pecci , f. , cacciari , c. , buonanno , r. , paltrinieri , b. , & hamilton , d. 1997 , , 320 , 757 frogel , j. a. , persson , s. e. , & cohen , j. g. 1981 , , 246 , 842 frogel , j. a. , persson , s. e. , & cohen , j. g. 1983 , , 53 , 713 grebel , e. k. , & roberts , w. j. 1995 , , 109 , 293 harris , w. e. 1997 , catalog of parameters for milky way globular clusters , http://physun.physics.mcmaster.ca/gc/ hesser , j. e. , harris , w. e. , vandenberg , d. a. , allwright , j. w. b. , shott , p. , & stetson , p. b. 1987 , , 99 , 739 kaluzny , j. 1997 , , 122 , 1 kaluzny , j. , wysocka , a. , stanek , k. z. , & krxeminsky , w. 1998 , aca , 48 , 439 king , i. r. , anderson , j. , cool , a. m. , & piotto , g. 1998 , , 492l , 37 kurucz , r. l. 1995 ( private communication ) lee , s - g .
, lee , m. g. , & kim , e. 1996 , jkas , 29 , 171 lejeune , t. , cuisinier , f. , & buser , r. 1997 , , 125 , 229 lejeune , t. , cuisinier , f. , & buser , r. 1998 , , 130 , 65 montegriffo p. , ferraro f. r. , fusi pecci f. , origilia l. 1995 , , 276 , 739 piotto , g. , cool , a. m. , & king , i. r. 1997 , , 113 , 1345 ridgway , s. t. , joyce , r. r. , white , n. m. , & wing , r. f. 1980 , , 235 , 126 rieke , g. h. , & lebofsky , m. j. 1985 , , 288 , 618 salaris , m. , chieffi , a. , & straniero o. 1993 , , 414 , 580 sandage , a. 1970 , , 162 , 841 sandquist , e. l. , bolte , m. , stetson , p. b. , & hesser , j. e. 1996 , , 470 , 910 scholz , m. 1997 ( private communication ) stetson , p. b. , & harris , w. e. 1988 , , 96 , 909 | although stellar sed models are constantly improving and seem to show similar differential tendencies as measured spectra , their absolute properties ( eg . absolute flux in different wavelength regions ) can still differ significantly from the measured ones .
therefore , it is necessary to `` color - calibrate '' the theoretical spectra to obtain realistic model - based stellar spectra , which can then be used for many purposes , such as population synthesis models . + in order to calibrate a hybrid library ( consisting of the kurucz - , the allard and hauschildt - , and the scholz - models ) of theoretical seds , @xmath0 photometric data of solar neighborhood stars were collected from the literature , and the solar metallicity models were calibrated using an algorithm developed by cuisinier et al .
the corrections were then propagated through the whole grid of models , resulting in the `` semi - empirical '' models ( available on ftp from the university of basel ) . +
in order to extend the calibration to non - solar abundances , we have collected @xmath1 photometric data from well - known galactic globular clusters , covering a wide range of metallicities . + we shall present first results , using the metallicity- calibrated library for calculating isochrones of metal - poor clusters .
in particular , the observed slope and curvature of the rgb appears to be reproducible in a systematically correct manner for the first time . |
the concept of inverse orthogonalization of a sequence of functions emerged from the study of spectral methods for numerical simulation of near - wall turbulent flows @xcite .
turbulence in the boundary layer near a flat plate is an intensive physical phenomenon that exhibits an exponential decay of disturbances far from the wall .
thus , the sequence of exponential functions @xmath3 , being orthogonalized , can be employed for numerical simulation of the phenomenon .
it was noticed that applying the alternative algorithm of inverse orthogonalization to the sequence of exponential functions is as constructive as applying the direct orthogonalization algorithm to the sequence of monomials @xmath4 $ ] , and the bidirectional algorithm of orthogonalization was introduced in @xcite for defining systems of orthogonal exponential polynomials on the semi - axis .
later , the alternative legendre polynomials were presented , and two associated quadrature rules were obtained @xcite . the alternative jacobi polynomials are described in this paper . by taking advantage of the bidirectional orthogonalization we establish general properties of the polynomials .
we show that the shifted jacobi polynomials corresponding to the weight function with a translated parameter are a special subset of the alternative jacobi polynomials .
also , we consider an exceptional case of singular orthogonality ; this feature takes place in some of the alternative systems .
we find that the alternative jacobi polynomials in an exponential form are the system of functions on the semi - axis @xmath5 that has the same algorithmic capacity as the jacobi polynomials have on the interval @xmath6 $ ] .
such a deduction is illustrated with two special subsystems of the exponential polynomials possessing nice properties .
the orthogonal exponential polynomials on the semi - axis may serve a useful purpose in approximation of a continuous function on a closed interval .
following @xcite , we introduce two parameterized systems of discretely almost orthogonal functions on the interval @xmath0 $ ] .
the systems are constituted of the orthogonal exponential polynomials and unity . for the second system , supplementary computations are required to meet the definition of the functions , and properties of such a special construction have not been studied . to obtain these results we followed the classical theory of orthogonal polynomials .
different aspects of the theory that we were inspired by are represented in monographs @xcite .
let @xmath8 be a fixed whole number and @xmath9 is the system of polynomials defined by rodrigues type formula as @xmath10 with @xmath11 . from ( [ j1 ] )
it immediately follows that @xmath12 notation @xmath13 represents the falling factorial .
the last formula contains the term of lowest order in @xmath14 such that @xmath15 where @xmath16 is the gamma function . for @xmath17 @xmath18 substituting ( [ j1 ] ) to ( [ j1aaa ] ) and integrating by parts we confirm ( [ j1aaa ] ) .
the alternative jacobi polynomials @xmath19 satisfy @xmath20 @xmath21 here @xmath22 is the kronecker delta .
since orthogonality is verified by lemma 1.1 , we only need to consider the case @xmath23 . substituting ( [ j1 ] ) and ( [ j1aa ] ) to ( [ j0 ] ) and integrating by parts we calculate @xmath24 . substituting @xmath25 and @xmath26 to @xmath24 we confirm the restrictions on @xmath27 and @xmath28 in ( [ jd ] ) . for @xmath29 orthogonality relations ( [ j1aaa ] ) - ( [ j2 ] )
hold the invariance @xmath30 and @xmath31 polynomials @xmath32 are orthogonal on the interval @xmath0 $ ] with the weight function @xmath33 . following the proof of theorem 2.2
we also evaluate @xmath34 making use of ( [ j1 ] ) , cauchy s integral formula for analytic functions , and reciprocal substitutions one can obtain the integral representation @xmath35 where @xmath36 is a closed contour encircling the point @xmath37 .
being constructed by the procedure of inverse orthogonalization , polynomials @xmath19 can be represented in terms of the jacobi polynomials .
below we derive two relationships of such kind .
first we determine the relationship between @xmath38 and polynomials from the sequence @xmath39 that are defined by the procedure of direct orthogonalization as @xmath40 @xmath41 @xmath42 since @xmath8 is fixed , by verifying ( [ jsign ] ) , ( [ j2b ] ) , and ( [ j2a ] ) , the polynomials @xmath43 can be immediately connected to a fixed set of the jacobi polynomials @xmath44 @xcite .
precisely , @xmath45 this formula can be used directly for describing properties of @xmath46 , and one of the results that follow from ( [ j3 ] ) is the integral representation @xmath47 here @xmath48 is a closed contour encircling the point @xmath49 .
comparison of ( [ j21 ] ) and ( [ j5 ] ) leads to @xmath50 thus , relation of reciprocity ( [ j6 ] ) associates two sequences of polynomials that contain different powers @xmath51 i.e. , with @xmath52 and with @xmath53 respectively .
let us now consider the relationship between @xmath38 and polynomials from the sequence @xmath54 such that @xmath55 shifted jacobi polynomials ( [ jj1 ] ) are orthogonal on the interval @xmath0 $ ] with the weight function @xmath56 .
the polynomials can also be defined by the rodrigues formula as @xmath57 comparing ( [ j1 ] ) and ( [ jj2 ] ) we find that @xmath58 relation of reciprocity ( [ j6 ] ) and formula ( [ j3 ] ) , as well as ( [ jj3 ] ) and ( [ jj1 ] ) , enable employing the classical results for establishing properties of @xmath19 . following ( [ j6 ] ) and ( [ j3 ] ) we obtain the three - term recurrence relation @xmath59 @xmath60 @xmath61 @xmath62\ ] ] @xmath63 and the differential - difference relations @xmath64\ ] ] @xmath65 @xmath66 @xmath67x{\cal p}_{nk}^{(\alpha,\beta)}(x)\ ] ] @xmath68 following ( [ jj3 ] ) and ( [ jj1 ] ) we find the differentiation formula @xmath69 for @xmath70 , and the differential equation @xmath71y^{\prime}(x)\ ] ] @xmath72y(x)=0 \label{de1}\ ] ] that has a solution @xmath73 .
let us consider the case @xmath25 .
it follows from ( [ jj3 ] ) that @xmath74 thus , the shifted jacobi polynomials with translated @xmath27 , @xmath75 , are a subset of the alternative jacobi polynomials @xmath19 .
relation ( [ al2 ] ) can also be directly traced from equation ( [ de1 ] ) . for @xmath25
the degrees of the polynomial coefficients in ( [ de1 ] ) are reduced by one , and this makes @xmath76 the sequence with classical properties . for @xmath77 polynomials @xmath19
are less perfect . by considering property ( [ dx ] ) , or by differentiating ( [ de1 ] )
, we find that the first derivative of a @xmath19 is not a constant multiple of a @xmath78 with an @xmath79 , and so @xmath80 do not belong to the original class of polynomials .
still , other properties of @xmath19 described above are similar to properties of the classical orthogonal polynomials .
the restriction @xmath81 for the the weight function @xmath82 in ( [ j1aaa ] ) , ( [ j0 ] ) is stronger than the restriction on @xmath27 in ( [ al2 ] ) .
let @xmath83 . in this exceptional case
the polynomial @xmath84 is not integrable with respect to the weight function @xmath85 and therefore non - normalizable .
however , it is consistent with orthogonality relations ( [ j1aaa ] ) , ( [ j0 ] ) for @xmath86 .
thus , @xmath84 is a singular term of the sequence @xmath87 , which we call a marginal system .
assuming @xmath77 one can employ the sequences @xmath88 for approximation of a continuous function @xmath89 on the interval @xmath0 $ ] .
such an approximation is effective near @xmath90 only if @xmath91 .
below we describe two noteworthy orthogonal sequences that hold singular terms . considering ( [ j1 ] ) we find that @xmath92 are the shifted legendre polynomials .
this classical set of polynomials is a singular subset of the system @xmath93 the system @xmath94 was introduced in @xcite , and it is the first example of an inversely orthogonalized sequence of monomials @xmath95 .
the polynomials @xmath96 obey the orthogonality relations @xmath97 and @xmath98 .,width=226,height=136 ] they can be calculated using the expansion @xmath99 or following the three - term recurrence relation @xmath100 @xmath101 @xmath102{\mathscr a}_{nk}(x)\ ] ] @xmath103 they satisfy the differential - difference relations @xmath104 @xmath105x{\mathscr a}_{nk}(x)-(n - k)(n+k+1)x{\mathscr a}_{n , k+1}(x),\ ] ] and @xmath106 @xmath107x{\mathscr a}_{nk}(x)+(n+k)(n - k+1)x{\mathscr a}_{n , k-1}(x).\ ] ] we also find that @xmath108 is a solutions to the differential equation @xmath109y(x)=0.\ ] ] from ( [ j1 ] ) it follows that @xmath110 are the ( shifted ) chebyshev polynomials .
this classical set of highly demanded in approximation theory orthogonal polynomials is a singular subset of the system of polynomials @xmath111 where @xmath112 and @xmath113 .
the polynomials obey the orthogonality relation @xmath114 @xmath115 and @xmath116 .,width=226,height=136 ] they can be calculated using the three - term recurrence relation @xmath117 @xmath118[2(n - k)+1](4k+1){\mathscr t}_{n , k-1}(x)\ ] ] @xmath119{\mathscr t}_{nk}(x)\ ] ] @xmath120 we also find that @xmath121 is a solution to the differential equation @xmath122y(x)=0.\ ] ]
the classical orthogonal polynomials have a property of discrete orthogonality and the alternative orthogonal polynomials have it also ( see , for example , @xcite ) .
let us consider the sequence @xmath123 since 1 is not the term of the sequence , direct orthogonalization of ( [ e00 ] ) on the semi - axis with a chosen weight function does not result in gauss type quadrature and , consequently , in the discrete orthogonality of the sequence .
thus , inverse orthogonalization is the only procedure for constructing the exponential polynomials possessing such a property ( see , for example , @xcite ) .
this makes the set of orthogonal functions introduced below a unique system .
let @xmath124 be the polynomials of exponential function defined as @xmath125 from ( [ e0 ] ) it immediately follows that @xmath126 is the orthogonal system , such that @xmath127 @xmath128 for @xmath81 , @xmath129 .
properties of @xmath130 that are of interest for computations can be easily derived from the results of previous sections .
in particular , ( [ e1 ] ) leads to the three - term recurrence relationship , which sets the associated orthogonal function @xmath131 .
since @xmath131 is not integrable on the semi - axis with the weight function 1 , we do not include it in @xmath126 .
functions @xmath132 form the system with nice properties , because the zeros of @xmath131 are the abscissas of gauss type quadratures for the exponential functions on the semi - axis . in the next subsection we present two systems that are the exponential form of polynomials @xmath96 and @xmath133 .
let @xmath134 .
following ( [ e0 ] ) and ( [ lk0 ] ) we define @xmath135 the orthogonal exponential polynomials @xmath136 were first introduced in @xcite .
since they are of interest for approximation of a function in @xmath137 , we repeat some properties of @xmath96 in the exponential form as follows @xmath138 @xmath139 @xmath140 @xmath141 @xmath142{\cal e}_{nk}(t)\ ] ] @xmath143 @xmath144=-(k-1){\cal e}_{n , k-1}(t)+k{\cal e}_{nk}(t).\ ] ] let @xmath145 . following ( [ e0 ] ) and ( [ tlk0 ] )
we define @xmath146 for convenience , we also repeat some properties of @xmath133 in the exponential form as follows @xmath147 @xmath148 @xmath149 @xmath118[2(n - k)+1](4k+1){\cal e}_{n , k-1}^{\mathscr t}(t)\ ] ] @xmath150{\cal e}_{nk}^{\mathscr t}(t)\ ] ] @xmath151 the zeros of @xmath152 is easy to calculate , and this advantage can be a motivation for approximation of a function on the semi - axis by @xmath153 . employing exponential polynomials @xmath154
can be of interest for approximation of a function on the interval @xmath0 $ ] .
one may try to apply the system @xmath155 as the basis functions . here , the scale parameter @xmath156 is the maximum non - trivial zero of the zeros @xmath157 of the function @xmath158 @xcite .
another way is the use of parameters @xmath1 .
below we describe a formal construction of functions that involves computation of @xmath159 for desirable distribution of the zeros @xmath160 on the interval @xmath161 $ ] .
first , by considering @xmath162 , we introduce parameter @xmath163 that sets the abscissa of the weight function maximum , @xmath164 .
then , by subjecting choice of @xmath165 to the requirements @xmath166 we introduce parameter @xmath167 . in general ,
the equality in ( [ a1 ] ) can be reached if @xmath168 ; the inequality holds if @xmath169 are selected from the set of whole numbers or from a set of rational numbers of interest .
accordingly , for @xmath170 we define the system of functions @xmath171 and , for @xmath172 , @xmath173 similarly , we define the sequences @xmath174 and @xmath175 .
choice of @xmath176 sets distributions of the zeros of the functions under consideration , and @xmath177 .
the system of functions @xmath178 is easier to construct , in particular for @xmath179 .
an interpolating sequence of functions corresponding to approximation by z - functions with @xmath180 , @xmath181 and @xmath182 was employed in @xcite for solving a stiff initial value problem .
making use of the algorithm of bidirectional orthogonalization we obtained two systems of orthogonal functions , i.e. the alternative jacobi polynomials @xmath19 and the orthogonal exponential polynomials @xmath154 .
algorithmic properties of the systems are very similar to the properties of the jacobi polynomials .
this indicates that each of the systems may be a convenient tool for applied analysis .
in addition , we described two marginal sequences @xmath96 and @xmath133 , as well as their exponential counterparts @xmath136 and @xmath153 .
similarly , other important special systems of polynomials can be derived from the alternative jacobi polynomials .
it should be mentioned that there are regular orthogonal exponential polynomials that are often used for approximation on the semi - axis , but usually they are not considered as special functions sui generis ( say , in @xcite ) .
description of these orthogonal exponentials can be found in @xcite .
also , a special orthogonal sequence of exponentials that are defined on the semi - axis in a particularly appealing form is constructed in @xcite .
the concept that is developed in this paper is quite consistent with the maxim of c.g.j .
jacobi :
invert , always invert " .
many years ago we unknowingly subscribed to this dictum , and now we recognize its notability @xcite . 2 , _ on approximation by exponentials _ , annals of num .
maths , no 4 ( 1997 ) , pp .
39 - 54 . , _ sequences of exponential polynomials , which are orthogonal on the semi - axis _ , reports of the academy of sciences of the ukssr , ( doklady an ukssr ) , ser . a , no 1 ( 1987 ) , pp .
( in russian ) , _ a spectral method for the cauchy problem solving _ , in arxiv:1105.1563v1 [ math.na ] ( a corrected version of the paper published in proceedings of mcme international conference , problems of modern applied mathematics , n. mastorakis , ed . , wses press , 2000 , pp .
227232 ) . | sequences of orthogonal polynomials that are alternative to the jacobi polynomials on the interval @xmath0 $ ] are defined and their properties are established . an @xmath1-parameterized system of orthogonal polynomials of the exponential function on the semi - axis @xmath2 is presented .
two subsystems of the alternative jacobi polynomials , as well as orthogonal exponential polynomials are described .
two parameterized systems of discretely almost orthogonal functions on the interval @xmath0 $ ] are introduced .
alternative orthogonal polynomials , recurrence relations , orthogonal polynomials of the exponential function , discretely almost orthogonal functions 33c45 |
entanglement is one of the most important resources for quantum communication and information processing including quantum cryptography , teleportation , and superdense coding . on this account ,
the research on entanglement has considerably been developed and has improved the quantum information science in recent years . in particular , the quest for proper measures of entanglement has received a great deal of attention , and several measures of entanglement , such as the negativity and the entanglement of formation , have been proposed @xcite . peres - horodeckis criterion for separability @xcite leads a natural entanglement measure , called the _ negativity _
@xmath3 , defined by @xmath4 where @xmath5 is the partial transpose of a state @xmath6 in a hilbert space @xmath7 and @xmath8 is the trace norm . vidal and
werner @xcite defined the negativity of a state @xmath6 as @xmath9 , which corresponds to the absolute value of the sum of negative eigenvalues of @xmath5 , and which vanishes for separable states . for a maximally entangled pure state such as one of the bell states ,
this quantity is strictly less than one . in order for any maximally entangled pure state in @xmath1 quantum system to have the negativity one , it must be defined as .
we note that the negativity is an entanglement monotone @xcite under local quantum operations and classical communication ( locc ) @xcite , and that it is a measure of entanglement which can be computed effectively for any state .
however , although the positivity of the partial transpose is a necessary and sufficient condition for nondistillability in @xmath1 quantum system @xcite , there exist entangled states with positive partial transposition ( ppt ) in any bipartite system except for in @xmath10 and @xmath11 quantum systems @xcite , and hence it is not sufficient for the negativity to be a good measure of entanglement even in @xmath1 quantum system .
in this paper we consider @xmath1 quantum systems for @xmath2 , and exhibit a two - parameter class of states in @xmath1 quantum system , which can be obtained from an arbitrary state by means of locc and are invariant under all unitary operations of the form @xmath12 on @xmath1 quantum system .
we show that a state in the two - parameter class has a ppt if and only if the state is separable , so that the negativity can be a measure to quantify the amount of entanglement of states in the class . we remark that a finite dimensional truncation of a single two - level atom interacting with a single - mode quantized field @xcite , can be regarded as a @xmath1 quantum system
. the _ entanglement of formation _ is defined to be the convex - roof extension of the pure - state entanglement , that is , the minimum average of the pure - state entanglement over all ensemble decompositions of a given state @xmath6 , @xmath13 here , the pure - state entanglement @xmath14 is defined as the entropy of subsystem @xmath15 , @xmath16 , @xmath17 is the von neumann entropy .
since the concept of the pure - state entanglement @xmath14 has most widely been accepted as an entanglement measure for pure states and the entanglement of formation is its natural extension as seen in , the entanglement of formation is one of the best measures for bipartite quantum systems which have been known so far .
nevertheless , there is no known explicit formula for the entanglement of formation of states in a general quantum system except for states in @xmath10 quantum system @xcite , the isotropic states and the werner states in @xmath18 quantum system @xcite , and states of the specific form @xcite . for @xmath1 quantum system ,
only a lower bound on the entanglement of formation is given by decomposing a @xmath1 dimensional hilbert space into many @xmath19 dimensional subspaces @xcite .
in this paper we present a lower bound and a tight upper bound on the entanglement of formation for the two - parameter class of states .
the explicit calculations of the negativity and bounds on the entanglement of formation show that the entanglement of formation of any state in the two - parameter class can not exceed its negativity .
this paper is organized as follows : in we exhibit a two - parameter class of states and a procedure involving only locc which transform an arbitrary state into one of states in the class , and show that the states in the class are invariant under all unitary operations of the form @xmath12 . in
we explicitly calculate the negativity , and a lower bound and a tight upper bound on the entanglement of formation for the two - parameter class , and compare its negativity with the entanglement of formation .
finally , in we summarize our results and discuss a generalization of the two - parameter class into a higher dimensional system .
we consider the following class of states with two real parameters @xmath20 and @xmath21 in @xmath1 quantum system : @xmath22 where @xmath23 is an orthonormal basis for @xmath1 quantum system , @xmath24 and the parameter @xmath25 is dependent on @xmath20 and @xmath21 by the unit trace condition , @xmath26 from the unit trace condition in we can readily obtain the domain for the parameters @xmath20 and @xmath21 , @xmath27 and @xmath28 .
we note that the states of the form @xmath29 for @xmath30 are equal to werner states @xcite in @xmath10 quantum system , that the states are entangled and distillable if and only if @xmath31 , and that for @xmath32 , @xmath33 where @xmath34 is the binary shannon entropy .
we are going to show now that an arbitrary state @xmath6 in @xmath1 quantum system can be transformed to a state of the form @xmath0 in by locc .
in other words , we will show that there exist unitary operators @xmath35 and probabilities @xmath36 such that @xmath37 for some @xmath27 and @xmath28 , using the method similar to those presented by bennett _ et al
. _ @xcite and dr _ et al .
_ @xcite .
we define the operation @xmath38 as @xmath39 where @xmath40 .
we first perform @xmath41 with probability @xmath42 , while with probability @xmath42 no operation is performed , that is , @xmath43 let us now define the operation @xmath35 by @xmath44 for @xmath45 , and then for each @xmath46 , perform @xmath47 with probability @xmath42 , while applying the identity operation with probability @xmath42 , respectively . here , we remark that @xmath48 .
we now perform @xmath49 with probability @xmath42 as in , and then perform the swap operator @xmath50 ( @xmath51 for @xmath52 ) with probability @xmath42 .
then we obtain a state of the following form : @xmath53 let @xmath54 be the unitary operator defined by @xmath55 , @xmath56 , @xmath57 , @xmath58 , @xmath59 , @xmath60 , and @xmath61 . now we perform the following operation : @xmath62 here , we also remark that @xmath63 for any @xmath64 .
then a state in now has the form @xmath65 let @xmath66 be the unitary operator defined as @xmath67 , @xmath68 , and @xmath69 for @xmath70 .
after performing as follows : @xmath71 let us perform the sequence of the previous operations again .
then one can easily check that one of states with two parameters in is obtained , that is , there exist unitary operators @xmath35 and probabilities @xmath36 satisfying , and that furthermore if a state @xmath6 is given by @xmath72 then the two - parameter state transformed by the above procedure becomes @xmath0 in where @xmath73 and @xmath74 . noting that @xmath75 for any @xmath76 in the group @xmath77 defined in , we can readily show that @xmath0 is invariant under all @xmath12 , that is , for any @xmath78 , @xmath79 we now define the ( @xmath12)-_twirling _ superoperator @xmath80 as @xmath81 and @xmath82 is the normalized haar measure on @xmath83 .
then it follows from and that @xmath84 for some @xmath20 and @xmath21 , and that @xmath85 , respectively .
we remark that the negativity and the entanglement of formation are entanglement monotones @xcite .
thus , it follows that the negativity and the entanglement of formation for a given @xmath6 are not less than those for @xmath86 , respectively .
in this section , we consider two measures of entanglement for @xmath0 , the negativity and the entanglement of formation , and explicitly calculate the value of its negativity and bounds on its entanglement of formation . in order to calculate the negativity of @xmath0
, we have to compute its partial transpose:@xmath87 since @xmath88 if and only if @xmath89 is positive , the negativity of @xmath0 is @xmath90 , whose graph is shown in
. then the domain of the parameters @xmath20 and @xmath21 for @xmath0 consists of two triangular regions , the ppt region satisfying @xmath91 and the nonpositive partial transposition ( npt ) region satisfying @xmath92 , as shown in .
we note that all states in the ppt region are separable since three states , @xmath93 , @xmath94 , and @xmath95 , corresponding to the vertices of the ppt region are separable and all states in the region are convex combinations of the three states .
we now consider the entanglement of formation for @xmath0 .
even though it has not been known whether the explicit formula of the entanglement of formation for states in @xmath1 quantum system can be computed or not , one can readily compute one of its lower bounds @xcite , @xmath96 where @xmath97 @xmath98 is the wootters cuncurrence @xcite in @xmath10 dimensional subsystem which is supported by the bases @xmath99 , @xmath100 , @xmath101 , and @xmath102 .
it is straightforward to check that the lower bound in is @xmath103 since @xmath104 and @xmath105 unless @xmath106 and @xmath107 . for @xmath108 quantum system ,
the lower bound on @xmath109 is plotted in . in order to obtain a tight upper bound of @xmath109 , we consider two - parameter states satisfying @xmath110 , that is , the states corresponding to the line through two points @xmath111 and @xmath112 in .
then the states are of the following form : @xmath113 we note that is an eigenvalue decomposition of a state @xmath114 , and that @xmath115 in view of the convexity of @xmath116 .
let @xmath117 be an optimal decomposition for its entanglement of formation when @xmath118 .
by hughston , jozsa and wootters theorem @xcite , there exists a @xmath119 unitary matrix @xmath76 , @xmath120 being greater than or equal to @xmath121 , the rank of @xmath122 , such that for each @xmath123 @xmath124 where @xmath125 .
thus it can be obtained from that @xmath126 where @xmath127 then it follows from and the concavity of @xmath128 that @xmath129 since the von neumann entropy of a pure state is vanished . hence , from and the unitarity of @xmath76 , we obtain the following inequality : @xmath130 by two inequalities ( [ eq : less_than_gamma ] ) and ( [ eq : eof_s ] ) we conclude that @xmath131 .
we now consider the states corresponding to @xmath132 in the interior of the npt region , that is , @xmath132 satisfying @xmath133 . by virtue of the convexity of the entanglement of formation , for a given @xmath132
the following inequality can be obtained : @xmath134/(n-2),t[2n_{(\alpha,\gamma)}-1]+(1-t)n_{(\alpha,\gamma)}\right)}\right ) \nonumber\\ & \le & t\cdot e_f\left(\rho_{\left([1-n_{(\alpha,\gamma)}]/(n-2),2n_{(\alpha,\gamma)}-1\right)}\right)\nonumber\\ & & + ( 1-t)\cdot e_f\left(\rho_{\left(0,n_{(\alpha,\gamma)}\right)}\right)\nonumber\\ & = & t\cdot [ 2n_{(\alpha,\gamma)}-1 ] + ( 1-t)\cdot \mathcal{e}(2n_{(\alpha,\gamma ) } ) \label{eq : upperbound1}\end{aligned}\ ] ] where @xmath135 , and @xmath136 is chosen by @xmath137/(n-2),\nonumber\\ \gamma & = & t\cdot[2n_{(\alpha,\gamma)}-1]+(1-t)\cdot n_{(\alpha,\gamma)}. \label{eq : t}\end{aligned}\ ] ] choosing the appropriate @xmath136 satisfying , we are straightforwardly able to calculate the following upper bound on @xmath109:@xmath138 for @xmath108 quantum system the upper bound ( [ eq : upperbound2 ] ) on @xmath109 is shown in .
we remark that the entanglement of formation and its upper bound ( [ eq : upperbound2 ] ) have the same values for the states corresponding to three edges of the npt region in .
therefore , we can not derive any upper bound tighter than the upper bound ( [ eq : upperbound2 ] ) from the method using the convexity of the entanglement of formation as in . from this viewpoint
, we can say that the upper bound ( [ eq : upperbound2 ] ) is a tight upper bound on @xmath139 .
we note that @xmath140 and that @xmath141 for all @xmath28 .
thus , it follows that the entanglement of formation of any @xmath0 can not exceed its negativity , that is , @xmath142 for any @xmath0 .
in this paper we exhibited a two - parameter class of states @xmath0 in @xmath1 quantum system , found that an arbitrary state can be transformed into @xmath0 by means of locc , and showed that @xmath0 is invariant under unitary operations of the form @xmath12 on @xmath1 quantum system .
we finally investigated the entanglement for @xmath0 , by computing two measures of entanglement , the negativity and the entanglement of formation . for a higher dimensional quantum system , that is , an @xmath143 quantum system ( @xmath144 ) , we can exhibit a two - parameter class given by @xmath145 where @xmath146 and @xmath147 furthermore , we can show that any state in @xmath143 quantum system can be transformed into a state of the form of by locc , and that any state in this class is ( @xmath12)-invariant for all unitary @xmath76 in the group @xmath148 defined in .
since every state in this class has properties analogous to those of @xmath0 , one could investigate its entanglement . | we exhibit a two - parameter class of states @xmath0 , in @xmath1 quantum system for @xmath2 , which can be obtained from an arbitrary state by means of local quantum operations and classical communication , and which are invariant under all bilateral operations on @xmath1 quantum system .
we calculate the negativity of @xmath0 , and a lower bound and a tight upper bound on its entanglement of formation .
it follows from this calculation that the entanglement of formation of @xmath0 can not exceed its negativity . |
this article is concerned with the inverse spectral problem for metrics of revolution on @xmath0 .
we will assume that our metrics are real analytic and belong to a class @xmath1 of rotationally invariant metrics which are of ` simple type ' and which satisfy some generic non - degeneracy conditions ( see definition ( 0.1 ) ) .
in particular , we will assume they satisfy the generalized ` simple length spectrum ' condition that the length functional on the loop space is a clean bott - morse function which takes on distinct values on distinct components of its critical set ( up to orientation ) .
denoting by @xmath2 the spectrum of the laplacian @xmath3 , our main result is the following : * theorem i * _ spec : @xmath4 is 1 - 1 . _ thus , if @xmath5 are isospectral surfaces of revolution in @xmath1 , then @xmath6 is isometric to @xmath7
. it would be very desirable to strengthen this result by removing the assumption that @xmath8 , thereby showing that metrics in @xmath1 are spectrally determined within the entire class of analytic metrics on @xmath0 with simple length spectra .
the only metric on @xmath0 presently known to be spectrally determined in this sense is the standard one ( which is known to be spectrally determined among all @xmath9 metrics ) .
a metric @xmath7 satisfying @xmath10 for some @xmath11 must have many properties in common with a surface of revolution of simple type ; it would be interesting to explore whether it must necessarily be one .
let us now be more precise about the hypotheses .
first , we will assume that there is an effective action of @xmath12 by isometries of @xmath13 .
the two fixed points will be denoted @xmath14 and @xmath15 will denote geodesic polar coordinates centered at @xmath16 , with @xmath17 some fixed meridian @xmath18 from @xmath16 to @xmath19 .
the metric may then be written in the form @xmath20 where @xmath21 \rightarrow { { \rm i\!r}}^+$ ] is defined by @xmath22 , with @xmath23 the length of the distance circle of radius @xmath24 centered at @xmath16 .
for any smooth surface of revolution , the function @xmath25 satisfies @xmath26 and two such surfaces @xmath27 @xmath28 are isometric if and only if @xmath29 and @xmath30 or @xmath31 we will then assume that the metrics belong to the following class @xmath32 of _ simple analytic _ surfaces of revolution [ bl ] : * ( 0.1 ) definition * _ @xmath32 is the moduli space of metrics of revolution @xmath33 with the properties : + ( i ) @xmath6 ( equivalently @xmath25 ) is real analytic ; + ( ii ) @xmath25 has precisely one non - degenerate critical point @xmath34 , with @xmath35 , corresponding to an ` equatorial geodesic ' @xmath36 ; + ( iii ) the ( non - linear ) poincare map @xmath37 for @xmath36 is of twist type ( cf . 1 ) .
+ we denote by @xmath38 the subset of metrics with ` simple length spectra ' in the sense above . _ regarding ` simple length spectra , ' we recall that the closed geodesics of a surface of revolution come in one - parameter families of a common length , filling out invariant torii @xmath39 for the geodesic flow .
the canonical involution @xmath40 takes @xmath39 to its ` time reversal ' @xmath41 , and takes the closed geodesics of @xmath39 to their reversals on @xmath41 .
a closed geodesic and its reversal have the same length , so the length spectrum is automatically double except for the length @xmath42 of the torus of meridians , which is @xmath43-invariant .
the simple length spectrum hypothesis is that up to time reversal , the common lengths of the closed geodesics on distinct torii are distinct ( cf . definition 1.2.2 ) .
in fact , it would be sufficient for the proof that the length @xmath42 of the ` meridian ' closed geodesics is not the length of closed geodesics on any other torus . in any case , it is not hard to show that @xmath1 is residual in @xmath32 ( cf .
proposition 1.2.4 ) .
the condition ( iii ) actually appears in the proof in the following way : the quadratic coefficient ` @xmath44 ' of the classical birkhoff normal form of the metric @xmath45 at the torus of meridian geodesics must be non - vanishing ( see definition ( 1.4.5 ) for the precise meaning of @xmath46 ) .
this condition is used in proposition ( 4.1.2 ) and corollary ( 4.1.3 ) to evaluate the wave invariants for the meridian torus . as will be explained further below
, the main purpose of the non - degeneracy and simplicity conditions is to insure that there are global action - angle variables for the geodesic flow .
these conditions rule out several types of surfaces of revolution : first , they rule out zoll surfaces of revolution , which are degenerate in every possible sense .
it is indeed unknown at this time whether real analytic zoll surfaces of revolution are determined by their spectra .
they also rule out ` peanuts of revolution ' ( which have hyperbolic waists ) and other natural rotational surfaces such as liouville torii . to our knowledge
, the strongest prior result on the inverse spectral problem for surfaces of revolution is that of bruning - heintze [ b.h ] : smooth surfaces of revolution with a mirror symmetry thru the @xmath47 plane are spectrally determined among metrics of this kind .
there are also a number of proofs that a surface of revolution is determined by the joint spectrum of @xmath48 and of @xmath49 , the generator of the rotational symmetry [ kac][b][gur ] .
the method - of - proof of bruning - heintze was based on the observation that the invariant spectrum can be heard from the entire spectrum . hence by separating variables the problem
can be reduced to the inverse spectral problem for 1d even singular sturm - liouville operators , which was solved by gelfand - levitan and marchenko .
our proof of the main theorem is based on different kind of method , and the inverse result presented here is hopefully just one illustration of it .
we begin with the facts that simple surfaces of revolution are completely integrable on both the classical and quantum levels and that the laplacian has a global quantum normal form in terms of action operators .
we then study the trace of the wave group and prove that from its singularity expansion we can reconstruct the global quantum normal form .
finally we show that this normal form determines the metric .
this approach is suggested by the recent inverse result of guillemin [ g.1 ] , which shows that the microlocal normal form of @xmath48 around each non - degenerate elliptic closed geodesic can be determined from the wave invariants ( see also [ z.1,2 ] ) .
however , there is but one non - degenerate closed geodesic on a simple surface of revolution , so the direct application of this inverse result does not take full advantage of the situation .
rather , it is natural to start from the fact that the wave group is completely integrable in the following strong sense : namely , it commutes with an effective action of the torus @xmath50 by fourier integral operators on @xmath51 .
that is , there exist global _
action operators _ @xmath52 and a polyhomogeneous symbol @xmath53 of degree 1 on @xmath54 such that @xmath55 .
this is the global quantum birkhoff normal form alluded to above .
our principal tool is the following inverse result : * main lemma * _ the wave trace invariants of @xmath33 with @xmath56 determine the quantum normal form @xmath57 _
this lemma does not actually require that @xmath33 be a surface of revolution , but only that the geodesic flow is toric integrable , i.e. commutes with an effective hamiltonian torus action .
it immediately implies that the principal symbol @xmath58 of @xmath59 is a spectral invariant .
since @xmath58 is essentially a global birkhoff normal form for the metric , the wave invariants determine the symplectic equivalence of the geodesic flow .
thus we have : * corollary 1 * _ from the wave trace invariants of @xmath33 with @xmath60 we can determine the symplectic equivalence class of @xmath61 .
_ corollary 1 does not however determine the isometry class of a general @xmath62 : as will be discussed in 1 ( see also [ c.k ] ) , simple surfaces of revolution are not symplectically rigid unless they are mirror symmetric .
otherwise put , recovery of the classical birkhoff normal form only determines the _ even part _ of the metric in the following sense : * corollary 2 * _ a metric @xmath63 may be written in the form @xmath64 ^ 2 du^2 + [ sin u]^2 d\theta^2 $ ] ( 1 , [ besse ] ) . from the wave invariants
one can determine the _ even part _ of @xmath65 .
in particular if @xmath6 is mirror symmetric , one can determine @xmath6 among mirror symmetric metrics in @xmath1 from its spectrum .
_ it is interesting to note that symplectic rigidity in the mirror - symmetric case gives a new and self - contained proof of the bruning - heintze result , without the use of marchenko s inverse spectral theorem for singular sturm liouville operators .
although this result is superceded by the theorem , it may have some future relevance to other inverse problems . to complete the proof of the theorem
, we therefore have to study the subprincipal terms in @xmath53 .
the result is : * final lemma * _ from @xmath53 one can determine the isometry class of @xmath6 . _
it is in this last lemma that we use in full that @xmath33 is a surface of revolution rather than just a surface with toric integrable geodesic flow .
we also use in full that @xmath53 is a global quantum normal form rather than a microlocal one at @xmath36 . in subsequent work we will investigate the analogues of the final lemma for the microlocal normal form at @xmath36 , for more general toric integrable metrics and for metrics isospectral to toric integrable ones . to close this introduction
, we discuss some background and some open problems connected with this work .
the principal motivation for studying the inverse spectral problem for surfaces of revolution is its simplicity .
there are to date very few inverse results which determine a metric from its laplace spectrum within the entire class of metrics , or even within concrete infinite dimensional families .
to our knowledge , only the standard @xmath66 for @xmath67 and flat 2-torii are known to be spectrally determined .
hence it is desirable to have a simple model of how an inverse result might go .
a second motivation is a somewhat loose analogy between surfaces of revolution and planar domains .
namely , in both cases the unknown is a function of one variable ( the profile curve , resp .
the boundary ) which completely determines the first return times and angles of geodesics emanating from a transversal .
that is , paths of billiard trajectories ( broken geodesics ) on a bounded planar domain are determined from collisions with the boundary , while paths of geodesics on surfaces of revolution are determined from collisions with a meridian ( or with the equator ) . in analytic cases , it is plausible that the unknown function may be determined in large part by the spectrum of first return times from the local transversal . in the case of an analytic surface of revolution , this length spectrum determines the corresonding birkhoff normal form and hence by corollary 2 it determines the even part of the profile curve .
similarly , in the case of an analytic plane domain , it is proved in [ cv.2 ] that the even part of the boundary of an analytic domain may be determined from the birkhoff normal form of the billiard map at a bouncing ball orbit .
hence , there is a similarity in the relation between the unknown function and the local classical birkhoff normal forms .
it is interesting to observe in this context that the rigidity result of colin de verdiere [ cv.2 ] is quite analogous to the inverse result of bruning - heintze . some immediate open problems : first , there is the symplectic conjugacy problem mentioned above .
second , can one relax analyticity to smoothness in the theorem above ?
this is likely to follow from a more intensive analysis of the wave invariants .
third , can one extend it to other ` non - simple ' types of surfaces of revolution ?
the main obstacle is that one will generally not have global action - angle variables or global quantum normal forms .
what about completely integrable systems in higher dimensions ?
finally , we would like to thank d.kosygin for several helpful conversations on this paper , and b.kleiner for giving us up to date information about the status of the conjugacy problem for geodesic flows on surfaces of revolution .
from a geometric ( or dynamical ) point of view , the principal virtue of metrics in @xmath32 is described by the following : * ( 1.1.1 ) proposition * _ suppose @xmath6 is a real analytic metric of revolution on @xmath0 such that @xmath25 has precisely one non - degenerate critical point at some @xmath68 then the hamiltonian @xmath69 on @xmath70 is completely integrable and possesses global real analytic action - angle variables .
_ * proof * : the complete integrability of @xmath45 ( i.e. of the geodesic flow ) is classical , and follows from the existence of the clairaut integral @xmath71 . since the poisson bracket @xmath72 , the geodesics are constrained to lie on the level sets of @xmath73 ; and since both @xmath45 and @xmath73 are homogeneous of degree one , the behaviour of the geodesic flow is determined by its restriction to @xmath74 . with the assumption on @xmath25 , the level sets are compact and
the only critical level is that of the equatorial geodesics @xmath75 ( traversed with either orientation ) .
the other level sets are well - known to consist of two - dimensional torii .
( had we allowed the existence of at least two critical points in @xmath25 , there would exist a saddle level , i.e. an embedded non - compact cylinder ) .
the existence of global action - angle variables follows from the general results of [ d][g.s ] and have been constructed explicitly for simple surfaces of revolution in [ cv.1 ] .
the general formula is as follows : let @xmath76 be the moment map of the hamiltonian @xmath77-action defined by the geodesic flow and by rotation .
the singular set of @xmath78 is the closed conic set @xmath79 , i.e. @xmath80 is the cone thru the equatorial geodesic ( in either orientation ) .
the image of @xmath80 is the boundary of @xmath81 ; the map @xmath82 is a trivial @xmath83 bundle over the open convex cone @xmath84 ( the interior of @xmath81 ) . for each @xmath85 ,
let @xmath86 denote the homology of the fiber @xmath87 this lattice bundle is trivial since @xmath81 is contractible , so there exists a smoothly varying homology basis @xmath88 which equals the unit cocircle @xmath89 together with the fixed closed meridian @xmath18 when @xmath90 is on the center line @xmath91 .
the action variables are given by [ cv.1 , 6 ] @xmath92 where @xmath93 are the extremal values of @xmath24 on the annulus @xmath94 ( with @xmath95 the standard projection ) .
on the torus of meridians in @xmath96 , the value of @xmath97 equals @xmath98 and it equals one on the equatorial geodesic .
so extended , @xmath99 are smooth homogeneous functions of degree 1 on @xmath70 , and generate @xmath100-periodic hamilton flows .
it follows that the pair @xmath101 generates a global hamiltonian torus ( @xmath83)-action commuting with the geodesic flow .
the singular set of @xmath102 equals @xmath103 , corresponding to the equatorial geodesics . the map @xmath104 for the torus @xmath105 with @xmath106 and let @xmath107 denote the closure of @xmath108 as a convex cone .
the symplectically dual angle variables @xmath109 then give , by definition , the flow times ( mod @xmath100 ) along the orbits of @xmath110 from a fixed point on @xmath111 , which we may take to be the unique point lying above the intersection of the equator and the fixed meridian on @xmath111 with the geodesic pointing into the northern hemisphere .
so far , we have only assumed the metric to be @xmath9 .
we now observe that if @xmath6 is real analytic , then so are @xmath112 this is obvious in the case of @xmath113 and follows from the explicit formula ( 1.1.3 ) for @xmath114 ' '' '' since the metric norm function @xmath45 commutes with @xmath99 , it may be expressed as a function @xmath115 of the action variables .
hamilton s equations for the geodesic flow then take the form @xmath116 where @xmath117 is the _ frequency vector _ of the torus @xmath118 with action coordinates @xmath119 . the geodesic flow on @xmath118 is then given by @xmath120 so that all the geodesics are quasi - periodic in action - angle coordinates .
the frequency vector @xmath121 is homogeneous of degree 0 on @xmath122 , and hence is constant on rays of torii @xmath123 . to break the @xmath124 symmetry we restrict to the level set @xmath125 in action space and view the frequency vector as the map : @xmath126 since @xmath127 the frequency map is more or less the gauss map of @xmath128 ( although it is not normalized to be of unit length ) . as a map of the global action space ,
the frequency map is the legendre transform associated to @xmath129 ( cf .
[ f.g , p.338 ] ) . *
( 1.1.6 ) definition * _ we say that the simple surface of revolution @xmath33 is _ globally non - degenerate _ if @xmath130 is an embedding .
_ this is the natural homogeneous analogue of the non - degeneracy condition of [ f.g , loc.cit ] to the effect that the legendre transform be a global diffeomorphism , and has previously been studied in some detail by bleher in the setting of simple surfaces of revolution [ bl,6 ] .
as will be seen in 1.3 , the curve @xmath128 is the graph of a smooth function of the form @xmath131 in the cone @xmath132 and the non - degeneracy condition ( 1.1.6 ) will follow as long as @xmath133 is a convex or concave function .
in fact , in the proof of the theorem we will only need to use that @xmath128 is non - degenerate at the one point @xmath134 this is sufficient because we assume the metric to be real analytic .
[ figure1 ] we now come to the definition of length spectrum and simple length spectrum for a completely integrable geodesic flow .
we first observe that the orbit thru @xmath135 is periodic of period @xmath136 if and only if @xmath137 for some @xmath138 the minimal positive such @xmath136 will be called the primitive period ; the corresponding @xmath139 is known as the vector of winding numbers of the torus @xmath118 .
@xmath139 parametrizes the homology class of the closed orbit @xmath140 since the latter has the form @xmath141 relative to the homology basis @xmath142 . due to the homogeneity ,
the period and vector of winding numbers are constant on the ray @xmath143 . by euler
s formula we then have @xmath144 hence the length is given in terms of the winding vector by @xmath145 or simply @xmath146 on the unit tangent bundle @xmath147 .
it is clear that the periodicity condition @xmath148 is independent of @xmath149 .
hence , all of the geodesics on @xmath118 are closed if any of them are .
( this also follows , of course , from the transitivity of the torus action on each invariant torus . )
we therefore say : * ( 1.2.2 ) definition * _ _
\(a ) a torus @xmath118 is a _ periodic torus _ if all the geodesics on it are closed .
+ ( b ) the _ period _ @xmath136 of the periodic torus is then the common period of its closed geodesics .
+ ( c ) the _ length spectrum _ @xmath150 of the completely integrable system is the set of these lengths .
+ ( d ) the completely integrable system has a _ simple length spectrum _ if there exist a unique periodic torus ( up to time reversal ) of each length @xmath151 . in the last statement
( d ) we are referring to the canonical involution @xmath152 , which reverses the orientation of the geodesics .
it is obvious that if @xmath118 is a periodic torus of period @xmath136 , then so is @xmath153 .
let @xmath154 denote the set of periodic points for the geodesic flow on @xmath155 , i.e. the set of points which lie on a closed geodesic .
it is a union of periodic torii in @xmath155 together with points along the equatorial geodesics ( which are degenerate torii ) .
the set of all periodic points in @xmath156 is then equal to @xmath157 since the invariant torii are parametrized by the points @xmath158 of action space , it is convenient to parametrize @xmath159 by a subset of the level set @xmath160 * ( 1.2.3 ) definition*_the set of points @xmath161 such that @xmath162 will be called , with a slight abuse of notation , the set of _ periodic _ points on @xmath163 and will be denoted by @xmath164 .
that is , @xmath165 _ the following proposition will be needed later on ( proposition ( 4.1.4 ) . * ( 1.2.4 ) proposition * _ if @xmath33 is non - degenerate ( 1.1.6 ) , then @xmath164 is dense in @xmath163 . _ * proof * : let @xmath166 equal the image of @xmath164 under the frequency map @xmath167 . then by definition , @xmath168 ( for ` rational points ' ) is the projection to the curve @xmath169 of the integer lattice in @xmath170 .
it is clear that @xmath168 is a dense set in @xmath169 ; since @xmath167 is an embedding , @xmath164 is dense in @xmath171 ' '' '' the next proposition shows that our inverse result is valid for a residual set of simple analytic surfaces of revolution . *
( 1.2.5 ) proposition * _ let @xmath38 be the subset of metrics with simple length spectra .
then @xmath1 is a residual subset of @xmath32 . _ * proof * : if @xmath172 , then there exist @xmath173 and @xmath174 such that @xmath175 so @xmath176 implies that @xmath177 hence that the @xmath119-coordinates of @xmath178 are dependent over the rationals . since the length spectrum moves continuously and non - trivially under deformations in @xmath32 , such a dependence for fixed @xmath179 can only hold on a closed nowhere dense set .
the proposition follows . ' '' '' let us consider more carefully the geometric interpretation of the clairaut integral on a torus @xmath180 since @xmath113 and @xmath181 are independent commuting coordinates , and since there are global action - angle variables , the different invariant torii in @xmath96 are parametrized by the values @xmath182 $ ] of the clairaut integral along @xmath163 . indeed by ( 1.1.3 ) , the second action coordinate @xmath97 is determined from @xmath113 on @xmath128 by the formula @xmath183 where @xmath184 are the roots of @xmath185 the torus in @xmath96 with @xmath186 is therefore @xmath187 , which we will denote simply by @xmath188 the projection of @xmath189 to @xmath0 is an annulus of the form @xmath190 .
the geodesics on @xmath191 project to @xmath0 as almost periodic curves oscillating between the two extremal parallels @xmath192 and @xmath193 we observe then that @xmath189 contains a unique geodesic @xmath194 which passes thru the intersection of the reference geodesics @xmath18 and @xmath36 ; @xmath195 equals @xmath196 where @xmath197 is the angle between @xmath194 and @xmath18 . in other words
, @xmath198 is the common angle with which the geodesics on @xmath189 intersect the meridians as they pass thru the equator in the direction of the northern hemisphere . since the length @xmath199 of the equator @xmath36 equals @xmath200 , and since this length is a symplectic invariant of the geodesic flow , the coordinates @xmath195 and @xmath201
are related in an essentially universal fashion .
hence , either @xmath113 or @xmath202 could be used as an action coordinate on @xmath163 ; @xmath201 is perhaps more geometric but @xmath113 is more convenient in calculations .
the picture is the same for any invariant torus @xmath118 : under the @xmath124 action on @xmath203 , it scales to a torus @xmath204 and all features of its geometry are identical to that of @xmath189 .
thus it carries a unique ( parametrized ) geodesic @xmath205 such that @xmath206 is at the intersection @xmath207 , and so that @xmath208 points to the northern hemisphere .
the initial angle variables of @xmath209 are therefore @xmath210 at time @xmath211 we denote the angle variables of @xmath212 by @xmath213 where @xmath214 essentially measures the meridianal angle and @xmath215 measures the equatorial angle .
we then introduce the following ` first return times ' : * ( 1.3.2 ) definition * _ + ( ei ) the equatorial first return time is the minimal time @xmath216 such that @xmath217 ; + ( eii ) the equatorial first return angle @xmath218 is the change in angle along the equator of a geodesic on @xmath118 leaving @xmath207 at @xmath219 , upon its first return time to @xmath36 ; + ( mi ) the meridianal first return time is the minimal time @xmath220 such that @xmath221 + ( mii ) the meridianal first return angle is the angle change @xmath222 along the meridian of a geodesic on @xmath18 leaving @xmath207 at @xmath219 , upon its first return time to @xmath18 . _ the terminology ` first return time ' is taken from dynamics .
note that @xmath223 is the equation of the curve on @xmath118 which lies over the equator .
hence , @xmath224 is the time of first return of @xmath225 to the equator in the direction of the northern hemisphere .
this is actually the second time of intersection of @xmath205 with the equator , the first one occurring when @xmath226 is heading to the southern hemisphere .
this intersection is not in the projection of @xmath223 .
similarly , @xmath227 is the equation of the fixed meridian @xmath18 , and so @xmath228 is the time of first return to the arc @xmath18 ( half of the closed geodesic ) .
[ figure1 ] the following gives some relations between the various angles and return times . * ( 1.3.3 ) proposition * _ let @xmath229 be the frequency vector of the invariant torus @xmath118 . then : + ( a ) @xmath230 ; + ( b ) @xmath231 ; + ( c ) @xmath232 _ * proof * : the equation of the geodesic @xmath212 on @xmath118 beginning at @xmath233 in angle variables is @xmath234 . when the torus is projected to the base
, then @xmath214 measures the meridianal angle and @xmath215 measures the equatorial angle .
so by definition of return times , [ cols= " < " , ] the statements in the proposition follow immediately . ' '' '' these first return times ( and angles ) are closely related to the ( non - linear ) poincare maps of the geodesic flow .
we recall that for each closed geodesic @xmath140 , the poincare map @xmath235 is defined as the first return map of the geodesic flow , restricted to a symplectic transversal @xmath236 ( surface of section ) .
it is well - known @xmath237 is a symplectic map [ k ] .
since the symplectic form on a cotangent bundle equals @xmath238 ( with @xmath201 the action form ) , it follows that @xmath239 is a closed 1-form on @xmath240 since @xmath241 may be assumed contractible , it follows that there exists a function @xmath242 such that @xmath243|_{s_{\gamma } } = d \tau_{\gamma}.$ ] since the integral of @xmath201 over an arc of a unit speed geodesic just gives its length , @xmath242 is the first return time of geodesics near @xmath140 to @xmath240 in particular , let @xmath244 be the equator ( in one of its orientations ) , and let @xmath245 denote a symplectic transversal at the point @xmath246 thru the fixed meridian @xmath18 . since @xmath18 is transverse to the equator , we may define @xmath245 to consist of a small variation of @xmath247 moved up and down a small arc of @xmath18 .
we see then that the first return time @xmath248 is precisely the first return time @xmath228 defined above .
we also observe that the foliation of @xmath249 by invariant torii restricts to a foliation of @xmath245 by invariant circles for @xmath37 , closing in on the center point where @xmath36 intersects @xmath245 .
as noted above , the action coordinate @xmath186 gives a natural action ( radial ) coordinate on @xmath241 . since @xmath236 , the @xmath124 homogeneity is broken and we may reformulate the twist condition ( 0.1 ( iii ) ) as follows : * ( 1.3.4 ) proposition * _ the poincare map @xmath37 is a twist map of @xmath245 iff @xmath250 , where @xmath251 _ * proof * : the coordinates @xmath252 restrict to a system of symplectic coordinates on @xmath245 , in terms of which the poincare map takes the form @xmath253 by definition it is a twist map if @xmath254 in a neighborhood of @xmath255 in @xmath245 . ' '' '' for background on the twist condition in a related context see [ f.g][p ] .
the situation for the other periodic orbits is different since they come in one - parameter families .
thus , for the ( closed ) meridian geodesic @xmath18 , a transversal @xmath256 is given by the equator @xmath36 and a small variation of @xmath257 along it .
the foliation by invariant torii restricts to @xmath256 to a foliation by invariant lines ( non - closed curves ) , including the curve of closed geodesics thru @xmath18 .
the poincare map @xmath258 is then of parabolic type ; the first return time is @xmath259 above . for the inverse problem it will be necessary to have expressions for these return times and angles in terms of the metric
this will also make the twist condition more transparent .
for ease of quotation , it is convenient to make a change of dependent and independent variables , following [ besse ] and darboux [ d ] .
equivalent expressions in the original polar coordinates can also be easily derived , and will be given below . *
( 1.3.5 ) proposition * _ suppose that @xmath33 is a simple surface of revolution
. then there exists a coordinate system @xmath260 on @xmath261 , with @xmath262 and a smooth function @xmath65 on @xmath263 $ ] such that @xmath264 and such that @xmath265 ^ 2 du^2 + sin^2 u d\theta^2.\ ] ] _ * proof * : first , define @xmath266 \\ \pi - sin^{-1 } ( a ( r ) ) & r \in [ r_o , l ] \end{array } \right .\ ] ] and @xmath267})^{-1 } ( \sqrt{1 - v^2 } ) & v \in [ 0 , 1 ] \\ ( a |_{[r_o , l]})^{-1 } ( \sqrt{1 - v^2 } ) & v \in [ -1 , 0 ] \end{array } \right.\ ] ] then define @xmath268 by @xmath269},\;\;\;(v\not= 0 ) \\
f(0 ) = \frac{1}{- a^{''}(r_o ) } \end{array } \right .\ ] ] since @xmath270 we have @xmath271 dr = cos u du$ ] and @xmath272 the smoothness properties of @xmath65 follow from those of @xmath25 and from the fact that @xmath273 and @xmath274 have the same qualitative shapes [ besse , loc.cit . ] . ' '' '' the geometric result is the following [ besse , theorem 4.11 ] : * ( 1.3.6 ) proposition * _ in the above coordinates , the equator @xmath275 has the equation @xmath276 any other geodesic @xmath277 in @xmath261 is contained between two parallels @xmath278 and @xmath279 and the angle @xmath280 between two consecutive points of contact with these parallels is given by : @xmath281 the length of this arc of @xmath140 is given by @xmath282 _
[ figure3 ] * sketch of proof : * using the clairaut integral , the equations of the geodesic have the form : @xmath283 the formulae above follow by integration . ' '' '' the following is geometrically obvious : * ( 1.3.7 ) proposition * _ let @xmath284 denote an invariant torus with @xmath285 and let @xmath286 be the @xmath287-coordinate of the extremal parallel closest to @xmath16 in the projection of @xmath118 to @xmath0 . then : + ( i ) @xmath288 ; + ( ii ) @xmath289 ; + ( iii ) @xmath290 + ( iv ) @xmath291 _ * proof : * \(i ) by definition , @xmath292 is the change in angle along the equator between a geodesic @xmath205 on @xmath118 , starting at the equator on a fixed meridian and heading towards the northern hemisphere , and the fixed meridian , upon second intersection with the equator .
we can view this arc of the geodesic as consisting of three pieces : one from the equator to the northern extremal parallel , one on the ` back - side ' between the two extremal parallels , and one on the ` front - side ' from the southern parallel to the equator .
since the lengths of the ` front - side ' arcs are unchanged by rotation , we can rotate one until the two make up a smooth geodesic arc between the parallels .
the length of the geodesic arc is therefore twice that of an arc between the parallels , i.e. @xmath293 \(ii ) for the angle change : in the same way , the change in @xmath294 along this arc of @xmath205 is the change in @xmath294 of two arcs between the extremal parallels .
we subtract @xmath295 since @xmath121 measures the addition to one full revolution .
\(iii ) since @xmath296 and @xmath297 , we have @xmath298 \(iv ) from propositions 1.3.2 we have that @xmath299 since @xmath300 implies that @xmath301 we get from ( 1.3.1 ) that @xmath302 ' '' '' * ( 1.3.8 ) corollary * _ the non - degeneracy condition ( 1.1.6 ) is satisfied if @xmath37 is globally twisted , i.e. if @xmath303 or @xmath304 . _ * proof * : in these cases , @xmath133 is convex ( or concave ) . since the set @xmath128 is the graph of @xmath133 in @xmath108 , the gauss map ( and hence the frequency map )
is an embedding . ' '' '' * remark * both cases , of concavity and convexity , occur for ellipsoids of revolution , see [ bl ] .
the separating case of the round sphere is of course degenerate .
the classical birkhoff normal form of hamiltonian @xmath129 near a non - degenerate periodic orbit @xmath140 is a germ of a completely integrable system to which @xmath129 is symplectically equivalent in a ` formal neighborhood ' of @xmath140 ( see e.g. [ f.g ] for a detailed discussion ) . in the case at hand , where @xmath129 is already completely integrable ,
the birkhoff normal form is simply @xmath129 itself expressed in terms of action - angle variables .
the ` birkhoff invariants ' of the hamiltonian at a torus @xmath305 may then be identified with the germ of @xmath58 or of @xmath121 at @xmath306 since @xmath129 is homogeneous of degree 1 , it is equivalent and somewhat clearer to define the birkhoff invariants after first breaking the @xmath124 symmetry .
that is , we would like to introduce a ` base ' to the cone @xmath108 .
the most natural one may appear to be the energy level @xmath163 ; but for the purpose of calculating wave invariants at @xmath305 it is more convenient to use the tangent line @xmath307 at a point @xmath308 let us first consider the level set @xmath163 as the transversal .
the homogeneous function @xmath129 is obviously determined by the curve @xmath309 whose equation is given by ( 1.3.1 ) .
hence we can define the birkhoff invariants at a torus @xmath305 to be the taylor coefficients of the function @xmath310 at @xmath311 . by ( 1.3.7 ( iv ) ) it is equivalent to put : * ( 1.4.1 ) definition * _ the ( first ) birkhoff invariants of @xmath129 at an invariant torus @xmath189 with @xmath312 are the taylor coefficients of @xmath313 at @xmath195 in the coordinate @xmath113 . _
secondly , let us consider the tangent lines as transversals : we fix a point @xmath314 , let @xmath315 denote the common frequency vector of the ray of torii @xmath316 and put @xmath317 the equation of the tangent line to @xmath128 at @xmath318 in the action cone @xmath108 is then given by @xmath319 .
note that @xmath320 is homogeneous of degree 1 and hence equals @xmath115 along the ray @xmath321 ; consequently it is elliptic ( non - vanishing ) in a conic neighborhood @xmath322 of it .
the conic neighborhood will be parametrized in the following way : we fix a basis ( i.e. a non - zero vector ) @xmath323 of the line @xmath324 , and define the map @xmath325 for sufficiently small @xmath326 , this map sweeps out a conic neighborhood @xmath327 of @xmath318 with inverse given by @xmath328 since @xmath329 and since @xmath330 we may write @xmath331 where @xmath332 is the function on @xmath333 defined by @xmath334 the @xmath9 taylor expansion of @xmath335 around @xmath336 is then a symplectic invariant of @xmath129 . * ( 1.4.5 ) definition * _ the second ( tangential ) classical birkhoff invariants of @xmath129 associated to the periodic torus @xmath337 are the taylor coefficients @xmath338 _ in the real analytic case the taylor coefficients determine @xmath332 and hence @xmath129 by homogeneity .
it is more or less obvious that the first and second birkhoff invariants also carry the same information in the @xmath9 case . to be sure ,
we prove : * ( 1.4.6 ) proposition * _ the first birkhoff invariants canonically determine the tangential birkhoff invariants and vice versa .
_ * proof * : by definition @xmath339 therefore , @xmath340 or equivalently @xmath341 writing @xmath342 , this says @xmath343 hence the knowledge of taylor coefficients of @xmath133 is equivalent to knowledge of the taylor coefficients of @xmath332 . ' '' '' [ figure4 ] we may reformulate the non - degeneracy condition ( 0.1 ( iii ) ) in terms of @xmath332 where @xmath337 is the meridian torus : * ( 1.4.7 ) proposition * _ @xmath33 satisfies the non - degeneracy condition ( 0.1 ( iii ) ) as long as @xmath344 _ * proof * : by definition , @xmath345 . at the meridian torus , @xmath346 so @xmath347 and @xmath348 .
hence @xmath349 since @xmath350 it follows that @xmath351 hence , the condition @xmath352 is equivalent to the condition that @xmath353 the equivalence of this to ( 0.1 ( iii ) ) is proved in propositions ( 1.3.4 ) and again in corollary ( 1.3.8 ) . ' '' '' the birkhoff normal form of a hamiltonian @xmath129 at a closed orbit ( or family of closed orbits ) is a symplectic conjugacy invariant of @xmath129 in a neighborhood of the orbit(s ) .
hence the global birkhoff normal form @xmath58 of a completely integrable hamiltonian is a symplectic conjugacy invariant .
the purpose of this section is to show that it is a complete conjugacy invariant .
we begin by showing that the homogeneous hamiltonian torus actions commuting with geodesic flows of simple surfaces of revolution are all symplectically equivalent : * ( 1.5.1 ) proposition * _ suppose @xmath354 and @xmath355 are smooth surfaces of revolution of simple type and let @xmath110 resp .
@xmath356 denote their global action variables as above .
then there exists a homogeneous symplectic diffeomorphism @xmath357 such that @xmath358 .
_ * sketch of proof :* let @xmath359 , resp .
@xmath360 be the dual angle variables on @xmath354 resp . @xmath355 . except on the equators of @xmath361 a point in @xmath70
is uniquely specified by its action - angle coordinates .
define @xmath362 to be the identity map in action - angle coordinates with respect to the two metrics .
it is obvious that @xmath362 is a homogeneous symplectic diffeomorphism on the complement of the equators , so to prove the corollary it suffices to show that @xmath362 extends to the equators with this property .
since @xmath362 is homogeneous of degree 1 , it is necessary and sufficient to define it on the unit cotangent bundles .
moreover , since it commutes with the hamilton flow of @xmath73 on the regular set , its extension must also do so .
hence it must be the lift of a diffeomorphism @xmath363 on the orbit space @xmath364 of the rotation action .
this action is free so the natural projection @xmath365 must be diffeomorphic to the standard projection from @xmath366 the image of the torus foliation defined by level sets of @xmath73 is a singular foliation of @xmath367 formed by level sets of the function @xmath368 induced by @xmath73 on @xmath367 , and the two singular points @xmath369 are the images of the equators @xmath370 .
since @xmath368 is , by assumption , a perfect morse function on @xmath371 for each metric , the quotient map @xmath363 on the punctured quotient @xmath372 of @xmath363 extends smoothly to the completion .
it follows that @xmath362 extends smoothly as a rotationally equivariant map on the completion of @xmath373 and the homogeneous extension must be symplectic .
( see [ cv.1 ] for more on the behaviour near the poles ) . ' '' '' * remark * : in fact , this proposition can be sharpened to say : there exists only one homogeneous hamiltonian @xmath83 action on @xmath203 ( up to symplectic equivalence ) .
this follows from a homogeneous analogue of the delzant classification of completely integrable torus action on compact kahler manifolds .
we hope to report on more general results of this kind at a later time .
thus , the torus actions defined above on the cotangent bundles of simple surfaces of revolution are always symplectically equivalent .
the question arises when the geodesic flows are symplectically equivalent .
the answer can be given by expressing the norm functions @xmath45 of the metrics in terms of the global action variables . before doing so
, we note that the action variables are not quite uniquely defined above because the choice of generators @xmath374 is not unique .
for instance , one might have permuted the roles of @xmath16 and @xmath19 .
hence we have : * ( 1.5.2 ) proposition * _ let @xmath375 be a simple surfaces of revolution , and let @xmath376 resp .
@xmath356 be the global action variables defined above .
let @xmath377 resp .
@xmath378 be the expressions of the metric norms of @xmath379 in terms of action variables . then
: the geodesic flows of @xmath375 are homogeneously symplectically equivalent if and only if there a linear map @xmath380 such that @xmath381 _ * proof * if such a choice exists , then the map @xmath362 above obviously defines a symplectic conjugacy .
conversely , suppose the geodesic flows are symplectically conjugate by a homogeneous symplectic diffeomorphism @xmath362 , i.e. @xmath382 then @xmath383 are global action variables for the geodesic flow of @xmath384 . but
global action variables are almost unique : they correspond to a trivialization of the lattice bundle @xmath86 [ d][g.s ] .
therefore there exists @xmath385 so that @xmath386 then @xmath387 . ' '' '' the following gives a geometric interpretation of the conjugacy condition : * ( 1.5.3 ) proposition * _ suppose @xmath388
. then their geodesic flows are symplectically equivalent if and only if their equatorial first return times @xmath259 and angles @xmath313 are equal .
_ * proof * : suppose first that the flows are symplectically equivalent .
after choosing compatible bases for the homology we may then assume that the expressions for @xmath389 and @xmath390 in global action - angle variables are the same .
then the frequency maps are the same and hence by proposition ( 1.3.3 ) the equatorial return angles are the same .
also , the poincare maps @xmath37 are conjugate and hence the equatorial return times are equal .
conversely , if the equatorial return times and angles are the same , then the flows have the same frequency vectors ( as functions of the global action angle variables ) and hence have the same global birkhoff normal forms . by proposition ( 1.5.2 ) the flows are symplectically equivalent . ' '' ''
it is owing to the following notion that simple surfaces of revolution are so manageable . *
( 2.1 ) definition * _ the wave group @xmath391 of a compact , riemannian n - manifold @xmath392 is _ quantum torus integrable _ if there exists a unitary fourier - integral representation @xmath393 of the n - torus and a symbol @xmath394 such that @xmath395 _ in the formula above , we follow physics notation in indicating operators ( as opposed to their symbols ) with a ` hat ' . thus , the generators @xmath396 are first order pseudodifferential operators with the property that @xmath397 for some constant @xmath398 of modulus one .
since @xmath53 is a first order elliptic symbol on @xmath399 it has an asymptotic expansion in homogeneous functions of the form : @xmath400 the quantum action operators are uniquely defined up to a choice of basis of @xmath401 , the terms @xmath402 are uniquely determined up to the same ambiguity .
the principal symbols @xmath403 of the @xmath396 s generate a classical hamiltonian torus action , so any quantum torus action automatically induces a classical one .
conversely , it is a theorem of boutet de monvel - guillemin and weinstein that any classical hamiltonian torus action can be quantized [ bm.g , appendix , proposition 6.6 ] .
since metrics in @xmath32 give rise to quantum torus actions , we have , in particular : * ( 2.3 ) proposition ( cf .
[ cv.1 ] ) * _ suppose @xmath63
. then @xmath404 is quantum torus integrable .
_ we also observe that the following holds for any laplacian commuting with a quantum torus action : * ( 2.4 ) proposition * _ for any @xmath48 , @xmath405 _ * proof : * since the subprincipal symbol @xmath406 equals zero , the same is true @xmath407 now @xmath408 is invariantly defined ( independent of a choice of symplectic coordinates ) ; hence it may be expressed in action - angle coordinates in the form @xmath409 but the mixed derivative term automatically vanishes since @xmath389 is a function only of the classical action variables . ' '' '' let us specialize to the case of @xmath404 with @xmath410 from the fact that @xmath411 for a quantum torus action , it follows that the joint spectrum of the quantum moment map @xmath412 is the set of integral lattice points , translated by @xmath413 , in the closed convex conic subset @xmath414 .
the vector @xmath415 can be identified with the maslov indices of the homology basis of the invariant torii . in the case of @xmath404 , @xmath416 [ cv.1 ] .
[ figure5 ] expressing @xmath404 in the form @xmath417 we have that @xmath418 thus , the eigenvalues of @xmath404 have the form : @xmath419 and the wave trace takes the form @xmath420 the symbol @xmath53 may be regarded as a global quantum birkhoff normal form . as in the classical case ,
it has a germ at any periodic torus , so these may be regarded as the microlocal birkhoff canonical forms . to be more precise ,
we imitate the definition of the classical tangential birhoff normal forms and write @xmath421 where @xmath422 is a linear coordinate relative to a vector @xmath323 generating @xmath423 we then taylor expand @xmath424 around @xmath336 : @xmath425 * ( 2.7 ) definition * _ the homogeneous quantum birkhoff normal form coefficients of @xmath53 at the periodic torus @xmath337 are the taylor coefficients @xmath426 for @xmath427 _ * ( 2.8 ) proposition * _ assume that @xmath11 . then all of the functions @xmath402 are real analytic in the interior of the cone @xmath108 and all of the functions @xmath428 are real analytic near @xmath336 . _
* proof * : first consider @xmath429 we know that @xmath129 is a @xmath9 homogeneous function on @xmath108 . on the other hand , from the formula @xmath430
we see that @xmath97 is a real analytic function of @xmath431 since @xmath432 is the inverse function to @xmath129 with respect to the first variable , @xmath129 must also be a real analytic function of @xmath112 next recall that @xmath433 is defined ( up to a smoothing term ) as the function @xmath434 which has principal symbol @xmath435 and which has integral spectrum .
more precisely , one begins with @xmath436 , which has the property that @xmath437 with @xmath438 of order -1 .
one defines @xmath439 and puts @xmath440 since @xmath432 is a real analytic function and we only apply the holomorphic functional calculus in the steps of the construction , it follows that @xmath441 is an analytic function of @xmath442 .
since @xmath443 has the inverse function @xmath444 , it follows again by the inverse function theorem for analytic functions that @xmath53 is a real analytic function .
the real analyticity of the @xmath428 s follows from that of the @xmath402 s . ' '' ''
to relate the wave invariants to the coefficient of the normal form , it will be helpful ( as in [ g.1][z.1,2 ] ) to express the wave invariants as non - commuative residues of the wave operator and its time - derivatives .
let us recollect how this goes .
the non - commutative residue of a fourier integral operator is an extension of the well - known non - commutative residue of a pseudodifferential operator @xmath445 on a compact manifold @xmath139 , defined by @xmath446 where @xmath447 here , @xmath48 is a laplacian ( or any positive elliptic operator ) on @xmath139 . from work of seeley , wodzicki and guillemin
, one knows that @xmath448 is holomorphic in re @xmath449 ord@xmath450 and admits a meromorphic continuation to @xmath451 , with simple poles at @xmath452ord @xmath453 ( @xmath454 ) .
the residue at @xmath455 has a number of remarkable properties ( not shared by the residues of the other poles ) : =
res@xmath456 res@xmath457 , i.e. res is a trace on the algebra @xmath458 of pseudodifferential operators over @xmath139 ; + res@xmath450 is independent of the choice of @xmath459 ; + there is a local formula for the residue , res@xmath460 , where @xmath461 is the term of degree @xmath462 in the complete symbol expansion @xmath463 for @xmath445 ; @xmath464 is the canonical symplectic volume measure on @xmath465 .
these results may be extended to fourier integral operators as follows : let @xmath445 be a fourier integral operator in @xmath466 for some homogeneous canonical relation @xmath467 and @xmath468 .
below , diag@xmath469 denotes the diagonal in @xmath470 . below we will sketch a proof of the following : * ( 3.1 ) theorem * _
suppose @xmath471 and diag@xmath472 intersect cleanly .
then @xmath473 tr @xmath474 has a meromorphic continuation to @xmath451 , with simple poles at @xmath475 , where @xmath476 diag@xmath477 , and @xmath478 .
_ the clean intersection hypothesis above is that @xmath479 diag@xmath480 is a clean intersection .
it is satisfied in the case where @xmath481 if and only if the fixed point set of @xmath61 is clean .
hence , theorem ( 3.1 ) implies : * ( 3.2 ) corollary * _ let @xmath482 tr @xmath483 .
_ if the fixed point set of @xmath484 is clean , then @xmath485 has a meromorphic continuation to @xmath451 , with simple poles at @xmath486 ( @xmath487 ) . here , @xmath19 fix@xmath488 is the set of unit vectors in the fixed point set of @xmath61 . in the case of a completely integrable system , @xmath19 fix@xmath489 is the union of the periodic torii with period @xmath136 .
we assume here , and henceforth , that the periodic torii are all clean fixed point sets for @xmath61 on @xmath490 .
the non - commutative residue of the fourier integral operator @xmath445 is then defined by : * ( 3.3 ) definition * _ @xmath491 _ just as in the case of pseudodifferential operators . and just as in that case , res@xmath450 has some remarkable properties : - it is independent of the choice of @xmath48 ; - if either @xmath445 or @xmath81 is associated to a local canonical graph , then res@xmath456 res@xmath457 - there is a local formula for res@xmath450 .
the basic properties of res@xmath450 may be deduced from a singularity analysis of the closely related distribution trace @xmath492 tr@xmath493 ( cf .
[ z.4 ] ) . under the cleanliness assumption above
, @xmath494 is a lagrangean distribution on @xmath495 with singularities at the set of ` sojourn times ' , @xmath496 for a given sojourn time @xmath497 , the corresponding set @xmath498 of sojourn rays fills out a submanifold of @xmath490 of some dimension @xmath499 . with @xmath500
, we have * ( 3.4 ) proposition [ z.4 , proposition 1.10 ] . * _ if @xmath501 with @xmath502 , supp @xmath503 and @xmath504 near 0 , then @xmath505 where @xmath506 is a universal constant .
the coefficients have the form , @xmath507 where @xmath508 diag@xmath509 , @xmath510 , @xmath511 is a canonical density on @xmath512 and the functions @xmath513 are determined by the @xmath514-jet of the ( local ) complete symbols of @xmath515 along @xmath512 . _
_ proof of theorem ( 3.1 ) .
_ as in the pseudodifferential case [ dg , proposition 2.1 ] , we have @xmath516 hence , @xmath517 & = & c_n\displaystyle{\sum^\infty_{j=0 } \frac{\alpha_j}{m+1+\frac{e_0 - 1}{2}-(s+j ) } + \mbox { entire}}\;,\end{array}\ ] ] completing the proof . ' '' '' in particular , if @xmath518 , then @xmath519 is a sojourn time if and only if @xmath520 .
if @xmath521 lsp@xmath392 , @xmath522 is regular at 0 .
now let us return to the case where the geodesic flow is completely integrable .
then the dimension of each periodic torus @xmath39 of period @xmath136 equals @xmath523 .
hence we have : @xmath524 where the sum runs over the periodic tori in @xmath490 and where @xmath525 more precisely , it takes this form if @xmath526 is even ; if @xmath526 is odd , the positive powers of @xmath527 should be multiplied by @xmath528 . in the following corollarly we use the notation @xmath41 for the time reverse torus @xmath529 * ( 3.6 ) corollary * _ @xmath530 _ _ proof . _ in the notation of ( possibly negative ) fractional derivatives in @xmath211 , we have @xmath531 .
the claim follows from the facts that @xmath532 and that the non - commutative residue is the coefficient of @xmath528 . ' '' '' * ( 3.7 ) examples * : \(a ) if @xmath533 , then the wave trace expansion at a torus @xmath39 has the form @xmath534 hence @xmath535 \(b ) if @xmath536 , the wave trace expansion at @xmath39 has the form @xmath537 hence @xmath538 * ( 3.8 ) remark * : the wave invariants for a closed geodesic @xmath140 ( or periodic torus @xmath539 ) are exactly the same as for its time reversal , hence the same residue formulae also give the individual wave invariants .
for this reason it is often not necessary to resolve the ambiguity between a torus and its time reversal .
we now prove that the wave trace invariants of a metric @xmath11 determine its quantum normal form @xmath53 .
we begin by specializing corollary ( 3.6 ) to the case of a surface of revolution @xmath33 in @xmath540 since the length spectrum @xmath150 is simple , there is a unique periodic torus @xmath541 of each length @xmath542 ( up to time reversal ) . by the existence of global action - angle coordinates , it may be expressed in the form @xmath543 for a unique point @xmath544 ( up to reflection ) .
let @xmath545 denote the frequency vector at @xmath546 .
then we have @xmath547 where @xmath548 is the vector of winding numbers of the periodic orbits on @xmath549 the other periodic torii of period @xmath136 lie on the rays @xmath550 .
their action coordinates lie on the rays @xmath551 and they have the same frequency vector , @xmath545 , as for @xmath552 .
since the wave invariants at @xmath553 depend only on the microlocalization of the wave group to a conic neighborhood of @xmath554 , we introduce a microlocal cut off operator @xmath555 with @xmath556 homogeneous of degree 0 , equal to 1 in a small conic neighborhood of the ray @xmath557 and zero off of a slightly larger conic neighborhood .
the singularity of @xmath558 at @xmath559 is then the same as the singularity of @xmath560 to calculate the wave trace as a residue , we also introduce the first order pseudodifferential operator @xmath561 .
we emphasize that @xmath562 is a linear combination with constant coefficients of the action operators .
since @xmath563 , the principal symbol @xmath564 of @xmath562 takes the value 1 at @xmath565 and therefore @xmath566 is elliptic in a conic neighborhood of @xmath553 .
we will use it as the gauging elliptic operator in the residue formula for the wave invariants . *
( 4.1.1 ) proposition * _ let @xmath11 and let @xmath151 .
then we have : @xmath567 where as above @xmath548 is the vector of winding numbers of @xmath553 .
_ * proof * : since @xmath444 and since @xmath568 is a function of the action operators , we have by corollary ( 3.6 ) that @xmath569 we then apply the poisson summation formula for re @xmath570 to replace the sum over @xmath571 by @xmath572 where @xmath573 it follows from theorem 3.1 and by simplicity of the length spectrum that only the term with @xmath574 has a pole at @xmath575 this can be seen more directly from the fact that only in this term does the phase @xmath576 have a critical point , since @xmath577 implies that the torus with actions @xmath578 is periodic of period @xmath136 ( cf .
1.2 ) . ' '' '' to calculate the residue of the integral @xmath579 , we rewrite the integrals in terms of the @xmath580 coordinates introduced in ( 1.4.3 ) in a conic neighborhood of the point @xmath581 : recall that we parametrize the tangent line @xmath582 by @xmath583 where @xmath323 is a non - zero vector along the line @xmath584 and parametrize a conic neighborhood of @xmath557 by @xmath585 since @xmath552 is fixed , we abbreviate @xmath586 by @xmath7 in the next proposition .
we also denote the cutoff in these coordinates by @xmath587 * ( 4.1.2 ) proposition * _ with the above notation , @xmath588 equals the term of order @xmath589 in the asymptotic expansion of the integral @xmath590^ { - \frac{1}{2 } + k } d\xi.\ ] ] here , @xmath591 , and @xmath592 is the third order remainder in the taylor expansion of @xmath46 at @xmath593 _ * proof * we first change variables @xmath594 in the expression for @xmath595 in the preceding lemma and then further change variables to @xmath596 .
thus we get @xmath597 } [ \rho h(\xi ) + \rho^{-1 } h_{-1}(\xi ) + \dots]^{-{\frac{1}{2}}+ k } \rho^{-s+1 } d\rho d\xi.\ ] ] taylor expanding @xmath598 and using that @xmath599 , that @xmath600 and that @xmath601 by ( 1.2.1a - b ) , we get @xmath602 } [ \rho h(\xi ) + \rho^{-1 } h_{-1}(\xi ) + \dots]^{-{\frac{1}{2}}+ k } \rho^{-s+1 } d\rho d\xi.\ ] ] now change variables again , @xmath603 and pull the factor @xmath604 in front of @xmath46 in the bracketed expression outside the @xmath605 -integral .
we get : @xmath606}\ ] ] @xmath607^{-{\frac{1}{2}}+ k } d\xi\ } \rho^{-s+1 + k - { \frac{1}{2 } } } d\rho .\ ] ] the pole at @xmath455 is produced by the terms of order @xmath608 in the @xmath609-integrals , hence by the terms of order @xmath610 in the asymptotic expansion of the @xmath611-integral . ' '' '' to determine the terms of order @xmath612 in @xmath613 we apply the method of stationary phase .
cancelling factors of @xmath614 we get : * ( 4.1.3 ) corollary * _ in the above notation , @xmath588 equals the term of order @xmath615 in the asymptotic expansion of @xmath616^ { - \frac{1}{2 } + k}\}|_{\xi = 0}\ ] ] _ we now want to analyse the terms of the expansion coming from the ray of meridian torii with @xmath617 the special property of the meridian torii is that they are invariant under the canonical involution @xmath618 of @xmath156 . as the following proposition shows
, this leads to a useful symmetry property of the quantum normal form .
as above , the notation @xmath619 etc .
refer to the functions @xmath586 ( etc . ) where @xmath552 is the point on @xmath163 corresponding to the meridian torus in @xmath96 . *
( 4.2.1 ) proposition * _ the complete symbol of @xmath53 is invariant under @xmath43 . hence @xmath428 is even for all j. _ * proof *
let @xmath620 denote the operator of complex conjugation : @xmath621
since @xmath622 commutes with @xmath620 , so does @xmath623 since @xmath620 is a conjugate - linear involution , this implies that @xmath624 where the bar denotes complex conjugation .
we claim that @xmath625 and that @xmath626 .
moreover , that @xmath627 the statements regarding @xmath628 are obvious since @xmath629 changes sign under complex conjugation and since @xmath630 this latter also follows from the fact @xmath631 .
regarding @xmath97 we note that its @xmath43-invariance follows immediately from the explicit formula ( 1.3.1 ) .
the non - obvious claim is that @xmath433 is invariant under @xmath620 .
but from the invariance of the principal symbol we have @xmath632 where @xmath633 is of order 0 .
since @xmath433 is a function of @xmath634 , it is clear that @xmath620 must take joint @xmath635 eigenfunctions into joint eigenfunctions .
hence @xmath620 determines an involution on the joint spectrum @xmath636 which , we recall , is simple .
consequently , the involution ( still denoted @xmath620 ) must take the form @xmath637 where @xmath65 is a bounded function .
moreover , since @xmath638 and since the @xmath639 are quasi - modes associated to bohr - sommerfeld - maslov torii [ cv.3 ] , it is clear that @xmath640 must be a polyhomogeneous function of order 0 on @xmath108 .
since it is also integral - valued on the semi - lattice of joint spectral points , it must be constant .
additionally it must satisfy the involution condition , and one sees that the constant must be 0 .
returning to @xmath59 we have @xmath641 so that @xmath642 we then observe that @xmath53 is a real function ( at least modulo terms of order @xmath643 ) to see this , we recall that the eigenvalues @xmath644 are real and argue by induction on the symbol expansion using that @xmath645 first , the principal symbol @xmath129 is real so we may drop it from the expansion without affecting reality . assuming that @xmath646 are real , we may drop them all and get that the tail sum is real . since it is dominated by @xmath647 , this function
must be real for all points @xmath648 sufficiently far from @xmath649 .
but by homogeneity , @xmath647 is then real on the projection of these lattice points to @xmath650 the projected points form a dense set by proposition 1.2.4 and hence @xmath647 is everywhere real .
we conclude that @xmath53 is @xmath43-invariant .
now consider the ray of meridian torii , or more precisely the ray @xmath651 in the action cone @xmath652 we note that this ray is invariant under the involution @xmath653 of @xmath654 , and moreover the meridian torus is invariant under @xmath43 ( it ` rotates ' the torus by angle @xmath655 ) from the above , the level set @xmath163 is invariant under @xmath656 and hence the tangent line at @xmath657 is invariant .
evidently it is horizontal in the @xmath110-plane and @xmath656 restricts to it to the map @xmath658 since the complete symbol of @xmath53 is @xmath656-invariant , the @xmath428 s must be even . ' '' '' we now go back to the wave invariants associated to the meridian torus and its iterates .
let us write the amplitude for the kth wave invariant of the iterate of length @xmath136 , namely @xmath659^ { - \frac{1}{2 } + k},\leqno(4.2.2)\ ] ] in the form @xmath660 thus , @xmath661 is the term of order @xmath612 in @xmath662.\leqno(4.2.3)\ ] ] expanding the derivatives and using that @xmath663 is even and of order @xmath664 , we may rewrite ( 4.2.3 ) in the form @xmath665 here , the multi - index @xmath666-sum runs over @xmath667-tuplets with @xmath668 with @xmath669 in order that @xmath670 it is necessary that @xmath671 these restrictions follow from the fact that it takes at least 4 derivatives of @xmath592 to make a non - zero contribution . for future reference
we assemble some notation and identities regarding the coefficients @xmath672 and their relations to the wave invariants . *
( 4.3.1 ) notation * l ( 1 ) define : @xmath673 by @xmath674^ { - \frac{1}{2 } } = $ ] + @xmath675 $ ] + + ( 2 ) define : @xmath676 by @xmath677 + + ( 3 ) define @xmath678 by @xmath679 $ ] + + ( 4 ) define @xmath680 by @xmath681 $ ] + + * ( 4.3.2 ) identities * l \(1 ) @xmath682 = @xmath683 $ ] + + ( 2 ) @xmath684 = @xmath685^ { k } a_{l0}.$ ] + + ( 3 ) @xmath686 + + @xmath687.$ ] + ( 4 ) @xmath688 + the coefficients @xmath689 are universal and hence up to the prefactor of @xmath690 , @xmath691 is a ( non - homogeneous ) polynomial of degree @xmath514 in @xmath136 , in @xmath692 and in the @xmath693 s with universal coefficients . in particular , the first few @xmath691 s are given by : @xmath694.\ ] ] it follows that the principal wave invariant is given by @xmath695 and that the higher wave invariants @xmath661 are given by @xmath696 times polynomials in @xmath136 and in the derivatives of @xmath697 at @xmath336 .
for instance , the subprincipal wave invariant in dimension 2 is given given in terms of universal coefficients @xmath698 by : @xmath699\ ] ] which is easily seen to equal @xmath700.\ ] ] above the indices in the coefficients @xmath701 have the following meaning : @xmath702 corresponds of course to the @xmath702 index ; @xmath703 is the power of @xmath136 and @xmath704 are the jet orders of @xmath705 , with with exception of @xmath706 which is the jet order of @xmath7 . *
( 4.3.3 ) proposition * _ we have _
@xmath707 where @xmath708 is a polynomial with the following properties : \(i ) it involves only the first @xmath709 taylor coefficients of @xmath7 at @xmath649 , the first @xmath710 of @xmath711 , ... , the first @xmath712 of @xmath705 ... , the first @xmath713 of @xmath714 and the 0th of @xmath715 \(ii ) it is of degree 1 in the variables @xmath716 , @xmath717 , and each occurs in precisely one term .
\(iii ) the @xmath136-order of the monomials containing these terms is respectively @xmath718 * proof * : to prove these claims we combine the formula for @xmath719 in ( 4.2.4 ) with the formulae for the @xmath720 given in ( 4.3.2 ) . since we may replace the factors of @xmath592 by @xmath7 in ( 4.2.4 ) and since ( 4.3.2 ) expresses @xmath672 as @xmath690 times a polynomial in @xmath136 and in the @xmath693 s , it is clear that @xmath719 is given by a polynomial in the data stated above .
it remains to prove that the polynomial has the properties claimed in ( i)(iii ) .
we prove these claims by proving the stronger statement that @xmath721 where @xmath722 has the properties ( i)-(iii ) for variable @xmath723 the proof is by induction on @xmath702 .
the properties are visibly true for the principal and subprincipal wave invariants .
assume then that they are correct for @xmath724 and consider how things change as @xmath725 .
first , the amplitude @xmath726 is given by @xmath727a_{l n}$ ] .
second , we are looking at the term of order @xmath728 rather than @xmath729 in the asymptotic series ( 4.2.4 ) . in going one further order into the asymptotic series
, only two new things happen : @xmath730 the term @xmath731 appears for the first time , arising from the linear term in the taylor expansion of the exponential in ( 4.2.2 ) .
the linear term in the binomial expansion of the power in ( 4.2.2 ) does not contribute at this stage because its @xmath608 factor has one higher power .
@xmath730 two additional derivatives are allowed to fall on the previous @xmath428 s .
a priori , there could be between two and six additional derivatives in a method of stationary phase expansion .
however , the cases of three to six derivatives do not contribute any new data . indeed
, the cases of three - four derivatives occur when @xmath703 goes up by one .
one has to remove the extra factor of @xmath608 by applying at least one derivative to the @xmath732 factor .
but in fact all four have to be applied to @xmath592 to make a non - zero contribution , and hence no derivatives are left to apply to the @xmath733 s . in the case of five or six derivatives , where @xmath703 goes up to two
, one needs to remove two extra factors of @xmath608 by applying derivatives to the @xmath734 factor , and there is no non - zero way to do this . claim ( i ) follows immediately from these observations . to prove ( ii ) we note that by the induction hypothesis , @xmath735 occur linearly in @xmath736 since it requires both of the two new derivatives to fall on these factors to produce @xmath737 etc . in @xmath738 , these factors will also occur to order 1 . as for @xmath731 , we observed above that it comes only from the linear term in the exponential in ( 4.2.2 ) and hence it too appears to order 1 .
the proof of ( iii ) follows the proof of ( ii ) .
since the terms in the exponent of the exponential factor in ( 4.2.2 ) have the form @xmath739 , the new @xmath740 term has the coefficient @xmath741 .
similarly , the other terms under discussion , e.g. @xmath742 , originated as @xmath743 at the @xmath744th stage with the factor of @xmath745 .
we would like to show that they remain with just this factor of @xmath136 as we move inductively up to the @xmath746st stage .
to this end , we note that in order to produce the term @xmath747 at the ( n+1)st stage , @xmath748 derivatives must fall on the original @xmath743 .
but there are exactly @xmath24 stages intermediate between the @xmath749th stage and the @xmath746st stage and at each stage at most two new derivatives can fall on a factor .
hence , each of the two new derivatives at each stage must fall on the factor of concern .
let us also consider what can be multiplied against the original @xmath743 in the course of producing @xmath742 .
we observe that each new pair of derivatives is accompanied by a factor of @xmath750 since @xmath751 , at the @xmath752th stage , comes with the factor @xmath753 , the @xmath748 further derivatives will bring its @xmath608-order up to @xmath754 hence no other factors of @xmath608 could have fallen on this factor .
therefore in the repeated multiplications by @xmath727 $ ] , only the repeated choice of the term @xmath7 , of order 0 in @xmath608 , can have given rise to the term @xmath755 .
it follows that it retains its original @xmath136-order , namely @xmath756 ' '' '' our purpose is now to show that by using the joint @xmath757 asymptotics one can recover the complete taylor expansions of all the @xmath428 s from the wave invariants of the meridian torus and its iterates .
that is , to complete the * proof of the main lemma * : we will prove by induction that from the wave invariants @xmath758 with @xmath724 and all @xmath759 , we can determine the @xmath760 -jet of @xmath7 , the @xmath710-jet of @xmath711 , ... , the @xmath712-jet of @xmath705 at @xmath593 ( for @xmath761 ) we note that , unlike in the non - degenerate case , the principal wave invariant determines the 2-jet of @xmath7 , and gives no new information under iteration .
we therefore begin the induction with the subprincipal wave invariant @xmath762 . from the explicit formula above for the subprincipal wave invariant
@xmath763 it is evident that the 2-jet term in @xmath7 is old information , while the other two terms differ in the power of @xmath136 .
hence they decouple under iteration @xmath764 and we can determine the 4-jet of @xmath7 and the 0-jet of @xmath711 from the first two wave invariants .
hence the induction hypothesis is correct at the first stage .
assume the induction hypothesis for the @xmath765st stage .
then the only new information at the kth stage is that contained in the terms @xmath716 , @xmath717 . by the proposition above
, they occur linearly in the monomials containing them and the monomials have also the factor of @xmath136 to the powers @xmath766 hence the terms containing the new data decouple as @xmath767 , and the new data can be determined as stated .
this completes the inductive argument .
it follows that we can determine the full taylor expansions of the @xmath768 s at @xmath593 since they are real analytic they are completely determined . then from the homogeneity of @xmath402
, we can determine @xmath402 from @xmath428 and hence the entire function @xmath53 is determined . ' '' '' * remark * in the above argument , we are able to drop the terms involving only previously known data because of the universal nature of the polynomials @xmath708 .
this universality depends on the fact that we are only comparing wave data for quantum torus integrable laplacians .
we now show that the symplectic equivalence class of a metric in @xmath1 is spectrally determined among metrics in this class . * ( 5.1 ) proof of corollary 1 * by the main lemma , the principal symbol @xmath769 as a function of the action variables is spectrally determined for a metric @xmath6 in @xmath540 by propositions ( 1.5.2 - 3 ) , @xmath770 determines the geodesic flow up to symplectic equivalence among metrics in @xmath1 . ' '' '' * remark * the corollary could also be proved by noting that the first return time @xmath259 is spectrally determined .
but geodesic flows of in the class @xmath1 are symplectically equivalent if and only if they have the same first return times @xmath259 .
( see also [ c.k ] for an equivalent statement . )
also , it should be noted that the first return time could be determined from the wave invariants at the equator ; hence corollary 1 would also follow from guillemin s inverse result for non - degenerate elliptic closed geodesics . * ( 5.2 ) proof of corollary 2 * : since @xmath224 is spectrally determined , so is the function @xmath771 of proposition 1.3.6 .
it is given by @xmath772 or , putting @xmath773 for @xmath774 $ ] and @xmath775 for @xmath776 $ ] , by @xmath777 ( cos(i(i))^2 - x^2)_+^{-{\frac{1}{2 } } } dx.\ ] ] it therefore suffices to show that @xmath778 ( u^2 - x^2 ) _
+ ^{-{\frac{1}{2 } } } dx\ ] ] determines @xmath779 $ ] .
but @xmath780 $ ] is smooth and even so may be written as @xmath781 for a smooth @xmath6 ; changing variables @xmath782 , and @xmath783 we get @xmath784 thus @xmath497 is a standard abel transform , which is well - known to be invertible .
it follows that @xmath785 is spectrally determined ; hence so is the even part of @xmath65 . ' '' ''
to complete the proof of the theorem , we need to show that @xmath53 determines a metric in @xmath786 it is plausible that this can be done for the following reason : the spectrum of @xmath48 is the set of values @xmath787 of @xmath788 on the integer points of the action cone @xmath652 on the other hand since @xmath789 , the set of values @xmath790 for fixed @xmath526 , is just the spectrum @xmath791 of the the singular sturm - liouville operator @xmath792 obtained by separating variables in @xmath48 , fixing the eigenvalue of @xmath793 equal to @xmath526 , and putting the radial part in normal form . hence ,
from the coefficients of @xmath53 we can determine @xmath794 for each @xmath526 .
that is , from @xmath795 we have determined the _ joint _ spectrum of @xmath796 it would thus remain to show that the metric can be determined from this joint spectrum , a much more elementary inverse result which has been stated several times in the literature ( [ kac],[b ] [ gur ] ) . since these prior discussions seem to us somewhat sketchy and incomplete , we give a self - contained proof below which was found before we were aware of these references .
* proof of final lemma * the proof is basically to write down explicit expressions for @xmath129 and @xmath797 in terms of the metric ( i.e. in terms of @xmath273 ) and then to invert the expressions to determine @xmath273 .
the first step is therefore to construct an initial part of the quantum normal form explicitly from the metric . up to now , we only know that a polyhomogeneous normal form exists . to be sure , the principal term @xmath129 has already been implicitly expressed in terms of the metric : knowledge of @xmath129 is equivalent to knowledge of the level set @xmath798 and hence to knowledge of the function @xmath799 in ( 1.3.1 ) .
unfortunately , we have seen in corollary ( 5.2 ) that @xmath129 only determines the ` even part ' of the metric .
hence we need to determine at least one of the subprincipal terms @xmath800 .
it turns out that only @xmath797 is needed in addition to @xmath129 to determine @xmath6 .
since the calculation of @xmath797 requires a new calculation of @xmath129 , we start calculating both from scratch .
to determine @xmath129 and @xmath797 in terms of @xmath6 , we are going to study the spectral asymptotics of @xmath801 along ` rays of representations ' of the quantum torus action , i.e. along multiples of a given lattice point @xmath802 such rays are the quantum analogue of rays @xmath803 thru invariant tori and are basic to homogeneous quantization theory [ g.s.3 ] .
the basic idea is that the lattice points @xmath804 parametrize tori @xmath805 satisfying the bohr - sommerfeld quantization condition . to each such quantizable torus
one can construct a joint eigenfunction @xmath806 of @xmath635 by the wkb method .
the @xmath807 are eigenfunctions of @xmath48 with complete asymptotic expansions along rays . by studying the eigenvalue problem as @xmath808
we can determine the @xmath800 s .
the wkb method we employ is closely related to the classical wkb method for constructing quasi - modes ( cf .
[ cv.3 ] and the appendix ) , except that we have an internal rather than an external planck constant .
let us recall the relevant terminology and notation from the latter case .
for each torus @xmath118 , we denote by @xmath809 the space of oscillatory integrals ( semi - classical lagrangean distributions ) associated to @xmath118 , with semi- classical parameters @xmath810 in a set @xmath811 to be specified by the quantization condition .
such oscillatory integrals have the form @xmath812 where the projection of @xmath118 is covered by open sets @xmath813 , where @xmath814 and where the phase @xmath815 with @xmath816 , parametrizes a part of @xmath118 .
the amplitude is a classical symbol in @xmath702 of order @xmath413 . for further details
, we refer to [ c.v.3 , 8 ] . as discussed above , the torii @xmath118 project @xmath817 to the annuli @xmath818 in @xmath0 and have fold singularities along the extremal parallels . away from the parallels ,
an associated quasi - mode is given by a sum of two simple wkb functions @xmath819 since the actual @xmath48- eigenfunctions @xmath806 are quasi - modes attached to the bohr - sommerfeld torii @xmath820 , they have such a form modulo @xmath821 and , since @xmath806 is an exact @xmath822-eigenfunction , its phases must take the form @xmath823 with @xmath294-independent amplitudes in polar coordinates .
it follows that @xmath824 has the form @xmath825 , where @xmath826 is an oscillatory integral in the @xmath24-variable .
it is of course associated to the pushed - forward lagrangean @xmath827 where @xmath828 $ ] is the projection induced from the map @xmath829 in the case of the meridian torii @xmath830 , the pushforward is just a horizontal line @xmath831 in @xmath832 $ ] . in the other cases , the @xmath833 is a closed curve projecting to an interval @xmath834 $ ] with fold singularities at the turning points ( endpoints ) .
the curve is given by an equation of the form @xmath835 where @xmath836 is the radial hamiltonian and where @xmath837 was the level of the torus .
the radial part @xmath838 of @xmath806 , is an eigenfunction of the radial operator @xmath839 arising from separating variables in @xmath48 . before proceeding ,
we simplify by conjugating @xmath840 to the 1/2-density radial laplacian @xmath841 where @xmath842 a(r)^ { - { \frac{1}{2}}}.$ ] ( note that the volume form @xmath843 of @xmath33 projects to @xmath844 ) thus , we view the radial eigenfunction as having the form @xmath845 and apply the wkb method to the eigenvalue problem @xmath846 although the coefficients become singular at @xmath847 , the standard wkb theory applies in the interior because it applies to quasi - modes of @xmath48 on @xmath0 .
away from the turning points , the radial part of the 1/2-density eigenfunction therefore has the form @xmath848\leqno(6.3)\ ] ] where the phase @xmath849 is homogeneous of degree 1 and where the amplitude @xmath850 is homogeneous of degree @xmath851 in @xmath852 these homogeneities replace the powers of @xmath702 in the non - homogeneous theory described above .
the bohr - sommerfeld quantization condition on @xmath118 is that : @xmath853 where @xmath854 is the maslov index ( cf .
[ cv.3 , 4 ] ) .
it is satisfied by @xmath805 and hence by the radial lagrangean @xmath855 .
the local wkb ansatz ( 6.3 ) therefore extends to a quasi - mode of infinite order associated to the global @xmath856 .
our purpose now is to write down and solve the first two transport equations , which are needed to determine @xmath129 and @xmath797 . for background on the relevant aspects of the wkb method
, we refer to the appendix .
the transport equations are obtained by separating out terms of like order in the asymptotic eigenvalue problem @xmath857 \sim ( h(n_o , k_o + { \frac{1}{2 } } ) + h_{-1 } ( n_o , k_o + { \frac{1}{2 } } ) + \dots ) ^2 \sum_{\pm } [ e^ { \pm s_{n_o k_o}(r ) } \sum_{m=0}^{\infty } \alpha_{n_o , k_o ; m } ( r ) ] .\ ] ] the leading term , of order 2 , is the eikonal equation @xmath858 whose solution is the first order phase function @xmath859 the bohr - sommerfeld quantization condition on @xmath860 thus reads : @xmath861 the first transport equation , @xmath862 , is solved by @xmath863^{-\frac { 1}{4 } } \leqno ( 6.5)\ ] ] away from the turning points . here , @xmath864 .
the solution is of course determined only up to a constant , and we have normalized it so that @xmath865 is homogeneous of order 1 . in the usual way ( see the appendix ) , we will interpret @xmath866 as the coefficient of the @xmath867- invariant 1/2-density @xmath868^{-\frac { 1}{4 } } \sqrt{dr}\ ] ] on @xmath869 where @xmath867 is the hamilton vector field of @xmath870 .
we then re - write the higher coefficients @xmath871 in the form @xmath872 the second transport equation ( of order zero ) then has the form : @xmath873 where @xmath24 denotes the local coordinate on @xmath833 obtained by pulling back the base coordinate under the projection .
the integral of the left hand side over the closed curve @xmath874 with respect to the @xmath867-invariant density @xmath875 must equal zero .
this gives a formula for the first correction to the bohr - sommerfeld eigenvalue @xmath876 : @xmath877 dr .
\leqno(6.7)\ ] ] here , @xmath878 denotes the period of the @xmath867 flow on @xmath833 and @xmath879 are the turning points . plugging in ( 6.5 ) we get , formally , the expression @xmath880
@xmath881 actually , the integral in ( 6.8 ) diverges at the turning points , and the correct formula is the regularization obtained ( for instance ) by the method of the maslov canonical operator . for the sake of completeness
, we provide an exposition of this method in the appendix . roughly speaking ,
it regularizes ( 6.8 ) by formally integrating the @xmath882 derivatives by parts and by moving them outside the integral ( in the appropriate way ) as derivatives in the energy level @xmath837 .
a crucial consequence of the regularization is that only first derivatives of @xmath25 appear in the formulae for @xmath883 thus the first term of ( 6.8 ) is regularized by @xmath884 here and below @xmath885 denote ( non - zero ) constants which we will not need to determine . for the second term of ( 6.8 ) , we note that @xmath886 after some simplification , this regularizes to : @xmath887 @xmath888 now let us return to the inverse problem .
we begin from the fact that @xmath889 is a known function of the variable @xmath890 for each @xmath891 .
its principal symbol @xmath892 is then a known function and hence the inverse function @xmath893 satisfying @xmath894 is a known of @xmath837 .
this of course presupposes that @xmath895 , which follows from the non - degeneracy assumption ( 1.1.6 ) .
we may write the integral in the form @xmath896 where @xmath413 is the distribution function of @xmath897 , i.e. @xmath898 with @xmath899 the lebesgue measure .
the above integral is an abel transform and as mentioned above it is invertible .
hence @xmath900 is a spectral invariant .
some simplification leads to the conclusion that the function @xmath901 is known from the spectrum . by the simplicity assumption on @xmath25
, there are just two solutions of @xmath902 ; the smaller will be written @xmath903 and the larger , @xmath904 thus , the function @xmath905 is a spectral invariant . now let us turn to the @xmath797 expression .
since @xmath58 is a spectral invariant , the factors @xmath906 and @xmath907 are spectral invariants .
hence we may remove them from the expression for @xmath797 and still get a spectral invariant .
for various universal constants @xmath908 it takes the form @xmath909
@xmath910 |_{e = \frac{h(i_1,i_2)}{i_1 ^ 2}}.\ ] ] by a change of variables , we may rewrite ( 6.10 ) in the form @xmath911 @xmath912|_{e = \frac{h(i_1,i_2)}{i_1 ^ 2}}\ ] ] where @xmath913 all values of @xmath837 which occur as ratios @xmath914 give spectral invariants , so ( 1.10 ) ( as a function of the variable @xmath837 ) is a spectral invariant .
we now claim that @xmath438 itself is a spectral invariant . to determine it from ( 6.11 ) we rewrite ( 6.11 ) in terms of the fractional integral operators ( cf .
[ g.sh , ch.1 5.5 ] ) @xmath915 on the half - line @xmath916.$ ] these operators satisfy @xmath917 thus ( 6.11 ) equals @xmath918 where @xmath150 is the fractional integral operator @xmath919 to solve for @xmath438 we apply @xmath920 to @xmath921 to get @xmath922 this equation determines @xmath438 up to a solution @xmath65 of the associated homogeneous equation , essentially an euler equation . but also @xmath923 on @xmath924 $ ] since no homogeneous solution can have this property , @xmath438 is uniquely determined by this boundary condition .
it follows that both ( 6.9 ) and ( 6.12 ) are spectral invariants . but from @xmath925 and @xmath926 one can determine the pair @xmath927 .
hence @xmath928 is determined from the spectrum .
since @xmath929 this determines @xmath25 and hence the surface . ' '' ''
the purpose of this appendix is to give an algorithm for calculating the higher order terms in the quasi - classical approximation of eigenvalues for 1 d schrodinger operators @xmath930 with confining potentials . in particular , we carry out the calculation of the @xmath931 term , which was used in the proof of the final lemma . the algorithm is based on the maslov method of canonical operators .
expositions and refinements of this method can be found , among other places , in maslov s book [ m ] , in the article of colin de verdiere [ cv.3 ] ( and in its references ) , and in the recent book of bates- weinstein [ b.w ] .
although these references explain the construction of the canonical operator and prove the existence of complete quasi - classical eigenvalue expansions , they do not generally go on to describe the calculation of the terms .
an exception is the original book of maslov [ m ] , which does calculate the first two or three terms ; but the method of canonical operators is abandoned at this point in favor of some methods of special functions . as we will show , the canonical operator method gives the required corrections quite efficiently .
the unknown function @xmath933 is an oscillatory associated to a lagrangean of the form @xmath934 where @xmath935 is determined by the bohr - sommerfeld - maslov quantization condition : @xmath936 here , @xmath413 is the maslov index of @xmath937 ; it equals @xmath713 for connected level sets of hamiltonians of the form @xmath938 to find the higher order corrections @xmath939 we consider the maslov canonical operator @xmath940 we follow here the notation and terminology of [ b.w][cv.3 ] : @xmath941 is the space of symbolic sections of the bundle of 1/2-densities times maslov factors and @xmath942 is the space of oscillatory integrals associated to @xmath937 . there is a natural symbol map in the reverse direction ; any inverse modulo @xmath943 is a quantization or canonical operator .
its existence is equivalent to the condition that @xmath937 satisfy the bsm quantization condition .
for background we again refer to [ cv.3][b.w ] .
we also denote by @xmath944 the hamilton vector field of @xmath129 , by @xmath945 the lie derivative on any bundle , by @xmath946 a nowhere vanishing section of @xmath947 , and by @xmath948 a @xmath945-invariant density on @xmath937 ( for a fixed @xmath526 ) . by surjectivity of @xmath949 ,
we can write an oscillatory integral associated to @xmath937 in the form @xmath950.\ ] ] our aim is to determine the quasi - classical series @xmath951 and coefficient functions @xmath952 for which the asymptotic eigenvalue problem is solvable .
we begin by constructing local solutions .
thus we first consider @xmath953-projectible pieces of @xmath954 : pieces which projects regularly from a neighborhood of @xmath955 to a neighborhood of @xmath956 restricted to densities supported on such pieces , the maslov canonical operator is truly canonical : if @xmath957 is a phase locally parametrizing @xmath937 , then @xmath958 = e^{\frac{i}{h } s } \cdot \sum_{j=0}^{\infty } h^j a_j\ ] ] for some smooth coefficients @xmath959 .
we may then substitute this expression into the eigenvalue problem and obtain eikonal and transport equations .
the eikonal equation @xmath960 has been solved by our choice of phase , so the transport equations become : @xmath961 \end{array } \right.\ ] ] and so on . as is well - known ( cf .
e.g. [ b.w ] ) , these equations may be put into geometric form by observing that @xmath962 is the projection to @xmath495 of @xmath945 and that @xmath963 |dx| = div ( a^2 \nabla s ) |dx| = { \cal l}_{\xi_h } ( a^2 |dx|).$ ] hence the transport equations become @xmath964 |dx|^{{\frac{1}{2 } } } \end{array } \right.\ ] ] the invariant 1/2-density is given by the well - known formula @xmath965 , or equivalently by @xmath966 on @xmath967 if we write @xmath968 then @xmath969 and the transport equations become @xmath970 \end{array } \right.\ ] ] here , the expressions in @xmath971 are understood to have been lifted up to @xmath937 .
the eigenvalue corrections @xmath939 are determined by integrating both sides of the transport equations over the level @xmath972 .
since the equations are solvable and since the left hand sides will integrate to zero , we get @xmath973 \rho \\
e_n^{(3 ) } = \frac{1}{2i t(e_n^{(1 ) } ) } \int_{\{h = e_n^{(1)}\ } } [ e_n^{(2 ) } \frac{a_1}{a_o } + i a_o^{-1 } \frac{d^2}{dx^2 } a_1 ] \rho \end{array } \right.\ ] ] here , @xmath974 is the period of @xmath944 at level @xmath837 .
parametrizing @xmath972 as a graph over the @xmath953-axis away from the turning points , the invariant density takes the form @xmath975 with @xmath976 .
hence at least formally the eigenvalue corrections are given by @xmath977 \frac{dx}{\sqrt{e - v } } = \frac{1}{2i t(e ) } \int_{x_{-}(e)}^{x_{+}(e ) } ( e - v(x))^{-\frac{1}{4 } } \frac{d^2}{dx^2 } ( e - v(x))^{-\frac{1}{4 } } dx.\ ] ] unfortunately the integral is ill - defined due to the singularities at the turning points .
the problem is that the maslov operator can not be defined near these points as a simple pull - back operator .
rather it should be defined as the composition of the fourier transform with the pull - back operator defined over the @xmath422-projection .
this problem and its solution constitute a key aspect of the maslov method ( in one dimension ) ; we refer to [ cv.3][b.w ] for extended discussions the point which is important for us is that the maslov method gives a regularization of the divergent integral .
it works in the following way : we introduce a cut - off @xmath978 supported away from a @xmath979-neighborhood of the turning points @xmath980 more precisely we define @xmath981 on @xmath937 , equal to one on @xmath982 resp . @xmath983 and equal to zero on @xmath984 resp .
@xmath985 we then put @xmath986 where @xmath987 is the pull - back to @xmath495 under the phase parametrization by @xmath988 and where @xmath989 is the fourier transform of the @xmath422-parametrization .
the notation @xmath978 stands for @xmath981 . for details on @xmath990 ,
see [ b.w ]
. returning to the previous calculation of eigenvalue corrections , we see that what is missing is the cut - off @xmath978 in the integrals and the contributions from @xmath989 .
we wish to avoid confronting the latter .
fortunately , it is not necessary to do so : the fact that the eigenvalues are independent of @xmath979 allows us to determine the @xmath989 ( i.e. the turning point ) contribution indirectly . to see this
, we substitute the cut - off into the formula for @xmath991 to get @xmath992 dx + ii_{\delta}\ ] ] with @xmath993 the contribution from @xmath989 .
since the integral is now nicely convergent we can integrate by parts and simplify to the form @xmath994 the first term tends to @xmath995 as @xmath996 .
the second expression only involves the taylor expansion of @xmath997 near the turning points .
its singular part must be cancelled by the singular part of @xmath993 , leaving a possible ` residue ' as @xmath998 we claim that this residue is zero : in fact , this is known to be the case since the first term is well - defined , independent of @xmath979 , and agrees with the formula given in [ m ] .
to give an independent proof that it vanishes , without analysing the @xmath989-term in detail , we observe that the limit contribution involves only the 2-jet of @xmath997 at the turning points .
hence it must agree with the corresponding expression for a harmonic oscillator at its turning points .
but no such correction occurs .
v.i.arnold , _ mathematical methods of classical mechanics _ , springer - verlag , new york ( 1977 ) .
s.bates and a.weinstein , _ lectures on the geometry of quantization _ , berkeley lecture notes ( 1996 ) .
p.berard , quelques remarques sur les surfaces de revolution dans @xmath999 .
c.r.acad.sci.paris t.282 ( 1976 ) , 159 - 161 .
a.besse , _ on manifolds all of whose geodesics are closed _ ,
93 , springer - verlag , new york ( 1978 ) .
p.bleher , distribution of energy levels of a quantum free particle on a surface of revolution , duke math .
j. 74 ( 1994 ) , 45 - 93 .
l. boutet de monvel and v. guillemin : _ the spectral theory of toeplitz operators _ , ann.math .
studies * 99 * , princeton u. press ( 1981 ) .
j.bruning and e.heintze , spektrale starrheit gewisser drehflachen , math.ann .
269 ( 1984 ) , 95 - 101 .
g.darboux , _ theorie des surfaces iii _
, gauthier - villars , paris ( 1894 ) .
y.colin de verdiere , spectre conjoint doperateurs pseudo - differentiels qui commutent i : le cas non - integrable , duke math .
j. 46 ( 1979 ) , 169- 182 .
y.colin de verdiere , sur les longuers des trajectoires periodiques dun billard , in : p.dazord and n. desolneux - moulis ( eds . ) _ geometrie symplectique et de contact : autour du theoreme de poincare - birkhoff .
travaux en cours , sem .
sud - rhodanien de geometrie iii _ pairs : herman ( 1984 ) , 122 - 139 .
y.colin de verdiere , quasi - modes sur les varietes riemanniennes , inv . math .
43 ( 1977 ) , 15 - 52 .
c.croke and b.kleiner , conjugacy and rigidity for manifolds with a parallel vector field , j.d.g .
39 ( 1994 ) 659 - 680 .
duistermaat , on the existence of global action - angle variables , comm.pure.appl.math .
33 ( 1980 ) , 687- 706 .
j.j.duistermaat and v.guillemin , the spectrum of positive elliptic operators and periodic bicharacteristics , invent.math .
29 ( 1975 ) , 29 - 79 .
francoise and v. guillemin : on the period spectrum of a symplectic mapping , j. fun .
anal . , ( 1991 ) 317358 .
i.m.gelfand and g.e.shilov , _ generalized functions , volume 1 _ , academic press , new york ( 1964 ) .
v.guillemin , wave trace invariants , duke math j. 83 ( 1996 ) , 287- 352 .
v.guillemin and s.sternberg , _ symplectic techniques in physics_. cambridge univ . press , new york ( 1985 ) .
v.guillemin and s.sternberg , homogeneous quantization and multiplicities of group representations , j.funct.anal .
47 ( 1982 ) , 344- 380 .
d.gurarie , semiclassical eigenvalues and shape problems on surfaces of revolution , j.math.phys .
36 ( 1995 ) 1934 - 1944 .
m.kac , on applying mathematics : reflections and examples , in mark kack : probability , number theory , and statistical physics , k.baclawski and m.d.donskder ( eds . ) , mit press , cambridge ( 1979 ) .
w. klingenberg , _ lectures on closed geodesics _ , grundlehren der .
w. * 230 * , springer - verlag ( 1978 ) .
v.p.maslov , _ theorie des perturbations et methods asymptotiques _ , dunod , paris ( 1972 ) .
g.popov , length spectrum invariants of riemannian manifolds , math.zeit . 213 (
1993 ) , 311 - 351 .
s.zelditch , wave invariants at elliptic closed geodesics , to appear in gafa .
, wave invariants at non - degenerate closed geodesics ( preprint , 1996 ) .
, isospectrality in the fio category , j. diff .
* 35 * , 689 - 710 ( 1992 ) . ,
kuznecov sum formulae and szeg limit formulae on manifolds , comm .
pde , * 17 * ( 1992 ) , 221 - 260 . | we prove that isospectral simple analytic surfaces of revolution are isometric . the inverse spectral problem for surfaces of revolution steve zelditch + johns hopkins university , baltimore , maryland 21218 + [email protected] research partially supported by nsf grants # dms-9404637 and # dms-9703775 . |
we would like to thank s. pratt for providing the crab program and acknowledge support by the frankfurt center for scientific computing ( csc ) .
we also thank m gyulassy , h j drescher , and s haussler for helpful discussions .
q. li thanks the frankfurt institute for advanced studies ( fias ) for financial support .
this work is partly supported by gsi , bmbf , and volkswagenstiftung .
t. d. lee , nucl .
a * 750 * ( 2005 ) 1 .
m. gyulassy and l. mclerran , nucl .
a * 750 * ( 2005 ) 30 .
e. v. shuryak , nucl .
phys . a * 750 * ( 2005 ) 64 .
l. m. satarov , h. stoecker and i. n. mishustin , phys .
b * 627 * ( 2005 ) 64 .
r. a. janik and r. peschanski , phys .
d * 73 * ( 2006 ) 045013 .
t. hirano and m. gyulassy , nucl .
phys . a * 769 * ( 2006 ) 71 .
d. h. rischke and m. gyulassy , nucl .
a * 608 * ( 1996 ) 479 .
m. a. lisa , s. pratt , r. soltz and u. wiedemann , ann .
. part .
* 55 * ( 2005 ) 357 .
p. chung and p. danielewicz [ na49 collaboration ] , arxiv:0707.2690 [ nucl - ex ] .
p. danielewicz and s. pratt , phys .
c * 75 * ( 2007 ) 034907 .
q. li , m. bleicher and h. stoecker , j. phys .
g * 34 * ( 2007 ) 2037 .
p. huovinen , arxiv : nucl - th/0305064 .
p. f. kolb and u. w. heinz , arxiv : nucl - th/0305084 .
p. huovinen , nucl .
a * 761 * ( 2005 ) 296 .
e. frodermann , u. heinz and m. a. lisa , phys .
c * 73 * ( 2006 ) 044908 .
s. a. bass _ et al .
_ , [ urqmd - collaboration ] , prog .
* 41 * ( 1998 ) 255 .
j. czyzewski , acta phys .
b * 21 * ( 1990 ) 41 .
a. accardi , v. muccifora and h. j. pirner , nucl .
a * 720 * ( 2003 ) 131 ; w. cassing , k. gallmeister and c. greiner , nucl .
a * 735 * ( 2004 ) 277 ; f. arleo , eur .
j. c * 30 * ( 2003 ) 213 ; b. z. kopeliovich , j. nemchik and i. schmidt , nucl .
phys . a * 782 * ( 2007 ) 224 .
a. bialas and m. gyulassy , nucl .
b * 291 * , 793 ( 1987 ) ; a. bialas and j. czyzewski , phys .
b * 222 * , 132 ( 1989 ) .
t. j. humanic , int .
j. mod .
e * 15 * ( 2006 ) 197 .
m. hofmann , m. bleicher , s. scherer , l. neise , h. stoecker and w. greiner , phys .
b * 478 * ( 2000 ) 161 ; s. haussler , s. scherer and m. bleicher , arxiv : hep - ph/0702188 .
q. li , z. li , s. soff , m. bleicher and h. stoecker , j. phys .
g * 32 * ( 2006 ) 151 .
b. a. li , c. b. das , s. d. gupta and c. gale , phys . rev .
c * 69 * ( 2004 ) 011603 .
m. isse , a. ohnishi , n. otuka , p. k. sahu and y. nara , phys .
c * 72 * ( 2005 ) 064908 .
m. a. lisa _ et al . _
[ e895 collaboration ] , phys .
* 84 * ( 2000 ) 2798 .
s. kniege _ et al . _
[ na49 collaboration ] , aip conf .
* 828 * ( 2006 ) 473 . c. adler _ et al . _
[ star collaboration ] , phys .
* 87 * ( 2001 ) 082301 .
j. adams _ et al . _ [ star collaboration ] , phys . rev .
c * 71 * ( 2005 ) 044906 .
s. s. adler _ et al . _
[ phenix collaboration ] , phys .
* 93 * ( 2004 ) 152302 .
k. adcox _ et al . _
[ phenix collaboration ] , phys .
* 88 * ( 2002 ) 192302 .
h. weber _ et al .
_ , phys .
b * 442 * ( 1998 ) 443 .
[ arxiv : nucl - th/9808021 ] .
h. petersen and m. bleicher , pos * cpod2006 * ( 2006 ) 025 .
q. li , m. bleicher and h. stoecker , phys .
c * 73 * ( 2006 ) 064908 .
q. li , m. bleicher , x. zhu and h. stoecker , j. phys .
g * 33 * ( 2007 ) 537 .
j. g. cramer , g. a. miller , j. m. s. wu and j. h. s. yoon , phys .
* 94 * ( 2005 ) 102302 [ erratum-_ibid_. * 95 * ( 2005 ) ] 139901 .
j. adams _ et al . _ [ star collaboration ] , phys .
c * 72 * ( 2005 ) 014904 . | the nuclear stopping , the elliptic flow , and the hbt interferometry are calculated by the urqmd transport model , in which potentials for pre - formed " particles ( string fragments ) from color fluxtube fragmentation as well as for confined particles are considered .
this description provides stronger pressure at the early stage and describes these observables better than the default cascade mode ( where the pre - formed " particles from string fragmentation are treated to be free - streaming ) .
it should be stressed that the inclusion of potential interactions pushes down the calculated hbt radius @xmath0 and pulls up the @xmath1 so that the hbt time - related puzzle disappears throughout the energies from ags , sps , to rhic .
one of the most important discoveries from sps and rhic experiments in the beginning of this century is that a new form of matter is produced in high energy ion collisions that seems to behave as a strongly interacting quark gluon plasma ( sqgp ) @xcite .
furthermore , at rhic energies , the sqgp seems to be a nearly perfect liquid " with small viscosity @xcite .
although hydrodynamical studies are a useful tool for the theoretical investigations at rhic energies , hydrodynamics fails at lower energies .
therefore , a relativistic dynamic transport approach , in which an entire equilibrium is not pre - assumed , is necessary to explore the excitation functions of various observables .
in addition , the transport model has the advantage of observing the whole phase - space evolution of all particles involved in a microscopic fashion . in order to detect this new matter - quark gluon plasma ( qgp ) - many theoretical suggested observables
have been discussed and argued frequently . among these ,
the hanbury - brown - twiss interferometry ( hbt ) or femtoscopy technique has been used widely to extract the spatio - temporal information of the particle freeze - out source .
it would be a vital discovery if there is an nontrivial transition in the spatio - temporal characteristics of the source , when going from low to high beam energies @xcite .
experimentally , the hbt parameters have been scanned thoroughly over the energies from sis , ags , sps , up to rhic , unfortunately , no obvious discontinuities do appear @xcite . however , this null " result does not mean that the hbt technique comes to an end . non - gaussian effects might shadow the possible energy dependence of the hbt parameters ( _ e - puzzle _ " ) , and this topic has been quickly improving in the recent years @xcite . secondly ,
even with a gaussian parametrization , we found that the hbt time - related puzzle ( _ t - puzzle _ " ) is present at almost all energies from ags to rhic @xcite . in order to understand the origins of these puzzles " , it is necessary to dig deeper into the dynamics of the heavy ion collisions .
the hydrodynamic model , in which various kinds of equations of state ( eos ) with latent heats are considered , successfully explained the elliptic flow @xmath2 at transverse momentum @xmath3gev@xmath4 at rhic energies @xcite .
however , hydrodynamics failed to explain the experimental @xmath5 data ( @xcite and the references therein ) , i.e. the ratio between the hbt radii in outward and sideward directions , where the t - puzzle " initially occurred . but
recently , it has been realized that a direct ( fully apple - to - apple " ) comparison between hydro calculations and data is required @xcite . in the transport models ,
the effective eos is softened by the existence of resonances , but the re - scattering between particles will reinforce to some extent the stiffness of the dense matter . at sps and rhic energies ,
the color - string excitation and fragmentation is further considered for particle production and subsequent evolution in the ultra - relativistic / relativistic quantum molecular dynamics ( urqmd / rqmd ) hadron - string transport models .
the formation time of the hadrons from string fragmentation is determined by a
yo - yo " mode @xcite . during the formation time , the pre - formed " particles ( string fragments that will be projected onto hadron states later on ) are usually treated to be free - streaming , while reduced cross sections are only included for leading hadrons .
the idea of pre - formed " hadrons is a well - known concept in the description of deep inelastic scattering data ( dis ) @xcite .
it has also been found that the pre - formed " hadrons from the color fluxtube fragmentation behave already very hadronlike , before the hadron formation is fully completed @xcite . in previous calculations
@xcite , the interaction of the newly created
pre - formed " particles from the string fragmentation is usually not taken into account for simplicity .
the facts have shown that the effective eos at the early stage in the past model simulations is too soft and the early pressure is inadequate to match the calculated elliptic flow to the experimentally observed flows at rhic energies @xcite . in order to make up for this discrepancy in the early - stage eos ,
it is essential to modify the dynamics of the early stage with / without new degrees of freedom , such as done in a multi - phase transport ( ampt ) @xcite , the hadronic rescattering ( hrm ) @xcite , the quark molecular dynamics ( qmd ) @xcite , and the parton hadron string dynamics ( phsd ) @xcite model .
besides the new matter which is dominant at the early stage , the potentials between confined hadrons which play a role during almost all time of the collision process deserve attention as well . in the previously low energy urqmd calculations ,
the hamiltonian @xmath6 served for the eos consists of the kinetic energy @xmath7 and the effective two - body interaction potential energy @xmath8 .
the potential energy @xmath8 is indispensable , which includes the two - body and three - body ( which can be approximately written in the form of two - body interaction ) skyrme- ( also called as the density dependent terms ) , yukawa- , and coulomb - terms @xcite . recently , in order to be more successfully applied in the intermediate energy region ( @xmath9a gev ) , more terms are considered @xcite , those are , the density - dependent symmetry potential ( essential for isospin - asymmetric reactions at intermediate and low energies ) and the momentum - dependent term . at higher energies ,
i.e. , ags / fair and sps , the yukawa- and symmetry- potentials of baryons becomes negligible , while the skyrme- and the momentum - dependent part of potentials still influence the whole dynamic process of hics though weakly . in ref .
@xcite , the effects of the mean field with momentum dependence on collective flows from hics at 2 - 158a gev energies were studied by a jet aa microscopic transportation model ( jam ) and it has been found that the momentum dependence in the nuclear mean field is important for the understanding of the proton collective flows at ags and even at sps energies .
further , by adopting the same soft eos with momentum dependence ( sm - eos ) on confined baryons , we also found that the calculated @xmath5 ratio from hics at ags energies can be vastly improved when compared to the measured data @xcite . in this work
, we will follow this idea and investigate the hbt radii at higher beam energies , i.e. sps and rhic .
furthermore , the other observables , such as the rapidity distributions of net protons ( @xmath10 ) and pions , and the elliptic flows of charged hadrons , are checked at the same time . as an initial attempt
, we consider potentials for both formed and pre - formed " particles : 1 ) , for formed baryons , the sm - eos @xcite is chosen . the coulomb potential is also considered , as in the low - energy calculations @xcite .
2 ) , for formed mesons , only the coulomb potential is optional in calculations , while other potentials are not considered as in previous calculations at low energies @xcite .
3 ) , for pre - formed " particles from string fragmentation , the similar density dependent ( skyrme - like ) terms , where @xmath11 is the hadronic density , @xmath12 is the normal nuclear density . @xmath13 , @xmath14 , and @xmath15 are parameters , in this work for the sm - eos , they are , @xmath16 mev , @xmath17 mev , and @xmath18 , respectively . ] as the formed baryons are used , but without the yukawa , the coulomb , and the momentum dependent terms .
4 ) , the pre - formed " mesons act like pre - formed " baryons but with a reduction factor ( @xmath19 ) due to the quark - number difference .
furthermore , 5 ) , there is no potential interaction between pre - formed " baryons and formed baryons so far .
correspondingly , the hadronic density is calculated by @xmath20 , where @xmath21 for baryons , @xmath19 for pre - formed " mesons , and @xmath22 for formed mesons .
@xmath23 is a gaussian in coordinate space . as stated in @xcite , the relativistic effect ( lorentz transformation ) on the relative distance and relative momentum between two particles @xmath24 and
@xmath25 has been considered . to set the stage , fig .
[ fig1 ] shows the rapidity dependence of the multiplicities of net protons ( top ) and negatively charged pions ( bottom ) for central ( @xmath26 of total cross section @xmath27 ) pb+pb collisions at @xmath28a gev .
the experimental data ( solid stars ) are taken from @xcite , and the open stars represent the data reflected at mid - rapidity .
the dash - dot - dotted lines are the cascade mode calculations , the dashed lines represent the calculations with sm - eos of confined baryons , while the solid lines illustrate the results with the additional density - dependent potentials for pre - formed " particles .
while the cascade calculations give a reasonable rapidity distribution for the pions , the rapidity distribution of net protons is gaussian - as seen in previous urqmd calculations @xcite , and not fully consistent with the data ( which has a two - bump structure at @xmath29 ) . after considering the baryonic potentials , the distribution only changes slightly at projectile and target rapidities .
however , when the density - dependent eos for pre - formed " particles is additionally taken into account , one observes that two bumps in the rapidity distribution of net protons appear .
this implies that , due to the interactions of pre - formed " particles at early stage , the additional pressure increases the effective stopping power as well as the central density at the compression stage .
the total pion multiplicity remains basically unchanged on a level below @xmath30 at this beam energy .
thus , the current pion distribution with potentials is acceptable for further investigations . of total cross section )
pb+pb collisions at @xmath28a gev .
the experimental data ( solid stars ) are taken from @xcite , while the open stars represent the data reflected at @xmath31 in the center - of - mass system .
the dash - dot - dotted lines are the calculations with cascade mode , the dashed lines represent the calculations with potentials for confined baryons ( sm - eos ) , while in the solid lines , the density - dependent potentials for pre - formed " particles are further taken into account .
, scaledwidth=40.0% ] using the analyzing program correlation after - burner " ( crab v3.0@xmath14 ) , which is contributed by s. pratt @xcite , we calculate the hbt correlator and further fit it as a three - dimensional gaussian form under the pratt convention , i.e. , the longitudinally comoving system ( lcms ) , which is expressed as @xmath32 in eq . ( [ fit1 ] ) the @xmath33 is the incoherence factor , @xmath34 , @xmath0 , and @xmath1 are the pratt radii in longitudinal , outward , and sideward directions , while the cross - term @xmath35 plays a role at large rapidities .
@xmath36 is the pair relative momentum in the @xmath24 direction .
fig .
[ fig2 ] shows the transverse momentum @xmath37 dependence ( @xmath38 ) of the hbt - radii @xmath34 , @xmath0 , and @xmath1 ( at midrapidity ) for central hics at ags ( @xmath39a , @xmath40a , @xmath41a , and @xmath42a gev ) , sps ( @xmath43a , @xmath44a , @xmath45a , @xmath46a , and @xmath47a gev ) , and rhic ( @xmath48 and @xmath49 gev ) energies .
the physical cuts used are same as those listed in @xcite .
the data are from the e895 , na49 , star , and phenix collaborations @xcite .
the calculations with cascade mode are shown by lines with down - triangles while the calculations with various potential treatments are shown with other lines with different symbols . at @xmath39a gev
, we show a comparison of calculations between with and without coulomb potentials for confined mesons .
it is seen that the two - body mesonic coulomb potential before freeze - out affects the hbt radii only very weakly at about @xmath50mev@xmath4 , and the ratio between @xmath0 and @xmath1 values is not altered consequently . at ags energies , besides the nuclear potentials , the coulomb potential of confined mesons is also considered ( solid lines with circles ) , while at sps and rhic energies , the coulomb potential for confined mesons is switched off in order to save computing times and to avoid problems with non - local interactions .
it is known that , below ags energies hadrons are dominantly produced from the decay of resonances so that the considered potentials for pre - formed " particles from string fragmentations have no effect on the hbt radii @xcite . at higher energies
pre - formed " particles play a dominant role for the early stage dynamics , which can be seen , e.g. , by the comparison of solid lines and the lines with triangles at @xmath51 gev . so far
the question of a steeper @xmath37-dependence of hbt radii than data at high sps and rhic energies , similar to the ampt transport model calculations with parton scatterings @xcite , is still open .
the most exciting results show up in the transverse space . roughly speaking
, the nuclear potentials lead to a smaller @xmath0 ( especially at large @xmath37 ) but a larger @xmath1 ( especially at small @xmath37 ) generally allowing for a better description of the @xmath37 dependent data .
certainly , our present crude approach is only semi - realistic as a description of the dynamic processes at the early stage of the reaction .
however , one can still come to the conclusion that the potentials help to obtain a better description of the pion freeze - out in the transverse space . in fig .
[ fig3 ] we depict the excitation function of the @xmath5 ratio at small @xmath37 : at ags and sps energies , the results at @xmath52mev@xmath4 are shown , while at rhic energies , the @xmath37 is set to @xmath53mev@xmath4 .
the experimental data within the transverse momentum regions are compared with the calculations with and without potentials . as seen before @xcite , in the cascade mode ( dash - dot - dotted lines with open rectangles )
the @xmath5 ratio , which is related to the duration of the emission @xmath54 ( @xmath55 ) , is larger than the experimentally observed values at all investigated energies .
furthermore , the increase of the @xmath5 ratio with beam energies at ags is not seen in these calculations .
when the sm - eos is considered for confined baryons ( dashed lines with open up - triangles ) , the @xmath5 ratio decreases compared to the result in the cascade mode and reproduces the energy dependence of the data up to the lower sps energies . at high sps and rhic energies , however , the @xmath5 ratio increases further with increasing beam energies and deviates strongly from the data at top rhic energies . here
, the ratio nearly approaches the value obtained in the cascade mode .
it implies that the potential of confined hadrons is increasingly losing its importance with increasing beam energy . at sps and rhic
energies the deviation from data can be interpreted by the absence of interactions of pre - formed " particles from string fragmentation since at @xmath56a gev the measured @xmath5 ratio begins to flatten / decrease in contrast to the aforementioned calculations .
however , with the consideration of a density dependent eos for pre - formed " hadrons ( solid lines with open down - triangles ) this behavior can be reproduced in the calculations . with this approach , at rhic energies ,
the ratio @xmath5 is about unity ( slightly below data ) which implies that the compressibility provided by the currently employed potential is too strong .
nevertheless , since we are not aiming at fitting the whole excitation function of the data , we leave this detail aside for further detailed explorations . therefore , our explanation of the hbt time related puzzle is different from the investigations in ref .
@xcite where cramer , et al . , considered a single - pion optical potential which simulates its interactions with the dense medium .
while in this work , the microscopic
nuclear " potentials are considered for pre - formed " hadrons from the string fragmentation .
the hamilton s equations of motion of the particles represented by gaussian wave packets are solved microscopically .
after that , the freeze - out phase space is analyzed numerically by the crab program , in which the plain and distorted ( by coulomb and nuclear potentials after freeze - out ) wave functions are optional .
ratio at small @xmath37 .
the data are indicated by solid stars .
the dash - dot - dotted lines with open rectangles are results under cascade mode , the dashed lines with open up - triangles represent the calculations with potentials for confined baryons , while the solid lines with open down - triangles are results with potentials for both pre - formed " and confined particles .
, scaledwidth=40.0% ] let us finally check the modification of the early transverse expansion by a look into the elliptic flow . in the upper plots of fig .
[ fig4 ] we present the transverse momentum @xmath57 dependence of the elliptic flow @xmath2 ( @xmath58 , where @xmath59 ) of charged pions from au+au collisions at rhic @xmath51 gev .
two sets of centralities are chosen : @xmath60 ( in ( a ) ) and @xmath61 ( ( b ) ) . to avoid non - flow effects @xcite ,
the four - particle integral cumulant @xmath2 data ( @xmath62 ) are shown with stars .
similar to those in fig.[fig3 ] , the urqmd calculations without and with potentials are illustrated for comparison with data .
it is clear that due to the absence of early pressure , the elliptic flow in the cascade mode ( dash - dot - dotted lines with open rectangles ) can not grow as fast as the data with the increase of transverse momentum in both centralities . after switching on the sm - eos for confined baryons ( dashed lines with open up - triangles ) ,
the flow is seen to increase slightly especially at high transverse momenta . with
the density dependent potentials for pre - formed " particles ( solid lines with open down - triangles ) , the elliptic flow at large transverse momenta rises further , however , at small transverse momenta it is driven down and deviates from data especially at large centralities .
this phenomenon calls for a more complete description of the dynamics of hics at high energies in which the ( in)elastic scatterings of particles ( confined hadrons and deconfined partons ) should be considered in more detail . to make a whole comparison , we show in fig .
[ fig4 ] ( c ) the centrality dependence of the ratio between different theoretical @xmath2 values of charged hadrons and the four - particle cumulant data ( @xmath63 " , here the flows are integrated over @xmath57 and @xmath64 ) .
as stated before , the flow obtained from the cascade calculation is only about @xmath65 of the experimental one .
the contribution of potentials of confined baryons to flow is only at the level of about @xmath66 .
while the potentials of pre - formed " hadrons increase the flow in more central collisions . at large centrality , the potential effects are relatively weak .
dependence of the elliptic flow @xmath2 for au+au collisions at rhic @xmath51 gev .
the pseudo - rapidity cut is @xmath67 .
two sets of centralities are chosen : @xmath60 ( in ( a ) ) and @xmath61 ( ( b ) ) .
stars represent the four - particle integral cumulant @xmath2 data ( @xmath62 ) .
dash - dot - dotted lines with open rectangles represent the cascade calculations . dashed lines with open up - triangles
are the results with confined baryonic potentials , while the solid lines with open down - triangles are the results with potentials for both
pre - formed " particles and confined baryons .
bottom plot ( ( c ) ) : centrality dependence of the ratio between theoretical and experimental flows of charged hadrons , which are integrated over @xmath57 and @xmath64.,scaledwidth=40.0% ] in summary , besides a soft equation of state with momentum dependence , which is required especially at low energies , a density dependent ( skyrme - like ) potential has been tested for the pre - formed " particles from string fragmentation in the urqmd transport model .
although the form of the potential for the new phase is simple and rough ( ideally , the eos of the new phase should be based on the first - principle lattice qcd calculations @xcite ) , it provides new insights into the dynamics of hics .
it reproduces the proper stopping power for net protons .
it also improves the elliptic flow at large transverse momenta .
most importantly however , it decreases the hbt radius @xmath0 but increases the @xmath1 so that the experimental duration time related quantity here the @xmath5 ratio can be reasonably reproduced throughout the energies from ags , sps , up to rhic . although some open problems remain due to the simple treatment of the interaction of the particles at the early stage . |
numerical simulations of qcd can be performed by discretizing euclidean space - time onto a lattice and working within a finite volume .
physical measurements are then obtained by extrapolating to zero lattice spacing and infinite lattice volume .
there are many lattice actions which correspond to continuum qcd in the appropriate limit , and a lot of recent work has focussed on finding actions which reduce the errors due to discrete lattice spacing .
the goal is to find actions which provide accurate results on coarsely - spaced lattices and so allow lattice sizes ( and therefore computing resources ) to be reduced .
a progress report on improving actions may be found in ref .
@xcite . to demonstrate the value of an improved action
, one would like to show that accurate results can be obtained from coarse lattices for a wide variety of quantities . as a first step beyond the mass spectrum , where significant improvements on coarse lattices have already been demonstrated , we consider the decay constants of light vector mesons and the associated vector current renormalization factor .
( a preliminary discussion was given in ref .
an advantage of studying vector currents is the existence of a conserved vector current even at nonzero lattice spacing , which does not get renormalized .
neglecting lattice spacing errors , the matrix element of the conserved current between any initial and final states is identical to the matrix element of the local vector current between those same states , except for the multiplicative `` renormalization factor '' , @xmath4 . in this way
, @xmath4 can be easily computed . in practice
, it has been known for a long time@xcite that the determination of @xmath4 is quite sensitive to lattice spacing errors , which is another reason that the vector current is valuable for the present work , since it is precisely these errors that should be reduced by an improved action .
the action@xcite that will be used here is improved at the classical level up to and including @xmath3 , and is also improved at the quantum level via the `` tadpole factor '' of lepage and mackenzie@xcite .
the action has the added feature of being quite simple , since the gauge sector contains only @xmath5 and @xmath6 plaquettes , and the fermion sector does not contain the so - called `` clover '' term and has no need of the associated field transformations .
the mass spectrum of this improved action has been studied numerically in ref .
@xcite , where it was concluded that data from the original wilson action@xcite at a lattice spacing of @xmath7fm can be reproduced with the improved action on lattices with @xmath8fm , representing a very substantial saving of computing resources .
other studies of the mass spectrum with the same improved action have produced similar results@xcite . in this paper
it is shown that a comparable degree of improvement is also obtained in the computation of the vector meson decay constants , @xmath9 , @xmath10 and @xmath11 , and the renormalization factor , @xmath4 .
[ sec : method ] begins with a review of the general methods for calculating vector current matrix elements , and then discusses the particular improved action chosen for this work .
[ sec : treelevel ] is used to show explicitly that the quantities to be computed are classically improved .
the results of numerical simulations using the improved action are compared to wilson action results in sec .
[ sec : simulation ] , and a summary is given in sec .
[ sec : conclude ] .
the @xmath0 meson decay constant , @xmath9 , is defined from a continuum matrix element of the local vector current as follows .
@xmath12 @xmath13 a similar definition applies to @xmath10 and @xmath11 . recall that this current
is conserved in the continuum . at nonzero lattice spacing the local current
is not conserved , and its matrix element between any initial and final states gets renormalized by a factor @xmath4 , @xmath14 @xmath15 is the lattice coupling , and @xmath16 represents the contributions from operators of higher mass dimension which are multiplied by powers of the lattice spacing , @xmath17 . for the wilson action these corrections begin at @xmath18 , but for our improved action no @xmath19 classical terms can appear , and the tadpole factor is expected to approximately minimize terms proportional to the coupling , for example @xmath20 terms .
if all quark flavours have the same mass , then both the improved action and the wilson action have an exact vector ( flavour ) symmetry , which implies the existence of a conserved `` noether '' current , @xmath21 .
( actually , there is not a unique conserved current , and care must be used to choose one that is classically improved to the appropriate order
. this will be addressed below . ) because of the nonrenormalization of conserved currents , @xmath22 which allows @xmath4 to be computed from a ratio of matrix elements as follows .
@xmath23 although @xmath4 itself is independent of the initial and final states , the ratio of matrix elements in eq .
( [ ratio ] ) is not , so the size of the lattice spacing errors can be estimated by varying @xmath24 and @xmath25 . in this work , we will consider four separate choices for the ratio in eq .
( [ ratio ] ) : a vector current insertion between pseudoscalar meson , vector meson or spin-1/2 baryon states , @xmath26 and the decay process , @xmath27 in our lattice simulations , states will be created via the following local interpolating fields
. @xmath28 \label{interpol2}\end{aligned}\ ] ] ( @xmath29 is the charge conjugation matrix . ) the decay constant itself is obtained from simulations of the two - point vector correlator , @xmath30 which is derived by multiplying eq .
( [ definefv ] ) with its hermitian conjugate , converting from continuum to lattice quantities , and performing the appropriate summations .
the explicit dependence on @xmath4 can be changed by replacing one ( or both ) of the local currents in eq .
( [ calcfv ] ) by a conserved current . on a lattice with periodic boundary conditions
, another term must be added to the right - hand side of eq .
( [ calcfv ] ) which decays exponentially from the ( periodic ) image of the source . following the work of lscher and weisz@xcite ,
the gauge field term of the improved action involves a sum over @xmath6 rectangular plaquettes ( @xmath31 ) as well as @xmath5 elementary plaquettes ( @xmath32 ) , @xmath33~,\ ] ] where a tadpole factor , defined by @xmath34 has been introduced to absorb the lattice tadpole effects and thereby improve the matching to perturbation theory@xcite .
the fermion part , which was first introduced more than a decade ago@xcite , also contains next - nearest - neighbour interactions .
@xmath35 \nonumber \\ & & -\frac{\kappa}{6u_0}\sum_{x,\mu}\left[\bar{\psi}(x)(2-\gamma_{\mu } ) u_{\mu}(x)u_{\mu}(x+\mu)\psi(x+2\mu ) \right .
\nonumber \\ & & \phantom{-\frac{\kappa}{6u_0}\sum_{x,\mu}}\left .
+ \bar{\psi}(x+2\mu)(2+\gamma_{\mu } ) u_{\mu}^{\dagger}(x+\mu)u_{\mu}^{\dagger}(x)\psi(x)\right]\end{aligned}\ ] ] the presence of next - nearest - neighbour timesteps is a generic consequence of classical improvement beyond @xmath18 , and generates well - known artifacts , e.g. unphysical branches in the free dispersion relations , discussed for gauge fields in ref .
the effect is also seen for fermions , as discussed in following sections . the noether vector current corresponding to eq .
( [ s_f ] ) is @xmath36 although this current is conserved , an explicit calculation shows that it is not classically improved to the level of the action because of extra total derivative terms .
an improved conserved current could be written as follows@xcite , @xmath37 but the offending total derivatives in @xmath21 do not contribute to the quantities we wish to compute , as will be shown explicitly in sec .
[ sec : treelevel ] , so @xmath38 is not required for this work .
the matrix elements that are used to obtain @xmath9 , @xmath10 , @xmath11 and @xmath4 can be evaluated analytically at the classical level .
we will follow closely the approach and notation of el - khadra , kronfeld and mackenzie@xcite , paying particular attention to the effects of next - nearest - neighbour interactions , which were not present in the actions of ref .
@xcite .
the equation of motion that comes from eq .
( [ s_f ] ) can be written in the form @xmath39\psi(y ) = 0\ ] ] where the fourier transforms of @xmath40 and @xmath41 are @xmath42 respectively .
the on - shell condition is @xmath43 which is a quartic equation in @xmath44 .
[ fig : evsp ] shows the solutions for 3-momenta in the ( 1,1,0 ) direction and @xmath45 , compared to the wilson action result and the continuum @xmath46 line .
notice that the improved action has artifacts which are restricted to the region of large @xmath47 , and that the low - energy solution is closer to the continuum for a greater range of @xmath48 than is the wilson curve . also related to our discussion of decay constants is the limit of vanishing 3-momentum , @xmath49 , and fig .
[ fig : evsm ] shows the solutions of eq .
( [ kkll ] ) versus @xmath50 .
again , improvement is found with respect to the wilson action . in this case as well there are next - nearest - neighbour timestep artifacts which , however , remain at high energies for the values of @xmath50 that we will use in numerical simulations .
note that eq .
( [ kkll ] ) is only quadratic in the mass , and becomes the following unique expression for @xmath50 when @xmath51 .
@xmath52\ ] ] we now seek spinor solutions to the equations of motion , @xmath53u(\xi,{\bf p } ) & = & 0 \\ \left[-i\sum_\mu\gamma_\mu{k}_\mu(p)+l(p)\right]v(\xi,{\bf p } ) & = & 0\end{aligned}\ ] ] where @xmath54 is the spin index .
the equation of motion for @xmath55-spinors has been modified so that from now on @xmath56 is always a positive quantity .
the solutions are @xmath57}}\right]u(\xi,{\bf 0 } ) \\ v(\xi,{\bf p } ) & = & \left[\frac{i\gamma_\mu{k}_\mu(p ) + l(p)}{\sqrt{2l(p)\left[l(p)+{\rm sinh}ea\left(4-{\rm cosh}ea \right)/3\right]}}\right]v(\xi,{\bf 0})\end{aligned}\ ] ] where @xmath58 the general solution of eq .
( [ eom ] ) is @xmath59\ ] ] where @xmath60 annihilates a quark and @xmath61 creates an antiquark .
the normalization factor @xmath62 can be determined , for example , from the normalization of the quark propagator . by choosing the creation and annihilation operators to be normalized in the following way ,
@xmath63 the normalization factor is found to be @xmath64^{1/2}~.\ ] ] the vector meson decay amplitude and the insertion of a vector current between hadron states can now be calculated at the classical level using the matrix elements @xmath66 respectively . in each case
, kinematics are chosen that corresponds to the numerical simulations of sec .
[ sec : simulation ] , which means that the decay amplitude is studied in the rest frame ( @xmath67 ) , whereas the vector current insertion has no 4-momentum transfer ( @xmath68 ) .
a consequence of these choices is that only @xmath69 are nonzero for the decay amplitude , and only @xmath70 is nonzero for the vector current insertion .
using @xmath71 as the vector current , the decay amplitude is @xmath72u(\xi,{\bf p } ) \\ & = & n^2({\bf p})\left[\frac{4}{3}{\rm cos}p_ja-\frac{1}{3}{\rm cos}2p_ja \right]\bar{v}(\xi,{\bf 0})\gamma_ju(\xi,{\bf 0 } ) \\ & = & \bar{v}(\xi,{\bf 0})\gamma_ju(\xi,{\bf 0})\left[1+{\cal o}(a^3)\right ] \label{classdecay}\end{aligned}\ ] ] where eq .
( [ useful ] ) has been used .
a similar derivation shows that eq .
( [ classdecay ] ) also holds for the local vector current , @xmath73 , so the decay amplitude is classically improved up to @xmath3 for both vector currents .
the expression for the current insertion is @xmath74u(\xi,{\bf p } ) \\ & = & \delta^{\xi\xi^\prime}~.\end{aligned}\ ] ] notice that there are no finite-@xmath17 errors here as required by current conservation . for the local current
, it is found that @xmath75\end{aligned}\ ] ] which again displays the desired classical improvement .
thus , although we are using an _ un_improved conserved current , the particular calculations being considered in this work are still improved to the level of the action
. this conclusion will be valid for the numerical simulations as well , provided we create the external hadron states using operators that maintain the improvement .
local interpolating fields , such as those of eqs .
( [ interpol1]-[interpol2 ] ) which will be used in the present work , guarantee that classical improvement is not destroyed .
calculations were performed at two values of @xmath76 with the improved action , corresponding to lattice spacings of 0.4fm and 0.27fm as derived from the string tension . for comparison ,
the wilson action was also used to calculate at the same lattice spacings .
some details of the simulations are provided in table [ tab : details ] .
ccccccc ' '' '' lattice & @xmath77 & @xmath76 & @xmath78[fm ] & @xmath79 & @xmath80 & @xmath81 + + @xmath82@xmath83 & 200 & 6.25 & 0.4 & 0.162 , 0.168 , 0.174 & 0.1803(2 ) & 0.1658(8 ) + @xmath84@xmath85 & 100 & 6.8 & 0.27 & 0.150 , 0.154 , 0.158 & 0.1638(2 ) & 0.1561(5 ) + + @xmath82@xmath83 & 200 & 4.5 & 0.4 & 0.189 , 0.201 , 0.213 & 0.2189(2 ) & 0.2050(3 ) + @xmath84@xmath85 & 100 & 5.5 & 0.27 & 0.164 , 0.172 , 0.180 & 0.1857(3 ) & 0.1779(7 ) + in all cases , we used pseudo - heatbath updating with periodic boundary conditions for the gauge fields .
the lattices were thermalized with 4000 sweeps , and then 250 sweeps ( 200 sweeps ) were discarded between each pair of improved ( wilson ) configurations that was kept .
a stabilized biconjugate gradient algorithm was used for the fermion matrix inversion . to maximize the number of useful time slices , dirichlet boundary conditions were used for fermions in the time direction , with periodic boundary conditions in all spacial directions .
the source was placed two timesteps away from the time boundary in every case , and for the 3-point correlators of @xmath86 , @xmath87 and @xmath88 we attached the vector current insertion to the lattice s centre time slice .
[ fig : effmass ] shows the effective mass plot for a pseudoscalar meson with the improved action .
the non - monotonic behaviour near the source ( @xmath89 in fig .
[ fig : effmass ] ) is a consequence of next - nearest - neighbour interactions , as discussed by lscher and weisz@xcite .
the effect of these high - energy oscillations decreases as @xmath79 increases , and fig .
[ fig : effmass ] shows that there is already an identifiable plateau at our smallest value of @xmath79 . to extract the mass from an effective mass plot such as fig .
[ fig : effmass ] , or to obtain @xmath4 or @xmath90 from the appropriate plot , we require that the plateau contains at least three neighbouring points , and we never use the two time slices near the boundary ( 12 and 13 in fig .
[ fig : effmass ] ) which can be affected by the dirichlet boundary condition . the systematic uncertainties associated with choosing a plateau region
will not be presented since we do not expect these uncertainties to affect the principal conclusions of our study .
all statistical errors will be computed using half the distance between the 16th and 84th percentiles from 500 bootstrap samples .
the results for the renormalization factor , obtained from the improved action at @xmath91 , are shown for degenerate - mass quarks in fig .
[ fig : zrawequ ] . the value of @xmath4 that comes from using meson decay amplitudes ( @xmath92 ) in eq .
( [ ratio ] ) is significantly smaller than the values that come from inserting a vector current between pion , @xmath0 meson or nucleon states ( @xmath86 , @xmath87 , @xmath88 ) , and this difference is a reflection of the finite-@xmath17 errors that are present in eq .
( [ ratio ] ) . to decide
whether these finite-@xmath17 errors should be considered as large or small , a comparison is made to results from the wilson action .
[ fig : zrawequ ] shows that the improved action results agree remarkably well with wilson data at @xmath93 , while being clearly distinguishable from wilson data ( for @xmath92 ) at @xmath94 . by defining an effective hopping parameter
which averages over the two quark flavours in a meson or baryon , @xmath95 we can plot all of our data , involving degenerate and nondegenerate quarks , together .
[ fig : zrawtot ] shows that our improved action results for @xmath4 appear to be linear in @xmath96 . the chiral limit , @xmath97 , is reached by extrapolating @xmath4 , @xmath90 , pseudoscalar meson masses _ squared _ , vector meson masses and baryon masses linearly in @xmath98 , using three data points ( degenerate and nondegenerate quarks are used separately ) .
in this way , we obtain the results of fig .
[ fig : zlight ] for the renormalization factor without strange quark involvement . also shown
are the chiral limit values corresponding to the wilson results of fig .
[ fig : zrawequ ] , and our own wilson calculations at smaller @xmath76 s .
note that the improved action approximately resembles the trajectory of wilson data , but translated horizontally to lattice spacings about three times as large .
this is very similar to what was also found in a comparison of the mass spectra@xcite , and provides the first indication that the accuracy of calculations with tadpole - improved actions on coarse lattices persists beyond spectral quantities .
[ fig : zfv ] shows the @xmath0 meson decay constant without including the effect of @xmath4 , and with all wilson quarks normalized by @xmath99 .
again , it is found that the improved action gives results which resemble the wilson data , but at a lattice spacing about three times as large .
our best determination of the @xmath0 meson decay constant comes from the product of @xmath4 ( obtained from @xmath92 ) and @xmath100 , which is shown in fig . [ fig : fv ] .
the wilson calculations can be divided into two categories : those which compute @xmath4 nonperturbatively as we have done in this work , and those which employ a perturbative estimate of @xmath4 , denoted by @xmath101 .
( the data in fig .
[ fig : fv ] with @xmath101 use quarks which are _ not _
normalized by the naive @xmath99 , see the original papers for details . )
the graph shows that both the improved and wilson actions give nonperturbative results which are quite independent of lattice spacing over a large range of @xmath17 , and which are near the experimental value . in contrast , the wilson data using @xmath101 approaches a similar continuum limit but with a noticeable dependence on @xmath17 .
having now discussed @xmath90 in the chiral limit , let s return to our unextrapolated results , plotted in fig .
[ fig : fvraw ] .
notice that the slope of our data is consistent with the observed slope from the two experimental points @xmath102 and @xmath103 , but as @xmath104 the simulations appear to favour values larger than the experimental ones .
we find it interesting that the ukqcd data , also shown in fig .
[ fig : fvraw ] , displays the same tendency . in their paper@xcite
, the ukqcd group concludes that this shift toward larger @xmath90 values is the opposite of what would be expected from quenching , and they suggest that finite volume effects or lattice spacing effects may be responsible .
the testing of such possibilities within the context of our @xmath3 classically - improved , tadpole - improved action is much less computer - intensive than for the @xmath105 clover action , and might offer some valuable insight .
the hopping parameter corresponding to the strange quark as given in table [ tab : details ] , was determined from the experimental relation @xmath106 by linear interpolation in @xmath98 between the two neighbouring data points . using this
, table [ tab : fv ] gives our results for the three light vector meson decay constants and compares them to predictions from other groups . lcccc ' '' '' action & @xmath107[fm ] & @xmath102 & @xmath108 & @xmath103 + improved , @xmath109 & 0.32(1 ) & 0.278(11 ) & 0.255(6 ) & 0.236(2 ) + improved , @xmath91 & 0.242(5 ) & 0.301(8 ) & 0.284(5 ) & 0.265(3 ) + wilson , @xmath93 , ref .
@xcite & 0.0845(1 ) & 0.33(1 ) & & + wilson , @xmath110 , ref .
@xcite & 0.071(1 ) & 0.34(1 ) & & 0.38(1 ) + clover , @xmath105 , ref .
@xcite & 0.073(3 ) & @xmath111 & @xmath112 & @xmath113 + wilson , @xmath114 , ref .
@xcite & ? & 0.32(2 ) & 0.32(1 ) & + clover , @xmath114 , ref .
@xcite & ? & 0.25(2 ) & 0.24(1 ) & + experiment & & 0.28 & & 0.23 +
in this paper we have calculated vector current matrix elements and the renormalization factor in the light quark sector for an improved action on lattices with spacings of about 0.3 to 0.4fm .
this is the first step in going beyond the mass spectrum with improved actions on coarse lattices .
the action used here is particularly simple .
it has next - nearest - neighbour couplings to remove @xmath18 and @xmath3 errors at the classical level and also incorporates tadpole improvement .
no special tuning of parameters and no field transformations are required with this action .
vector meson decay constants and the vector current renormalization factor calculated with the improved action were found to be comparable to those with the wilson action on lattices with about 1/3 the spacing .
this is the same as the improvement found previously in mass calculations and it represents a very substantial saving in computer resources .
f. niedermayer , hep - lat/9608097 , to appear in the proceedings of the 14th international symposium on lattice field theory , st .
louis , u.s.a . , 1996 .
r. lewis and r. m. woloshyn , hep - lat/9607059 , to appear in the proceedings of the 14th international symposium on lattice field theory , st .
louis , u.s.a . ,
1996 . g. heatlie , c. t. sachrajda , g. martinelli , c. pittori , and g. c. rossi , nucl .
b * 352 * , 266 ( 1992 ) .
m. lscher and p. weisz , comm .
* 97 * , 59 ( 1985 ) .
h. hamber and c. m. wu , phys .
b * 133 * , 351 ( 1983 ) ; w. wetzel , phys .
b * 136 * , 407 ( 1984 ) ; t. eguchi and n. kawamoto , nucl .
b * 237 * , 609 ( 1984 ) .
g. p. lepage and p. b. mackenzie , phys .
d * 48 * , 2250 ( 1993 ) .
h. r. fiebig and r. m. woloshyn , phys .
b ( in press ) .
k. g. wilson , in _ new phenomena in subnuclear physics _ ,
( erice , 1975 ) , ed .
a. zichichi , plenum press , 1977 .
f. x. lee and d. b. leinweber , hep - lat/9606005 , to appear in the proceedings of panic 96 , world scientific .
a. borii and ph .
de forcrand , hep - lat/9608105 , to appear in the proceedings of the 14th international symposium on lattice field theory , st .
louis , u.s.a . , 1996 .
peter lepage , private communication .
a. x. el - khadra , a. s. kronfeld and p. b. mackenzie , report no .
fermilab - pub-96 - 074-t , hep - lat/9604004 ( unpublished ) .
m. lscher and p. weisz , nucl .
b * 240 * , 349 ( 1984 ) .
g. martinelli and c. t. sachrajda , nucl .
b * 316 * , 355 ( 1989 ) .
l. maiani and g. martinelli , phys .
b * 178 * , 265 ( 1986 ) .
qcdpax collaboration , y. iwasaki _ et .
d * 53 * , 6443 ( 1996 ) .
f. butler , h. chen , j. sexton , a. vaccarino and d. weingarten , nucl .
b * 421 * , 217 ( 1994 ) .
t. bhattacharya and r. gupta , phys .
d * 54 * , 1155 ( 1996 ) .
ukqcd collaboration , c. r. allton _ et .
d * 49 * , 474 ( 1994 ) . c. r. allton , v. gimnez , l. giusti , f. rapuano , hep - lat/9609017 , to appear in the proceedings of the 14th international symposium on lattice field theory , st .
louis , u.s.a . , 1996 . | the @xmath0 , @xmath1 and @xmath2 decay constants and the vector current renormalization factor are studied by using an @xmath3 classically - improved , tadpole - improved action .
tree - level calculations are used to show how the classical improvement of the action , involving next - nearest - neighbour timesteps , is transferred to the matrix elements .
simulations are performed on coarse lattices and compared to wilson results from both coarse and fine lattices .
the improved action data are found to resemble wilson data obtained at 1/3 of the lattice spacing , which is the same degree of improvement that is seen by comparing the mass spectra . |
qcd theory predicts that an abundance of low-@xmath5 scattered gluons ( minijets ) should be produced in relativistic nuclear collisions at rhic energies .
those gluons are believed to drive formation of the colored medium in heavy ion collisions @xcite .
if so , we may discover remnants of low-@xmath5 ( @xmath6 1 - 5 gev ) partons in the correlation structure of final - state hadrons .
initial studies of two - particle correlations in p - p collisions emphasized momentum subspace @xmath3 ( pseudorapidity and azimuth ) @xcite described at smaller @xmath4 in terms of string fragmentation @xcite .
angular correlations of fragments from hard - scattered partons ( jets ) were observed on @xmath3 at larger @xmath4 and with increasing @xmath7 @xcite . the _ two - component _ model of p - p collisions , with string and parton fragments forming distinguishable soft and hard components , describes particle multiplicity distributions at large @xmath7 @xcite and has also been introduced to the heavy - ion context @xcite .
two issues arise in a conventional study of parton fragment distributions : 1 ) the distribution of fragment momenta along the jet thrust axis ( fragmentation function ) and 2 ) the angular distribution of fragments relative to the thrust axis .
the thrust axis ( estimating the parton direction ) may be inferred from a collection of hadrons ( jet ) identified with the scattered parton , or ( especially in heavy ion collisions ) a ` high-@xmath4 leading particle ' may be used to estimate the parton momentum direction and magnitude . in the present analysis
we adopt no _ a priori _ jet or factorization hypothesis .
we study minimum - bias two - particle distributions on transverse rapidity space @xmath8 to obtain fragment _ distributions _ ( not fragmentation _ functions _ ) and on angle space @xmath9 to obtain fragment angular correlations .
particles in pairs are treated symmetrically , as opposed to asymmetric ` trigger ' and ` associated ' particle combinations .
we observe that hard - component correlations obtained with this minimum - bias analysis , in contrast to jet correlations obtained with the conventional leading- or trigger - particle approach , represent the _ majority _ of parton fragment pairs , those with @xmath10 gev / c .
analysis of nonperturbative low-@xmath5 parton fragment correlations has motivated some novel techniques , including use of transverse rapidity @xmath1 rather than momentum @xmath4 and formation of _ joint _ ( 2d ) angular _ autocorrelations_. symmetric analysis of fragment pairs requires a generalized treatment of fragmentation functions and jet angular correlations .
an immediate consequence has been observation of substantially asymmetric angular correlations about the thrust axis for low-@xmath5 parton collisions .
while we attempt to understand qcd in a - a collisions we should also revisit its manifestations in elementary collisions , where novel phenomena are still emerging .
in a study of @xmath4 spectra from p - p collisions at @xmath0 gev we observed that the spectra can be separated into a soft component ( string fragments - longitudinal fragmentation ) described by a lvy distribution on transverse mass @xmath12 and a ` semi - hard ' component ( parton fragments - transverse fragmentation ) described by a gaussian distribution on transverse rapidity @xmath1 , based on systematic variation with event multiplicity @xmath13 .
the hard component is interpreted as fragments of _ minimum - bias _ ( mainly low-@xmath5 ) partons . when a soft - component lvy distribution ( defined as the limiting case of the @xmath13 dependence ) is subtracted from @xmath4 spectra for ten multiplicity classes
we obtain distributions in fig .
[ lowq2 ] ( first panel ) , described by hard - component model distributions ( solid curves ) with gaussian shape essentially independent of @xmath13 .
we have transformed to transverse rapidity @xmath14 with @xmath15 assigned as the pion mass to provide a ` native ' description of jets as hadron fragments from a moving source .
hard components of p - p @xmath4 spectra _ vs _
@xmath13 plotted on transverse rapidity @xmath1 ; low-@xmath5 parton and string fragment distributions on @xmath8 and angular difference variables @xmath11.,title="fig : " ] hard components of p - p @xmath4 spectra _ vs _
@xmath13 plotted on transverse rapidity @xmath1 ; low-@xmath5 parton and string fragment distributions on @xmath8 and angular difference variables @xmath11.,title="fig : " ] hard components of p - p @xmath4 spectra _ vs _
@xmath13 plotted on transverse rapidity @xmath1 ; low-@xmath5 parton and string fragment distributions on @xmath8 and angular difference variables @xmath11.,title="fig : " ] hard components of p - p @xmath4 spectra _ vs _
@xmath13 plotted on transverse rapidity @xmath1 ; low-@xmath5 parton and string fragment distributions on @xmath8 and angular difference variables @xmath11.,title="fig : " ] that novel single - particle result motivated a follow - up study of two - particle correlations on transverse rapidity @xcite . the minimum - bias distribution in fig .
[ lowq2 ] ( second panel ) represents all event multiplicities and charge combinations within the star @xmath3 detector acceptance .
separate soft and hard components are evident , as is the correspondence with hard - component gaussians in the first panel .
the @xmath8 space can thus be used as a cut space to study trends of corresponding angular correlations on pseudorapidity @xmath16 and azimuth @xmath17 for soft ( string fragment ) and hard ( parton fragment ) components , as shown in the third and fourth panels of fig .
[ lowq2 ] .
a sequence of several analysis steps has led from study of single - particle @xmath4 spectra in a two - component context to isolation of soft and hard spectrum components to corresponding two - particle correlations on transverse rapidity to jet - like angular correlations obtained without imposition of a jet hypothesis .
we thus achieve a model - independent analysis of parton scattering and fragmentation .
angular _ autocorrelations _ are not dependent on a leading or trigger particle and provide unprecedented access to low-@xmath5 partons . the lower limit on fragment @xmath4
is determined by the fragmentation process itself , not the analysis method .
we observe jet correlations in p - p collisions down to @xmath4 = 0.35 gev / c for both particles .
the main subject of this paper is the properties of fragment distributions on rapidity and angle from low-@xmath5 partons in p - p collisions .
correlation measure @xmath18 is closely related to pearson s normalized covariance or _ correlation coefficient _ @xcite . for event - wise particle counts @xmath19 and @xmath20 in histogram bins @xmath21 and @xmath22 on single - particle space @xmath23 and pair counts in corresponding bin @xmath24 on two - particle space @xmath25
we obtain pearson s coefficient @xmath26 averaged over an event ensemble .
pearson s coefficient is approximated by density ratio @xmath27 , where @xmath28 is the histogram bin size on @xmath23 and poisson values of the number variances in the denominator have been substituted .
@xmath18 , estimating the density of correlated pairs _ per particle _ , is our basic correlation measure for two - particle distributions on @xmath8 , @xmath29 and @xmath30 . to view angular correlations compactly
we combine spaces @xmath29 and @xmath30 by means of a _ joint autocorrelation _ , providing projection ( _ by averaging _ ) to a lower - dimensional space with essentially no information loss .
an autocorrelation for 1d primary space @xmath23 is constructed by averaging coefficients @xmath31 over index @xmath21 along the @xmath32 diagonal on @xmath25 . for angle space @xmath3
we average simultaneously along diagonals on @xmath29 and @xmath30 to obtain a _ joint _ autocorrelation , a projection by averaging of the full two - particle angle space onto its _ difference variables _ @xmath33 and @xmath34 .
data for this analysis were selected from 12 million p - p minimum - bias collision events at @xmath35 gev obtained with the star detector at rhic @xcite .
non - single - diffractive collisions were defined by a trigger coincidence between beam - beam counters positioned symmetrically in @xmath36 @xcite .
particle tracks were reconstructed with the star time projection chamber ( tpc ) operating within a uniform 0.5 t magnetic field parallel to the beam axis .
tracks were accepted within full azimuth , @xmath37 and @xmath38 gev / c .
tpc tracks and transverse beam position were used to reconstruct the primary or collision vertex .
use of transverse rapidity @xmath1 to study parton fragmentation insures more compact distributions and relates directly to fundamental qcd aspects such as the double log approximation .
it also provides a common basis , with longitudinal rapidity , for comparing string and parton fragmentation .
particle pairs from nuclear collisions can be separated on azimuth difference variable @xmath34 into _ same - side _ ( ss ) ( @xmath39 ) and _ away - side _ ( as ) ( @xmath40 ) pairs . fig .
[ ytyt ] shows same - side ( left two panels ) and away - side ( right two panels ) correlations of the form @xmath18 ( normalized covariance ) distributed on transverse rapidity subspace @xmath8 .
each pair of panels represents like - sign ( ls ) and unlike - sign ( us ) charge combinations ( left and right respectively ) .
structure can be separated into a soft component ( @xmath41 for both pair partners ) and a hard component ( @xmath42 for both partners ) .
both components are strongly dependent on charge - sign combination and azimuth opening angle ( same - side or away - side pairs ) .
description of the correlation structure is simplest in terms of sum and difference variables @xmath43 and @xmath44 .
number correlations on transverse rapidity @xmath45 from string and parton fragments for same - side ( left ) and away - side ( right ) pairs , and for like - sign and unlike - sign combinations.,title="fig : " ] number correlations on transverse rapidity @xmath45 from string and parton fragments for same - side ( left ) and away - side ( right ) pairs , and for like - sign and unlike - sign combinations.,title="fig : " ] number correlations on transverse rapidity @xmath45 from string and parton fragments for same - side ( left ) and away - side ( right ) pairs , and for like - sign and unlike - sign combinations.,title="fig : " ] number correlations on transverse rapidity @xmath45 from string and parton fragments for same - side ( left ) and away - side ( right ) pairs , and for like - sign and unlike - sign combinations.,title="fig : " ] in the left ( ss ) panels , the ls soft component is interpreted as quantum correlations ( hbt ) . the ls hard component along the diagonal is small and may itself be dominated by quantum correlations ( from parton fragmentation ) .
the us hard component is a peak at @xmath46 ( @xmath47 1 gev / c ) elongated along @xmath48 .
the us hard component runs continuously into the us soft component at lower @xmath1 , which is suppressed due to transverse momentum conservation . in the right ( as ) panels , the hard - component peaks for ls and us pairs are nearly symmetric about their centers and much broader on @xmath49 than their ss us counterpart , with rapid falloff below @xmath50 ( hadron @xmath51 gev / c ) and nearly equal amplitudes .
the large us soft component represents longitudinal string fragmentation , subject to local momentum and charge conservation ( charge ordering on @xmath52 ) @xcite and therefore contributing negligibly to as ls correlations .
the @xmath8 correlations in the previous section are directly related to angular correlations on @xmath3 . to isolate soft and hard components of p - p angular correlations we define soft pairs by @xmath41 ( @xmath53 gev / c ) and hard pairs by @xmath42 for each particle . fig .
[ etaphi ] shows minimum - bias correlations ( all pairs for all multiplicity classes ) of the form @xmath18 on @xmath11 for the soft component ( left two panels ) and hard component ( right two panels ) , with charge combinations like - sign ( ls ) and unlike - sign ( us ) ( left and right respectively ) .
the first panel is dominated by a 2d gaussian peak at the origin representing quantum correlations .
the us combination in the next panel is dominated by a 1d gaussian peak on @xmath54 arising from local charge conservation during string fragmentation ( producing charge ordering along the collision axis ) .
that trend is suppressed near the origin ( the depression on azimuth of the @xmath54 gaussian ) due to local transverse - momentum conservation .
the narrow peak at the origin represents electron - positron pairs from photon conversions .
except for the hbt contribution all structure in the soft component is apparently a consequence of local measure conservation during string fragmentation .
the resulting ` canonical suppression ' is dependent on local particle density ( event multiplicity ) .
number correlations on @xmath11 for soft - component ( left panels ) and hard - component ( right panels ) pairs , and for like - sign and unlike - sign charge combinations.,title="fig : " ] number correlations on @xmath11 for soft - component ( left panels ) and hard - component ( right panels ) pairs , and for like - sign and unlike - sign charge combinations.,title="fig : " ] number correlations on @xmath11 for soft - component ( left panels ) and hard - component ( right panels ) pairs , and for like - sign and unlike - sign charge combinations.,title="fig : " ] number correlations on @xmath11 for soft - component ( left panels ) and hard - component ( right panels ) pairs , and for like - sign and unlike - sign charge combinations.,title="fig : " ] the hard - component correlations in the right panels consist of a same - side peak at the origin and an away - side ridge .
the ls same - side peak may be due to quantum correlations rather than jet fragmentation _ per se_.
the us same - side peak represents angular correlations of parton fragments ( jet cone ) .
the away - side hard - component correlations for ls and us pairs are essentially identical in shape and amplitude and reflect momentum conservation between scattered partons ( dijets ) , including lack of structure on @xmath54 due to the broad distribution of parton - collision centers of momentum .
these hard - component @xmath3 systematics , fully consistent with conventional expectations for high-@xmath4 jet angular correlations , are observed in this study for pairs of particles with _ both _ @xmath4s as low as 0.35 gev / c ( @xmath55 ) , _ much lower than previously observed with leading - particle methods_. in what follows we emphasize low-@xmath5 parton fragmentation .
we first consider parton fragments distributed on transverse momentum . a _ fragmentation function
_ describes the _ conditional _ distribution of fragment momenta along a jet thrust axis given the parton momentum ( as estimated directly from collision kinematics or from the observed fragment distribution or jet ) .
hadron momentum is conventionally normalized by the parton momentum / energy ( _ e.g. , _
@xmath56 ) , represented by @xmath57 ( reconstructed jets ) , @xmath58 ( e@xmath59-e@xmath60 ) or @xmath61 ( e - p deep - inelastic scattering or dis ) .
alternatively , relative fragment momentum is measured logarithmically by @xmath62 .
dis fragmentation functions on a semi - log format , @xmath63 ( soft ) and @xmath64 ( hard ) components of the p - p @xmath4 spectrum on a similar format , fragmentation functions on logarithmic momentum variable @xmath2 and p - p @xmath4 spectrum hard - component on transverse rapidity @xmath1.,title="fig : " ] dis fragmentation functions on a semi - log format , @xmath63 ( soft ) and @xmath64 ( hard ) components of the p - p @xmath4 spectrum on a similar format , fragmentation functions on logarithmic momentum variable @xmath2 and p - p @xmath4 spectrum hard - component on transverse rapidity @xmath1.,title="fig : " ] dis fragmentation functions on a semi - log format , @xmath63 ( soft ) and @xmath64 ( hard ) components of the p - p @xmath4 spectrum on a similar format , fragmentation functions on logarithmic momentum variable @xmath2 and p - p @xmath4 spectrum hard - component on transverse rapidity @xmath1.,title="fig : " ] dis fragmentation functions on a semi - log format , @xmath63 ( soft ) and @xmath64 ( hard ) components of the p - p @xmath4 spectrum on a similar format , fragmentation functions on logarithmic momentum variable @xmath2 and p - p @xmath4 spectrum hard - component on transverse rapidity @xmath1.,title="fig : " ] one semilog plotting convention is a logarithmic frequency distribution on linear transverse momentum , or normalized ratio @xmath65 , as shown in fig .
[ linlog ] ( first panel ) from e - p deep - inelastic scattering @xcite .
those distributions are similar to the hard component @xmath64 of the p - p single - particle @xmath4 spectrum when plotted on @xmath4 shown as the dashed curve in fig .
[ linlog ] ( second panel ) .
the fragmentation function plotted on linear @xmath4 is _ approximately _ exponential .
an alternative convention is a linear frequency distribution on logarithmic transverse - momentum variable @xmath66 as shown in fig .
[ linlog ] ( third panel ) @xcite . those fragmentation functions ( for two @xmath5 intervals )
can be compared with the hard components obtained from rhic p - p spectra plotted on transverse rapidity in the fourth panel ( the sense of increasing @xmath4 is reversed in the two cases ) .
the conditional ( on parton @xmath61 ) fragmentation functions on @xmath2 in the third panel are approximately gaussian ( further discussed below ) , as are the minimum - bias fragment distributions ( no parton condition ) in the fourth panel .
the comparisons above motivated us to replot measured fragmentation functions on transverse rapidity @xmath1 as an _ infrared safe _ logarithmic momentum variable . in fig . [ fragonyt ]
( first two panels ) we have replotted fragmentation functions from several collisions systems @xcite on transverse rapidity @xmath67 , with @xmath15 assigned as the pion mass for all particles . we find that fragmentation functions so plotted are well described by beta distribution @xmath68 , with beta function @xmath69 .
the _ standard _ beta distribution defined on @xmath70 $ ] is rescaled to @xmath71 on @xmath72 $ ] .
we observe that @xmath73 is 0.5 ( @xmath74 gev / c ) for e - e collisions and 1.5 ( @xmath75 gev / c ) for p - p collisions , the latter endpoint presumably reflecting the presence of the spectator parton system in the p - p case .
the parton momentum is represented by @xmath76 , with @xmath77 , @xmath58 or @xmath61 depending on the collision system .
@xmath78 values for each fragmentation function in fig .
[ fragonyt ] are marked with short vertical lines . for e - e collisions
the beta parameters are @xmath79 and @xmath80 for the cases examined ; for p - p collisions the values are @xmath81 and @xmath82 .
the hard component from the star two - component analysis of p - p collisions is included in the first panel as the small gaussian curve ( describing _ minimum - bias _ parton fragments with no @xmath78 condition ) .
the energies in that panel are _ dijet _ energies @xmath83 .
the data in the second panel are also plotted on @xmath84 ( comparable to @xmath2 but with the sense of increasing @xmath4 reversed ) in the third panel
. the solid curves are mlla predictions @xcite , which reflect momentum conservation and qcd branching above the peaks but deviate strongly from data at lower momenta .
in contrast , the beta distribution provides a precise description of the entire fragment distributions on transverse rapidity in the left panels .
fragmentation functions from cdf , opal and tasso plotted on transverse rapidity @xmath1 and momentum as @xmath84 ; lep jet fragment multiplicities plotted on parton rapidity @xmath78 . ] fragmentation functions from cdf , opal and tasso plotted on transverse rapidity @xmath1 and momentum as @xmath84 ; lep jet fragment multiplicities plotted on parton rapidity @xmath78 .
, title="fig : " ] fragmentation functions from cdf , opal and tasso plotted on transverse rapidity @xmath1 and momentum as @xmath84 ; lep jet fragment multiplicities plotted on parton rapidity @xmath78 .
, title="fig : " ] the distribution systematics for e - e collisions in the second panel , interpreted in the context of the double - log approximation ( dla ) , imply that the e - e jet multiplicity trend can be described by @xmath85 , whereas the modified leading - log approximation ( mlla ) prediction is @xmath86 @xcite .
the fourth panel shows measured jet multiplicities _ vs _ jet energy ( @xmath78 ) for e - e collisions ( points ) @xcite , our parameterization with @xmath87 ( solid curve ) and the mlla prediction with @xmath21 chosen to provide the best fit ( dashed curve ) .
the hatched region corresponds to the low-@xmath5 partons which dominate a minimum - bias p - p analysis . by a similar argument
we can represent the most probable point ( mode ) as @xmath88 , where for the beta distribution and e - e collisions @xmath89 , whereas the mlla gives @xmath90 .
fragment distributions on transverse rapidity are consistent with conventional treatments ( _ e.g. , _ @xmath2 ) but are ` infrared safe ' as @xmath91 .
the conventional asymmetric treatment of parton fragments in terms of trigger and associated particles can not access the low-@xmath5 partons of greatest interest to us .
we therefore symmetrize the analysis of fragment correlations , first for two - particle fragment distributions on @xmath1 and then for angular correlations . the same - side unlike - sign ( us ) @xmath45 correlations in sec . [ ytytsect ]
can be interpreted as a two - particle jet fragment distribution which we now model : we combine what is known about single - particle fragmentation functions with expectations for two - particle correlations to sketch the parameterization of a two - particle fragment distribution . in the previous section we presented single - particle fragmentation functions plotted on transverse rapidity @xmath1 .
an example of fragmentation functions plotted on @xmath2 from the h1 experiment at hera is shown in fig .
[ linlog ] ( third panel ) for two intervals on @xmath61 . based on those trends we sketch a 2d distribution relating parton momentum @xmath61 and fragment momentum @xmath4 , both in rapidity format , as shown in the first two panels of fig . [ consym ] .
the first panel shows the locus of modes ( most probable points , dashed line with smaller slope ) for a continuum of fragmentation functions varying with parton rapidity on the horizontal axis , with fragment rapidity on the vertical axis . to construct the two - particle distribution we require a curve smoothly joining that line of modes and the diagonal with unit slope ( solid line ) given by the solid curve .
the plot is then symmetrized . in the second panel we sketch the corresponding 2d distribution of fragmentation functions _ vs _ parton rapidity @xmath78 .
the diagonal dotted line represents parton momentum @xmath61 plotted as a rapidity on the fragment axis .
the dashed line is the line of modes .
the fragmentation function for a given parton momentum would be a vertical slice from this distribution . locus of most - probable points for pair fragment distribution , fragment rapidity vs parton rapidity , symmetrized distribution , measured same - side unlike - sign fragment distribution . ]
locus of most - probable points for pair fragment distribution , fragment rapidity vs parton rapidity , symmetrized distribution , measured same - side unlike - sign fragment distribution . ] locus of most - probable points for pair fragment distribution , fragment rapidity vs parton rapidity , symmetrized distribution , measured same - side unlike - sign fragment distribution . ] in the third panel we symmetrize the distribution from the second panel because in our @xmath45 analysis we correlate fragments with fragments symmetrically , rather than fragments with partons .
that sketch can be compared with data in the fourth plot for unlike - sign same - side pairs the minimum - bias two - particle fragment distribution . for low-@xmath5 partons , which dominate the minimum - bias parton distribution ,
the two - particle fragment distribution becomes symmetric about the sum diagonal ( and the fragment number @xmath92 ) .
this parton - fragment distribution is biased by the requirement of fragment _ pairs _ : ` fragmentation ' to single hadrons can not contribute to this plot .
because of the symmetry , the more appropriate variables to describe the distribution are sum and difference rapidities @xmath93 and @xmath94 as illustrated in the first panel .
the comparison between sketch and data is informative .
the small excess yield at low @xmath1 in the fourth panel is string fragments , partially suppressed by transverse momentum conservation .
in the previous section we described parton fragment correlations on transverse rapidity . jet structure
is also characterized by two - particle angular correlations of hadron fragments , both intra - jet ( within one jet ) and inter - jet ( between opposing dijets ) . the conventional method
to describe angular correlations of fragments in the absence of full jet reconstruction is as a _ conditional _ distribution relative to a leading or trigger particle ( _ e.g. , _ the highest-@xmath4 particle in an event ) used as an estimator for the parton momentum .
the distribution of ` associated ' particles relative to the trigger particle is an approximation to full jet reconstruction .
angular correlations relative to the trigger reveal 2d same - side peaks with ( possibly different ) widths on @xmath16 and @xmath17 . from those widths the mean momenta of fragments transverse to the jet _ thrust axis _ in the two angular directions
@xmath3 are inferred . in this minimum - bias study
we encounter fragment distributions with a most - probable @xmath4 of about 1 gev / c .
the most probable fragment multiplicity is 2 , with approximately equal fragment momenta , and the parton @xmath61 is thus somewhat more than 2 gev_comparable _ to the intrinsic parton @xmath95 . for jets from low-@xmath5 partons we can not differentiate trigger and associated particles .
we want a symmetrized formulation of angular correlations relative to which the leading - particle conditional treatment is a _
special case_. for _ intra - jet _ correlations we define a symmetrized expression relating angular correlation width and transverse momentum @xmath96 relative to thrust as [ jt ] _ 12 \ { ^2 _ ss + 2 ^2 _ 1 ^2
_ 2 } + \ { ^2 ( _ / ) _ ss + 2 ^2 ( _ / 2 ) ^2 } for @xmath17 , with a similar expression for @xmath16 .
the inter - jet azimuth angular correlation of particle pairs from opposed jets ( dijets ) is wider than the intra - jet azimuth correlation .
the excess width or ` @xmath95 broadening ' is attributed to the initial ( intrinsic ) transverse momenta @xmath95 of partons prior to scattering , which produces an _ acoplanarity _ of the two jets with respect to the collision axis . the symmetrized relation between angular correlation width and
inferred @xmath95 is [ kt ] _ 12 \ { } + \ { } .
those symmetrized expressions are used to extract fragment transverse momenta relative to thrust axis from angular correlations from p - p collisions .
fragment angular correlations are thus obtained for previously inaccessible low-@xmath5 partons .
fig . [ jtkt ]
( first panel ) shows two - particle correlations on transverse - rapidity space @xmath8 described on sum and difference variables @xmath43 and @xmath44 .
hard - component fractions for angular correlation measurements are defined by the grid of rectangles or bins ( the narrow yellow lines ) along @xmath48 numbered @xmath97 .
the solid green boxes in the upper - right corner represent regions typically explored in _ leading - particle _ analyses based on high-@xmath4 ` trigger ' particles @xcite .
the dashed extensions represent cuts for extended associated - particle conditions recently applied to heavy ion collisions @xcite .
the first two panels of fig .
[ angcorr ] show angular autocorrelations for the second and eleventh bins on @xmath48 in fig .
[ jtkt ] as examples . in the first panel
the most probable combination is two particles each with @xmath98 gev / c .
the same - side peak ( jet cone ) is broad but well - defined .
the away - side ridge ( particle pairs from dijets ) is uniform on @xmath54 as expected .
this jet correlation is a remarkable result for particles with such low @xmath4 and illustrates the power of the autocorrelation technique .
the second panel shows correlations for the 11@xmath99 bin , corresponding to @xmath100 gev / c for each particle . the same - side cone is much narrower , as is the away - side peak on azimuth @xmath17 , a result more typical of a high-@xmath4 leading - particle analysis .
angular correlations for parton fragment pairs with smaller and larger geometric - mean @xmath4 ; measured angular widths of same - side peaks on @xmath16 and @xmath17 _ vs _ inverse mean @xmath4.,title="fig : " ] angular correlations for parton fragment pairs with smaller and larger geometric - mean @xmath4 ; measured angular widths of same - side peaks on @xmath16 and @xmath17 _ vs _ inverse mean @xmath4.,title="fig : " ] angular correlations for parton fragment pairs with smaller and larger geometric - mean @xmath4 ; measured angular widths of same - side peaks on @xmath16 and @xmath17 _ vs _ inverse mean @xmath4 . ]
the general trend is monotonic reduction of peak widths on @xmath11 with increasing @xmath48 .
the angular widths are determined by the mean fragment momentum @xmath101 perpendicular to the parton momentum ( jet thrust ) axis , relative to the particle @xmath4 in the form @xmath102 .
we extract from those angular correlations near - side widths on @xmath54 and @xmath103 , and an away - side width on @xmath103 .
we fit the same - side ( @xmath39 ) angular correlations ( jet cone ) with a 2d gaussian and extract widths on pseudorapidity and azimuth for like - sign ( ls ) , unlike - sign ( us ) and total ( ci ) pairs .
those widths are plotted in the last two panels of fig .
[ angcorr ] _ vs _ @xmath104 , that is , proportional to the inverse geometric mean of the pair momenta .
the corresponding mean particle @xmath4 is given at the top of the third panel .
the dashed diagonal lines in those panels represent ` @xmath101 scaling ' : the assumption that @xmath101 ( _ r.m.s . _ @xmath96 ) is independent of hadron momentum as a limiting case for larger @xmath4 .
we observe a dramatic departure from that expectation for these low fragment momenta ( @xmath105 gev / c ) .
however , with the autocorrelation technique we observe that jet structure is very well defined , even down to hadron pairs with particle @xmath106 gev / c .
the hatched regions to the left in those panels show the values measured for minimum - bias hard - component hadrons , which are weighted toward lower @xmath4 . in the following section
we examine the systematics of @xmath101 _ vs _ @xmath48 separately on @xmath16 and @xmath17 , and for values of parton @xmath5 previously inaccessible .
by applying eq . ( [ jt ] ) with the values of @xmath108 defined on the @xmath8 cut space as in fig . [ consym ] ( first panel )
we obtain @xmath101 values for @xmath109 from the corresponding angular widths . in fig .
[ jtkt ] ( second and third panels ) we observe a strong decrease in both @xmath101 components with decreasing @xmath48 ( @xmath110 parton @xmath5 ) .
we also observe substantial differences in the @xmath101 @xmath16 and @xmath17 components , converging however to the ` @xmath101 scaling ' expectation at larger @xmath48 ( hatched bands ) .
it is notable that the orientation of the @xmath101 asymmetry ( major axis lying in the @xmath111 plane ) is parallel to the ` contact plane ' of the parton collision .
that asymmetry orientation is orthogonal to what is observed for low-@xmath5 jet fragments in central heavy ion collisions @xcite .
minimum - bias fragment distribution on @xmath8 showing cut grid along the diagonal used for the angular correlation study ; @xmath101 and @xmath107 values inferred from angular correlation study . ]
minimum - bias fragment distribution on @xmath8 showing cut grid along the diagonal used for the angular correlation study ; @xmath101 and @xmath107 values inferred from angular correlation study . ]
minimum - bias fragment distribution on @xmath8 showing cut grid along the diagonal used for the angular correlation study ; @xmath101 and @xmath107 values inferred from angular correlation study . ]
we also measured the angular width of the _ away - side _ azimuth correlation of hadron pairs ( @xmath112 ) and compared it with the same - side azimuth correlation to infer the mean transverse momentum @xmath107 of partons prior to scattering .
we converted the angular widths to @xmath107 values according to eq .
( [ kt ] ) , where @xmath52 estimates the mean value of the ratio of ` trigger particle ' momentum to parton momentum . at larger @xmath4
that value is approximately 0.75 , but is a falling function of @xmath4 for smaller fragment @xmath4 , as in this case .
below @xmath113 ( hadron @xmath114 gev / c ) we observe that @xmath107 values fall below the @xmath95-scaling trend ( hatched band ) by an increasing amount with smaller @xmath48 ( smaller @xmath5 ) .
the trend of pythia @xmath107 values ( dotted line ) is substantially displaced below the data .
we have presented a broad survey of newly - obtained two - particle correlations from 200 gev p - p collisions at rhic .
correlations from longitudinal string fragmentation and transverse scattered parton fragmentation are clearly distinguished .
the former is dominated by local momentum and charge conservation .
the latter provides new access to scattering and fragmentation of _ minimum - bias _ partons . using newly - devised analysis techniques we find that low-@xmath5 parton fragmentation is accessible down to hadron @xmath4 = 0.35 gev / c for both hadrons of a correlated pair .
jet morphology for low-@xmath5 partons requires a more general treatment of fragment @xmath4 distributions and angular correlations .
fragment distributions on transverse rapidity @xmath1 are ` infrared safe ' and exhibit interesting systematic behaviors which can be compared with the mlla .
jet angular correlations show a dramatic asymmetry about the thrust axis at low-@xmath5 ( up to 2:1 aspect ratio , with larger width in the azimuth direction ) , possibly related to nonperturbative details of soft parton collisions .
these measurements of p - p correlations provide an important reference for the study of _ in - medium modification _ of parton scattering and fragmentation in heavy ion collisions .
mueller a h 2000 _ nucl .
_ b * 572 * 227 whitmore j 1976 _ phys .
* 27 * 187 + kafka t et al 1977 _ phys .
_ d * 16 * 1261 andersson b , gustafson g , ingelman g and sjstrand t 1983 _ phys . rep . _ * 97 * 31 angelis a l s et al ( ccor ) 1980 _ phys .
_ b * 97 * 163 + albajar c et al ( ua1 ) 1988 _ nucl .
_ b * 309 * 405 giovannini a and ugoccioni r 1999 _ phys .
_ d * 59 * 094020 + dias de deus j and ugoccioni r 2000 _ phys .
_ b * 494 * 53 kharzeev d and nardi m 2001 _ phys .
_ b * 507 * 121 porter r j and trainor t a ( star collaboration ) 2005 _ acta phys .
_ b * 36 * 353 | we describe a comprehensive survey of two - particle correlations from p - p collisions at @xmath0 gev .
correlations on transverse rapidity @xmath1 are separated into a soft component ( string fragments ) and a hard component ( parton fragments ) .
fragment correlations on transverse rapidity @xmath1 are compared to fragmentation functions on logarithmic momentum variable @xmath2 .
cuts on transverse rapidity isolate angular correlations on @xmath3 for the two components .
hard - component angular correlations are compared to results from conventional leading - particle jet analysis .
angular correlations for low-@xmath4 fragments , corresponding to low-@xmath5 parton scattering , reveal a large asymmetry about the jet thrust axis . . |
large hadron collider ( lhc ) collider at cern is expected to start data taking in 2009 with the luminosity ramp up to the design goal of 10@xmath2/@xmath3/s over the next few years .
after 2016 a proposal to extend the physics potential of the lhc with a major luminosity upgrade to the super lhc ( slhc ) has been endorsed by the cern council strategy group @xcite . the goal is a factor ten increase in luminosity .
this imposes severe requirements on the silicon tracking devices , which need to be completely redesigned in order to cope with the order of magnitude increase in occupancy and radiation damage .
the increase in occupancy requires a corresponding increase in granularity , while the radiation hardness can be improved by a different sensor design combined with cooling to lower temperatures in order to reduce the leakage current .
radiation damage to silicon sensors creates defects with energies between the valence and conduction band .
this increases both the leakage current at a given temperature and the depletion voltage .
if one uses n - strips on p - type wafers the depletion will start from the n - strips , so there will be no undepleted layer between the sensor and the strips .
furthermore , it was shown that even after 10@xmath4 n/@xmath3 , which is the fluence expected at the slhc at the innermost layers during its lifetime , the collected charge is still around 7000 electrons , which is roughly one quarter of the non - irradiated sensor @xcite . if one reduces the noise both , by cooling the sensor and by reducing the sensor capacitance ( by reducing the strip length ) it seems feasible to get a similar signal / noise ratio as for non - irradiated samples .
shorter strips ( `` strixels '' ) are anyway required to reduce the channel occupancy for the high luminosity of the slhc .
another requirement is the reduction of the material budget in order to reduce the interactions in the tracker , which lead to a distortion of the momentum measurements and an increase of the multiplicity by photon conversions .
one expects up to 400 interactions per bunch crossing at the slhc with typically 1000 charged tracks per bunch crossing .
if most photons convert into electron - positron pairs , as they will do with the present material budget , the charged multiplicity will be doubled to tripled .
have been indicated as well . , width=336 ] originates from the service connections for cooling and power to the inner barrel and disk detectors , which pass in front of the endcap detectors @xcite .
this can be avoided by having a long barrel ( see fig .
[ f3 ] ) with all service connections outside the tracking volume .
this leads to a material budget indicated by the light shaded ( yellow ) area.,width=336,height=336 ] here we present a tracker design based on co@xmath1 two - phase cooling .
co@xmath1 allows for long cooling pipes with still negligible temperature gradients , because of the small pressure drop ( owing to the large latent heat and correspondingly small mass flow and the low viscosity ) .
co@xmath1 cooling is efficient between 20 and -50 @xmath0c , so the leakage currents can be made negligible , while the long cooling pipes allow the use of 3 m long ladder type detectors , thus paving the way for having all service connections outside the tracking volume . by combining the functions of cooling pipe , mechanical support and
current leads the material budget can be additionally minimized .
first tests with a simple co@xmath1 blow system prove the feasibility of the approach .
a traditional layout of a silicon tracker has a barrel with endcaps and a pixelated detector near the beam pipe , as shown in fig .
[ f1 ] using the 200 m@xmath5 silicon tracker of cms as an example @xcite .
the main reason for the endcaps is that in the forward direction the particles traverse horizontal detectors under a small angle @xmath6 , thus increasing the material thickness by a factor @xmath7
. therefore vertical detectors are much better with respect to the material budget for small angles @xmath6 , since in this case the traversed material is @xmath8 . however , in this case the cooling pipes , current leads and endflanges of the horizontal detectors are just in front of the endcaps , which increases the material budget in the forward region by as much as 0.8 @xmath9 , as shown in fig .
[ f2 ] for the present cms silicon tracker . to avoid
this would require long barrel detectors , which seemed difficult to cool .
however , with co@xmath1 two - phase cooling this is possible and therefore it seems better to use only horizontal detectors , as shown in fig .
[ f3 ] . in this case
all service connections and interconnect boards for the read out can be at the end of the barrel , i.e. outside the tracking volume . also optocouplers on the interconnect boards would be further away from the center , which is important , since these are known to suffer severe radiation damage after a fluence of @xmath10 .
furthermore , they consume a significant amount of power , which need not to be entered into the tracking volume .
all signals from the hybrids have then to be transferred to the interconnect board , e.g. by lvds signals via aluminized kapton cables .
an additional reduction of the material budget can be obtained by using the cooling pipes as power lines , as sketched in fig .
details will be discussed in the next section . to get enough tracking points for tracks in the forward direction
requires then to have disks at small radii . a disk layer for radii between 5 and 35 cm
is shown in fig .
the rings can be constructed in the same way as the ladders of the barrel , i.e. with cooling pipes at each side which act at the same time as power lines .
the only difference is that the cooling pipes are now bent in half circles with the connections on the inside , i.e. outside the tracking volume .
half circular disks are are needed to be able to install or replace the detectors with the beampipe installed .
the four inner horizontal layers and four vertical disks are pixelated , most likely similar to the present pixel detectors .
note that the pixel detectors as well as the strixel disks can be exchanged without moving the major part of the detector , the barrel detectors .
this is important , since for these detectors the radiation damage is much higher than for the barrel , so they may need replacement .
having all service connections outside the tracking volume reduces the material budget by 40% or more , as shown by the light shaded ( yellow ) area in fig .
note that in the center the total material budget is similar , around 0.4 x@xmath11 , as expected , since in this area there are no service connections and the number of layers is similar .
two - phase cooling system .
, width=336 ] the luminosity at the slhc will be an order of magnitude higher than at the lhc . requiring the same occupancy implies that the number of channels increases also by a factor between 5 to 10 , e.g. by having pixel detectors or shorter strips ( `` strixels '' ) .
the power will not increase by an order of magnitude , since the smaller feature size of today s and future electronics is expected to reduce the power per channel by a factor 5 or more@xcite . so the total power for the cms upgrade is expected to be of the order of 35 kw , similar to the present cms tracker .
if one opts for a long barrel of identical ladders , one needs only one type of sensors with eventually different pitches and strixel lengths for the inner and outer ladders in order to reduce the number of channels .
this simplifies the detector construction enormously in comparison with the present detector .
long ladders require efficient cooling .
we first discuss the module design integrated on a long ladder with cooling pipes as power lines integrated in the same mechanical structure and then the cooling system .
as mentioned in the introduction , the material budget can be reduced by long ladders having all service connections outside the tracking volume as well as combining the mechanical support , the cooling pipe , which acts simultaneously as power line , into a single mechanical structure .
a possible design is sketched in fig .
[ f4 ] . the pair of cooling pipes on each side of the detector ( inlet and outlet ) are most easily realized by extruded al tubes .
pure al , which is easy to extrude , has the additional advantage that the electrical resistivity drops by about 30% , if one cools the tube from + 40 to -40 @xmath0c . for a 2.5 m long pair of cooling pipes with
an i d of 1.5 mm and outer od of 3 mm the total resistance is about 5 m@xmath12 , so they form excellent power lines . in order to avoid a contact resistance to the hybrid , the al tubes have to be chromatized , which is a standard process .
the sensors are glued to a piece of al , shown in red , which serves simultaneously as power connector , the mechanical support and the cooling point .
the hybrids and the sensors are both screwed on the cooling pipes , but there is no further mechanical or thermal contact between them , so the hybrid and sensor are cooled independently and thus can operate at different temperatures .
thermal stresses are reduced , since the hybrid and sensor move together with the cooling pipes during cooldown . for the tests the hybrids consisted of a simple kapton printed circuit with a resistive line between the power pads and a pt100 thermometer , which can be read out via an usb interface .
the power pads of copper on the kapton are pressed against the cooling block by the mounting screw , which forms simultaneously the power connector . a picture is shown in fig .
[ f6 ] . in total 56 hybrids
are screwed between the cooling pipes .
of course , in such a scheme all modules are powered in parallel , which raises several questions .
first , what is the voltage drop along the power line .
the current needed for a single ladder is estimated as follows .
each sensor of 9x10 cm is assumed to have strixels of 2.2 cm with bond pads at the outer edge , as indicated in fig .
[ f7 ] . assuming a pitch of 130 @xmath13 leads to a total of 12x256 channels , i.e. 6 front end chips with 256 channels bonded to each side of the detector .
if we assume new front ends need 0.5 mw per channel @xcite , if the 0.13 @xmath13 or even smaller feature sizes will be used and add 25% for the control power , one needs 1.6 a at 1.2 v per sensor . for 56 hybrids this corresponds to about 50 a and 60 w per ladder . to get all connections at large z requires that the current is returned , e.g. via the cooling pipes of the neighboring ladder , as sketched in fig .
[ f8 ] . in this case
the current is used twice , thus reducing the total amount of current in the power lines between the power supplies and the detector .
the total current needed for such a design is of the order of 12 ka ( 33 kw ) for a total of 64 million channels .
this is to be compared with 15 ka and 33 kw in the present cms detector .
if the input current flows in both power lines in the same direction by returning the current in the neighboring ladder , one has the same voltage drop on each power line , so one naively expects no influence of the voltage drop on the hybrids , since each hybrid would see the same voltage difference between the power lines .
however , at the beginning of the ladder the positive power line carries 50 a and the negative only the current from the first hybrid , i.e. around 1 a , so the voltage drops are not equal in both power lines .
this can be improved by increasing the thickness of the positive power line at the beginning , as shown at the bottom of fig .
[ f6 ] by the additional piece of al between the power connectors ( = cooling points ) . in the middle of the ladder
both power lines carry a similar current , so here nothing needs to be done , but at the end the returning power line needs to have additional material , since here the full current flows . by this voltage compensation in the power lines all hybrids can have the same voltage with a precision better than 20 mv .
the hybrids have not exactly a common ground because of the voltage drop of about 0.1 v along the power lines , so the control and signal lines better be differential lvds type lines with a standard ac coupling to the interconnect boards at the end of the ladder .
alternatively , a differential lvds receiver with a somewhat higher source voltage can easily cope with common mode offsets of 100 mv .
the detectors have been mounted on a 5 mm thick support plate made from rohacell foam , a strong lightweight material , which has been used before as support in silicon detectors @xcite . with the attached cooling structure this forms a stable double layer sandwich structure which can easily take up the magnetic forces of the power lines , if the ladder is inside a solenoid .
the radiation length of rohacell is above 5 m , so it does not contribute significantly to the material budget .
however , it is hygroscopic .
water absorption can be prevented by a parylene coating @xcite .
alternatively , the support structure can be made out of carbon fiber .
+ .,width=336 ]
the leakage current of heavily irradiated sensor can be reduced sufficiently by using sensor temperatures below -25 @xmath0c .
the high cooling power of co@xmath1 makes it a natural choice as refrigerant , especially since co@xmath1 becomes increasingly popular in replacing the climate unfriendly freons .
this leads to many off - the - shelf components , like co@xmath1 liquid pumps .
the temperature range for co@xmath1 2-phase cooling can be read off from the pressure - temperature diagram in fig .
[ f9 ] : -57 to + 31 @xmath0c for pressures between 5.2 and 73 bar , respectively .
this implies that during normal operations the pressure is below 20 bar , but after warm up to room temperature the pressure increases to around 50 bar at 20 @xmath0c
. showing that a two - phase liquid - gas system exists between -57 and 31 @xmath0c , so this is the range of cooling liquid temperatures with vapor pressures between 5 and 73 bar .
, width=336 ] what makes co@xmath1 so interesting for low mass cooling systems is its large enthalpy , which allows to cool close to 300 j for each g of co@xmath1 evaporated .
this is to be compared with non - evaporative systems , like cms , where only 5 j / g are obtained , if a temperature increase of 5 k is accepted
. therefore the flow rate of co@xmath1 liquid can be one to two orders lower compared to non - evaporative systems , thus reducing the size of tubing , pumps etc . by a similar amount .
in addition , the viscosity of co@xmath1 is low , so the pressure drop and the corresponding temperature drop in the cooling tubes is small . for evaporative systems one can either have a cooling plant with a gas compressor to liquify the evaporated gas or
pump the liquid around with a liquid pump instead of using the compressor to build up pressure .
both schemes have been used at the lhc , as shown in fig .
[ f10 ] for the atlas @xcite and lhcb @xcite cooling systems .
the atlas system is running with a low vapor pressure c3f8 coolant , while lhcb has opted for a high pressure co2 cooling system , in which the liquid is pumped around .
great features of the co@xmath1 cooling system are : i ) no active elements like heaters or electric valves inside detector ii ) standard liquid co@xmath1 pumps iii ) standard primary chiller iv ) the temperature of the whole system is controlled by only one parameter , namely the vapor pressure in the accumulator , which can be increased by the heater and decreased by the chiller . blow system .
the co@xmath1 bottle is precooled to -40 @xmath0c in a household freezer .
the co@xmath1 liquid is sent through 2 mm i d al cooling pipes of a 2.5 m ladder of hybrids , consisting of kapton with resistors and pt100 temperature sensors ( see fig .
the cooling pipes act at the same time as power lines and mechanical support .
bottom : photographs of the system .
the freezer is shown on the left .
the temperatures of different sensors as function of time are shown at the right .
the blue curve shows the temperature of a sensor just be cooled by the arriving liquid.,title="fig:",width=336 ] blow system .
the co@xmath1 bottle is precooled to -40 @xmath0c in a household freezer .
the co@xmath1 liquid is sent through 2 mm i d al cooling pipes of a 2.5 m ladder of hybrids , consisting of kapton with resistors and pt100 temperature sensors ( see fig .
the cooling pipes act at the same time as power lines and mechanical support . bottom : photographs of the system .
the freezer is shown on the left .
the temperatures of different sensors as function of time are shown at the right .
the blue curve shows the temperature of a sensor just be cooled by the arriving liquid.,title="fig:",width=336 ] another advantage of having a liquid pump is that the system works very well without heat load , e.g. when the detector is switched off .
liquid co@xmath1 membrane pumps are widely used in the food industry and elsewhere , so they are cheap and easy to obtain including sensors in case the membrane would be leaking @xcite .
large liquid pumps with capacities of around 100 m@xmath14/h are available , which are e.g. used on off - shore gas platforms to pressurize the natural gas with co@xmath1 . for the cooling power of an slhc tracker of about 50
kw pumps with a capacity of about 1 m@xmath14/h are enough .
of course , pumping the liquid requires thermally isolated outlets , but this is easily accomplished , especially since the temperature gradients are not large .
since co@xmath1 is non - toxic and non - flammable one can easily design small `` blow''-cooling - systems , in which the co@xmath1 is blown to the air .
the main precautions , which have to be taken , is that co@xmath1 gas has a higher density than air , so it will collect at the bottom .
therefore it has to be blown outside the window , if one wants to prevent the risk of not having enough oxygen in the lab .
a simple blow - system is shown in fig .
the co@xmath1 bottle is precooled in a commercial house - hold freezer to -40 @xmath0c , which reduces the pressure to around 10 bar . in this case
nylon tubing with standard swagelok connectors can be used to transport the liquid to and from the detector .
the nylon tubing is already a good isolator , so only little additional isolation is required . after the detector a simple ball flow meter , consisting of a little ball inside a flow tube with a needle valve at the entrance for regulating the flow rate , is installed .
the environmental heat into the nylon tube is enough to evaporate the rest of liquid from the detector before entering the flow meter .
the whole system is regulated by just two knobs : i ) the pressure reducer regulates the temperature of the liquid - gas mixture coming out of the bottle , which has a riser tube sticking into the liquid at the bottom ; ii ) the cooling power is regulated by the needle valve in the ball flow meter . the long ladder , shown in fig . [ f6 ] , with two pairs of 2x2.5 m long cooling tubes with an inner diameter ( i d ) of 2 mm and outer diameter ( od ) of 3 mm has been tested .
the temperature was set by the pressure reducer on the bottle , which kept the temperature on the whole ladder very well constant , as shown in fig .
[ f12 ] . increasing
the power in the hybrid resulted in evaporation of the liquid .
if the liquid had evaporated before the end of the ladder , the temperature at the end of the ladder increased , i.e. dry - out occurred .
increasing the flow rate restored the cooling everywhere .
[ f13 ] shows the power against the flow rate required to be just above the dry - out condition .
the minimum power estimated for a 2.5 m long ladder is 50 w , but clearly much higher powers can be sustained with 2 mm i d cooling tubes .
the maximum power was not limited by the detector system , but by the maximum throughput of the flow meters .
bottle.,width=336 ] at a temperature of -31 @xmath0c as function of the flow rate of co@xmath1 needed to prevent dry - out .
the total power @xmath15 is proportional to the flow rate @xmath16 , so the negative offset for @xmath17 represents environmental heat load @xmath18 of approximately 15 w. this is rather large , since the ladder was only inside a box with a transparent cover to prevent condensation from the air , but this yields minimal isolation.,width=336 ]
the successful operation of a co@xmath1 cooling system in lhcb proves its feasibility @xcite .
this paves the way for a rather different design of a large radiation hard silicon tracker , since co@xmath1 cooling has two advantages : it allows to cool easily to temperatures below -30 @xmath0 , thus eliminating the leakage currents in a harse radiation environment and secondly , the high cooling power allows for tiny and long cooling tubes , which in turn allows to have long ladder type detectors with all service connections and interconnect boards outside the tracking volume
. the number of layers for tracks at small angles with respect to the beampipe can be increased by a larger number of pixel layers and strixel disks at small radii . for pitches between 100 and 200 @xmath13 and strixels around 2 cm
the strixels can have their bonding pads all at the outside of a 9x10 cm sensor , so a similar module construction can be used as in present lhc detectors , except for having hybrids on each side of the detector .
although this increases the number of channels by a factor 5 to 10 , the total power is expected to be similar because of the expected power reduction in future front end electronics with a smaller feature size .
serial powering of two ladders is proposed , which would keep the total current similar to the ones in present detectors , thus allowing to reuse the existing services and power supplies without the need for dc - dc converters .
it is shown that the material budget can be reduced by 40% or more in the forward region of the cms detector , if one adopts the strawman design of long barrel detectors with inner disks at small radii ( instead of the traditional endcaps ) , especially if one combines the functions of cooling pipes , power lines and mechanical support in a single structure .
a simple co@xmath1 blow system has been designed and first tests regarding powering via cooling pipes are encouraging .
a final pair of ladders with real sensors and all interconnect boards at the end of a long ladder needs to be tested in order to check the stability of the proposed powering via cooling pipes and the transfer of signals to and from the interconnect board at the end of the long ladder .
special thanks go to bart verlaat for helpful discussions on co@xmath1 cooling .
the project was supported by the bundesministerium fr bildung und forschung under grant 05 hs6vk1
. 99 slhc strategy , see http://council-strategygroup.web.cern.ch/council-strategygroup/ g. casse , a. affolder and p. p. allport , ieee trans .
* 54 * ( 2007 ) 1695 .
r. adolphi _ et al . _
[ cms collaboration ] , jinst * 3 * ( 2008 ) s08004 m. raymond , http://indico.cern.ch/conferencedisplay.py?confid=22827 d. attree _ et al .
_ , jinst * 3 * ( 2008 ) p07003 .
b. verlaat , proceedings 22nd iir international congress of refrigeration , beijing , china , 21 - 26 aug .
icr07-b2 see e.g. lewa herbert ott gmbh , leonberg , germany , http://www.lewa.com/main/de/2_2_proc0/ e. bosze , j. simon - gillo , j. boissevain , j. chang and r. seto , nucl . instrum .
a * 400 * ( 1997 ) 224 , http://p25ext.lanl.gov / phenix / mvd/. | silicon trackers at the slhc will suffer high radiation damage from particles produced during the collisions , which leads to high leakage currents .
reducing these currents in the sensors requires efficient cooling to -30 @xmath0c .
the large heat of evaporation of co@xmath1 and the low viscosity allows for a two - phase cooling system with thin and long cooling pipes , because the small flow of liquid needed leads to negligible temperature drops . in order to reduce the material budget a system is proposed in which a large scale tracker requiring ca .
50 kw of power is powered via 1 - 2 mm diameter aluminum cooling pipes with a length of several m. these long cooling pipes allow to have all service connections outside the tracking volume , thus reducing the material budget significantly .
the whole system is designed to have negligible thermal stresses .
a co@xmath1 blow system has been designed and first tests show the feasibility of a barrel detector with long ladders and disks at small radii leading to an optimized design with respect to material budget and simplicity in construction . |
i am indebted to t. csrg for the helpful comments on the manuscript and to g. jancs for the constructive criticisms .
this work was supported by the hungarian science foundation under grant no .
otka - t024094/1997 .
the fox function @xmath74 is defined by the mellin - barnes type integral @xmath75\right.= { \sf h^{m , n}_{p,\;q } } \left[x\left| \begin{array}{c } ( a_1 , \alpha_1),\ldots,(a_p , \alpha_p ) \\ ( b_1 , \beta_1),\ldots,(b_q , \beta_q ) \end{array } \right]\right.= \frac{1}{2\pi i}\int_{\cal c}\,\frac{a(s)b(s)}{c(s)d(s)}\,x^s\,ds \eqnum{a.1}\ ] ] with @xmath76 and @xmath77 above @xmath78 are integers satisfying @xmath79 and @xmath80 . in the cases
@xmath81 , @xmath82 and @xmath83 ( a.2 ) has to be interpreted as @xmath84 , @xmath85 and @xmath86 , respectively . the parameters @xmath87 and @xmath88 are complex , whereas parameters @xmath89 and @xmath90 are positive numbers .
they are restricted by the condition that @xmath91 for @xmath92 ; @xmath93 ; @xmath94 .
the contour @xmath95 in the complex @xmath96 plane is such that the points @xmath97 and @xmath98 lie to the right and left of @xmath95 respectively while @xmath95 extends from @xmath99 to @xmath100 where @xmath101 is a constant with @xmath102 .
elementary properties of the fox function very useful to manipulate them are @xmath108\right.= { \sf h^{n , m}_{q,\;p}}\left[\,x\left| \begin{array}{c } ( 1-b_q,\ ; \beta_q ) \\ ( 1-a_p,\ ; \alpha_p ) \end{array } \right]\right . \eqnum{a.3}\ ] ] @xmath109\right.= { \sf h^{m , n}_{p,\;q}}\left[\,x\left| \begin{array}{c } ( a_p+c\,\alpha_p,\ ; \alpha_p ) \\ ( b_q+c\,\beta_q,\ ; \beta_q ) \end{array } \right]\right .
\eqnum{a.4}\ ] ] @xmath110\right.= \frac{1}{c}\,{\sf h^{m , n}_{p,\;q}}\left[\,x\left| \begin{array}{c } ( a_p,\ ; \alpha_p / c ) \\ ( b_q,\ ; \beta_q / c ) \end{array } \right]\right.\quad c>0 . \eqnum{a.5}\ ] ] many of the integral transforms of @xmath19 yield fox functions again with altered parameters . for example
, the laplace transform is given by the equation pair @xmath111\right.\quad\mbox{for $ 0\leq\kappa\leq1 $ } \eqnum{a.6}\ ] ] and @xmath112\right.\quad\mbox{for $ \kappa\geq1 $ } \eqnum{a.7}\ ] ] with scale parameter @xmath113 .
the mellin transform of @xmath19 reads @xmath114 an important special case of @xmath19 is the generalized hypergeometric function @xmath115 .
its relation to the fox function is given by @xmath116= { \sf h^{1,p}_{p , q+1 } } \left[-x\left| \begin{array}{c } ( 1-a_p,1 ) \\ ( 0,1),\;(1-b_q,1 ) \end{array } \right]\right.\nonumber\\=\ ; & & { \sf h^{p,1}_{q+1,p } } \left[-\frac{1}{x}\left| \begin{array}{c } ( 1,1),\;(b_q,1 ) \\ ( a_p,1 ) \end{array } \right]\right .
\quad\mbox{for $ p\leq q$ , or $ p = q+1 $ and $ |x|<1$. } \eqnum{a.9}\end{aligned}\ ] ] for more details on the h - function as well as for a rich collection of its particular cases the reader is referred to [ 9 ] .
z. koba , h.b .
nielsen and p. olesen , nucl .
* b40 * , 314 ( 1972 ) .
a. giovannini , nuovo cim . * 10a * , 713 ( 1972 ) .
p. carruthers and c.c .
shih , int .
phys . * a2 * , 1447 ( 1987 ) .
dremin , physics uspekhi * 37 * , 715 ( 1994 ) .
r. ugoccioni , a. giovannini and s. lupia , in proc .
symp . on multiparticle dynamics ,
a. giovannini , s. lupia and r. ugoccioni ( world scientific , 1995 ) .
n.l . johnson and s. kotz , distributions in statistics vol . 2 . , continuous univariate distributions ( wiley , 1970 ) . s. hegyi , phys
. lett . * b387 * , 642 ( 1996 ) . s. hegyi , phys . lett . *
b388 * , 837 ( 1996 ) .
a.m. mathai and r.k .
saxena , the h - function with applications in statistics and other disciplines + ( wiley eastern , 1978 ) .
f. oberhettinger , tables of mellin transforms ( springer , 1974 ) .
l. bondesson , scand .
j. 48 ( 1978 ) .
press , b.p .
flannery , s.a .
teukolsky and w.t .
vetterling , numerical recipes the art of scientific + computing ( cambridge univ . press , 1986 ) .
s. wolfram , mathematica a system for doing mathematics by computer ( addison - wesley , 1991 ) . c. bromberg et al .
* d15 * , 64 ( 1977 ) .
+ j. erwin et al .
* 32 * , 254 ( 1974 ) .
+ c. bromberg et al .
lett . * 31 * , 1563 ( 1973 ) .
+ d. brick et al .
, phys . rev .
* d25 * , 2794 ( 1982 ) .
brenner et al . , phys . rev . * d26 * , 1497 ( 1982 ) .
+ s. barish et al . , phys . rev .
* d9 * , 2689 ( 1974 ) .
+ m. adamus et al . , z. phys .
* c32 * , 475 ( 1986 ) .
dao et al . , phys .
* 29 * , 1627 ( 1972 )
bailly et al . , z. phys .
* c23 * , 205 ( 1984 ) .
+ r. ammar et al . , phys . lett . * b178 * , 124 ( 1986 ) .
+ a. breakstone et al . , phys . rev .
* d30 * , 528 ( 1984 )
alner et al . , phys .
rep . * 154 * , 247 ( 1987 ) .
s. aid et al .
, z. phys .
* c72 * , 573 ( 1996 ) .
ansorge et al . , z. phys .
* c43 * , 357 ( 1989 ) .
r. barate et al .
, cern - ppe/96 - 186 , to appear in phys .
+ p. abreau et al . , z. phys . *
c50 * , 185 ( 1991 ) .
+ b. adeva et al . , z. phys . * c55 * , 39 ( 1992 )
acton et al . , z. phys .
* c53 * , 539 ( 1992 ) .
+ k. ackerstaff et al .
, cern - ppe/97 - 015 , submitted to z. phys . c. d. buskulic et al .
, z. phys . *
c69 * , 15 ( 1995 ) .
r. szwed , g. wrochna and a.k .
wrblewski , mod .
phys . lett .
* a5 * , 1851 ( 1990 ) .
+ s. carius and g. ingelman , phys .
* b252 * , 647 ( 1990 ) . s. hegyi , manuscript in preparation .
.results of hnbd fits ( weibull case , @xmath42 ) to the multiplicity distributions measured in inelastic @xmath3 reactions [ 14 ] .
the last row corresponds to the inelastic @xmath4 data of the ua5 collaboration .
[ cols="^,^ , > , > " , ] | the recently introduced h - function extension of the negative binomial distribution is investigated .
the analytic form of @xmath0 is rederived by means of the mellin transform .
applications of the hnbd are provided using experimental data for @xmath0 in @xmath1 , @xmath2 , inelastic @xmath3 and non - diffractive @xmath4 reactions . in multiparticle
phenomenology we celebrate the 25th anniversary year of the invention of two ideas which turned out to be highly influential : the koba - nielsen - olesen ( kno ) scaling of the multiplicity distributions @xmath0 [ 1 ] and their negative binomial distribution ( nbd ) type parametrization [ 2 ] .
both subjects received a great deal of attention during the past 25 years , see refs .
[ 3 - 5 ] for review articles .
a popular way of introducing the nbd as a model of @xmath0 is the poisson transform defined by @xmath5 for the nbd the continuous probability density @xmath6 is the gamma distribution providing the asymptotic kno scaling form of @xmath0 [ 3 ] . at high energies
the agreement between the nbd and observations is often not quite satisfactory therefore it is of interest to look for possible generalizations .
we shall consider here an extension of the nbd that can be obtained via the scaling transformation @xmath7 of the gamma variate with scaling exponent @xmath8 .
the modified @xmath6 can be written in the form @xmath9^\mu\right)\ ] ] which is the generalized gamma distribution [ 6 ] with shape parameter @xmath10 and scale parameter @xmath11 .
its moments are given by @xmath12 for @xmath13 the ordinary gamma distribution is recovered and the poisson transform of eq .
( 2 ) yields the nbd , @xmath14 where @xmath15 is the average multiplicity .
for @xmath16 the basic properties of the poisson transformed generalized gamma density have been determined in two previous letters [ 7,8 ] .
since the analytic form of @xmath0 can be expressed in terms of the h - function of fox [ 9 ] ( see the appendix for a summary ) we will call the distribution as hnbd for short .
the main goal of the present paper is the analysis of various multiplicity data which are known to be in disagreement with the @xmath13 special case . in ref .
[ 7 ] the analytic form of the hnbd was obtained utilizing the laplace transform of the fox function given by the equation pair ( a.6 - 7 ) .
this is a straightforward method since @xmath17 in eq .
( 1 ) is the laplace transform of the product @xmath18 which can be expressed in terms of @xmath19 for the generalized gamma distribution . to illustrate the simplicity of manipulating h - functions ,
first we provide a slightly different approach to the analytic form of the newly developed probability law .
the laplace and mellin transforms of a function @xmath20 , @xmath21 , are defined by @xmath22 and by @xmath23 respectively .
they are related to each other through @xmath24 which can be justified by fubini s theorem [ 10 ] . from the moments of @xmath6 given by eq .
( 3 ) one obtains @xmath25 for the mellin transform of @xmath18 and the use of eq .
( 7 ) yields @xmath26 comparison with the mellin transform ( a.8 ) of the fox function shows that @xmath27 , providing @xmath17 , is expressible in terms of @xmath19 with the following parameters : @xmath28\right .
\quad\mbox{for } \mu>1\ ] ] where the restriction on @xmath29 follows from the existence condition @xmath30 of @xmath19 discussed in the appendix .
thus , using identity ( a.4 ) , one arrives at @xmath31\right .
\quad\mbox{for } \mu>1\ ] ] and with the help of identity ( a.3 ) we can write the analytic form of @xmath0 in terms of a legal h - function also for @xmath32 : @xmath33\right .
\quad\mbox{for } 0<\mu<1.\ ] ] a reparametrization of eqs .
( 11 ) and ( 12 ) using @xmath15 in place of the scale parameter @xmath34 is more convenient .
the factorial moments of @xmath0 are equivalent to the ordinary moments of @xmath6 for eq .
( 1 ) therefore @xmath35 and eq .
( 3 ) yields @xmath36 for unit scaling exponent @xmath29 the hnbd reduces to the pure nbd given by eq .
( 4 ) which is the @xmath13 marginal case of eq . ( 11 ) for @xmath37 and of eq . ( 12 ) for @xmath38 .
this can be deduced by comparing eq .
( 10 ) with ( a.9 ) and recalling that @xmath39 yields the binomial function @xmath40 for @xmath41 . at first glance
it may seem strange that a three - parameter discrete probability law requires three different expressions for @xmath0 each having a restricted domain of validity over the parameter space .
but this complication is a reasonable price for the high degree of flexibility of the hnbd .
it involves as a special or limiting case the poisson transform of many widely used probability densities such as the * gamma distribution for @xmath13 and @xmath10 , * weibull distribution for @xmath42 and @xmath8 , * chi distribution for @xmath43 and @xmath10 , * pareto distribution for @xmath44 and @xmath45 , * log - normal distribution for @xmath46 and @xmath47 to mention but a few .
therefore one may think the hnbd as a family of different discrete probability laws rather than a single distribution . since the infinite divisibility of @xmath6 is preserved by @xmath0 for eq .
( 1 ) we can use bondesson s theorem [ 11 ] to deduce that the hnbd is infinitely divisible if @xmath48 , otherwise this feature of @xmath0 is violated . before turning our attention to data analysis
let us consider briefly the computational aspects of the hnbd .
unfortunately , the fox function is not available in software packages for the computation of special functions .
a straightforward way to calculate @xmath0 in fitting procedures is the evaluation of the integral eq .
( 1 ) numerically with @xmath20 given by eq .
( 2 ) , requireing @xmath49 and changing @xmath50 to @xmath51 in the poisson weight . for @xmath46 and @xmath47
the hnbd converges to the poisson transformed log - normal distribution which lacks a representation in terms of known functions . to include this important limiting case into the fits eq .
( 2 ) should be written in the following form [ 7 ] : @xmath52\qquad\mbox{for $ p\neq0$}\ ] ] where @xmath53 with new parameters @xmath54 , @xmath55 and @xmath56 .
this reparametrization of the generalized gamma density allows @xmath57 as well ( @xmath58 ) and the log - normal distribution can be mapped to the origin in @xmath55 , @xmath59\quad\mbox{\ \ for $ p=0$}.\ ] ] restricting the location parameter @xmath56 to @xmath60 @xmath0 is computed by inserting eqs .
( 14 ) and ( 15 ) into eq .
( 1 ) , changing @xmath50 to @xmath51 in the poisson weight and evaluating the integral by the romberg method or gaussian quadrature [ 12 ] or using the numerical integration package of mathematica [ 13 ] . applying the hnbd to experimental data we shall consider typical reactions and energies where the pure nbd is claimed to be in contradiction with observations .
significant deviations from @xmath13 are expected in 1 .
inelastic @xmath3 reactions up to isr energies , 2 .
deep - inelastic @xmath2 scattering at hera energies , 3 .
@xmath1 annihilations at lep energies , 4 .
non - diffractive @xmath4 collisions at top sps energy @xmath61 gev . items 2 and 3
have already been investigated in refs .
[ 7,8 ] and here only a few additions will be made .
but let us first consider inelastic @xmath3 reactions , perhaps the earliest typical example for the failure of nbd fits . without the aim of completeness we have analysed multiplicity distributions in the c.m .
energy range @xmath62 gev [ 14 ] including the inelastic @xmath4 data of the ua5 collaboration . in the fitting procedure numerical evaluation of the integral eq .
( 1 ) was carried out as described above .
after some initial playing with the parameters it was found that the @xmath42 special case of the hnbd , i.e. the poisson transformed weibull distribution , provides good description of the investigated data .
the results of fits are collected in table 1 .
it is worth noticing that the best - fit value of @xmath29 slowly decreases with increasing energy , further , @xmath63 for each data set . according to some preliminary results the @xmath42 special case produces similar quality fits for inelastic @xmath64 and @xmath65 reactions .
the poisson transformed weibull law works successfully in deep - inelastic @xmath2 scattering at hera energies as well . in ref .
[ 8 ] the multiplicity distributions measured by the h1 collaboration [ 15 ] were compared to the weibull case of the hnbd neglecting systematic errors of @xmath0 .
to avoid possible misunderstanding we repeated the analysis with the inclusion of both sources of experimental uncertainties .
the average multiplicity @xmath15 was fixed at its observed value and only @xmath29 was treated as free parameter in the fitting procedure .
the results are collected in table 2 , the quoted errors of @xmath29 are statistical and systematic .
the success of the @xmath42 special case of the hnbd ( even with omitted systematic errors of @xmath0 ) suggests that the inelastic @xmath3 and deep - inelastic @xmath2 multiplicity distributions are similarly shaped .
this seems to be confirmed by the observation [ 15 ] that the h1 data and the best - fit nbd deviate from each other at small multiplicities . due to the diffractive fraction of events
, the same happens for inelastic @xmath3 data too . in both reactions
the rise of @xmath0 at small @xmath66 is less rapid than that of the nbd and this property is well described by the poisson transform of the weibull distribution with @xmath63 .
a completely different type of deviation from @xmath13 arises in non - diffractive @xmath4 collisions at @xmath61 gev .
the multiplicity distributions measured by the ua5 collaboration [ 16 ] are at present the highest energy published data for @xmath0 .
the negative binomial fits are satisfactory in narrow pseudorapidity intervals but for @xmath67 the nbd fails to reproduce the observed shape of @xmath0 . according to widespread opinion the failure of the nbd is caused by the much quoted shoulder structure . in narrow bins
this is not visible but for @xmath67 some kind of structure can indeed be seen in the tail of @xmath0 . however , the dominant source of discrepancy between the nbd and observations has nothing to do with the shoulder .
it was demonstrated already in [ 16 ] that the real difficulty for wide @xmath68-intervals is the pronounced narrow peak of the heavy tailed distributions .
performing hnbd fits reveals clearly that the @xmath13 special case is unable to reproduce a highly skewed shape which is so characteristic of the ua5 data for @xmath67 . but letting @xmath29 to vary freely in the fits it turns out that the @xmath46 , @xmath47 limiting case of the hnbd , i.e. the poisson transformed log - normal distribution , yields reasonable description of @xmath0 possessing a narrow peak and extended tail .
the results for full phase - space and wide @xmath68-windows are collected in table 3 .
although systematic deviations between the experimental and theoretical @xmath0 inevitably occur due to the shoulder structure ( the hnbd is unimodal ) the overall shape of the multiplicity distributions can be reproduced remarkably well by the two - parameter poisson transformed log - normal law . for the full phase - space data the fit
is illustrated in fig . 1 . interestingly , the log - normal limit of the hnbd was found to be successful also in @xmath1 annihilations at the @xmath69 peak . in ref . [ 7 ] the results of three - parameter hnbd fits were presented favouring @xmath70 for the full phase - space multiplicity distributions .
here we provide the outcome of two - parameter poisson transformed log - normal fits , see table 4 .
the analysed data sets now include the aleph data with updated systematic errors and the opal data corresponding to @xmath71 gev [ 17 ] .
as is seen the quality of fits are satisfactory for each experiment .
similar results can be obtained for the aleph data measured in central rapidity windows [ 18 ] , these are quoted in table 5 .
we have to mention that the delphi data in @xmath72-intervals exhibit significant shoulder structure which can not be reproduced by unimodal distributions and therefore the hnbd fails as well .
the importance of the log - normal distribution to approximate @xmath0 in @xmath1 and @xmath4 collisions has already been stressed in the literature [ 19 ] .
one may think that the observed success of the poisson transformed log - normal law is only another manifestation of the above cited results .
but the widely known and accepted evidences for the log - normality of @xmath0 at present energies are not rarely questionable
. their comparison to our findings ( which indicate that log - normality emerges asymptotically ) will be treated in a separate paper [ 20 ] . in conclusion , we have investigated the h - function extension of the nbd obtained by the poisson transform of the generalized gamma distribution eq .
( 2 ) .
the analytic form of @xmath0 was rederived via mellin transform . since eq .
( 2 ) involves as special and limiting cases many classical probability densities , the hnbd is expected to obey high degree of flexibility .
our applications confirmed this expectation .
fitting the hnbd to various multiplicity data claimed to be in disagreement with the pure nbd we have obtained satisfactory results . according to these findings ,
two major types of departure from @xmath13 arise in multiparticle production . for inelastic @xmath3 and deep - inelastic @xmath2 scattering
the dominant source of discrepancy is the less steep rise of @xmath0 at small multiplicities caused by the diffractive component of the underlying dynamics ( at least for @xmath3 data ) .
the multiplicity distributions can be reproduced successfully by the weibull case of the hnbd with @xmath63 .
the other @xmath16 type behaviour arises in non - diffractive @xmath4 collisions at @xmath61 gev where the nbd is incompatible with the highly skewed shape of @xmath0 for @xmath67 .
these heavy tailed distributions with a narrow peak are best described if @xmath46 and @xmath47 , i.e. by the log - normal limit of the hnbd . despite of the absence of skew multiplicity curves , the same type of departure from the pure nbd has been observed also in @xmath1 annihilations at lep energies .
this remarkable ubiquity of the @xmath46 limit for the @xmath1 and @xmath4 data at top energies has a natural explanation based on renormalization group arguments for asymptotic kno scaling [ 20 ]
. it will be interesting to see how precisely can the poisson transformed log - normal law reproduce the forthcoming @xmath4 data at @xmath73 gev at tevatron .
the observed success of the hnbd indicates that the majority of existing multiplicity data can be interpreted on the basis of the asymptotic kno scaling form of the negative binomial distribution with the extension of its validity to positive powers of the scaling variable .
it is hoped that in the next 25 years the hnbd will provide a helpful unifying framework to study the nb regularity and the origin of possible deviations . |
quantum spin ladder systems have attracted much attention in the past few years@xcite .
strong quantum fluctuations prevent any long range antiferromagnetic ( af ) order in quasi - one dimension .
the magnetic phases of the ladder systems are rich and strongly dependent on their geometric structures .
several types of disordered `` quantum spin liquid '' phases are known @xcite .
typical examples of two - leg ladders are the railroad ladder and the zigzag ladder .
the railroad ladder has a singlet ground state with elementary triplet excitations ( magnons ) @xcite .
depending on the ratio of the leg to rung exchange integrals , the zig - zag ladders may have gapless ground states with algebraically decaying spin correlations or spontaneously broken dimerized ground states@xcite .
the gapped dimer ground state is degenerate , and the elementary excitations are pairs of spinons .
the spin ladders have been studied experimentally in compounds such as @xmath4 and @xmath5 @xcite .
the two - leg zigzag ladder , which has been well investigated , is perhaps the simplest example of the frustrated spin model and highlights the role played by frustration .
however , less attention has been paid on asymmetric spin ladders where the exchange integrals on both legs are different .
only the extreme case where one leg of a zig - zag ladder is missing entirely ( sawtooth or @xmath6-chain ) has been solved@xcite . in this paper , we perform a systematic study of an asymmetric zig - zag spin ladder , which is a heisenberg model defined on the structure shown in fig .
[ mod_fig ] . in general , it is convenient to represent the zigzag ladder as a spin chain with nearest - neighbor ( nn ) exchange @xmath0 and next - nearest - neighbor ( nnn ) exchange @xmath7 corresponding to the inter - leg exchange @xmath0 and intra - leg exchange @xmath7 .
equivalently , the asymmetric ladder model can be represented as a chain with an alternating nnn exchange @xmath8 { \bf s}_{l } \cdot { \bf s}_{l+2 } \right\ } \ ; , \label{spinchain}\ ] ] where @xmath9 and @xmath10 are the nearest - neighbor ( nn ) and alternating nnn coupling constants , respectively .
the introduction of @xmath3 makes the exchanges on top and bottom legs different .
@xmath11 is the ordinary zig - zag ladder or frustrated spin chain@xcite and @xmath12 is the extreme case with one leg completely missing .
fixing @xmath13 gives an exactly solved continuous manifold of hamiltonians @xcite .
the majumdar - ghosh ( m - g ) @xcite and the sawtooth chain @xcite are extreme cases with @xmath14 and @xmath15 respectively .
= 2.5 cm for @xmath16 the model ( [ spinchain ] ) is well understood for general @xmath7 @xcite .
frustration due to @xmath7 is irrelevant when @xmath17 @xcite , and the system renormalizes to the heisenberg fixed point @xcite , whose ground state is described as a spin fluid or luttinger liquid with massless spinon excitations .
as @xmath18 , the frustration term is relevant and the ground state is doubly degenerate .
particularly , the ground state has simple product form at @xmath19@xcite .
the critical value of @xmath20 can be determined numerically@xcite . for @xmath21 ,
quite different field theory treatments are required depending on the ratio of @xmath22 @xcite . since relatively less
is known about the asymmetric ladder , we study physical effects brought about by the leg - asymmetry . before solving the quantum problem , we start with the classical problem which can give us an instructive insight into properties of the asymmetric spin model . in the classical limit ,
the ground state of this model is a nel state for @xmath23 and a spiral with a pitch @xmath24 for @xmath25 .
both ground states and the critical ratio of exchange integrals separating them , are independent of @xmath3 .
the excitations may depend on @xmath3 , however .
certainly , the quantum case is much more complicated .
when the asymmetric exchange interaction is introduced , some unexpected phenomena will appear .
we found that the heisenberg fixed point is no longer stable and flows to an intermediate - coupling fixed point with gapless excitations and a vanishing spin velocity .
@xcite however , there is still reminiscence of classical results .
it was found that the ground state is independent of @xmath3 when @xmath26 , but the excited gap is decreased by @xmath3 .
part of the work on weak frustration regime has been reported in our previous letter@xcite , and here we quantitatively study the crossover of the excitation spectrum from the symmetric m - g model to the extremely asymmetric sawtooth model in detail .
an extension to a two - dimensional double layer model is also presented .
the outline of this paper is as follows . in section
ii , the effective low energy theory of the asymmetric ladder is derived by using bosonization .
we qualitatively discuss the effect of alternating nnn interaction and compare our model with the well - known spin - peierls model .
the phase diagram of the system is discussed with the help of the renormalization group analysis . in section iii , we study the asymmetric model at the special point @xmath27 .
the ground state and excitation properties are also discussed . in section iv
, we generalize our model to a two - dimensional double layer model , whose ground state is a dimer product state and the excitations are magnons .
section v contains our conclusions .
following the general procedure of transforming a spin model to an effective model of continuum field @xcite , we convert the spin hamiltonian to a hamiltonian of spinless fermions using jordan - wigner transformation , then map it to a modified luttinger model with umklapp and backscattering - type interactions . using the standard dictionary of bosonization@xcite , we obtain the effective boson hamiltonian @xmath28 with @xmath29 \ ; , \label{h0 } \\ h_1 & = & \int dx \left [ { \frac{g_{3}}{2(\pi a)^{2}}}\cos 4\phi+{\frac{g_{1}}{\pi ^{2}a } } \left ( \partial_{x } { \phi } \right ) \cos 2\phi \right ] \ ; , \label{h1 } \end{aligned}\ ] ] where @xmath30 is a bosonic phase field and @xmath31 its canonically conjugate momentum . here
@xmath32 is a short - distance cutoff , @xmath33 is the umklapp - scattering amplitude and @xmath34 is the amplitude of the alternating nnn field .
the parameters @xmath35 and @xmath36 are the effective spin velocity and coupling constants which are given by @xmath37 in genereal , these values are only valid near the free fermion point ( @xmath38 ) , whereas @xmath39 is fixed by the symmetry at the isotropic point .
the corresponding spin - correlation functions can be calculated from the boson representation , which gives @xmath40 it is clear that the su(2 ) symmetry is restored at the isotropic point with @xmath39 . for small @xmath41 and @xmath42 ,
@xmath43 could be considered as a perturbation to @xmath44 . without the @xmath42 term
, @xmath45 represents a standard sine - gorden model @xcite .
the @xmath41 term is either marginally irrelevant , which leads to the weak - coupling heisenberg fixed point , or relevant , which drives the system to a strong - coupling dimer state .
therefore , what we are interested in is how the alternating nnn interaction @xmath42 changes the physical properties of the system .
qualitative results on the influence of the new interaction ( @xmath42 ) can be obtained from scaling analysis and physical considerations .
an important information about perturbative operators is whether they are relevant , marginal or irrelevant . in general ,
only the most relevant perturbation is important , because the irrelevant operator will scale to zero at large lengths .
we can give an approximate estimate by comparing scaling dimensions of the given operators .
it follows that @xmath46 has scaling dimension of @xmath47 and @xmath48 has scaling dimension of @xmath49 .
therefore , the scaling dimensions of the umklapp and the alternating nnn terms @xmath41 and @xmath42 are @xmath50 at the isotropic heisenberg fixed point , @xmath51 is marginal with @xmath52 , while the @xmath53-term with @xmath54 is relevant .
we conclude that @xmath42 destabilizes the isotropic heisenberg fixed point and the spin liquid ground state .
on the other hand , for @xmath55 , the @xmath41 term is marginally relevant and renormalizes to a strong coupling fixed point in the long - wavelength limit . near the strong coupling dimer fixed point ,
the @xmath41 term is much more relevant than the @xmath42 term .
usually ( e.g. @xmath56 ) , the boson field @xmath57 locks into a constant value with small fluctuations , and an associated excitation gap . here
, the constant solution @xmath58 corresponds to the degenerate ground state at the strong coupling fixed point for @xmath59 .
the standard @xmath60 sine - gordon equation has a pair of solutions of kink and antikink , which describe the elementary excitations ( a pair of spinon ) for the degenerate dimer phase . even including the less relevant @xmath42 term
, the soliton solutions will survive .
however , the phase locking of @xmath42 term is forbidden by the @xmath61-pre - factor to the @xmath62-term in @xmath63 . in this sense
, there is no standard strong coupling theory for the @xmath42-term . from known results on the sawtooth chain@xcite and the m - g model@xcite
, we expect that the @xmath42 term , induced by the alternating nnn interaction , does not confine spinons and plays a quite different role than the dimerization by other degrees of freedom .
moreover , the difference in the size of excitation gaps in these two models implies that the @xmath42 term quite generally competes with the umklapp term whereas an external nn dimerization would cooperate .
as we will show , it turns out that @xmath42 opens no spin gap despite being a relevant perturbation of the heisenberg fixed point .
this result is also corroborated by the absence of a magnetization plateau in our model in small magnetic fields@xcite .
for an alternating nn exchange , a magnetization plateau is observed in small magnetic fields , but for alternating nnn exchange , it is only observed in high fields @xcite .
to get an instructive insight , we would like to compare the alternating nnn interaction in our model with the alternating nn interaction in the well - known spin - peierls ( sp ) model . in the language of field theory
, the external dimerization corresponds to a relevant term @xmath64 .
this term is always much more relevant than the @xmath41 term and lifts the degeneracy ( @xmath58 ) of the ground state . with the presence of the @xmath65 term ,
the lowest energy configuration is @xmath66 if @xmath67 or @xmath68 if @xmath69 .
since the most relevant term is @xmath70 , the corresponding sine - gordon system has a pair of soliton and antisoliton solutions ( excitations with @xmath71 ) as well as two breather solutions ( excitations with @xmath72).@xcite the lowest breather is precisely degenerate with the soliton and antisoliton excitations , and they form a bound state which corresponds to @xmath73 triplet . in this case , the elementary excitation should be a spin triplet and a spin singlet , no spinons exist as elementary excitation.@xcite however , the alternating nnn interaction , @xmath74 , does not lift the degenerate phases @xmath58 , due to the existence of the pre - factor @xmath75 .
we now perform a perturbative renormalization group ( rg ) analysis to our model by following standard procedures.@xcite introducing the reduced variables @xmath76 and @xmath77 , we obtain the rg equations @xmath78 under a change of length scale @xmath79 . here
we define @xmath80 .
the rg equation for the spin velocity @xmath35 is a consequence of the anisotropy of the @xmath42-interaction in the classical 2d xy - model , i.e. its non - retarded but non - local character in the quantum field theory ( [ h1 ] ) . for @xmath81 equations ( [ dy1 ] ) and
( [ du ] ) would not appear , thus @xmath35 is not renormalized .
the rg equations ( [ dk ] ) and ( [ dy3 ] ) with @xmath81 describe the symmetric spin ladder system with a kosterlitz - thouless transition.@xcite spin - rotation invariant models scale along the separatrix between nel and spin liquid phases , or along its continuation into the dimer regime . linearizing the rg equations ( [ dk])-([du ] ) around the isotropic heisenberg fixed point ( @xmath82 ) and defining @xmath83
, we obtain the linearized rg equations @xmath84 a family of solutions of the rg equations , projected on the @xmath85-plane , are shown in fig .
[ rg - fig ] .
we choose the initial value of @xmath86 as @xmath87 and find the trajectories are not sensitive to the choice of the initial values of @xmath86 . from the rg equations , one can directly find that there exist two intermediate fixed points given by @xmath88 . here
@xmath86 takes the values @xmath89 , which reflects that our rg equations is symmetric to @xmath86 . as shown in fig .
[ rg - fig ] , the intermediate fixed point @xmath90 on the plane of @xmath91 is stable along the line @xmath92 where the spin - rotation invariance is protected . for points near the intermediate fixed points , but not exactly on the line of @xmath92 , the spin - rotation invariance is broken , thus they will flow to the spin - fluid fixed point @xmath93 or the strong coupling fixed point @xmath94 .
= 6.0 cm at the fixed point , the rg equation ( [ du ] ) implies that the spin velocity @xmath95 is renormalized .
the robustness of the existence of this intermediate fixed point against higher order perturbations does not depend on the exact value of the fixed point as long as the fixed point is located on the rg separatrix with @xmath96.@xcite the intermediate fixed point is thus described by an effective fixed - point hamiltonian .
inverting the definitions of @xmath97 , we rewrite the effective fixed - point hamiltonian as a product of @xmath95 and a term independent of @xmath95 , @xmath98
. then the vanishing of @xmath95 leads to a trivial fixed point hamiltonian @xmath99 . as a result of the vanishing of the renormalized spin velocity , the elementary excitations at the fixed point , spinon and antispinon , are still gapless
the vanishing of velocity is usually interpreted as a sign of ferromagnetism , however we interpret this as our spins effectively decoupling at the lowest energy scales , i.e. a kind of asymptotic freedom in this spin - rotation invariant ladder@xcite . also , the numerical results of wiessner et al.@xcite indicate a paramagnetic susceptibility . when @xmath7 increases beyond a critical value @xmath100 ( now depending on @xmath3 ) , the rg flows to a strong coupling fixed point , which corresponds to the quantum dimer phase .
for @xmath101 , this critical point is indicated in @xmath102-coordinates in fig .
[ rg - fig ] by an asterisk . for small @xmath3 and @xmath103 ,
our rg equations show that the system will remain in the universality class of the dimer solid corresponding to the strongly fixed point , however the spin gap is decreased by increasing @xmath3 . basically , the gap size @xmath104 where @xmath105 is the length of the scaling trajectory from the initial values to the point where the most relevant perturbation is of order unity . this length is increased , and @xmath6 therefore decreased , by the @xmath86 contributions to @xmath36 and @xmath106 being opposite in sign to those of @xmath36 and @xmath106 in eqs .
( [ dk ] ) and ( [ dy3 ] ) .
recently , sarkar and sen @xcite studied the same model by using a nonlinear @xmath107-model field theory and abelian bosonization .
however , the main discrepancy between our work and theirs @xcite is that we kept the bosonized operator of the alternating nnn operator and analyzed it by rg , while they just discarded it by giving an argument of the irrelevance of the operator . for an anisotropic xxz chain
, our rg result indeed shows that the @xmath42 operator is irrelevant @xcite in the meaning that it does not drive the system to a new phase , and this is consistent with that of sarkar et .
but the main difference lies in the question whether an intermediate fixed point exists and whether this fixed point corresponds to a phase different from a luttinger liquid . in our previous work
the argument of a vanishing spin - wave velocity plays a crucial role in the existence of such an unusual phase .
if the spin - velocity does not vanish , one should explain the fixed point as a spin - liquid phase as in the @xmath108 model with the spin velocity renormalized .
our result also suggests that the quantum phase transition parameter ( @xmath109 ) to the dimer phase is changed by the alternating nnn operator , which may be verified directly by numerical simulations like the density - matrix renormalization group .
we also notice that the magnetization curve of the heisenberg model with an additional alternating nnn operator @xcite gives an obvious different magnetization susceptibility from the one without it .
their susceptibility is enhanced by this new interaction with respect to an equivalent heisenberg chain , indicating a reduced spin velocity .
furthermore , a complete scheme to deal with the alternating nnn operator should give a correct description of the induced effect not only on the weak frustration regime but also on the strong frustration regime .
the omission of the @xmath42 term could not give any explanation why the operator shrinks the spin gap sizes in the regime of strong frustration as we will study in detail in the next section .
however , our rg analysis gives a qualitative explanation of the influence of the @xmath42 term on spin gap sizes .
therefore , we think that the scheme of discarding the @xmath42 operator based on its irrelevance seems to be oversimplified .
we hope that more numerical simulations will eventually be able to resolve this disagreement and study quantitatively the phase diagram of the ground state as a function of @xmath110 and @xmath3 .
it is generally believed that continuum field theory can not give a good description for the behavior of the system far away from the critical point . in the case of @xmath111 ,
the correlations extend only to a distance of one lattice spacing , thus the continuum field description is not a good approach .
as we have shown , the phase corresponding to large @xmath7 is the dimer phase .
this is consistent with our knowledge from the models with exactly known ground states , say , majumdar - ghosh ( m - g ) model and sawtooth model . in this section
, we prove that there exists a continuous manifold of hamiltonians with dimer product ground states as long as @xmath112 .
we start with the asymmetric ladder model @xmath113 { \bf s}_{l } \cdot { \bf s}_{l+2 } \ ; .\ ] ] the m - g model and sawtooth model are corresponding to @xmath11 and @xmath114 respectively . for the m - g model@xcite , the two linearly independent ground states , say ,
the left and right dimer ground states , are products of nearest - neighbor singlets , respectively @xmath115,~~\mid \phi _ { r}\rangle = \prod_{l = even}[l , l+1],\ ] ] where @xmath116= ( \alpha _ { i}\beta _ { j}-\beta _ { i}\alpha _ { j } ) / { \sqrt{2 } } $ ] denotes the singlet combination of spin i and j with the direction of dimers defined as @xmath117 . here
@xmath118 represents the up - spin and @xmath119 the down - spin state at site @xmath120 .
@xmath121 also represent the degenerate ground states of the sawtooth model@xcite . for the asymmetric ladder model
, we notice that the nnn exchange alternation does not modify the product states of nearest - neighbor singlets @xmath122 this is induced by the fact that the alternating nnn couplings along the upper leg and the lower leg of the ladder cancel out each other , when they operate on @xmath123 .
it is obvious that @xmath124 and @xmath125 are eigenstates of the hamiltonian ( [ saw - mg ] ) .
in fact , as we will prove , they are exactly the ground states of ( [ saw - mg ] ) .
to see this more clearly , we rewrite the asymmetric ladder model as a sum of projection operators @xmath126 : @xmath127 .\ ] ] with @xmath128 .\ ] ] here , we introduce @xmath129 to indicate the center position of three neighboring sites @xmath130 . such an operator is a special case of the general positive semidefinite lwdin s projection operators @xcite @xmath131 where @xmath132 and @xmath133 are the maximum and minimum values of the total spin @xmath134 .
as long as @xmath135 , the coefficients in eq .
( [ p - asylad ] ) are non - negative .
therefore , the hamiltonian ( [ saw - mg ] ) is a linear combination of projection operators with positive coefficients .
since @xmath136 projects a state composed of three spins , @xmath137 , into a subspace of total spin @xmath138 , its eigenvalues are @xmath139 ( if the total spin is @xmath140 ) and @xmath141 ( if the total spin is @xmath138 ) . by virtue of the properties of the positive semidefinite projection operator ,
whose lowest eigenvalue is zero , the ground state of the hamiltonian ( [ p - asylad ] ) can be constructed by choosing states with such configurations that each projection operator has the lowest eigenvalues @xmath139 when operating on these states .
it is easy to prove that @xmath123 are the exact ground states of the asymmetric ladder model ( [ saw - mg ] ) .
the ground state energy is independent of @xmath3 and given by @xmath142 there is thus an entire manifold of hamiltonians with fixed @xmath143 , parameterized by @xmath3 , with doubly degenerate ground states of nn - dimer product ground states @xmath144 . for convenience ,
in our following discussion we will shift our model by an energy of @xmath145 , @xmath146 which is equivalent in taking the ground state of the system as zero . in the following , we consider the excited state of our system ( [ saw - mg ] ) .
the elementary excitation of the system is a pair of spinons known as the kink or antikink@xcite .
first results were obtained by shastry and sutherland for the m - g model@xcite within a variational ansatz . since the ground state of ( [ saw - mg ] ) is independent of the alternating nnn exchange , the construction of the excited states for the m - g model can be directly extended to the asymmetric model ( [ saw - mg ] ) .
breaking a singlet pair in the ground state would give rise to two unpaired `` defect '' spins .
therefore , the simplest excitation consists of a pair of spinons .
the spinons can be thought as domain - walls separating different dimer ground state configurations . from symmetry consideration , the kink and antikink are identical in the m - g model , and specifically they have the same dispersions . with alternating nnn interaction ,
the symmetry between legs is broken , therefore some properties of kinks and antikinks are different , in particular their dispersion .
however , they still survive as elementary excitations of the asymmetric spin ladder system .
= 2.5 cm in general , we call a spinon at the odd site @xmath147 a kink and the other one at the even site @xmath148 an antikink ( see fig .
the kink and antikink always appear in pairs in a periodic system , however , a single spinon can be realized in the open boundary systems .
taking the variational wavefunction with one `` defect '' spin,@xcite one can easily obtain the spinon dispersion of the m - g model @xmath149 the energy gap @xmath6 is therefore @xmath150 . for @xmath151 ,
i.e. the sawtooth model , the kink excitation is much different from the antikink excitation . as shown by d.
al . and t. nakamura et .
@xcite , the kink ( k ) excitation in the sawtooth chain is exactly a single spin on odd site and dispersionless @xmath152 however , an antikink propagates with an effective mass along the lattice .
the antikink is not a free spin and spreads out to an extended region , because it is not a eigenstate of the local hamiltonian . in the first approximation ,
the antikink ( @xmath153 ) is supposed to be a single defect spin at the even site , and the dispersion obtained by variational calculation has similar form as the spinon dispersion of the m - g model @xmath154 with the corresponding energy gap of @xmath155 .
despite the variational nature of the dispersion , the results agree very well with exact numerical results @xcite .
we will explicitly calculate the change of the gap size with increasing @xmath3 .
as expected , we found that the @xmath3 term changes the energy gap size of excitations , which is consistent with our conclusion obtained by the renormalization group analysis .
following d. sen et .
al @xcite , we assume both the kink and antikink to be a @xmath156-cluster block with spin @xmath157 . it is known that for the m - g and sawtooth chain there is no closely bound kink - antikink pair whose energy is lower than that of a widely separated pair .
thus we can deal with the kink and antikink separately .
the gap of the lowest excitation is a sum of the gaps of kink and antikink @xmath158 where the subscripts , @xmath36 and @xmath153 , represent the kink and the antikink respectively .
it should be noticed that both the m - g and sawtooth model have the same energy gap size @xmath6 under the first approximation ( 1-cluster approximation ) .
that is not true as we take more precise n - cluster approximation .
= 4.5 cm = 4.5 cm the lowest energy is obtained by finding parameters which minimize the energy expectation @xmath171 since @xmath172 and @xmath173 ( @xmath174 and @xmath175 ) are symmetric about the site @xmath147 ( @xmath148 ) , there is no reason to discriminate between these configurations and it is reasonable to choose @xmath176 . the computation of eq .
( [ e_var ] ) is straightforward although a little bit lengthy , we will not give the detail here but refer to the literature @xcite .
it is found that the minimum value of @xmath177 occurs at @xmath178 and is given by @xmath179 and the minimum value of @xmath180 is @xmath181 where we take @xmath182 for convenience .
for any given value of @xmath3 , the excited gap @xmath183 ( @xmath184 ) of a kink ( an antikink ) corresponds to the minimum of eq .
( [ deltak ] ) ( eq .
( [ deltaak ] ) ) .
our result is shown in fig .
[ exc ] , which indicates that the elementary excitation gap decreases from @xmath185 in the m - g to @xmath186 in the sawtooth with the increase of the coupling constant @xmath3 , while the ground state energy is constant .
in particular , for @xmath187 , i.e. the sawtooth chain , eq .
( [ deltak ] ) reduces to @xmath188 which has the minimum @xmath189 for @xmath190 , while eq .
( [ deltaak ] ) becomes @xmath191 which has a minimum @xmath192 at @xmath193 .
it is clear that the kink excitation is exactly dispersionless , while an antikink is still a domain wall propagating with an effective mass@xcite . for the m - g model , @xmath194 , eq .
( [ deltak ] ) and eq .
( [ deltaak ] ) have the same form ( @xmath195 ) , thus @xmath196 whose minimum value is @xmath197 . in these limits , the results are consistent with the known results of the m - g and the sawtooth model , as well as our qualitative conclusion obtained from the field theory .
= 5.cm if we introduce additionally an alternating nn exchange to our asymmetric model ( [ saw - mg ] ) , the degeneracy of @xmath198 or @xmath199 will be lifted and the singlets would be pinned along the stronger nn external dimer potential .
the elementary excitations are no longer separated kinks and antikinks . in the presence of an external dimer potential , a kink and an antikink separated by a distance of @xmath129
give rise to a region in the incorrect `` ground state '' , which effectively produces a confining potential between the kink and antikink .
this potential is proportional to the distance of @xmath129 , thus the kink and antikink can not escape from each other and behave analogously as quark - antiquark pair .
@xcite the kink - antikink bound state corresponds to a magnon with spin 1 .
the interesting topic of how the confined spinons develop to magnons has been investigated by uhrig et .
recently , it was found that the 2d shastry - sutherland model@xcite can be used to explain the experimentally realized material @xmath200 , hence such kind of model with exact dimer ground state@xcite attracted much attention again . in this section
, we will show that the asymmetric spin ladder model can be generalized to a double layer model , whose ground state is a simple direct product of singlet dimers .
the double layer model is constructed from two coupled spin layers shown in fig .
[ layer - fig ] , where each layer has @xmath201 sites and couples to the other layer by the inter - layer exchange interactions @xmath202 and @xmath203 .
the intra - layer exchange interactions @xmath204 and @xmath205 on top and bottom layers may have different strengths .
the hamiltonian of our model is given by @xmath206 where the superscripts @xmath207 denote labels of the top and bottom layers .
@xmath202 is the perpendicular inter - layer exchange interaction and @xmath203 is the diagonal inter - layer exchange interaction . here
all the exchanges are taken to be positive .
a similar model has been investigated in ref .
@xcite , where the layer model is a direct generalization of the bose - gayren ladder model@xcite .
it is clear that every slice of the double layer net is just a ladder whose hamiltonian has the same form of eq .
( [ saw - mg ] ) .
thus we find that the ground state of the layer model is given by a product of all perpendicular singlet pairs @xmath208 when the condition @xmath209 is fulfilled .
a rigorous proof can be made directly by representing the layer model as a sum of the projection operators as in the spin ladder case .
the corresponding ground state energy is @xmath210 = 5.cm it is obvious that the dimer product state @xmath211 remains the ground state for @xmath212 .
the properties of the ground state are independent of the specific values of @xmath204 and @xmath205 as long as the constraint condition ( [ constraint ] ) is satisfied .
since the dimerized ground state is not degenerate , the lowest excitation is expected to be a triplet excitation , corresponding to breaking of a singlet bond , with a gap size proportional to @xmath213 .
however , the many - particle excitation spectra might be very complicated because of the effective interactions among the triplet excitations@xcite .
the spin - isotropic , asymmetric zig - zag ladders are studied using the field theory method and the variational approach . when the leg exchange integrals are small compared with the nn exchange , the spin model is mapped to a revised double frequency sine - gorden model .
renormalization group analysis shows that there are two fixed points , say , an intermediate - coupling fixed point and a strong coupling fixed point . in the weak frustration limit
, the system is described by the intermediate - coupling fixed point with gapless excitations .
the vanishing of spin velocity at the intermediate - coupling fixed point is likely to indicate a decoupling of spins at low energy scales .
apart from the isotropic separatrix , we find gapless spin liquid and gapped nel states with easy - plane and easy - axis anisotropy . for large frustration , a more usual dimer solid phase is realized corresponding to the strong coupling fixed point .
the rg analysis also predicts that the spin gap is decreased by increasing the leg - asymmetry @xmath3 . a continuous manifold of hamiltonians with the same singlet
product ground state interpolates between the majumdar - ghosh model and the sawtooth spin chain .
starting from the exact ground state wavefunction , we construct the variational wavefunction of the excited state and investigate the change of spin gap with the change of leg - asymmetry @xmath3 . in the spirit of constructing hamiltonian in the form of a sum of positive semidefinite projection operators , extension to the double layer model is carried out .
we propose an exactly solved two - dimensional double layer model with a ground state of a product of inter - layer dimers .
the authors would like to thank prof .
k. h. mtter and dr .
m. nakamura for interesting discussions .
s. chen would like to specially thank f. siano and b. han for their critical reading of the manuscript .
this research was supported by deutsche forschungsgemeinschaft through grants no .
vo436/6 - 2 and vo436/7 - 2 .
99 e. dagotto and t. m. rice , science , * 271 * , 618 ( 1996 ) e. dagotto , rep .
. phys . * 62 * , 1525 ( 1999 ) a. a. nersesyan and a. m. tsvelik , phys .
lett . * 78 * , 3939 ( 1997 ) ; a. k. kolezhuk and h. j. mikeska , phys . rev
. lett . * 80 * , 2709 ( 1998 ) y. wang , phys .
b. * 60 * , 9236 ( 1999 ) .
f. d. m. haldane , phys .
b. * 25 * , 4925 ( 1982 ) ; * 26 * , 5257 ( 1982 ) .
m. azuma , z. hiroi , m. takano , k. ishida , and y. kitaoka , phys .
lett . * 73 * , 3463 ( 1994 ) ; g. castilla , s. chakravarty and v. j. emery , phys . rev . lett . * 75 * , 1823 ( 1995 ) .
d. sen , b. s. shastry , r. e. walstedt and r. cava , phys .
b. * 53 * , 6401 ( 1996 ) .
t. nakamura and k. kubo , phys .
b. * 53 * , 6393 ( 1996 ) .
s. chen , h. bttner and j. voit , phys .
87 * , 087205 ( 2001 ) . c. k. majumdar and d. k. ghosh , j. math
. phys . * 10 * , 1388 ( 1969 ) ; * 10 * , 1399 ( 1969 ) . k. okamto and k. nomura , phys
a. * 169 * , 433 ( 1992 ) d. allen and d. snchal , phys .
b. * 51 * , 6394 ( 1995 ) s. r. white and i. affleck , phys .
b. * 54 * , 9862 ( 1996 ) a. luther and i. peschel , phys .
b. * 12 * , 3908 ( 1975 ) .
i. affleck , in les houches , session * xlix * , fields , strings and critical phenomena , elsevier , new york , 1989 j. voit , rep . prog . phys .
* 58 * , 977 ( 1995 ) .
j. v. delft and h. schoeller , annalen der physik , * 4 * 225 ( 1998 ) .
r. m. wiessner , a. fledderjohann , k. h. mtter , and m. karbach , eur .
j. b * 15 * , 475 ( 2000 ) a. fledderjohann , c. gerhardt , m. karbach , k. h. mtter , and r. wiessner , phys .
b. * 59 * , 991 ( 1999 ) .
r. f. dashen et al .
d. * 10 * , 4130 ( 1974 ) ; phys .
rev . d. * 11 * , 3424 ( 1975 ) .
a. m. tsvelik , phys .
b. * 45 * , 486 ( 1992 ) .
i. affleck , cond - mat/9705127 .
a. dobry and d. ibaceta , phys .
b. * 59 * , 8660 ( 1999 ) g. s. uhrig and h. j. schulz , phys .
b. * 54 * , r9624 ( 1996 ) . s .-
chui and p. a. lee , phys .
* 35 * , 325 ( 1975 ) j. m. kosterlitz and d. j. thouless , j. phys . c.
* 6 * , 1181 ( 1973 ) ; j. m. kosterlitz , j. phys .
c. * 7 * , 1046 ( 1974 ) .
t. giamarchi and h. j. schulz , phys . rev . b. * 37 * , 325 ( 1988 ) . j. voit and h. j. schulz , phys . rev . b. * 36 * , 968 ( 1987 ) ; phys .
b. * 37 * , 10068 ( 1988 ) .
s. sarkar and d. sen , phys .
b. * 65 * , 172408 ( 2002 ) .
p. o. lwdin , rev .
phys . * 36 * , 966 ( 1964 ) .
b. s. shastry and b. sutherland , phys .
* 47 * , 964 ( 1981 ) .
e. s. srensen , i. affleck , d. augier , and d. poilblanc , phys .
b. * 58 * , r14701 ( 1998);. r. r. p. singh and zheng weihong , phys .
b. * 59 * , 9911 ( 1999 ) .
g. s. uhrig , f. s. schnfeld , m. laukamp and e. dagotto , eur .
j. b * 7 * , 67 ( 1998 ) .
s. miyahara and k. ueda , phys .
lett . * 82 * , 3701 ( 1999 ) b. s. shastry and b. sutherland , physica b * 108 * , 1069 ( 1981 ) ; j. richter , n.b . ivanov and j. schulenburg , j.phys . : condens .
matter * 10 * , 3635 ( 1998 ) .
s. chen and h. bttner , eur .
j. b. * 29 * , 15 ( 2002 ) . h. q. lin and j. l. shen , j. phys .
. jpn . * 69 * , 878 ( 1998 ) .
i. bose and s. gayen , phys .
b. * 48 * , 10653 ( 1993 ) .
o. p. sushkov and v. n. kotov , phys .
lett . * 81 * , 1941 ( 1998 ) . | we perform a systematic investigation on an asymmetric zig - zag spin ladder with inter - leg exchange @xmath0 and different exchange integrals @xmath1 on both legs . in the weak frustration limit
, the spin model can be mapped to a revised double frequency sine - gorden model by using bosonization .
renormalization group analysis shows that the heisenberg critical point flows to an intermediate - coupling fixed point with gapless excitations and a vanishing spin velocity .
when the frustration is large , a spin gap opens and a dimer ground state is realized .
fixing @xmath2 , we find , as a function of @xmath3 , a continuous manifold of hamiltonians with dimer product ground states , interpolating between the majumdar - ghosh and sawtooth spin - chain model . while the ground state is independent of the alternating next - nearest - neighbor exchange @xmath3 , the gap size of excitations
is found to decrease with increasing @xmath3 .
we also extend our study to a two - dimensional double layer model with an exactly known ground state .
+ 2 |
the generation of harmonics by atoms and molecules in a laser field is a topic that has been thoroughly studied during the last fifteen years ( see @xcite for recent reviews ) .
the big interest lies in the potential of atoms and molecules in laser fields as sources of short - wavelength and short pulse radiation .
it is well known that when a strong linearly polarized laser field interacts with an atom only odd harmonics are generated . in more complex systems a detailed study of the symmetry transformations is required for predicting which harmonics are allowed or not @xcite . in a recent work
@xcite it was shown that peculiar selection rules can be obtained for molecules which are invariant under a discrete rotational symmetry @xmath4 ( like benzene where @xmath1 ) .
when this kind of molecules interacts with a circularly polarized laser pulse of infinite duration , the selection rules for the harmonics are @xmath5 , where @xmath6 and @xmath7 is the harmonic order .
this class of molecules is particularly interesting because they can work as a filter , i.e. , the first allowed harmonic is @xmath8 , and therefore , if @xmath0 is large , the emitted frequency can be very high .
all systems whose hamiltonian is invariant under a `` dynamical rotation '' , i.e. , a discrete spatial rotation plus a time translation , generate selection rules of the kind @xmath2 .
together with cyclic molecules another example of such a system is the single atom interacting with two - color laser fields with circular polarizations and frequencies one an integer multiple of the other . in both the molecular and the atomic case it is possible to obtain these selection rules also from the conservation of the angular momentum component perpendicular to the polarization plane . in a previous paper @xcite
we have presented results from numerical simulations for a benzene model - molecule showing that , when a realistic , finite pulse is taken into account , other lines than those expected from the @xmath9 selection rule can be present . in this paper
we will show that , if embedded in a general group theory treatment , also this case can be understood in a systematic way .
our analysis will allow us to achieve a full understanding of the molecule harmonic spectrum .
moreover , we will show that the field - dressed level scheme can be inferred from the harmonic spectrum .
the particular features and differences of a finite pulse compared to an infinite one will be also addressed .
all the laser fields considered are circularly polarized in the @xmath10-plane , and atomic units are used throughout the paper .
we apply the dipole approximation which is excellent for the laser frequencies and intensities under consideration .
the hamiltonian @xmath11 of a cyclic molecule interacting with a single laser field and the hamiltonian @xmath12 of a single atom interacting with two circularly polarized lasers are , from a symmetry point of view , equivalent . both are invariant under certain dynamical rotations .
we will consider only a single electron active .
this assumption does not introduce any loss of generality to our derivations @xcite .
therefore @xmath11 in cylindrical coordinates @xmath13 reads @xmath14 where @xmath15 is the amplitude of the electric field .
if @xmath16 is a model potential for a cyclic molecule with @xmath0 ions the hamiltonian @xmath11 is invariant under the transformation @xcite
@xmath17 in the case of the single atom in the two - color field @xmath12 is @xmath18 where @xmath19 and @xmath20 are the electric field amplitudes of the two laser fields and opposite polarization has been assumed .
the second laser has a frequency that is @xmath21 times the frequency of the first laser , with @xmath21 integer .
it is nice to see that it is really the second laser which generates the particular symmetry .
in fact , all the terms except the last one of @xmath12 , are invariant under the continuous transformation @xmath22 where @xmath23 has , so far , an arbitrary value .
if no other term is present in @xmath12 no harmonics are generated .
forcing also the last term to be invariant under such a transformation leads to @xmath24 therefore the hamiltonian @xmath12 is invariant under the discrete transformation @xmath25 and this is equivalent to the case of a cyclic molecule with @xmath26 .
hence the selection rules for this case are equal to those derived in @xcite for the cyclic molecules , @xmath5 .
the case where the two lasers have the same polarization can be handled with the same method , leading to similar selection rules .
further details will be discussed elsewhere @xcite . in the following the two lasers
are always considered with opposite polarization .
an experiment with the two frequencies @xmath27 and @xmath28 was performed @xcite , and its theoretical aspects were recently discussed within a semiclassical three - step model @xcite .
selection rules for the case of an atom interacting with a circularly polarized field of frequency @xmath27 and a linearly polarized field of frequency @xmath29 were also derived @xcite , confirming numerical results obtained previously @xcite . in the derivation of the @xmath9-selection
rule one assumes that only a single non degenerate , bound state is involved in the harmonic generation process . with this restriction the selection rules
can be also derived simply with a change of reference frame . in the rotating laser field
frame the observer `` sees '' an ionic potential oscillating with a frequency @xmath29 in a static field ; in the two - color atomic case instead , after the change of frame , the atom is at rest in the static field and it experiences an oscillating field of frequency @xmath30 with @xmath31 .
hence , in this frame the dipole has only fourier components of frequency @xmath32 with @xmath33 .
going back to the laboratory frame where the molecule is at rest , the fourier components are shifted @xmath34 and the selection rules presented in @xcite are obtained . keeping the assumption of a single bound state involved in the process , it is , moreover , possible to derive the same selection rules also arguing about the angular momentum only .
let us first discuss about the cyclic molecule : because of the symmetry of the molecule , an expansion of the ground state in spherical harmonics @xmath35 has to look like @xmath36 , i.e , the ground state is expanded over functions that have the required azimuthal symmetry .
the wavefunction is then invariant under a rotation of @xmath37 .
let us assume that the laser field has a polarization @xmath38 , i.e. , the projection of the angular momentum of a photon on the field propagation direction is equal to 1 .
if a harmonic of polarization @xmath38 is emitted , the molecule , for keeping the symmetry , hence has to absorb @xmath39 photons so that the final change in the projection of the angular momentum is @xmath40 . on the contrary , if the harmonic emitted has polarization @xmath41 , the molecule must absorb @xmath42 photons .
the consequence is that for each @xmath43 the higher harmonic @xmath39 has the same polarization as the laser field , and the lower @xmath42 the opposite one , as it was obtained with the dynamical symmetry method in @xcite . for the atom
the derivation is similar @xcite .
let us assume an electron is in an @xmath44 state ( the quantum numbers @xmath45 and @xmath46 are zero ) .
let the first laser of frequency @xmath27 have a polarization @xmath38 and the second laser of frequency @xmath47 a polarization @xmath41 .
if a harmonic @xmath38 is emitted the sum of all the components along @xmath48 of the angular momentum carried by absorbed photons has to be @xmath49 .
if from the first laser the atom absorbs @xmath50 photons , it must absorb @xmath51 from the second laser .
the total absorbed energy is @xmath52 = \omega [ ( p-1)(1+\eta ) + 1]$ ] . since @xmath50 is arbitrary , with @xmath53 and @xmath54 we see that the @xmath55 harmonic is emitted . with the same argument ,
starting from a harmonic of polarization @xmath41 , we obtain @xmath42 .
from a group theory point of view the hamiltonian @xmath56 of a cyclic molecule without the laser field has a symmetry described by the group @xmath57 @xcite .
when the molecule interacts with a laser field which has no temporal envelope its total hamiltonian , given by ( [ ham_mo ] ) , can be written as @xmath58 where @xmath59 is the molecule - laser interaction .
first we study the symmetry group of the total hamiltonian @xmath11 .
@xmath56 has the symmetry @xmath57 and @xmath60 is described by @xmath61 , i.e , an extension to infinity of the so called `` cyclic dynamical symmetry group '' @xcite , defined by @xmath62 where @xmath63 is the order of the group and the number of its elements .
@xmath64 is the operator already introduced in ( [ rotat ] ) which acts in space and time , and is defined as @xmath65 combining @xmath57 and @xmath61 we have that the symmetry group of the hamiltonian @xmath11 is @xmath66 where @xmath0 is the number of ions in the molecule which determines the discrete rotational symmetry . because @xmath67 is a cyclic group , it is also abelian , and therefore it has the following properties : ( _ i _ ) each element forms a class by its own and hence there are @xmath0 irreducible representations , ( _ ii _ ) each irreducible representation is one - dimensional , i.e. , the elements are complex numbers .
the group @xmath67 is isomorphic to the well known group @xmath4 and therefore the same character table apply .
it is interesting to look how , from a physical point of view , the symmetry of the molecular hamiltonian changes when the laser field is present .
the group @xmath57 contains @xmath68 elements , i.e. , @xmath0 rotations about an axis of the @xmath0th order , @xmath0 rotations through an angle @xmath69 about horizontal axes , @xmath0 reflections @xmath70 at the vertical planes and @xmath0 rotary - reflection transformations @xmath71 @xcite .
looking at the field - dressed hamiltonian ( [ ham_mo ] ) it appears that only the rotations about the vertical axis are still symmetry transformations .
in fact , the reflection at the vertical planes and the rotations about horizontal axes would require a space - dependent time transformation .
the reflection @xmath72 at the horizontal plane would change the polarization of the laser field .
hence from the @xmath68 possible operations we had for @xmath56 only @xmath0 , i.e. , the rotations about the vertical axis , keep the hamiltonian @xmath11 invariant . for a certain group @xmath4
the type of its irreducible representations depends on @xmath0 .
if @xmath0 is odd there are one real and @xmath73 complex representations , if @xmath0 is even there are two real and @xmath74 complex representations .
half of the complex representations is the complex conjugate of the other half .
usually in the literature , the two complex conjugated representations are considered equivalent and indicated with e , normally reserved for two - dimensional representations @xcite . in our case , because the hamiltonian ( [ ham_mo ] ) is not time - reversal invariant , these representations are not equivalent .
therefore we have to keep all the @xmath0 representations separated .
we call each of the @xmath0 irreducible representations @xmath75 where @xmath76 is an index between @xmath77 and @xmath73
. the generator of the @xmath76-th representation is given by @xmath78 .
therefore , it follows that the @xmath79-th entry of the @xmath76-th representation is @xmath80 ( the first entry is given by the identity ) . because of the unidimensionality of the representations the product of two representations , @xmath81 and @xmath82 , is the irreducible representation @xmath83 . if @xmath84 , @xmath83 is equivalent to @xmath85 .
thus each entry of the product table is again a representation @xmath75 with @xmath86 .
three conclusions can be drawn now : ( i ) from the definition of the representation generator we have that the complex conjugated representation of @xmath75 is @xmath87 , ( ii ) the representation @xmath88 has all characters equal to 1 ( this is the well known `` totally symmetric irreducible representation '' ) and it is present for every @xmath0 , ( iii ) when @xmath0 is even @xmath89 is the other real representation .
the description given so far in terms of group theory will prove particularly useful later on when we analyze the transitions between the field - dressed states of the molecule .
indeed , whether an electronic transition between two states , each of them corresponding to an irreducible representation , is allowed or not can be determined examining the product of the symmetries of the initial state , the inducing operator , and the complex conjugated of the final state .
if the product leads to the totally symmetric irreducible representation the transition is possible , i.e. , @xmath90 in our notation .
provided the laser pulse is sufficiently adiabatic and the emitting molecules can be considered uncorrelated it is a good approximation to calculate the harmonic spectrum from the fourier - transformed dipole expectation value . with the wavefunction
@xmath91 we can write the probability to get the @xmath7th harmonic as @xmath92 for a system whose hamiltonian is periodic in time with a period @xmath93 we can apply the floquet theorem ( see , e.g. , @xcite ) .
the wavefunction of a certain state of the system can be written as @xmath94 where @xmath95 is the so - called quasi - energy and @xmath96 .
the functions @xmath97 are eigenfunction of @xmath98 $ ] , i.e. , @xmath99 in what follows we refer to @xmath97 as a floquet state .
if the system is in a pure floquet state we can derive the selection rules as presented in @xcite .
however , due to pulse shape effects or other perturbations the system might be not in a pure floquet state , but rather in a superposition of them , @xmath100 here , for simplicity , we assume the @xmath101s constant ( or slowly varying ) in time .
the quasi - energies are given by @xmath102 where the real part @xmath103 is the field - dressed energy of the state and the imaginary part @xmath104 corresponds to half the ionization rate .
the latter would lead to finite harmonic line widths of lorentzian shape . in what follows
we will neglect this effect keeping only the real part of @xmath95 for our symmetry analysis .
( [ proba ] ) then becomes @xmath105 we introduce @xmath106 and the following operator @xmath107 noting that the functions @xmath97 form an extended hilbert space @xcite eq .
( [ expan ] ) can be rewritten as @xmath108 where the double brackets stand for integration over time and space .
we can think about the different operators @xmath109 as operators which induce transitions where the initial and the final states can be the same or different .
such transitions are accompanied by the emission of harmonics .
if the system is stable in one floquet state the expansion in ( [ expan2 ] ) reduces to only one term . in particular , when we consider the system being in the floquet ground state @xmath110 , @xmath111 is proportional to the fourier - transform of the expectation value of the dipole calculated with the floquet ground state , i.e. , @xmath112 from ( [ expan2 ] ) , the structure of the dipole spectrum can be derived .
because @xmath113 , @xmath114 and writing the operator @xmath115 explictly we have @xmath116 \\ & = & \hat{g}_\mp^{(i , j ) } \left [ \rme^{\rmi \frac{2\pi } { n}(n -\frac{\delta_i^j}{\omega } \mp 1 ) } \right ] .
\end{aligned}\ ] ] applying @xmath117 times @xmath118 to the operator @xmath119 we obtain @xmath120^m \,\ ,
\hat{g}_\mp^{(i , j ) } [ \hat{p}_n^{-1}]^m = \hat{g}_\mp^{(i , j ) } \left [ \rme^{\rmi \frac{2\pi m}{n}(n - \frac{\delta_i^j}{\omega } \mp 1 ) } \right ] .\ ] ] from a group theory point of view the term in square brackets in ( [ gener ] ) is simply the generator of the corresponding irreducible representation . from ( [ power ] ) we see how the harmonic generation operator behaves under the dynamical rotations of the symmetry group of the hamiltonian . given the irreducible representations of the states @xmath97 and @xmath121 we can derive to which representation the operator @xmath119 must correspond in order to satisfy eq.([cond ] ) . in representation notation , we must have @xmath122 , where @xmath123 is the irreducible representation of the operator @xmath119 . from the definition of the character table of an abelian group @xcite we know that @xmath124 is equal to the phase @xmath125 times @xmath121 and similar for @xmath97 .
also @xmath126 generates a phase . for allowing the emission of a harmonic the product of these three phases
must be the real unity .
consequently , the operator @xmath119 has to belong to the representation @xmath127 . imposing the term in square brackets of eq .
( [ gener ] ) to be equal to the generator of the representation @xmath128 leads to @xmath129 from this follows @xmath130 where @xmath131 .
introducing the extended index @xmath132 , i.e. , the symmetry index minus the field - dressed energy of the state in frequency units , we can rewrite ( [ rules ] ) in the final form@xmath133 where @xmath134 .
varying @xmath7 we force the generic operators @xmath109 to behave under the dynamical rotations according to the different representations @xmath75 of the symmetry group , and for each @xmath135 certain harmonics are emitted . for the special case @xmath136 one obtains @xmath5 , reproducing the results in @xcite .
they have been calculated taking @xmath137 and @xmath138 because the initial and the final state are the same .
@xmath136 behaves according to the totally symmetric representation @xmath88 .
it is worth to be noted that if the energy difference of the field - dressed states corresponding to the irreducible representations @xmath139 and @xmath140 is equal to @xmath141 we have a kind of accidental degeneracy : the lines corresponding to such transition can not be distinguished from those arising from the single state processes .
the operators @xmath109 can behave according to any representation .
hence all couplings between the floquet states of the molecule are in principle possible .
this of course does not mean that they have equal probability to occur .
considering that the ground state corresponds to @xmath88 , and the other @xmath75 represent excited states we can derive the following symmetry properties for the relevant single transitions : ( _ i _ ) @xmath142 : symmetry @xmath88 , ( _ ii _ ) @xmath143 : symmetry @xmath75 , ( _ iii _ ) @xmath144 : symmetry @xmath145 , ( _ iv _ ) @xmath146 : symmetry @xmath147 .
it is evident that when the frequency @xmath27 ( or one of its integer multiples ) of a finite laser pulse is nearly resonant with a transition to a certain state , this state becomes more easily accessible , and therefore transitions involving that state are more likely to happen . on the other hand ,
when the laser frequency is tuned off from all the resonances mostly the field - dressed ground state , i.e. @xmath88 , is involved and it will be the only one giving a significant contribution to harmonic generation . the fact that selection rules for different initial and final states contain the terms @xmath103 is particularly interesting for the case where the rotating field removes a degeneracy .
the measurement of the distance between the satellite lines arising from the two states gives an estimation of the field - dressed level - splitting .
the benzene molecule has a symmetry @xmath148 , but as long as we consider only functions which are all even or odd with respect to reflection at the molecule plane ( @xmath72 in the standard notation ) it is sufficient to consider only the symmetry operations of the reduced group @xmath149 .
the character table of the group @xmath150 is given in tab .
1 . connecting the four lowest states of the benzene molecule and the irreducible representations of the group @xmath150 ,
i.e. , @xmath151 , @xmath152 , @xmath153 , @xmath154 leads to the following energetic sequence : @xmath151 ( ground state ) , @xmath154 ( first excited , two - dimensional , i.e. , two - fold degenerated ) , @xmath153 ( second excited , two - fold degenerated ) , @xmath152 ( third excited , non - degenerated ) @xcite . in the ground state @xmath151 and
@xmath154 contain two and four @xmath69-electrons , respectively , and the other levels are empty @xcite .
as it was discussed in the previous section , when the circularly polarized laser field has an infinite duration , the total hamiltonian @xmath155 would belong to the dynamical symmetry group @xmath156 which is isomorphic to @xmath157 .
the character table of the groups @xmath157 and @xmath156 is shown in tab 2 .
for the group @xmath157 the standard notation is reported , for the group @xmath156 instead we use the notation that has been previously introduced . with the laser field
switched on the degeneracies connected to the representations @xmath154 and @xmath153 are removed . if we take for example the degenerated level @xmath158 of the field - free system , in an infinite pulse it is reducible , and knowing the irreducible representations of the group @xmath157 we can easily decompose it into irreducible parts : @xmath159 . the same argument can be applied to the level @xmath160 , @xmath161 .
what is worth to be stressed is that , contrary to the standard case in quantum mechanics textbooks , here the two representations which are complex conjugated , like @xmath162 , @xmath163 and @xmath164 , @xmath165 , correspond to two different energy levels and therefore the so - called " physically irreducible representations really generate two independent states .
this particular feature is due to the non - invariance of the total hamiltonian @xmath166 under a time inversion .
a similar case is obtained when a magnetic field is considered . at this point , with an oscillating field infinite in time , we can think about the field - dressed molecule states in pure terms of @xmath167 , where @xmath168 runs from @xmath77 to @xmath169 .
this system is stable , i.e. , there are no transitions between the states corresponding to different representations , unless external perturbations are present . the selection rules in this case can be derived from the product table , tab .
3 . when the pulse has a finite length we can consider the pulse shape as a perturbation of a system that has a @xmath150 symmetry .
a transition between two states is possible only if the operator inducing the transition has the `` proper '' symmetry . as it can be derived from tab .
1 the representations @xmath151 , @xmath154 , @xmath153 and @xmath152 correspond to rotations of angle @xmath77 , @xmath170 , @xmath171 and @xmath69 , respectively . in terms of @xmath76
they correspond to @xmath172 , @xmath173 , @xmath174 and @xmath175 .
in this section we illustrate and test our group theoretical method by applying it to harmonic spectra obtained with the help of _ ab initio _ numerical simulations . in order to keep the numerical effort feasible
we restrict ourselves to a two - dimensional ( 2d ) model where the molecule plane and the rotating electric field are properly represented .
we study a single active electron in a ring - shaped potential with @xmath4 symmetry ( @xmath176 in the benzene case ) .
results from a fully 3d calculation could be quantitatively different but the structure of the spectrum would be the same .
we use the potential @xcite @xmath177\ ] ] where @xmath178 is the radius of the molecule ( @xmath179 ) , and @xmath180 and @xmath181 are polar coordinates .
@xmath182 is the parameter which gives the `` degree of smoothness '' of the potential and determines the width of the ground state along the ring .
@xmath183 moves the local maxima of the potential keeping the minima constant ( the parameter @xmath183 is introduced in order to avoid the presence of points where @xmath184 for finite @xmath180 , because that could generate non - physical ionization ) .
finally , @xmath185 is the `` strength '' of the potential . for our simulations we chose @xmath186 and @xmath187 .
we used @xmath185 for tuning the ionization energy of the molecule . for more details about the model and its properties
see @xcite .
when the correct ionization energy for the benzene molecule is taken , the four states are relatively weakly bound and hence , as it has been already discussed in @xcite , for inducing a multiphoton process with a low ionization rate a low frequency is required . as the main aim of the numerical simulations presented in this work is to test _ ab initio _ the analytical derivations discussed in the previous sections , we chose @xmath188 which is less demanding for what concerns the run - time because a higher frequency can be used @xcite . if with @xmath189 we indicate the field - free energy of the level corresponding to the @xmath190 representation of the @xmath150 group , with @xmath188 the energies of the four states are : @xmath191 , @xmath192 , @xmath193 and @xmath194 , and the energy gaps between the excited states and the ground state are @xmath195 , @xmath196 and @xmath197 . in order to have a convenient way of referencing we introduce the notation @xmath198 for indicating a transition between the state @xmath199 and @xmath200 and its corresponding line in the spectrum .
@xmath201 is the @xmath76 which gives the transition symmetry .
the lower index distinguishes between the two signs in equation ( [ rules ] ) .
the upper indices specify the initial and the final state , respectively .
the order of the states in the upper index gives information about the sign of the line shift due to the distance between the states .
most of the non - single - state transitions discussed here connect an excited state with the ground state .
transitions between excited states are possible but much less likely to occur .
the position of the lines due to a `` transition '' from the ground to an excited state is @xmath202 , and for the opposite process @xmath203 , where @xmath103 is the distance between the ground state and the state @xmath199 .
when the laser is switched on the states move because of the dynamical stark effect and therefore @xmath103 is given by @xmath204 , where @xmath205 is the separation between the two states when no field is present and @xmath206 is the relative shift due to the presence of the oscillating field . at this point
, as we know the selection rules and the values of the field - free energy intervals @xmath205 , we can predict the structure of a spectrum where also non - single state processes are involved .
the only uncertainty is due to the values of the field shifts @xmath206 .
however , those shifts can be determined by performing several simulations with different laser intensities : when the field is low @xmath206 is small ( in the limit of zero field @xmath207 ) and the lines are located in the proximity of the position expected from the unperturbed level scheme . with increasing field the shift of the states can be followed .
those shifts can become relatively large but the interpretation of the spectrum remains always unambiguous .
an example of a spectrum , where also lines occurring from processes between different states are present , is shown in fig . 1 . there , a sine - square pulse of 40 cycles and frequency @xmath208 interacting with the model molecule has been simulated
the peak field strength @xmath209 was 0.14 .
it is possible to figure out for each peak the transition which generates it .
the same symbol indicates the same process but with different multiplier @xmath43 , i.e. , different number of photons .
a certain process is repeated every @xmath176 frequency units .
the states that together with the ground state play a role are the first and the second excited , i.e. , @xmath210 , and @xmath211 .
the first state generates transitions of type @xmath212 , the second state of type @xmath213 . with the parameters used for the simulation of fig .
1 the role of the second state is more important than that one of the first excited state .
this is mainly related to how strong the resonance with a particular state is . with a different frequency
the relative intensity of the lines connected to the two processes can differ strongly . from the fact that the degeneracies of the states @xmath210 and @xmath211
are removed one could expect more lines than present in fig . 1 .
let us discuss for example about the states @xmath211 , that in our case are the most important role players .
each of the states @xmath211 for a given @xmath43 generates two lines given by @xmath214 and two given by @xmath215 .
therefore eight lines are expected . however , using @xmath216 and @xmath217 to distinguish the two states @xmath153 , we find the shifts @xmath218 and @xmath219 .
considering that @xmath220 it follows that if the difference of the shifts @xmath206 due to the laser field is very small the eight lines are gathered in four couples , each of them containing two very close lines , difficult to resolve .
moreover , what makes the distinction between the two lines in a couple possible is not the absolute value of @xmath206 but the difference between the two shifts . as the two states without field are degenerate , this difference is the splitting generated by the field . with our notation
we have for the first couple of lines , e.g. , @xmath221 .
the same applies for the other three couples .
this explains why in fig .
1 only four lines due to the states @xmath153 are observed . in order to obtain a larger splitting ,
more easily detectable , we have to study how the shift @xmath206 behaves with respect to the different field parameters .
it was found that with lower frequencies the separation among the states which without field would be degenerate becomes more significant , as one would expect from a ponderomotive scaling .
moreover , it has been observed that the value of each @xmath206 can become relatively large compared to @xmath205 . in fig .
2 we show how the location of certain lines moves when the field is increased .
the dipole spectrum obtained with the lower frequency @xmath222 ( all other parameters the same as before ) is shown in fig .
3 . each of the lines in fig .
1 is splitted in two , confirming our predictions . in order to investigate the role of pulse shape effects simulations using trapezoidal pulses with different rampings
have been performed . in fig .
4 two spectra for pulses of the same length but with a @xmath223 cycle and @xmath224 cycle ramping , respectively , are reported .
it appears that in both case the overall structure is like in the case of the sine - square pulse .
it is worth to note that when the ramping is shorter , fig .
4b , the relevance of the `` extra '' lines with respect to the @xmath225-lines originating from the single state transitions is higher than in the case of longer ramping , fig 4a .
this is due to the fact that when the ramping is shorter the pulse is less adiabatic and therefore transitions between different floquet states are induced more likely . in the limit of an infinitely long ramping only the lines @xmath226 would be present .
we presented a general group theoretical approach to harmonic generation by systems possessing a discrete rotational symmetry @xmath227 , in particular a ring - shaped molecule in a circularly polarized laser field ( e.g. , benzene ) . in the simple case where
the initial and the final field - dressed electronic states are the same the known selection rule @xmath9 , @xmath3 was obtained .
we demonstrated that this selection rule follows also from a change of reference frame as well as angular momentum conservation in a straight forward manner .
since due to this selection rule less harmonics are allowed within a fixed frequency range such systems might be beneficial for generating short wavelength radiation more efficiently , i.e. , without wasting a vast amount of laser energy in undesired laser harmonics .
however , we showed that in general other harmonics different from the expected ones at the positions @xmath228 are present as well .
those lines carry important information about the laser field - dressed level scheme of the molecule since energy differences between different states are involved .
transitions between field - dressed states occur because the laser pulse has a temporal shape instead of being infinite as assumed in the derivation of the @xmath9-selection rule . for the case of the ring - shaped molecule we showed that our group theoretical method is capable to predict also the position of those extra lines .
this was demonstrated with the help of harmonic spectra obtained by an _ ab initio _ numerical simulation of a benzene model molecule interacting with a finite laser pulse .
the extra harmonics are particularly pronounced in the cases where the laser frequency ( or a multiple of it ) becomes resonant with electronic transitions between field - dressed states .
harmonic peak splittings and shifts were fully understood in the framework of the group theory approach . from the symmetry point of view
the case of a cyclic molecule in a circularly polarized laser field is equivalent to a single atom in a two - color field with one frequency an integer multiple of the other .
of course , the groups representing the unperturbed systems are different .
99 p salires , a lhuillier , p antoine and m lewenstein , adv . at . , mol .
, opt . , phys . * 41 * , 83 ( 1999 ) ; c j joachain , m drr , and n j kylstra , adv . at .
phys . * 42 * , 225 ( 2000 ) ; t brabec , and f krausz , rev .
phys . * 72 * , 545 ( 2000 ) .
o alon , v averbukh and n moiseyev , phys .
lett . * 80 * , 3743 ( 1998 ) . o alon , v averbukh , n moiseyev , phys .
lett . * 85 * , 5218 ( 2000 ) .
f ceccherini and d bauer , phys .
a * 64 * , 033423 ( 2001 ) .
f ceccherini _ et al .
_ , in preparation .
h eichmann , a egbert , s nolte , c momma , b wellegehausen , w becker , s long and j k mciver , phys .
rev . a , * 51 * r3414 ( 1995 ) d b milosevic , w becker and r kopold , phys .
a * 61 * 063403 ( 2000 ) .
v averbukh , o e alon and n moiseyev phys .
a * 60 * , 2585 ( 1999 ) .
x -m tong and s -i chu , phys .
a * 58 * r2656 ( 1998 ) . s long , w becker and j k mciver , phys .
a * 52 * , 2262 ( 1995 ) .
d b milosevic in _
atoms , molecules and quantum dots in laser fields : fundamental processes _ , edited by n. bloembergen et .
al , italian physical society , ed .
compositori , bologna , 2001 .
l d landau , e m lifshitz , _ quantum mechanics , non - relativistic theory _ , ( pergamon press , oxford , 1977 ) .
michael tinkham , _ group theory and quantum mechanics _ , ( mcgraw - hill book company , new york , 1964 ) .
f h m faisal , _ theory of multiphoton processes _ ( plenum press , new york , 1987 ) .
h sambe , phys .
rev . a * 7 * , 2203 ( 1973 ) .
j s lomont , _ applications of finite groups _ , ( accademic press , new york , 1959 ) .
@r|rrrrrr @xmath150 & @xmath229 & @xmath230 & @xmath231 & @xmath232 & @xmath233 & @xmath234 @xmath151 & @xmath235 & @xmath235 & @xmath235 & @xmath235 & @xmath235 & @xmath235 @xmath236 & @xmath235 & @xmath235 & @xmath235 & @xmath235 & @xmath237 & @xmath237 @xmath152 & @xmath235 & @xmath237 & @xmath235 & @xmath237 & @xmath235 & @xmath237 @xmath238 & @xmath235 & @xmath237 & @xmath235 & @xmath237 & @xmath237 & @xmath235 @xmath153 & @xmath239 & @xmath239 & @xmath237 & @xmath237 & @xmath77 & @xmath77 @xmath154 & @xmath239 & @xmath240 & @xmath237 & @xmath235 & @xmath77 & @xmath77 @rr|rrrrrr @xmath156 & @xmath157 & @xmath229 & @xmath241 & @xmath242 & @xmath230 & @xmath243 & @xmath244 @xmath88 & a & @xmath235 & @xmath235 & @xmath235 & @xmath235 & @xmath235 & @xmath235 + @xmath245 & b & @xmath235 & @xmath237 & @xmath235 & @xmath237 & @xmath235 & @xmath237 + @xmath246 & @xmath247 & @xmath235 & @xmath27 & @xmath248 & @xmath249 & @xmath250 & @xmath251 + @xmath252 & @xmath253 & @xmath235 & @xmath251 & @xmath250 & @xmath249 & @xmath248 & @xmath27 + @xmath254 & @xmath216 & @xmath235 & @xmath248 & @xmath250 & @xmath235 & @xmath248 & @xmath250 + @xmath255 & @xmath217 & @xmath235 & @xmath250 & @xmath248 & @xmath235 & @xmath250 & @xmath248 + @c|rrrrrr @xmath256 & @xmath88 & @xmath246 & @xmath252 & @xmath254 & @xmath255 & @xmath245 @xmath88 & @xmath88 & @xmath246 & @xmath252 & @xmath254 & @xmath255 & @xmath245 + @xmath246 & @xmath246 & @xmath254 & @xmath88 & @xmath245 & @xmath252 & @xmath255 + @xmath252 & @xmath252 & @xmath88 & @xmath255 & @xmath246 & @xmath245 & @xmath254 + @xmath254 & @xmath254 & @xmath245 & @xmath246 & @xmath255 & @xmath88 & @xmath252 + @xmath255 & @xmath255 & @xmath252 & @xmath245 & @xmath88 & @xmath254 & @xmath246 + @xmath245 & @xmath245 & @xmath255 & @xmath254 & @xmath252 & @xmath254 & @xmath88 + | we discuss harmonic generation in the case of laser field - dressed hamiltonians that are invariant under so - called dynamical symmetry operations .
examples for such systems are molecules which exhibit a discrete rotational symmetry of order @xmath0 ( e.g. benzene with @xmath1 ) interacting with a circularly polarized laser field and single atoms in a bichromatic field , with the two lasers having circular polarizations . within a general group theory approach
we study the harmonics one obtains from the interaction of a laser pulse and a circular molecule .
when the system is in a pure field - dressed state the known selection rule @xmath2 , @xmath3 results .
however , other lines are observed when recombinations with states of a symmetry different from the initial one become important .
this is the case for realistic laser pulses ( i.e. , with a finite duration ) . in particular when the fundamental laser frequency ( or one of its multiples ) is resonant with a transition between field - dressed states .
numerical _ ab initio _
simulations , confirming our analytical calculations and illustrating the power of the group theory approach , are presented . |
elliptical galaxies in clusters are old .
dispersions of colours around colour - magnitude ( _ c - m _ ) relation of cluster ellipticals are small at low and intermediate redshifts ( bower , lucey & ellis 1992b ; ellis 1997 ) .
this strongly suggests that almost all stars in ellipticals formed at a high redshift , at least @xmath3@xmath4 .
colour change of ` red - envelope ' galaxies in high redshift clusters suggests that these galaxies formed at @xmath3 and evolved passively since then ( aragn - salamanca 1993 ) .
a similar implication was given by mg@xmath5 indices of ellipticals in clusters at intermediate redshift ( bender , ziegler & bruzual 1996 ) .
recent analyses of the _ c - m _ relations of cluster e / s0s at high redshift , all of which are morphologically classified by hst imaging , also require that ellipticals should form at @xmath3 ( kodama 1997 ; stanford , eisenhardt & dickinson 1997 ) .
an implication derived from these observational evidence is in full agreement with a picture predicted by a galactic wind model for the formation of elliptical galaxies ( larson 1974 ; arimoto & yoshii 1987 ; kodama & arimoto 1997 ) .
elliptical galaxies should form by dissipative collapse of proto - galactic massive cloud , during which stars were born intensively and the star formation ceased when the gas in a galaxy was expelled due to the super - novae driven galactic wind .
galaxies then evolved passively without experiencing any significant event of star formation .
nevertheless , there is a growing piece of evidence showing that elliptical galaxies , in particular in the low density environment , might have experienced secondary formation of stars in the recent past ( e.g. , schweizer 1990 ; schweizer & seitzer 1992 ; rose 1994 ; bressan , chiosi & tantalo 1996 ; barger 1996 ) .
schweizer & seitzer ( 1992 ) analysed colours and line indices of e / s0 galaxies in the field as well as in groups and found that @xmath6 and @xmath7 become systematically bluer and h@xmath1 indices become stronger as the number of fine structure increases , which suggests that the secondary episode of star formation is related to dynamical disturbances of galaxies , probably caused by mergers and/or interactions with other galaxies ( see also schweizer 1990 ) .
hierarchical clustering models of galaxy formation based on the cdm hypothesis ( kauffmann , white & guiderdoni 1993 ; cole 1994 ) suggest that the star formation took place recently in elliptical galaxies due to galaxy merging .
the cdm based models predict significant fraction of star burst at @xmath8 in elliptical galaxies especially in the field . with a help of population synthesis model by worthey ( 1994 ) , de jong & davies ( 1996 )
have recently analysed the h@xmath1 and [ mgfe ] indices of elliptical galaxies , spectra of which were taken by gonzlez ( 1993 ) .
most of galaxies are the members of small groups .
they estimate the amount of star burst that is required to reproduce the h@xmath1 and disk - to - total light ratio ( d / t ) .
the resulting fraction of young stars is larger than 10 % if they are 2 gyr old .
barger ( 1996 ) demonstrate that at least 30 % of cluster galaxies in ac 103 , ac 114 , and ac 118 ( @xmath9 ) , seem to have undergone a secondary burst of star formation during the last @xmath10 2 gyr prior to the epoch of observation , though many of them are regular e / s0s .
this implies that the frequency of secondary burst is very high even for cluster ellipticals , where the galaxy - galaxy merger is supposed to be less frequent than in the field .
independently , broad band colours can put much stronger constraints on the nature of secondary star burst in cluster e / s0s .
bower ( 1992b ) analysed @xmath0 colour dispersions around the _ c - m _ relations of coma and virgo clusters , and found that the amount of mass involved in the secondary burst should be less than 10 % if the age of burst population is 5 gyr .
if a younger age of 2 gyr is assumed , the burst population should be less than only a few percent .
the problem arising here is an apparent inconsistency between the two types of analyses ; the line indices against the colour dispersion .
the lines always indicate larger fraction of the secondary burst population .
recent studies tend to heavily rely on the line indices data to derive the contribution of the secondary burst population ( e.g. , worthey 1994 ; de jong & davies 1996 ) .
however , the burst strength thus derived should also explain the observed @xmath0 , since at least the synthesized broad band colours are much more reliable than the line indices which are usually calculated independently of colours by using empirical polynomial fitting functions ( e.g. , worthey 1994 ; vazdekis et al .
1997 ) . in this paper
, therefore , we analyse @xmath0 and h@xmath1 indices of ellipticals in gonzlez ( 1993 ) sample , and derive new constraints on the amount of young population . in
2 we compile the data used in this study . in
3 we describe the method to compute models for elliptical galaxies with the secondary star burst .
the resulting constraints for the star formation history in ellipticals are given in 4 . in
5 we discuss possible causes for the discrepancy between @xmath0 and h@xmath1 analysis , and conclude this paper .
we have analysed a whole sample of 41 elliptical galaxies to which gonzlez ( 1993 ) gives measurements of the conspicuous lick line strengths .
h@xmath1 line strengths ( at @xmath11 ) and central velocity dispersions @xmath12 ( within the inner 5 arcsec ) are taken from gonzlez ( 1993 ) and broad band colours are taken from rc3 ( de vaucouleurs et al .
following schweizer & seitzer ( 1992 ) , we define @xmath13 as @xmath14,\ ] ] where subscript @xmath15 refers to global colours , subscript @xmath16 indicates colours corrected for the extinction due to interstellar dust in our galaxy as well as for the redshift , and subscript @xmath17 refers to average colours within an effective radius .
internal extinction is neglected because they are not given in rc3 but an argument is given later on the possible influence of the internal extinction .
observational errors are taken into account in @xmath18 , but errors involved in the correction procedure in deriving @xmath13 are not included because they are not given in rc3 . therefore the errors of @xmath13 could be substantially larger than those given in figs .
[ fig : uv],[fig : uv_tb2 ] , [ fig : uv_hbeta_tb2 ] , and [ fig : uv_av_tb2 ] .
this would not affect our arguments significantly , however , since our main concern is to understand why the dispersion in colours is so small along the @xmath13 versus @xmath19 relation ( see fig . [
fig : uv ] ) while some galaxies are showing strong evidence of young stars in the measured h@xmath1 lines .
errors in the velocity dispersion are very small ( @xmath10 1 % ; gonzlez 1993 ) , and thus are ignored .
1.0 the 10 galaxies among gonzlez s sample show strong h@xmath1 absorption , which may be due to contamination of some young stars in these galaxies .
in analogy to h@xmath20 strong galaxies defined by couch & sharples ( 1987 ) , we define h@xmath1 strong galaxies as those having equivalent width ew(h@xmath1 ) larger than 2 . gonzlez ( 1993 )
gives h@xmath1 indices for three different apertures ; nuclear 5 arcsec , @xmath21 , and @xmath22 , where @xmath23 is the effective radius of elliptical galaxies .
we adopt the @xmath22 aperture for h@xmath1 indices , since it is the closest to @xmath24 within which the colours in rc3 are defined .
this would minimize the effect of h@xmath1 line strength gradient which is sometimes shown for elliptical galaxies ( gorgas , efstathiou & aragn - salamanca 1990 ; davies , sadler & peletier 1993 ; gonzlez 1993 ) . figure [ fig : hbeta ] shows the h@xmath1 indice versus central velocity dispersion ( @xmath25 ) for the gonzlez s whole sample .
the filled circles represent the h@xmath1 strong galaxies .
the ngc number is attached to each of them .
the solid line corresponds to a sequence of model ellipticals ( see 3 ) .
note that the h@xmath1 strong galaxies are away from the sequence more than 0.3 .
1.0 figure [ fig : uv ] shows the @xmath13 versus @xmath25 ( km s@xmath26 ) diagram for 33 ellipticals .
8 galaxies among gonzlez s sample are not shown in fig .
[ fig : uv ] because @xmath13 are not available in rc3 .
the h@xmath1 strong galaxies are indicated by filled circles .
we note that a well defined correlation between @xmath13 and @xmath25 is clearly seen ( the _ c_-@xmath27 relation ) in this diagram .
the solid line gives a regression line for the sampled galaxies .
the regression line to elliptical galaxies in coma and virgo clusters ( bower 1992b ) is also shown by the dot - dashed line . since the @xmath28-band filter used by bower , lucey & ellis ( 1992a ) is slightly different from the standard johnson system , we apply small corrections , although it is negligibly small compared to the errors in colours ( @xmath29 mag ) . in spite of a difference in the aperture size between the two photometries , agreements both in slope and zero - point are remarkable .
the standard deviations are also comparable ; 0.046 mag for bower ( 1992b ) and 0.054 mag for our sample .
true that most of the gonzlez galaxies locate in small groups , yet they have similar _
c_-@xmath27 relation to that of cluster ellipticals .
considering the fact that cluster ellipticals have an almost universal metallicity sequence as a function of galaxy luminosity or size ( kodama et al .
1998 ) , the similarity in the _ c_-@xmath27 relation we see above suggests that gonzlez sample would also share the same metallicity sequence .
this reinforces our coming assumption that the colour of the underlying old population in ellipticals should locate along the _
c_-@xmath27 relation thus defined ( see 3.2 ) .
although larson , tinsley , & caldwell ( 1980 ) claimed that the _ c - m _ relation of field ellipticals are less tight than that of cluster ellipticals , we do not find this in the case for the gonzlez galaxies .
it is clearly needed to settle this problem with new systematic photometry data of nearby field ellipticals .
the other notable thing is that the h@xmath1 strong galaxies are not always bluer in colours compared to the _
c_-@xmath27 relation , although they are expected to be appreciably bluer if suffering from the secondary burst of star formation recently .
the 5 galaxies out of 10 h@xmath1 strong galaxies , ngc 507 , ngc 584 , ngc 1700 , ngc 3377 , and ngc 7454 , locate below ( bluer ) the line of _ c_-@xmath27 relation in fig . [
fig : uv ] .
these galaxies are likely to have considerable number of young stars .
however , except ngc 3377 , these galaxies are bluer no more than 0.1 mag , which indicates little contamination of young stars to the galactic light .
the h@xmath1 strong galaxies should contain certain amount of young stars and yet some of them do not show any evidence in colours for young stars . in the following 3
, we build a series of models for elliptical galaxies with the young population superposed onto the underlying old populations and examine how much young stars can be in hiding to give compatible h@xmath1 line strengths while keeping the _ c_-@xmath27 relation tight .
we build models for elliptical galaxies , having experienced the secondary burst of star formation , by superposing the young stellar population onto the old underlying populations . for the underlying populations ,
we adopt the metallicity sequence models of elliptical galaxies calculated by kodama & arimoto ( 1997 ) and kodama ( 1997 ) .
these models are calibrated to reproduce the _ c - m _ relation of elliptical galaxies in coma cluster by changing a galactic wind epoch as a function of initial mass of a galaxy .
this is equivalent to change the mean stellar metallicity of galaxies along the _ c - m _ relation .
the age of the galaxies is assumed to be 12 gyr .
all stars formed very quickly before the wind ( @xmath300.5 gyr ) .
the time scale of star formation is assumed to be 0.1 gyr .
a salpeter - like initial mass function ( imf ) is assumed with a slope @xmath31 ( the salpeter imf corresponds to @xmath32 in our definition ) .
the lower and the upper mass cutoffs @xmath33 and @xmath34 are set to be 0.1 and 60 @xmath35 , respectively .
the mean stellar metallicity @xmath36 decreases from 0.0 to @xmath370.5 for the range of 6 magnitudes covered by the _ c - m _ relation ( @xmath38 to @xmath39 mag ) .
chemical evolution is taken into account under the so - called infall model scheme ( kppen & arimoto 1990 ) .
the models are shown to match very well with the observed _ c - m _ relations of elliptical galaxies in distant clusters ( kodama & arimoto 1997 ; kodama et al . 1998 ) .
the velocity dispersion is assigned to each model galaxy with a help of the @xmath13 versus @xmath40 relation defined in 2.3 ( solid line in fig . [ fig : uv ] ) as : @xmath41,\ ] ] where we use theoretical @xmath0 for @xmath13 .
the lick spectral indices are calculated by using the polynomial analytical fits given by worthey , faber & gonzlez ( 1992 ) .
the synthesized h@xmath1 indices of the underlying galaxy models with various initial masses are shown in fig .
[ fig : hbeta ] by the solid line .
1.0 we assume that the secondary burst happened instantaneously .
in such a case , the spectrum of the burst population is approximated by the so - called simple stellar population ( ssp ) model ( buzzoni 1989 ) .
thus we ignore the metal enrichment during the secondary burst .
the imf slope of the secondary population is assumed to be @xmath32 ( the salpeter imf ) . @xmath33 and @xmath34 are the same as those of the underlying population .
note we adopt different imf slope for the burst popualtion from that of the underlying population , simply because there is no reliable estimate available .
however , our conclusion would not be affected by this choice of imf slope .
if we were to use @xmath31 for the burst population instead , the acceptable burst strength would become even smaller due to more numerous young stars .
therefore , our coming conclusion would be more strengthened .
the evolution of @xmath0 and h@xmath1 index of the ssp models are illustrated in fig .
[ fig : uv_hbeta ] .
the thick dashed lines and the thick solid ones represent the ssp models with @xmath42 and @xmath43 , respectively .
the thin lines represent various ssp models with the solar abundance ( @xmath44 0.02 ) available in the literature ( worthey 1994 ; bressan et al .
1996 ; vazdekis et al .
the age of the burst population @xmath45 is varied from 0.1 gyr to 5 gyr as a free parameter .
if the secondary burst is induced by capturing a small gas - rich galaxy , as is most likely , the mean stellar metallicity of the burst population should not exceed the solar metallicity , since typical metallicities of dwarf irregulars are well below the solar value ( skillman , kennicutt & hodge 1989 ) . even if the chemical enhancement during the burst is considered , the average metallicity of the bust population hardly exceeds the solar value unless the imf slope is significantly flatter than the salpeter imf
. therefore we adopt @xmath46 and @xmath47 as the representative metallicities of the secondary burst .
the burst strength ( @xmath48 ) is defined by a _
mass _ fraction of burst population to the whole galaxy at the present epoch ( 12 gyr ) , and are allowed to vary from 0.1 % to 20 % .
we assume that the velocity dispersion does not change before and after the burst , since the mass involved in the secondary burst is usually small compared to that of the underlying population .
even if the fraction of the burst is as high as 20 % in mass , the resulting increase of the velocity dispersion is at most @xmath49 if the galaxy is in virial equilibrium before and after the burst .
if the dark matter dominate the galaxy potential , the influence of the burst population to the velocity dispersion should be less significant . along the _
c_-@xmath27 relation , the young ssp model is superposed in such a way that the mass fraction of young population to the total mass equals to a given burst fraction @xmath48 .
the @xmath0 colour and h@xmath1 index of a total galaxy are calculated from the resulting composite spectrum .
1.0 1.0 figure [ fig : uv_tb2 ] illustrates the colour change due to an increase of the burst strength in the @xmath13 versus @xmath25 diagram .
burst age @xmath45 is fixed to be 2 gyr .
the open and filled circles show the same galaxies as in fig . [
fig : uv ] , and the dotted line gives the regression line for the data , along which we assign the underlying galaxies .
the dashed line and the solid one correspond to the metallicity of the burst @xmath500.008 and 0.02 , respectively .
the burst strength @xmath48 is increasing from the top to the bottom . apart from one exception , ngc 3377 , the maximum burst strength allowed to be consistent with the data is less than @xmath10 7 % for @xmath512 gyr . in the same way
, models with various combination of @xmath45 , @xmath48 , and @xmath52 are confronted to the data and the upper limit of the burst strength is investigated for each case . the results are summarized in table 1 . the acceptable amount of colour change @xmath53 due to the young population is set to @xmath370.1 at log @xmath54 .
the maximum burst strength @xmath48 is adjusted in such a way that @xmath53 reaches this limit for a given set of @xmath45 and @xmath52 . as a result
, we find that the permitted range of the burst strength should be very small ; typically less than a few percent in mass , at most 47 % within recent 2 gyr .
[ table-1 ] [ cols="<,<,<,^,^,^,^ " , ] the above number is considerably smaller than that given by de jong & davies ( 1996 ) , who obtained 1015 % young population for @xmath55 gyr burst to reproduce the strong h@xmath1 line strengths ( @xmath2 ) . independently , as described in 3 , we calculated h@xmath1 indices for the secondary star burst models to see the consistency .
the results are shown in fig .
[ fig : hbeta_tb2 ] for the case of @xmath562 gyr .
it is clearly seen that the h@xmath1 strong galaxies can only be reproduced by adding more than 1015 % young star burst , which is consistent with de jong & davies ( 1996 ) . however , this number is more than twice as much as that expected from the @xmath0 analysis presented above .
we also estimated h@xmath1 values that correspond to @xmath48 values constrained from @xmath53 .
the results are tabulated in the fifth column of table 1 . as a consequence ,
the h@xmath1 indices are always smaller than 2 .
thus the models constrained from the @xmath0 colours fail to reproduce the h@xmath1 strong galaxies ( h@xmath1 @xmath57 2 ) .
since the strong balmer absorption lines are characteristic of a - type stars , the contribution from such stars becomes maximum at around @xmath10 0.20.5 gyr after the onset of burst ( fig .
[ fig : uv_hbeta ] ) .
however , even with @xmath58 0.20.5 gyr , the resulting h@xmath1 indices are again less than 2 .
true one can obtain stronger h@xmath1 strengths by imposing much larger burst strength @xmath59 , but the resulting galaxy becomes too blue in @xmath0 .
we will discuss this @xmath0 versus h@xmath1 discrepancy later .
if the age of secondary burst is as young as 0.51 gyr as suggested from the numerical simulation of shells around dynamically disturbed ellipticals ( hernquist & quinn 1987 ) , the burst strength should be quite small , such as less than 12 % .
conversely , if we assume older burst ages , the limit of @xmath48 becomes larger ( more than 20 % for @xmath515 gyr , for example ) , but @xmath60
- h@xmath1 discrepancy becomes larger . for reference , the @xmath28- and @xmath61-band luminosity weighted mean age of stars in a galaxy @xmath62
are also given in table 1 , which are defined as : @xmath63 where summation is taken for all stars in a galaxy .
thus , the lower limit of @xmath64 is 7.39.1 gyr and that of @xmath65 is 8.410.5 gyr .
these are much older than the age of bulk of ellipticals that worthey , trager & faber ( 1996 ) assigned from the balmer line indices .
these authors find significant number of extremely young elliptical galaxies with age only 23 gyr .
the mean age is also quite young such as 45 gyr , nearly half of what we find .
the discrepancy of the results between the analysis based on the colours and the one based on the line indices can be seen here as well . in conclusion , apart from ngc 3377 , which could be the only candidate that might actually have high burst strength of upto 1520 % for @xmath55 gyr and could be consistent with its h@xmath1 strength ,
the intensity of the secondary star burst should be less than only a few percent in mass to be compatible with the observed @xmath0 .
it is not strong enough to explain the h@xmath1 strong galaxies , however .
so far we have calibrated the models of the underlying old populations using the linear regression line of the data on the _
c_-@xmath27 diagram , supported by the existence of mass - metallicity relation for elliptical galaxies as we discussed in 2.3 .
however , one could alternatively take the location of the underlying population as the reddest envelope of the sample galaxies , if all galaxies bluer than the red envelope are regarded as being contaminated by young stars .
although we believe this is unlikely as discussed above , we estimate how much the upper limit of @xmath59 could be increased if this effect is the case .
if we take the red envelope at 0.07 mag redder than the regression line , the acceptable burst strength increases to @xmath6610 % .
therefore , even in this case , the maximum burst strength in mass fraction should be less than 10 % at most , within recent 2 gyr .
1.0 furthermore , one may claim that low mass systems have more often disky shapes and can contain more young stellar populations ( e.g. , kormendy and bender 1996 ) , therefore the slope of the _ c_-@xmath27 relation of the underlying old population might possibly be flattened at the faint end than that defined in this paper .
this is equivalent to weaken the metallicity variation as a function of galaxy mass and to introduce some age difference instead .
in fact , as shown in fig .
[ fig : uv_hbeta_tb2 ] where the sample galaxies are plotted in the @xmath0 versus h@xmath1 diagram with the secondary burst models superposed , the four h@xmath1 strong ellipticals with low dynamical masses ( @xmath67 ) could be explained consistently both in @xmath0 and h@xmath1 at the same time with some combination of metallicity and burst strength , if one could neglect the _
c_-@xmath27 relation .
the other h@xmath1 strong ellipticals with larger masses locate out of the model grid , and the @xmath60-h@xmath1 discrepancy is quite evident for these galaxies in anycase . however , if lower mass systems tend to be more contaminated by young stars , the slope of the _ c_-@xmath27 relation and hence the _ c - m _ relation is expected to vary from cluster to cluster , given the colour is very sensitive to a small difference of the amount of young stars .
this trend is far from what we actually observe both locally and at high redshifts ( garilli et al .
1996 , stanford et al .
1997 , kodama et al .
1998 , lpez - cruz 1997 ) .
it has been shown that most of clusters have a common slope when compared to coma or virgo cluster . even if it were the case that low mass galaxies would contain more young stars , the @xmath60-h@xmath1 discrepancy could not be compensated . in order to be consistent with strong h@xmath1 indices of these low mass galaxies ,
the secondary star burst of 1520 % in mass is required with @xmath68 gyr .
if they were to have such strong star burst ( 15 % ) , the colour of the underlying population should be @xmath69 , just comparable to the typical colour of the rest of ellipticals , and we would no longer see the _
c_-@xmath27 relation for the underlying old population .
the 15 % burst with @xmath68 gyr gives luminosity weighted ages of @xmath70 and @xmath71 , which are 67 gyr younger than the bulk of more massive ellipticals ( which corresponds to about 4 magnitude brighter using the fabor - jackson law ) .
as expected , the 67 gyr difference in 4 magnitude range of _ c - m _ relation is almost comparable to a pure age sequence discussed in kodama & arimoto ( 1997 ) and kodama et al .
( 1998 ) , which is absolutely rejected by the evolution of _ c - m _ slope as a function of redshift . therefore ,
the observed _ c - m _ relation of clusters does not support the idea xthat the low mass h@xmath1 strong ellipticals have experienced much stronger secondary burst than our estimate .
after all , the @xmath60-h@xmath1 discrepancy does remain ; i.e. , the integrated @xmath0 colour of gonzalez s sample of ellipticals permit only 10% of the secondary stellar population at most , while the strong h@xmath1 indices suggest much larger burst strengths .
we here consider possible reasons for this discrepancy .
\(1 ) _ model uncertainty_. one possible reason is that the h@xmath1 synthesis model based on the empirical calibration of worthey ( 1992 ) may be still premature partly because the scatter around the polynomial fitting lines might be too large and/or because the zero - point of the model might be uncertain .
@xmath0 could have uncertainties as well ( charlot , worthey & bressan 1996 ) , but they are expected to be small at ages 12 gyr or younger , since at such young ages , @xmath0 colour is driven primarily by the evolution of the main sequence turn - off stars , and is relatively simple to model .
( 2 ) _ emission correction_. gonzlez ( 1993 ) corrected the contamination of h@xmath1 emission for the equivalent width of h@xmath1 absorption by using the intensity of [ oii ] emission .
although this empirical procedure is reasonable , it causes uncertainty in the h@xmath1 indices . if the correction tends to be overestimated , the real h@xmath1 absorption could be much lower and consistent with @xmath0 analysis .
\(3 ) _ aperture effect_. h@xmath1 indices are defined in the @xmath72 aperture , while the colours are defined in the @xmath23 aperture .
if all stars of the secondary population are formed locally within @xmath72 , the allowed burst strength could be 1.56 times larger within @xmath72 following de vaucouleurs ( 1948 ) @xmath73-law . this could rise the model h@xmath1 indices by about 0.10.15 , and it could marginally make up the discrepancy . according to davies , sadler & peletier ( 1993 ) , however , h@xmath1 absorption is constant or increasing with a radius in almost all cases of elliptical galaxies in their sample . though this could be partly due to the dilution of the h@xmath1 absorption feature by emission in the centers of galaxies ( davies et al .
1993 ) , h@xmath1 would be hardly changed between @xmath72 and @xmath23 aperture .
\(4 ) _ internal dust extinction_. the internal dust extinction is expected to have more effect on @xmath0 colour than on h@xmath1 indices .
it would redden @xmath0 of galaxies without reducing h@xmath1 indices much and could dissolve the discrepancy . to see the effect in some more detail ,
the internal extinction is taken into account to the model spectra in a simple way .
the extinction law is taken from mathis ( 1990 ) .
we assumed that the dust distribution in the galaxies is localized only in the region where the secondary burst occurred .
hence we applied the extinction correction only for the burst population , and not for the old underlying population . in figures [ fig : uv_av_tb2 ] and [ fig : hbeta_av_tb2 ] , the effect of extinction is shown for @xmath0 and h@xmath1 , respectively .
@xmath74 is changed as 0.0 , 0.5 , and 1.0 mag . the model with @xmath59=20 % and @xmath74=0.5 mag , for example
, could possibly reproduce some of the h@xmath1 strong galaxies both in @xmath0 and h@xmath1 .
the effect of internal extinction certainly allows a higher upper limit to the burst strength .
a study of infrared colours of gonzlez sample would help since the sed at longer wavelengths is less vulnerable to the dust extinction .
1.0 .
1.0 in summary , although considerable dust extinction for burst population could solve the discrepancy between @xmath0 and h@xmath1 analyses , we rely on @xmath0 at the moment , considering that the h@xmath1 might have great uncertainties both in models and in observational data . in this case
, we can put strong constraint on the secondary episode of star formation in ellipticals : elliptical galaxies , regardless of their environments , experience little secondary star formation in the recent past .
for example , the mass fraction of the young burst population should be quite small , typically less than 47 % , and 10 % at most within recent 2 gyr , this suggests that the dynamical disturbance found in some ellipticals is more likely to be caused by a capture of dwarf galaxy with a mass of order @xmath75@xmath76 , rather than a merging of two galaxies of comparable size .
if ellipticals formed by the recent mergings of comparable galaxies , as hierarchical clustering models of the universe suggest , it should have proceeded without any significant star formation .
however , it is yet to be studied in detail how such process can create the observed _ c - m _ relation of ellipticals .
we would like to thank a. aragn - salamanca , r.l .
davies , r.g .
bower , r.s
. de jong , h. kuntschner , t.j .
ponman , and d.a .
forbes for very useful discussions .
thanks the japan society for promotion of science ( jsps ) postdoctoral fellowships for research abroad for supporting his stay in the institute of astronomy , cambridge , uk .
thanks the jsps for supporting his stay in observatoire de paris , section de meudon , france .
this work was financially supported in part by a grant - in - aid for the scientific research ( no.09640311 ) by the japanese ministry of education , culture , sports and science .
aragn - salamanca a. , ellis r. s. , couch w. j. , carter d. , 1993 , mnras , 262 , 764 arimoto n. , yoshii y. 1987 , a&a , 173 , 23 barger a. j. , aragn - salamanca a. , ellis r. s. , couch w. j. , smail i. , sharples r. m. , 1996 , mnras , 279 , 1 bender r. , ziegler b. , bruzual a. g. , 1996 , apj , 463 , l51 bower r. g. , lucey j. r. , ellis r. s. , 1992a , mnras , 254 , 589 bower r. g. , lucey j. r. , ellis r. s. , 1992b , mnras , 254 , 601 bressan a. , chiosi c. , tantalo r. , 1996 , a&a , 311 , 425 buzzoni a. , 1989 , apjs , 71 , 817 charlot s. , worthey g. , bressan a. , 1996 , apj , 457 , 625 cole s. , aragn - salamanca a. , frenk c. s. , navarro j. f. , zepf s. e. , 1994 , mnras , 271 , 781 couch w. j. , sharples r. m. , 1987 , mnras , 229 , 423 davies r. l. , sadler e. m. , peletier r. f. , 1993 , mnras , 262 , 650 de jong r. s. , davies r. l. , 1997 , mnras , 285 , l1 de vaucouleurs g. , 1948 , ann .
dastrophys . , 11 , 247 de vaucouleurs g. , de vaucouleurs a. , corwin h. g. , buta r. j. , paturel g. , fouqu p. , 1991
, third reference catalog of bright galaxies ( springer;berlin ) ( rc3 ) ellis r. s. , smail i. , dressler a. , couch w. c. , oemler jr a. o. , butcher h. , sharples r. m. , 1997 , apj , 483 , 582 garilli b. , bottani d. , maccagni d. , carrasco l. , recillas e. , 1996 , apjs , 105 , 191 gonzlez j. j. , 1993 , ph.d .
thesis , univ . of california , santa cruz gorgas j. , efstathiou g. , aragn - salamanca a , 1990 , mnras , 245 , 217 hernquist l. , quinn p. j. , 1987 , apj , 312 , 1 kauffmann g. , white s. d. m. , guiderdoni b. , 1993 , mnras , 264 , 201 kodama t. , arimoto n. , 1997 , a&a , 320 , 41 kodama t. , arimoto n. , barger , a. j. , aragn - salamanca , a. , 1998 , a&a , in press kodama t. , 1997 , ph.d . thesis , univ . of tokyo kppen j. , arimoto n. , 1990 , a&a , 240 , 22 larson r. b. , 1974 , mnras , 169 , 229 larson r. b. , tinsley , b. m. , caldwell , c. n. , 1980 , apj , 237 , 692 lpez - cruz , 1997 , ph.d . thesis , univ . of
toronto mathis j. s. , 1990 , ara&a , 28 , 37 rose j. a. , bower r. g. , caldwell n. , ellis r. s. , sharples r. m. , tegue p. , 1994
, aj , 108 , 2054 schweizer f. , seitzer p. , faber s. m. , burstein d. , ore c. m. d. , gonzlez j. j. , 1990 , apj , 364 , l33 schweizer f. , seitzer , p. , 1992
, aj , 104 , 1039 skillman , e .
d. , kennicutt r. c. , hodge p. w. , 1989 , apj , 347 , 875 stanford , s .
a. , eisenhardt p. r. m. , dickinson , m. , 1997 apj , in press vazdekis a. , casuso e. , peletier r. f. , beckman j. e. , 1996 , apjs , 106 , 307 worthey g. , 1994 , apjs , 95 , 107 worthey g. , trager s. c. , faber , s. m. , 1996 , in : fresh views of elliptical galaxies , ed .
a. buzzoni , a. renzini , a. serrano ( asp conf .
86 ) , p.203 worthey g. , faber s. m. , gonzlez j. j. , 1992 , apj , 398 , 69 | we put upper limits of secondary burst of star formation in elliptical galaxies of the gonzalez s ( 1993 ) sample , most of which locates in small groups , based on the colour dispersion around the @xmath0 versus central velocity dispersion relation , and the equivalent width of h@xmath1 absorption .
there are significant number of h@xmath1 strong galaxies which have ew(h@xmath1)@xmath2 , however they do not always have bluer colours in @xmath0 . to be consistent with small colour dispersion of @xmath0 , the mass fraction of secondary burst to the total mass should be less than only 10 % at the maximum within recent 2 gyr .
this result suggests that even if recent galaxy merging produce some ellipticals , it should not have been accompanied by an intensive star burst , and hence it could not involve large gas - rich systems .
a capture of dwarf galaxy is more likely to explain dynamical disturbances observed in some elliptical galaxies .
the above analysis based on the @xmath0 is not compatible with the one based on the line indices , which requires that more than 10 % of mass is present in a 2 gyr old star burst to cover the full range of the observed h@xmath1 ( de jong & davies 1997 ) .
the discrepancy might be partly explained by the internal extinction localized at the region where young stars form .
however , considering that the h@xmath1 index might have great uncertainties both in models and in observational data , we basically rely on @xmath0 analysis .
[ firstpage ] galaxies : elliptical galaxies : evolution galaxies : stellar content . |
we are carrying out a large project to study the spiral structure and kinematics of the milky way by measuring trigonometric parallaxes and proper motions of star forming regions .
the target sources are 12 ghz methanol masers and we use the national radio astronomy observatory s very long baseline array ( vlba ) . details of this program can be found in @xcite , hereafter called paper i. here we present observations of ( iras 19410 + 2336 ) and . depending on its distance ,
could be in either the carina - sagittarius or local spiral arm of the milky way .
is a very well studied region of high - mass star formation ; its radial velocity is very close to the maximum allowed by circular rotation for standard models of the milky way , and it has generally been assigned the tangent - point distance , @xmath2 , in the carina - sagittarius arm .
our direct measurements of distance and proper motion indicate the true location of these star forming regions in the galaxy and their departures from circular galactic orbits .
we conducted phase - referenced observations of and with respect to two extragalactic radio sources with the vlba under program br100d in order to measure parallaxes .
paper i describes the general observational setup and data calibration procedures , so here we only describe details specific to the observations of and .
the time between epochs was planned to be three months , matching the eastward and northward extrema of the earth s orbit as seen by the sources .
the observations were conducted on 2005 jul .
13 and oct .
20 , 2006 jan 15 , apr .
23 and oct . 19 , and 2007 apr . 19 . however , the data for the epoch of 2006 apr .
23 were not correlated at the position of and were lost for this source .
background compact extragalactic sources were chosen as follows . for
, we selected j1946 + 2300 ( with separation of 1.0 from the maser target ) from the icrf source list @xcite and j1941 + 2307 ( separation 0.7 ) and j1943 + 2330 ( separation 0.3 ) , based on a vla survey of compact nvss sources @xcite .
ultimately , we used only data from the first two sources , as we failed to detect j1943 + 2330 . for , we chose two sources from the vlba calibrator survey @xcite , j1922 + 1530 ( separation 1.0 ) and j1924 + 1540 ( separation 1.2 ) , augmented by j1922 + 1504 ( separation 0.6 ) , which was found in our vla survey .
two strong sources ( j1800 + 3848 and j1922 + 1530 ) were observed near the beginning , middle and end of the observations in order to monitor delay and electronic phase differences among the if bands . after applying the basic calibration procedures
described in paper i , we used the maser features toward at @xmath3 and toward at @xmath4 for interferometer phase - reference data .
when imaging the data referenced to , we adopted a round restoring beam of 1.7 mas ( fwhm ) , slightly larger than the geometric mean of the interferometer response ( `` dirty beam '' ) of @xmath5 mas at a position angle of 14 east of north .
for the data referenced to , we adopted a round restoring beam of 3.3 mas ( fwhm ) , slightly larger than the dirty beam of @xmath6 mas at a position angle of 134 .
[ [ ppm_g59 ] ] a map of the maser emission made by integrating all channels with detectable emission at the first epoch is shown in fig .
[ g59_masers ] .
millimeter and centimeter wavelength continuum data associated with the maser are also presented .
we analyzed 8.4 ghz data , which we obtained from the vla archival database , and find a weak continuum source with flux density of @xmath7 mjy at ( @xmath8 ) = ( 19 43 11.21 , + 23 44 03.32 ) , with a probable position uncertainty of about 0.2 arcsec .
we conclude that the methanol maser ( table 1 ) and the compact continuum sources are associated .
the continuum source has an upper limit of @xmath9 for its size .
one millimeter wavelength core is also associated with the maser and the weak cm continuum source @xcite .
we show the first epoch images of each background continuum source in fig .
[ g59_qsos ] .
one can see that they both appear dominated by a single compact component .
oh maser position ( see table 1 ) . _ middle panel _ : weak 8.4 ghz emission toward imaged with the vla .
contours represent 30 , 50 , 70 , and 90% of the peak brightness of 0.68 mjy beam@xmath10 .
the lowest contour represents 2 times the rms noise level , while the _ thick _
50% contour closely follows the size of the synthesized beam ( @xmath11 ) .
the ( 0,0 ) position corresponds to the position of the 12 ghz methanol maser . _
bottom panel _ : map made by integrating all channels with detectable methanol maser emission at the first epoch .
the @xmath3 maser spot , located at the origin , was used for the parallax fits . the maser spot located near ( @xmath12 ) is at @xmath13 .
contours represent 30 , 50 ( bold ) , 70 , and 90%of the velocity - integrated flux density of 0.82 jy .
the restoring beam is 1.7 mas fwhm for the maser image .
[ g59_masers ] ] for j1941 + 2307 ( upper panel ) and 0.7 mjy beam@xmath10 for j1946 + 2300 ( lower panel ) .
other contours represent increases by a factor of 2 starting from that level .
the peak brightness and integrated flux density are 10.8 mjy beam@xmath10 and 16.3 mjy , respectively , for j1941 + 2307 and 74.8 mjy beam@xmath10 and 114 mjy , respectively , for j1946 + 2300 .
the upper limit of source sizes is 3.0 mas for j1941 + 2307 and 2.9 mas for j1946 + 2300 .
the fwhm of the synthesized beams are represented in the lower left corner of each panel .
[ g59_qsos ] ] when conducting phase - referenced observations , it is important that the position of the reference source matches the interferometer phase center in order to minimize second - order positional errors and improve image quality .
since the icrf source j1946 + 2300 has a position accuracy of @xmath14 mas , we used its position as the basis for all absolute positions given in table [ table : positions ] .
we fitted elliptical gaussian brightness distributions to two maser spots and the two background radio sources for all seven epochs . in fig .
[ g59_parallax ] , we plot the positions of two maser spots ( at of 26.6 and 27.4 ) relative to two background sources .
the measured positions of the g59.7 + 0.1 masers were then modeled as a linear combination of the elliptical parallax and linear proper motion signatures . because systematic errors ( owing to small uncompensated atmospheric delays and , in some cases , varying maser source structure ) typically dominate over signal to noise considerations when measuring relative source positions , we added `` error floors '' in quadrature to the formal position uncertainties .
we used different error floors for the right ascension and declination data and adjusted them to yield post - fit residuals with @xmath15 per degree of freedom near unity for both coordinates .
individual fits are given in table 2 .
fitting for the parallax and proper motion for both sources simultaneously , we obtain @xmath16 mas . the quoted parallax uncertainty is the formal fitting uncertainty , multiplied by @xmath17 to account for possible correlations between the position data for the two maser spots .
this parallax corresponds to a distance of @xmath0 kpc , which is smaller than the `` near '' kinematic distance of 2.7 kpc , and rules out the far distance of 5.8 kpc .
the average proper motions in the eastward and northward directions are @xmath18 and @xmath19 mas y@xmath10 , respectively , as listed in table [ table : g59fits ] .
similarly , the uncertainties were also multiplied by @xmath17 .
[ [ ppm_w51 ] ] we made a 12 ghz methanol maser map of by integrating all channels with detectable emission at the first epoch .
this map is shown in fig .
[ fig4 ] , superposed on a continuum ( 23.8 ghz ) image from archival vla data ( as724 ) . in fig .
[ w51_qsos ] , we show maps of the background continuum sources , phase referenced to the @xmath4 maser spot in , from the first epoch . these background sources are dominated by a single compact component .
absolute positions for the maser reference spot and the background sources are given in table [ table : positions ] .
these are based on the position of j1924 + 1540 , which is uncertain by @xmath14 mas @xcite . .
contours are -5 , 5 , 10 , and 20 times the rms noise of 0.28 mjy beam@xmath10 .
the methanol maser is , as indicated , clearly associated with the hypercompact hii region .
the separation between them is around 0.07 arcsec .
the inset shows the velocity - integrated methanol maser emission with contours representing 5 , 10 , 20 , 40 , 80 , and 160 times 5 mjy beam@xmath10 km s@xmath10 .
the x- and y - axes give east and west offset , respectively , relative to the position given in table 1 .
[ fig4 ] ] for j1922 + 1530 ( top panel ) , 1.8 mjy beam@xmath10 for j1922 + 1504 ( middle panel ) and 2.9 mjy beam@xmath10 for j1924 + 1540 ( bottom panel ) .
other contours represent increases by a factor of 2 starting from that level .
the peak brightness and integrated flux density are 22.5 mjy beam@xmath10 and 43 mjy , respectively , for j1922 + 1530 , 24 mjy beam@xmath10 and 37 mjy , respectively , for j1922 + 1504 , and 465 mjy beam@xmath10 and 615 mjy , respectively , for j1924 + 1504 .
the upper limit of source sizes is 2.6 mas for j1922 + 1530 , 3.8 mas for j1922 + 1504 , and 0.0 mas for j1924 + 1504 .
the fwhm of the synthesized beams are represented in the lower left corner of each panel .
[ w51_qsos ] ] in order to determine the parallax and proper motion of , we used positions of two strong maser spots relative to all three background sources . following the fitting procedures discussed for
, individual spot / background source parallax solutions are listed in table [ table : w51fits ] .
a combined parallax solution yielded @xmath20 mas , corresponding to a distance of @xmath1 kpc , which will be discussed in [ disc ] .
the data and model used for this fit are shown in fig .
[ w51_parallax ] .
the formal parallax uncertainty has been multiplied by @xmath17 , in order to account for possible correlations between the position data for the two maser spots .
some of the relative positions for the fifth epoch ( 2006.80 ) appear to be outliers .
were we to drop the data from this epoch , the parallax estimate decreases to @xmath21 mas , suggesting a slightly greater distance .
the average proper motions of the two maser spots in the eastward and northward directions are @xmath22 and @xmath23 mas y@xmath10 , respectively , as listed in table [ table : w51fits ] .
it is interesting to compare our directly measured distance of @xmath1 kpc with other distance determinations that have recently published for w51 irs 2 .
@xcite performed multi - epoch vlbi observations of the intense h@xmath24o maser outflow in the region ( `` w 51 n '' ) .
a kinematic model they fitted to the measured internal proper motions of the outflow contains its distance as free parameter , for which they find a best fit value of @xmath25 kpc .
@xcite combine near infrared spectroscopy and radio continuum and recombination line observations to characterize the exciting star of w 51 irs 2 . using available data and calculations modeling the temperatures , radii and lyman continuum production rates of zero age main sequence ( zams ) stars ,
they propose a classification of its spectral type as o3 or o4 ( zams ) . using the source s observed bolometric luminosity , they derive @xmath26 and 5.8 kpc for these two choices , respectively , which compares very well with our distance .
we note that the same group recently also obtained a distance of @xmath27 from spectroscopic and photometric observations of four objects in w 51 a classified as o - type stars @xcite .
w 51 a and w 51 irs 2 , separated by an angle of roughly 1 arc minute ( or 1.5 pc ) , are commonly thought to belong to the same complex and a ( near ) kinematic distance of 5.5 kpc has been found by @xcite for w 51 a , comparable to our parallax distance of irs 2 .
currently , large - scale infrared surveys are leading to the discovery of many new open star clusters throughout the milky way ( see , e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , why may trigger a renaissance in efforts to use spectro / photometry of o - type stars for distance estimates with the goal of constraining galactic structure ( see , e.g. , * ? ? ?
* ) . in view of this , we find that the discrepancy discussed above deserves further investigation .
we note that in the famous case of w3oh in the perseus arm , the o - star method yielded @xmath28 kpc for the close - by per ob 1 association @xcite , roughly half the kinematic distance implied by a model of galactic rotation .
this turned out to be in excellent agreement with the 2.0 kpc directly determined via two vlbi parallax measurements of ch@xmath29oh and h@xmath24o masers in w3oh much later @xcite .
future maser vlbi trigonometric parallax measurements of w 51 a could certainly contribute to the perplexing question described above . in order to study the 3-dimensional motion of the maser sources in the galaxy , we converted the proper motions and radial velocities to a galactocentric reference frame .
we adopt the iau standard constants of @xmath30 kpc and @xmath31 , and the hipparcos solar motion values @xmath32 , @xmath33 , and @xmath34 from @xcite . for these values and a flat rotation curve for the galaxy ,
the galactocentric distance of is 7.64 kpc .
its velocity in the direction of galactic rotation is @xmath35 slower than for a circular orbit .
its velocity toward the galactic center is @xmath36 , and its velocity toward north galactic pole is @xmath37 .
these uncertainties include measurement errors , but do not include systematic terms from uncertainty in @xmath38 and @xmath39 . thus we find that has a peculiar motion of @xmath40 directed mostly counter to galactic rotation and toward the galactic center . our trigonometric parallax places in the milky way between the carina - sagittarius and perseus spiral arms .
it seems to be a fairly distant member of the local ( orion ) arm or spur , located close to the point where the spur joins the carina - sagittarius arm .
is near the open cluster ngc 6823 , which is also thought to be located in the local arm @xcite .
spur - like structures have been observed for many galaxies @xcite .
these spurs may form as a consequence of gravitational instabilities inside spiral arms or / and effects of magnetic fields @xcite .
@xcite showed that the growth of spurs can occur due to the mutual contributions of self - gravity and magnetic fields via the so - called magneto - jeans instability . adopting a distance of 5.1 kpc from the sun , places 6.5 kpc from the galactic center , in the the carina - sagittarius spiral arm and
reasonably close the spiral arm tangent point . converting the proper motions and radial velocity of to a galactocentric reference frame
, we find a velocity in the direction of galactic rotation that is @xmath41 slower than for a circular orbit .
its velocity toward the galactic center is @xmath42 , and its velocity toward north galactic pole is @xmath43 .
these peculiar motion uncertainties are fairly large , primarily because of the uncertain parallax measurement .
we have measured the parallax and proper motion of methanal masers in two regions of high - mass star formation
. lies at a distance of @xmath0 kpc in the local ( orion ) arm or spur .
its space motion , relative to a frame rotating with the milky way , is about 12 counter to galactic rotation and toward the galactic center .
our parallax for , while less accurate than for , indicates that is in the carina - sagittarius spiral arm .
this work was supported by chinese nsf through grants nsf 10673024 , nsf 10733030 , nsf 10703010 and nsf 10621303 , and nbrpc ( 973 program ) under grant 2007cb815403 .
andreas brunthaler was supported by the dfg priority programme 1177 .
aalto , s. , h@xmath44ttemister , s. , scoville , n. z. & thaddeus , p. 1999
, , 522 , 165 balbus , s. a. 1988 , , 324 , 60 barbosa , c. l. , blum , r. d. , conti , p. s. , damineli , a. , & figuerdo , e. 2008 , , 678 , l55 beuther , h. , schilke , p. & stanke , t. 2003 , , 408 , 601 basharina , t. s. , pavlovskaya , e. d. , & filippova , a. a. 1980 , _ sva _ , 24 , 559 dutra , c. m. , bica , e. , soares , j. , & barbuy , b. 2003 , , 400 , 533 bica , e. , dutra , c. m. , soares , j. , & barbuy , b. 2003 , , 404 , 223 dehnen , w. , & binney , j. j. , 1998 , , 298 , 387 figurdo , e. , blum , r. d. , damineli , a. , conti , p. s. & barbosa , c. l. 2008 , , 136 , 221 hachisuka , k. , brunthaler , a. , menten , k. m. , reid , m. j. et al . 2006 , , 645 , 337 humphreys , r. m. 1978 , , 38 , 309 imai , h. , watanabe , t. , omodaka , t. , nishio , m. , kameya , o. , miyaji , t. , & nakajima , j. 2002 , , 54 , 741 ivanov , v. d. , borissova , j. , pessev , p. , ivanov , g. r. , & kurtev , r. 2002 , , 394 , l1 kim , w .- t .
, & ostriker , e. c. 2002 , , 570 , 132 la vigne , m. a. , vogel , s. n. , & ostriker , e. c. 2006 , , 650 , 818 ma , c. 1998 , , 116 , 516 messineo , m. , petr - gotzens , m. , menten , k. m. , schuller , f. , & habing , h. j. 2006 , journal of physics conference series , 54 , 238 petrov , l. , kovalev , y. y. , fomalont , e. , & gordon .
d. 2006 , , 131 , 1872 reid , m. j. 2008 , submitted to , ( paper i ) .
russeil , d. 2003 , , 397 , 133 scoville , n. z. , polletta , m. , ewald , s. , stolovy , s. r. , thompson , r. , & rieke , m. 2001 , 122 , 3017 shetty , r. & ostriker , e. c. 2006 , , 647 , 997 xu , y. , reid , m. j. , menten , k. m. & zheng , x. w. 2006a , , 166 , 526 xu , y. , reid , m. j. , zheng , x. w. , & menten , k. m. 2006b , science , 311 , 54 lllllcc g59.7 + 0.1 ...... & 19 43 11.2470 & 23 44 03.315 & & 1.4 & 27.4 & 1.7 + j1946 + 2300 ...... & 19 46 06.25140 & 23 00 04.4145 & 1.0 & 0.073 & & [email protected] @ 33 + j1941 + 2307 ....... & 19 41 55.1114 & 23 07 56.525 & 0.7 & 0.010 & & [email protected] @ 30 + + w51 ......... & 19 23 39.8244 & 14 31 04.953 & & 2.2 & 56.4 & 3.3 + j1924 + 1540 ... & 19 24 39.45588 & 15 40 43.9417 & 1.2 & 0.46 & & [email protected] @ -51 + j1922 + 1530 ... & 19 22 34.6993 & 15 30 10.0327 & 1.0 & 0.23 & & [email protected] @ -48 + j1922 + 1504 ... & 19 22 33.2728 & 15 04 47.537 & 0.6 & 0.022 & & [email protected] @ -43 + [ table : positions ]
llllll 26.6 ...... & j1941 + 2307 & @xmath46 & @xmath47 & @xmath48 + 26.6 ...... & j1946 + 2300 & @xmath49 & @xmath50 & @xmath51 + 27.4 ...... & j1941 + 2307 & @xmath52 & @xmath53 & @xmath54 + 27.4 ...... & j1946 + 2300 & @xmath55 & @xmath53 & @xmath56 + + 26.6 ...... & combined & @xmath57 & @xmath50 & @xmath58 + 27.4 ...... & & & @xmath59 & @xmath60 + [ table : g59fits ] llllll 56.0 ...... & j1922 + 1504 & @xmath61 & @xmath62 & @xmath63 + 56.0 ...... & j1922 + 1530 & @xmath64 & @xmath65 & @xmath66 + 56.0 ...... & j1924 + 1540 & @xmath67 & @xmath68 & @xmath69 + 56.4 ...... & j1922 + 1504 & @xmath70 & @xmath71 & @xmath72 + 56.4 ...... & j1922 + 1530 & @xmath73 & @xmath74 & @xmath75 + 56.4 ...... & j1924 + 1540 & @xmath76 & @xmath77 & @xmath78 + + 56.0 ...... & combined & @xmath79 & @xmath80 & @xmath81 + 56.4 ...... & & & @xmath82 & @xmath83 + [ table : w51fits ] | we report trigonometric parallaxes for and , corresponding to distances of @xmath0 kpc and @xmath1 kpc , respectively . the distance to is smaller than its near kinematic distance and places it between the carina - sagittarius and perseus spiral arms , probably in the local ( orion ) spur .
the distance to , while subject to significant uncertainty , is close to its kinematic distance and places it near the tangent point of the carina - sagittarius arm .
it also agrees well with a recent estimate based on o - type star spectro / photometry .
combining the distances and proper motions with observed radial velocities gives the full space motions of the star forming regions .
we find modest deviations of 5 to 10 from circular galactic orbits for these sources , both counter to galactic rotation and toward the galactic center . |
investigation of co adsorption on well - defined transition metal surfaces has been of great academic interest for several decades @xcite because of the molecule s obvious relevance to many industrial processes and as a prototype reactant in studies aiming to provide an understanding of catalytic reactions @xcite .
these experimental and theoretical @xcite studies have considered the importance of identifications of active sites based on examination of adsorption and desorption energies and sticking coefficients @xcite . since
real catalysts consist of small metal clusters with microfacets of various orientations , the catalyst surface is generally far from that of one with a low miller index , rather it contains defects like steps and kinks which may play specific roles in determining its reactivity @xcite . a plausible way to understand systematically
the effect of steps and kinks on chemisorption is to undertake the examination of co adsorption on a set of vicinal surfaces , as has been done in a recent thermal deposition spectroscopy ( tds ) study on cu surfaces .
these experimental studies find a dependence of the co binding energy on the coordination of the adsorption site , but surprisingly the effect does not extend to the least undercoordinated sites , namely that of co on the kink - sites on cu(532 ) .
_ @xcite find hardly any difference in the adsorption energy for co on a step and a kink site atom implying that coordination alone may not account for the adsorption energy on this vicinal surface .
as proposed by bagus _ et .
@xcite in their study of co adsorption on cu(100 ) , the repulsive interaction between the o atom and the _ s _ states of the substrate can also play an important role in determining the trends in the adsorption energy . similarly in the extensive study of co adsorption on the ( 111 ) surface of several transition metals , gajdos _
_ et . al . _ _
@xcite argue that the extent of the shift of the @xmath0 and @xmath1 orbital charge density from the c - o bond to the region below the carbon ( metal surface ) may control the adsorption energetics . of course
, an in - depth understanding of the effect of the local electronic and geometric structure on co adsorption may be obtained from the application of established theoretical methods such as those based on density functional theory @xcite to the set of cu surfaces studied in these experiments .
such calculations may raise some concerns since on the low miller index surfaces of cu they have been found to : a ) show preference for the hollow site @xcite , while the top site is preferred in the analysis of experimental data @xcite ; b ) overestimates the co adsorption energy @xcite .
very recently several groups @xcite have attempted to supplement dft functionals in various ways so as to remedy the two above shortcomings of dft for this system with some success .
our intention in this paper is , however , to carry out a systematic , comparative study of co chemisorption on the on - top sites of a set of low and high miller index surfaces of cu to see the trends in adsorption energies as a function of local coordination and co coverage .
while these calculations have been motivated by experimental data @xcite , we are also aware of theoretical work @xcite on co adsorption on a few of the surfaces that we are considering here .
naturally , we will compare our results to all available information that exists to date .
in addition to the trends in co adsorption energy on cu surfaces , we are also interested in the diffusion barrier for co , since this is a necessary step in any chemical reaction on the surface . a comparative study of the diffusion of co on cu(111 ) and cu(211 ) is thus presented .
also , given the interest in reactions at and near steps and kinks on surfaces we have carried out an examination of co on several sites near these defects .
we are also including in this work the effect of co adsorption on the nature of atomic relaxations of the vicinal surface , to see how the trends in the relaxation patterns correlate with those in the surface electronic structure .
after giving the computational details of our work in next section , we present a detailed analysis of the calculated co adsorption energies as a function of local coordination and their comparison to experiments and other calculations .
this is followed by investigations of the implication of increasing co coverage on adsorption energetics .
rest of the paper is devoted to the characterization of the changes in the surface electronic structure ( local density of states , workfunctions ) on co adsorption .
the vibrational frequencies of co are also presented on the set of surfaces .
_ ab initio _ calculations performed in this study are based on the well - known density functional theory ( dft ) @xcite . for purposes here ,
a systematic study of the energetics and the electronic structure of a set of cu surfaces of varying geometry was made by solving kohn - sham equations in plane - wave basis set using the vienna _ ab initio _ simulation package ( vasp ) @xcite .
the electron - ion interactions for c , o and cu are described by ultrasoft pseudopotentials proposed by vanderbilt @xcite . a plane - wave
energy cut - off of 400 ev was used for all calculations and is found to be sufficient for these systems @xcite . in all calculations reported in this article , the generalized gradient correction of perdew and wang @xcite ( pw91 ) was used since it has been shown to give results for adsorption energetics and structural parameters which are in better agreement with experiment as compared to those obtained using the local density approximation ( lda ) @xcite . the calculated bulk lattice constant for cu was found to be 3.6471 and the brillouin zone sampling of the total energy was based on the technique devised by monkhorst and pack @xcite for all bulk calculations with a _ k_-point mesh of @xmath2 the supercell approach with periodic boundaries is employed to model the surface systems . to calculate the total energies for several coverages of co , we have used surface unit cells consisting of @xmath3 atoms per layer . while the definition of the coverage for the low miller index surfaces is conceptually easy and simply defined as the ratio of number of adsorbate and substrate atoms in the surface unit cell , for surfaces with steps and kinks we need also to take into account the specific areas of the unit cells , for consistency . for vicinal surface
we have thus adopted a definition for the coverage @xmath4 which is related to the one on the ( 111 ) surface @xmath5 according to @xmath6 @xmath7 , where @xmath8 is the area of the surface unit cell of cu(111 ) and @xmath9 is the area of the vicinal surface .
we have used 2.579 as the nearest neighbor distance for evaluating the area of the surface unit cell .
cu(100 ) and cu(111 ) surface systems were modeled by a 4 layer ( 16 atoms ) tetragonal and hexagonal supercell , respectively .
these 4 layers are separated with 11 of vacuum . for cu(100 ) , calculations were performed for @xmath10 overlayer corresponding to 50% coverage of co and for a @xmath11 structure corresponding to full coverage even though such a high coverage is unrealistic . for cu(111 ) , the @xmath12 structure corresponding to 33% co coverage , and those corresponding to coverages of 66% and 100% were also studied . for all coverages considered on cu(100 ) and cu(111 ) ,
the co molecule was taken to be adsorbed on the on - top sites , with the molecule sitting perpendicular to the surface and the c atom forming a bond with the cu surface , as also reported in number of experiments @xcite . a @xmath13 monkhorst - pack _
k_-point mesh was used for cu(100 ) and of @xmath14 for cu(111 ) .
the cu(110 ) , cu(211 ) and cu(221 ) surfaces were modeled by an orthorhombic supercell of 8 , 17 and 22 layers , respectively , separated with approximately 12 vacuum . for cu(110 ) , a @xmath15 surface unit cell ( 16 cu atoms ) with 1 and 2 co molecules were chosen to model 50% and 100% coverage of co , respectively , as shown in the fig .
1c . a monkhorst - pack _
k_-point mesh of @xmath16 was used for this surface .
both cu(211 ) and cu(221 ) being regularly stepped surfaces with monoatomic steps and terrace width of 3 and 4 atoms , needed 34 and 44 atoms , respectively , to model a @xmath15 surface unit cell . for cu(211 ) , this supercell corresponds to 17.69% coverage , while for cu(221 ) the corresponding coverage is 14.44% .
calculations for twice the coverage were also performed by incorporating an additional co molecule in the same supercell .
the detailed analysis was performed for the on - top and the bridge site , also for cu(221 ) , for which the @xmath15 supercell corresponds to 14.44% co coverage and an additional atom in the same supercell boosts the coverage to 28.88% . a monkhorst - pack _
k_-point mesh of @xmath17 was used for ( 211 ) and @xmath18 was used for the ( 221 ) surface .
finally , cu(532 ) whose surface consists of regularly spaced kinks , was modeled by a simple monoclinic supercell of five layers , where each layer has 8 non - equivalent atoms .
these five layers were separated with 12 of vacuum . a monkhorst - pack _
k_-point mesh of @xmath19 was used and co was adsorbed on top of the cu atom at the kink site which is the experimentally proposed preferred site @xcite .
two coverages ( 10.76% and 21.52% ) of co on cu(532 ) were modeled by adsorbing one and two molecules , respectively , on the kink site ( and on a site very next to kinked site ) .
since vicinal surfaces provide local environments with a range of coordinations and a hierarchy of adsorption sites may exist on them @xcite , we have carried out calculations for one of the stepped surfaces , cu(211 ) , for eight different sites as indicated in fig .
1d . since these results point to a preference for the bridge site , for cu(221 ) and
cu(532 ) we have calculated adsorption energies for the bridge sites , in addition to the on - top site . in order to calculate the total energy of the system for a relaxed configuration , atoms on all surfaces considered were allowed to move in all three directions and the structures were relaxed until the forces acting on each atom were converged to better that 0.02 ev / .
the adsorption energies were calculated by subtracting the total energy of the co - molecule and that of the corresponding fully relaxed clean - cu surface system from the total energy of co / cu surface system @xmath20 .
another quantity of interest , the workfunction , was calculated by taking the difference of the average vacuum potential and fermi energy for each surface .
finite difference method was used to obtain vibrational frequencies of the co molecule in the gas phase and on the surfaces .
co internal stretching and co - metal stretch frequencies were calculated in the direction perpendicular to the surface . to analyze the nature of the bonding between co and the cu - atoms , local density of states
were also obtained .
a gaussian function of 0.2 ev width was used to smoothen the local density of states ( ldos ) .
as already mentioned above , the collection of surfaces studied here provide variation of adsorption sites with coordination ranging from 6 to 9 .
the kink sites with coordination 6 on cu(532 ) are particularly interesting because of their unexpected results from tds measurements @xcite . also of significance
is the fact that the step edges of cu(211 ) and cu(221 ) represent the two different microfacets of monoatomic steps on fcc(111 ) surfaces .
cu(211 ) has the ( 100)-microfacet while cu(221 ) has the ( 111)-microfacet
. it will be interesting to see if such differences in their geometry affect the energetics of co on these surfaces . in table
i , we have compared our calculated co bond length and the corresponding surface - carbon distance for these three surfaces .
while the c - o bond length remains insensitive to the local geometry and coordination , c - cu bond is much shorter for bridge ( b ) then ontop ( t ) site adsorption . as for cu(532 ) there are two non equivalent bridge sites labeled by b@xmath21 and b@xmath22 in table i and fig .
the c - o bond lengths are very close to the ones calculated by others @xcite .
when co is adsorbed on the bridge site on the three vicinal surfaces , the co bond length slightly increases and the carbon - surface distance decreases .
earlier dft calculations for co on cu(211 ) also find the same trend @xcite .
the only available date @xcite of bond lengths is for the low miller index surfaces and is in good agreement with our calculated value of 1.16 for the c - o bond length and 1.85 and 1.86 c - cu bond for cu(111 ) and cu(100 ) , respectively .
when compared to vicinal surfaces , we do not see any difference for c - o bond length and a small variation within 0.02 for c - cu bonds .
a small increase of 0.02 was seen in c - cu bond when co coverage was increased from 33% to 100% whereas when coverage is doubled from 50% in cu(100 ) and cu(110 ) , we see only 0.01 increase in c - cu bonds and no change was found in c - o bond lengths . our calculated co adsorption energies on various cu surfaces and their surface atomic coordination is summarized in table ii .
the lowest coordinated and in turn the most favorable for co adsorption is the kink atom on cu(532 ) ( fig .
1f ) , with the highest adsorption energy of 0.98 ev at a coverage of 10.75% co ( i.e. one co molecule / per kink atom ) .
the calculated adsorption energy for this case is found to be particularly higher than what has been seen experimentally .
although overestimation of adsorption energies is typical of dft based calculations , it is expected that the qualitative behavior would be similar . for co adsorption on the steps of cu(110 ) , cu(211 ) and
cu(221 ) with atomic coordination 7 , our calculations find @xmath23 to be about @xmath24 ev , while experimental values are around 0.6 ev . for adsorption on the kink site
we predict an increase in @xmath23 while experiments @xcite find it to be same as for the steps . as is clear from fig .
2 our calculated values scale nicely with the surface coordination and the variation is larger than that extracted from experimental data , which is plotted in fig . 2 for comparison . the natural question is why this difference between theory and experiment .
we offer here a few rationale .
for example , there is the possibility that site blocking by another co could lead to smaller measured value on an open surface like cu(532 ) . to test the viability of this proposition , we adsorbed an additional co molecule on a site next to the kink site ( sc1 in fig .
1f ) which increased the coverage to 21.5% . as we see in table ii , this leads to a decrease in the adsorption energy to 0.85 ev which is very close to what we find for cu stepped surfaces .
the variation of the adsorption energy with coverage is quite remarkable for all surfaces considered . for cu(221 )
@xmath25 changes from 0.85 ev for 14.5% coverage to 0.65 for 29% coverage .
although dft studies of co adsorption on cu(110 ) and cu(211 ) have already been reported in the literature @xcite , for consistency we have included them in table ii . though we have used the same @xmath15 cell for cu(211 ) , this leads to a different coverage ( 17.69% ) from that on cu(221 ) because of the difference in the terrace width . the same @xmath15 cell used for cu(110 )
( with [ 110 ] being _ x _ and [ 100 ] being _ y_-directions ) , as shown in fig .
1c , results in a 50% coverage .
the co molecule is adsorbed on the top site on cu(110 ) and cu(211 ) , as shown in fig .
1d and 1e .
our calculated adsorption energies are within 20 mev of each other for all of cases involving atoms with coordination 7 . for cu(211 )
our calculated value is somewhat smaller than in ref .
@xcite which may be associated with the differences in calculational details . for co adsorption on cu(110 ) liem _ et .
@xcite obtained an adsorption energy of 0.95 ev for the same coverage and on - top site .
the discrepancy with our results could be due to the usage of fewer layers and smaller _ k_-point mesh and energy cut - off in their calculations .
for example , we find the adsorption energy to change from 0.87 ev to 0.89 ev when we reduce the number of layers in our supercell from 8 to 6 ( note that a three layer slab was used in ref .
also , we have found an energy cut - off of 300 ev to be not sufficient to provide converged results for the total energy for open surfaces such as the ( 110 ) .
a cut - off of at least 400 ev ( value used in all our calculations ) is required for accurate determination of adsorption energies .
for the next surface in this hierarchy cu(100 ) whose surface atoms have coordination 8 , we obtain an adsorption energy of 0.77 ev for a 50% co coverage .
this result is again higher than the experimental value which ranges , between @xmath26 and @xmath27 ev @xcite . on the other hand ,
our result is close agreement with the calculated value of 0.863 ev found for a smaller coverage of 0.25 ml by gajdos and hafner @xcite , using the same dft technique .
finally , we turn to cu(111 ) which has a surface atoms with coordination 9 and displays the lowest adsorption energy of all surfaces discussed here . for this particular surface , we have used the experimentally studied @xmath12 overlayer which corresponds to 33% co coverage . for 0.25 ml coverage an adsorption energy of 0.74 ev has already been reported @xcite .
the small difference from our result of 0.634 ev may be assigned to the difference in coverage , and is consistent with a decrease of adsorption energy with an increase of coverage .
the clear trend of increasing adsorption energy with decreasing local coordination can be seen from the plot in fig
. 2 where solid triangles are our calculated values and empty triangles are the one from the experiment @xcite .
as expected we find a hierarchy of adsorption sites on cu(211 ) . the bridge site ( labeled 2 in fig .
1d ) was found to be slightly preferred over on - top by @xmath28 0.06 ev . the sites labeled 3 ( fcc - hollow ) and 4 ( hcp hollow near step edge ) , in fig 1d were the next preferred with adsorption energies of 0.78 and 0.89 ev , respectively . the least favorable site on this surface was found to be the site labeled 6 which is between two step edges and two corner atoms and has highest number of bonds with the carbon atom and an adsorption energy of 0.56 ev . on site 5 , 7 and 8
adsorption energies are very close to each other ranging from 0.6 - 0.65 ev .
the adsorption site with the highest coordination is the corner atom ( cc ) on cu(211 ) with effective surface coordination of 10 , and labeled as site # 8 in fig .
1d for which we find the adsorption energy to be the lowest of all ( 0.617 ev ) , consistent with the above discussion .
note that we find the bridge site to be also preferred by about 0.08 ev over the on - top site for co adsorption on cu(221 ) . on the other hand , for cu(532 ) which has two non - equivalent bridge sites referred as b@xmath21 and b@xmath22 in table ii and fig .
1f , we find the adsorption energy to be 0.94 ev which is less than that on the kink site . in all cases , adsorption energies drop by @xmath29 mev when the coverage is doubled for these high miller index surfaces .
various experiments already show the strong dependence of adsorption energy on co coverage on cu surface @xcite .
the electron energy loss spectroscopy ( eels ) experiment of peterson _ et .
@xcite shows a 50% decrease in adsorption energy for 0.3ml increase in co coverage for cu(100 ) .
one of the experimental techniques to determine the adsorption energies is the temperature programmed desorption ( tpd ) .
it is hence desirable to develop a techniques by which tpd spectra are calculated for a given system .
for the case of low miller index surfaces , the task may be trivial as only a limited number of processes are involved .
but for the case of real surfaces with steps and kinks , the situation become more cumbersome . to achieve a realistic description of tpd spectra from these surfaces , one needs to calculate not only adsorption energies for sites with different local environments , but also activation energies associated with different diffusion paths .
as we have reported above , the co molecules adsorb preferably near kinks and steps , and it is hence rarely that co molecules sit on the down side of the terrace ( sites 6 and 8 in fig .
the relevant energies are hence those associated with diffusion on cu(111 ) and near the step of a cu vicinal surface . here
we have in mind systems dominated by the presence of ( 111 ) facets and step edges .
the knowledge of these activation energies along with the different adsorption energies will constitute the base for kinetic monte carlo simulations that will determine the tpd spectra . in order to reach this goal , the energy landscape for the diffusion of co on cu(111 ) and cu(211 )
surfaces were calculated . since we were interested in describing the motion of the co molecule between the equilibrium sites , the path on cu(111 )
was chosen to be from one top site to the adjacent one , passing through fcc - hollow , bridge , and hcp hollow sites ( fig .
all layers of the metal surface , and c and o atoms were allowed to relax fully in all three directions on high symmetry sites . for the intermediate sites
only carbon atom was fixed in one direction and rest of the system was allowed to relax fully .
three intermediate points between the top and the bridge site and two points for each bridge - fcc - hollow and bridge - hcp - hollow were calculated , as shown in fig.4a . the relative energy profile of the co molecule on cu(111 )
is shown in fig .
4a . as in previous studies ,
our calculations based on dft with gga indicate that co prefers to sit on fcc hollow site instead of the top site .
however , our calculations show a barrier of about 200 mev to go from the hollow to the top site and less than 100 mev to go from the hollow to the bridge site . fig . 4b and 4c ,
show the one dimensional energy landscape for the co molecule on cu(211 ) .
two diffusion paths are chosen in this case : a ) top - bridge - top ( along the step edge ) and b ) top - fcc - hollow ( away from the step ) .
4b indicates that diffusion from bridge to top is more favorable than from top to hollow with diffusion barrier about 60 mev compared about 200 mev barrier for diffusion from top to hollow site .
these results indicate that co molecules are more free to roam on the step edge of cu(211 ) than away from it .
in table iii , we present a comparison of the interlayer relaxations of the stepped and kinked cu surfaces as introduced by the adsorption of the co molecule . the results for the corresponding clean surfaces
are also included .
the cu(532 ) surface has the most dramatic effect of co adsorption as @xmath30 is found to have a large outward relaxation of @xmath31 compared to an inward relaxation of @xmath32 for the clean surface .
the effect is local and as seen from table iii , there is only a small change in the relaxations for the rest of the layers on co adsorption .
there is so far no experimental data available for interlayer relaxation on cu(532 ) , however , our calculated trend for clean cu(532 ) is in agreement with those calculated with many body potentials @xcite . on the stepped surface cu(221 ) , the effect of co adsorption is also mostly on @xmath30 in which large inward relaxation of @xmath33 on the clean surface is overtaken by a small outward relaxation of @xmath34 .
calculated interlayer relaxation on clean cu(221 ) matches well with those obtained with the full - potential linear augmented plane - wave method ( fplapw ) @xcite and many body potentials @xcite for this surface .
the interlayer relaxations of clean cu(211 ) agrees with those obtained from low energy electron diffraction measurements and previous theoretical techniques @xcite .
like other two surfaces , cu(211 ) shows that the strong contraction of @xmath35 between layer 1 and 2 ( @xmath30 ) of the clean surface is overtaken by a small inward relaxation of @xmath36 on co chemisorption .. this general trend of the change in the top layer relaxation from a large contraction to expansion upon adsorption of co is in well accord with the observation by ibach and bruchman @xcite who reported a surface vibrational mode above the bulk band for the vicinal surface pt(775 ) ( due to the large contraction ) that disappears when co is adsorbed ( reflecting the change of the large contraction to an expansion ) .
vibrational properties of a ` free ' co molecule were calculated by fully relaxing a single molecule in a large super cell of size , approximately @xmath37 .
the stretching mode was calculated to be 264.5 mev .
this can be compared with the experimentally measured frequency of isolated co molecule of 257.8 mev with eels and of 257.7 mev with infrared ( ir ) spectroscopy @xcite . for co - covered cu surfaces
, we have calculated the frequencies of two modes : the metal - molecule ( @xmath38 ) stretch and the intra - molecular stretch ( @xmath39 ) .
we find a small drop in the vibrational frequency as compared to that of the gas phase co due to the bonding on co with the cu surface atom .
the frequencies of the co vibrational modes calculated for all surfaces are summarized in table iv .
for all cases in which co is adsorbed on the top site we see almost the same value of the co stretching frequency ( @xmath39 ) , as expected
. however , there is a small increase in ( @xmath38 ) , with a decrease in the coordination of the substrate atom .
this is also to be expected since molecules adsorbed on low coordinated sites will have stiffer bonds than those on highly coordinated sites . whereas , for the case of the bridge site , there is a decrease of approximately @xmath40 mev in co stretch frequency , from that for top site , for all cases
in this case understood due to the molecule strongly bonded with two surface atoms and being not as free to vibrate as on top site .
the drop in @xmath38 mode is also @xmath41 mev greater than top site . in table
v , we have summarized the calculated workfunctions for all cu surfaces considered here , with and without the adsorbate along with available experimental values for clean surfaces of cu(100 ) , cu(111 ) and cu(110 ) .
our calculated workfunctions are smaller than the experimental values of the known surfaces @xcite .
our calculated values showed a decrease in workfunctions for clean surfaces with decrease in coordination of substrate .
a similar kind of behavior can be seen in the experiments for ( 111 ) , ( 100 ) and ( 110 ) surface @xcite .
our calculations also showed an increasing trend on workfunction when co is adsorbed on the surface .
this effect is strongest for cu(100 ) surface with workfunction change of 0.57 ev for all other surfaces this change is less than 0.17 ev .
when an electronegative element like c is adsorbed on a metallic surface , it tends to take charges away from the surface .
this would have a strong dependence upon the coordination of the adsorbate as reported in many studies @xcite .
since we have considered co adsorption on a variety of local atomic environments it is important that we examine the differences in the local electronic structure and the nature of the bonding of the co with the substrate . in fig .
5 - fig . 7 , we have plotted the local density of _ d_-state ( ldos ) of cu atoms and _ sp_-states for c and o atoms to study the _ sp - d _ hybridization , for three high miller index surfaces of interest here .
first of all , we see narrowing ( fig
. 5 - fig .
7 ) of the _ d_-band for the lower coordinated surfaces , as seen by tersoff _ et . al . _
@xcite for their calculations of ni and cu surfaces .
the _ sp_-state of c and o are split and have almost filled bonding band .
these states make sub - bands with _ d_-state of cu surface in all case which can be seen to appear at the same low energies as for co case .
this affect is found to be more localized on flat surfaces but on stepped surfaces , even though co is adsorbed on top site we see a small change in redistribution of ldos for next layer atom , which can be caused due to the small inter layer separation i.e. 0.296 for ( 532 ) as compared to the biggest 2.1 for ( 111 ) case .
the next layer atom being so close to co also experiences some affect of the adsorbate .
this could well describe the different adsorption energies with different coordination . along with this electronic charge density distribution and ldos for these atoms
we can safely say that this difference for flat ( high coordinated sites ) and stepped and kinked ( low coordinated sites ) could serve as the basis of different adsorption energies .
in this paper , we have presented the results of a detailed theoretical investigation of co adsorption on three low miller index surfaces , namely , ( 111 ) , ( 100 ) , ( 110 ) , two stepped surfaces , ( 211 ) and ( 221 ) , and a kinked ( 532 ) surface of cu .
the co was adsorbed on experimentally observed preferred sites to explain trends in adsorption energies due to different geometrical and chemical structure .
our calculated values of adsorption energies show the increasing trend with decrease in local coordination of surface atoms with cu(532 ) being the most favorable surface for co adsorption with lowest coordination
. however bridge sites of cu(211 ) and cu(221 ) were found to be slightly preferred over top site which was not the case for cu(532 ) .
we also found a strong dependence of adsorption energy on coverage of the adsorbate .
our calculations show a large decrease in adsorption energy with increase in coverage .
our calculations of diffusion of co molecule on metal surfaces indicate that it diffuses much more easily along the step edge as compared to away from the step edge ( terrace ) .
a very small drop in vibrational frequency of free molecule was noted when it was adsorbed on the metal surface but difference within different surfaces was negligible due to same on top adsorption site for all surfaces .
workfunction of adsorbate covered surface was found to increase in all cases compared to the clean surface .
the authors would like to thank dr .
sergey stolbov and dr .
claude r. henry for interesting and very helpful discussions .
we also like to acknowledge financial support from us department of energy under grant no .
de - fgo3 - 03er15445 and computational resources provided by national science foundation cyberinfrastructure and teragrid grant no .
tg - dmr050018n .
.calculated structural properties of co adsorbed on the top ( t ) and bridge ( b ) sites on cu(211 ) , cu(221 ) and cu(532 ) .
the two non equivalent bridge sites on cu(532 ) ( see fig .
1f ) are labeled as b@xmath42 and b@xmath42 .
values in square brackets are reported by others [ 21 ] .
[ key ] [ cols="^ " , ] top view of a ) 111 , b)100 , c)110 , d)211 , e ) 221 and f ) 532 surface .
different coverages at different sites are shown .
dark colors represents top layer .
note that in 532 surface all eight atoms are non - equalent and belongs to eight different layers .
, width=453 ] | our _ ab initio _ calculations of co adsorption on several low and high miller index surfaces of cu show that the adsorption energy increases as the coordination of the adsorption site decreases from 9 to 6 , in qualitative agreement with experimental observations . on each surface
the adsorption energy is also found to decrease with increase in coverage , although the decrement is not uniform . calculated
vibrational properties show an increase in the frequency of the metal - c mode with decrease in coordination , whereas no such effect is found for the frequency of the co stretch mode .
examination of the surface electronic structure shows a strong local effect of co adsorption on the local density of state of the substrate atoms .
we also provide some energetics of co diffusion on cu(111 ) and cu(211 ) . |
* natural extension and regular part . * let @xmath2 be a rational function of degree @xmath3 .
it generates a non - invertible dynamical system @xmath4 but it also generates an invertible dynamics in the space of backward orbits " ( the inverse limit ) @xmath5 with action @xmath6 we say @xmath7 ( with dynamics by @xmath8 ) is the _ natural extension _ of @xmath9 , with topology induced by @xmath10 .
note that the _ projection _
@xmath11 defined by @xmath12 semiconjugates @xmath8 and @xmath9 . in 1990
s , m.lyubich and y.minsky @xcite introduced the theory of hyperbolic 3-lamination associated with rational functions , which is analogous to the theory of hyperbolic 3-manifold for kleinian groups .
the theory is based on the study of the natural extension , in particular the subset called the _
regular part _ ( or _ regular leaf space _ ) , defined as follows : the point @xmath13 is _ regular _ if there exists a neighborhood @xmath14 of @xmath15 whose pull - back @xmath16 along @xmath17 ( i.e. , @xmath18 is the connected component of @xmath19 containing @xmath20 ) is eventually univalent .
regular part _
@xmath21 of @xmath7 is the set of all regular points , and we say the point in @xmath22 is _ irregular_. the regular part is invariant under @xmath8 , and each path - connected component ( leaf " ) of the regular part possesses a riemann surface structure isomorphic to @xmath23 , @xmath24 , or an annulus .
( the annulus appears only when @xmath9 has a herman ring . ) * type problem . * when the critical orbits of @xmath9 behave nicely , we may regard @xmath21 as a riemann surface lamination with all leaves isomorphic to @xmath23 .
such a situation yields some nice properties of dynamics , like rigidity , or existence of conformal invariant measures on the lamination .
for example , this is the case when @xmath9 has no recurrent critical points in the julia set ( * ? ? ?
* prop.4.5 ) .
another intriguing case is when @xmath9 is an infinitely renormalizable quadratic map with a persistently recurrent critical point ( * ? ? ?
* lem.3.18 ) . for general cases ,
it is questioned in ( * ? ? ?
* , 10 ) when we have leaves of _ hyperbolic type _ , especially leaves isomorphic to @xmath24 .
( the counterpart , leaves isomorphic to @xmath23 , are conventionally called _
parabolic_. ) easy examples of hyperbolic leaves are provided by invariant lifts of siegel disks and herman rings .
non - rotational hyperbolic leaves ( that are rather non - trivial ) are found by j.kahn and by j.rivera-letelier .
readers may find details in the paper by j.kahn , m.lyubich , and l.rempe @xcite , that can be summarized as follows : when the julia set is contained in the postcritical set , there are infinitely many hyperbolic leaves . in this case
such a leaf may not touch the ( lifted ) julia set in the natural extension .
however , by using the tuning technique and _ the gross star theorem _ , which gives a necessity condition of a leaf to be parabolic , we may construct a quadratic polynomial with hyperbolic leaves that do intersect with the julia set . in both cases ,
a recurrent critical point plays a crucial role . in the quest of new non - rotational hyperbolic leaves ,
it is natural to ask the following question : _ is there any non - rotational hyperbolic leaf when @xmath9 has an irrationally indifferent fixed point ? _ because existence of such a fixed point implies existence of a recurrent critical point whose postcritical set is a continuum , and it seems really close to the situations in @xcite .
the aim of this note is to give some results on this question .
* siegel disk of bounded type .
* @xmath0 with irrational @xmath1 of bounded type has a siegel disk @xmath25 centered at the origin , whose boundary @xmath26 is a quasicircle . in section 2 of this note
we give a proof of [ thm_siegel ] in the regular part of the natural extension @xmath27 , the only hyperbolic leaf is the invariant lift @xmath28 of the siegel disk . in the proof we use
lyubich and minsky s criteria for parabolic leaves , _ uniform deepness _ of the postcritical set , and one of mcmullen s results on bounded type siegel disks .
* feigenbaum maps .
* it would be worth mentioning that the same method as the proof of theorem [ thm_siegel ] can be applied to a class of infinitely renormalizable quadratic maps , called _
feigenbaum maps _ ( precise definitions will be given later . ) in section 3
we give an alternative proof a result in @xcite on non - existence of hyperbolic leaves : [ thm_feigenbaum_intro ] the regular part @xmath21 of a feigenbaum map @xmath9 has only parabolic leaves .
* cremer points and hedgehogs . *
section 4 is devoted for rational functions with cremer fixed points and their hedgehogs . for any small neighborhood of cremer fixed point @xmath29
, there exists an invariant continuum @xmath30 ( a hedgehog " ) containing @xmath29 , equipped with invertible sub - dynamics " @xmath31 . in section 4 , we will present a proof of the next result by following a.chritat : [ thm_hhat_intro ] the invariant lift @xmath32 of @xmath30 is a continuum contained in the irregular part of the natural extension . since this natural extension has a continuum of irregular points
, one may expect to apply the classical gross star theorem to find a hyperbolic leaf , as in @xcite .
however , the actual situation is not that good .
it is still difficult to show the ( non-)existence of hyperbolic leaves without assuming the same conditions as @xcite .
indeed , we will show that the irregular points in hedgehogs are not big enough to apply the gross star theorem .
* acknowledgment .
* the authors would like to thank a.chritat , j.kahn , m.lyubich , and l.rempe for stimulating discussions on this subject .
here we give a proof of theorem [ thm_siegel ] .
the idea is fairly simple and we will apply it to feigenbaum quadratic maps in the next section .
let us start with some terminologies : * postcritical sets .
* let @xmath33 be a polynomial of degree more than one and we denote the set of its critical points by @xmath34 .
the _ postcritical set _ of @xmath9 is defined by @xmath35 for example , if @xmath0 with @xmath1 of bounded type , then its siegel disk @xmath25 is a quasidisk and its boundary contains the only critical point @xmath36 .
since the dynamics of @xmath37 is topologically conjugate to an irrational rotation , we have @xmath38 . *
deep points and uniform deepness .
* let @xmath39 be a compact set in @xmath23 . for @xmath40 ,
let @xmath41 denote the radius of the largest open disk contained in @xmath42 .
( when @xmath43 , we define @xmath44 . )
then it is not difficult to check that the function @xmath45 is continuous .
we say @xmath40 is a _ deep point _ of @xmath39 if @xmath46 as @xmath47 .
is a _ deep point _ if @xmath48 for some @xmath49 .
here we mildly generalize it . ] for a subset @xmath50 of @xmath39 , we say @xmath50 is _ uniformly deep _ in @xmath39 if for any @xmath51 there exists an @xmath52 such that for any @xmath53 and @xmath54 , we have @xmath55
. * deepness and measurable deepness . * for a given measurable set @xmath56 , let @xmath57 denote its area . for a compact set @xmath39 in @xmath23 and an @xmath58 with @xmath59 , we define the _ density _ of @xmath39 in @xmath58 by @xmath60 we say @xmath40 is a _ measurable deep point _ of @xmath39 if @xmath61 as @xmath47 .
we say a subset @xmath50 of @xmath39 is _ uniformly measurable deep _ in @xmath39 if for any @xmath51 , there exists an @xmath62 such that for any @xmath53 and @xmath54 we have @xmath63 [ lem_deepness ] let @xmath39 be a compact set in @xmath23
. then 1 .
if @xmath40 is a measurable deep point of @xmath39 , then @xmath64 is a deep point .
if @xmath65 is uniformly measurable deep in @xmath39 , then @xmath50 is uniformly deep .
it is enough to show ( 2 ) : suppose that @xmath50 is uniformly measurable deep .
then for any @xmath51 , there exists an @xmath52 such that @xmath66 for any @xmath53 and @xmath54 . since @xmath67 we have @xmath68 .
this implies that @xmath50 is uniformly deep .
@xmath69 * uniform deepness of the postcritical set .
* now let us go to the proof of theorem [ thm_siegel ] .
set @xmath0 , where @xmath1 is of bounded type .
its siegel disk is denoted by @xmath25 .
we will use the following result by c.mcmullen @xcite : [ thm_mcmullen ] the postcritical set @xmath70 is uniformly measurable deep in @xmath71 . in particular
, @xmath72 is uniformly deep by lemma [ lem_deepness ] above . indeed , it is shown in @xcite that there exist two positive constants @xmath73 and @xmath74 such that for any @xmath75 , @xmath76 for sufficiently small @xmath77 . as a quasidisk .
see also ( * ? ? ?
* cor.5 , p.42 ) for a bit more elementary proof for the uniformity . ]
let @xmath78 be the regular part of @xmath7 , and @xmath28 be the invariant lift of the siegel disk @xmath25 .
we will show that for any leaf @xmath79 of @xmath80 is parabolic .
we first claim : _ any leaf @xmath79 of @xmath80 contains a backward orbit @xmath17 that stays in @xmath81 .
_ in fact , if @xmath79 is parabolic , the projection @xmath82 has at most two exceptional values by picard s theorem , and we can find such a @xmath17 . if @xmath79 is hyperbolic and @xmath83 , the claim is true because @xmath84 is an open map . in the remaining case @xmath85
is contained in a fatou component @xmath86 .
since there is no critical point in the fatou set , the projection @xmath87 has to be a conformal isomorphism that can not be extended .
this happens only when @xmath88 , because @xmath89 .
hence the claim is justified .
* case 1 . *
when @xmath90 does not accumulate on @xmath91 , the leaf @xmath92 is parabolic by a criterion of parabolicity by lyubich and minsky ( * ? ? ?
* cor.4.2 ) . * case 2 .
* now let us assume that @xmath90 accumulates on @xmath70 . since @xmath20 is contained in the basin at infinity , none of @xmath20 hits the filled julia set @xmath71 , in particular , none of @xmath20 hits @xmath26 either . by another criterion of parabolicity by lyubich and minsky (
* lem 4.4 ) , it is enough to show : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ @xmath93 , where @xmath94 is the derivative of the branch of @xmath95 sending @xmath15 to @xmath20 , and the norm is measured in the hyperbolic metric of @xmath96 . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ then @xmath92 is a parabolic leaf .
now set @xmath97 .
then @xmath20 is contained in @xmath98 for all @xmath99 .
since @xmath98 is topologically a punctured disk , it has a unique hyperbolic metric @xmath100 induced by the metric @xmath101 of constant curvature @xmath102 on the unit disk . to show the claim , it is enough to show @xmath103 where the norm in the left is measured in the hyperbolic metric @xmath104 .
here is a well - known property on @xmath104 that plays an important role ( see for example , ( * ? ? ?
* thm . 1 - 11 ) ) : [ lem_metric ] the hyperbolic metric @xmath105 of any hyperbolic domain @xmath106 is bounded by @xmath107-metric .
more precisely , @xmath108 for any @xmath109 .
we can check this by comparing @xmath110 and the hyperbolic metric on the disk of radius @xmath111 centered at @xmath112 .
of course this lemma holds for @xmath113 .
now it is enough to show : @xmath114 set @xmath115 . by assumption , @xmath116 tends to @xmath117 by taking @xmath99 in a suitable subsequence .
let @xmath118 denote the disk of radius @xmath119 centered at @xmath120 , and let @xmath121 denote the connected component of @xmath122 containing @xmath20 . since @xmath123 , we have a univalent branch @xmath124 of @xmath95 .
set @xmath125 . by the koebe @xmath126 theorem
, @xmath127 contains the disk of radius @xmath128 centered at @xmath20 , and since @xmath129 we have @xmath130 . * case 2 - 1 : * first assume that @xmath131 .
if @xmath99 ranges over a suitable subsequence , we have @xmath132 and thus ( [ eq : df ] ) holds .
( note that by the proof of ( * ? ? ?
* lem 4.4 ) we need ( [ eq : df ] ) only for a subsequence . ) * case 2 - 2 : * next consider the case when @xmath133 .
we may assume that @xmath99 ranges over a subsequence with @xmath134 . for @xmath135 ,
let @xmath136 denote the disk @xmath137 .
since @xmath138 is centered at a point in @xmath139 , we can choose an @xmath140 such that @xmath141 . by the koebe 1/4-theorem
, @xmath142 contains @xmath143 .
let us take a point @xmath144 in @xmath145 such that @xmath146 .
then we have @xmath147 and thus @xmath148 .
recall the assumption @xmath149 for @xmath150 .
this implies that the ratio @xmath151 is bounded by a positive constant from below .
however , @xmath152 by assumption and it contradicts to the uniform deepness of @xmath72 ( theorem [ thm_mcmullen ] ) . @xmath69
in this section we give an alternative proof of theorem [ thm_feigenbaum_intro ] ( ( * ? ? ? * lem 4.6 ) ) . * infinitely renormalizable quadratic maps .
* let @xmath86 and @xmath153 be a topological disk with @xmath154 .
a proper branched covering @xmath155 of degree two is called a _
quadratic - like map_. a quadratic map @xmath156 is _ infinitely renormalizable _ if there is a sequence of quadratic - like maps @xmath157 such that : @xmath158 and it is the critical point of @xmath159 ; @xmath159 is a restriction of @xmath160 on @xmath121 for some @xmath161 ; and for each @xmath99 the ratio @xmath162 is an integer @xmath163 .
* feigenbaum maps .
* an infinitely renormalizable quadratic map @xmath156 is called _ feigenbaum _ ( or _ feigenbaum - like _ ) if there exist positive constants @xmath164 and @xmath165 independent of @xmath99 such that @xmath166 and we can choose @xmath121 and @xmath167 satisfying @xmath168 .
now let us restate theorem [ thm_feigenbaum_intro ] : [ thm_feigenbaum ] the regular part @xmath21 of a feigenbaum quadratic map @xmath9 has only parabolic leaves .
to apply the same argument as theorem [ thm_siegel ] , we first show : [ prop_feigenbaum ] the postcritical set @xmath72 of a feigenbaum quadratic map @xmath9 is uniformly deep in the filled julia set @xmath71 . by mcmullen (
* thm.8.3 ) , each point in the postcritical set is a deep point of @xmath169 , so we only need to show the uniform deepness . here
we follow the idea of ( * ? ? ?
* cor . 5 ) .
let us prove it by contradiction : suppose that there exist an @xmath170 , and sequences @xmath171 and @xmath172 such that @xmath173 for all @xmath174 .
then for each disk @xmath175 we can find a disk @xmath176 in @xmath177 .
we may also assume that @xmath178 as @xmath179 . for @xmath180
set @xmath181 then @xmath182 is a closed set with @xmath183 and @xmath184 .
it is known that the postcritical set @xmath72 is a cantor set with minimal invertible dynamics @xmath185 ( * ? ? ? * chap.8 ) . in particular
, @xmath72 itself is a complete metric space by the euclidean metric on @xmath23 .
hence by the baire category theorem there exists an @xmath186 such that @xmath187 contains an open ball of @xmath72 .
more precisely , there exists a round disk @xmath153 such that @xmath188 .
next we take a neighborhood @xmath86 of @xmath64 and a univalent map @xmath189 such that @xmath190 , @xmath191 , and @xmath192 . in fact ,
when @xmath64 does not land on the critical point , we just take a small disk @xmath86 around @xmath64 and @xmath193 such that @xmath194 .
( here we used the minimality of the dynamics on @xmath72 . )
then we have a univalent map @xmath195 as desired .
otherwise , there exist a @xmath196 and a @xmath193 such that @xmath197 ( again by minimality ) with @xmath198 .
then we take a small disk @xmath86 where univalent branch @xmath199 of such @xmath200 is defined . for both cases
, we may take @xmath86 as a disk of certain radius @xmath52 centered at @xmath64 .
now let us work with the deepness : let @xmath201 denote @xmath202 and @xmath203 denote the disk @xmath204 . by the koebe distortion theorem , we can find a @xmath205 such that @xmath206 for any @xmath207 .
when @xmath208 is sufficiently large , @xmath209 is contained in @xmath210 and we have @xmath211 moreover , by the koebe 1/4 theorem , @xmath212 since @xmath213 , we have @xmath214 and thus @xmath215 this is a contradiction , since @xmath216 and @xmath217 .
@xmath69 we just follow the proof of theorem [ thm_siegel ] . take a leaf @xmath79 of @xmath21 . to show @xmath79
is parabolic we only need to check : a. we can always take a backward orbit @xmath218 contained in @xmath219 . b. the hyperbolic metric on @xmath220 is hyperbolic in order to apply lemma [ lem_metric ] . c. the postcritical set @xmath72 is uniformly deep in @xmath71 .
\(i ) is clear since @xmath71 has no interior .
( ii ) is also clear because @xmath72 contains at least two points .
( iii ) is checked by proposition [ prop_feigenbaum ] . @xmath69 * remark . * as lyubich and
minsky s original proof of theorem [ thm_feigenbaum_intro ] in ( * ? ? ?
4.6 ) , it would be possible to show theorem [ thm_siegel ] without using the deepness of the postcritical set .
the virtue of our proofs presented here is that we only concern with the geometry of the julia set and the postcritical set , and there is no need to look at the dynamics in detail .
one may expect to apply it to the cremer case .
but so far the authors do not have any example of cremer quadratic map with uniformly deep postcritical set .
let @xmath27 be a rational function of degree more than one that has a cremer fixed point @xmath29 . here is a fundamental result due to prez - marco ( see ( * ? ? ?
* thm . 1 ) and ( * ? ? ?
2 ) ) : [ thm_pm ] let @xmath118 be any neighborhood of @xmath29 where @xmath9 is univalent and there exists a univalent branch @xmath221 of @xmath222 with @xmath223 .
then for any jordan neighborhood @xmath224 of @xmath29 , there exists a compact set @xmath225 containing @xmath29 with the following properties : we say @xmath30 a _ hedgehog _ of the cremer point @xmath29 . by this theorem
we have an invertible sub - dynamics " @xmath231 .
let @xmath232 be the natural extension of @xmath27 and @xmath32 be the _ lifted hedgehog _ in @xmath233 , that is , the set of backward orbits in remaining in @xmath30 . by replacing @xmath15 by @xmath20 with sufficiently large @xmath99
, we may assume that whole pull - back @xmath16 along @xmath17 is univalent .
then @xmath236 makes sense for all @xmath99 and @xmath237 does not touch the critical set @xmath34 and its preimage @xmath238 . since @xmath9 has degree more than one , the family @xmath239 is normal and thus it has a holomorphic sequential limit @xmath240 .
if @xmath241 is not constant , then the set @xmath242 is an open set intersecting with @xmath30 .
since @xmath243 , @xmath244 is normal on @xmath242 .
this implies that @xmath242 is contained in the fatou set , but it contradicts the fact that @xmath30 is contained in the julia set . by the theorem
above , we have a fairly big set of irregular points in the natural extension generated by the cremer hedgehogs .
hence , according to ( * ? ? ?
* lem.3.3 ) , one would expect to apply the gross star theorem to find hyperbolic leaves .
however , we will see that these hedgehogs are not big enough to apply the gross star theorem .
for simplicity we consider the case when @xmath247 is a quadratic polynomial .
let @xmath50 and @xmath248 denote the postcritical set and the julia set .
( conventionally we remove @xmath249 from quadratic postcritical sets . ) for the natural extension @xmath232 , let @xmath250 denote the projection @xmath251 , the @xmath99-th entry of @xmath17 . instead of
@xmath252 we use the notation @xmath84 as before . fix any @xmath253 .
then each @xmath254 is regular in @xmath233 . in particular ,
the projection @xmath255 is locally univalent near @xmath256 . 0
contains such a @xmath257 by the same reasoning as the proof of theorem .indeed , we can take it such that @xmath258 . ]
let @xmath259 denote the half - line given by @xmath260 . by using the gross star theorem , (
* lem.3.3 ) claims : _
if @xmath261 is isomorphic to @xmath23 , then for almost every angle @xmath262 the locally univalent inverse @xmath263 has an analytic continuation along the whole half - line @xmath264 .
_ hence to show @xmath261 is hyperbolic , we should show : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ @xmath265 : _ there exist a @xmath266 and a set of @xmath267 of positive length such that the analytic continuation of @xmath263 along @xmath264 hits an irregular point at some @xmath268 .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ now we assume that @xmath265 holds in our setting
. we will claim that for almost every @xmath1 , the irregular point corresponding to the half - line @xmath264 does not belong to the lifted hedgehog @xmath32 : that is , the hyperbolic leaf is not generated by @xmath32 .
since the critical orbit @xmath269 is a countable set , the set of angles @xmath1 with @xmath270 is at most countable , i.e. , a null set of angles .
we may forget about such angles for our purpose .
so we define the set @xmath271 as follows : @xmath272
if @xmath273 and @xmath274 . since @xmath30 is a continuum , @xmath275 is a set of angles with positive length .
one more fact we need to recall is that the hedgehog @xmath30 does not intersect with @xmath276 .
( but @xmath30 is contained in @xmath277 .
see @xcite . )
in fact , if @xmath278 , the only critical point of @xmath247 eventually captured in @xmath279 for all @xmath280 . however , this contradicts the fact that larger hedgehogs are also contained in @xmath50 . since @xmath264 does not contain the points of @xmath276 , its preimage by @xmath286 consists of exactly @xmath287 disjoint curves .
we can uniquely choose one of them that touches @xmath30 , since @xmath274 , @xmath288 , and @xmath289 is a homeomorphism .
now the backward orbits remaining in such a curve for each @xmath99 form a unique path in the natural extension that has a unique initial point @xmath290 and passes though a point in @xmath32 .
this is the desired path @xmath284 .
@xmath69 recall that every @xmath272 uniquely determines @xmath284 and @xmath295 . by the proposition above ,
if there is a @xmath296 as in @xmath265 , then for any @xmath297 , the analytic continuation of the locally univalent inverse @xmath298 along @xmath299 does not hit @xmath32 .
hence it can only hit an irregular point that is _ not _ contained in @xmath32 .
let us take @xmath272 and @xmath300 with @xmath301 .
then within the bounded region enclosed by @xmath302 , we have at least one point in @xmath276 .
( actually infinitely many . )
let @xmath303 denote @xmath304 , i.e. , the starting point of the @xmath99-th pull - back of @xmath264 along @xmath32 .
now it is enough to claim : _ there exists an @xmath99 such that @xmath305 is different from @xmath306 .
_ let @xmath86 be the connected component of the region enclosed by @xmath302 containing @xmath15 on its boundary .
suppose that @xmath307 for all @xmath99 .
then we have a pull - back @xmath308 of @xmath309 that has @xmath310 on its boundary .
thus @xmath18 does not contain the critical point for all @xmath99 , and we have a univalent branch @xmath311 of @xmath95 .
since @xmath312 avoids @xmath117 , @xmath313 , and @xmath249 , @xmath314 forms a normal family . by a similar argument as in the theorem above
, any sequential limit can not be an open map and they must be constant . however , this is impossible , because the compact set @xmath315 contains a continuum of definite diameter , and the action on @xmath316 will be arbitrarily close to identity . @xmath69 | in this note we show that the regular part of the natural extension ( in the sense of lyubich and minsky @xcite ) of quadratic map @xmath0 with irrational @xmath1 of bounded type has only parabolic leaves except the invariant lift of the siegel disk .
we also show that though the natural extension of a rational function with a cremer fixed point has a continuum of irregular points , it can not supply enough singularity to apply the gross star theorem to find hyperbolic leaves . |
the exponential function is often appeared in every scientific field . among many properties of the exponential function ,
the linear differential function @xmath1 is the most important characterization of the exponential function .
a slightly nonlinear generalization of this linear differential equation is given by@xmath2(see the equation ( 17 ) at page 5 of @xcite and the equations ( 22)-(23 ) at page 8 of @xcite . )
this nonlinear differential equation is equivalent to@xmath3then we define the so - called @xmath0__-logarithm _ _ @xmath4@xmath5as a generalization of @xmath6 . applying the property : @xmath7 to ( [ diffequ1 ] )
, we obtain@xmath8where @xmath9 is _ any _ constant @xcite .
then we define the so - called @xmath0__-exponential _ _ @xmath10 as the inverse function of @xmath11 as follows:@xmath12 ^{\frac{1}{1-q } } & \text{if } 1+\left ( 1-q\right ) x>0 , \\ 0 & \text{otherwise.}% \end{array}% \right .
\label{q - exponential}\]]note that the @xmath0-logarithm and @xmath0-exponential recover the usual logarithm and exponential when @xmath13 , respectively ( see the pages 84 - 87 of @xcite for the detail properties of these generalized functions @xmath14 and @xmath10 ) .
thus , the general solution to the nonlinear differential equation ( [ nonlinear differential equation ] ) becomes@xmath15where @xmath9 is _ any _ constant satisfying @xmath16 dividing the both sides by @xmath17 of the above solution , we obtain @xmath18under the following scaling:@xmath19we obtain@xmath20this means that the solution of the nonlinear differential equation ( nonlinear differential equation ) obtained above is _ scale - invariant _ under the above scaling ( [ scaling ] ) .
moreover , we can choose _ any _ constant @xmath9 satisfying @xmath21 because @xmath9 is an integration constant of ( [ diffequ1 ] ) .
note that the above scaling ( [ scaling ] ) with respect to both variables @xmath22 and @xmath23 can be observed only when @xmath24 and @xmath25 .
in fact , when @xmath26 i.e. , @xmath27 ( [ scaling ] ) reduces to the scaling with respect to only @xmath28 i.e. , @xmath29 , and when @xmath30 both scalings in ( [ scaling ] ) disappears @xcite .
the above scaling property of the nonlinear differential equation ( nonlinear differential equation ) is very significant in the fundamental formulations for every generalization based on ( [ nonlinear differential equation ] ) .
we summarize the important points in the above fundamental result . 1
. @xmath9 in the scaling ( [ scaling ] ) is _ any _ constant satisfying @xmath21 because @xmath9 is an integration constant of ( diffequ1 ) .
this means that the scaling ( [ scaling ] ) is _ arbitrary _ for _ any _ @xmath0 and @xmath9 if @xmath0 and @xmath9 satisfies @xmath31 @xmath32 .
2 . in general studies of differential equation , @xmath9 is determined by the _
initial condition _ of the nonlinear differential equation ( nonlinear differential equation ) .
this means , when an observable in a dynamics grows according to the nonlinear differential equation ( nonlinear differential equation ) , _ the initial condition determines the scaling of the dynamics .
_ this is applicable to the analysis of the chaotic dynamics @xcite@xcite@xcite@xcite@xcite@xcite@xcite .
3 . in general , for a mapping @xmath33 , @xmath34 is called a restriction of a mapping @xmath35 to @xmath36 , which is denoted by @xmath37 .
let @xmath38 be a @xmath0-exponential function ( [ q - exponential ] ) from @xmath39 .
for the restricted domain @xmath40 defined by@xmath41a restriction of a mapping @xmath38 to @xmath42 is denoted by@xmath43then @xmath44 becomes a power function:@xmath45 in the above formulations the only case @xmath46 is discussed . as shown in the section iv , the @xmath0-generalizations along the line of ( [ nonlinear differential equation ] )
has a symmetry @xmath47 , i.e. , @xmath48 .
therefore , the above @xmath49 can be replaced by a restriction of a mapping @xmath50 to @xmath51 @xmath52 in accordance with the symmetry , which implies that the case @xmath53 can be discussed . in this way ,
the _ restriction _ of the @xmath0-exponential function to the domain @xmath54 or @xmath55 coincide with a power function , which has been often appeared and discussed in science .
in general , the restricted domain @xmath42 or @xmath56 is called _ scaling domain _ and its corresponding range is called _ scaling range_. 4 .
a power function @xmath57 is known to be characterized by the following functional equation , i.e. , there exists a function @xmath58 such that @xmath59holds for any @xmath60 .
the above functional equation uniquely determines a power function @xmath61for choosing @xmath62 .
see some references such as @xcite for the proof . on the other hand ,
a @xmath0-exponential function is characterized by the nonlinear differential equation ( [ nonlinear differential equation ] ) as similarly as a exponential function . moreover ,
the solutions of ( [ nonlinear differential equation ] ) are scale - invariant under the scaling ( [ scaling ] ) and reduce to power functions when the domain is restricted to @xmath42 or @xmath56 as shown above .
in fact , by restricting the domain of ( [ scaling-1 ] ) to @xmath63 the general solution ( [ scaling-1 ] ) of the nonlinear differential equation ( [ nonlinear differential equation ] ) reduces to the following power function according to ( [ f_q]).@xmath64that is,@xmath65which is equivalent to ( [ power - func ] ) if @xmath66therefore , many discussions on exponential versus power - law , i.e. , @xmath67 versus @xmath68 should be replaced by exponential versus @xmath0-exponential , i.e. , @xmath69 versus @xmath70 , which is more natural from mathematical point of view . as shown in these discussions , the fundamental nonlinear differential equation ( [ nonlinear differential equation ] ) provides us with not only the characterization of the @xmath0-exponential function but also the scaling property in its solution .
as similarly as the relation between the exponential function @xmath71 and shannon entropy , we expect the corresponding information measure to the @xmath0-exponential function @xmath10 .
there exist some candidates such as rnyi entropy , tsallis entropy and so on .
but the algebra derived from the @xmath0-exponential function uniquely determines tsallis entropy as the corresponding information measure . in the following sections of this paper , we present the two mathematical results to uniquely determine tsallis entropy by means of the already established formulations such as the @xmath0-exponential law , the @xmath0-multinomial coefficient and @xmath0-stirling s formula .
the exponential law plays an important role in mathematics , so this law is also expected to be generalized based on the @xmath0-exponential function @xmath72 . for this purpose ,
the new multiplication operation @xmath73 is introduced in @xcite and @xcite for satisfying the following identities:@xmath74the concrete form of the @xmath0-logarithm or @xmath0-exponential has been already given in the previous section , so that the above requirements as @xmath0-exponential law leads us to the definition of @xmath75 between two positive numbers .
for two positive numbers @xmath76 and @xmath23 , the @xmath0-_product _
@xmath75 is defined by@xmath77 ^{\frac{1}{1-q } } , & \text{if } % x>0,\,y>0,\,x^{1-q}+y^{1-q}-1>0 , \\ 0 , & \text{otherwise.}% \end{array}% \right .
\label{def of q - product}\ ] ] the @xmath0-_product _ recovers the usual product such that @xmath78 .
the fundamental properties of the @xmath0-product @xmath75 are almost the same as the usual product , but @xmath79the other properties of the @xmath0-_product _ are available in nmw03 and @xcite . in order to see one of the validities of the @xmath0-product ,
we recall the well known expression of the exponential function @xmath80 given by@xmath81replacing the power on the right side of ( [ def of expx ] ) by the @xmath82 times of the @xmath0-product @xmath83@xmath84@xmath10 is obtained . in other words ,
@xmath85 coincides with @xmath86 @xmath87the proof of ( [ repre of q - exp ] ) is given in the appendix of @xcite .
this coincidence ( [ repre of q - exp ] ) indicates a validity of the @xmath0-product .
in fact , the present results in the following sections reinforce it .
we briefly review the @xmath0-multinomial coefficient and the @xmath0-stirling s formula by means of the @xmath0-product @xmath75 .
as similarly as for the @xmath88-product , @xmath0__-ratio _ _ is introduced as follows : for two positive numbers @xmath76 and @xmath23 , the inverse operation to the @xmath0-product is defined by @xmath89 ^{\frac{1}{1-q } } , & \text{if } % x>0,\,y>0,\,x^{1-q}-y^{1-q}+1>0 , \\ 0 , & \text{otherwise}% \end{array}% \right.\]]which is called @xmath0__-ratio _ _ in @xcite .
@xmath90 is also derived from the following satisfactions , similarly as for @xmath91 @xcite@xcite . @xmath92
the @xmath0-product and @xmath0-ratio are applied to the definition of the @xmath0-multinomial coefficient @xcite . for @xmath93 and @xmath94 the @xmath0-multinomial coefficient
is defined by@xmath95 _
{ q}:=\left ( n!_{q}\right ) \oslash _ { q}\left [ \left ( n_{1}!_{q}\right ) \otimes _ { q}\cdots \otimes _ { q}\left ( n_{k}!_{q}\right ) % \right ] .
\label{def of q - multinomial coefficient}\ ] ] from the definition ( [ def of q - multinomial coefficient ] ) , it is clear that@xmath96 _ { q}=\left [ \begin{array}{ccc } & n & \\
n_{1 } & \cdots & n_{k}% \end{array}% \right ] = \frac{n!}{n_{1}!\cdots n_{k}!}.\ ] ] in addition to the @xmath0-multinomial coefficient , the @xmath0-stirling s formula is useful for many applications such as our main results . by means of the @xmath0-product ( [ def of q - product ] ) , the @xmath0-factorial @xmath97
is naturally defined as follows . for a natural number @xmath98 and @xmath99 ,
the @xmath0-factorial @xmath100 is defined by@xmath101 using the definition of the @xmath0-product ( [ def of q - product ] ) , @xmath102 is explicitly expressed by @xmath103if an approximation of @xmath104 is not needed , this explicit form should be directly used for its computation . however , in order to clarify the correspondence between the studies @xmath105 and @xmath24 , the approximation of @xmath106 is useful .
in fact , using the following @xmath0-stirling s formula , we obtain the unique generalized entropy corresponding to the @xmath0-exponential function @xmath10 , shown in the following sections .
let @xmath100 be the @xmath0-factorial defined by ( [ def of q - kaijyo ] ) .
the rough @xmath0-stirling s formula @xmath106 is computed as follows:@xmath107 the proof of the above formulas ( [ rough q - stirling ] ) is given in su04b .
in this section we show that tsallis entropy is uniquely and naturally derived from the fundamental formulations presented in the previous section . in order to avoid separate discussions on the positivity of the argument in ( [ def of q - multinomial coefficient ] ) , we consider the @xmath0-logarithm of the @xmath0-multinomial coefficient to be given by @xmath108 _ { q}=\ln _ { q}\left ( n!_{q}\right ) -\ln _ { q}\left ( n_{1}!_{q}\right ) \cdots -\ln _ { q}\left ( n_{k}!_{q}\right ) .
\label{lnq of multinomial}\]]the unique generalized entropy corresponding to the @xmath0-exponential is derived from the @xmath0-multinomial coefficient using the @xmath0-stirling s formula as follows @xcite .
when @xmath82 is sufficiently large , the @xmath0-logarithm of the @xmath0-multinomial coefficient coincides with tsallis entropy ( [ d - tsallis entropy ] ) as follows:@xmath108 _
{ q}\simeq \left\ { \begin{array}{ll } \dfrac{n^{2-q}}{2-q}\cdot s_{2-q}\left ( \dfrac{n_{1}}{n},\cdots , \dfrac{n_{k}% } { n}\right ) & \text{if}\quad q>0,\,\,q\neq 2 \\ -s_{1}\left ( n\right ) + \sum\limits_{i=1}^{k}s_{1}\left ( n_{i}\right ) & \text{% if}\quad q=2% \end{array}% \right .
\label{important0 - 2}\]]where @xmath109 is tsallis entropy @xcite:@xmath110and @xmath111 is given by @xmath112 the proof of this theorem is given in @xcite .
note that the above relation ( [ important0 - 2 ] ) reveals a surprising _ symmetry _ : ( [ important0 - 2 ] ) is equivalent to@xmath113 _
{ 1-\left ( 1-q\right ) } \simeq \frac{n^{1+\left ( 1-q\right ) } } { % 1+\left ( 1-q\right ) } \cdot s_{1+\left ( 1-q\right ) } \left ( \frac{n_{1}}{n}% , \cdots , \frac{n_{k}}{n}\right ) \label{symmetry}\]]for @xmath114 and @xmath115 .
this expression represents that there exists a _ symmetry _ with a factor
@xmath116 around @xmath105 in the algebra of the @xmath0-product .
substitution of some concrete values of @xmath0 into ( important0 - 2 ) or ( [ symmetry ] ) helps us understand the meaning of this symmetry .
remark that the above correspondence ( [ important0 - 2 ] ) and the symmetry ( [ symmetry ] ) reveals that the @xmath0-exponential function ( q - exponential ) derived from([nonlinear differential equation ] ) is consistent with tsallis entropy only as information measure .
this section shows another way to uniquely determine the generalized entropy .
more precisely , the identity derived from the @xmath0-multinomial coefficient coincides with the generalized shannon additivity which is the most important axiom for tsallis entropy .
consider a partition of a given natural number @xmath82 into @xmath117 groups such as @xmath118 .
in addition , each natural number @xmath119 is divided into @xmath120 groups such as @xmath121 where @xmath122 .
[ ptbh ] fig1.eps then , the following identity holds for the @xmath0-multinomial coefficient.@xmath123 _
{ q}=\left [ \begin{array}{ccc } & n & \\ n_{1 } & \cdots & n_{k}% \end{array}% \right ] _ { q}\otimes _ { q}\left [ \begin{array}{ccc } & n_{1 } & \\
n_{11 } & \cdots & n_{1m_{1}}% \end{array}% \right ] _ { q}\otimes _ { q}\cdots \otimes _ { q}\left [ \begin{array}{ccc } & n_{k } & \\
n_{k1 } & \cdots & n_{km_{k}}% \end{array}% \right ] _ { q } \label{q - identity}\]]it is very easy to prove the above relation ( [ q - identity ] ) by taking the @xmath0-logarithm of the both sides and using ( [ lnq of multinomial ] ) . on the other hand , the above identity ( [ q - identity ] )
is reformed to the generalized shannon additivity in the following way . taking the @xmath0-logarithm of the both sides of the above relation ( [ q - identity ] )
, we have @xmath124 _ { q}=\ln _ { q}\left [ \begin{array}{ccc } & n & \\ n_{1 } & \cdots & n_{k}% \end{array}% \right ] _ { q}+\sum_{i=1}^{k}\ln _ { q}\left [ \begin{array}{ccc } & n_{i } & \\ n_{i1 } & \cdots & n_{im_{i}}% \end{array}% \right ] _ { q}.\]]from the relation ( [ important0 - 2 ] ) , we obtain@xmath125then , by means of the following probabilities defined by@xmath126the identity ( [ n - gene - shanaddi ] ) becomes@xmath127the formula ( [ generalizedshannon ] ) obtained from the @xmath0-multinomial coefficient is exactly same as the generalized shannon additivity ( see [ gsk3 ] given below ) which is the most important axiom for tsallis entropy @xcite .
in fact , the generalized shannon - khinchin axioms and the uniqueness theorem for the nonextensive entropy are already given and rigorously proved in su04d . the present result ( [ generalizedshannon ] ) and the already established axiom [ gsk3 ]
perfectly coincide with each other .
let @xmath128 be defined by the @xmath82-dimensional simplex : @xmath129the following axioms [ gsk1]@xmath130[gsk4 ] determine the function @xmath131 such that @xmath132where @xmath133 satisfies properties ( i ) @xmath130 ( iv ) : 1 .
@xmath133 is continuous and has the same sign as @xmath134i.e .
, @xmath135 2 .
@xmath136 3 .
there exists an interval @xmath137 such that @xmath138 and @xmath133 is differentiable on the interval @xmath139 4 .
there exists a constant @xmath140 such that@xmath141 [ gsk1 ] _ continuity _ :
@xmath109 is continuous in @xmath128 and @xmath99 , [ gsk2 ] _ maximality _ : for any @xmath99 , any @xmath142 and any @xmath143 , @xmath144 [ gsk3 ] _ generalized shannon additivity _ : if @xmath145then the following equality holds:@xmath146 [ gsk4 ] _ expandability _ :
@xmath147 note that , in order to uniquely determine the tsallis entropy ( d - tsallis entropy ) in the above set of the axioms , @xmath148 should be removed from ( [ constrant4 ] ) , that is , @xmath149 ( i.e. , @xmath150 ) should be used instead of ( [ constrant4 ] ) .
the general form @xmath151 perfectly corresponds to tsallis original introduction of the so - called tsallis entropy in 1988 @xcite . see his original characterization shown in page 9 of @xcite for the detail ( @xmath152 corresponds to @xmath153 in his notation .
his simplest choice of @xmath153 coincides with the simplest form of @xmath152 i.e. , @xmath154 ) .
when one of the authors ( h.s . ) submitted the paper @xcite in 2002 , nobody presented the idea of the @xmath0-product .
however , as shown above , the identity on the @xmath0-multinomial coefficient @xcite which was formulated based on the @xmath0-product @xcite@xcite coincides with one of the axioms ( _ _ _ _ [ gsk3 ] : generalized shannon additivity ) in @xcite .
this means that the whole theory based on the @xmath0-product is self - consistent .
moreover , other fundamental applications of the @xmath0-product , such as law of error @xcite and the derivation of the unique non self - referential @xmath0-canonical distribution @xcite@xcite , are also based on the @xmath0-product .
starting from a fundamental nonlinear equation @xmath155 , we present the scaling property and the algebraic structure of its solution .
moreover , we prove that the algebra determined by its solutions is mathematically consistent with tsallis entropy only as the corresponding unique information measure based on the following 2 mathematical reasons : ( 1 ) derivation of tsallis entropy from the @xmath0-multinomial coefficient and @xmath0-stirling s formula , ( 2 ) coincidence of the identity derived from the @xmath0-multinomial coefficient with the generalized shannon additivity which is the most important axiom for tsallis entropy . 1
. axioms and the uniqueness theorem for the nonextensive entropy su04d 2 .
law of error in tsallis statistics @xcite 3 .
@xmath0-stirling s formula in tsallis statistics @xcite 4 .
@xmath0-multinomial coefficient in tsallis statistics @xcite 5 .
central limit theorem in tsallis statistics ( numerical evidence only ) @xcite 6 .
@xmath0-pascal s triangle in tsallis statistics @xcite 7 . the unique non self - referential @xmath0-canonical distribution in tsallis statistics @xcite@xcite 8 .
scaling property characterized by the fundamental nonlinear differential equation [ the present paper ] .
all of the above fundamental results are derived from the algebra of the @xmath0-product and mathematically consistent with each other .
this means that the @xmath88-product is indispensable to the formalism in tsallis statistics .
more important point is that the @xmath0-product originates from the fundamental nonlinear differential equation ( [ nonlinear differential equation ] ) with _ scale - invariant _ solutions . in any manuscripts and papers up to now ,
this constant @xmath9 is set to be @xmath156 , i.e. , @xmath157 ( see the equation ( 17 ) at page 5 of @xcite and the equations ( 22)-(23 ) at page 8 of @xcite ) .
but the arbitrariness of this constant @xmath9 plays a very important role in scaling ( [ scaling ] ) .
h. suyari , the unique non self - referential @xmath0-canonical distribution and the physical temperature derived from the maximum entropy principle in tsallis statistics , prog .
* 162 * , pp.79 - 86 ( 2006 ) .
( see the author s website ( http://www.ne.jp/asahi/hiroki/suyari/publications.htm ) for the detail derivation . ) | we derive a scaling property from a fundamental nonlinear differential equation whose solution is the so - called @xmath0-exponential function .
a scaling property has been believed to be given by a power function only , but actually more general expression for the scaling property is found to be a solution of the above fundamental nonlinear differential equation .
in fact , any power function is obtained by restricting the domain of the @xmath0-exponential function appropriately .
as similarly as the correspondence between the exponential function and shannon entropy , an appropriate generalization of shannon entropy is expected for the scaling property .
although the @xmath0-exponential function is often appeared in the optimal distributions of some one - parameter generalized entropies such as rnyi entropy , only tsallis entropy is uniquely derived from the algebra of the @xmath0-exponential function , whose uniqueness is shown in the two ways in this paper . |
herbig ae ( hae ) stars are widely recognized as intermediate mass counterparts of low mass t tauri stars ( ttss ) @xcite . in the last few decades , it has been revealed that these stars are commonly accompanied by a circumstellar disk like ttss by direct imaging at infrared and radio wavelengths @xcite .
these disks are believed to be the precursors of planetary systems around hae stars , and are the important targets to study the formation process of a planetary system around an intermediate - mass star and to construct a general scenario for the formation of planetary systems .
recent advanced instruments have revealed their physical properties , such as the disk mass , the radial temperature and surface density distributions , the outer radius of the disk , and the velocity field .
the number of well - studied disks around hae stars , however , is still insufficient to establish a firm understanding of their physical conditions and their statistics ; only a few of limited examples of their structure have deeply been analyzed ( e.g. , @xcite ) .
temperature distribution along the vertical direction of the disk should also be taken into account when one examines how the gas and dust inside the disk evolve .
theoretical studies predict that flared circumstellar disks should have two or more layers with distinct temperature in vertical direction that are formed by diffusion of stellar radiation scattered at the disk surface @xcite .
to examine vertical structure of the disks based on observations , it requires examples as many as possible , though several studies have already tried to make vertical temperature or density distributions unveiled @xcite .
hd163296 is one of the best targets for this purpose because it exhibits strong emission from the disk in many co isotopologue lines .
hd163296 is a star with a3v in spectral type @xcite .
its luminosity , mass , age , and distance are 30 @xmath5 , 2.3 @xmath6 , 4 myr , and 122 pc , respectively @xcite .
observations at infrared and radio wavelengths revealed the existence of a circumstellar disk with its mass of 0.028 @xmath6 and 45@xmath7 in inclination angle @xcite .
hd163296 is categorized in group ii by the feature of spectral energy distribution ( sed ) and expected to have a self - shadowed disk @xcite .
more recently , imaging studies with iram / pbi , sma , and vla both in continuum and in the @xmath0co , @xmath2co , and c@xmath8o emission lines were made @xcite , arguing that the disk radius is 550au and that the gas kinematics can be well explained by keplerian rotation .
this paper presents multi - line observations of hd163296 with the nobeyama 45-meter ( nro 45 m ) radio telescope and the atacama submillimeter telescope experiment ( aste ) sub - millimeter telescope , and discusses vertical temperature distribution inside of the disk .
the outline is as follows : details of the observations are described in section [ sec : observation ] , and the results are shown in section [ sec : results ] . in section [ sec : model fitting ] ,
physical properties of the disk are estimated by model fitting in which power - law distributions of temperature and surface density are assumed .
interpretation of the model fitting results as well as the comparisons with previous interferometric observations are discussed in section [ sec : discussion ] .
observations of @xmath0co ( @xmath1=10 ) emission line were performed in winter 2006 by the nro 45 m telescope , operated by nobeyama radio observatory ( nro ) , a branch of national astronomical observatory of japan ( naoj ) . to separate the emission originated from the disk associated with the central star from surrounding components such as an envelope or a remnant cloud ,
a profile map including the stellar position and its vicinity with sufficient spacial resolution is decisive in identifying the origin of the emission . to obtain such a map ,
the multi - beam receiver , 25 beam array receiver system ( bears : @xcite ) that is capable of simultaneously detecting 25 different locations in double sideband ( dsb ) operation was employed as the frontend and tuned at the rest frequency of @xmath0co ( @xmath1=10 ) , 115.271 ghz .
the half - power beam width ( hpbw ) is 15@xmath9 at this frequency , corresponding to 1800 au at a distance of 122 pc from the sun .
the beam grid spacing of bears is 40@xmath9 ( @xmath10 5000au ) in both r.a . and
dec . directions , but this is much larger than the hpbw and therefore is insufficient to isolate the star+disk system from the ambient . to fill these gaps among the beams , we also took data at positions shifted by 20@xmath9 in both r.a . and
dec . directions from the central star .
the integration time for the observation was 7 hours .
we only show 80@xmath9 @xmath11 80@xmath9 ( @xmath10 0.05 pc @xmath11 0.05 pc ) region near the star that covers a typical size of the envelope .
a digital spectrometer with 32 mhz in bandwidth and 37.8 khz in frequency resolution was used as the backend .
correction for atmospheric absorption during observations was made by the chopper - wheel method @xcite , and the intensities were obtained in antenna temperature , @xmath12 , in kelvin .
correction for the main beam efficiency , @xmath13 , was made when we compared model calculation results ( section [ sec : model fitting ] ) with the brightness temperature , @xmath14 , where @xmath15 for @xmath0co ( @xmath1=10 ) emission line .
the dsb system temperature in @xmath0co ( @xmath1=10 ) mapping observation was between 250 k and 600 k. the accuracy for telescope pointing was regularly checked against the sio maser vx sgr . , which was about 30@xmath16 away from hd163296 , and pointing deviation was achieved in less than @xmath173@xmath9 during whole observation period .
data reduction and analysis were made with the newstar software package developed by nro , which is a front end of aips developed in the national radio astronomy observatory ( nrao ) .
observations were made in position - switching mode .
it was found that the extended co ( @xmath1=10 ) emission was present around the region of interest , hence careful search is needed to find an off point as close to hd163296 as possible .
the search for off points was conducted with bears based on the fact that the systemic velocity of hd163296 is 6 km s@xmath18 @xcite . the emission free position applicable as an off point
was finally found at about 10@xmath16 north from hd163296 .
figure [ fig : figure1 ] is a profile map toward this position , centered at and ( j2000 ) and covering 80@xmath9 @xmath11 80@xmath9 ( @xmath10 0.05 pc @xmath11 0.05 pc ) region , showing that there is no prominent emission detected around 6 km s@xmath18 in @xmath19 .
( 180mm,120mm)figure1.eps observations of @xmath2co ( @xmath1=10 ) emission line were conducted by the nro 45 m telescope in winter 2008 . as will be shown in section [ sec :
result of 12co(j=1 - 0 ) mapping observations ] , the @xmath0co ( @xmath1=10 ) emission detected around 6 km s@xmath18 in @xmath19 at the central star is isolated .
therefore , the @xmath2co ( @xmath1=10 ) observations were made only at the stellar position by a pair of single - beam receivers that enable to simultaneously detect both the polarization components in single sideband ( ssb ) mode .
the receiver was tuned at @xmath2co ( @xmath1=10 ) rest frequency of 110.201 ghz and the hpbw is 15@xmath9 at this frequency .
the @xmath13 in 2008 at this frequency was 0.4 .
acousto optical spectrometer ( aos ) whose bandwidth and frequency resolution are 40 mhz and 20 khz respectively was used .
the system noise temperature during observation was between 300 k and 600 k in ssb mode .
the pointing was regularly checked by the same manner as in @xmath0co ( @xmath1=10 ) observations described in section [ sec:12co(j=1 - 0 ) mapping observation ] .
the same off point as in the @xmath0co ( @xmath1=10 ) observations was used .
all observational data when the wind speed exceeded 5 m s@xmath18 were flagged out before obtaining the final profile of @xmath2co ( @xmath1=10 ) emission .
the resultant total integration time was 8.5 hours on source .
data reduction and analysis were made in the same manner as in @xmath0co ( @xmath1=10 ) observation in section [ sec:12co(j=1 - 0 ) mapping observation ] .
observations of @xmath0co and @xmath2co ( @xmath1=32 ) line were carried out in july 2006 with atacama submillimeter telescope experiment ( aste ) , a 10 m sub - millimeter telescope in chile @xcite .
hpbws of the aste telescope at 345.79 and 330.59 ghz were 22@xmath9 and 23@xmath9 , respectively , which corresponds to about @xmath20 au at the distance to the target , and the @xmath13 was 0.6 around these frequencies .
sis mixer receiver having dsb response was employed .
typical atmospheric opacity at 220 ghz toward the zenith was 0.05 during the observations , and the system noise temperature of the telescope was @xmath21 k in dsb during observations .
the digital auto - correlator with 1024 frequency channels was configured so that the total bandwidth was 128 mhz , resulting in a frequency resolution of 125 khz , or @xmath22 0.11 km s@xmath18 at the observed frequencies .
telescope pointing was checked every @xmath23 hours by cross scan of jupiter , and the error was proved to be @xmath24 .
all the spectra were obtained in a position - switching mode : the location of the off - position was @xmath25 from the star . in @xmath0co ( @xmath1=32 ) observations , we took spectra not only at the stellar position but also at the four adjacent points , 22@xmath9 apart from the stellar position in r.a . or dec .
direction , to check out whether the emission of the stellar position is really isolated .
@xmath2co ( @xmath1=32 ) data were taken only toward the stellar position since the @xmath0co ( @xmath1=32 ) at the systemic velocity ( @xmath26 km s@xmath18 ) was proven to be isolated , as will be shown in section [ sec : results of aste observations ] .
data reduction and analysis were conducted in the same manner as in @xmath0co ( @xmath1=10 ) observations described in section [ sec:12co(j=1 - 0 ) mapping observation ] .
figure [ fig : figure2 ] shows a @xmath0co ( @xmath1=10 ) profile map within the 80@xmath9 @xmath11 80@xmath9 region centered at hd163296 .
a clear double - peaked profile with 4.8@xmath27 is seen only at the center panel at 6 km s@xmath18 in @xmath19 that matches with the systemic velocity of hd163296 @xcite .
the double - peaked profile has peak intensity of 1.0 k in @xmath28 and [email protected] km s@xmath18 in fwhm velocity width .
the achieved mean rms noise level after data reduction was 0.21 k. figure [ fig : figure3]a displays the same profile on the central panel in figure [ fig : figure2 ] in an enlarged view .
several positions in figure [ fig : figure2 ] , especially 40@xmath9 away to the west from the center , show an emission feature at @xmath29 6 km s@xmath18 , but there seems no emission above the 3@xmath27 level at this velocity at four positions adjacent to the star ( 20@xmath9 apart from the star ) . to confirm whether there is no significant emission at 6 km s@xmath18 in these regions ,
these four profiles were averaged into one spectrum ( figure [ fig : figure4]a ) .
there is no emission above the 3@xmath27 level at 6 km s@xmath18 in figure [ fig : figure4]a .
the profile detected at the stellar position ( figure [ fig : figure4]b , or the central panel in figure [ fig : figure2 ] ) is compared with that after subtracting the profile in figure [ fig : figure4]a ( figure [ fig : figure4]c ) .
although the noise level in figure [ fig : figure4]c increases due to subtraction , there is no remarkable difference between these two profiles .
these comparisons strongly suggest that the double - peaked emission component at the star is isolated and originates from the compact circumstellar disk in keplerian rotation .
the components at @xmath29 6 km s@xmath18 in south and west regions @xmath30 apart from the star in figure [ fig : figure2 ] might be the emissions from a remnant cloud partially remained around the star+disk system .
( 180mm,120mm)figure2.eps the dip at 8 km s@xmath18 around the star+disk system ( figure [ fig : figure2 ] ) is due to weak emission seen at the off points , especially in the northern region ( figure [ fig : figure1 ] ) ; in position switching mode , the emission at the off point appears like an absorption dip .
the dip dose not affect further analysis since it does not corrupt the original emission profile at 6 km s@xmath18 .
in addition , the strong emission components between 10 and 15 km s@xmath18 and at 20 km s@xmath18 appear regardless of any positions : these two components probably come from foreground or background clouds that exist coincidentally along a line of sight .
( 160mm,110mm)figure3.eps ( 80mm,110mm)figure4.eps figure [ fig : figure3]b shows the profile of @xmath2co ( @xmath1=10 ) emission line .
again , a double - peaked profile was clearly detected .
the peak intensity and velocity width are 0.36 k in @xmath28 and [email protected] km s@xmath18 at fwhm , respectively .
the achieved rms noise level is 0.061 k. since the @xmath0co ( @xmath1=10 ) and @xmath2co ( @xmath1=10 ) emissions are detected in the same velocity range , both emissions should have the same origin . the intensity ratio between @xmath0co(@xmath1=10 ) and @xmath2co(@xmath1=10 ) is less than 5 , suggesting that the disk is optically thick . @xcite evaluated the optical depth of the co emission lines and concluded that the disk is optically thick .
our model fitting shown in section [ sec : fitting results ] also provided that the disk is optically thick for both the co lines .
figure [ fig : figure5 ] shows a @xmath0co ( @xmath1=32 ) profile map within the 44@xmath9 @xmath11 44@xmath9 region centered at hd163296 ; the central panel of figure [ fig : figure5 ] is also shown in figure [ fig : figure3]c in an enlarged view .
the achieved rms level was 0.122 k. a clear double - peaked profile at @xmath316 km s@xmath18 is seen only in the center panel .
the peak intensity and velocity width are 1.35 k in @xmath28 and [email protected] km s@xmath18 in fwhm , respectively .
the emission components between 10 and 15 km s@xmath18 and at @xmath2220 km s@xmath18 in @xmath19 appear regardless of positions as in the case of @xmath0co ( @xmath1=10 ) ( figure [ fig : figure2 ] ) .
figure [ fig : figure3]d presents the @xmath2co ( @xmath1=32 ) spectrum at the stellar position , showing its double - peaked profile .
the peak intensity and velocity width are 0.38 k in @xmath28 and 3.85 km s@xmath18 in fwhm , respectively .
as in the case of the @xmath1=10 lines , the disk seems optically thick for @xmath1=32 line from the intensity ratio between @xmath0co ( @xmath1=32 ) and @xmath2co ( @xmath1=32)(see also section [ sec : fitting results ] ) .
model fitting is one of the effective methods to estimate physical properties of disks around hae / be stars @xcite .
we fitted all the observational profiles , which are two co isotopologues emission lines with two different rotational transitions , with those calculated by the disk model ( details in section [ sec : disk model ] ) .
we adopted a classical disk model which includes power - law forms of temperature and surface density distributions in radial direction with inner and outer cutoffs @xcite .
the results are compared with the ones obtained by millimeter and sub - millimeter interferometers ( @xcite , details in section [ sec : interferometers ] ) .
although the interferometric observations can provide detailed geometrical structure with high spacial resolution , it might be unable to detect the emission from the diffuse gas uniformly extending away due to spatial filtering effect .
the single dish telescopes with large diameter , on the other hand , have higher sensitivity to the weak extended emission which interferometers can not detect well .
interferometric and single dish observations are , therefore , complementary to each other , and comparing their results is a fruitful way to gain a better understanding of the disk nature .
several assumptions for the disk structure and parameters are applied in the model fitting .
first , the disk is assumed to have keplerian rotation as was proven to be the cases in other hae stars such as hd169142 and hd31648 @xcite .
@xcite pointed out the existence of keplerian rotating disk around hd163296 from `` a butterfly shape '' seen in its position velocity diagram .
secondly , the temperature distribution over the disk , @xmath32 , is assumed to follow a power - law form of the radius with its index of @xmath33 , as described by the following equation : @xmath34 where @xmath35 is the radius in cylindrical coordinate system centered on the star and @xmath36 is the temperature at @xmath37 au @xcite .
this assumption implies that the temperature is constant along the vertical direction ( @xmath38 ) at a certain @xmath35 . in reality ,
a circumstellar disk should have temperature distribution along the @xmath38-direction . an elaborate disk model ( e.g. , @xcite )
predicts that the surface regions where the stellar radiation can penetrate and heat directly will have higher temperature than that in the interior regions , forming 2-layer temperature structure .
single layer temperature model , however , can still be applicable because an optically thick line should have `` photosphere '' at each @xmath35 and its characteristic temperature can uniquely be determined .
that is , vertical temperature distribution can be traced by a series of a different temperature of each co emission line .
thirdly , the radial distribution of surface density over the disk , @xmath39 , is also assumed to follow a power - law form of the radius with its index of @xmath40 @xcite , as described by the following equation : @xmath41 where @xmath42 is the surface density at @xmath37 au .
fourthly , we assumed that density distribution in vertical direction can be achieved by hydrostatic equilibrium .
the density distribution @xmath43 is , therefore , expressed by @xmath44\ , \label{eq : rho}\ ] ] where @xmath45 is the density at the mid - plane of the disk , and @xmath46 is the scale height given by @xmath47 where @xmath48 is boltzmann constant , @xmath49 is gravitational constant , @xmath50 is the mass of the central star , and @xmath51 is the mean mass of gas molecules .
thus , @xmath45 in equation ( [ eq : rho ] ) becomes @xmath52 fifthly , local thermodynamic equilibrium ( lte ) is assumed .
the number density of h@xmath53 at the position of photosphere of each co line can be estimated by the equation ( [ eq : rho ] ) . as shown in section [ sec : fitting results ] , the number density of h@xmath53 at the @xmath0co ( @xmath1=32 ) photosphere is calculated to be @xmath22 10@xmath54 @xmath55 at @xmath56 100 au , and it is higher than the typical value of critical density for @xmath0co ( @xmath1=32 ) ( @xmath22 5 @xmath11 10@xmath57 @xmath55 at 60 k , fitting temperature of this line discussed in section [ sec : fitting results ] ) by more than two orders of magnitude .
lte assumption is sometimes inadequate for analysis of protoplanetary disks since it tends to overestimate the excitation temperature for a higher transition with a high critical density .
however , @xcite evaluated many types of model calculations and found that lte approximation seems valid for the cases of @xmath1=43 or a lower transition of co. the lte assumption is , therefore , applicable in our model calculation for the @xmath1=10 and @xmath1=32 transitions of @xmath0co and @xmath2co .
finally , the abundance ratios of h@xmath58 to @xmath0co and @xmath0co to @xmath2co are assumed to be 10@xmath59 and 60 , respectively , which are the typical values in the interstellar medium @xcite .
the disk is assumed to be heated only by a stellar radiation and accretion heating is neglected .
this is a reasonable approximation for a disk with modest accretion rate such as hd163296 case ( 7.6 @xmath11 10@xmath60 @xmath6 yr@xmath18 ; @xcite ) .
all parameters regarding the object , listed in table [ tb : parameters ] , should be taken into account in the model calculation for the best accuracy , but some of them rarely affect the disk property .
emission profiles are mostly characterized by three parameters , @xmath36 , @xmath42 , and the outer radius of the disk ( @xmath61 ) , and these are regarded as variables .
the other parameters are fixed to be constants as listed in table [ tb : parameters ] . in actual calculations ,
@xmath42 was fixed in each run , and the possible fitting range of @xmath36 was searched .
@xmath42 was set to be between 0.003 and 3 g @xmath62 .
the search for @xmath36 in upper range of @xmath42 is stopped at 3 g @xmath62 because @xmath36 in each co line is well confined due to asymptotic behavior of fitting temperature in large @xmath42 region and the result remains almost the same .
intensity is affected by both @xmath36 and @xmath42 in the case of optically thin disk and by only @xmath36 in the case of optically thick disk . among three variables ,
@xmath61 can be guessed from the separation between two peaks in the profiles obtained from observation , since the keplerian velocity at @xmath61 ( @xmath63 ) roughly corresponds to half of the separation between the two peaks of the profile .
@xmath641000 au is adopted as an initial guess and then adjusted .
the best fit temperatures and outer radius at 0.1 , 0.3 , 1 , and 3 g @xmath62 in @xmath42 where fitting solution for all the co lines simultaneously exist are provided in table [ tb : best fit parameters ] , and the fitting results are shown in figure [ fig : figure6 ] .
note that the observed profiles shown in figure [ fig : figure6 ] were re - sampled with velocity resolution of 0.08 km s@xmath18 for @xmath0co ( @xmath1=10 ) line and 0.11 km s@xmath18 for the others to directly compare with the model calculations .
as shown in table [ tb : best fit parameters ] , the fitting temperature of @xmath0co ( @xmath1=32 ) shows the highest value among all the co lines .
it is essential to ascertain the goodness of the solution obtained in the model fitting .
figure [ fig : figure7 ] shows fitting temperature at several specific @xmath42 between 0.003 and 3 g @xmath62 .
we regarded the results as an acceptable fit when the following two conditions were met simultaneously : ( i ) to exclude the case of large deviation in amplitude of profiles , the residual in absolute value integrated over the range of @xmath65 2@xmath27 should be less than the uncertainty ( 1@xmath27 ) for the integrated intensity , ( ii ) to exclude the case of large deviation in line shape , more than 80% of the channels over @xmath66 should agree with the model calculations within 2@xmath27 .
the best fit @xmath67 for the @xmath0co ( @xmath1=32 ) line in table [ tb : best fit parameters ] are between 58.5 @xmath17 9.5 k , and this range is significantly higher than those estimated for the other co lines , 31 @xmath17 15 k when @xmath42 is greater than 0.1 g @xmath62 .
this difference in best fit @xmath67 implies that the disk is composed of at least two layers with distinct temperatures .
the uncertainty of fitting temperature is smaller in higher surface density or large optical depth in the disk . in lower surface density ,
the fitting temperature range becomes drastically larger or no solution exists ; @xmath2co ( @xmath1=10 ) emission line , for instance , dose not have any fitting temperature at @xmath42 less than 0.1 g @xmath62 .
the fitting errors in temperature at 0.1 g @xmath62 in @xmath42 are more than 30 % for @xmath2co ( @xmath1=10 ) line and more than 20 % for @xmath2co ( @xmath1=32 ) line , respectively .
the ones at greater than 0.3 g @xmath62 in @xmath42 , on the other hand , are less than 20 % for all the co lines and the fitting temperatures are gradually confined .
the @xmath42 that is greater than 0.1 g @xmath62 obtained by model fitting corresponds to @xmath68 10 ( @xmath22 5 even at @xmath61 = 550 au ) in optical depth , and therefore the disk is surely optically thick for all the co lines .
ccc & variables + @xmath69 & 2.3 @xmath70 & @xmath36 + @xmath71 & 122 pc&@xmath42 + @xmath72 & 45@xmath73&@xmath61 + @xmath74 & 0.1au & + @xmath75 & 1.0 & + @xmath76 & 0.5 & + @xmath77 & 1/10000 & + @xmath78 & 60 & + [ tb : parameters ] ( 180mm,80mm)figure6.eps [ cols="^,^,^,^,^ " , ] [ tb : best fit parameters ] ( 170mm,150mm)figure7.eps
high - resolution images of hd 163296 were taken with interferometers at multi - wavelengths ( the iram / pbi , sma , and vla ) both in continuum and in the emission lines of @xmath0co , @xmath2co , and c@xmath8o @xcite .
they compared the observational results of the co emissions with those predicted by self - consistent disk model to estimate physical disk properties .
although the assumptions in model calculations were slightly different from each other , the temperature and minimum surface density of the disk derived in section [ sec : fitting results ] showed a reasonable agreement with the values obtained by @xcite ; the temperature at 100 au and the minimum surface density at 10 au from the central star were estimated to be 26@xmath79 k and 1 g @xmath62 from our @xmath2co ( @xmath1=10 ) observations and they were within the possible range of the fitting values obtained by @xcite : 30@xmath80 k and 4@xmath81 g @xmath62 , respectively .
the radius derived from single dish observations , which generally have higher sensitivity to weaker emission , was 550@xmath82 au and was consistent with the one derived by interferometric observations .
as described in section [ sec : fitting results ] , the disk is arguably optically thick in co emissions . taking advantage of different mass opacity of isotopologues whose
emission lines reflect the temperature at each photosphere having different depth , the vertical thermal structure of the disk can be analyzed .
as shown in figure [ fig : figure7 ] , model fittings in the case of @xmath0co ( @xmath1=10 ) , @xmath2co ( @xmath1=10 ) , and @xmath2co ( @xmath1=32 ) emission lines were succeeded in almost the same temperature when @xmath83 0.1 g @xmath62 , but the @xmath0co ( @xmath1=32 ) results require almost double of the fitting temperature of the other three lines .
these different temperatures can be explained by temperature distribution along the vertical direction , as predicted by theoretical studies ( e.g. , @xcite ) . to examine the vertical temperature distribution in detail , we derived the column densities of h@xmath58 that satisfy optical depth for each co emission line to be unity ( @xmath84 ) from @xmath85\nonumber\\ \times \left[1-{\rm exp}\left(\frac{h{\nu_0}}{{k_b}t_{\rm ex}}\right)\right]^{-1 } , \label{eq : n}\end{aligned}\ ] ] where @xmath86 is the rotational constant of a molecule
, @xmath87 is the permanent dipole moment , @xmath88 is the excitation temperature , @xmath89 is the velocity width of a gas molecule , @xmath72 is the inclination angle , @xmath90 is planck constant , @xmath1 is a rotational quantum number , and @xmath91 is the frequency of transition @xcite . note that @xmath84 is the value obtained by integrating the number density from the disk surface to the disk interior along a line of sight . in this paper
, we adopted @xmath92 for simplicity , where @xmath93 is the thermal velocity width of gas molecules expressed by @xmath94 turbulent velocity ( @xmath95 ) can also contribute to @xmath89 .
@xcite derived a turbulent line width of @xmath22 0.3 km s@xmath18 for the disk around hd163296 by fitting profile with high spectral resolution .
previous observations also revealed that @xmath95 in disks around tts or hae star is comparable to @xmath93 : 0.07 - 0.15 km s@xmath18 in dm tau @xcite and 0.38 km s@xmath18 in ab aur @xcite .
in such cases , however , the line profile calculated by the model is not altered significantly by @xmath96 because the line shape is mainly determined by keplerian rotation velocity .
this assumption probably produces a factor of uncertainty for the estimation of @xmath84 but it does not significantly affect the following discussion .
@xmath84 for each co emission line provides the information about the vertical location of the photosphere of each co emission line in the disk .
figure [ fig : figure8 ] shows @xmath84 and vertical locations of photospheres of each co emission line .
the results are provided in the case of @xmath97 k and 60 k , which are the representative values of the low temperature interior and the high temperature surface , respectively , as described in section [ sec : fitting results ] . to derive the location of each photosphere ,
on the other hand , we select @xmath98 0.1 g @xmath62 as the representative case since this value is consistent with the results by @xcite . in both cases shown by figure [ fig : figure8 ] , @xmath84 results in the descending order of @xmath0co ( @xmath1=32 ) , @xmath0co ( @xmath1=10 ) , @xmath2co ( @xmath1=32 ) and @xmath2co ( @xmath1=10 ) , implying that the photosphere of @xmath0co ( @xmath1=32 ) emission is located uppermost among the four co lines .
the total geometric cross section of grains per hydrogen atom ( @xmath99 ) in interstellar medium is @xmath2210@xmath100 @xmath101 at visible wavelength @xcite , hence the column density that makes the optical depth for the stellar radiation unity is 10@xmath102 @xmath62 .
the grazing angle is assumed to be 0.05 in @xcite calculation , hence the column density of the upper layer should be about 5 @xmath11 10@xmath103 @xmath62 .
this column density is consistent with the one above the photosphere of @xmath0co ( @xmath1=32 ) at @xmath104 au , as shown in figure [ fig : figure8 ] .
@xmath84 of @xmath0co ( @xmath1=10 ) decreases faster than that of @xmath0co ( @xmath1=32 ) at @xmath105 au in the case of @xmath106 = 30 k because co molecules is not sufficiently excited to the @xmath1=3 level due to lower temperatures .
the same phenomenon is seen in @xmath2co as well . from the comparisons between the fitting temperature and the vertical locations of the photosphere of each co line , the general trend is that the farther away the layer is from the mid - plane of the disk , the warmer it reaches .
this is quite reasonable because the gas around the disk surface should be heated more by the radiation from the central star .
when we derive @xmath84 in figure [ fig : figure8 ] , we assumed that the disk temperature is uniform in vertical direction .
these considerations , however , might be invalid as a consequence of vertical temperature distribution ; the @xmath0co ( @xmath1=32 ) emission line originates from an upper layer of the disk , but this could also contribute significantly to the other co lines .
we therefore evaluate how strongly such an upper layer affects the other co emissions .
figure [ fig : figure9 ] shows how strongly the @xmath0co ( @xmath1=10 ) is emitted from above the photosphere of @xmath0co ( @xmath1=32 ) . in this figure
, we first calculate the optical depth of @xmath0co ( @xmath1=10 ) emission line in the region above the photosphere of @xmath0co ( @xmath1=32 ) , and then the intensity of @xmath0co ( @xmath1=10 ) at the fitting temperature was compared with that originated from the region . the contribution of high temperature layer is less than 10@xmath107 of the @xmath0co ( @xmath1=10 ) intensity and can be safely neglected .
therefore , the estimated vertical location of each co emission layer in figure [ fig : figure8 ] is validated .
( 170mm,150mm)figure8.eps ( 90mm,100mm)figure9.eps as shown in figure [ fig : figure7 ] , the temperature of @xmath0co ( @xmath1=10 ) and @xmath2co ( @xmath1=32 ) layers are almost the same when @xmath83 0.1 g @xmath62 .
the temperature of @xmath2co ( @xmath1=10 ) layer seems to be even lower . such vertical thermal structure of irradiated accretion disks around tts has been theoretically predicted .
@xcite constructed the detailed vertical structure models of irradiated accretion disks and extensively explored the dependence of the structure and emission properties on mass accretion rate , viscosity parameter , and disk radius .
@xcite also propose a 3-layer disk model .
the middle layer can form where the optical depth for the radiation reprocessed by the upper layer is less than unity .
our results suggest the existence of such a three - layer structure .
this paper presents observational results of the disk around hd163296 in @xmath0co ( @xmath1=10 ) , @xmath0co ( @xmath1=32 ) , @xmath2co ( @xmath1=10 ) , and @xmath2co ( @xmath1=32 ) emission lines .
double - peaked profiles originated from a rotating circumstellar disk were successfully detected in all the co lines .
physical parameters of the disk such as temperature distribution , surface density distribution and outer radius are estimated by model fitting utilizing a disk model .
these physical values obtained by single dish observation were confirmed to be consistent with the results of interferometric observations @xcite .
the disk must be optically thick for all the co lines .
taking advantage of difference in position of the photosphere among the co lines , we revealed temperature distribution in vertical direction .
it is proven that there are at least two distinct temperature layers , possibly three layers , from the model fitting .
the temperature in the uppermost @xmath0co ( @xmath1=32 ) emitting layer is estimated to be @xmath3 k at 100 au from the central star , and it is about twice higher than that in the inner regions emitting the other co lines , @xmath4 k at 100 au . since the vertical temperature distribution inside of a disk
is commonly suggested for both tts and hae stars ( dm tau , ab aur , and hd31648 ) , such a temperature distribution may be ubiquitous in protoplanetary disks @xcite .
the temperature and density structure of protoplanetary disks are still open questions .
better knowledge such as the radial and vertical density distribution will be provided by high - resolution observations in upcoming alma era .
acke , b. & van den ancker , m. e. , 2004 , , 426 , 151 beckwith , s. v. w. , sargent , a. i. , chini , r. s. , & guesten , r. 1990 , , 99 , 924 chiang , e. i. & goldreich , p. 1997
, , 490 , 368 dalessio , p. , calvet , n. , hartmann , l. , lizano , s. , & cant@xmath108 , j. 1999 , , 527 , 893 dartois , e. , dutrey , a. , & guilloteau , s. , 2003 , , 399 , 773 dent , w. r. f , , greaves , j. s. & coulson , i. m. 2005 , , 359 , 663 dullemond , c. p. 2002 , , 395 , 835 dullemond , c. p. , & dominik , c. , 2004 , , 417 , 159 dullemond , c. p. , dominik , c. , & natta , a. , 2001 , , 560 , 957 ezawa , h. , kawabe , r. , kohno , k. , & yamamoto , s. 2004 , proc .
spie , 5489 , 763 frerking , m. a. , langer , w. d. , & wilson , r. w. 1982 , , 262 , 590 fukagawa , m. et al .
2004 , , 606 , 53 garcia lopez , r. , natta , a. , testi , l. , & habart , e. , 2006 , , 458 , 837 grady , c.a .
et al . 2000 , , 544 , 895 hayashi , c. , nakazawa , k. , & nakagawa , y. , 1985 , protostars & planets ii ( the university of arizona press ) , 1100 hughes , a. m. , wilner , d. j. , qi , c. , & hogerheijde , m. r. 2008 , , 678 , 1119 hughes , a. m. , wilner , d. j. , andrews , s. m. , qi , c. , & hogerheijde , m. r. 2011 , , 727 , 85 inoue , a. , oka , a. , & nakamoto , t. , 2009 , , 393 , 1377 isella , a. , testi , l. , natta , a. , neri , r. , wilner , d. , & qi , c. 2007 , , 469 , 213 kitamura , y. , omodaka , t. , kawabe , r. , yamashita , t. , & handa , t. 1993 , , 45 , 27 kutner , m. l. , & ulich , b. l. 1981 , , 250 , 341 mannings , v. & sargent , a. i. 1997 , , 490 , 792 mannings , v. , koerner , d. w. & sargent , a. i. 1997 , , 388 , 555 meeus , g. , waters , l.b.f.m . , bouwman , j. , van den ancker , m. e. , waelkens , c. , & malfait , k. , 2001 , , 365 , 476 natta , a. , testi , l. , neri , r. , sheperd , d. s. , & wilner , d. j. 2004 , , 416 , 179 pavlyuchenkov , ya . ,
semenov , d , henning , th . , guilloteau , st . , pi@xmath109tu , v. , launhardt , r. , dutrey , a. 2007 , , 669 , 1262 pani@xmath110 , o. & hogerheijde , m. r. , wilner , d. , & qi , c. 2008 , , 491 , 219 pi@xmath109tu , v. , dutrey , a. , & guilloteau , s. , 2007 , , 467 , 163 pi@xmath109tu , v. , guilloteau , s. , & dutrey , a. , 2005 , , 443 , 945 pi@xmath109tu , v. , dutrey , a. , & kahane , c. , 2003 , , 398 , 565 scoville , n. z. , sargent , a. i. , sanders , d. b. , claussen , m. j. , masson , c. r. , lo , k. y. , & phillips , t. g. 1986 , , 303 , 416 simon , m. , dutrey , a. , & guilloteau , s. , 2001 , , 545 , 1034 stahler , s. w. & palla , f. 2004 , the formation of stars ( wiley - vch verlag gmbh & co.kgaa ) sunada , k. , yamaguchi , c. , kuno , n. , okumura , s. , nakai , n. , & ukita , n. 2000 , aspc , 217 , 19 tanaka , h. , himeno , y. , & ida , s. 2005 , , 625 , 414 testi , l. , natta , a. , shepherd , d.s .
, & wilner , d. j. , 2003 , , 403 , 323 van den ancker , m. e. , de winter , d. , & tjin a djie , h.r.e . 1998 , , 330 , 145 wannier , p. g. , penzias , a. a. , & jenkins , e. b. 1982 , , 254 , 100 waters , l. b. f. m. , & waelkens , c. , 1998 , ara&a , 36 , 233 wisniewski , m. , clampin , m. , grady , c. , ardila , david . , ford , h. , golimowski , d. , illingworth , g. , krist , j. & 2008 , , 682 , 548 yamaguchi , c. , sunada , k. , iizuka , y. , iwashita , h. , & noguchi , t. 2000 , spie , 4016 , 614 | this paper presents observations of a protoplanetary disk around herbig ae star hd163296 in @xmath0co ( @xmath1=10 ) , @xmath0co ( @xmath1=32 ) , @xmath2co ( @xmath1=10 ) , and @xmath2co ( @xmath1=32 ) emission lines .
double - peaked emission profiles originating from the rotating circumstellar disk were detected in all the lines .
the disk parameters were estimated from model calculation in which the radial distribution of temperature or surface density inside the disk has a power - law form
. the surface density should be sufficiently high so that the disk is optically thick for all the co lines , as discussed in previous studies based on interferometric observations .
the temperature and outer radius of the disk were also confirmed to be consistent with the previous results .
taking advantage of difference in position of the photosphere among the co lines , we revealed temperature distribution in vertical direction .
the temperature of @xmath0co ( @xmath1=32 ) emitting region is about twice higher than that of any other co emitting region ; the former is @xmath3 k while the latter is @xmath4 k at 100 au from the central star , suggesting that there are at least two distinct temperature regions .
the best fit temperature for @xmath2co ( @xmath1=10 ) that should trace the deepest region of the disk is even lower , implying that there is also a different temperature region in deep inside of the disk .
such vertical temperature distribution in a disk was identified both in t tauri and herbig ae stars ( e.g. , dm tau , ab aur , and hd31648 ) , and this should be a common feature in protoplanetary disks . |
has been a large body of work involving fiber bragg grating sensors over the past two decades @xcite .
early demonstrations were based on changes in the gross bragg wavelength as the gratings were perturbed due to strain and temperature . as interrogation techniques became more sophisticated , various signal processing and active fringe side locking schemes were employed , which dramatically improved their resolution @xcite .
this was further enhanced by refinement of grating design , enabling fabry - perot resonators to be fabricated @xcite , which effectively multiply the phase change due to fiber optical path displacements . with careful control of the grating writing process and appropriate choice of glass material , a bragg grating fiber fabry - perot ( ffp )
can now have a finesse of well over 1000 and a linewidth of a few mhz @xcite .
a fiber distributed feedback ( dfb ) laser can be fabricated when the ffp is written in a fiber amplifier and pumped optically @xcite .
these lasers have attracted significant interest for use in various schemes @xcite as active sensing elements , where changes in lasing wavelength due to environmental perturbations are used as the sensor signal . the past decade has also seen intense international effort in attaining direct gravitational wave detection , which demands unprecedented interferometric sensitivity to measure the strain of space - time . towards achieving this ultra resolution ,
the pound - drever - hall ( pdh ) laser frequency locking scheme @xcite is widely used .
it is adopted for laser frequency stabilization @xcite , interferometer longitudinal control , as well as gravitational wave signal extraction @xcite .
.,width=288 ] while the pdh frequency locking technique is well - established with free - space bulk - optical resonators and solid - state lasers within the gravitational wave community , it can readily be extended to diode laser stabilization @xcite , and guided - wave optics .
it has previously been utilized in a fiber laser stabilization scheme , where an erbium doped fiber laser was referenced to a coated micro resonator @xcite .
the pdh locking scheme can be adapted for both low frequency ( @xmath010hz ) quasi - static strain sensing , and dynamic measurements at higher frequencies .
in this paper we will discuss this technique in some detail , and demonstrate pdh locking for signal extraction in a fiber sensor , with a bragg grating ffp as the sensing resonator .
remote interrogation of passive resonators has significant advantages over active dfb lasers as sensing devices .
the problems relating to dopant clustering and relaxation oscillations in erbium doped dfb lasers @xcite are completely avoided , and the undesirable effects of optical pump noise is eliminated @xcite .
in addition , the interrogating laser can be housed in a controlled environment , and any residual introduction of phase noise due to laser intensity fluctuations are secondary effects . when the chosen interrogating wavelength is 1550 nm , telecoms grade smf-28 fiber can be used for both laser delivery as well as grating sensor fabrication , with the dual benefit of low cost and low loss .
this removes the need for more exotic fibers which requires cutoff below the pump wavelength for single - mode pump delivery . at 0.1 - 1mw output power ,
erbium doped fiber dfb lasers are inherently inefficient compared with commercially available extra - cavity diode lasers , used in this work , with @xmath110mw of output power .
this higher laser power improves the signal to shot noise ratio by up to an order of magnitude . while this output power can potentially be matched by er / yb codoped dfb lasers , they require complex fiber geometry to achieve sufficient photosensitivity @xcite .
in addition , frequency instability due to thermal pump absorption continues to limit their sensing performance @xcite .
remote interrogation , therefore , presents itself as an elegant and superior sensing solution .
for the purpose of this discussion , we will simplify our treatment of the ffp as similar to that of a free space resonant cavity , ie , within the bandwidth of concern , the bragg reflectors are broadband , and both the reflectors and resonator refractive index are non - dispersive . at the optical carrier frequency @xmath2 , the complex reflection response of a lossless ffp formed by two matched reflectors separated by distance @xmath3 ,
both with amplitude reflection coefficient @xmath4 , can be expressed as @xmath5 , \end{aligned}\ ] ] where @xmath6 and @xmath7 are the reflected and incident electric fields ; @xmath8 is the round - trip phase in a material of refractive index n ; @xmath9 and @xmath10 are , respectively , the amplitude and phase response .
the ffp has a full - width half - maximum ( fwhm ) bandwidth of @xmath11 .
the pdh locking scheme involves interrogating the ffp with the laser carrier phase modulated at @xmath12 , while measuring the reflected power with a photodetector , as illustrated in figure [ schematic ] .
after electronic demodulation and low - pass filtering , this signal can be reduced to @xcite @xmath13\cos(\psi ) \nonumber \\ & & + \im[\tilde{f}(\nu)\tilde{f}^{*}(\nu_{+ } ) -\tilde{f}^{*}(\nu)\tilde{f}(\nu_{-})]\sin(\psi)\ } , \label{err_v}\end{aligned}\ ] ] where the cross term @xmath14\ } \nonumber \\ & & -a(\nu)a(\nu_{-})\exp\{i[\phi(\nu_{-})-\phi(\nu)]\ } ; \label{cross_term } \end{aligned}\ ] ] @xmath15 and @xmath16 ; @xmath17 is the power in the carrier while @xmath18 is the power in each sideband .
the phase shift @xmath19 is set to optimize the demodulated error signal . in general
this is achieved when @xmath20\big]/d\nu } { d\big[\re[\tilde{c}(\nu_{\pm})]\big]/d\nu}\bigg\}_{\theta(\nu)=m2\pi},\end{aligned}\ ] ] where m is an integer .
the round - trip phase @xmath21 when the carrier is resonant with the ffp . from equation [ cross_term ]
, we can deduce that in the case of @xmath22 , @xmath23 and @xmath24 are both very small , and so the expression is dominated by its real part .
conversely , when @xmath25 , the sidebands are well outside of the ffp linewidth when the carrier is near resonance . in this case
these phase difference terms approach @xmath26 and the expression is dominated by its imaginary part
. if the ffp lineshape is symmetric and the carrier is at resonance , @xmath27 and @xmath28 for both cases , implying that equation [ cross_term ] , and hence equation [ err_v ] , become zero .
this is the usual lock point of the frequency servo . from equation [ err_v ] , it is clear that when the cross term equals 0 ( locked to resonance ) , the output @xmath29 is equal to zero and independent of @xmath17 and @xmath18 .
hence , when locked , the pdh system is immune to variations in laser intensity noise to the first order . in comparison , a fringe - side locking technique shows no implicit immunity to intensity noise , and requires an additional intensity monitor and subtraction electronics @xcite .
figure [ theo_errsig]a illustrates the theoretical error signal for the case of @xmath30 , while figure [ theo_errsig]b is for the case of @xmath31 , when @xmath2 is scanned across the resonance of a ffp . figure [ theo_errsig]c shows the intermediate case where @xmath32 .
the two satellite error signals in figure [ theo_errsig]b are due to the sidebands undergoing the ffp resonance , whereas in figure [ theo_errsig]c , the error signals due to the carrier and sidebands merge to form a single and almost square error signal .
the plots assume a resonance linewidth of 150mhz , and it is interrogated by phase modulation frequencies 15mhz , 1500mhz and 300mhz respectively .
.,width=288 ] the case where @xmath25 describes the classic pdh locking regime , involving high finesse fabry - perot cavities .
the principle of operation behind both extremes are similar and , for the sake of brevity , we shall refer to both as pdh locking in this treatment .
subsequent experimental results for our ffp to be presented in this paper will show that we were operating nearer the @xmath22 regime . .
inset : normalized experimental operating regimes for two resonances , overlaid with expanded theoretical plot.,width=288 ] for a given resonance fwhm , @xmath11 , the frequency separation between the turning points of a pdh error signal is dependent on @xmath12 . it approaches asymptotic values for both cases of @xmath22 and @xmath25 , as illustrated by the theoretical plot in figure [ theo_errsig_bw ] .
the plot is calculated with @xmath19 optimized for each @xmath12 .
on the other hand , for a given modulation frequency , the size and , therefore , slope of the error signal is dependent on the fwhm bandwidth @xmath11 .
figure [ theo_errsig_size ] shows the theoretical plot of peak - to - peak normalized error signal size vs normalized fwhm bandwidth .
the error signal size approaches zero when @xmath33 , but reaches an asymptotic value when @xmath34 .
the topology of our pdh interrogation of an ffp is shown in figure [ schematic ] .
the laser carrier is provided by a new focus vortex 6029 , which was an extra - cavity diode laser with a factory - estimated linewidth of 1mhz , and an intrinsic linewidth of @xmath35300khz .
its optical wavelength was centered around 1550.15 nm , with about 0.40 nm tuning range , which corresponds to a frequency range of @xmath3550ghz .
the frequency tuning of the laser was actuated by applying a voltage to the piezo - electric transducer ( pzt ) , which changed the laser cavity length .
the factory calibration specified that its laser pzt actuator had a gain of 12.5ghz / v . after passing through the optical isolator
, the laser polarization was adjusted to vertical by a half - wave plate before being modulated at 15mhz by the resonant phase modulator ( new focus 4003 ) .
the phase modulator was driven by a radio - frequency ( rf ) signal generator , which also provided the local oscillator for the demodulation electronics .
the modulated laser beam was coupled with an aspheric lens into a fiber - pigtailed polarization - independent optical circulator , which was spliced to the ffp .
the ffp was held between a pair of magnetic clamps , with one of the clamps in turn mounted on a translation stage , so that the bragg wavelength could be stretch - tuned to within the laser frequency range .
our ffp consisted of a pair of nominally matched 13.5db bragg gratings ( r@xmath36 ) each 15 mm long , spaced 10 mm apart , fabricated in a single phase - coherent writing process .
the schematic for the uv exposure along the length of the fiber is illustrated in figure [ ffp_schematic ] .
they were written in hydrogenated smf-28 fiber with no apodization .
both the transmitted and reflected light were collimated back into free space with ashperic lenses and then focussed onto photodetectors tx and rx , respectively , each with electronic bandwidth of @xmath3520mhz .
the optical isolator in the transmitted port eliminated any parasitic etalon effects due to residual back reflections from the collimating asphere .
the rf local oscillator was phase shifted before being used to mix down the electronic signal from the reflected port .
the mixed signal was low - pass filtered to provide the pdh error signal .
the local oscillator phase shift @xmath19 was optimized experimentally by maximizing the error signal .
a 95hz voltage ramp of 2vp - p and 50:50 symmetry was applied to the laser pzt input to sweep the laser carrier frequency , which equates to a slope of 380v / s .
the intensities transmitted and reflected by the ffp , as measured by the photodetectors , and the corresponding mixed down experimental error signal were recorded using a digital oscilloscope while the laser frequency was scanned .
they are displayed in figures [ exp_scan_wide]a , [ exp_scan_wide]b and [ exp_scan_wide]c , respectively .
there were two ffp resonances within the bragg grating bandwidth with differing peak heights and @xmath11 s .
these differences were mainly due to the frequency dependent reflectivity of the bragg grating pair , thus resulting in differing finesse at the two resonances .
since the gratings were not apodized during the fabrication process , we can expect higher reflectivity near the center of their bandwidth , which is confirmed by the higher finesse and thus narrower fwhm of the first resonator mode .
further , by comparing the heights of the two peaks in figure [ exp_scan_wide]a , we can see that the lower finesse resonance is closer to impedance matching . at this mode ,
nearly all of the laser light was transmitted and the reflection approached zero .
this difference in transmitted intensity , compared with the under - coupled high finesse mode , can be explained by uv induced loss in the resonator , particularly in the 10 mm spacing between the grating pair .
the higher finesse resonance transmitted less intensity due to its greater resonator round - trip number , or total storage time , which resulted in greater total loss while circulating within the resonator . to reduce this loss ,
the uv laser can be easily controlled to avoid fiber exposure between the grating pair during the resonator fabrication process .
the transmission scan and the reflected error signal for the narrower resonance is enlarged in figure [ exp_scan_zoom]a and [ exp_scan_zoom]b .
the fwhm time for the pzt scan in figure [ exp_scan_zoom]a was @xmath37s , which corresponds to 11.4mv on the pzt . recalling that the factory calibration specified that its laser pzt input provided 12.5ghz / v of tuning , the fwhm bandwidth of this mode can be determined to be 143mhz . for comparison ,
the broader resonance had a fwhm time of 66@xmath38s , which implies a @xmath11 bandwidth of 314mhz .
the separation between the two peaks can be seen to be @xmath39ms in figure [ exp_scan_wide ] , which infers a free spectral range of 9ghz .
hence , the narrower mode had a finesse of 63 , while the broader resonance had a finesse of 29 .
the @xmath40 ratio for the higher finesse mode was @xmath41 .
the corresponding peak - to - peak time for its error signal in figure [ exp_scan_zoom]b was @xmath42s , which yields an error signal turning point frequency separation to @xmath11 ratio of @xmath43 .
on the other hand , the lower finesse resonance had an error signal peak - to - peak time of 38@xmath38s , which corresponds to @xmath40 of @xmath44 , and an error signal turning point separation to @xmath11 ratio of @xmath45 .
the error signal turning point separation to @xmath11 ratios for the two modes are close to each other , and agree with the values as predicted in figure [ theo_errsig_bw ] . at these linewidths
, @xmath12 is small enough relative to @xmath11 to approach the asymptotic value of the lower limit .
the peak - to - peak error signal size for the higher finesse mode was larger than that of the lower one , as seen in figure [ exp_scan_wide]c , since the @xmath40 for the higher finesse mode was twice that of the lower finesse mode .
this was predicted by the theoretical plot in figure [ theo_errsig_size ] .
the error signal peak - to - peak voltage for the high finesse mode was measured to be 1.4v , while that for the lower finesse resonance was 0.63v .
these two points , for @xmath40 of 0.1 and 0.05 , are normalized and overlaid with the theoretical plot in the inset of figure [ theo_errsig_size ] , to illustrate the region where these two modes were operated . assuming an effective refractive index of 1.45 , a free spectral range of 9ghz yields a resonator length of 11.5 mm , implying that the effective reflection point of the gratings was @xmath46 mm inside each grating .
we tested for polarization dependence of the ffp response with a second half - wave plate before the laser was coupled into the fiber . no visible shift in resonance frequencies
were observed as the waveplate was rotated .
this implies that for the intent and purpose of this application , the uv illumination of the grating core during the fabrication process can be regarded as isotropic .
any degeneracy due to parasitic birefringence was beyond the linewidth resolution of the ffp resonance , as the two modes provided well behaved error signals free from input polarization wander effects .
it is evident from figures [ exp_scan_wide ] and [ exp_scan_zoom ] that pzt scanning to sweep the frequency of the laser , as demonstrated in this experiment , is a simple alternative to the single - sideband modulation technique for high resolution spectral characterization of fiber gratings @xcite .
the slope of the error signal through resonance was @xmath47nv / hz for the higher finesse mode , and @xmath48nv / hz for the lower finesse mode .
hence the higher finesse resonance was our preferred mode for pdh locking , as it provided a larger signal as a frequency and displacement discriminator in sensing applications .
one should note , however , that while higher ffp finesse is preferred for superior sensitivity , the free running laser frequency noise sets a limit to interferometer sensitivity . to lock the laser , the voltage ramp from the signal generator
was turned off , and the pzt dc offset voltage tuned slowly while the transmitted and reflected laser intensities were monitored with an oscilloscope . when the laser is nearly resonant with the chosen ffp peak , the transmitted intensity approaches its maximum , and the feedback loop was then engaged to acquire lock .
this process was recorded by the digital oscilloscope traces shown in figure [ exp_lock_acq ] .
the servo amplifier used in this experiment had a single real pole response with a corner frequency of 0.03hz .
the total feedback loop had a dc gain of @xmath351000 and a unity gain bandwidth of around 40hz .
lock acquisition was straight forward and once it was acquired , the system stayed locked for several hours even when subjected to large environmental noise events .
lock termination occured when the grating drifted outside the laser tuning range , and this typically happened after over 3 hours of locked operation .
5.5sec.,width=288 ] in a pdh locking scheme , the sensor signal is extracted by either monitoring the pzt feedback voltage required to maintain lock within the servo unity gain bandwidth , or by monitoring the mixer output at frequencies above the unity gain bandwidth .
environmental stimulations , such as temperature drift as well as stress and strain due to mechanical or acoustic perturbation , change the resonance condition of the ffp .
this results in both dc and ac voltage change in the mixer output and the pzt feedback voltage . when the mixer output is read with a dynamic signal analyzer , information about these perturbations can be extracted .
the signal analyzer performs a fast fourier transform of the mixer output voltage , and provides a trace with units in volts@xmath49 . the quotient of this trace by the slope of the error signal ( 19nv / hz ) yields the callibrated measurement in hz@xmath49 .
the low frequency measurement of this mixer output is shown in figure [ exp_sig_analyser_lowfreq ] .
there was a large component of ambient noise at low frequencies as the ffp was not isolated from laboratory acoustic and thermal noise .
we were able to identify fiber violin modes , broadband acoustic noise , and pzt resonances in this frequency regime .
for example , the large feature at @xmath50 khz seen in figure [ exp_sig_analyser_lowfreq ] is due to closed loop excitation of a laser pzt mode .
figure [ exp_sig_analyser ] shows a wider frequency scan of the ambient frequency noise .
it is overlaid with the calculated shot noise and measured electronic noise . at frequencies above ambient excitation ,
the free running frequency noise of the laser limits this measurement to @xmath51300hz@xmath49 .
assuming the laser has a lorentzian lineshape with white spectral density of frequency noise @xmath52 , the 3db linewidth of the laser @xmath53 can be estimated by @xcite @xmath54 where @xmath52 has units of hz@xmath49 .
thus , the broadband frequency noise of @xmath51300hz@xmath49 corresponds to an intrinsic laser linewidth of @xmath55khz , which is consistent with the manufacturer s estimate of 300khz .
according to the empirical model determined by kersey et al .
@xcite , bragg grating responsivity @xmath56 where @xmath57 is the strain perturbation , and @xmath58 is the bragg wavelength , 1 pm of induced grating wavelength shift corresponds to a strain of @xmath59 . at @xmath60 nm , equation ( [ kerseymodel ] )
can be rearranged to arrive at the conversion factor @xmath61 where @xmath62 is the equivalent induced grating frequency shift . since 1 pm is equivalent to 125mhz at 1550 nm , we can infer from the high frequency noise floor that the ffp sensor has a broadband strain sensitivity of @xmath63 .
the shot noise in figure [ exp_sig_analyser ] was calculate as follows @xcite : @xmath64 where @xmath65 is the equivalent shot noise voltage ; @xmath66 c is the electronic charge ; @xmath67 is the @xmath68 output voltage of the photodetector when the system is locked ; g is the transimpedance gain of the photodetector : and @xmath69 is the mixer conversion gain .
the quotient of @xmath65 by the error signal slope then gives the shot noise in units of hz@xmath49 .
this was calculated to be 16 hz@xmath49 , which corresponds to a limiting shot - noise sensitivity of ( 16 hz@xmath70 ) @xmath51 100 f@xmath71 .
the electronic noise is the dark noise measured at the mixer output . within the unity
gain bandwidth of the feedback system , the sensor dynamic range depends on the laser optical frequency tuning range .
since our laser had a pzt tuning range of 50 ghz , the low frequency dynamic range of this system is limited to ( @xmath72 ) 330 @xmath73 . assuming a breaking stress of @xmath1 100 kpsi @xcite , and a young s modulus of @xmath74 kpsi @xcite for fused silica , the breaking strain is @xmath1 9800 @xmath73 .
this means that typically , the breaking strain is well beyond the limited tuning range of the laser used in this experiment . above the unity gain bandwidth ,
the sensor dynamic range is limited by the fwhm bandwidth of the resonator to ( @xmath75 ) 0.9 @xmath73 .
hence , for large dynamic range applications , the preferred operating approach is to expand the unity gain bandwidth out to a maximum , and perform in - loop measurements at the laser pzt actuator input .
we have presented a passive fiber sensor interrogation technique which was adapted from pdh locking , used in gravitational wave detection .
we demonstrated the robust and stable operation of the simple locking system in fiber . in many applications
, we believe this passive technique is superior to active methods using fiber lasers , due to its better efficiency , improved signal to shot - noise ratio , lower cost , and its suitability for remote sensing .
it has an implied broadband strain sensitivity limit of 2 p@xmath57 , which is due to the free - running frequency noise of our laser . with appropriate laser stabilization prior to ffp interrogation , however ,
it has the potential to surpass the pico - strain regime and approach the fundamental shot noise limit .
june 29 , 2004
the authors would like to thank adrian l. g. carter , of nufern , for useful discussions .
alan d. kersey , michael a. davis , heather j. patrick , michel leblanc , k. p. koo , c. g. askins , m. a. putnam , and e. joseph friebele , `` fiber grating sensors , '' _ j. lightwave technol .
_ , vol . 15 , pp . 1442 - 1463 , 1997 .
anthony dandridge , alan b. tveten , and thomas g. giallorenzi , `` homodyne demodulation scheme for fiber optic sensors using phase generated carrier , '' _ ieee j. quantum electron .
_ , qe-18 , pp . 1647 - 1653 , 1982 .
n. e. fisher , d. j. webb , c. n. pannell , d. a. jackson , l. r. gavrilov , j. w. hand , l. zhang , and i. bennion , `` ultrasonic hydrophone based on short in - fiber bragg gratings , '' _ appl . opt .
8120 - 8128 , 1998 .
michel leblanc , alan d. kersey , and tsung - ein tsai , `` sub - nanostrain strain measurements using a pi - phase shifted grating , '' in _ proc .
fiber sensors ofs 97 _ , williamsburg va , 1967 pp .
28 - 30 , 1997 .
y. j. rao , m. r. cooper , d. a. jackson , c. n. pannell , and l. reekie , `` absolute strain measurement using an in - fibre - bragg - grating - based fabry - perot sensor , '' _ electron .
708 - 709 , 2000 .
sigurd weidemann lvseth , jon thomas kringlebotn , erlend rnnekleiv , and kjell bltekjr , `` fiber distributed - feedback lasers used as acoustic sensors in air , '' _ appl .
4821 - 4830 , 1999 .
a. frank , k. bohnert , k. haroud , h. brndle , c. v. poulsen , j. e. pedersen , and j. patscheider , `` distributed feedback fiber laser sensor for hydrostatic pressure , '' _ ieee photon .
15 , pp . 1758 - 1760 , 2003 .
r. w. p. drever , j. l. hall , f. v. kowalski , j. hough , g. m. ford , a. j. munley , and h. ward , `` laser phase and frequency stabilization using an optical resonator , '' _ appl .
b _ , vol . 31 , pp .
97 - 105 , 1983 .
bram j. j. slagmolen , malcolm b. gray , karl g. baigent , and david e. mcclelland , `` phase - sensitive reflection technique for characterization of a fabry - perot interferometer , '' _ appl .
3638 - 3643 , 2000 .
timothy day , eric k. gustafson , and robert l. byer , `` sub - hertz relative frequency stabilization of two - diode laser - pumped nd : yag lasers locked to a fabry - perot interferometer , '' _ ieee j. quantum electron .
1106 - 1117 , 1992 .
kenneth a. strain , guido mller , tom delker , david h. reitze , david b. tanner , james e. mason , phil a. willems , daniel a. shaddock , malcolm b. gray , conor mow - lowry , and david e. mcclelland , `` sensing and control in dual - recycling laser interferometer gravitational - wave detectors , '' _ appl .
1244 - 1256 , 2003 .
daniel a. shaddock , malcolm b. gray , conor mow - lowry , and david e. mcclelland , `` power - recycled michelson interferometer with resonant sideband extraction , '' _ appl . opt .
42 , pp . 1283 - 1295 , 2003 .
a. schoof , j. grnert , s. ritter , and a. hemmerich , `` reducing the linewidth of a diode laser below 30 hz by stabilization to a reference cavity with a finesse above @xmath78 , '' _ opt .
1562 - 1564 , 2001 .
t - c zhang , j - ph poizat , p. grelu , j - f roch , p. grangier , f. marin , a. bramati , v. jost , m. d. levenson , and e. giacobino , `` quantum noise of free - running and externally - stabilized laser diodes , '' _ quantum semiclass .
_ , vol . 7 , pp . 601 - 613 , 1995 | we discuss a phase - sensitive technique for remote interrogation of passive bragg grating fabry - perot resonators .
it is based on pound - drever - hall laser frequency locking , using radio - frequency phase modulation sidebands to derive an error signal from the complex optical response , near resonance , of a fabry - perot interferometer .
we examine how modulation frequency and resonance bandwidth affect this error signal .
experimental results are presented that demonstrate when the laser is locked , this method detects differential phase shifts in the optical carrier relative to its sidebands , due to minute fiber optical path displacements .
fiber fabry - perot , fiber resonator , bragg grating resonator , fiber sensor , strain sensor , nanostrain , picostrain , frequency locking , bragg grating interrogation . |
maximal entangled states are the basic resource for the encoding and communication on quantum information @xcite . in open systems entanglement
degrades and it is necessary to design strategies that preserve or partially recover the entanglement .
error correction codes@xcite , filtering and distillation of entanglement @xcite and the use of decoherence free sub - spaces @xcite , are some of such strategies . in a typical protocol for distillation of entanglement
there are two observers , alice and bob and two o more copies of bipartite entangled system .
each observer has access only to one part of the system on which he or she may applicate local operations .
these operations can be independent ( filtering ) or collectives ( proper distillation ) . in a proper distillation protocol at the end of the process , after the collective operations have been completed , the observers execute a set of local measurements on some copies ( the ancillas ) leaving one copy ( the source ) untouched . by the effect of measurements the ancillas collapse to separable states , transferring its entanglement to the undisturbed copy .
the most used protocols , i.e. bennett s protocol @xcite and deutsch s protocol @xcite , use bell diagonal states as the resource from which to extract entanglement .
in fact , bennett s protocol starts with a werner state obtained by applying a random bilateral rotation to an arbitrary state .
maximal entangled states are obtained only after an infinite distillation steps . in @xcite ,
verstraete _ et .
_ show that all the mixed states of a 2-qubits system are classified in two equivalence classes . the first class is formed by the states which are equivalent to a bell - diagonal state under local @xmath1 transformations .
the second class is formed by states which are represented by non - diagonal matrices ( in the bell basis ) parametrized by four real parameters .
it comprises four sub - sets depending on the values of the parameters . in @xcite ,
the authors advocate the use of non quasi - separable states ( see section 2 ) to distill states with maximal entanglement in a finite number of distillation steps . in this work
we first review the structure of quasiseparable states for @xmath0 systems and show that all non - quasiseparable states are non diagonal and belong to one of the four sub - sets mentioned before .
this study allows to define a finite round protocol for distilling , at any time , a maximal entangled state in a two qubits system at zero temperature .
starting with singlet , it is shown that the system evolves inside the set of non - quasiseparable states , allowing the application of the protocol at arbitrary times . for thermal baths at non zero temperature
the states of the systems are always quasi - separable , so it is not possible to distill a bell state . nevertheless under certain conditions it is possible to increase the entanglement applying the same protocol used for vacuum .
consider a mixed state @xmath2 , with non vanishing probabilities @xmath3 and the transformations that change the probabilities without adding or suppressing any pure state in the mixture : @xmath4 the new probabilities @xmath5 should be also non vanishing , and we will refer to @xmath6 as a new state of @xmath7 . by definition a state @xmath8
is called quasi - separable if at least one of its new states is separable .
if a given final state is obtained through a distillation process using @xmath9 copies of the original state , then the same final state may be obtained with different probability from any of the new states of @xmath8 .
this implies in particular that is not possible to distill a maximal entangled state using a finite number of copies if the source or the ancilla are in quasiseparable states . in order to distill bell states ,
is useful to determine which states of the @xmath0 system are not quasiseparable . on the other hand ,
if one can distill a bell state from a given state obviously one can distill the same bell state from any other state obtained by local operations and classical communication .
now we consider the states of the system inside the equivalent classes defined by verstraete _
et al _ in @xcite .
the first class is composed of bell - diagonal states , they can be written as : @xmath10 with @xmath11 the four bell states .
for any mixed bell diagonal state one can observe that it is possible to find a new state for which all the probabilities are the same , this new state is separable without mattering the rank of the original matrix .
so , excluding bell states , all bell diagonal states are quasi - separable and are not useful in order to distill a bell state . following the classification , the second class is formed by non diagonalizable states represented by matrices of the form : @xmath12 here the entries @xmath13
satisfy one and only one of the following four conditions : 1 .
@xmath14 2 .
@xmath15 and @xmath16 3 .
@xmath17 and @xmath18 4 .
@xmath19 and @xmath20 the cases 2,3 and 4 correspond to separable states from which it is not possible to distill any entanglement . in the first case one
is compelled to consider two kind of states . for a non conditioned @xmath21
the rank of the matrix is 3 and the state takes the form : @xmath22 requiring normalization one puts @xmath23 and @xmath8 can be decomposed as follows , @xmath24 with @xmath25 the state @xmath6 defined by @xmath26 is a new separable state of @xmath8 , and so all the states of form ( [ uto1 ] ) with unconditioned @xmath21 are quasiseparable .
consider now the states with matrices of rank 2 defined by setting the coefficients @xmath27 .
the density matrix in this case is : @xmath28 any new state of this @xmath8 can be written as @xmath29 the concurrence of this state is @xmath30 .
since by definition @xmath31 , it is clear that the concurrence is different from zero for any choice of @xmath32 .
the main concussion is that these states are not quasiseparable .
in this section we review briefly the distillation protocol introduced by cheng _ et .
al . _ in ref .
suppose that one has a set of @xmath0 systems which are in the state ( [ ito2 ] ) .
use one half of the systems as a source and the other half as the ancilla .
alice execute a unilateral not operation on the source particles , then state of the source becomes : @xmath33 the ancilla remains in the original state .
now alice and bob execute unilateral c - not operation on their respective particles regarding the source as control and the ancilla as target .
finally alice and bob make a measurement in the computational basis of the ancilla . if they get the @xmath34 result the ancilla collapses to @xmath35 and the source to @xmath36 .
the success probability in this case is @xmath37 .
one can increase the success probability allowing the use of the @xmath38 as a valid result . in that case
the final reduced density matrix of the source is not more the bell state but an state with same structure than the original one : @xmath39 the concurrence of this state is @xmath40 and the probability of success is @xmath41 . in figure
( 1 ) we plot the concurrence of the distilled state and the concurrence of the non - distilled state . as one can see the protocol that include both results increases the entanglement only when @xmath42 .
[ fig11 ] and @xmath38 result ] in figure ( 2 ) we compare the probability when one takes only the result @xmath43 with that when one takes the result @xmath43 and @xmath38 [ fig2 ] and when one takes @xmath43 and @xmath38 results ]
the master equation for a system composed of two two - level particles in vacuum is @xmath44 where @xmath45 and @xmath46 .
we assume that the baths associated with each particle are independent .
if the initial state is the singlet the solution is : @xmath47 in this case the protocol studied in the previous section allows one to distill the singlet @xmath48 at an arbitrary time with probability @xmath49 .
the protocol run as follow : a ) divide the particle into two similar sets , one as source and the other as ancilla .
b ) alice applies a not operation on the ancilla c ) both observers apply a c - not operation regarding the source as control and the ancilla as target .
b ) each observer measures the ancilla waiting for the result @xmath43 .
c ) if the result is successful the source is left in the state @xmath50 d ) alice applies a @xmath51 pauli rotation in order to obtain the singlet @xmath48 as final state .
the master equation for a system composed of two two - level particles in the presence of two independent thermal baths is @xmath52 where we are assuming that both baths have the same temperature so they have the same average number @xmath9 of thermal photons .
the solution of this equation with a singlet as initial condition is @xmath53 with @xmath54 @xmath55 @xmath56 the concurrence of this state is given by : @xmath57 with @xmath58 if one now applies exactly the same protocol described in the previous section the source is left in the following distilled state @xmath59 where @xmath60 is the success probability and @xmath61 @xmath62 @xmath63@xmath64 are the eigenvalues of @xmath8 . the concurrence of the distilled state is given by @xmath65 with @xmath66 in the figure ( 3 ) we plot both the original and the distilled concurrence .
as one can see the gain in entanglement is appreciable .
we have presented a classification of quasi - separable states in @xmath67 systems that identifies the set of non quasi - separable states with part of the non diagonal mixed states class . more precisely : a non quasiseparable state is given by setting the four parameters of the non diagonal state equal to @xmath68 .
having characterized non quasiseparable states , we presented a distillation protocol which allows to increase the entanglement of the original state using a finite number of copies . for systems in vacuum
it is possible to recover a maximally entangled state with probability greather than zero but smaller than @xmath69 . by changing the accepted results in the measure of the ancilla , it is possible to obtain a final state with increased entanglement ( but not maximally entangled ) and probability greater than @xmath69 .
finally , this protocol can also be applied in the case of non zero temperature .
in this situation the system evolves with a comparatively sustained entanglement before the entanglement sudden dead time , although the probability rapidly falls bellow one half .
we wish to thank j. stephany for his constructive remarks and precious help .
this work was supported by did - usb grant gid-30 and by fonacit grant no g-2001000712 .
a.einstein , b.podolsky y n.rosen , _ phys .
_ , * 47 * , 777 , ( 1935 ) .
r.horodecki , p.horodecki , m.horodecki and k.horodecki _ rev .
_ , * 81 * , no . 2 , 865 ( 2009 )
c.h.bennett , g.brassard , c.crpeau , r.jozsa , a.peres and w.k.wootters , _ phys rev lett .
_ , * 70 * , 1895 , ( 1993 ) . c. h. bennett , d. p. divincenzo , j. a. smolin , and w. k. wootters , _ phys
rev . _ a * 54 * , 3824 ( 1996 ) d. a. lidar , i. l. chuang and k. b. whaley , _ phys .
_ , * 81 * , 2594 , ( 1998 ) .
f.verstraete , j.dehaene and b.demoor,_phys .
_ , * a64 * , 010101(r ) ( 2001 ) .
bennett , g .
brassard , s. popescu , j b. schumacher , .
a. smolin and w. k. wootters,_phys . rev . lett_*76 * , 722 , ( 1996 ) .
d. deutsch , a. ekert , r. jozsa , c. macchiavello , s. popescu and a. sanpera,_phys .
rev . lett_*77 * , 2818 , ( 1996 ) .
p. chen , c. li and m. huang , _ phys .
_ , * a65*,012317 ( 2001 ) .
f.verstraete , j.dehaene and b.de moor,_phys .
_ , * a65 * , 032308(2002 ) . | in this work we review the entire classification of @xmath0 distillable states for protocols with a finite numbers of copies .
we show a distillation protocol that allows to distill bell states with non zero probability at any time for an initial singlet in vacuum .
it is shown that the same protocol used in non zero thermal baths yields a considerable recovering of entanglement . |
the full power of the renormalization group ( rg ) concept is perhaps most clearly seen in its application to quantum impurity systems.@xcite the classic paradigm is the kondo model,@xcite being the simplest to capture the fundamental physics associated with all quantum impurity models : universal rg flow from an unstable fixed point ( fp ) to a stable one on successive reduction of the temperature or energy scale .
the kondo model describes a single local spin-@xmath0 ` impurity ' , coupled by antiferromagnetic exchange to a single channel of noninteracting conduction electrons . here
, perturbative scaling arguments@xcite indicate an rg flow from a high - energy unstable ` free fermion ' fp ( describing a free impurity decoupled from a free conduction band ) , to a low - energy stable ` strong coupling ' fp ( where the impurity is screened by conduction electrons via formation of a ` kondo singlet ' ) .
this rg flow is characterized by a scaling invariant the kondo temperature @xmath1 which sets the crossover energy scale .
but analysis of the crossover itself goes beyond simple scaling ideas and the conventional rg picture .
wilson s numerical renormalization group@xcite ( nrg ) allows an exact nonperturbative calculation of certain thermodynamic and dynamical quantities which show the crossover ( for a review , see ref . ) .
universal scaling of all physical quantities in terms of the crossover scale @xmath1 confirms the basic rg structure of the problem .
however , a different rg flow occurs when the impurity is coupled to two or more independent conduction channels.@xcite in this multichannel kondo model , the frustration inherent when several channels compete to screen the impurity spin renders the strong coupling fp unstable . a third fp at _ intermediate _
coupling@xcite then dictates the low - energy physics .
this fp exhibits non - fermi liquid ( nfl ) behavior , including notably a residual entropy@xcite of @xmath2 in the two - channel kondo ( 2ck ) model .
the crossover from the free fermion fp to the 2ck fp has been the focus of much theoretical attention . in particular , solution of the model using the bethe ansatz
yields the exact crossover behavior of thermodynamic quantities;@xcite while nrg has been used to calculate thermodynamics@xcite and dynamics@xcite numerically .
it was also shown recently that this 2ck physics can arise in odd - membered quantum dot rings@xcite and chains,@xcite and in quantum box systems.@xcite indeed , the same type of nfl behavior@xcite arises in the two - impurity kondo ( 2ik ) model.@xcite the tendency to form a trivial local singlet state is favored by an exchange coupling acting directly between the impurities ; while the coupling of each impurity to its own conduction channel favors separate single - channel kondo screening .
the resulting competition gives rise to a critical point@xcite that is identical to that of the 2ck model with additional potential scattering.@xcite 2ik physics is also expected to appear in certain double quantum dot systems,@xcite and other even - impurity chains.@xcite a description of the nfl fps of such two - channel models in terms of an effective boundary conformal field theory ( cft ) shows that the operator controlling the fp has an anomalous scaling dimension.@xcite this implies unconventional energy / temperature dependences of physical quantities such as conductance @xmath3 , measured as a function of bias voltage @xmath4 and temperature @xmath5 . in the 2ck device of ref .
, @xmath6 and @xmath7 corrections to the nfl fp conductance predicted by cft were directly observed in experiment .
similar signatures are expected@xcite in the channel - asymmetric 2ik model ; although leading linear behavior emerges in the symmetric 2ik.@xcite this behavior is of course in marked contrast to @xmath8 and @xmath9 fermi liquid ( fl ) behavior obtained ubiquitously in the single - channel case.@xcite schematic phase diagram for the 2ck and 2ik models , as a function of temperature @xmath5 , external energy scale @xmath10 , and symmetry - breaking perturbation strength @xmath11 .
the three fps of each model give rise to three distinct regimes : free fermion , nfl and fl .
we considered the nfl to fl crossover at @xmath12 in ref .
, indicated by arrow ( a ) .
here we generalize the results to finite temperature , arrow ( b ) .
, width=328 ] the nfl fp itself ( and the crossover to it ) has now been rather well studied . however , nfl physics is extremely delicate : various symmetry - breaking perturbations destabilize the nfl fp@xcite and generate a new crossover scale @xmath13 . at @xmath12 , the impurities are thus completely screened and all residual entropy is quenched .
indeed , regular fl behavior,@xcite including the standard @xmath14 and @xmath15 corrections to conductance , must appear at low temperatures @xmath16 and energies @xmath17 .
therefore , no evidence of nascent nfl physics can manifest in the immediate vicinity of the fl fp . only on fine - tuning the perturbation strength @xmath18 to the critical point so that @xmath19 does one obtain nfl physics on the lowest energy scales .
but rg analysis in the vicinity of the free fermion , nfl and fl fps implies two successive crossovers , with @xmath1 setting the energy scale for flow to the nfl fp , and @xmath13 characterizing flow away from it . even in the fl phase away from the critical point ( which is the generic case relevant to experiment ) , nfl behavior
can be observed at _ higher _ temperatures and energies , provided there is good scale separation @xmath20 ( see fig .
[ fig : pd ] for a schematic phase diagram ) . in this case ,
conductance @xmath3 through the 2ck quantum dot device of ref . , or
proposed 2ik devices,@xcite should exhibit a clean nfl to fl crossover . in ref .
, we considered this conductance crossover at @xmath12 as a function of bias @xmath4 , corresponding to the crossover labelled by arrow ( a ) in fig .
[ fig : pd ] ( and @xmath4 here playing the role of the external energy scale @xmath10 ) .
it was shown@xcite that the full crossover is wholly characteristic of the high - energy nfl state .
the @xmath12 crossover is expected to describe the behavior at very low temperatures . from a scaling perspective
, rg flow is cut off on the energy scale @xmath21 , so @xmath22 for @xmath16 since there are no further crossovers below @xmath23 . by contrast , at higher temperatures @xmath24 ( @xmath25 ) , no evidence of the nfl to fl crossover will be observed ( see fig .
[ fig : pd ] ) , and only the nfl fp is probed .
indeed , this is the likely scenario in the experiment of ref . : rather than tuning to the critical point @xmath26 , signatures of the true fl ground state are simply washed out by temperature .
but the behavior as a full function of temperature @xmath5 and energy scale @xmath10 for finite perturbation strength @xmath11 is much more subtle , and naturally strengthens connection to experiment .
exploring the third temperature axis in fig .
[ fig : pd ] , and considering the resultant nfl to fl crossover [ eg . arrow ( b ) ] , is thus the focus of the present work . in this paper
we combine abelian bosonization methods@xcite with the powerful machinery of cft@xcite to obtain an exact description of the nfl to fl crossover in two - channel kondo models .
in particular , we calculate the full electron green function at finite temperature , from which conductance follows.@xcite the field theoretic description links the 2ik model with a classical ising model on a semi - infinite plane.@xcite application of a boundary magnetic field @xmath27 in this boundary ising model ( bim ) results in rg flow from an unstable fp with free boundary condition @xmath28 to a stable fp with fixed boundary condition @xmath29.@xcite this rg flow due to @xmath27 is identical to that occurring between nfl and fl fps in the 2ik model due to a small perturbation @xmath11.@xcite indeed , an emergent symmetry of the nfl fp in the 2ik model,@xcite together with the common cft description of 2ik and 2ck models,@xcite implies the existence of a single universal nfl to fl crossover function for both models , resulting from any combination of relevant perturbations.@xcite exact results@xcite for the bim are the source of our solution , which becomes exact when there is good scale separation @xmath30 , as sought experimentally .
the exact crossover green function at @xmath12 was calculated in ref . by exploiting the above connection .
however , ambiguities appear at finite temperature which prevent straightforward generalization of those results .
thus we take a different route here : the bim solution is used to calculate real - space friedel oscillations around the impurities at finite temperature , which are themselves related by integral transformation@xcite to the green function .
the problematic analytic continuation is avoided in this way .
the paper is organized as follows . in sec .
[ model ] we introduce the 2ck and 2ik models , together with representative symmetry - breaking perturbations that generate the nfl to fl crossover .
we then present and discuss our main results for the exact finite - temperature green function along the crossover .
the corresponding conductance crossover for quantum dot systems that might realize 2ck or 2ik physics is then calculated . in secs .
[ 2ikgf][gengf ] we derive the analytic results .
first we consider the 2ik model at @xmath12 with a single detuning perturbation . in sec .
[ 2ikgf ] we calculate the resulting crossover green function , exploiting the analogy to the bim . in sec .
[ finitetderiv ] we extend the calculation to finite temperatures , extracting the desired t matrix from friedel oscillations .
the results are generalized to the 2ck model in sec .
[ 2ckgen ] and to an arbitrary combination of perturbations in sec .
[ gengf ] .
exact results are compared with finite - temperature nrg calculations in sec .
[ nrgcomp ] .
other quantities showing the crossover such as entropy and nonequilibrium transport are then briefly considered in sec . [ other ] .
the paper concludes with a general discussion in sec .
[ concs ] , and details of certain calculations can be found in the appendices .
we consider the standard 2ck and 2ik models , @xmath31 where @xmath32 describes two free conduction electron channels @xmath33 , with spin density @xmath34 ( and @xmath35 ) coupled to one spin-@xmath0 impurity @xmath36 ( 2ck ) or two impurity spins @xmath37 ( 2ik ) . for @xmath38 , the nfl ground state of @xmath39
is described by the 2ck fp .
likewise , a critical inter - impurity coupling @xmath40 can be found such that the ground state of @xmath41 is again a nfl,@xcite and is similarly described by the 2ck fp for @xmath42.@xcite relevant perturbations to the above models ( embodied by @xmath43 and @xmath44 ) are those that destabilize the nfl fp , and result in a fl ground state .
a new scale @xmath23 is thus generated , characterizing rg flow from nfl to fl fps .
the relevance of such perturbations can be traced to the breaking of certain symmetries,@xcite such as parity or particle - hole symmetries .
in fact , there are many possible perturbations to the 2ck and 2ik models ; but two perturbations may be considered ` equivalent ' if they break the same underlying symmetry and their effect on the low - energy physics will be identical.@xcite for concreteness , we consider now the simplest perturbations which exemplify such symmetry - breaking , and which in combination generate all possible nfl to fl crossovers at low energies / temperatures .
specifically , @xmath45 describes @xmath46 channel asymmetry in the 2ck model for @xmath47 , while charge transfer between the leads is embodied in the @xmath48 and @xmath49 components of the first term [ here @xmath50 are the pauli matrices in the channel ( spin ) sector ] .
the second term describes a magnetic field acting on the impurity . for the 2ik model , the critical point
is destabilized by finite @xmath51 , and also through @xmath52 where @xmath53 describes electron tunneling between the leads and @xmath54 the application of a staggered magnetic field .
channel anisotropy could also be included in the 2ik model , but the critical point can always be recovered@xcite on retuning @xmath55 .
similarly , spin - assisted tunneling between channels @xmath56 ( as arises in a two - impurity anderson model@xcite ) is expected to have the same destabilizing effect as the @xmath53 term in eq .
[ delta2ik ] , since they both have the same symmetry at the nfl fp@xcite ( although the resulting crossover energy scales may themselves be different@xcite ) .
thus we do not consider such perturbations explicitly here .
signatures of the nfl to fl crossover on the scale of @xmath23 should appear in all physical quantities . in 2ck or 2ik quantum dot devices which could access this physics,@xcite
the quantity of interest is the @xmath57 conductance @xmath58 through channel @xmath59 or @xmath60 . here
, @xmath61 describes the relative strength of coupling to source and drain leads ( see fig .
[ fig : condpic ] for an illustration of the setup ) . at zero - bias @xmath62 ,
the conductance is given exactly by,@xcite @xmath63 where @xmath64^{-1}$ ] is the fermi function , and @xmath65 is the energy - resolved local density of states ( or ` spectrum ' ) , @xmath66 itself related to the t matrix@xcite @xmath67 , describing scattering of a @xmath68 or @xmath69 electron within channel @xmath59 or @xmath60 , with bare lead density of states per spin @xmath70 .
note that for equal hybridization to source and drain leads , @xmath71 , @xmath72 is maximal ; while in the asymmetric limit @xmath73 , @xmath74
. in the latter case , the weakly - coupled source lead probes the system perturbatively , so the system remains near equilibrium , even at finite bias @xmath75 .
the resulting conductance is then simply,@xcite @xmath76 where @xmath77 is the equilibrium ( zero - bias ) spectrum .
schematic illustration of possible 2ck ( a ) and 2ik ( b ) setups to measure conductance .
the left lead is ` split ' into source and drain , allowing the resulting conductance through the attached impurity to be measured.,width=283 ] the t matrix itself must show signatures of the nfl to fl crossover since scattering is purely inelastic at the nfl fp,@xcite but inelastic scattering must cease at energies @xmath78 , where the impurity degrees of freedom are fully quenched.@xcite thus the crossover also shows up in conductance .
our goal here is to calculate the full crossover t matrix , and hence conductance , at finite temperature for the 2ck and 2ik models in the presence of symmetry - breaking perturbations described by eqs .
[ 2ck][delta2ik ] . in the next sections we derive an exact expression for the desired t matrix , describing the universal crossover from nfl to fl behavior in the 2ck and the 2ik models at finite temperature and which as such generalize the results of our previous work in ref . .
here we pre - empt the full derivation , and present our key results .
the nfl to fl crossover is characterized by a low - energy scale @xmath23 arising due to the presence of symmetry - breaking perturbations to the 2ck and 2ik models .
it is given generically by@xcite @xmath79 where @xmath80 .
the eight contributions correspond to relevant perturbations which have distinct symmetry at the nfl fp .
two perturbations which have the same symmetry correspond to the same @xmath81 .
the perturbations given in eqs .
[ delta2ck ] and [ delta2ik ] are classified viz , .classification of perturbations [ cols="^,^,^ " , ] [ table : pert ] where @xmath82 .
the perturbations associated with coupling constants @xmath83 and
@xmath84 do not conserve total charge,@xcite and so are ignored here ( although we note that such operators can be of importance , for example , in the context of strongly correlated superconductors@xcite ) .
@xmath85 are fitting parameters@xcite which depend on the model and on @xmath86 , and @xmath87 is the kondo temperature , characterizing rg flow from the high - energy free fermion fp to the nfl fp.@xcite we do not discuss the high - energy crossover in the present work .
the various perturbations described by eqs .
[ delta2ck ] and [ delta2ik ] describe very different physical processes but the resulting crossover scale eq .
[ tstar ] , has a simple form due to an _ emergent _ @xmath88 symmetry of the effective nfl fp hamiltonians , as discussed in the following sections .
the main result of this paper is the nfl to fl crossover t matrix , given by , @xmath89 where @xmath90 is the scattering s matrix , which is an @xmath91 and @xmath12 quantity characterizing the fl fp . for the 2ck model it is given by , @xmath92/\lambda,\ ] ] with @xmath93 . for the 2ik model
it is , @xmath94/\lambda .
\end{split}\ ] ] the single function @xmath95 describes the crossover due to a generic combination of relevant perturbations in both 2ck and 2ik models .
it does not depend on details of the model or the particular perturbations present , except through the resulting crossover scale @xmath13 .
thus @xmath96 is a _
universal _ function of rescaled energy @xmath97 and temperature @xmath98 .
our exact result at finite temperature is , @xmath99 , \end{split}\ ] ] where @xmath100 is the gamma function , and @xmath101 is the gauss hypergeometric function.@xcite at @xmath12 , eq .
[ analyticformula ] reduces@xcite to the result of ref . , @xmath102,\ ] ] where @xmath103 $ ] is the complete elliptic integral of the first kind , yielding asymptotically @xmath104 for @xmath105 ; and @xmath106 ) \tilde{\omega}^{-1/2 } + \mathcal{o}(\tilde{\omega}^0)$ ] for @xmath107 .
below we consider the local density of states ( spectrum ) @xmath108 , from which conductance can be calculated ( see eqs .
[ zbcond ] and [ asymcond ] ) .
it is related to the t matrix via eq .
[ specdef ] , and is thus given exactly along the nfl to fl crossover by eqs .
[ exactresult][analyticformula ] : @xmath109,\ ] ] where the required diagonal elements of the full s matrix ( eqs .
[ s2ck ] and [ s2ik ] ) are more simply expressed as , @xmath110 with @xmath111 for spins @xmath112 and @xmath113 for channel @xmath46 ( and we use @xmath114 for simplicity ) . for @xmath115 and @xmath116 scattering preserves channel and spin , and
the fl phase shift @xmath117 then follows from @xmath118 $ ] .
these exact results for the crossover are compared with finite - temperature nrg calculations in sec .
[ nrgcomp ] , with excellent agreement .
we now examine the generic behavior of the spectral function at finite temperatures in the crossover regime .
although we consider explicitly @xmath119-channel spectra @xmath120 in the following , note from eq .
[ smatrix ] that @xmath121 upon exchanging @xmath122 in the 2ck model , or @xmath123 in the 2ik model .
also , @xmath124 on reversing the magnetic field , @xmath125 ( and in the zero - field case , @xmath68 and @xmath69 spectra are of course identical ) .
spectrum @xmath126 vs @xmath127 for @xmath128 , approaching @xmath129 from above ( @xmath130 ) or below ( @xmath131 ) .
circles show @xmath12 result of eq .
[ exactt=0 ] .
, width=302 ] in fig .
[ fig : k ] , we take the representative case of finite channel anisotropy @xmath132 in the 2ck model , or finite detuning @xmath133 in the 2ik model , and plot @xmath126 as a full function of @xmath127 for different temperatures @xmath134 .
since only @xmath135 acts in either case , @xmath136 is diagonal ( see eqs . [ s2ck ] , [ s2ik ] ) , meaning that an electron in channel @xmath137 scatters elastically at low energies , and stays in channel @xmath137 . by eq .
( [ specexact ] ) , the spectrum @xmath138 then probes the real part of the universal function @xmath139 because @xmath140 is real .
general scaling arguments suggest that rg flow stops on an energy scale given by the temperature . as seen from fig .
[ fig : k ] , this is indeed the case , with the spectrum @xmath141 essentially constant for @xmath142 .
mutatis mutandis , for @xmath143 one obtains @xmath144 , corresponding to the @xmath12 limit considered previously.@xcite at @xmath12 and @xmath91 , eq .
[ specexact ] yields @xmath145 , which is determined solely by the s matrix and hence the phase shift associated with the stable fl fp .
when only @xmath135 acts , the spectrum is thus @xmath146 or @xmath147 only ( with corresponding phase shifts @xmath148 or @xmath149 ) . in particular ,
the kondo phase is characterized by unitarity @xmath150 , obtained in the more strongly coupled channel for the 2ck model , and in both channels for @xmath151 in the 2ik model . in the opposite limit @xmath24 ( @xmath25 ) rg flow to the fl fp is completely cut off , and inelastic scattering@xcite at the nfl fp results in @xmath152 for all @xmath153 . the generic rg picture illustrated in fig .
[ fig : pd ] and supported by fig .
[ fig : k ] , suggests an approximate complementarity between @xmath10 and @xmath5 .
this is explored further in fig .
[ fig : tw ] , where we compare the zero - frequency value of the spectrum as a function of temperature , with the zero - temperature spectrum as a function of frequency .
as immediately seen , there is a striking similarity .
indeed , a classic signature of fl physics ( arising at low - energies / temperatures @xmath154 ) is the common quadratic dependence of the spectrum on both frequency and temperature , @xmath155 with @xmath156 and @xmath157 constants @xmath158 that depend on the particular model under consideration .
perturbation theory with respect to the fl fp in the spirit of nozires,@xcite yields @xmath159 for the 2ik model ( see appendix [ pertfl ] ) .
this same ratio also follows directly from the limiting behavior of the full crossover function , eq .
[ analyticformula ] , which yields @xmath160 and @xmath161 ; and as such provides a stringent check of our results .
@xmath162 and @xmath163 vs @xmath164 for @xmath130 .
common fl asymptote eq .
[ flspec ] shown as dot - dashed line ; nfl asymptotes eq .
[ nflspec ] shown as dashed and dotted lines . , width=302 ] but the exact symmetry between @xmath10 and @xmath5 in eq .
[ flspec ] is a special property of the fl fp itself , and does not in general apply at higher energies ; although as seen from fig .
[ fig : tw ] , the qualitative behavior over the full crossover is in fact rather similar . in the vicinity of the nfl fp ( arising for @xmath165 ) , eq .
[ analyticformula ] gives asymptotically , [ nflspec ] @xmath166 with @xmath167 for @xmath168 ; and @xmath169 ( where @xmath170 is euler s constant ) , @xmath171 , and @xmath172 .
terms of the form @xmath173 and @xmath174 in eq .
[ nflspec ] signal the scaling dimension @xmath175 of the relevant perturbation . whereas such powerlaws occur in both the frequency and temperature dependence , additional logarithmic corrections appear in the frequency - dependence only .
this difference can be understood by comparing the full @xmath12 result ( eq . [ exactt=0 ] ) with the high-@xmath5 behavior captured by perturbation theory around the nfl fp ( see ref . and appendix [ nflpert ] ) . the full dependence on @xmath10 and @xmath5 described by eq .
[ analyticformula ] naturally leads to more subtle behavior when @xmath176 and @xmath5 are of the same order , as shown in fig .
[ fig : k ] . when some degree of inter - channel charge transfer is also present , the nfl to fl crossover is generated by the _
combination _ of relevant perturbations @xmath135 and @xmath177 . in the 2ck model
, this corresponds to finite channel anisotropy @xmath132 and impurity - mediated tunneling @xmath48 ( see eq . [ delta2ck ] ) ; while for 2ik it corresponds to finite detuning @xmath133 and direct tunneling @xmath53 ( see eq .
[ delta2ik ] ) . the resulting behavior in the 2ck model can be simply understood because the perturbations @xmath132 and @xmath48 are related by a ` flavor ' rotation of the bare hamiltonian , as discussed further in sec .
[ flavorrot ] .
the 2ik model does not possess such a flavor symmetry , although an _ emergent _ symmetry@xcite of the nfl fp hamiltonian can be exploited when @xmath20 .
in fact , as shown in sec .
[ unitarytrans ] , this symmetry allows all the relevant perturbations in either 2ck or 2ik models to be simply related,@xcite implying the existence of a _ single _ crossover function .
the rotation @xmath178 can be used to relate systems where both @xmath135 and @xmath177 act , to those in which @xmath135 alone acts ( as in fig .
[ fig : k ] ) . but
@xmath65 probes the system in the original unrotated basis , and hence the spectra undergo a rescaling when finite @xmath177 is included : @xmath179 , with @xmath180 at both fl and nfl fps . at the fl fp , electrons thus scatter elastically between @xmath59 and @xmath60 channels ; and the corresponding s matrix @xmath181 is purely _
off_-diagonal when inter - channel charge transfer dominates ( see eqs .
[ s2ck ] , [ s2ik ] ) .
thus , no crossover shows up in the spectrum or conductance , although @xmath13 is of course finite ( see eq .
[ tstar ] ) ; and the crossover can still appear in other physical quantities.@xcite @xmath126 vs @xmath127 for @xmath128 , in the presence of finite uniform ( staggered ) magnetic field in the 2ck ( 2ik ) model , @xmath182 .
spectra approach @xmath129 from above ( @xmath68 ) or below ( @xmath183 ) .
circles show @xmath12 result .
, width=302 ] when a uniform ( 2ck ) or staggered ( 2ik ) magnetic field acts ( finite @xmath184 only ) , @xmath185 is pure imaginary ( with phase shifts @xmath186 ) , and again we obtain @xmath187 at the fl fp .
however , @xmath65 now probes the imaginary part of the universal function @xmath139 ( see eq . [ specexact ] ) ; and so the full spectrum along the nfl to fl crossover due to @xmath184 is simply the hilbert transform of the spectrum due to @xmath135 compare figs .
[ fig : k ] and [ fig : mag ] .
a spectral feature in consequence appears on the intermediate scale of @xmath13 for finite @xmath184 , even though @xmath188 at both nfl and fl fps ; as shown in fig .
[ fig : mag ] .
the existence of such a feature can be understood physically from the impurity magnetization @xmath189 arising for small applied field @xmath190 in the 2ck model ( or staggered magnetization @xmath191 due to a staggered field in the 2ik model ) . since the magnetization @xmath192\ne 0 $ ] is finite for finite applied field , @xmath193 .
a ` pocket ' thus opens between the ` up ' and ` down ' spin spectra at @xmath194 , whose area is proportional to the magnetization .
in particular , the temperature - dependence of the magnetization can be extracted from the universal function , eq .
[ analyticformula ] , viz @xmath195 valid for small perturbations @xmath184 , such that @xmath20 as usual ( the high - frequency cutoff @xmath196 then being justified since @xmath197 for @xmath198 , as confirmed directly from nrg ) .
magnetization @xmath199/\sqrt{t_kt^*}$ ] vs @xmath134 for @xmath182 and @xmath200 .
nfl asymptote eq .
[ magnfl ] shown as dashed line ; fl asymptote eq .
[ magfl ] shown as dotted line in the inset .
, width=264 ] using eq .
[ exactt=0 ] , the zero - temperature magnetization depends on @xmath1 via , @xmath201 and since @xmath202 ( see eq . [ tstar ] ) , it follows that @xmath203 .
the full temperature dependence @xmath204 vs @xmath5 is shown in fig .
[ fig : magn ] , demonstrating scaling collapse for different kondo scales @xmath1 .
the asymptotic behavior in the vicinity of the fl and nfl fps follows as , yielding in particular @xmath206 when @xmath207 .
this asymptotic behavior can again be understood from perturbation theory around the fl and nfl fps .
furthermore , since @xmath208 , when @xmath209 , one obtains @xmath210 for the uniform ( staggered ) magnetic susceptibility of the 2ck ( 2ik ) model .
this diverging susceptibility is a classic signature of the nfl fp , known for example from the bethe ansatz solution of the 2ck model,@xcite or from cft for the 2ik model.@xcite however , at lower temperatures @xmath211 , @xmath212 does not correspond to the magnetic susceptibility : here the nfl to fl crossover itself is being probed .
indeed , for @xmath213 , we obtain a quadratic @xmath15 temperature - dependence of magnetization , eq .
[ magfl ] , characteristic of the fl fp .
conductance @xmath214 vs bias @xmath4 and temperature @xmath5 for @xmath130 .
black lines connect regions of constant conductance .
light colors correspond to high conductance near the fl fp ; dark colors correspond to lower conductance near the nfl fp .
, width=283 ] conductance @xmath214 arising for finite @xmath135 .
_ upper panel : _ vs temperature @xmath134 for @xmath215 , approaching @xmath216 from above ( @xmath130 ) or below ( @xmath131 ) .
circles show the exact zero - bias result . _
lower panel : _ vs bias @xmath217 for @xmath128 .
diamonds show zero - temperature result.,width=283 ] finally , we turn to conductance @xmath218 , obtained from the spectrum @xmath65 by combining eqs .
[ asymcond ] , [ analyticformula ] and [ specexact ] .
it follows as @xmath219 \bigg \ } , \end{split}\ ] ] where the integral over @xmath10 can be evaluated using contour methods , @xmath220 , \end{split}\ ] ] with rescaled @xmath221 , @xmath222 , @xmath97 as before . in particular at zero - bias , @xmath223 a color plot of conductance @xmath218 along the nfl to fl crossover
is shown in fig .
[ fig : condcp ] , for the representative case of @xmath130 .
the black lines connect regions of equal conductance . in the fl regime @xmath224 , the quadratic form of eq .
[ flspec ] thus yields a simple ellipse ; while in the nfl regime there is a pronounced @xmath4-@xmath5 anisotropy .
the detailed behavior of eq .
[ condexact ] is seen on taking cuts through fig .
[ fig : condcp ] at constant @xmath4 and @xmath5 , as shown in fig .
[ fig : condcut ] .
returning to the 2ck quantum dot system of ref .
, we comment now on the possible strength of symmetry - breaking perturbations present in the experiment . due to the coulomb blockade physics of the quantum box ,
inter - channel charge - transfer was effectively suppressed . in the absence of a magnetic field ,
the dominant perturbation is thus _ channel anisotropy _ , @xmath132 ( see eq . [ delta2ck ] ) . the experiment showed@xcite
2ck scaling of conductance around @xmath1 , but no fl crossover at @xmath13 .
this implies @xmath207 see for example fig .
[ fig : k ] for @xmath225 , which shows little sign of the crossover . since @xmath226 in the experiment and taking @xmath227 , from eq .
[ tstar ] and table [ table : pert ] it follows that @xmath228 could be at most @xmath229 ( @xmath230 depends on details of the model / device setup ) .
the observed 2ck physics is thus an impressive testament to the tunability and control available in such quantum dot devices .
having presented our main results and discussed their physical implications , we turn in the following sections to the formal derivation .
our goal is an exact expression for the nfl to fl crossover t matrix , which is related to the electron green function . in ref . , we calculated the crossover at @xmath12 ; further details of that calculation are presented here , providing as they do the necessary foundations for our generalization of the results to finite temperature .
the structure of the nfl fixed point hamiltonian of the 2ik model allows for an elegant description of the nfl to fl crossover.@xcite before presenting that derivation , we discuss first some relevant preliminaries which will be of later use .
in particular , we consider now the representation of the fixed point hamiltonians within cft , and the structure of the corresponding green functions .
our starting point is a description of the free conduction electron hamiltonian @xmath231 in terms of chiral dirac fermions .
a 1d quadratic dispersion relation @xmath232 can be linearized near the fermi points @xmath233 , @xmath234 ( with @xmath235 the fermi velocity ) .
this is the standard case@xcite and applies to arbitrary dimension within the assumption that the bare density of states is flat at low energies.@xcite conduction electron operators can be fourier transformed and expanded near the fermi points , focusing on states within width @xmath236 around @xmath237 , @xmath238 we thus define left and right movers,@xcite @xmath239 defined for @xmath240 , with @xmath241 the distance from the impurities located at the ` boundary ' @xmath242 . with a boundary condition @xmath243 we introduce a single left - moving chiral dirac fermion for all @xmath241 , @xmath244 .
since the new operators satisfy the usual fermionic anticommutation relations @xmath245 , the free hamiltonian @xmath231 can be expressed as a chiral hamiltonian,@xcite @xmath246 hereafter , we set @xmath247 [ and the fermi level density of states is then @xmath248 .
the free electron green function then follows as,@xcite @xmath249 } \\
\equiv & \delta_{\sigma \sigma ' } \delta_{\alpha \alpha'}g^0(\tau , x_1-x_2 ) , \end{split}\ ] ] where @xmath250 is imaginary time .
the corresponding matsubara green function is then defined by the transformations @xmath251 with matsubara frequencies @xmath252 for integer @xmath253 , and where @xmath254 is inverse temperature .
direct evaluation of eq .
[ g0matsubara ] using eq .
[ freeg ] then yields simply , @xmath255 .
\end{split}\ ] ] of course , the interesting behavior arises when coupling to the impurities is switched on . as usual,@xcite the full green function is related to the t matrix via @xmath256 where @xmath257 , @xmath258 and @xmath259 are @xmath260 matrices with indices taking the values @xmath261 .
in particular , it should be noted that the t matrix is local .
[ g0 ] and [ tdefinition ] also imply that if @xmath262 and @xmath263 have equal sign then the full green function reduces to the free green function , reflecting the chiral nature of the dirac fermion , eq .
[ h0 ] . since all such fermions are now left - moving , information about scattering from the impurities located at the boundary @xmath242 and hence the t matrix
is obtained from the full green function with @xmath262 and @xmath263 located on _ opposite _ sides of the boundary .
the free green function eq .
[ freeg ] is naturally obtained at fps where the free boundary condition pertains .
but at any conformally invariant fp , the powerful machinery of boundary cft gives nonperturbative information about the green function .
specifically , when the boundary condition is obtained from fusing with some primary field @xmath156 , then correlation functions are given generically by@xcite @xmath264^{2 d}},\ ] ] where @xmath265 is the scaling dimension of the primary field @xmath266 , and @xmath267 are elements of the modular s matrix.@xcite using the conformal mapping from the plane to the cylinder with circumference @xmath268 , one obtains@xcite the generalization to finite temperature @xmath269 , @xmath270)^{2 d}}.\ ] ] in the present context , we are interested in the electron green function at the conformally invariant free fermion , nfl and fl fps of the 2ik model . with @xmath271 the @xmath272 fermion field , the full fp green functions takes the form,@xcite @xmath273},\ ] ] where @xmath274 can be understood as the one - particle to one - particle scattering matrix , and can be calculated from the modular s matrix in the case of boundary conditions obtained by fusion.@xcite thus , the effective fp theory is identical to that of the free theory discussed above , but with a modified boundary condition that determines the scattering matrix @xmath274 . choosing @xmath275 and @xmath276 in eq .
[ tdefinition ] yields @xmath277 , or equivalently @xmath278 , where the above convention @xmath279 was used . at a fl fp ,
the scattering matrix is unitary,@xcite @xmath280 , and as such describes purely elastic scattering .
by contrast , at the nfl fp of the 2ik model , it has been shown@xcite that @xmath281 , which implies fully inelastic scattering : a single electron sent in to scatter off the impurities decays completely into collective excitations , and no single - particle state emerges .
such behavior is manifest by a half - unitarity spectrum , @xmath282 .
however , along a crossover between fps , the green function does not in general take the form of eq .
[ cibc ] .
further insight into the fps of the 2ik model is provided by a representation in terms of majorana fermions ( mfs ) .
considering again the free theory described by @xmath231 , four nonlocal fermions can be defined by abelian bosonization and refermionization@xcite of the four original dirac fermions @xmath284 with spin @xmath285 and channel index @xmath286 .
8 mfs are then obtained by taking the real and imaginary part of each .
specifically , four bosonic fields @xmath287 are defined , viz @xmath288 where @xmath289 are klein factors.@xcite linear combinations of these bosonic fields are then used to construct new fields , @xmath290 where @xmath291 and @xmath292 .
refermionizing yields , @xmath293 where the klein factors @xmath294 are related to @xmath289 as described in appendix [ quadratic ] .
thus , four new fermionic species @xmath295 are defined , with @xmath296 corresponding to ` charge ' , ` spin ' , ` flavor ' and ` spin - flavor ' .
the real and imaginary parts of each , @xmath297 fulfill the majorana property @xmath298 and satisfy separately the fermionic anticommutation relation @xmath299 , and so are referred to as mfs . the free fermion fp hamiltonian ( corresponding to eqs .
[ 2ck ] and [ 2ik ] with @xmath300 and @xmath301 ) then follows as @xmath302 where @xmath303 , and the scattering states are defined by the trivial boundary condition @xmath304 .
the fixed point thus possesses a large @xmath88 symmetry in terms of these mfs .
likewise , the fl fp hamiltonian ( in which the impurity degrees of freedom are quenched ) is similarly described by eq .
[ mffreefp ] .
the corresponding boundary condition is encoded in the single - particle fl scattering s matrix @xmath305 .
although it depends on the specific perturbations generating the crossover , the boundary condition is thus trivial at the fl fp . in particular ,
finite detuning @xmath306 results in @xmath304 , as obtained at the free fermion fp .
the remarkable fact is that the nfl fp hamiltonian also takes the form of eq .
[ mffreefp ] ; and its nontrivial boundary condition@xcite is again simple in terms of the mfs . it can be accounted for by defining the scattering states @xmath307 , and @xmath308 for @xmath309 . thus , 7 of the 8 mfs are described by the free theory at the nfl fp .
the fp hamiltonian eq .
[ mffreefp ] describes a higher @xmath88 symmetry than is present in the original hamiltonian .
the explicit symmetries of the 2ik model also allow a separation of the theory into different symmetry sectors .
specifically , the @xmath310 symmetry sectors comprise a bose - ising representation,@xcite describing a coset construction of three wess - zumino - witten ( wzw ) nonlinear @xmath311 models , together with a @xmath312 ising model . the two @xmath313 theories with central charge @xmath314 correspond to conserved charge in the left and right channels .
the @xmath315 theory with @xmath316 corresponds to conserved total spin .
finally , the ising model @xmath312 is a @xmath317 theory corresponding to a single mf .
this non - abelian representation is connected with the 8 mfs , as discussed in ref . .
the symmetry ` currents ' of those sectors , such as the spin current @xmath318 , are represented _ quadratically _ in terms of mfs as described in appendix [ quadratic ] .
specifically the @xmath319 charge currents in left and right channels are represented in terms of 4 mfs @xmath320 and @xmath321 ; while the @xmath322 spin current @xmath323 is represented in terms of three mfs , @xmath324 .
the @xmath312 theory corresponds to the remaining mf , @xmath325 .
the important implication for our purposes is that the green function can be factorized into pieces coming from different sectors associated with the various mfs .
we now exploit the above bose - ising construction,@xcite in terms of which the fermion field can be expressed by the bosonization formula , @xmath326_1(x ) g_\sigma(x ) \sigma_l(x).\ ] ] here , the dimension @xmath272 fermion field has been decomposed into a dimension @xmath327 factor @xmath328_1 $ ] representing the @xmath329 spin-@xmath0 primary field of the @xmath313 charge theories , a dimension @xmath330 factor @xmath331 representing the spin-@xmath0 primary field of the @xmath315 spin theory , and the dimension @xmath332 factor @xmath333 originating from the ising sector .
the subscript @xmath119 emphasizes that @xmath333 is only the left - moving chiral component of the full spin operator @xmath311 of the ising sector , arising here because @xmath284 is the chiral left - moving fermion field .
since the nfl fp of the 2ik model is conformally invariant,@xcite we may use eqs .
[ bosonizationf ] and [ corrfunct ] to determine the contribution to the full green function coming from each of the sectors : @xmath334^{\tfrac{7}{8}}\\ & \times \langle \sigma_l(\tau , ix_1)\sigma_l(0,ix_2 ) \rangle . \end{split}\ ] ] the free boundary condition in the charge and spin sectors yields the first factor , corresponding to the free green function eq .
[ freeg ] but with power @xmath335 arising because 7 of the 8 mfs are associated with these sectors .
the nfl boundary condition is expressed in terms of fusing with the @xmath336 field @xmath333 from the ising sector in eq .
[ bosonizationf ] .
the second factor thus comes from the remaining ising sector . at the nfl
fp it follows from eq .
[ corrfunct ] that @xmath337^{\tfrac{1}{8}}$ ] .
since the modular s matrix for fusion with the @xmath338 ising operator has a vanishing element @xmath339 , the entire green function thus vanishes at the nfl fp;@xcite consistent with eq .
[ cibc ] with @xmath340 . in summary , the nontrivial boundary condition at the nfl fp affects only the _ ising _ sector of the 2ik model .
the function @xmath341 in eq .
[ factorg ] is a quantity pertaining to the @xmath312 ising model and contains the nontrivial physics ; while the spin and charge sectors are simply spectators . in the next section
we exploit eq .
[ factorg ] and a connection between the 2ik model and a classical ising model@xcite to determine the green function along the _ crossover _ from the nfl fp to the fl fp .
as highlighted above , the nfl fp of the 2ik model is described by the free theory in all sectors except the ising sector , which takes a modified boundary condition . when the small perturbation @xmath135 is included ( corresponding to detuning @xmath342 ) ,
the nfl fp is destabilized .
the effective hamiltonian in the vicinity of the nfl fp is then @xmath343+\delta h_{qcp}$ ] , with @xmath344 the fp hamiltonian itself , given in eq .
[ mffreefp ] and parametrized in terms of the scattering states @xmath345 which encode the boundary condition .
the correction is given by , @xmath346 where @xmath156 is a local mf involving impurity spin operators@xcite which satisfies @xmath347 and anticommutes with all other mfs , @xmath348 .
the perturbation @xmath135 thus acts only in the ising sector .
indeed , the difference in boundary conditions between the nfl and fl fps is also confined to the ising sector ( all other sectors have free boundary conditions at both fps ) .
the _ entire crossover _ from nfl to fl fp thus occurs completely within the ising sector because the @xmath135 perturbation does not spoil the decoupling of the sectors , which becomes exact@xcite in the limit @xmath349 .
the other sectors then act as spectators along the crossover .
this implies a generalization of eq .
[ factorg ] to the full crossover .
one can still interpret the first factor in eq .
[ factorg ] with power @xmath350 as the product of autocorrelation functions of 7 spin fields which undergo no change in the boundary condition .
but the @xmath311 field originating from the @xmath312 sector in the @xmath351 construction has _ flowing _ boundary condition ( which is not conformally invariant ) .
we consider first the equal - time green function with @xmath352 positioned symmetrically on either side of the boundary : @xmath353^{\tfrac{7}{8 } } \times \langle \sigma_l(\tau , ix)\sigma_l(\tau ,- ix ) \rangle , \end{split}\ ] ] where the factor @xmath354 now describes the _ crossover _ in the ising sector in terms of the chiral ( left - moving ) ising magnetization operator @xmath355 .
we now use standard boundary cft methods@xcite to relate the two - point chiral function @xmath356 living in the full plane [ see fig . 9 ( b ) ] to the product of chiral holomorphic @xmath357 and antiholomorphic @xmath358 operators living in the halfplane with a boundary . but bulk operators in cft may be expressed as @xmath359 , and so the desired correlator is simply @xmath360 in terms of the bulk ising magnetization operator , evaluated at a distance @xmath241 from the boundary [ see fig . [
fig : images ] ( a ) ] . finally , we note that @xmath361 is independent of @xmath250 due to translational invariance along the boundary .
the green function along the nfl to fl crossover then follows as , @xmath362^{\tfrac{7}{8 } } \langle \sigma(x ) \rangle,\\ \equiv & g_{\sigma\alpha,\sigma'\alpha'}(x ) , \end{split}\ ] ] ( with the factor @xmath363 required for normalization ) . at long distances where the green function describes the fl fp , eq .
[ factorgising ] implies @xmath364 as @xmath365 .
we now exploit a connection@xcite between the 2ik model at criticality with a simpler ising model to obtain @xmath366 and hence the exact green function @xmath367 along the nfl to fl crossover . from the boundary in ( a )
is mapped to the two point function of the associated chiral fields in the absence of a boundary shown in ( b ) and evaluated at image positions with respect to the line @xmath242 .
( c ) generalization of the two point function away from image points .
[ fig : images],width=321 ] when a small magnetic field @xmath27 is applied to the boundary of a classical ising model on a semi - infinite plane , the local magnetization shows a crossover@xcite as a function of distance from the boundary .
the crossover in this boundary ising model ( bim ) can be understood as an rg flow@xcite from an unstable fp at short distances ( with free boundary condition @xmath28 ) , to a stable fp at large distances ( with fixed boundary condition @xmath368 ) .
the universal crossover is characterized by an energy scale @xmath369 ( or a corresponding lengthscale @xmath370 ) .
importantly , it was shown in ref . that the rg flow in the bim due to small @xmath27 at the critical temperature is identical to the rg flow in the 2ik model due to small detuning @xmath371 at @xmath12 . as such , the nfl and fl fps of the 2ik model can be understood in terms of the bim fps with free and fixed boundary conditions .
the crossover energy scale in the 2ik model can then be identified as @xmath372 with @xmath230 as in table [ table : pert ] . in ref .
chatterjee and zamolodchikov derived an exact expression for the ising magnetization @xmath366 on the semi - infinite plane geometry in the continuum limit .
their result@xcite is @xmath373 with @xmath374 the modified bessel function of the second kind ; and @xmath167 for @xmath375 .
we take now @xmath376 ( corresponding to @xmath377 ) for concreteness . note that eq .
[ czmag ] yields asymptotically @xmath378 at long distances @xmath379 , consistent with the normalization of eq .
[ factorgising ] .
we now show that the analyticity of the green function and the local nature of the kondo interaction implies a generalization of @xmath367 in terms of spatial coordinate @xmath241 [ fig .
[ fig : images ] ( b ) ] , to @xmath380 in terms of general complex coordinates @xmath381 and @xmath382 [ fig .
[ fig : images ] ( c ) ] . using the free chiral green function eq .
[ g0 ] in the definition of the t matrix eq .
[ tdefinition ] , one obtains @xmath383 , \end{split}\ ] ] where @xmath275 and @xmath276 as before , and the t matrix is local in space .
the matsubara transform then yields @xmath384\\ \equiv & g_{\sigma\alpha,\sigma'\alpha'}\left ( \frac{z_1-z_2}{2i}\right ) .
\end{split}\ ] ] thus , the green function only depends on @xmath385 .
this is a somewhat counter - intuitive result , because the boundary breaks the translational invariance along the spatial coordinate @xmath241 . in appendix
[ app_analyticity ] we give an alternative proof , showing that eq .
[ tmatz1z2 ] holds to all orders in perturbation theory around the nfl fp . comparing eq .
[ tmatz1z2 ] and eq .
[ factorgising ] , it follows that @xmath386 when @xmath387 . employing this substitution in eqs .
[ factorgising ] and [ czmag ] , then taking the matsubara transform , we obtain @xmath388 where we have used @xmath389 as appropriate for eq .
[ czmag ] , and so @xmath252 is continuous . setting @xmath390 , @xmath391
, we define the infinitesimal @xmath392 , such that @xmath393 . with the integral representation of the bessel function @xmath394 ( for @xmath395 ) , we obtain @xmath396 with the integral over @xmath250 evaluated by contour methods , @xmath397 for negative matsubara frequencies @xmath398 , the green function then follows as @xmath399 where the @xmath400 integral has been expressed more simply in the last line in terms of the complete elliptic integral of the first kind , @xmath401 . since @xmath402 has a branch cut discontinuity in the complex @xmath403-plane running from @xmath147 to @xmath404 , the analytic continuation to real frequencies , @xmath405 can be performed without crossing any singularity .
thus , if one has @xmath367 as an _ analytic _ function of spatial coordinate @xmath241 ( as in eq . [ factorgising ] ) , then a full knowledge of both space and time dependences of the green function is implied by analytic continuation . using eq .
[ tdefinition ] , one recovers our earlier result@xcite for the @xmath12 crossover t matrix in the 2ik model due to perturbation @xmath342 , @xmath406\right ) , \ ] ] in terms of @xmath369 ( and with @xmath407 as before ) . here
@xmath408 is used for @xmath409 , corresponding to local singlet or kondo screened phases of the 2ik model , respectively .
[ tmatrixt=0 ] is thus equivalent to eqs .
[ exactresult ] and [ exactt=0 ] with the scattering s matrix @xmath410 . in the next section ,
we generalize these results to finite temperature .
our starting point for the derivation of the finite - temperature crossover green function is eq .
[ factorgising ] .
@xmath367 thus follows from the ising magnetization @xmath411 evaluated at temperature @xmath412 . in ref .
we considered the magnetization at temperature @xmath413 and distance @xmath241 from the boundary in a quantum 1d transverse field ising critical chain , with a magnetic field @xmath27 applied to the first spin at the point boundary .
it is given by @xmath414 with @xmath415}{\gamma[1 + 2\beta h^2]},\ ] ] and where @xmath416 is the result of leclair , lesage and saleur in ref .
, who generalized the @xmath12 result of chatterjee and zamolodchikov@xcite for the semi - infinite plane geometry ( eq .
[ czmag ] ) to the geometry of a semi - infinite _ cylinder _ with circumference @xmath268 , and magnetic field @xmath376 applied to the circular boundary .
the latter is equivalent to the quantum ising chain with transverse field .
we showed@xcite that while @xmath417 gives the full and highly nontrivial @xmath241 dependence of @xmath411 , it misses the multiplicative scaling function of the variable @xmath418 , given in eq .
[ ffunc ] . in the low - temperature limit @xmath389 , one obtains @xmath419 and @xmath420 such that eqs .
[ finitetisingmag ] , [ lls ] reduce as they must to eq .
[ czmag ] . in the limit @xmath421
, one recovers the fixed boundary condition , describing the fl fixed point .
again @xmath419 but eq . [ lls ] reduces now to @xmath422 this limiting behavior can also be obtained using the conformal mapping from the semi - infinite plane geometry with boundary @xmath423 , to the semi - infinite cylinder geometry parametrized by @xmath424 , with boundary @xmath425 , @xmath426 on the semi - infinite plane , the ising magnetization in the limit @xmath421 is known@xcite to decay as @xmath364 , yielding precisely eq . [ largeh ] on the semi - infinite cylinder . combining eqs .
[ factorgising ] , [ finitetisingmag][lls ] we obtain the crossover green function at finite temperature , @xmath427}\times \\ \frac{\gamma[\tfrac{1}{2 } + 2\beta h^2]}{\gamma[1 + 2\beta h^2]}~_{2}f_1\left(\frac{1}{2},\frac{1}{2};1 + 2 \beta h^2,\frac{1-\coth \frac{2 \pi x}{\beta}}{2 } \right ) .
\end{split}\ ] ] for @xmath421 , this gives correctly @xmath428)$ ] , as expected from the boundary cft result for the fl fp green function , eq .
[ cibc ] ( with @xmath429 ) .
of course , the function @xmath430 does become important when considering the behavior of the green function over the entire range of @xmath418 ( or equivalently , @xmath134 ) .
in particular , at high temperatures @xmath431 ( and finite @xmath27 ) , one obtains @xmath432 .
thus @xmath433 , again correctly recovering the known boundary cft result@xcite for the nfl fp green function , eq . [ cibc ] ( with @xmath434 ) .
the factor @xmath430 is indeed necessary to cancel the unphysical logarithmic divergence of @xmath435 as @xmath436 ( see also appendix [ pertfl ] ) .
the quantity of interest is of course the t matrix @xmath437 , related via eq .
[ tdefinition ] to the matsubara green function , itself obtained from eq .
[ gfinitetx ] via @xmath438 with @xmath381 and @xmath382 as usual . using the integral representation of the hypergeometric function , @xmath439}{\gamma[b ] \gamma[c - b ] } \int_0 ^ 1 dt ~\frac{t^{b-1}(1-t)^{c - b-1}}{(1-tz)^a},\ ] ] we then obtain @xmath440 where @xmath441 for @xmath442 and @xmath443 such that @xmath444 ; and with @xmath445 a negative integer . using contour integration , it can be shown that @xmath446 however , the naive substitution @xmath447 ( or @xmath448 ) is problematic here and leads to unphysical divergences . indeed , such analytic continuation always involves ambiguities due to the fact that @xmath449 on the integers , but it becomes @xmath450 upon analytic continuation .
thus , it is hard to find the function @xmath451 which gives the physical analytic continuation .
[ gfinitetx ] describes the chiral electron green function @xmath452 .
the information contained in such green functions is directly linked to the physical density oscillations around impurity ( friedel oscillations ) , which in turn can be calculated from the t matrix , @xmath453.@xcite indeed , in ref .
the real - space densities and hence green function @xmath454 for the nfl to fl crossover in the 2ck model was explicitly calculated at @xmath12 using the exact t matrix announced in ref . .
it was also highlighted@xcite that far from the impurity , the integral transformation relating the t matrix to the friedel oscillations can be inverted .
in this section we exploit these connections to calculate @xmath453 directly from the density oscillations described by @xmath454 , and thus circumvent the need for problematic analytic continuation . for simplicity we restrict ourselves to 1d in this section ; although we note that the resulting t matrix is general , because at low energies the standard flat band situation of most interest is recovered .
the density of the 1d fermion field @xmath455 at position @xmath241 is given by @xmath456 .
expanding around the left @xmath457 and right @xmath458 fermi points at low energies using eq .
[ chiralft ] , one obtains @xmath459 with the oscillating part of the density following as @xmath460 in the presence of particle - hole symmetry , @xmath461 for all @xmath241 ( with lattice spacing set to unity ) ; and there are no density oscillations , @xmath462 .
however , introduction of potential scattering breaks particle - hole symmetry and leads generically to real - space density oscillations , which contain information about the t matrix .
such potential scattering produces a phase - shift @xmath463 at the fermi energy , independent of the underlying kondo physics , but which does modify the boundary condition at @xmath242 , according to @xmath464 . as before ,
we define a chiral left - moving field on the infinite line , but now take into account this change in the boundary condition : @xmath465 using the imaginary time ordering of the chiral green function defined in eq .
[ factorgising ] , we have @xmath466 and @xmath467 , from which it follows that @xmath468 + { \rm{h.c.}}. \end{split}\ ] ] the oscillating part of the density given by eq .
[ denosc ] can also be obtained@xcite from the t matrix .
generalizing to finite temperatures , we have @xmath469 ^ 2 \mathcal{t}_{\sigma\alpha,\sigma\alpha}(\omega,\delta_p ) \right \ } , \end{split}\ ] ] where @xmath470 is the fermi function , @xmath471 is the free green function eq .
[ freeg ] as a function of real frequency @xmath10 ; and @xmath472 is the scattering t matrix , defined in the presence of the potential scattering . as shown in ref . , at low energies @xmath473 -i\ ] ] in terms of the desired t matrix defined _ without _ the potential scattering .
indeed , far from the impurity one obtains asymptotically@xcite @xmath474 ^ 2=-\tfrac{1}{v_f^2}e^{2i k_f
x + 2i \omega x /v_f}.\ ] ] the oscillating part of the density can then be expressed as @xmath475 e^{2i k_fx + 2i \delta_p + 2i\omega x / v_f } \\ & \times\big [ i \mathcal{t}_{\sigma\alpha,\sigma\alpha}(\omega ) -\tfrac{1}{2\pi\nu } \big ] + \text{h.c.}. \end{split}\ ] ] comparing now to eq .
[ denosc ] and inverting the fourier transform by operating on the resulting equations with @xmath476 , we obtain @xmath477 , \end{split}\ ] ] where @xmath454 is given as an analytic function in eq .
[ gfinitetx ] .
we now set @xmath247 and @xmath478 as before .
note that the hypergeometric function @xmath479 has a branch cut discontinuity in the complex @xmath403 plane running from @xmath147 to @xmath404 .
the discontinuity occurs only in the imaginary part of the function , with @xmath480 for @xmath481 .
furthermore @xmath482 for @xmath483 .
thus , integrating symmetrically above and below the real @xmath241 axis , as per eq .
[ resultt ] , amounts to taking only the _ real _ part of @xmath479 ; whence we obtain our final result @xmath484 using the definition of the crossover scale @xmath485 ( see eq .
[ bim2ikt * ] ) , this gives the announced result eqs .
[ exactresult ] , [ analyticformula ] for the 2ik model in the special case of perturbation @xmath377 , where @xmath486 . by simple extension , for @xmath151 ( corresponding to @xmath487 )
one obtains the same crossover but with @xmath488 .
finally , we note that taking the limit @xmath389 of eq .
[ result ] yields correctly eq . [ tmatrixt=0 ] . in the next section
we show that an _ identical _ crossover occurs in the 2ck model due to channel anisotropy .
the nfl fixed point hamiltonians of the 2ik model and the 2ck model have the same basic structure.@xcite although the underlying symmetries of the 2ck model are different from those of the 2ik model , the free conduction electron hamiltonian can be written in terms of the same mfs in both cases ( see eqs . [ bosonization][mfdecomp ] ) .
the cft decomposition@xcite of the 2ck model into @xmath489 symmetry sectors ( corresponding to conserved charge , spin and flavor ) , can then be expressed in terms of these mfs : the @xmath490 theory with central charge @xmath314 consists of a free boson or equivalently two mfs @xmath491 ; the spin @xmath315 theory with @xmath492 consists of three mfs @xmath493 ; similarly the flavor @xmath315 theory consists of three mfs @xmath494 .
the charge , spin and flavor currents can also be written in terms of the mfs corresponding to those symmetry sectors , as given in eq .
[ jchi2 ] of appendix [ quadratic ] . in particular , the nfl fixed point hamiltonian is of the form of eq .
[ mffreefp ] , with a boundary condition that is again simple in terms of the mfs . in the 2ck model ,
the nfl physics arises due to a modification of the boundary condition in the spin sector only ( the free boundary condition pertains in charge and flavor sectors ) .
the nontrivial boundary condition can be accounted for by defining the scattering states @xmath495 and @xmath496 , @xmath497 for @xmath498 .
indeed , the finite - size spectrum at the nfl fp@xcite can be understood in terms of excitations of a free majorana chain.@xcite the nfl fixed point of the 2ck model is destabilized by certain symmetry - breaking perturbations .
these perturbations can again be matched to mfs , with the correction to the nfl fixed point hamiltonian being of the form of eq .
[ deltahlambda1 ] in the simplest case of channel anisotropy @xmath499 ( see table [ table : pert]).@xcite importantly , it was shown recently in ref . that the nfl fixed points of the 2ck and 2ik model are in fact _ identical _ in the sense that they both lie on the same marginal manifold parametrized by potential scattering .
indeed , the low - energy effective model for the 2ik model in the limit of strong channel asymmetry is the 2ck model,@xcite but with an additional @xmath149 phase shift felt by the conduction electrons of one channel.@xcite for concreteness , we consider now a variant of the standard 2ik model in which channel asymmetry appears explicitly : @xmath500 one thus recovers eq .
[ 2ik ] at the symmetric point @xmath501 . in the limit @xmath502 ,
the left impurity is kondo screened by the left lead on the single - channel scale @xmath503 . at energies @xmath504 the right impurity is still essentially free .
however , it feels a renormalized coupling to its attached right lead , and an effective coupling to the remaining fermi liquid baths states of the left lead ( which suffer a full @xmath149 phase shift due to the first - stage kondo effect in that channel ) .
the relative effective coupling strengths between left and right channels can be tuned by the interimpurity coupling @xmath55 .
tuning @xmath55 to its critical value @xmath40 yields the 2ck critical point;@xcite while deviations @xmath342 correspond to finite channel anisotropy @xmath505 in the effective 2ck model . in consequence ,
one expects the nfl to fl crossover in the two models to be simply related .
since the nfl fixed point itself is the same in both 2ck and 2ik models ( up to potential scattering),@xcite and because the correction to the fixed point hamiltonian due to the @xmath135 perturbation is the same,@xcite the rg flow along the nfl to fl crossover is identical . to calculate the corresponding crossover green function , we simply incorporate the additional @xmath149 phase shift felt by the left - channel conduction electrons into our scattering states definition . using @xmath506 ,
one straightforwardly obtains @xmath507 in terms of the analytic crossover green function for the 2ik model given in eq .
[ gfinitetx ] . following the steps of sec .
[ friedel ] , the t matrix follows as @xmath508.\ ] ] this result was obtained at @xmath12 in ref . , where the correspondence was checked by explicit numerical renormalization group calculation .
in the 2ck model with @xmath509 , the left lead is more strongly coupled , and completely screens the impurity at the fl fp on the lowest energy scales ; while the right lead decouples asymptotically .
the physical interpretation of eq .
[ 2ik2cktmat ] is thus that a kondo resonance appears in the spectral function on the scale of @xmath13 in the @xmath59 channel , while the resonance is destroyed in the @xmath510 channel ( hence the dependence on the flavor - space pauli matrix @xmath511 ) .
in sec . [ finitetderiv ] we considered the finite - temperature crossover green function in the 2ik model due to the detuning perturbation @xmath342 ; while in sec
. [ 2ckgen ] we calculated the analogous crossover green function in the 2ck model due to channel anisotropy @xmath505 . in this section
we generalize the results to an arbitrary combination of perturbations in either model . before discussing the full calculation
, we motivate the general approach by exploiting a bare symmetry of the 2ck model , in a simple intuitive example . unlike the 2ik model , the 2ck model possesses a bare _ flavor _ symmetry ( see eq . [ 2ck ] ) .
the perturbations @xmath48 , @xmath49 and @xmath132 break this symmetry , but are themselves related by rotations in flavor - space .
a canonical transformation of the conduction electron operators of the bare hamiltonian is defined , viz @xmath512 such that the unitary matrix @xmath513 satisfies @xmath514 . with the parametrization @xmath515 ,
one obtains explicitly @xmath516 .
it follows that @xmath517 , with @xmath518 ( cf eq .
[ tstar ] ) .
the physical interpretation is that the impurity couples more strongly to one linear combination of channels than the other .
thus the perturbations @xmath135 , @xmath177 and @xmath519 in table [ table : pert ] are simply related , and their combined effect enters only through @xmath520 . in particular
this implies only one fitting parameter @xmath521 for the different components of the perturbation @xmath522 in the 2ck model .
the physical behavior in the case of arbitrary @xmath520 can now be understood in terms of the situation where @xmath135 alone acts using the flavor rotation eq .
[ cantrans ] .
for example , the green function @xmath523 probes the physical channel @xmath524 with @xmath59 in the original basis . using the transformation eq .
[ cantrans ] it can be expressed as @xmath525 in terms of the green functions in the rotated basis , which correspond to those calculated for @xmath526 only , as considered in the previous section .
it is then easy to show that the t matrix for arbitrary @xmath520 is given by @xmath527 in terms of the t matrix @xmath528 due to @xmath135 given in eq .
[ 2ik2cktmat ] .
the simple rescaling of the spectral function discussed in sec .
[ results ] and the precise form of eq . [ s2ck ] follow immediately . in this section we make use of the field theoretical description of the nfl fixed point for both 2ck and 2ik models in the presence of relevant perturbations.@xcite a large @xmath88 _ emergent _ symmetry at the fixed point allows these perturbations to be related by a unitary transformation , in full analogy to the method demonstrated explicitly in the previous section for the case of the bare flavor symmetry in the 2ck model .
we express the nfl fixed point hamiltonian in terms the free mf scattering states , @xmath529 + \delta h_{qcp},\ ] ] where @xmath530 $ ] is given in eq .
[ mffreefp ] , and with @xmath345 a vector of the 8 mfs . as already commented ,
the structure of the nfl fixed point hamiltonian is the same for 2ck and 2ik models ; only the definition to the scattering states is different . to fix notation
, we define @xmath531 for the 2ik model ; and @xmath532 for the 2ck model . with this ordering of components , the correction to the nfl fixed point due to relevant perturbations
is given by @xmath533 with @xmath156 a local mf as before .
the @xmath135 perturbation corresponding to @xmath342 in the 2ik model or @xmath505 in the 2ck model was considered explicitly in eq .
[ deltahlambda1 ] .
the other coupling constants are defined in table [ table : pert ] ; with the single resulting crossover energy scale being the sum of their squares , eq .
[ tstar ] . for a detailed derivation of eq .
[ deltah ] , see refs . .
the two mfs @xmath534 corresponding to the real and imaginary parts of the total charge fermion , are not in fact allowed in the 2ck and 2ik models due to charge conservation . the crucial observation following from eq . [ deltah ]
is that only the linear combination @xmath535 of the 8 mf scattering states participates in the crossover .
the particular linear combination depends on the ratios of the various perturbations ( for example @xmath536 , @xmath53 , @xmath54 in the 2ik model ) . from secs .
[ finitetderiv ] and [ 2ckgen ] , we know the crossover green function caused by the @xmath135 perturbation .
the strategy is thus to fix the @xmath135 perturbation as the direction in the 8-dimensional space of perturbations along which the green function is known , then use an @xmath88 rotation to obtain the general crossover green function .
we search now for a unitary operator @xmath537 that transforms the full hamiltonian with an arbitrary combination of perturbations into one involving the single perturbation @xmath135 .
specifically we demand that , @xmath538 u^\dagger & = h_{fp}[\vec{\chi}],\\ u \delta h_{qcp } u^\dagger & = i \lambda \chi_1(0 ) a. \end{split}\ ] ] this transformation is accomplished by an operator that rotates the 8-component vector @xmath345 in the 8-dimensional space of perturbations .
the 28 generators of such rotations are of the form @xmath539 , where @xmath540 is a real antisymmetric @xmath541 matrix .
it is easy to verify that the desired operator satisfying eq .
[ rot ] is @xmath542 where @xmath543 one can apply this transformation to the expectation value of an operator written in terms of the original electrons , such as the green function @xmath544 .
the crucial property of the unitary transformation eq .
[ u ] is that it acts as a simple rotation also on the electron fields .
this occurs due to the existence of linear relations between the 28 quadratic forms of the original electron fields and of the mfs @xmath545 , as discussed in appendix [ quadratic ] and ref . .
as a simple relevant example , consider the 2ik model , perturbed by a combination of @xmath536 and tunneling @xmath53 , such that only @xmath135 and @xmath177 are finite . in this case the unitary operator reads @xmath546 with @xmath547 and @xmath548 .
using eq .
[ list ] , @xmath549 .
the quadratic form @xmath550 is related to a quadratic form for the original electrons @xmath551 , as shown in appendix [ quadratic ] .
the operator @xmath552 can now be understood as a simple rotation of electron fields , @xmath553 for @xmath554 ; and where the rotation matrix acts here in flavor space , @xmath555.\ ] ] the green function then follows as @xmath556 where @xmath557 and @xmath558 is assumed . in terms of complex coordinates @xmath559 and @xmath560 ( with @xmath275 , @xmath276 ) , the full green function @xmath561 is then obtained from @xmath562 , as given in eq .
[ gfinitetx ] for the case of finite @xmath131 in the 2ik model .
now we define a @xmath260 unitary fermi liquid scattering @xmath563 matrix for the 2ik model @xmath564 such that @xmath565 using eq .
[ m1 ] one obtains @xmath566 for the case of finite @xmath135 and @xmath177 in the 2ik model ; consistent with eq . [ s2ik ] . for arbitrary combination of @xmath567 ,
one has @xmath568 /\lambda_\perp } \\ & = e^{i \frac{\theta}{2 } \int_{-\infty}^\infty dx~{\rm{sign(x ) } } \psi^\dagger [ \lambda_2 \tau^x+ \lambda_3 \tau^y + ( \vec{\lambda}_b \cdot \vec{\sigma } ) \tau^z ] \psi /\lambda_\perp}. \end{split}\ ] ] from eq .
[ m ] , it then follows that @xmath569 ( suppressing spin and channel indices ) . using eq .
[ smm ] we recover our final result for the 2ik model , eq .
[ s2ik ] .
following the same steps for the 2ck model ( and noting eq .
[ 2ik2ckcorr ] ) , we obtain @xmath570 with @xmath571 yielding precisely eq .
[ s2ck ] for the 2ck model .
wilson s numerical renormalization group@xcite ( nrg ) has been firmly established as a powerful technique for the accurate solution of a wide range of quantum impurity problems.@xcite its original formulation provided access to numerically - exact thermodynamic quantities for the kondo@xcite and anderson impurity@xcite models .
an increase in available computational resources subsequently allowed straightforward extension to multi - impurity and multi - channel systems.@xcite more recently , the identification of a complete basis within nrg ( the ` anders - schiller basis ' comprising discarded states across all iterations@xcite ) has permitted rigorous extension to calculation of dynamical quantities . in particular , equilibrium spectral functions can be calculated using the full density matrix approach,@xcite yielding essentially exact results at zero - temperatures on all energy scales .
although discrete nrg data must be broadened to produce the continuous spectrum,@xcite artifacts produced by such a procedure are effectively eliminated by averaging over several interleaved calculations ( the so - called @xmath403-trick@xcite ) .
indeed , resolution at high - energies can be further improved by treating the hybridization term exactly.@xcite our exact analytic results were tested and confirmed by comparison to nrg at @xmath12 in ref . . due to the logarithmic discretization of the conduction band inherent to nrg,@xcite finite - temperature dynamical information
can not however be captured@xcite on the lowest energy scales @xmath572 .
but spectral functions for @xmath573 are accurately calculated , and the total normalization of the spectrum is guaranteed,@xcite implying that the total _ weight _ contained in the spectrum for @xmath574 can be deduced . from a scaling perspective ,
one expects rg flow to be cut off on the energy scale @xmath575 , so that there should be no further crossovers in spectral functions for @xmath574 .
the somewhat arbitrary strategy@xcite commonly employed is thus to smoothly connect the spectrum calculated at @xmath576 , in such a way as to preserve the total weight .
our exact finite - temperature results for the crossover t matrix of the 2ck and 2ik models thus offers the perfect opportunity to benchmark nrg calculations for interacting systems exhibiting a non - trivial temperature - dependence of their dynamics . for concreteness
, we consider now the 2ck model with channel - anisotropy @xmath509 ( see eqs . [ 2ck ] and [ delta2ck ] ) . to obtain the numerical results ,
we discretize flat conduction bands of width @xmath577 logarithmically using @xmath578 , and retain @xmath579 states per iteration in each of @xmath580 interleaved nrg calculations.@xcite all model symmetries are exploited . to ensure the desired scale separation @xmath20 , we take representative @xmath581 and small @xmath582 , yielding @xmath583 and @xmath584 ( @xmath585 is defined in practice here from the t matrix eq .
[ specdef ] , @xmath586 .
@xmath13 is defined according to eq .
[ exactresult ] , corresponding here to @xmath587 ) . from eq .
[ tstar ] and table [ table : pert ] , we thus obtain @xmath588 . the t matrix for this 2ck model can be expressed as , @xmath589 with @xmath590 for @xmath329 and where @xmath591 as usual @xmath592 is the fourier transform of the retarded correlator @xmath593 . the alternative expression given in ref .
is : @xmath594^{-1},\ ] ] where @xmath595 is the green function for the ` 0'-orbital of the @xmath329 wilson chain.@xcite both @xmath596 and @xmath597 can be obtained directly by nrg , but eq .
[ nrg_tmatrix2 ] gives much better numerical accuracy,@xcite and is employed in the following .
the desired spectral function @xmath65 is then obtained from eq .
[ specdef ] , and is plotted in fig . [
fig : finitetnrg ] as the dotted lines for temperatures @xmath598 , as in fig .
[ fig : k ] . the corresponding exact results for the nfl to fl crossover from eq .
[ exactresult ] are plotted as the solid lines .
as immediately seen , near - perfect agreement is obtained for all energies @xmath153 and temperatures @xmath599 where comparison between numerical and exact results can be made .
to obtain such an agreement , we found that high - accuracy nrg calculations must be performed .
in particular , the region @xmath600 was most problematical , with artifacts only being removed upon averaging over several band discretizations , and necessitating a large number of states to be kept at each nrg iteration .
the precise shape of the numerically - obtained spectrum then still depends on how the discrete data is smoothed .
we found that the broadening scheme described in ref .
produced the best results : for @xmath580 and @xmath578 as used here , a broadening parameter @xmath601 and kernel - crossover scale @xmath602 were optimal .
it should also be noted that if the correction factor@xcite @xmath603 is used in the nrg calculations ( in which case @xmath604 ) , then the many - particle energies used to calculate the density matrix must be accordingly scaled ( @xmath605 ) so that the results are independent of the discretization parameter @xmath606 and hence approximate accurately the desired @xmath607 limit .
as discussed in the previous sections , the hamiltonian controlling the nfl fl crossover in the 2ck or 2ik models has a free fermion structure in terms of mfs .
in fact , this feature allows calculation of various quantities along the crossover .
the difficulty of such calculations is dictated by the relation between the physical quantity of interest to the mfs . in the preceding sections we concentrated on the two - point function of the electron field ( the green function ) , related@xcite here to the one - point function of the magnetization operator in the ising model ( which is in turn related non - locally to the ising mfs@xcite ) .
more generally , @xmath608-point functions of the electron field are related@xcite to @xmath609-point functions of the magnetization operator .
multi - electron correlators can thus in principle be calculated in this way , but require knowledge of the corresponding multi - point correlation functions of the ising magnetization operator .
since thermodynamic quantities are local in the mfs , their calculation is rather straightforward . here
we will focus on the nfl fl crossover of the impurity entropy , following closely the earlier calculations for the 2ck model@xcite and the 2ik model,@xcite performed in the toulouse limit .
the toulouse limit corresponds here to maximal spin anisotropy in the exchange couplings , and as such breaks the overall @xmath610 spin symmetry of the models .
although the high - energy @xmath611 crossover to the nfl fixed point is strongly affected by large spin - anisotropy , we stress that for low energies @xmath25 ( and given a clear scale separation @xmath612 ) , the results become formally exact , and are universally applicable to the @xmath610 symmetric case of interest .
the key point is that the spin - anisotropy perturbation is rg _ irrelevant _ at the nfl fixed point . in particular ,
the effective theory obtained in the toulouse limit describing the nfl fl crossover due to relevant perturbations such as channel anisotropy or magnetic field in the 2ck model@xcite or detuning @xmath536 , staggered magnetic field , or left - right tunneling in the 2ik model,@xcite act exactly as in eq .
[ deltah ] . a detailed discussion for the 2ik model can be found in ref . .
turning now to the crossover in the impurity entropy , one finds@xcite that @xmath613 in terms of the universal function @xmath614- \log \left [ \frac{1}{\sqrt{\pi}}\gamma \left ( \frac{1}{2}+\frac{1 } { t } \right)\right ] , \ ] ] defined in ref . for the limit @xmath615 . here , @xmath616 is the psi ( digamma ) function and @xmath617 is a particular definition of the nfl fl crossover scale ( proportional to our definition , eq .
[ exactresult ] , such that @xmath618 with @xmath619 ) .
two regimes can thus be distinguished . in the fl regime , obtained for @xmath143 , the impurity is always completely screened : @xmath620 . by contrast , in the nfl regime , @xmath621 , the impurity entropy is close to @xmath622 .
interestingly , we find that independently of the relevant perturbations which act , the entropy crossover is always given by the universal function eq .
[ entropycross ] in the limit @xmath623 , in both 2ck and 2ik models .
this is of course not the case for the green function , because the fl scattering s matrix is affected differently by different perturbations ( see eqs .
[ exactresult][analyticformula ] and figs .
[ fig : k ] , [ fig : mag ] ) .
the 2ck model has also been solved exactly using the bethe ansatz,@xcite yielding the full evolution of thermodynamics in any parameter regime .
however , it can not be seen directly from the bethe ansatz equations that there is an emergent @xmath88 symmetry at the nfl fixed point , or that this leads to a single nfl
fl crossover function for the entropy , eq .
[ entropycross ] , regardless of the perturbation causing the crossover .
indeed , the fact that the same crossover occurs in the 2ik model can not be extracted using bethe ansatz since the 2ik model is not integrable . in fig .
[ fig : entcross ] we present numerically - exact nrg results for the temperature - dependence of the entropy due to various perturbations in the 2ck and 2ik models to confirm the validity of the field theoretic description . as in fig .
[ fig : finitetnrg ] , we exploit all model symmetries to obtain high - quality numerics , discretizing flat conduction bands of width @xmath577 logarithmically , using @xmath624 here , and retaining @xmath579 states per iteration in a single nrg calculation.@xcite at low temperatures @xmath599 ( and since @xmath20 ) , we obtain an essentially perfect agreement between the exact result eq .
[ entropycross ] ( points ) and nrg data ( solid line ) .
it was shown in refs . that the effective free fermion theory of the 2ik model allows to calculate certain _ non - equilibrium _ quantities .
finite conductance was found to arise in the weak coupling limit of 2ik systems close to the critical point @xmath20 at low energies @xmath625 .
this result was understood in terms of the growth under rg of the left - right tunneling perturbation @xmath53 . here , we generalize these results to the 2ck model , which has the same effective free fermion description .
related multichannel setups have been considered in refs . .
schematic illustration of a non - equilibrium 2ck setup .
we consider the case of @xmath626.,width=151 ] we consider a finite source - drain voltage @xmath4 across left and right metallic leads , which are exchange - coupled to a single impurity spin . to this system
we add a small but finite channel anisotropy perturbation , corresponding left / right tunneling mediated via the impurity spin .
the setup is illustrated in fig .
[ fig : dot ] .
the corresponding hamiltonian is given by eqs .
[ 2ck ] and [ delta2ck ] , with finite @xmath48 and possibly magnetic field @xmath627 , but now with left / right lead chemical potentials at @xmath628 .
the applicability of our exact solution is in the parameter regime @xmath626 , so that the system is close to the nfl critical point .
this situation is not in practice obtained in standard quantum dot devices , although more sophisticated experimental techniques such as those employed in ref . , do allow suppression of cotunneling perturbations such as @xmath48 . as per eq .
[ tstar ] , the crossover energy scale is @xmath629 . in the limit where @xmath630 is initially very small , we thus have @xmath612 . at higher energies
@xmath631 , we then expect conductance to be very small @xmath632 , corresponding to the weak coupling limit
. however , upon reducing the energy scale @xmath633 below @xmath1 , the conductance starts to increase since @xmath48 switches on a relevant operator with scaling dimension @xmath175 near the nfl fp .
below @xmath13 , a characteristic peak in the conductance is thus expected , signaling growth of the relevant operator to order one . scaling function for the nonlinear conductance of a 2ck device with small @xmath48 . from top to bottom at the peak : @xmath634.,width=264 ] the exact lineshape of the non - equilibrium conductance peak can be calculated from the fixed point hamiltonian , eq . [ hqcp ] ( including the correction due to relevant perturbations given by eq .
[ deltah ] ) .
the dependence on the ratio between the magnetic field and the tunneling perturbations is obtained using the @xmath88 rotation outlined in sec .
[ gengf ] .
the method of calculation , and the result for the universal lineshape of the peak , was obtained for the spin - exchange anisotropic version of the model by schiller and hershfield in ref . .
as argued in the previous subsection , we can borrow those toulouse - limit results if we restrict attention to the low energy crossover .
the final result for the nonlinear conductance is thus @xmath635,\\ f[t , v ] = & \frac{1}{4 \pi t } { \rm{re}}~ \psi_1 \left ( \frac{1}{2}+\frac{1}{4 \pi t}+\frac{i v}{2 \pi t}\right ) , \end{split}\ ] ] with @xmath636 and @xmath637 the trigamma function .
note that the definition of @xmath13 here is as in ref . .
at @xmath12 , one obtains @xmath638^{-1}$ ] ; while at zero - bias @xmath62 and low - temperatures @xmath143 , the asymptotic conductance is @xmath639 .
the full bias - dependence of conductance for various temperatures is shown in fig .
[ fig : cond ] .
in this paper we present a rare example of an exact nonperturbative result for the finite - temperature dynamics of a strongly correlated quantum many - body system .
we focus on the two - channel kondo and two - impurity kondo models ; although the same low - energy physics characterizes a wide class of quantum impurity problems in which competition between two conduction channels causes a frustration of screening .
the unusual non - fermi liquid critical points of these systems are destabilized by various symmetry - breaking perturbations , naturally present in experiment . in consequence ,
a crossover to regular fermi liquid behavior always occurs on the lowest energy scales . exploiting the connection@xcite to an exactly - solved classical boundary ising model,@xcite we calculated the exact finite - temperature crossover green function . in quantum dot systems which could access this crossover , the relevant experimental quantity is conductance , which we extract from the exact green function .
remarkably , we show that due to the free fermion structure of the effective low - energy theory in terms of majorana fermions and a large emergent @xmath88 symmetry , a _ single universal function _ pertains for any combination of perturbations in either model .
this single crossover is also starkly manifest in the behavior of thermodynamic quantities such as entropy ; as confirmed directly by nrg .
the method developed in this paper goes beyond the impurity models we considered explicitly , and finds powerful application to a wider family of systems . at heart
, our solution relies upon a formal separation of the theory into a sector containing all the universal crossover physics , and a sector acting as a spectator along this crossover .
importantly , the crossover is confined to a sector which can be identified with ising degrees of freedom , described by a minimal conformal field theory with central charge @xmath317 .
for example , in the two - channel kondo models studied here , the full set of degrees of freedom consist of a @xmath640 cft , but a large @xmath641 sector of the theory plays no role in the crossover from non - fermi liquid to fermi liquid physics .
interestingly there exist other models ( whose full set of degrees of freedom are not necessarily described by a @xmath640 cft ) which undergo precisely the same crossover due to their underlying @xmath317 ising sector .
those include certain luttinger liquids containing an impurity,@xcite and coupled bulk and edge states in certain non - abelian fractional quantum hall states@xcite ( see also ref . ) .
there are further interesting generalizations and questions arising from this work .
for example , the two - channel kondo effect evolves continuously as interactions are switched on in the leads , as was shown in the case of luttinger liquid@xcite and helical liquid@xcite leads .
it is an open question as to whether the low - temperature crossovers in the presence of such interacting leads are described by the same boundary ising model , or e.g. by coupled boundary ising models
. it would also be interesting to use the present formulation of the crossover in terms of a minimal ising theory to study time - dependent phenomena , quench dynamics , and other non - equilibrium physics .
we thank l. fritz , h. saleur and a. rosch for helpful discussions .
this work was supported by the a. v. humboldt foundation ( e.s . ) and by the dfg through sfb608 and for960 ( a.k.m . ) .
in this appendix we consider the 2ik model for @xmath151 , and its fl fixed point describing the ground state where each impurity forms a kondo singlet with its attached lead .
in particular , we calculate the t matrix for @xmath642 as a stringent consistency check of our full crossover t matrix , eq .
[ exactresult ] .
indeed , we also see that the multiplicative function @xmath430 that we included in eq .
[ finitetisingmag ] is precisely needed to reproduce the correct fl limit .
we use here the fermi liquid theory of nozires,@xcite applied to the 2ik model .
naturally , our derivation of the t matrix in the vicinity of the fl fixed point follows closely the analogous calculation for the simpler single channel kondo ( 1ck ) model .
thus , we first recap some of the basic concepts and results for the 1ck model.@xcite the irrelevant operator in the effective nozires hamiltonian@xcite for the fl fixed point of the 1ck problem may be written in cft language as@xcite @xmath643 where @xmath644 is the spin current for a single channel of conduction electrons ( implicit summation over repeated indices is implied ) .
the first term of the second line may be interpreted as an elastic single particle scattering , and the second term can be interpreted as a residual electron - electron interaction giving rise to inelastic scattering .
accordingly , one can separate the contributions to the t matrix as @xmath645 , where@xcite @xmath646 . \end{split}\ ] ] [ for simplicity we omit spin indices ; @xmath647 . here
@xmath648 , @xmath649 , and @xmath650 is a particular definition of the kondo temperature.@xcite the fermionic diagram yielding the inelastic contribution is shown in fig .
[ fig : fldiag ] . .
in the 1ck model @xmath651 ; in two - channel kondo models @xmath652 .
[ fig : fldiag],width=170 ] the imaginary part of the t matrix in the 1ck model can thus be expanded as @xmath653 where from eq .
[ tmatrixfl ] one obtains @xmath654 similar to the nozires fermi liquid theory for the 1ck model , an effective fermi liquid hamiltonian was constructed using cft methods based on the emergent so(7 ) symmetry of the crossover for the 2ik model.@xcite the leading irrelevant operator takes the form@xcite @xmath655 where @xmath656 , @xmath657 is a scattering state incorporating the @xmath149 kondo phase shift , @xmath658 , and @xmath13 is a particular definition of the low energy crossover scale.@xcite we now calculate the t matrix resulting from this hamiltonian in the fl regime @xmath642 . in the following
we suppress the indices @xmath659 , and note that the t matrix is proportional to @xmath660 in the present situation .
comparison of the irrelevant operators in the 1ck and 2ik models , eqs .
[ nozieres ] and [ fl ] , shows that the first two terms in eq .
[ fl ] are identical to the nozires irrelevant operator ( up to the exchange of energy scales @xmath661 ) . in consequence the elastic and inelastic scattering contributions _ within _ each channel of the 2ik model are the same as those arising in the 1ck model . indeed , from the diagram fig .
[ fig : fldiag ] we see that to second order they do not yield any mixed terms .
thus @xmath662 , where @xmath663 originates from the third term in eq .
[ fl ] , representing interaction _ between _ channels .
we separate the latter into @xmath664 and note that the second order @xmath665 contributions are equal , since the quantity of interest is invariant under spin rotations , and the hamiltonian is also @xmath610 spin - symmetric .
the latter can be written in terms of fermion fields as @xmath666 considering now the t matrix for a single electronic species with quantum numbers @xmath667 , eq .
[ flint ] describes the interaction with a second species of either @xmath668 or @xmath669 , and thus contributes a term proportional to the inelastic contribution @xmath670 . in fact , the amplitude for this interaction is @xmath671 , identical in absolute value to the intra - lead interaction amplitude between up and down electrons [ the second term of the second line in eq . [ nozieres ] ] .
but since the contributions to inelastic scattering are of second order ( see fig .
[ fig : fldiag ] ) , the _ sign _ of the scattering amplitude is unimportant in calculation of the t matrix . summing over the second species @xmath672 yields an extra factor of 2 , yielding @xmath673 .
putting all the contributions together , we have @xmath674 as a result , one again obtains @xmath675 , but with @xmath676 this result is in perfect agreement with a numerical evaluation of our full finite - temperature crossover t matrix , eq .
[ exactresult ] in the limit @xmath677 ; as demonstrated in fig .
[ fig : tw ] .
this calculation also confirms the need for the function @xmath430 used in eq .
[ finitetisingmag ] .
the asymptotic result eq .
[ abratio ] also follows from renormalized perturbation theory calculations presented recently in ref . for a related 2ik model
in sec . [ gft0 ]
we used the analyticity of the green function and the locality of the t matrix to argue that @xmath678 depends only on the difference @xmath385 .
this result should hold to all orders in perturbation theory , as shown explicitly in this appendix .
our starting point here is the nfl fp green function for the 2ik model , eq .
[ factorg ] , written as , @xmath679 } \right ) ^{\tfrac{7}{8 } } \\ \times \langle \sigma_l(z_1)\sigma_l(z_2 ) \rangle , \end{split}\ ] ] whose factorized form originates from the bose - ising decomposition of the 2ik model into spin , isospin and ising symmetry sectors.@xcite 7 of the 8 mfs ( those corresponding to the part of the fermion field carrying spin and isospin quantum numbers ) remain free at the nfl fp , and thus give rise to the first factor ( see eq . [ freeg ] ) .
the second correlator involves the chiral part @xmath680 of the ising magnetization operator @xmath681 , due to the remaining mf in the ising sector ( see fig . [
fig : images ] ) . when the detuning perturbation @xmath682 acts , the nfl fp is destabilized .
the perturbation appears as a correction to the action,@xcite @xmath683 where @xmath684 is the cft @xmath272 boundary operator from the ising sector interpreted as a boundary magnetic field.@xcite the resulting corrections to the green function can then be calculated within perturbation theory .
the full crossover green function is then @xmath685 , where the @xmath686-th order correction is given by ( suppressing spin and channel indices ) @xmath687 } \right)^{\tfrac{7}{8 } } \int_0^\beta \prod_{i=1}^n d \tau_i \times \\ & \langle \sigma_l(z_1 ) \sigma_l ( z_2 ) \prod_{j=1}^n \epsilon(0,\tau_j ) \rangle . \end{split}\ ] ] generically , correlation functions up to 3-point functions are determined by cft .
however , the ising cft is special because essentially _ all _ correlation functions are known exactly . in particular , ardonne and
sierra obtained explicit expressions@xcite for the correlators appearing in eq .
[ integral2 ] . in the case of even @xmath686 their result reads,@xcite @xmath688 where the sum is over all subsets of @xmath689 , containing an even number of elements @xmath690 .
@xmath691 denotes the haffnian of a symmetric @xmath692 matrix , and is given by @xmath693 , with @xmath311 a permutation .
the set @xmath694 ( containing @xmath695 elements ) is equal to @xmath696 . using the conformal mapping eq .
( [ confmap ] ) from the plane to the cylinder , each coordinate difference @xmath697 in eq .
( [ eqardonne ] ) is replaced by @xmath698 $ ] .
the dependence on @xmath699 and @xmath700 is through factors which explicitly depend on @xmath701 , and terms inside the square - root of the form @xmath702 with the aid of the trigonometric identity , @xmath703 and by shifting all @xmath704 integration variables by @xmath705 into the complex plane ( which can be done without encountering any singularities ) , the resulting integral in eq .
[ integral2 ] then depends only on @xmath706 .
the same conclusion is reached for odd-@xmath686 by a similar calculation .
thus @xmath707 for all @xmath686 , and hence the full crossover green function depends only on @xmath385 .
analyticity of the green function @xmath708 thus allows determination of @xmath709 in terms of general coordinates @xmath699 and @xmath700 by analytic continuation .
here we derive the nfl coefficients @xmath710 and @xmath711 of the asymptotic t matrix discussed in sec.[results ] .
since the green function vanishes at the nfl fixed point , the leading correction arises to first order .
we now use the first order result for @xmath712 derived in ref .
, @xmath713 note that @xmath714}{\pi } $ ] with @xmath55 the complete elliptic integral of the first kind .
we will also use the short distance @xmath715 limit of this formula , @xmath716,\ ] ] valid for @xmath717 .
[ factorgising ] and [ shortdt ] give @xmath718.\ ] ] to obtain the expansion of the t matrix at @xmath12 at large @xmath10 we use @xmath719 in eq .
[ resultt ] . recalling that @xmath720 , and writing @xmath721 and @xmath722 , eq .
[ resultt ] becomes @xmath723 note that the @xmath724 integral should be made symmetrically around the branch cut , as described in sec.[friedel ] .
one then obtains eq.[asymptotict ] with @xmath725 with the numerical values of these integrals announced below eq .
[ asymptotict ] . at finite temperature and
@xmath91 we use eq .
[ pt ] in eq .
[ factorgising ] . taking the limit @xmath726 of eq .
[ resultt ] and defining @xmath727 , we obtain eq.[asymptoticw ] with @xmath728
there are @xmath729 independent quadratic forms involving @xmath730 and @xmath731 , which together comprise the generators of the @xmath88 symmetry group .
these generators are linearly related to the @xmath732 quadratic forms of the mfs @xmath348 . in this appendix
we gather and re - derive some of these relations , which can also be found in e.g. ref . .
first we define a convention relating the klein factors @xmath294 for the new fermions @xmath295 , to the klein factors @xmath289 for the original fermions @xmath284 .
the relations are fully determined by@xcite @xmath733 and by the anticommutation relations @xcite @xmath734 , @xmath735 .
we now consider the instructive example of the operator @xmath736 , and use eqs .
[ refermionization ] and [ mfdecomp ] to relate it to a quadratic term involving the original fermions : @xmath737 using eqs .
[ ff ] and [ linear ] we obtain , @xmath738 finally using the bosonization formula eq .
[ bosonization ] we have @xmath739 in a similar fashion , all of the relations between quadratic forms can be determined .
conserved currents in the 2ck and 2ik models can be expressed in terms of the original fermions or the mfs , and the relations between them are needed for our generalization of the crossover t matrix to arbitrary perturbation , as considered in sec .
[ gengf ] .
the conserved currents of the 2ck model are , @xmath740 equivalently , one can define a 3-component spin vector @xmath493 and flavor vector @xmath494 such that , @xmath741 in the 2ik model , there is no flavor @xmath610 symmetry since each channel couples to a different impurity
. however , one can make use of the @xmath610 total spin current @xmath744 , as well as @xmath610 isospin currents for each channel @xmath745 and @xmath746 , where @xmath747 in terms of mfs , we have @xmath748 and @xmath749 hence one can understand the conformal embedding of the 2ik model as a @xmath750 decomposition into total spin , left / right channel isospin and ising sectors .
3 of the 8 mfs represent the total spin sector ; 4 represent the isospin symmetry sectors ( the charge and flavor mfs ) ; and the remaining mf , @xmath325 , is associated with the ising model ( and restores the total central charge @xmath640 ) . | an exact expression is derived for the electron green function in two - channel kondo models with one and two impurities , describing the crossover from non - fermi liquid ( nfl ) behavior at intermediate temperatures to standard fermi liquid ( fl ) physics at low temperatures .
symmetry - breaking perturbations generically present in experiment ensure the standard low - energy fl description , but the full crossover is wholly characteristic of the unstable nfl state .
distinctive conductance lineshapes in quantum dot devices should result .
we exploit a connection between this crossover and one occurring in a classical boundary ising model to calculate real - space electron densities at finite temperature . the single universal finite - temperature green function
is then extracted by inverting the integral transformation relating these friedel oscillations to the t matrix .
excellent agreement is demonstrated between exact results and full numerical renormalization group calculations . |
an increasing number of independent observations indicates that we are living in an expanding universe where the expansion itself is accelerating @xcite .
it has been accepted that this requires some additional negative - pressure matter source , named dark energy . the simplest model explaining this accelerated expansion is the cosmological constant @xmath0 which corresponds to an unusual equation of state @xmath1 .
although the standard model of cosmology , the @xmath0 cold dark matter ( @xmath0cdm ) model , fits the present data very well , the numerical value of the cosmological constant is about 120 orders of magnitude smaller than the vacuum expectation value predicted by quantum field theory .
this smallness problem can be addressed by considering dynamical models , where the field slowly rolls down some potential where the effective equation of state @xmath2 converges to @xmath3 . in the past
it has been believed that this value should be approached from above .
however , there has been a recent interest in phantom models where the dark energy equation of state is approached from below @xmath4 , see @xcite .
these models , as counter intuitive as they may appear , are not excluded by current data @xcite .
figure [ fig1 ] , taken from @xcite , shows data taken from the cluster abundance , supernovae , quasar - lensing statistics and the first acoustic peak in the cosmic microwave background radiation ( cmb ) power spectrum , which all imply that there is convergence to dark energy .
however , when the parameter space is expanded to include the phantom region the data does not rule out a convergence from below @xcite .
in fact , the data even seems to indicate that this is quite likely .
considering such phantom energy models , the outcome of such a universe dominated by phantom energy is very different to any outcome we are accustomed to . the evolution parameter increases quicker than the horizon and quickly starts to pull apart gravitationally bound objects and then finally any bound objects , including the strong force ! however , since cosmologists turned their attention from a constant equation of state for dark energy to one where it evolves through time it is believed that with or without crossing the divide the equation of state should eventually converge to @xmath5 due to the success of the @xmath0cdm model , and that observationally it is so close to this value at present .
the majority of such dynamical dark energy models is based on evolving scalar fields with a suitably chosen potential .
however , a canonical scalar field can not cross the phantom divide and therefore such models have to modified , for instance by considering fields with negative kinetic energy . within theories of modified gravity one
is left with the option of either changing the geometrical side or the matter side of the gravitational field equations , changing either gravity itself or altering the matter content which contributes to the forces , respectively .
those two approaches are not entirely independent as many modified theories bring the new geometrical quantities to the matter side , ultimately changing the energy content of the universe , however allowing for a new interpretation . here
, we follow a different approach , namely we model the evolving dark energy component of the universe to be of the form of a single dark spinor field .
dark spinors , originally dubbed the elko spinor @xcite , are similar to majorana spinors but acquire the full four degrees of freedom of a dirac spinor due to their helicity structure .
they couple to the higgs mechanism via @xmath6{\lambda}}}\lambda h^{\dagger}h$ ] and weakly to the electromagnetic field via @xmath6{\lambda}}}[\gamma^a,\gamma^b]\lambda f_{ab}$ ] , however in the latter case this coupling is heavily constrained because of the masslessness of the photon , making them a candidate for dark matter .
the idea of one field explaining both dark matter and dark energy has already been discussed in various approaches , see e.g. @xcite .
one possible mass range for our dark spinors is in the @xmath7 range .
the dark spinor field is based on the eigenspinors of the charge conjugation operator , for a general introduction we refer the reader to @xcite . for the subsequent developments of dark spinors as an alternative model to scalar field inflation , see @xcite . for instance , in @xcite it could be shown that dark spinors have the effect of suppressing the low multipole amplitude of the primordial power spectrum , while at the same time to provide the usual inflationary features and predicting a small scale power spectrum in agreement with scalar field inflation , see in particular @xcite for dark spinors in cosmology .
the standard model of cosmology is based upon the flat friedmann - lematre - robertson - walker ( flrw ) metric @xmath8 where @xmath9 is the scale factor and @xmath10 is cosmological time .
the dynamical behavior of the universe is determined by the cosmological field equations of general relativity @xmath11 where @xmath12 is the planck mass which we use as the coupling constant , @xmath13 and @xmath14 .
@xmath15 denotes the stress - energy tensor , which for a homogeneous and isotropic cosmology takes the form @xmath16 the cosmological field equations are given by @xmath17 where the dot denotes differentiation with respect to time @xmath10 and the hubble parameter @xmath18 is defined by @xmath19 .
let us consider a homogeneous single dark spinor field .
following @xcite , the effective lagrangian density of this field can be written in terms of the _ scalar _ field @xmath20 as @xmath21 if the potential @xmath22 contains a standard mass term @xmath23 , then we can rewrite the lagrangian as @xmath24 this allows us to interpret the explicit presence of the hubble parameter in the action as an effective mass term where the mass changes as the universe evolves , and we have @xmath25 if the universe undergoes a phase of accelerated expansion , the hubble parameter is approximately constant .
depending on the ratio @xmath26 , it is possible for models to attain a negative value for @xmath27 without creating ghost of having negative kinetic energy .
this is possible because of the additional coupling of geometry to the matter field we are considering , a spinor opposed to a scalar field .
the energy density and the pressure of the dark spinor field are given by @xmath28 these two equations have the important property of leaving the acceleration equation unchanged @xmath29 the spinor field s potential energy may yield an accelerated expansion of the universe .
it should be noted that the energy density and the pressure now explicitely depend on the hubble parameter .
these additional terms are present because the covariant derivative when acting on the spinor field also acts on its spinorial part . one way of interpreting this ` coupling ' in eq .
( [ eq : n1 ] ) is to regard the effective mass of the particle to depend on the hubble parameter @xcite and therefore on the evolution of the universe . another interesting way of reading this equation is to regard the gravitational coupling to be time dependent @xcite
. the effective equation of state of the dark spinor field is given by @xmath30 and its more complicated form , when compared with the scalar field , also shows that a crossing of the phantom divide is possible without negative kinetic energy . if we for the moment take the canonical mass term for the potential and assume that @xmath31 , the effective equation of state becomes @xmath32 therefore , we expect dark spinor models with canonical mass term to cross the phantom divide , however , we do not expect the effective equation of state to diverge since once the field is slow - rolling in its potential , there exists a lower bound to @xmath2 . while there is a large number of potentials considered in the literature in the context of dynamical dark energy models , we will restrict our attention to power counting renormalizable potentials . as the dark spinor field has mass dimension one , the two allowed potentials are @xmath33 and @xmath34 where the @xmath35 is the aforementioned canonical mass term and @xmath36 includes the self interaction term with @xmath37 being a dimensionless coupling constant
let us start by solving eq .
( [ eq5 ] ) for the hubble parameter , using eq .
( [ eq8 ] ) we find @xmath38 hence , using the first equation , the energy density of the dark spinor field is @xmath39 & = \bigl(\frac{1}{2}\dot{\varphi}^2 + v(\varphi)\bigr)\bigl(1-\frac{(\varphi/{m_{\rm pl}})^2/8}{1 + ( \varphi/{m_{\rm pl}})^2/8}\bigr ) .
\label{eq : n5}\end{aligned}\ ] ] it is precisely this latter form of the energy density which motivated @xcite to interpret ( [ eq : n5 ] ) as inducing a time - dependent gravitational coupling by considering @xmath40 where @xmath41 is the standard energy density of a scalar field .
now , we consider the conservation equation ( [ eq7 ] ) with ( [ eq : n4 ] ) and ( [ eq : n5 ] ) and numerically solve the resulting equation for @xmath42 and substitute into eqs .
( [ eq8 ] ) and ( [ eq9 ] ) to obtain the evolution of the effective equation of state @xmath43 and plot it as a function of the evolution parameter @xmath9 .
we chose our initial conditions to be @xmath44 , the first two representing radiation and dust , respectively .
the initial conditions with @xmath45 correspond to an initially accelerating universe .
small changes in these initial conditions do not alter the late - time asymptotic behavior of the solutions .
[ fig1w ] shows the dynamical behavior of the effective equation of state considering the potential @xmath46 with @xmath47 , @xmath48 and @xmath49 .
for the two different mass values it is possible to see that the effective equation of state almost immediately drops below the phantom divide . during the subsequent evolution
, @xmath50 begins to increase as further shown by fig .
[ fig2w ] to the desired dark energy value . from fig .
[ fig2w ] it is also evident that our model is practically indistinguishable from dark energy modeled by a cosmological constant , long before recombination when the scale factor is about @xmath51 .
we obtained very similar results for other initial values of the equation of state : @xmath52 , @xmath53 and @xmath54 .
we have shown , for comparison , results from @xmath52 in fig .
[ figwintsamemass ] and fig .
[ fig2wint ] , they qualitatively agree with the results presented in figs .
[ fig1w ] and [ fig2w ] , respectively .
the next natural step is to consider a potential which allows for a self - interaction terms . as discussed above
, the inclusion of this term also yields a power counting renormalizable field theory .
when we include the self interaction term in the potential , @xmath36 , we find , for all initial values of @xmath50 and @xmath55 , that the effective equation of state always diverges to @xmath56 , see fig .
( [ crazy ] ) .
we have also checked , although not included in here , that our numerical results for @xmath36 converge to results for @xmath35 as @xmath57 . although it might be possible to construct models with finely tuned initial conditions such that the divergence of the equation of state would happen in the future , such models are very unlikely .
hence , we are lead to conclude that a dynamical dark energy model based on dark spinors requires their potential to be of the simplest form , namely a canonical mass term , without self interaction .
lastly , we found another set of interesting results where the equation of state oscillated between @xmath58 and @xmath5 for all time .
the oscillation of the equation of state is very rapid , as can be seen in fig .
[ crazy1 ] , making them unphysical as a dark energy model .
this qualitative behavior does not change if we include the self - interaction term . however
, if such a model could be modified it would be a prime candidate for models where the field changes its characteristic from being dark matter at early times to become dark energy at late times , see also @xcite .
a dark spinor field is able to provide a possible model for dark matter as it couples mainly via the higgs mechanism , but has heavily constrained interactions with the electromagnetic field .
dark spinors have a predicted @xmath59 mass range and therefore experimental predictions can be formulated and possibly measured at the lhc .
our results now show that the dark spinor field is also capable of having a dynamical equation of state which crosses the phantom divide and asymptotes to @xmath5 .
this makes it a viable candidate for dark energy which can not be ruled out experimentally . unlike previous phantom models , dark spinors
do not obtain negative kinetic energy on crossing the phantom divide , due to both @xmath60 and @xmath61 depending on the hubble parameter and therefore these models do not create ghosts . according to @xcite the equation of state must not stay below the divide but converge to dark energy , therefore the dark spinors potential is of the simplest form , a canonical mass term @xmath62 .
our dark spinor model does not require a modification of general relativity , leaving one of the most successful models in theoretical physics untouched . due to the interesting nature of dark spinors
, they have been shown to give other unique properties not found with other matter sources considered in the past .
for now , in a cosmological setting , dark spinors are providing intriguing results in having the potential to be the best candidate dynamical dark energy model at hand . | ever since the first observations that we are living in an accelerating universe , it has been asked what dark energy is .
there are various explanations all of which with have various draw backs or inconsistencies . here
we show that using a dark spinor field it is possible to have an equation of state that crosses the phantom divide , becoming a dark phantom spinor which evolves into dark energy .
this type of equation of state has been mildly favored by experimental data , however , in the past there were hardly any candidate theories that satisfied this crossing without creating ghosts or causing a singularity which results in the universe essentially ripping .
the dark spinor model converges to dark energy in a reasonable time frame avoiding the big rip and without attaining negative kinetic energy as it crosses the phantom divide . |
it has been experimentally established that below a critical value of the magnetic field , vortex lattices undergo a first order transition in clean high temperature superconductors @xcite .
this has been seen in yba@xmath0cu@xmath1o@xmath2 ( ybco ) @xcite , bi@xmath0sr@xmath0cacu@xmath0o@xmath3 ( bscco ) @xcite and ( la@xmath4sr@xmath5)@xmath6cuo@xmath7 @xcite .
evidence for first order phase transitions comes from latent heat measurements @xcite and peaks in the specific heat @xciteas well as jumps in the resistivity @xcite and in the magnetization @xcite .
this transition is generally accepted as a melting transition from a vortex solid to a vortex liquid . a great deal of theoretical work @xcite has helped to establish that there is a first order melting transition .
brzin , nelson and thiaville @xcite showed that including fluctuation effects in abrikosov s mean field theory of the flux lattice transition would drive the transition first order .
the advent of the high temperature superconductors and the subsequent experimental indications of vortex lattice melting sparked intense theoretical activity .
early analytic efforts used the lindemann criterion @xcite , though such an approach could not show that the transition was first order .
studies of the mapping between vortex lines and the world lines of a 2d system of bosons have suggested a first order transition from an abrikosov lattice to an entangled vortex liquid @xcite .
numerical simulations @xcite have been able to show that the melting is first order by calculating quantities such as the magnetization jump @xcite and the delta function in the specific heat @xcite . however these simulations were done in the limit of high magnetic fields @xmath8 where @xmath9 is the coherence length , @xmath10 is the spacing between vortices , and @xmath11 is the penetration depth .
most @xcite assumed that the magnetic induction @xmath12 was spatially uniform and thus neglected the wavevector dependence of the elastic moduli .
this has made quantitative comparison with experimental data difficult , especially in the case of bscco whose vortex lattice melts at low fields .
in addition the mechanism for vortex lattice melting is still not well established .
there have been suggestions that topological defects @xcite such as vortex loops @xcite , vortex antivortex pairs @xcite , free disclinations @xcite , and dislocations @xcite may play a key role in triggering melting . in this paper
we show that melting can be induced by interstitial and vacancy line defects in the vortex lattice which soften the shear modulus @xmath13 .
this softening makes it easier to introduce more defects and increases the vibrational free energy .
the increased vibrations ultimately lead to melting .
there is good agreement with the experimental curve of transition temperature versus field , latent heat and magnetization jumps for ybco and bscco . using a viscoelastic approach
, we show that the shear modulus is frequency dependent . at zero frequency
the vortex liquid can not sustain a shear while at high frequency the liquid has a finite shear modulus .
since we can calculate the free energy for both the lattice and the liquid at melting , we have estimated the surface tension between a vortex line liquid and a vortex solid .
let us describe our scenario for melting .
our approach follows that of granato @xcite as well as previous work which showed that defects can lead to a first order phase transition @xcite .
we start with a vortex lattice in a clean layered superconductor with a magnetic field @xmath14 applied perpendicular to the layers along the c - axis .
we consider the vortices to be correlated stacks of pancake vortices .
we will assume that the transition is induced by topological defect lines , i.e. , vacancies and interstitials .
in a delaunay triangulation @xcite a vacancy or an interstitial in a triangular lattice is topologically equivalent to a pair of bound dislocations @xcite as well as to a twisted bond defect @xcite .
high temperature decoration experiments @xcite and monte carlo simulations @xcite have found such defects to be thermally excited .
the introduction of these defects softens the elastic moduli .
since the energy to introduce interstitials and vacancies is proportional to the elastic moduli , softening makes it easier to introduce more defects .
the softening also increases the vibrational entropy of the vortex lattice which leads to a melting transition .
the transition is driven by the increased vibrational entropy of the vortex lines of the lattice , and not by the entropy of the wandering of the defect lines .
in fact frey , nelson and fisher @xcite showed that a phase transition driven by the entropy of wandering flux lines occurs at a much higher magnetic field than what is observed experimentally . in the vicinity of the experimentally observed first order phase transition , wandering in the transverse direction by more than a lattice spacing is energetically quite costly and therefore rare .
such flux line bending also makes dislocations @xcite energetically costly at low dislocation densities .
( the energy scale is set by @xmath15 @xcite .
here @xmath16 is the interplane spacing and @xmath17 , the energy per unit length of a vortex , is given by @xmath18 where @xmath19 is the flux quantum and @xmath11 is the penetration depth for currents in the _ ab _ plane . for example , for ybco
@xmath20 k at @xmath21 k and for bscco @xmath22 k at @xmath23 k @xcite .
note that @xmath24 . )
the first order transition is nucleated in a small region by a local rearrangement of existing line segments . slightly above the melting temperature
@xmath25 a vortex line can distort and make an interstitial and a vacancy line segment that locally melt the solid .
this is the analog of a liquid droplet which nucleates melting of a crystal .
the role of the surface tension is played by the energy to connect the interstitial segment to the rest of the vortex line .
this connection can be a josephson vortex lying between planes or a series of small pancake vortex displacements spread over several layers .
when the length @xmath26 of the interstitial and vacancy segments equals the critical length @xmath27 , the energy gained by melting equals the energy cost of the connections . when @xmath28 , it is energetically favorable for the defect segments grow to the length of the system . we are ignoring the surface tension associated with the surface parallel to the c axis .
we shall show later that this is quite small .
to study melting we assume that we have a vortex lattice with interstitial and vacancy lines extending the length of the lattice .
our goal is to find the free energy density as a function of the concentration @xmath29 of defect lines .
the free energy density is @xmath30 where @xmath31 is the free energy density of a perfect lattice , @xmath32 is the work needed to introduce a straight interstitial or vacancy line into the lattice , @xmath33 is the vibrational free energy density of the system , and @xmath34 is the free energy due to the wandering of the defect lines over distances large compared to the lattice spacing .
we now examine these terms in detail .
@xmath31 , the free energy density of a perfect rigid flux lattice , is given by the london term @xcite : @xmath35 where @xmath36 is the spatially averaged magnetic induction , @xmath9 is the coherence length in the @xmath37 plane , and @xmath38 is 0.130519 for a hexagonal lattice and 0.133311 for a square lattice @xcite . for @xmath36 near @xmath39 , @xmath31
is given by the abrikosov free energy @xcite @xmath40 } \label{eq : abrikosov}\ ] ] where the ginzburg - landau parameter @xmath41 , and the abrikosov parameter @xmath42 is 1.16 for a triangular lattice and 1.18 for a square lattice .
to calculate @xmath33 , we follow ref . @xcite .
we denote the displacement of the @xmath43th vortex pancake in the @xmath29th plane from its equilibrium position by @xmath44 where @xmath45 and the pancake position @xmath46 .
the fourier transform @xmath47 $ ] .
@xmath48 and @xmath49 is the wavevector along the c - axis .
@xmath50 where @xmath51 is the volume and the vibrational partition function @xmath52 is given by @xmath53 where we have divided by the area @xmath54 of the normal core of a pancake @xcite . @xmath55 and @xmath56 are the real and imaginary parts of @xmath57 and @xmath58 .
the elastic free energy functional associated with these distortions is @xmath59 where @xmath60 and @xmath61 , the volume per pancake vortex is @xmath62 , and @xmath16 is the interplane spacing .
the @xmath63 sum is over a circular brillouin zone @xmath64 .
the matrix @xmath65 is given by @xmath66 where @xmath67 , @xmath13 , and @xmath68 are the bulk , shear , and tilt moduli , respectively .
@xmath69 for a hexagonal lattice .
diagonalizing @xmath65 leads to 2 eigenvalues : @xmath71 where @xmath72 is the diagonal matrix , the subscript @xmath26 denotes longitudinal and @xmath73 denotes transverse . using this leads to @xmath74 where @xmath75 . after integrating over @xmath76 in ( [ eq : z ] ) , the remaining sums over @xmath63 and @xmath49 are converted to integrals : @xmath77 where the volume of the sample is set to unity .
the integrals in ( [ eq : zintegrals ] ) are done numerically . at low fields ( @xmath78 ) ,
the elastic moduli are given by @xcite @xmath79 } { 4\pi[1+\lambda_{ab}^{2}(k^{2}+q^{2})](1+\lambda_{c}^{2}k^{2 } + \lambda_{ab}^{2}q^{2 } ) } \nonumber \\ c_{44 } & = & \frac{b^{2}}{4\pi(1+\lambda_{c}^{2}k^{2}+\lambda_{ab}^{2}q^{2 } ) } + \frac{b\phi_{o}}{32\pi^{2}\lambda_{c}^{2 } } \label{eq : moduli } \\ & & \mbox { } \times \ln\frac{\xi_{ab}^{-2}}{k_{o}^{2}+(q/\gamma)^{2 } + \lambda_{c}^{-2 } } + \frac{b\phi_{o}}{32\pi^{2}\lambda_{ab}^{4}q^{2 } } \ln(1+\frac{q^{2}}{k_{o}^{2 } } ) \nonumber \\ \nonumber\end{aligned}\ ] ] where @xmath80 is the penetration depth for currents along the c - axis , @xmath81 is the anisotropy , and @xmath82 . at high fields ( @xmath83 ) @xcite ,
@xmath13 is altered by the factor @xmath84 and the penetration depths in @xmath85 and @xmath68 are replaced by @xmath86 where @xmath87 denotes either @xmath11 or @xmath80 .
in addition the last two terms of @xmath68 are replaced by @xmath88 .
these replacements guarantee that the elastic moduli vanish at @xmath39 . for ybco the temperature dependence of the penetration depths and coherence lengths
are given by @xmath89 @xcite and @xmath90 , respectively .
for bscco whose melting field is two orders of magnitude below @xmath39 , @xmath91 and @xmath92 @xcite .
the free energy density @xmath32 due to the energy cost of adding a vacancy or interstitial vortex line is difficult to calculate accurately @xcite . however , we can write down a plausible form for @xmath32 by noting that a straight line defect parallel to the c - axis produces both shear and bulk ( but not tilt ) distortions of the vortex lattice .
for example , if a defect at the origin produces a displacement @xmath93 that satisfies @xmath94 where @xmath95 is a two dimensional delta function , then @xmath96 @xcite . inserting this in ( [ eq :
elastic ] ) , we find that @xmath97 where @xmath98 . generalizing this to allow for a more complicated distortion and for a concentration @xmath29 of line defects , we write @xcite @xmath99 where @xmath100 and @xmath101 are dimensionless constants .
we expect the isotropic distortion to be small , i.e. , @xmath102 , and the shear deformation to dominate , i.e. , @xmath103 .
integrating over @xmath29 allows the elastic moduli to depend on defect concentration .
we will assume that @xmath67 is independent of @xmath29 since we believe that the bulk modulus of the vortex solid is roughly the same as that of the liquid phase .
to find @xmath104 @xcite , we use its definition @xmath105 where @xmath106 is the shear strain . assuming that @xmath67 has negligible shear strain dependence , we find @xmath107 or @xmath108 if we shear the lattice in the @xmath37 plane along rows separated by a distance @xmath109 , the system must be unchanged if the displacement is equal to a lattice spacing .
this is a result of the discrete translational symmetry of the lattice .
the shear modulus should reflect this discrete translational symmetry and therefore must be periodic in displacements equal to the lattice constant @xmath110 .
we describe this with the simplest even periodic function : @xmath111 where the shear strain @xmath112 .
notice that this expression goes beyond the usual harmonic approximation . then taking the second derivative of eq .
( [ eq : cosinec66 ] ) , we obtain @xmath113 where @xmath114 . combining this with ( [ eq : c66 ] )
, we obtain @xmath115 where @xmath116 is given in eq .
( [ eq : moduli ] ) .
thus the shear modulus softens exponentially with the defect concentration @xmath29 .
this softening lowers the energy cost to introduce further defects , and increases the vibrational free energy @xmath33 when @xmath104 is used in @xmath65 .
substituting @xmath104 in eq .
( [ eq : c66exp ] ) into our expression ( [ eq : fw ] ) for @xmath32 yields @xmath117 the softening of the shear modulus with increasing defect concentration is well known in the case of atomic lattices @xcite .
there it has been shown both experimentally @xcite and theoretically @xcite that interstitials can substantially soften the elastic constants with the largest change being in the shear modulus .
linear extrapolation of the experimentally measured change of the shear modulus of copper would imply that the lattice becomes unstable for a concentration of about 3% interstitials @xcite .
an example of how interstitials can soften the shear modulus is illustrated in figure [ fig : dumbbell ] .
here we show a triangular lattice where an interstitial forms a dumbell aligned in the @xmath118 direction by sharing a site with another atom or flux line .
dumbbell displacements along the @xmath119 direction introduce a string like librational resonance mode consisting of displacements along the @xmath120 directions .
this mode couples strongly to an external shear stress and results in softening of the shear modulus @xcite .
the last term we need to consider is @xmath34 , the free energy due to the wandering of the defect lines over distances large compared to the lattice spacing .
we can estimate @xmath34 with the following expression @xcite @xmath121 where @xmath122 for a triangular lattice ( bscco ) and @xmath123 for a square lattice ( ybco ) .
@xmath124 can be thought of as the distance along the z axis that it takes @xcite for the defect line to wander a transverse distance of one lattice spacing @xmath125 . to go from one vacancy or interstitial site to the next , the defect line segment must jump over the barrier between the two positions .
this gives @xmath124 a thermally activated form : @xmath126 , where @xmath127 and @xmath128 .
@xmath129 is the line tension and is given by @xmath130 .
numerical simulations @xcite indicate that the barrier height @xmath131 is small and we use @xmath132 . @xmath34 itself
is quite small compared to the other terms because of the high energy cost of vortex displacements .
for example , at the transition @xmath34 is about two orders of magnitude smaller than @xmath32 or @xmath33 .
thus the transition is not driven by a proliferation of wandering defect lines because near the transition the high energy cost of vortex displacements is not sufficiently offset by the entropy of the meandering line @xcite .
before we plot @xmath133 versus @xmath29 , we note that the difference between @xmath36 and @xmath14 is negligible for ybco but can be a significant fraction of the melting field @xmath134 for bscco . to find the value of @xmath36 to use in the helmholtz free energy density @xmath133 , we minimize the gibbs free energy density @xmath135 , i.e. , @xmath136 where @xmath137 . since the concentration dependence of @xmath36 is negligible , we find @xmath36 for @xmath138 for each value of @xmath14 and @xmath139 .
typical plots of @xmath140 versus @xmath29 are shown in the inset of figure 1 .
the double well structure of @xmath141 is characteristic of a first order phase transition .
the equilibrium transition occurs when both minima have the same value of @xmath141 .
we associate the minimum at @xmath138 with the vortex solid and the minimum at finite @xmath29 with the vortex liquid .
the defect concentration at the transition is only a few percent .
( [ eq : c66exp ] ) implies that a finite value of @xmath29 yields a finite value for the shear modulus , e.g , for @xmath142% , @xmath143 for bscco .
previous work interpreted this to mean that the lattice did not melt .
however , they did not appreciate the fact that the shear modulus is frequency dependent and the @xmath13 used here is the high frequency response . at high frequencies
it is the elastic response which dominates and this is what enters into the expression for the free energy . for a liquid
the low frequency response is dominated by viscosity so that the zero frequency shear modulus is zero .
we will elaborate more on this later .
in figure [ fig : melting ] we fit the experimental first order transition curves in the @xmath144 plane using 2 adustable parameters : @xmath100 and @xmath101 .
as expected , @xmath103 and @xmath102 ( see figure [ fig : melting ] ) . the geometrical quantity @xmath145 can have several values for a given lattice structure , depending on
which planes are sheared .
we choose @xmath146 where @xmath147 is the angle between primitive vectors .
decoration experiments on bscco find a triangular lattice @xcite , so we use @xmath148 . for ybco
we choose @xmath149 which is very close to a square lattice which has @xmath150 .
maki @xcite has argued that the @xmath109-wave symmetry of the order parameter yields a square vortex lattice tilted by 45@xmath151 from the @xmath152axis .
experiments @xcite on ybco find @xmath147 ranging from @xmath153 to @xmath154 .
we can calculate the jump in magnetization @xmath155 at the transition using @xmath156 where @xmath157 . here
@xmath158 is the defect concentration in the liquid at the melting transition .
the jump in entropy @xmath159 is given by @xmath160 where @xmath159 is the entropy change per vortex per layer .
the results are shown in figure [ fig : mag - s ] .
we have checked that our results satisfy the clausius - clapeyron equation @xmath161 .
we obtain good agreement with experiment for ybco and the right order of magnitude for bscco .
the difference between theory and experiment in the temperature dependence of the entropy and magnetization jumps for bscco may be due to the decoupling of the planes @xcite .
this enhances the thermal excursions and hence the entropy of the pancake vortices @xcite .
decoupling may be brought about by other types of defects such as dislocations which we will discuss in section 5b .
we can compare our results with the lindemann criterion by calculating the mean square displacement @xmath162 at the transition using eq .
( [ eq : z ] ) : @xmath163 where @xmath72 is given by ( [ eq : a ] ) and @xmath164 labels the 2 eigenvalues . defining the lindemann ratio @xmath165 by @xmath166 , we find that @xmath167 for ybco at @xmath168 t and that @xmath169 for bscco at @xmath170 g. here we have used the same values of the parameters that were used to fit the phase transition curves in figure [ fig : melting ] .
these values of @xmath165 are consistent with previous values @xcite .
experiments have found little , if any , hysteresis @xcite .
this is consistent with our calculations .
we can bound the hysteresis by noting the range of temperatures between which the liquid minimum appears and the solid minimum disappears .
typical values for the width of this temperature range are 300 mk for ybco at @xmath171 and 1.3 k for bscco at @xmath172 g. another measure of the hysteresis can be found in the plots of @xmath141 versus @xmath29 ( see inset of fig . [
fig : melting ] ) . the barrier height @xmath173 between the minima is low ( @xmath174 mk ) which is consistent with minimal hysteresis . in going from the normal metallic phase to the vortex solid , two symmetries
are broken : translational invariance and gauge symmetry which produces the superconducting phase coherence along the magnetic field . in the liquid
, longitudinal superconductivity is destroyed by the wandering and entanglement of the vortex lines . even though line wandering is energetically costly and therefore rare ,
it does occur . as a result ,
the correlation length along the c axis will be quite long and of order @xmath124 .
this is consistent with measurements in ybco of the c axis resistivity which find that there is loss of vortex velocity correlations for samples thicker than 100 @xmath175 m @xcite . for an infinitely thick sample , the loss of longitudinal superconductivity coincides with the melting transition @xcite
this agrees with experiments which indicate that the loss of superconducting phase coherence along the c axis coincides with the first order transition @xcite .
the vortex line wandering renormalizes the coupling between the planes in the liquid phase , so it is difficult to estimate the surface tension parallel to the @xmath37planes which is primarily due to josephson vortices .
however , since we have expressions for the free energy in both the liquid and solid phases , we can estimate the surface tension parallel to the c axis along the melting curve .
we imagine a plane interface parallel to the c axis between the vortex liquid and the vortex lattice phases .
the surface tension @xmath176 is given by @xmath177 where @xmath178 is the gibbs free energy as a function of position and the constant @xmath179 is the gibbs free energy far away from the interface , for example within the solid phase where @xmath138 .
( at melting the vortex liquid and solid phases coexist because they have the same bulk value for the gibbs free energy . ) in the interface region the defect concentration @xmath29 changes from zero in the solid phase to a finite value in the liquid phase .
let us assume that in this region the concentration gradient @xmath180 is a constant @xmath181 where @xmath182 is the concentration of defects in the bulk liquid phase .
here we are assuming that the width of the interface is of order a vortex lattice constant @xmath10 .
then @xmath183 this is an integral of the area under the barrier between the solid and liquid phases in the plot of the gibbs free energy versus defect concentration ( see inset of fig . [
fig : melting ] ) . using the values for @xmath139 and @xmath36 along the melting curve
, we find the surface tension given in fig .
[ fig : surface ] .
the dependence of the surface tension on the melting temperature @xmath184 reflects that of the barrier height @xmath185 on @xmath184 .
the order of magnitude of the surface tension is given by @xmath186 , and as a result , the values are quite small . for example , at @xmath139=60.24 k and @xmath36=202.28 g , @xmath176=0.015 k/@xmath187 for bscco , and at @xmath139=80.9184 k and @xmath14=6.4807 t , @xmath188 for ybco , where @xmath16 is the interplane spacing and @xmath189 .
we believe these are correct order of magnitude estimates for the surface tension since the small values of the barrier height is consistent with the small amount of hysteresis found experimentally @xcite .
we now discuss the viscoelastic behavior of the vortex liquid .
as we mentioned earlier , the shear modulus is frequency dependent .
the low frequency response to a shear stress is flow and this is characterized by a viscosity @xmath38 . at high frequencies
the response is elastic and the crossover between the two occurs over a narrow frequency range so that the shear modulus as a function of frequency is rather like a step function .
this behavior can be simply modeled using the maxwell model @xcite for viscoelasticity in which a massless spring is damped by a viscous force .
the rate of shear strain @xmath190 is given by @xmath191 where @xmath176 is the shear stress , @xmath192 is the time derivative of the shear stress , and @xmath193 is the frequency . using the maxwell relation for the relaxation time @xmath194 and the definition of the frequency dependent shear modulus @xmath195 , we find that the real part of the shear modulus is given by @xmath196 \label{eq : shearmod}\ ] ] notice that @xmath197 which confirms that the vortex liquid can not sustain a shear stress . at high frequencies @xmath198
is given by @xmath199 . to estimate the crossover frequency we need to estimate the viscosity .
we are interested in the shear viscosity which arises from the interactions between vortices .
there are other sources of viscosity .
for example a single moving vortex line experiences a viscous drag due to the normal electrons in the core which produce resistance when they move with the vortex .
this is described by the bardeen
stephen model @xcite .
there is also viscosity which arises from pinning ; we will ignore this contribution since we are considering a clean lattice .
we can obtain a simple estimate for the viscosity following the approach of dyre , olsen and christensen @xcite .
shear flow occurs when some vortices push past other vortices .
the viscosity @xmath38 is given by @xmath200 \label{eq : visc}\ ] ] where the prefactor @xmath201 is the viscosity of a single noninteracting vortex line and is given by the bardeen stephen relation @xcite @xmath202 where @xmath203 is the normal state resistivity and @xmath204 is the speed of light . in ( [ eq : visc ] ) @xmath205 is the activation energy which is identified with the work done per vortex pancake in shoving aside the surrounding vortices .
the elastic energy associated with distorting the vortices is given by ( [ eq : elastic ] ) ; we identify @xmath205 with @xmath206 at the maximum distortion @xmath207 produced by the shoving .
the actual form of the distortion is difficult to determine analytically .
we will assume that the dominant contribution comes from tilt and shear ; and that there is no change in the density so that the contribution from the bulk modulus can be ignored .
in bscco in the vortex liquid , the planes are decoupled and the tilt modulus can be ignored . in ybco in the liquid
the correlation along the c axis can be quite long as we discussed earlier . in this case
the distortion involves various wavevectors ; the wavevector dependence of the tilt modulus @xmath68 is such that at small @xmath49 , @xmath68 is roughly comparable to the high frequency shear modulus @xmath199 and at large @xmath49 , @xmath208 .
( since we are considering the elastic response , it is appropriate to consider the high frequency shear modulus . )
so as a crude estimate we will assume the displacement is pure shear and write @xcite @xmath209 where @xmath210 is the volume change due to shoving and rearranging vortices .
since @xmath205 is the energy per vortex pancake , @xmath210 is some fraction @xmath211 of the volume @xmath212 per vortex pancake , i.e. , @xmath213 .
@xmath214 is given by eq .
( [ eq : c66exp ] ) . in fig .
[ fig : visc ] we show the reduced viscosity @xmath215 along the melting line for both ybco and bscco with @xmath216 .
as one can see , interactions enhance the viscosity @xmath38 over the noninteracting viscosity @xmath201 by a factor of 2 or less .
this is because @xmath217 along the melting line . using our estimate of the viscosity and eq .
( [ eq : shearmod ] ) , we can calculate the frequency dependence of the shear modulus @xmath198 as shown in fig .
[ fig : shearmod ] for a defect concentration of 5% .
as expected the shear modulus has the shape of a step function ; it is zero at low frequencies and rises quite sharply to its infinite frequency value @xmath199 .
notice that for bscco the crossover frequency is a few mhz and for ybco it is a few ghz . since 1
k corresponds to 20 ghz , this means that we made an excellent approximation in setting @xmath218 in the free energy density @xmath133 in eq .
( [ eq : freeng ] ) .
any theory that tries to describe melting has to contain two main ingredients : a satisfactory description of both the solid and the liquid phases and a mechanism by which the system goes from one to the other .
the difficulty has always been to describe two phases with wildly different properties within the same framework . in the present work
, we have achieve this by viewing the liquid as a solid with a finite concentration of vacancies and interstitial .
clearly this is an approximation . if one wishes to describe a liquid as a solid with defects , other types of defects have to be taken into account as well .
this is particularly true of dislocations .
while we do not believe that thermally excited dislocations play the key role in mediating the melting transition because of their high energy cost , we believe that they will proliferate as soon as the vacancies and interstitials appear .
there are two reasons for this .
first interstitial vortex lines are usually attractive @xcite and can aggregate to form dislocations that extend the entire length of the lattice parallel to the c axis .
the same is true for vacancies .
in particular , olive and brandt @xcite have done numerical simulations on line defects which were at least 5 lattice spacings apart .
they found that both centered and edge interstitials @xcite are attractive if @xmath219 . for @xmath220 , edge interstitials were attractive for separations less that 10 lattice spacings and repulsive at larger distances while centered interstitials were repulsive at distances larger than 5 lattices spacings .
vacancies were attractive in all cases .
the second reason is that the substantial softening of the shear modulus brought about by the vacancies and interstitials reduces the dislocation core energy as well as the elastic energy of creating dislocations .
for example , the core energy of a z directed dislocation goes as @xmath221 and the core energy of a screw dislocation goes as @xmath222 where @xmath223 is the magnitude of the burger s vector @xcite . in addition the long range interaction between dislocation loops is mediated by the strain field and depends on @xmath13 @xcite . once the dislocations have proliferated , the method and results of marchetti and radzihovsky @xcite can be used to provide a more detailed and accurate description of the liquid side of the transition .
for example they show that when dislocations at all length scales are present , the shear modulus vanishes in the long wavelength limit @xcite .
work in this direction is in progress .
to summarize we have presented a model for the melting of a vortex lattice into a vortex liquid .
the melting transition is induced by a few percent of vacancy and interstitial vortex lines that soften the shear modulus and increase the vibrational entropy .
the increased vibrational entropy leads to melting .
we obtain good agreement with the experimentally measured curve of transition temperature versus field , latent heat , and jumps in magnetization for bscco and ybco .
the lindemann ratio @xmath165 is @xmath224% for bscco and @xmath225% for ybco .
the hysteresis is small .
we find a very small surface tension between the vortex solid and the vortex liquid along an interface parallel to the c axis .
the shear modulus is frequency dependent ; it is zero at @xmath226 and plateaus at higher frequencies to its infinite frequency value .
we thank andy granato and lev bulaevskii for helpful discussions .
this work was supported in part by onr grant n00014 - 96 - 1 - 0905 and by funds provided by the university of california for the conduct of discretionary research by los alamos national laboratory .
although calculations indicate that fluctuations in the density of vacancies and interstitial defects do not renormalize @xmath13 @xcite , @xmath13 does decrease as the average number of defects increases . another way to estimate
the free energy @xmath205 in ( [ eq : visc ] ) is to use the energy difference between a centered interstitial and an edge interstitial @xcite .
a centered interstitial vortex line is at the center of a triangle whose vertices are the nearest neighbor lattice sites of a triangular lattice .
an edge interstitial lies at the midpoint of one of the sides of the triangle .
numerical calculations @xcite find this energy to be roughly @xmath235 k @xmath236 .
so on the melting curve , @xmath237 .
this estimate of @xmath205 is an order of magnitude smaller than the estimate we used in eq .
( [ eq : deltaf ] ) . as a result
the the crossover frequency , where @xmath198 goes from its low frequency to its high frequency value , increases by less than a factor of two compared to the crossover frequencies shown in fig .
[ fig : shearmod ] . | we show that thermally activated interstitial and vacancy defects can lead to first order melting of a vortex lattice .
we obtain good agreement with experimentally measured melting curve , latent heat , and magnetization jumps for yba@xmath0cu@xmath1o@xmath2 and bi@xmath0sr@xmath0cacu@xmath0o@xmath3 .
the shear modulus of the vortex liquid is frequency dependent and crosses over from zero at low frequencies to a finite value at high frequencies .
we also find a small surface tension between the vortex line liquid and the vortex lattice .
epsf 2 |
branching ratios of @xmath3 meson two - body nonleptonic decays have been measured by cleoiii , belle and babar collaborations @xcite .
cp violations in these modes may be observed in near future .
cognizant of this point , we have presented some theoretical anticipations for the @xmath10 @xcite , @xmath11 @xcite , and @xmath12 @xcite decays in the perturbative qcd ( pqcd ) framework . in particular
, @xmath13 cp violation is expected in the @xmath10 decays .
the @xmath14 transition form factors are the integral part of two - body nonleptonic decay amplitudes . in this paper
we shall convince readers that these form factors in the large recoil region of light mesons are calculable in pqcd .
this is where our approach starts to differ from other approaches to exclusive @xmath3 meson decays .
according to pqcd factorization theorem , a form factor is written as the convolution of a hard amplitude with initial - state and final - state hadron distribution amplitudes @xmath15 , where @xmath16 is the momentum fraction associated with one of the partons .
it has been pointed out that perturbative evaluation of the pion form factor suffers nonperturbative enhancement from the end - point region with a momentum fraction @xmath17 @xcite .
if this is true , the hard amplitude is characterized by a low scale , such that expansion in terms of a large coupling constant @xmath18 is not reliable .
more serious end - point ( logarithmic ) singularities have been observed in the twist-2 ( leading - twist ) contribution to the @xmath19 transition form factor @xcite .
the singularities even become linear at twist 3 ( next - to - leading twist ) @xcite . because of these singularities
, it was claimed that the @xmath19 form factor is dominated by soft dynamics and not calculable in pqcd @xcite .
we shall argue that this conclusion is false .
we shall show that at the end points , where the above singularities occur , the double logarithms @xmath20 should be resummed in order to justify perturbative expansion .
the result , called threshold resummation @xcite , leads to strong sudakov suppression at @xmath17 @xcite .
therefore , the end - point singularities do not exist in a self - consistent pqcd analysis . in this work we shall investigate contributions to the @xmath21 and @xmath22 transition form factors from twist-2 and from two - parton twist-3 distribution amplitudes . in sec .
ii we illustrate the pqcd formalism by studying the pion electromagnetic form factor .
we review the reasoning why one might conclude that the form factor is not calculable , and explain why these objections are not justified in qcd . in secs .
iii and iv we derive the @xmath3 meson transition form factors .
it will be shown that the twist-3 contributions , which seem to be proportional to @xmath23 or @xmath24 , do not vanish in the @xmath25 limit .
here @xmath26 , @xmath27 , and @xmath2 are the chiral symmetry breaking scale , @xmath28 meson mass , and @xmath3 meson mass , respectively .
we record our results of the form factors at large recoil : the @xmath19 form factor @xmath29 and the @xmath22 form factor @xmath30 .
meson distribution amplitudes are defined and the sudakov factor from threshold resummation is derived in the appendices .
the suggestion that a hadronic form factor is calculable in pqcd was first made in ref .
the rough idea is summarized as follows .
one expands the bound - state wave function for a pion in terms of fock states containing on - shell partons ( quarks or gluons ) @xcite , @xmath31 define a soft function @xmath32 at a typical hadronic scale @xmath33 as the initial wave function , @xmath34 the wave function @xmath35 can be related to @xmath32 via @xmath36 where @xmath37 is an irreducible kernel and @xmath38 the green function involving only hard loop momenta .
the pion electromagnetic form factor @xmath39 is then expressed as a convolution integral , @xmath40 with @xmath41 being the momentum of the initial - state pion , @xmath42 large momentum transfer , and @xmath43 . here
we have written the parton momenta associated with the initial state and final state as @xmath44 and @xmath45 , respectively , in the notation @xmath46 , and made explicit the dependence of the two - parton wave function @xmath47 on @xmath41 and @xmath48 . the first term in eq .
( [ fef ] ) contains leading contributions , and ellipses represent those from higher fock states , which are down by powers of @xmath49 in the light - cone gauge and by powers of @xmath18 .
the leading diagrams are displayed in fig . [ leading ] .
it can be shown that the large momentum transfer @xmath50 flows through the hard amplitude @xmath51 , and that all nonperturbative dynamics goes into wave functions .
one can therefore compute @xmath51 perturbatively .
however , it was pointed out that the above argument suffers a grave difficulty @xcite : the diagrams in fig .
1 may be infrared divergent , because important contribution to the form factor comes from the region , where the exchanged gluons are soft .
pqcd is then not applicable .
below we shall examine this difficulty in more details . according to eq .
( [ fef ] ) , the first diagram in fig .
[ leading ] gives @xmath52 with the color factor @xmath53 , the number of colors @xmath54 , and the hard amplitude @xmath55^{\gamma\delta } \frac{1}{(k_2-k_1)^2}\left[\gamma^\sigma\frac{\not k_2-\not p_1 } { ( p_1-k_2)^2}\gamma_\mu \right]^{\alpha\beta}\;. \label{perp}\end{aligned}\ ] ] write @xmath56 .
if we ignore @xmath57 , it has been shown that the integral in eq .
( [ pff ] ) is dominated by contributions from the end - point regions with @xmath58 , @xmath59 .
if the pion wave function does not vanish at @xmath17 , the integral will be even infrared divergent .
if we somehow regulate the infrared singularity by an appropriate choice of the wave function , the running coupling constant @xmath60 evaluated at the hard gluon momentum @xmath61 is still too large to make sense out of the perturbative expansion .
the above end - point singularity corresponds to the picture of the pion form factor feynman had in mind . in the so - called brick - wall frame the initial - state pion with momentum @xmath62
is struck by a space - like current of momentum @xmath63 , and turns around with momentum @xmath64 as shown in fig .
[ feynman ] .
feynman pointed out that the major contribution to the form factor comes from the region , where one of the partons carries the full pion momentum .
the rest of partons , being all wee partons , do not know in which direction they are moving .
the resulting configuration is essentially identical to the initial - state pion except that the momentum of the fast parton is reversed .
hence , feynman claimed that the @xmath50 dependence of the pion form factor is related to the probability of finding a single parton carrying all the pion momentum .
feynman s picture is consistent with the statement that the form factor is dominated by the singular part of eq .
( [ perp ] ) .
because it is singular , we can not compute the form factor .
we argue that feynman s picture of the pion form factor is false .
consider a qed example .
when an electron undergoes hard scattering , it can not help but emit infinitely many photons in the direction of the electron momentum . as a consequence ,
the elastic scattering cross section @xmath65 at finite angle vanishes at high energy , implying that the probability for the final @xmath66 state being accompanied by no photons diminishes . in other words ,
the final state must be accompanied by many photons . in the qcd case of the pion form factor , when a quark inside the pion gets hit by a current
, the final state will contain many gluons unless the spectator quark is nearby to shield the color charge .
when one of the quarks carries all the momentum , the rest of the pion can not shield the color charge of the fast quark , and many gluons will be emitted in arbitrary directions during the hard scattering .
thus , the final configuration ending up as a single pion is extremely unlikely .
this is so - called sudakov suppression on exclusive processes at kinematic end points .
therefore , the contribution from fig .
[ feynman ] is negligible , and the end - point singularity does not exist ! in the above argument we have ignored the term @xmath57 in eq .
( [ pff ] ) .
small @xmath4 in momentum space corresponds to large transverse distance @xmath67 of two valence quarks .
the color charge of the quark , which is struck by the current , is not shielded in this large @xmath67 region , and will emit many gluons . the probability for having a single pion in the final state is then vanishingly small .
that is , this configuration can not contribute to the form factor .
hence , the momentum space with @xmath68 , where the end - point singularity occurs , is also sudakov suppressed @xcite .
the typical behavior of the sudakov factor @xmath69 $ ] , @xmath70 , which is associated with the struck quark , is shown in fig .
[ suda2 ] .
we observe that the sudakov factor decreases fast at large @xmath67 for @xmath71 ( @xmath72 ) , which corresponds precisely to the end - point region in eq .
( [ perp ] ) . in conclusion ,
the end - point singularity is absent , and the major contribution to fig .
[ leading ] comes from the region with hard gluon exchanges .
as derived in the appendix a , a light - cone pion distribution amplitude is written as @xmath73_{\beta\gamma}\phi_\pi(x ) + [ \gamma_5]_{\beta\gamma}m_0\phi^p_\pi(x)\right .
\nonumber\\ & & \left.+m_0[\gamma_5(\not n_+\not n_--1)]_{\beta\gamma } \phi^t_\pi(x)\right\}\ ; , \label{fpd}\end{aligned}\ ] ] where @xmath74 is the pion momentum , the light - like vector @xmath75 the coordinate of the @xmath76 quark , and the dimensionless vector @xmath77 parallel to @xmath78 and @xmath79 parallel to @xmath80 .
here a four - vector has been expressed in terms of light - cone coordinates , @xmath81 the distribution amplitude @xmath82 is twist-2 , and @xmath83 and @xmath84 proportional to @xmath85 gev , where @xmath86 is the current quark mass of the quark @xmath42 , are twist-3
. the origin of these terms can be simply understood by means of the field - current identity from chiral symmetry , @xmath87 it is easy to observe that twist-3 contributions are suppressed by a power of @xmath88 .
the asymptotic behaviors of @xmath82 , @xmath83 and @xmath84 are known to be @xmath89 as the hard amplitude in eq .
( [ perp ] ) is convoluted with these distribution amplitudes , we find that twist-2 contribution is finite , while twist-3 ones are logarithmically divergent without sudakov suppression .
the sudakov factor then introduces an effective cut - off to the integral at @xmath90 , and the twist-3 contributions are proportional to @xmath91 .
that is , the power counting is not altered by a logarithmic divergence in the factorization formula .
as shown later , the power counting for contributions to the @xmath3 meson transition form factors is modified by linear divergences in the factorization formulas .
therefore , the different end - point behavior leads to different power counting rules for the pion form factor and for the @xmath3 meson transition form factors .
in the @xmath3 meson rest frame , we define the @xmath3 meson momentum @xmath41 and the pion momentum @xmath92 in the light - cone coordinates : @xmath93 with the energy fraction @xmath94 carried by the pion .
the spectator momenta @xmath48 on the @xmath3 meson side and @xmath95 on the pion side are parametrized as @xmath96 note that the four components of @xmath48 should be of the same order , @xmath97 , with @xmath98 , @xmath99 being the @xmath67 quark mass .
however , since @xmath95 is mainly in the plus direction with @xmath100 , the hard amplitudes will not depend on the plus component @xmath101 as explained below .
this is the reason we do not show @xmath101 in eq .
( [ k12 ] ) explicitly .
consider the configuration for the semileptonic decay @xmath102 depicted in fig .
[ aa ] , which corresponds to soft contribution to the @xmath19 form factor @xmath103 .
the @xmath104 quark and the lepton pair fly back to back with energy of @xmath105 .
the spectator quark @xmath76 carries a momentum of @xmath97 .
if this configuration is responsible for the decay , it is impossible to compute @xmath103 using pqcd .
however , applying an argument similar to that used for the pion form factor , we know that the @xmath104 quark recoiling against the lepton pair is bound to emit infinitely many gluons . thus , fig .
[ aa ] in fact corresponds to the inclusive decay @xmath106 .
the probability that the final state in fig .
[ aa ] contains only a single pion is suppressed by the sudakov form factors .
a quantitative estimate of sudakov suppression of the soft contribution to @xmath103 in the qcd sum rule formalism will be discussed later .
it has been explained that the internal @xmath107 quark involved in the hard amplitude becomes on - shell as the momentum fraction @xmath16 of the @xmath76 quark vanishes @xcite .
the contributions to the @xmath21 form factor @xmath103 are then logarithmically divergent at twist 2 and linearly divergent at twist 3 .
we argue that as the end - point region is important , the corresponding large double logarithms @xmath20 need to be organized into a jet function @xmath108 as a consequence of threshold resummation @xcite .
this jet function vanishes as @xmath109 , and modifies the end - point behavior of meson distribution amplitudes effectively .
this modification provides a plausible explanation for the model of the twist-3 pion distribution amplitude proportional to @xmath110 , which was adopted in @xcite .
our numerical study shows that the results of the @xmath19 form factor obtained in this work are almost the same as those obtained in @xcite . in the following analysis
we shall employ the approximate form , @xmath111^c\ ; , \label{trs}\end{aligned}\ ] ] where the parameter @xmath112 comes from the best fit to the next - to - leading - logarithm threshold resummation in moment space .
note that the jet function @xmath113 is normalized to unity . for details of the derivation ,
refer to the appendix d. similarly , the inclusion of @xmath4 regulates the end - point singularities , and large double logarithms @xmath114 are produced from higher - order corrections .
these double logarithms should be also organized to all orders , leading to @xmath4 resummation @xcite .
the resultant sudakov form factor , whose explicit expression can be found in our previous works @xcite , controls the magnitude of @xmath115 to be roughly @xmath116 by suppressing the region with @xmath117 .
the coupling constant @xmath118 is then small enough to justify the pqcd evaluation of heavy - to - light form factors @xcite .
we emphasize that the hard scale for heavy - to - light decays must be @xmath119 in order to define a gauge - invariant @xmath3 meson distribution amplitude @xcite .
we shall include the sudakov factor associated with the light spectator quark of the @xmath3 meson .
whether this factor is essential will be determined by the @xmath3 meson distribution amplitude . since the @xmath3 meson is dominated by soft dynamics with @xmath120 ,
the associated sudakov effect is minor compared to that from the energetic pion . with the possible order of magnitude of @xmath121 , a taylor expansion of the hard gluon propagator near the end point ,
@xmath122 is certainly not appropriate .
a more reasonable treatment is to keep @xmath115 in the denominators of internal particle propagators , and to drop @xmath115 in the numerators , which are power - suppressed compared to other @xmath123 terms . under this prescription ,
the sudakov factor from @xmath4 resummation can be introduced into pqcd factorization theorem without breaking gauge invariance of the hard amplitudes . for the same reason
, the terms proportional to @xmath124 in the numerators should be neglected .
it is then obvious from eq .
( [ fex ] ) that the hard amplitudes are independent of the component @xmath101 .
the @xmath101 dependence of the @xmath3 meson wave function can then be integrated out @xcite , leading to the parametrization in eq .
( [ k12 ] ) .
note that the mechanism of threshold and @xmath4 resummations is similar with the former responsible for suppression in the longitudinal direction and the latter for suppression in the transverse direction .
as shown below , both twist-2 and twist-3 contributions are well - behaved after including threshold and @xmath4 resummations .
hence , the contributions to @xmath103 from fig .
[ aa2 ] dominate in the large recoil region . in this configuration
the @xmath76 quark gains a large momentum parallel to the @xmath104 quark momentum by exchanging a hard gluon with the @xmath107 or @xmath104 quark .
we compute the @xmath19 form factors @xmath125 and @xmath126 defined by the following matrix element , @xmath127 + f_0(q^2)\frac{m_b^2-m_\pi^2}{q^2}q_\mu\;,\end{aligned}\ ] ] where @xmath128 is the lepton - pair momentum .
another equivalent definition is @xmath129 in which the form factors @xmath130 and @xmath131 are related to @xmath125 and @xmath126 by @xmath132 the factorization formula for the @xmath19 form factors is written as @xmath133 the pion distribution amplitude @xmath134 has been supplied in eq .
( [ fpd ] ) , and the @xmath3 meson wave function is given by ( see the appendix c ) @xmath135_{\alpha\delta}\;. \label{bwp1}\end{aligned}\ ] ] employing eqs .
( [ fpd ] ) and ( [ bwp1 ] ) , we derive , from eq .
( [ fbpi ] ) , @xmath136 e(t^{(1)})h(x_1,x_2,b_1,b_2)\ ; , \label{f1}\\ f_2&=&16\pi m_b^2c_f\int dx_1dx_2\int b_1db_1 b_2db_2\phi_b(x_1,b_1 ) \nonumber\\ & \times & \biggl\{\left[\phi_\pi(x_2)(1+x_2\eta ) + 2r_\pi\left((\frac{1}{\eta } -x_2 ) \phi_\pi^t(x_2 ) -x_2\phi_\pi^p(x_2 ) \right)\right]e(t^{(1)})h(x_1,x_2,b_1,b_2 ) \nonumber\\ & & + 2r_\pi\phi_\pi^p e(t^{(2})h(x_2,x_1,b_2,b_1)\biggr\}\ ; , \label{fpi}\end{aligned}\ ] ] with the ratio @xmath137 and the evolution factor @xmath138 in the above formulas we have dropped the terms proportional to the momentum fraction @xmath120 as argued before , which are power - suppressed compared to the leading terms such as @xmath139 in the form factor @xmath131 .
the explicit expressions of the sudakov exponents @xmath140 and @xmath141 are referred to @xcite .
the hard function is written as @xmath142\ ; , \label{dh}\end{aligned}\ ] ] where the factor @xmath113 suppresses the end - point behaviors of the pion distribution amplitudes , especially of the twist-3 ones .
the hard scales @xmath143 are defined as @xmath144 it is obvious that if turning off threshold and @xmath4 resummations with @xmath18 fixed , eqs .
( [ f1 ] ) and ( [ fpi ] ) are infrared divergent .
we argue that the two - parton twist-3 distribution amplitudes @xmath145 , though proportional to the ratio @xmath23 , need to be taken into account . as stated above
, the corresponding convolution integrals for the @xmath19 form factor are linearly divergent without including sudakov effects .
these integrals , regulated in some way with an effective cut - off @xmath146 , are proportional to the ratio @xmath147 . combining the two ratios @xmath23 and @xmath147 , the twist-3 contributions are in fact not down by a power of @xmath1 : @xmath148 and should be included in a complete leading - power analysis .
we emphasize that the presence of linear divergences modifies the power counting rules , making the difference between the @xmath3 meson transition form factors and the pion form factor .
various computing methods have been proposed for the evaluation of the @xmath19 transition form factors @xmath149 in the literature , such as the lattice technique @xcite , light - cone qcd sum rules @xcite , and pqcd @xcite .
obviously , lattice calculations become more difficult in the large recoil region of the light meson .
however , this region is the one where pqcd is reliable , indicating that the pqcd and lattice approaches complement each other .
this complementation will be explicitly exhibited in fig .
[ bpi ] below . in light - cone sum rules ,
dynamics of the @xmath19 form factors have been assumed to be dominated by the large scale of @xmath150 .
this is the reason twist expansion into fock states in powers of @xmath151 applies to the pion bound state .
if this assumption is valid , pqcd should be also applicable to the @xmath19 form factors . besides
, large radiative correction to the @xmath3 meson vertex , which reaches 35% of the full contribution , or about half of the soft ( zeroth - order ) contribution , has been noticed .
this @xmath152 correction renders the sum rule for @xmath153 , with @xmath154 being the @xmath3 meson decay constant , quite unstable relative to the variation of input parameters @xcite . to stabilize the sum rule ,
one considers another sum rule for @xmath154 at the same time , which also receives large radiative correction to the @xmath3 meson vertex .
the two large vertex corrections then cancel in the ratio @xmath155 . however , the radiative correction to @xmath154 is then large . a careful look at the light - cone - sum - rule analyses indicates that the soft contribution is more sensitive to the end - point ( @xmath17 ) behavior of the pion distribution amplitude than the @xmath152 correction @xcite .
hence , if the end - point behavior of the pion distribution amplitude is modified by the sudakov factor in this work , such that the end - point contribution is not important , perturbative contribution can become dominant .
the sudakov effect on the soft contribution to @xmath156 has been investigated in the qcd sum rule formalism @xcite ( without twist expansion for the pion bound state ) . in this analysis ,
the soft contribution without sudakov suppression was estimated to be between 0.15 ( corresponding to @xmath157 mev ) and 0.22 ( corresponding to @xmath158 mev ) .
the soft contribution to @xmath153 obtained in @xcite is consistent with the above range .
it was then shown that the sudakov effect decreases the soft contribution by a factor 0.4 - 0.7 , depending on infrared cut - offs for loop corrections to the weak decay vertex .
therefore , the soft contribution turns out to be about 0.06 - 0.15 . compared with the lattice results @xmath159 , it is reasonable to conclude that the soft contribution amounts to about 30% , which is consistent with the observation made in @xcite .
it is a fair opinion that the estimate of soft contribution is more model - dependent than perturbative one .
for example , perturbative contribution is less sensitive to the pion distribution amplitude or to other input parameters such as the borel mass in light - cone sum rules @xcite . in the pqcd approach
we calculate the perturbative contribution to @xmath103 , which is more model - independent , and show that the result can more or less saturate the value predicted by lattice technique . for the @xmath3 meson distribution amplitude ,
we adopt the model @xmath160\ ; , \label{os}\end{aligned}\ ] ] with the shape parameter @xmath161 gev @xcite .
the normalization constant @xmath162 is related to the decay constant @xmath163 mev through the relation @xmath164 it is easy to find that eq .
( [ os ] ) has a maximum at @xmath165 .
we employ the models for the pion @xcite , @xmath166\ ; , \label{pioa}\\ \phi_\pi^p(x)&=&\frac{f_\pi}{2\sqrt{2n_c } } \left[1 + 0.43c_2^{1/2}(2x-1)+0.09c_4^{1/2}(2x-1)\right]\ ; , \label{piob}\\ \phi_\pi^t(x)&=&\frac{f_\pi}{2\sqrt{2n_c } } ( 1 - 2x ) \left[1 + 0.55(10x^2 - 10x+1)\right]\ ; , \label{pioc}\end{aligned}\ ] ] with the pion decay constant @xmath167 mev .
the gegenbauer polynomials are defined by @xmath168 whose coefficients correspond to @xmath169 gev .
we first investigate the relative importance of the twist-2 and twist-3 contributions to @xmath170 , and the results are listed in table [ tbl - twist ] .
it is observed that the latter are in fact larger than the former , consistent with the argument that the twist-3 contributions are not power - suppressed .
the light - cone sum rules also give approximately equal weights to the twist-2 and higher - twist contributions to @xmath125 @xcite .
we then compare our results of @xmath170 and @xmath171 for @xmath172 gev@xmath173 with those derived from lattice qcd @xcite and from light - cone sum rules @xcite in fig .
[ bpi ] , where lattice results have been extrapolated to the small @xmath174 region .
different extrapolation methods cause uncertainty only of about 5% @xcite .
the good agreement among these different approaches at large recoil is explicit .
the fast rise of the pqcd results at slow recoil indicates that perturbative calculation becomes unreliable gradually .
the values of @xmath175 from pqcd for the parameter @xmath176 gev are listed in table [ bpitable ] .
the resultant range @xmath177 is in agreement with @xmath178 obtained in @xcite .
we shall adopt the same range of @xmath179 in the evaluation of the @xmath22 transition form factors below .
we also examine the uncertainty of our predictions from the parametrization of the jet function in eq .
( [ trs ] ) .
the values of @xmath170 vary about 15% for the choices of @xmath180 and @xmath181 as shown in table [ rsumtbl - pls ] .
the variation for @xmath171 is similar . in a future work
we shall incorporate the exact jet function into a convolution integrand in moment space .
.contributions to @xmath170 from the twist-2 and two - parton twist-3 pion distribution amplitudes .
[ cols="^,^,^,^,^,^,^,^,^,^,^,^",options="header " , ]
consider the semileptonic decay @xmath182 in the fast recoil region of the @xmath28 meson @xcite .
we define the @xmath3 meson momentum @xmath41 as in eq .
( [ pa ] ) , the momentum @xmath92 and the polarization vectors @xmath183 of the @xmath28 meson in light - cone coordinates as @xmath184 with the ratio @xmath185 and the energy fraction @xmath94 carried by the @xmath28 meson .
we first keep the @xmath186 dependence of the kinematic variables in eq .
( [ par ] ) , and extract the twist-3 terms proportional to @xmath187 .
the parametrization of @xmath92 and @xmath183 is chosen to make this extraction straightforward .
the @xmath22 form factors are defined through the following decompositions of hadronic matrix elements , @xmath188 \nonumber\\ & & -a_2(q^2)\frac{\epsilon^*\cdot q}{m_b+m_\rho } \left[p_1^\mu+p_2^\mu-\frac{m_b^2-m_\rho^2}{q^2}q^\mu\right]\;. \label{af}\end{aligned}\ ] ] to calculate the form factors @xmath189 , @xmath190 @xmath191 and @xmath192 , we adopt the following procedures . first , only the transverse polarization vectors @xmath193 are involved in eq .
( [ vf ] ) and associated with the definition of @xmath191 in eq .
( [ af ] ) , through which we evaluate the form factors @xmath189 and @xmath191 , respectively .
both the structures associated with @xmath191 and @xmath192 are orthogonal to the lepton pair momentum @xmath42 .
contracting eq .
( [ af ] ) with @xmath194 , we have @xmath195 which implies that only the form factor @xmath190 is relevant in two - body nonleptonic decays such as @xmath196 .
we calculate @xmath190 from eq .
( [ a0e ] ) using the distribution amplitudes associated with a longitudinally polarized @xmath28 meson . for the longitudinal polarization vector @xmath197 ,
the structures of @xmath191 and @xmath192 are in fact proportional to each other : @xmath198 = \frac{(\epsilon^*\cdot q)^2(m_b^2-m_\rho^2-q^2 ) } { ( m_b+m_\rho)[(\epsilon^*\cdot q)^2+q^2 ] } \left[\epsilon^{*\mu}-\frac{\epsilon^*\cdot q}{q^2}q^\mu\right]\;,\end{aligned}\ ] ] which can be easily derived via the relation , @xmath199 contracting eq .
( [ af ] ) with @xmath200 , we obtain @xmath201 \gamma_5 u(0)|b(p_1)\rangle \nonumber\\ & & = \frac{2p_2\cdot q}{m_b+m_\rho}\frac{(\epsilon^*\cdot q)^2}{q^2 } \left[a_2-\frac{(m_b+m_\rho)^2}{2p_2\cdot q } \left(1+\frac{q^2}{(\epsilon^*\cdot q)^2}\right)a_1\right]\ ; , \label{ea2}\end{aligned}\ ] ] from which the form factor @xmath192 can be computed .
it turns out that @xmath191 and @xmath192 have a simple relation , since the left - hand side of eq .
( [ ea2 ] ) is power - suppressed .
we derive the leading - power factorization formulas , @xmath202e(t^{(1)})h(x_1,x_2,b_1,b_2 ) \nonumber \\ & & + r_\rho\left[\phi_\rho^v(x_2)+\phi_\rho^a(x_2)\right ] e(t^{(2)})h(x_2,x_1,b_2,b_1)\biggr\}\ ; , \\
a_0&=&8\pi m_b^2 c_f\int dx_1dx_2\int b_1db_1 b_2db_2\phi_b(x_1,b_1 ) \nonumber\\ & & \times\biggl\{\left[(1+\eta x_2)\phi_\rho(x_2)+r_\rho \left((1 - 2x_2)\phi_\rho^t(x_2)+ \left(\frac{2}{\eta}-1 - 2x_2\right)\phi_\rho^s(x_2)\right)\right ] e(t^{(1)})h(x_1,x_2,b_1,b_2 ) \nonumber \\ & & + 2r_\rho\phi_\rho^s(x_2)e(t^{(2)})h(x_2,x_1,b_2,b_1)\biggr\}\ ; , \\
a_1&=&8\pi m_b^2 c_f\eta\int dx_1dx_2\int b_1db_1 b_2db_2\phi_b(x_1,b_1 ) \nonumber\\ & & \times\biggl\{\left[\phi_\rho^t(x_2)+r_\rho \left(\left(\frac{2}{\eta}+x_2\right)\phi_\rho^v(x_2 ) -x_2\phi_\rho^a(x_2)\right)\right]e(t^{(1)})h(x_1,x_2,b_1,b_2 ) \nonumber \\ & & + r_\rho\left[\phi_\rho^v(x_2)+\phi_\rho^a(x_2)\right ] e(t^{(2)})h(x_2,x_1,b_2,b_1)\biggr\}\ ; , \\ a_2&=&\frac{a_1}{\eta}\ ; , \label{a2f}\end{aligned}\ ] ] with the evolution factor @xmath203 the same as in eq .
( [ evol ] ) . taking the fast recoil limit with @xmath204 and assuming the asymptotic behavior @xmath205 , the above form factors are found to obey the symmetry relations @xcite , @xmath206 where the term @xmath207 , being higher - power
, does not appear in eq .
( [ a2f ] ) .
note that the form factors , treated as nonperturbative objects , are not calculated in @xcite .
instead , the diagrams we have calculated above are regarded as perturbative corrections to the relations in eq .
( [ sr ] ) .
we adopt the @xmath28 meson distribution amplitudes given in the appendix b @xcite , @xmath208\ ; , \label{pwr}\\ \phi_{\rho}^t(x)&=&\frac{f^t_{\rho}}{2\sqrt{2n_c } } \left\{3(2x-1)^2 + 0.3(2x-1)^2[5(2x-1)^2 - 3]\right
. \nonumber \\ & & \left.+0.21[3 - 30(2x-1)^2 + 35(2x-1)^4]\right\}\ ; , \label{pwt}\\ \phi_{\rho}^s(x ) & = & \frac{3f_\rho^t}{2\sqrt{2n_{c } } } ( 1 - 2x)\left[1 + 0.76(10x^2 - 10x+1)\right]\ ; , \label{pws}\\ \phi_\rho^t(x)&=&\frac{3f_\rho^t}{\sqrt{2n_c } } x(1-x)\left[1 + 0.2c_2^{3/2}(2x-1)\right]\ ; , \label{pwft}\\ \phi_{\rho}^v(x)&=&\frac{f_{\rho}}{2\sqrt{2n_c } } \bigg\{\frac{3}{4}[1+(2x-1)^2]+0.24[3(2x-1)^2 - 1 ] \nonumber \\ & & + 0.12[3 - 30(2x-1)^2 + 35(2x-1)^4]\bigg\}\ ; , \label{pwv}\\ \phi_{\rho}^a(x ) & = & \frac{3f_\rho}{4\sqrt{2n_{c } } } ( 1 - 2x)\left[1 + 0.93(10x^2 - 10x+1)\right]\ ; , \label{pwa}\end{aligned}\ ] ] with the decay constants @xmath209 mev and @xmath210 mev . the @xmath174 dependence of the form factors @xmath189 and @xmath211 with the same @xmath3 meson distribution amplitude in eq .
( [ os ] ) and @xmath212 gev employed , is displayed in fig .
our results are consistent with those from light - cone qcd sum rules @xcite at small @xmath174 .
it is found that the symmetry relation @xmath213 in eq .
( [ sr ] ) holds very well : @xmath191 is larger than @xmath189 only by 2% in the large recoil region , even after considering the pre - asymptotic forms of @xmath214 and @xmath215 in eqs .
( [ pwv ] ) and ( [ pwa ] ) , respectively . to compare our results with the second symmetry relation , we include next - to - leading power terms in eq .
( [ ea2 ] ) , obtaining @xmath216 e(t^{(1)})h(x_1,x_2,b_1,b_2 ) \nonumber \\ & & + 2r_\rho\phi_\rho^s(x_2 ) e(t^{(2)})h(x_2,x_1,b_2,b_1)\biggr\}\;.\end{aligned}\ ] ] because of the cancellation of the term @xmath217 and the second term in the above expression , the values of @xmath192 only slightly deviate from those in eq .
( [ a2f ] ) .
the numerical study shows that @xmath192 is larger than @xmath218 by about 40% , which can be regarded as the estimate of the symmetry breaking effect .
in this paper we have presented a complete leading - power and leading - order pqcd evaluation of the @xmath19 , @xmath28 transition form factors in the large recoil region .
it has been shown that under sudakov suppression arising from @xmath4 and threshold resummations , the end - point singularities ( logarithmic at twist 2 and linear at twist 3 ) do not exist . the soft contribution to the form factors ,
being sudakov suppressed , becomes smaller than the perturbative contribution .
the physical picture for the mechanism of sudakov suppression has been discussed .
we have emphasized that the twist-3 contributions are in fact not power - suppressed in the @xmath219 limit .
the treatment of the parton transverse momenta @xmath4 and the light spectator momentum @xmath48 in the @xmath3 meson in the computation of the hard amplitudes has been clearly explained : the hard amplitudes should not be expanded in powers of @xmath115 as the end - point region is important . using the light meson distribution amplitudes derived from qcd sum rules , and choosing an appropriate @xmath3 meson distribution amplitude ,
we have derived reasonable results for the @xmath19 , @xmath28 form factors , which are in agreement with those from light - cone qcd sum rules and from lattice calculations .
our study indicates that in a self - consistent perturbative analysis , the heavy - to - light form factors are calculable .
the jet function from threshold resummation needs more thorough exploration .
we shall investigate the relevant subjects , such as factorization theorem in moment space , threshold resummation up to next - to - leading logarithms , application to nonleptonic @xmath3 meson decays @xcite , and numerical effects elsewhere .
note that if considering only @xmath220 resummation @xcite , twist-3 contributions , though infrared finite , are still too large to give reasonable heavy - to - light transition form factors , because the large double logarithms @xmath20 have not yet been organized .
* acknowledgements * we thank s. brodsky and h.y .
cheng for helpful discussions .
the work was supported in part by grant - in aid for special project research ( physics of cp violation ) , by grant - in aid for scientific research from the ministry of education , science and culture of japan .
the work of h.n.l . was supported in part by the national science council of r.o.c . under the grant no .
nsc-89 - 2112-m-006 - 033 , and by theory group of slac .
the work of t.k was supported in part by grant - in aid for scientific research from the ministry of education , science and culture of japan under the grant no
. 11640265 .
it has been shown @xcite that the factorization in fermion flow between the pion distribution amplitude and the hard amplitude is achieved by inserting the fierz identity , @xmath221 into the quark and anti - quark lines of the pion , where @xmath222 represents the identity matrix , and @xmath223 is defined by @xmath224/2 $ ] .
the insertion of eq .
( [ fe ] ) then leads to various nonlocal matrix elements , @xmath225 each of which is characterized by different twists .
the light - like vector @xmath75 is the coordinate of the @xmath76 quark , and @xmath74 the pion momentum .
the general expressions of the relevant matrix elements are , quoted from @xcite , @xmath226 where @xmath227 and @xmath228 are the distribution amplitudes of unit normalization , @xmath229 the pion mass , @xmath16 the momentum fraction associated with the @xmath76 quark .
it is easy to observe that the contribution from @xmath230 , independent of the pion mass , is twist-2 , and the contribution from @xmath228 is twist-4 because of the factor @xmath231 .
the contributions from @xmath232 and @xmath233 , proportional to @xmath234 , are twist-3 .
we shall neglect the twist-4 terms and the term @xmath235 in eq .
( [ pt ] ) .
it is straightforward to read off the pseudo - vector and pseudo - scalar structures of the pion distribution amplitudes from eqs .
( [ pv ] ) and ( [ ps ] ) . to derive the pseudo - tensor structure from eq .
( [ pt ] ) , we need more effort . using integration by parts , eq .
( [ pt ] ) is rewritten as @xmath236 with the anti - symmetric tensor @xmath237 , @xmath238 .
the tensor @xmath237 in eq .
( [ pt3 ] ) contracts to the spin structure @xmath239 in the evaluation of the corresponding hard amplitude .
the factor @xmath240 comes from the extra factor @xmath240 associated with the pseudo - tensor structure compared to other structures in eq .
( [ fe ] ) .
we have @xmath241 therefore , up to twist-3 , the initial - state @xmath242 meson distribution amplitudes are written as @xmath243_{lj}\phi_\pi(x ) + [ \gamma_5]_{lj}m_0\phi^p_\pi(x)\right . \nonumber\\ & & \left.+m_0[\gamma_5(\not n_-\not n_+-1)]_{lj}\phi^t_\pi(x)\right\}\;,\end{aligned}\ ] ] with @xmath244 for the final - state @xmath242 meson , we consider the adjoints of eqs .
( [ pv ] ) , ( [ ps ] ) and ( [ pt ] ) : @xmath245 it is observed that the pseudo - tensor structure in eq .
( [ pto ] ) acquires an extra minus sign , compared to the other two structures .
the pseudo - tensor structure is then given by @xmath246 .
therefore , up to twist 3 , we have eq . ( [ fpd ] ) for the final - state @xmath242 meson .
note that there is an extra term in the definition of @xmath83 , which contains a differential operator applying to hard amplitudes @xcite .
this term , being power - suppressed , is negligible here .
the distribution amplitudes @xmath82 and @xmath84 are normalized according to @xmath247 the tensor distribution amplitude is normalized to zero , because of @xmath248 if @xmath233 vanishes at the end points of the momentum fraction .
we choose the @xmath28 meson momentum @xmath78 with @xmath249 , which is mainly in the plus direction .
the polarization vectors @xmath183 , satisfying @xmath250 , represent one longitudinal polarization vector @xmath197 and two transverse polarization vectors @xmath193 .
their explicit expressions in light - cone coordinates have been given in eq .
( [ par ] ) .
to arrive at the factorization in fermion flow , we insert the fierz identity into the quark and anti - quark lines of the @xmath28 meson . the spin structures in eq .
( [ fe ] ) lead to the following nonlocal matrix elements , @xmath251 characterized by different twists .
the definition of @xmath80 is the same as that for the pion distribution amplitudes in the previous appendix .
the general expressions of the above matrix elements are , quoted from @xcite , @xmath252\right\}\ ; , \label{v}\\ \langle \rho^-(p,\epsilon)|{\bar d}(z)\sigma_{\mu\nu}u(0)|0\rangle&= & -i\frac{f^t_{\rho}}{n_c}\left\{(\epsilon_{t\mu}p_{\nu}-\epsilon_{t\nu}p_{\mu } ) \int_0 ^ 1 dx e^{ix p\cdot z}\phi_t(x)\right .
\nonumber\\ & & + ( p_{\mu } z_\nu - p_{\nu } z_\mu)\frac{\epsilon\cdot z}{(p\cdot z)^2 } m^2_{\rho}\int_0 ^ 1 dx e^{ix p\cdot z}h^{(t)}_{\parallel}(x ) \nonumber\\ & & \left .
+ \frac{1}{2}(\epsilon_{t\mu}z_\nu-\epsilon_{t\nu}z_{\mu } ) \frac{m^2_{\rho}}{p\cdot z}\int_0 ^ 1 dx e^{ix p\cdot z}[h_3(x)-\phi_t(x ) ] \right\}\ ; , \label{t}\\ \langle \rho^-(p,\epsilon)|{\bar d}(z)i
u(0)|0\rangle&= & -\frac{i}{2n_c}\left(f^t_{\rho}-f_{\rho}\frac{m_u+m_d}{m_{\rho}}\right ) \epsilon\cdot zm^2_{\rho}\int_0 ^ 1 dx e^{ix p\cdot z } h^{(s)}_{\parallel}(x)\ ; , \nonumber\\ & = & \frac{1}{2n_c}\left(f^t_{\rho}-f_{\rho}\frac{m_u+m_d}{m_{\rho}}\right ) \frac{\epsilon\cdot z}{p\cdot z}m^2_{\rho } \int_0 ^ 1 dx e^{ix p\cdot z } \frac{d}{dx}h^{(s)}_{\parallel}(x)\ ; , \label{s}\\ \langle \rho^-(p,\epsilon)|{\bar d}(z)\gamma_5\gamma_\mu u(0)|0\rangle & = & -\frac{1}{4n_c}\left(f_{\rho}-f_{\rho}^t\frac{m_u+m_d}{m_{\rho}}\right ) m_\rho\epsilon_{\mu}^{~\nu\alpha\beta}\epsilon_{t\nu } p_{2\alpha } z_\beta\int_0 ^ 1 dx e^{ix p\cdot z}g^{(a)}_{t}(x)\ ; , \nonumber\\ & = & -\frac{i}{4n_c}\left(f_{\rho}-f_{\rho}^t\frac{m_u+m_d}{m_{\rho}}\right ) \frac{m_\rho}{p\cdot n_-}\epsilon_{\mu\nu\rho\sigma}\epsilon_t^\nu p^\rho n_-^\sigma\int_0 ^ 1 dx e^{ix p\cdot z } \frac{d}{dx}g^{(a)}_{t}(x)\ ; , \label{a}\end{aligned}\ ] ] where @xmath253 and @xmath254 are the decay constants of the @xmath28 meson with longitudinal and transverse polarizations , respectively , and @xmath16 the momentum fraction associated with the @xmath76 quark .
we adopt the convention @xmath255 for the levi - civita tensor @xmath256 . the distribution amplitudes @xmath227 , @xmath257 and @xmath258 are normalized to unity .
following the similar procedures , we derive the @xmath28 meson distribution amplitudes up to twist 3 , @xmath259_{lj}\phi_\rho(x)\right .
\nonumber\\ & & \left.+[\not \epsilon_l\not p]_{lj } \phi_{\rho}^{t}(x ) + m_\rho [ i]_{lj}\phi_\rho^s(x)\right\}\ ; , \label{lpf}\\ \langle\rho^-(p,\epsilon_t)|\bar d(z)_ju(0)_l|0\rangle & = & \frac{1}{\sqrt{2n_c}}\int_0 ^ 1 dx e^{ixp\cdot z } \bigg\{m_\rho[\not \epsilon_t]_{lj}\phi_\rho^v(x)+ [ \not\epsilon_t\not p]_{lj}\phi_\rho^t(x ) \nonumber\\ & & + \frac{m_\rho}{p\cdot n_- } i\epsilon_{\mu\nu\rho\sigma}[\gamma_5\gamma^\mu]_{lj}\epsilon_t^\nu p^\rho n_-^\sigma \phi_\rho^a(x)\bigg\}\ ; , \label{spf}\end{aligned}\ ] ] for longitudinal polarization and transverse polarization , respectively .
we have dropped the terms proportional to @xmath260 ( twist-4 ) and the terms @xmath261 in eqs .
( [ s ] ) and ( [ a ] ) .
the definitions of the above distribution amplitudes are , @xmath262
according to @xcite , the nonlocal matrix element associated with the @xmath3 meson is written as @xmath263 \right\}_{\delta\alpha}\ ; , \nonumber\\ & = & -\frac{i}{\sqrt{2n_c}}\left\{(\not p_1+m_b)\gamma_5 \left[\phi_b(k_1)-\frac{\not n_+-\not n_-}{\sqrt{2 } } { \bar \phi}_b(k_1)\right]\right\}_{\delta\alpha}\ ; , \label{bwp2}\end{aligned}\ ] ] with the wave functions , @xmath264 because the light meson momenta have been chosen in the plus direction , the hard amplitudes for the heavy - to - light transition form factors are independent of the component @xmath265 as explained in sec .
we construct the @xmath3 meson distribution amplitude , @xmath266 with @xmath267 .
the two @xmath3 meson distribution amplitudes @xmath268 and @xmath269 are related by the equation of motion @xcite , @xmath270 assuming that @xmath271 vanishes at both ends of the momentum fraction , @xmath17 and @xmath272 , we derive @xmath273 therefore , @xmath274 is normalized to zero .
we shall argue that the contribution from the distribution amplitude @xmath274 is negligible compared to that from @xmath275 .
consider the reasonable parametrizations , @xmath276\ ; , \nonumber\\ \bar\phi_b(x)&=&\frac{f_b}{2\sqrt{2n_c } } \left[-\frac{\bar\lambda}{2m_b}\delta'\left(x-\frac{\bar\lambda } { m_b}\right)+o\left(\frac{\bar\lambda^2}{m_b^2}\right)\right]\;,\end{aligned}\ ] ] whose moments satisfy eq .
( [ bep ] ) . as shown in sec .
iii , the hard amplitudes are approximated by @xmath277 at small @xmath58 .
a simple estimation indicates that the contribution from @xmath275 , proportional to @xmath278 , is numerically larger than that from @xmath274 , proportional to a constant .
hence , after taking into account eq .
( [ beq ] ) , we consider only a single @xmath3 meson distribution amplitude in this work .
in this appendix we supply details of the derivation of the sudakov factor in eq .
( [ trs ] ) .
threshold resummation introduces a jet function @xmath108 into the pqcd factorization of the @xmath19 form factors near the end points @xcite , @xmath279 where @xmath280 is an arbitrary real constant larger than all the real parts of poles involved in the integrand .
the factor @xmath281 comes from mellin transformation of the initial condition @xmath282 , @xmath283 the sudakov factor @xmath284 in the moment @xmath285 space has been derived explicitly to the accuracy of leading logarithms ( ll ) @xcite , @xmath286\;,\end{aligned}\ ] ] with the anomalous dimension @xmath287 . the contour integral in eq .
( [ mj ] ) leads to @xmath288 which vanishes at @xmath17 , since the integrand is an odd function in @xmath143 , and at @xmath272 due to the factor @xmath289 . in this paper
we consider threshold resummation up to next - to - leading logarithms . at this level of accuracy ,
the anomalous dimension @xmath290 contains two - loop contributions , and the coupling constant @xmath18 becomes running .
the sudakov factor @xmath284 is then given by @xmath291\ ; , \label{exa}\end{aligned}\ ] ] with @xmath292\ ; , \label{lk}\ ] ] @xmath293 being the number of quark flavors and @xmath294 a color factor . the anomalous dimension @xmath290 is the same as that for @xmath220 resummation @xcite .
it can be shown that @xmath108 still vanishes at the end points @xmath17 and @xmath272 . to simplify the analysis
, we propose the parametrization , @xmath111^c\ ; , \label{pst}\end{aligned}\ ] ] whose end - point behavior satisfies the above requirement . in the @xmath295 ( @xmath296 ) limit , _
i.e. _ , without qcd effects , eq
. ( [ pst ] ) approaches unity .
mellin transformation of @xmath108 gives @xmath297 the variable @xmath298 should be large enough to justify threshold resummation up to the next - to - leading logarithms @xmath299 , and small enough to avoid the divergent running coupling constant @xmath300 in eq .
( [ exa ] ) . performing the best fit of eq .
( [ trf ] ) to @xmath301 for @xmath302 , we determine the parameter @xmath303 . the difference @xmath304 is shown in fig .
[ thfit ] for @xmath305 and 0.4 .
equation ( [ pst ] ) implies that threshold resummation modifies the end - point behavior of the meson distribution amplitudes , rendering them vanish faster at @xmath17 .
cleo coll .
, y. kwon _
et al . _ , hep - ex/9908039 .
cleo collaboration , hep - ex/0001010 .
belle collaboration , talk presented at the 4th international workshop on @xmath3 physics and cp violation , ise , japan , feb .
babar collaboration , talk presented at the 4th international workshop on @xmath3 physics and cp violation , ise , japan , feb . , 2001 .
keum , h - n .
li , and a.i .
sanda , phys .
b * 504 * , 6 ( 2001 ) ; phys . rev .
* d63 * , 054008 ( 2001 ) .
l , k. ukai , and m.z .
yang , phys . rev . *
d63 * , 074009 ( 2001 ) ; c.d .
l and m.z .
yang , hep - ph/0011238 .
chen and h - n .
li , phys .
d * 63 * , 014003 ( 2001 ) .
n. isgur and c.h .
llewellyn - smith , nucl .
b317 * , 526 ( 1989 ) .
szczepaniak , e.m .
henley , and s.j .
brodsky , phys .
b * 243 * , 287 ( 1990 ) .
r. akhoury , g. sterman and y.p .
yao , phys .
d * 50 * , 358 ( 1994 ) .
m. beneke and t. feldmann , nucl . phys . *
b592 * , 3 ( 2000 ) .
a. khodjamirian and r. ruckl , phys .
d * 58 * , 054013 ( 1998 ) .
g. sterman , phys .
b * 179 * , 281 ( 1986 ) ; nucl .
b281 * , 310 ( 1987 ) .
s. catani and l. trentadue , nucl . phys . *
b327 * , 323 ( 1989 ) ; nucl . phys . * b353 * , 183 ( 1991 ) .
h - n . li , hep - ph/0102013 .
lapage and s.j .
brodsky , phys .
b * 87 * , 359 ( 1979 ) . g.p .
lepage and s.j .
brodsky , phys .
d * 22 * , 2157 ( 1980 ) .
a.v . efremov and a.v .
radyushkin , phys .
b * 94 * , 245 ( 1980 ) 245 .
li and g. sterman , nucl . phys . *
b381 * , 129 ( 1992 ) .
collins and d.e .
soper , nucl .
b193 * , 381 ( 1981 ) .
j. botts and g. sterman , nucl .
* b225 * , 62 ( 1989 ) .
li and h.l .
yu , phys .
lett . * 74 * , 4388 ( 1995 ) ; phys .
d * 53 * , 2480 ( 1996 ) .
t.w . yeh and h - n .
li , phys .
d * 56 * , 1615 ( 1997 ) .
li , hep - ph/0012140 , to appear in phys .
d. s. aoki , talk presented at the international conference on lepton - photon interaction at high energies , 1999 . p. ball j. high energy phys . * 09 * , 005 ( 1998 ) ; z. phys .
c * 29 * , 637 ( 1985 ) .
m. dahm , r. jakob , and p. kroll , z. phys .
c * 68 * , 595 ( 1995 ) .
a. khodjamirian _
_ , phys . lett .
b * 410 * , 275 ( 1997 ) .
a.p . szczepaniak and a.v .
radyushkin , phys .
d * 59 * , 034017 ( 1999 ) .
braun and i.e. filyanov , z phys .
c * 48 * , 239 ( 1990 ) ; p. ball , j. high energy phys . * 01 * , 010 ( 1999 ) .
et al . _ ,
b * 486 * , 111 ( 2000 ) .
d. becirevic , nucl .
. suppl . * 94 * , 337 ( 2001 ) .
h - n . li and h.l .
yu , phys .
b * 353 * , 301 ( 1995 ) .
j. charles _ et al .
d * 60 * , 014001 ( 1999 ) .
p. ball , v.m .
braun , y. koike , and k. tanaka , nucl .
* b529 * , 323 ( 1998 ) .
p. ball and v.m .
braun , phys .
d * 58 * , 094016 ( 1998 ) .
chen , y.y .
keum , and h - n .
li , in preparation .
d. du , d. yang , and g. zhu , hep - ph/0102077 ; hep - ph/0103211 .
a.g . grozin and m. neubert , phys .
d * 55 * , 272 ( 1997 ) .
li , phys .
b * 454 * , 328 ( 1999 ) ; chin .
* 37 * , 569 ( 1999 ) . | we calculate the @xmath0 transition form factors in the framework of perturbative qcd to leading power of @xmath1 , @xmath2 being the @xmath3 meson mass .
we explain the basic principle by discussing the pion electromagnetic form factor .
it is shown that the logarithmic and linear singularities occurring at small momentum fractions of light meson distribution amplitudes do not exist in a self - consistent perturbative analysis , which includes @xmath4 and threshold resummations .
epsfig.sty # 1[eq:#1 ] # 1#2#3 _ phys .
rev .
_ * d#1 * # 2 ( # 3 ) # 1#2#3 _ phys .
rev .
lett . _ * # 1 * # 2 ( # 3 ) # 1#2#3 _ phys .
lett . _ * # 1 * # 2 ( # 3 ) = -10 mm = -6 mm = -6 mm = 10000 * ncku - hep-01 - 04 + dpnu-00 - 13 * 1.cm 0.8 cm * .3 cm t. kurimoto@xmath5 , hsiang - nan li@xmath6 and a.i . sanda@xmath7
* * _ @xmath8 department of physics , toyama university , toyama 930 - 8555 , japan _ * * _ @xmath6 department of physics , national cheng - kung university , + tainan , taiwan 701 , republic of china _ * * _ @xmath6 national center for theoretical sciences , hsinchu , taiwan 300 , republic of china _ * * _ @xmath9 department of physics , nagoya university , nagoya 464 - 8602 , japan _ * 2.0 cm * pacs index : 13.25.hw , 11.10.hi , 12.38.bx , 13.25.ft * |
long - range interacting systems have attracted much attention because of a wide variety of physical systems and peculiarities of its dynamics and thermodynamics @xcite .
recent experimental achievement allows us to access artificial quantum lattice systems where the coupling parameters can be controlled .
dipolar gases in optical traps are suggested to be possible experimental realization of long - range interacting lattice systems .
moreover , it has been suggested @xcite that we can realize artificial gravitational systems in laboratory by irradiating off - resonant laser beam onto atoms in a bose - einstein condensate .
theoretically , long - range interactions exhibit rich equilibrium and non - equilibrium behavior like the ensemble inequivalence , negative specific heat , and the absence of thermalization .
these properties are atypical for short - range interacting systems .
therefore , studying long - range interacting systems will lead us to deepen our fundamental understanding of statistical mechanics .
those peculiar features of long - range interacting systems have been understood by using mean - field models @xcite . in mean - field models ,
the interaction between particles or spins is assumed to be independent of distance .
since exact results are available in mean - field models , they are important theoretically although they appear to be too simplified . thus exploring the condition on the realization of predictions by mean - field models should be actively proceeded .
kastner @xcite examined quantum spin systems with an infinite - range ferromagnetic interaction .
he demonstrated the ensemble inequivalence of these systems . in this paper ,
more general long - range interactions are treated ; ferromagnetic interactions whose range is comparable with the system size . an important class of these interactions is the power - law potential @xmath0 , where @xmath1 is the spatial distance between spins and @xmath2 satisfies @xmath3 ( @xmath4 is the spatial dimension ) .
it has been believed from the analysis of exact partition functions or the numerical calculations of some models that thermodynamic behavior does not depend on @xmath2 as long as it is less than @xmath4 @xcite .
as @xmath5 corresponds to the mean - field model , it implies that the results of the mean - field model might be exact for general long - range interacting systems as long as @xmath6 , and @xmath3 . in the previous papers @xcite
, it was shown that for classical spin systems , the free energy does not depend on @xmath2 in the canonical ensemble with non - conserved order parameters , but @xmath2 becomes relevant in some parameter region when there is a constraint of the value of the energy or the magnetization . a sufficient and a necessary condition for @xmath2 to be relevant
are derived in the previous works . in this paper , we show that it is also true in quantum spin systems .
an immediate conclusion is that if the inequivalence of the canonical and the microcanonical ensembles predicted by the analysis of the mean - field model can be observed in experiments , the violation of exactness of the mean - field theory will be also observed in the microcanonical ensemble .
inhomogeneous spin configurations in equilibrium are its sign .
although it is known that the mean - field models reproduce many properties of general long - range interacting systems even quantitatively , the infinite - range ( or curie - weiss type ) interaction does not become an adequate idealization of actual long - range interactions in such a situation . this paper is organized as follows . in sec .
[ sec : model ] , we introduce the model . in sec .
[ sec : cg ] , an important notion of coarse - graining is explained . in sec .
[ sec : classical ] , we express the free energy only by classical variables . in sec . [
sec : variational ] , the variational expression of the free energy is derived . in sec .
[ sec : exactness ] , we mention some results deduced from the previous sections , namely that the free energy is independent of the precise form of the interaction potential as long as there is no constraint such as fixed value of the energy or the magnetization . on the other hand , the free energy under the constraint of the fixed value of the magnetization and the microcanonical entropy may depend on the interaction potential .
this result is the same as in classical spin systems with long - range interactions . in sec .
[ sec : ex ] , we investigate the spin-1/2 xxz model in the canonical and the microcanonical ensemble as an illustrative example . in sec .
[ sec : summary ] , we summarize our results .
we study systems expressed by the following hamiltonian on @xmath4-dimensional regular lattice : h=-_ij^n^d(_ij)_a = x , y , z_as_i^as_j^a-_i^n_i .
[ eq : h ] we focus on the ferromagnetic interaction , @xmath7 and @xmath8 .
the parameter @xmath9 corresponds to the inverse of the interaction range and we take the limit of @xmath10 appropriately ( see below ) .
@xmath11 is the interaction potential and we will give the condition of @xmath12 later . @xmath13 is the spin operator at site @xmath14 whose length @xmath15 is independent of particular site @xmath14 and @xmath16 .
normalizations @xmath17 for the interaction strength and @xmath18 for the magnetic field @xmath19 are not essential . with these normalizations , the system is reduced to a classical continuous spin system in the limit of @xmath20 .
the lattice interval is set to be unity and @xmath21 denotes the length of the system ( the number of spins is @xmath22 ) .
the position of site @xmath14 is denoted by @xmath23^d$ ] .
the distance between two sites @xmath14 and @xmath24 is denoted by @xmath25 .
we take the two limits ; @xmath26 and @xmath10 . in the present paper
, we consider either of the following limiting procedures : * * nonadditive limit * : the condition for @xmath27 is that @xmath28 and @xmath29^d}\phi(\bm{r})d^dr=1 $ ] .
we take the limit @xmath10 , @xmath26 with @xmath30 fixed .
note that it includes power - law potentials @xmath31 , @xmath3 . * * van der waals limit * : the potential @xmath27 satisfies @xmath28 and @xmath32 .
we take the van der waals limit , that is @xmath10 after @xmath26 .
the case ( i ) corresponds to nonadditive interactions , namely the interaction range is comparable with the system size . on the other hand , in case ( ii ) ,
the interaction range is much longer than the lattice interval but much shorter than the system size @xmath21 .
hence , two limiting procedures treat different situations .
lebowitz and penrose @xcite proved that in the van der waals limit , the equation of state of the system is equal to the mean - field equation of state ( van der waals like equation ) with the so called maxwell s equal area rule ( or referred to be as the maxwell construction ) . in the case of
( ii ) , therefore , the free energy is independent of the precise form of the interaction potential @xmath11 .
extension to quantum many particle systems was done by lieb @xcite . the perturbative analysis around @xmath33
was suggested to understand the properties of a system with long but finite range interaction ( @xmath34 ) @xcite . even in the critical region ,
some exact results were obtained along this line @xcite .
on the other hand , the case of ( i ) covers much _ longer _ interactions than the van der waals limit .
similarly to the case of ( ii ) , it has been indicated that the free energy does not depend on the interaction potential , and it is equivalent to the mean - field free energy _ without _ the maxwell construction in the nonadditive limit .
it is called the exactness of the mean - field theory @xcite . therefore , we can regard the exactness of the mean - field theory in the nonadditive limit is a generalization of the lebowitz - penrose theorem in the van der waals limit .
actually , it is correct in classical spin systems unless there are some constraints such as fixed energy or fixed magnetization @xcite , as mentioned in the introduction .
in this section , we introduce a kind of coarse graining and show that the equilibrium free energy is exactly given by the free energy of the coarse - grained model .
we divide the system into many cells which are large enough to contain many spins , but small compared to the interaction range .
the length of each cell is denoted by @xmath35 and cells are labeled by @xmath36 , @xmath37 .
the resultant spin of a cell @xmath36 is denoted by @xmath38 .
eigenvalues of @xmath39 and @xmath40 are denoted by @xmath41 and @xmath42 , respectively ( we set @xmath43 to be unity ) . in each cell @xmath36 , there are @xmath44 spins , so they are properly labeled by @xmath45 .
we define @xmath46-partial resultant spin as _
p^(n)_i=1^n_p(i ) , and eigenvalues of @xmath47 are denoted by @xmath48 . then the full hilbert space is spanned by the tensor product of eigenstates @xmath49 of each cell , @xmath50 . here
@xmath51 is the `` trajectory '' of @xmath52 , @xmath53 .
it is explained later in more detail .
when the interaction is long - range , it is expected that the hamiltonian can be expressed only by the resultant spin operators @xmath54 approximately .
moreover , typically @xmath55 is considered to be very large when @xmath56 , so it is naturally expected that all the @xmath57 , @xmath37 , can be regarded as classical spins whose length is @xmath55 .
we shall construct the above argument rigorously .
the coarse - grained hamiltonian is introduced by = -_p , q^(l / l)^d^d_pq_a = x , y , zs_p^as_q^a-_p^(l / l)^d_p , [ eq : cg_h ] where _
pq_i_p_j_q(_ij ) .
the above coarse - grained hamiltonian is expressed only by the resultant spin operators of each cell .
the original hamiltonian is well approximated by this coarse - grained one . in order to see it , consider the operator g(h- ) .
we define the norm of the operator @xmath58 by a _ , |=1||a|| .
the norm of @xmath59 then is given by @xmath60 here @xmath61 , therefore g _ p , q^(l
/ l)^d_i_p_j_q|(_ij)-_pq|g_cl . in ref .
@xcite , it is shown that @xmath62 for the nonadditive limit and @xmath63 for the van der waals limit . therefore , the original hamiltonian is well approximated by the coarse - grained one . for convenience ,
we use the same notation `` @xmath64 '' for the nonadditive limit and the van der waals limit .
the above property then implies g=0 .
[ eq : lim_g ] by using this property , we shall show that the free energy per spin is exactly the same as that given by the coarse - grained hamiltonian .
the free energy is defined by f(,)=-e^-h . here
@xmath65 is the inverse temperature .
the coarse - grained free energy is given by ( , ) = -e^- .
from now on , we show that @xmath66 .
we use the bogoliubov - peierls inequality , e^-h_ie^- i|h|i , [ eq : peierls ] where @xmath67 is arbitrary orthonormal set . from @xmath68 and eq .
( [ eq : peierls ] ) , we obtain e^-h_ie^- i||ie^-l^d i|g|i .
[ eq : bp_bound ] by choosing the eigenstates of @xmath69 as @xmath67 , we have e^-h_ie^-_ie^-l^d i|g|ie^-l^d ge^- , where @xmath70 are eigenvalues of @xmath69 .
on the other hand , the upper bound of the partition function @xmath71 is obtained by using the golden - thompson inequality @xcite , e^-h = e^--l^dge^-e^-l^dg .
we have @xmath72 from eq .
( [ eq : bp_bound ] ) and eq .
( [ eq : gt_bound ] ) , we obtain e^-l^d ge^l^d g , and thus |f-| g. from eq .
( [ eq : lim_g ] ) , @xmath66 .
hence we can safely replace @xmath73 by @xmath69 to compute the free energy .
the coarse - grained hamiltonian depends only on the resultant spins @xmath74 .
since there are @xmath44 spins inside a cell and finally we take the limit of @xmath75 , it is expected that the typical length of resultant spins is huge .
we are then able to replace the quantum mechanical operator @xmath57 by the classical vector . in this section
, we develop a theory based on the above argument rigorously .
we consider the hilbert space of a cell @xmath36 .
the spins in this cell is labeled by @xmath76 , @xmath77 . as is mentioned in sec .
[ sec : cg ] , the hilbert space of the cell @xmath36 is spanned by the complete orthonormal basis @xmath78 .
we notice that when each spin @xmath13 is a spin-@xmath15 operator , @xmath79 the `` trajectory '' with length @xmath44 is defined by @xmath80 where @xmath81 satisfies eq .
( [ eq : process ] ) .
we will omit @xmath82 when there is no confusion .
the set of trajectories of all @xmath83 with a fixed @xmath55 is denoted by @xmath84 .
moreover , we define the _ weight function _ @xmath85 by w(j_p ) . explicit calculation of the weight function is given in appendix [ sec : w ] .
the identity operator in the hilbert space of a cell @xmath36 is expressed by 1_p=_j_p_m_p =- j_p^j_p__p^(l^d)_l^d(j_p)|j_p , m_p,_p^(l^d ) j_p , m_p,_p^(l^d)| . because the coarse - grained hamiltonian depends only on the redundant spins @xmath57 , the partition function of the coarse - grained hamiltonian
is written as @xmath86 e^{-\beta\tilde{h}}\left[\bigotimes_{p=1}^{(l / l)^d}|j_p , m_p,\gamma_p^{(l^d)}\>\right ] \nonumber \\ & = \prod_{p=1}^{(l / l)^d}\left\{\sum_{j_p}\sum_{m_p =- j_p}^{j_p}w(j_p)\right\ } \nonumber
\\ \times&\left[\bigotimes_{p=1}^{(l / l)^d}\
< j_p , m_p|\right]e^{-\beta\tilde{h}}\left[\bigotimes_{p=1}^{(l / l)^d}|j_p , m_p\>\right ] \nonumber \\ & \equiv \prod_{p=1}^{(l / l)^d}\left\{\sum_{j_p}w(j_p)\right\}\tr_{\{j_p\}}e^{-\beta\tilde{h}}. \label{eq : partition}\end{aligned}\ ] ] here @xmath87 is an arbitrary state vector such that @xmath88 for instance , we choose @xmath87 as @xmath89 means the trace over the subspace of fixed values of @xmath90 , _ j_pa_m_p =- j_p^j_pj_p , m_p|a|j_p , m_p .
lieb @xcite proved that for the general hamiltonian such as eq .
( [ eq : cg_h ] ) , if @xmath91 , z^cl(\{j_p})z^cl(\{j_p+1 } ) , [ eq : lieb1973 ] where @xmath92 is the _ classical _ partition function with spins @xmath90 , namely , @xmath93.\end{aligned}\ ] ] here , the classical hamiltonian is defined by the same form of eq .
( [ eq : cg_h ] ) but the spin operators @xmath57 are replaced by the classical vectors , @xmath94 lieb proved the inequality ( [ eq : lieb1973 ] ) by using a spin coherent state representation . from eqs .
( [ eq : partition ] ) and ( [ eq : lieb1973 ] ) , we obtain the classical bounds z_-zz_+ , where @xmath95 \nonumber \\ & \times\exp\left[-\beta\tilde{h}^{\rm cl}(\{j_p,\theta_p,\phi_p\})\right ] , \\ z_+\equiv&\sum_{\{j_p\}}{\cal w}(\{j_p\})\int d\omega\left[\prod_p(2j_p+1)\right ] \nonumber \\ & \times\exp\left[-\beta\tilde{h}^{\rm cl}(\{j_p+1,\theta_p,\phi_p\})\right].\end{aligned}\ ] ] here @xmath96 is defined by @xmath97 , and @xmath98 .
in this section , we evaluate the upper and the lower bounds of the free energy and derive its variational expression .
let us define _ \{_p,_p}^cl(\ { j_p,_p,_p})e^*(\ { j_p } ) .
we then find z_+_\ { j_p}w(\ { j_p})e^-e^*(\ { j_p+1 } ) . because @xmath99 for an arbitrary function @xmath100 , @xmath101\bigg\}.\end{aligned}\ ] ] therefore , we obtain @xmath102 .
\label{eq : lower1}\end{aligned}\ ] ] we show that we can replace @xmath103 by @xmath55 in eq .
( [ eq : lower1 ] ) .
it is verified by the following evaluation : @xmath104 where @xmath105 . as @xmath106
, we have @xmath107 here we note that _ pq^(l
/ l)^d^d_pq=_ij^d(_ij)= , and obtain @xmath108 thus we obtain the lower bound f_\{_p}. [ eq : f_lower ] next we estimate the upper bound .
we define ( _ p^*,_p^*)_\{_p,_p}^cl(\ { j_p,_p,_p } ) , @xmath109 , and @xmath110 .
in addition , the set @xmath111 is defined by _
p()\{(_p,_p):_p}. for sufficiently small @xmath112 , the volume of the set @xmath111 satisfies _ ( _ p,_p)_p()d_p=2(1- )
[ eq : area_omega_p ] moreover , we can show that _
( _ p,_p)_p()^cl(\ { j_p,_p,_p})-e^*(\ { j_p } ) l^d_max , [ eq : delta ] whose proof is given in appendix [ sec : delta ] . from eqs .
( [ eq : area_omega_p ] ) and ( [ eq : delta ] ) , @xmath113 \nonumber \\ \geq&\sum_{\ { j_p\}}{\cal w}(\ { j_p\})\frac{1}{(4\pi)^{(l / l)^d } } \nonumber \\ & \times\exp\left[-\beta\sup_{\{(\theta_p,\phi_p)\in\omega_p(\delta)\}}\tilde{h}^{\rm cl}(\ { j_p,\theta_p,\phi_p\})\right ] \nonumber \\ & \times\left[\int_{(\theta_p,\phi_p)\in\omega_p(\delta)}d\omega_p\right]^{(l / l)^d } \nonumber \\ \geq&\sum_{\ { j_p\}}{\cal w}(\ { j_p\})\left(\frac{\delta^2}{4\pi}\right)^{(l / l)^d } \nonumber \\ & \times\exp\left[-\beta e^*(\ { j_p\})-\beta l^d\lambda_{\rm max}\delta\right ] \nonumber \\ \geq&\max_{\ { j_p\}}\exp\bigg[-\beta e^*(\ { j_p\})+\ln { \cal w}(\ { j_p\ } ) \nonumber \\ & + \left(\frac{l}{l}\right)^d\ln\frac{\delta^2}{4\pi } -\beta l^d\lambda_{\rm max}\delta\bigg].\end{aligned}\ ] ] since @xmath114 is arbitrary , we choose @xmath115 .
we then find f_\{_p}. [ eq : f_upper ] from the upper bound ( [ eq : f_upper ] ) and the lower bound ( [ eq : f_lower ] ) , we obtain the desired variational expression for the free energy , f = lim_\{_p}. [ eq : f_variational ] the reason why we can reach the variational expression ( [ eq : f_variational ] ) is that the degrees of freedom to be considered is dramatically decreased by the coarse graining
. we can express the free energy by the continuous functional .
if we see the system in the length scale so the size of the system is unity , @xmath116 \nonumber \\ \equiv\inf_{\vec{s}(\bm{x})}&{\cal f}[\vec{s}(\bm{x } ) ] .
\label{eq : f_continuum}\end{aligned}\ ] ] here we defined the scaled potential @xmath117 and the entropy function @xmath118 with @xmath119 and @xmath120 , where @xmath121 is the central position of the cell @xmath36 .
notice that the length of the spin vector does not exceed unity , @xmath122 .
the _ free energy functional _
@xmath123 $ ] consists of two parts , @xmath124={\cal u}[\vec{s}(\bm{x})]-\frac{1}{\beta}\int_{[0,1]^d}\sigma(\vec{s}(\bm{x}))d^dx,\ ] ] where @xmath125=-\frac{1}{2}\int_{[0,1)^d}d^dx\int_{[0,1)^d}d^dy u(\bm{x}-\bm{y } ) \nonumber \\
\times\sum_{a = x , y , z}\lambda_as^a(\bm{x})s^a(\bm{y } ) -\vec{h}\cdot\int_{[0,1)^d}\vec{s}(\bm{x})d^dx \label{eq : internal}\end{aligned}\ ] ] is the _ internal energy functional_. it corresponds to the energy of a given configuration @xmath126 .
@xmath127 corresponds to the entropy of a given configuration @xmath128 .
because many microscopic states correspond to a single coarse - grained configuration @xmath128 , this entropic term arises .
thus , the calculation of the free energy is reduced to the minimization problem of the free energy functional .
it is noted that if the uniform spin configuration , @xmath129 independent of @xmath130 , is assumed , the free energy is exactly equal to that of the mean - field model ( the model with @xmath131 ) and independent of the potential @xmath132 ; [ ( ) = ] = f_mf ( , , ) .
the explicit form of @xmath133 is given in eq .
( [ eq : mf_free ] ) .
this observation implies that we can replace the interaction potential @xmath132 by the simple infinite range potential @xmath134 _ as long as the magnetization profile is uniform in equilibrium_. while , the spin configuration must be inhomogeneous whenever this replacement is not allowed .
finally , i make a few remarks on the relation to previous works .
gates and penrose @xcite obtained the similar variational expression of the free energy . however , the free energy functional obtained in the present paper [ eq .
( [ eq : f_continuum ] ) ] differs from that in those works with respect to the length scale of the free energy functional . in the present work
, the position variable @xmath130 runs over @xmath4-dimensional unit cube , @xmath135^d$ ] , and the system size is set to be unity .
it enables us to treat the long - range interaction potential like @xmath136 , @xmath137 . on the other hand , in the free energy functional obtained by ref .
@xcite , the position variable @xmath130 runs over the whole @xmath4-dimensional space , @xmath138 . in that case
, we can not treat the long - range interaction like @xmath136 with @xmath137 because the integrated value of the potential is divergent . if we consider only the van der waals limit , however , the free energy functional in ref .
@xcite has an advantage that it has information on the mesoscopic length scale . in the expression ( [ eq : f_continuum ] ) , we can not obtain any information on the length scale @xmath139 , _
e.g. _ the form of the interface in an ordered state with phase separation . in the work by kiessling and percus @xcite
, the conceptually different but mathematically equivalent limiting procedure was introduced to study the liquid - vapor interface in a many particle classical system .
that is , they took the limit of infinitely many particles in a finite domain . as a result , they derived the similar variational expression as eq .
( [ eq : f_continuum ] ) on the thermodynamic potential .
it shows the similarity between bulk properties of a system with a nonadditive long - range interaction and local properties of a system with a finite - range interaction and with a very large number of particles in a bounded domain . in the works of refs
@xcite , from the variational expression ( [ eq : f_continuum ] ) , simple sufficient and necessary conditions for the exactness of the mean - field theory are derived in classical spin systems , which is presented in the next section .
since the calculation of the free energy is reduced to the minimization of the free energy functional expressed by the classical variables , we can apply the same argument as the case of classical spin systems . in this section ,
we briefly review the result of the variational expression ( [ eq : f_variational ] ) or equivalently , the continuum expression ( [ eq : f_continuum ] ) .
hereafter , we assume periodic boundary conditions and apply the nearest image convention to pair interactions .
we assume that the system is in contact with a thermal reservoir at a temperature @xmath140 and there is no conserved quantity . in this case , the free energy is independent of the interaction potential @xmath11 and it is equal to the mean - field free energy @xmath141 \nonumber \\ & \equiv\min_{\vec{m}}f_{\rm mf}(\beta,\vec{m},\vec{h } ) , \label{eq : mf_free}\end{aligned}\ ] ] where @xmath142 this result is called `` exactness of the mean - field theory '' in long - range interacting systems . it is convenient to consider the canonical ensemble with the fixed magnetization , although the magnetization @xmath143 is not a conserved quantity .
roughly speaking , the free energy f(,,)=_\{():=_[0,1)^d()d^dx}f [ ( ) ] is related to the probability @xmath144 of the states with the magnetization @xmath143 by @xmath145 $ ] , although the meaning of `` the states with the magnetization @xmath143 '' is ambiguous in quantum spin systems because of the quantum fluctuation .
this function is well - defined and indeed useful when we analyze the case where there is some conserved quantity like the energy in sec .
[ sec : micro ] or @xmath146 in the xxz heisenberg model discussed in sec .
[ sec : ex ] .
later , we write @xmath147 simply as @xmath148 because the dependence on @xmath19 is trivial . in the van der waals limit
, the scaled potential is the @xmath149 function , @xmath150 , and in this case we can show f(,)=f_mf^ * * ( , ) .
this corresponds to the lebowitz - penrose theorem . in the nonadditive limit , @xmath151
the fourier expansion is given by u()=_^du_e^2i .
if we define u__^d\0u _ , we obtain @xcite @xmath152 where @xmath153 , and moreover f(,)<f_mf ( , ) when the matrix _
ab(u_max , ) [ eq : l ] is non - positive . from the inequality ( [ eq : f_ineq ] ) , we find that @xmath154 in the parameter @xmath155 such that @xmath156 .
we define @xmath157 by .
[ eq : tilde_beta ] it is then concluded that @xmath154 for any @xmath143 when @xmath158 .
this @xmath157 plays important roles in the following sections . in this section
we consider the microcanonical ensemble without any other conserved quantities .
the case in which there is another conserved quantity like @xmath146 is treated in sec .
[ sec : ex ] .
the microcanonical entropy is obtained by s()=_\{()}[_[0,1)^d(())d^dx:=u [ ( ) ] ] , where the internal energy functional @xmath159 $ ] is defined by eq .
( [ eq : internal ] ) .
this expression is common to classical spin systems , so the same argument as classical spin systems @xcite can be applied
. we can easily verify s_mf()s()s_mf^ * * ( ) , [ eq : entropy_ineq ] where @xmath160 is the microcanonical entropy in the mean - field model , @xmath161 , and @xmath162 is the concave envelope of @xmath163 or the function obtained by performing the legendre transformation twice on @xmath163 . from the inequality ( [ eq : entropy_ineq ] ) , we can conclude that the mean - field theory is exact in the microcanonical ensemble , @xmath164 , if the canonical and the microcanonical ensembles are equivalent in the mean - field model because @xmath165 in this case . in the case
where the canonical and the microcanonical ensembles are inequivalent in the mean - field model , more sophisticated treatment is necessary . in this case , it was shown in @xcite that @xmath166\right\}. \label{eq : s_inequality2}\end{aligned}\ ] ] here @xmath157 is given by eq .
( [ eq : tilde_beta ] ) .
moreover , it was shown that s()>s_mf ( ) [ eq : s_instability ] when the matrix @xmath167 is nonpositive .
the definition of the matrix @xmath168 is given by eq .
( [ eq : l ] ) .
the quantity @xmath169 is the inverse temperature of the mean - field model at the energy @xmath170 , and @xmath171 is the equilibrium magnetization of the mean - field model , _ eq ( ) _ .
the result of the inequality ( [ eq : s_inequality2 ] ) is as follows .
we consider the case that @xmath172 in the region of @xmath173 .
we define ^*. if @xmath174 , then @xmath164 holds for all the values of @xmath170 . therefore , the states with the negative specific heats in the mean - field model are maintained for systems with general long - range interaction @xmath11 as long as the condition @xmath174 is satisfied . on the other hand , the inequality ( [ eq : s_instability ] )
means that when the matrix @xmath167 is non - positive , the equilibrium state in the mean - field model , @xmath175 , becomes unstable and the inhomogeneity appears . in this case , the mean - field theory is not exact .
as an application of the theoretical framework , we consider the spin-1/2 xxz model , whose hamiltonian is given by eq .
( [ eq : h ] ) with @xmath176 and @xmath177 .
the hamiltonian is given by h=-_ij^d(_ij ) .
[ eq : h_xxz ] the mean - field counterpart of this model [ @xmath178 in eq .
( [ eq : h_xxz ] ) ] was studied by m. kastner @xcite . when the magnetization is not fixed , the exactness of the mean - field theory always holds in the canonical ensemble , @xmath179 , which is a general result from sec .
[ sec : canonical ] .
moreover , in this model , the mean - field theory is exact even in the microcanonical ensemble , @xmath164 .
it is resulted from the fact that the microcanonical entropy of the mean - field model corresponding to eq .
( [ eq : h_xxz ] ) is a concave function of @xmath170 .
it implies that @xmath180 and , thus , @xmath164 from eq .
( [ eq : entropy_ineq ] ) .
if the system is isolated , the @xmath181 component of the magnetization @xmath146 is conserved . in this section ,
we consider the canonical and the microcanonical ensembles with a fixed value of the @xmath181 component of the magnetization @xmath146 . in this case , as we will see below , the result is nontrivial and the exactness of the mean - field theory is violated for certain parameter region . first , we investigate the model in the canonical ensemble . the free energy as a function of @xmath182 and @xmath146 is given by f(,m_z)=_\ { m_x , m_y}f(,)f(,_eq(,m_z ) ) .
of course , @xmath183 .
the analysis of the previous sections leads us to the following result . from the inequality ( [ eq : f_ineq ] ) , we obtain @xmath184 therefore , if @xmath158 , @xmath185 holds because @xmath186 .
moreover , if the matrix @xmath187 has negative eigenvalues , the mean - field model is not exact , @xmath188 . since the entropy function of the spin-1/2 system is given by eq .
( [ eq : entropy_spin1/2 ] ) in appendix [ sec : w ] , the mean - field free energy with a fixed @xmath143 is given by @xmath189 \nonumber \\
+ \frac{1}{\beta}\left(\frac{1+|\vec{m}|}{2}\ln\frac{1+|\vec{m}|}{2}+\frac{1-|\vec{m}|}{2}\ln\frac{1-|\vec{m}}{2}\right ) .
\label{eq : f_xxz}\end{aligned}\ ] ] in order to obtain the mean - field free energy as a function of @xmath146 , it is necessary to minimize @xmath190 for @xmath191 and @xmath192 . by doing that
, we find in the mean - field model \ { & ( _ eq)_x=(_eq)_y=0 m_z(_m_z ) , + & |_eq|=(_|_eq| ) m_z<(_m_z ) . .
by differentiating twice the mean - field free energy @xmath190 , we obtain the hessian matrix @xmath193 ( see eq .
( [ eq : l ] ) ) .
we can obtain @xmath157 , which plays important roles to judge whether the mean - field theory is exact at a given parameter @xmath194 , by analyzing the eigenvalues of @xmath193 .
actually , we obtain = \{_^-1,_z^-1}. [ eq : tilde_xxz ] see appendix [ sec : tilde_xxz ] for the proof of eq .
( [ eq : tilde_xxz ] ) . by the reason mentioned in later analysis on the microcanonical ensemble , @xmath185 when @xmath195 .
the result is non - trivial when @xmath196 . in this case ,
@xmath197 , and the mean - field model gives the exact free energy , @xmath185 at least for @xmath198 . by calculating the minimum eigenvalue of @xmath199 , we obtain the sufficient condition under which the exactness of the mean - field theory does _ not _ hold . in fig .
[ fig : xxz_canonical ] , we plot the two regions in @xmath200 plane .
one is the region which is given by @xmath198 ( black region ) . in this region ,
the mean - field theory is exact .
the other is the region where the exactness of the mean - field theory is violated ( meshed gray region ) , which is given by the condition that the minimum eigenvalue of @xmath199 is negative .
the parameter @xmath201 is set to be 0.3 .
this value corresponds to the case that @xmath202 in the two - dimensional lattice . in the white region in fig .
[ fig : xxz_canonical ] , we can not determine whether the mean - field theory is exact from our analysis .
however , in the white region , the equilibrium state predicted by the mean - field model is at least locally stable because all the eigenvalues of @xmath199 are positive .
[ cols="^,^,^ " , ]
in this paper , we analyzed quantum spin systems with long - range interactions .
we reduced the calculation of the free energy to the minimization of the free energy functional , which is expressed only by the classical variables .
we obtained some results from this variational expression of the free energy .
one of them is that the free energy is always identical to that of the mean - field model when there is no conserved quantity .
when there is a conserved quantity such as @xmath146 , the mean - field model is exact at least for @xmath203 . the key quantity @xmath201 is determined by the form of the interaction potential @xmath204 and the spatial dimension @xmath4 .
moreover , if the matrix @xmath187 has a negative eigenvalue , the state with the uniform magnetization profile predicted by the mean - field model is locally unstable , and inhomogeneity which is not predicted by the mean - field model arises .
these results are the same as those in classical spin systems with long - range interactions .
we analyzed the systems also in the microcanonical ensemble , and obtained the same results as those in classical spin systems .
when the magnetization is not conserved , the microcanonical entropy is identical to that of the mean - field model as long as the canonical and the microcanonical ensembles are equivalent in the mean - field model .
however , if these two ensembles are inequivalent in the mean - field model , the microcanonical entropy may depend on the interaction potential @xmath132 .
the obtained results in this case are as follows .
the mean - field model is exact if @xmath205 for all @xmath170 such that @xmath172 . on the other hand , when @xmath167 has a negative eigenvalue , the microcanonical entropy is not identical to that of the mean - field model because the uniform state predicted by the mean - field model is locally unstable .
the analogous result holds when the magnetization @xmath146 is conserved . as a concrete example
, we investigated the spin-1/2 xxz model with a long - range interaction . when the magnetization is not conserved , the result is trivial ( the mean - field theory is exact ) both for the canonical and the microcanonical ensembles . therefore , we studied the model with a conserved @xmath146 . as a result
, it is shown that the exactness of the mean - field theory is violated in a certain parameter region for both the canonical and the microcanonical ensembles . in cold atom experiments
, we can set up the system which is well isolated from the environment , but the @xmath181-component of the magnetization is not a conserved quantity in general .
such a system will be described by the microcanonical ensemble without restriction of the value of @xmath206 . in this ensemble ,
the spin-1/2 xxz model with long - range interactions is equivalent to the corresponding mean - field model , as mentioned above . if we seek to observe the violation of exactness of the mean - field theory in long - range interacting systems , it is necessary to consider another model . in order to observe the violation of the exactness of the mean - field theory in the microcanonical ensemble
, the mean - field model corresponding to the system must satisfy @xmath207 .
this condition means that the canonical and the microcanonical ensembles are inequivalent in the mean - field model .. it implies that the canonical ensemble with a fixed value of the magnetization is not equivalent to the canonical ensemble with a fixed value of the magnetic field @xmath208 . ] therefore , if the actual interaction is not infinite - range ( or curie - weiss type ) , not only the ensemble inequivalence predicted by analysis of the mean - field model , but also the violation of the exactness of the mean - field theory will be observed .
although it is known that the mean - field models reproduce many properties of general long - range interacting systems even quantitatively , the infinite - range ( or curie - weiss type ) interaction does not become an adequate idealization of actual long - range interactions in such a situation .
kastner pointed out that the inequivalence of the canonical and the microcanonical ensembles can be observed in a long - range interacting system undergoing a temperature - driven first order phase transition @xcite .
a typical example of such spin models which possibly exhibits the violation of the exactness of the mean - field theory is the spin-1 ising model with anisotropy , h=-_ij^d(_ij)s_i^zs_j^z+d_i(s_i^z)^2 .
[ eq : d_ising ] the corresponding mean - field model , which is obtained by replacing @xmath209 by @xmath134 , undergoes the temperature - driven first order phase transition for certain range of the parameter @xmath210 . however , this model is a special one in the sense that it is a classical spin system .
it is expected that more general quantum spin systems with the anisotropic term , h=-_ij^d(_ij)_a = x , y , z_as_i^as_j^a+d_i(s_i^z)^2 . behaves similarly , although it is not proven straightforwardly from the formalism given in this paper .
the presence of the nonlinear external field @xmath211 makes it difficult to treat by the formalism given in this paper .
a future work should treat nonlinear external fields such as @xmath211 , which will promote an experimental realization of characteristic properties of long - range interacting systems .
the author thanks seiji miyashita for valuable discussion and acknowledges jsps for financial support ( grant no .
we seek to calculate the weight function @xmath85 . from the definition
, @xmath85 is the total number of trajectories with length @xmath44 from 0 to @xmath55 .
we can obtain the expression of @xmath85 by counting such trajectories , but we can do it more easily .
after all , @xmath85 is the number of states with the total spin @xmath55 and the total magnetization @xmath212 .
the states with @xmath213 have @xmath214 or @xmath215 .
states with @xmath215 and @xmath213 are obtained by multiplying @xmath216 to states with @xmath217 .
therefore , if we refer to @xmath218 as the number of states with @xmath213 , the weight function is given by w(x)=w_m(x)-w_m(x+1 ) .
for instance , in the case of spin-1/2 , @xmath218 is given by w_m(x)=. therefore , @xmath219 applying the stirling formula , we obtain @xmath220 the weight function and the entropy function are calculated similarly for higher spins ( @xmath221 ) , though the calculation becomes complex . in the case of @xmath222 ,
the result is w_m(x)=_n=0^l^d - x , and @xmath223 where @xmath224
we derive the inequality ( [ eq : delta ] ) . for simplification ,
we write @xmath225 if we can express the unit vector @xmath226 by @xmath227 with @xmath228 . first
, we show that @xmath229 for arbitrary unit vectors @xmath230 .
we have as the direction of the vector @xmath231 .
any @xmath232 is represented as @xmath233 with @xmath234 , which yields @xmath235 .
therefore , from eq .
( [ eq : ab ] ) , we have |-||-_p|+|-_p|2 .
we now show eq .
( [ eq : delta ] ) . from the definition ,
we have @xmath236 \nonumber \\ & \leq\sup_{(\theta_p,\phi_p)\in\omega_p(\delta)}\frac{1}{2s^2}\sum_{p , q}^{(l / l)^d}\gamma^d\phi_{pq}j_pj_q\lambda_{\rm max } \nonumber \\ & \qquad\times|\vec{e}_p\cdot\vec{e}_q-\vec{e^*}_p\cdot\vec{e^*}_q|.\end{aligned}\ ] ] here we define @xmath237 and @xmath238 . because @xmath106 and @xmath239 we obtain the inequality ( [ eq : delta ] ) , @xmath240
in the model described by eq .
( [ eq : h_xxz ] ) , the mean - field free energy is given by eq .
( [ eq : f_xxz ] ) . the hessian matrix @xmath193 defined by eq .
( [ eq : l ] ) determines the convexity of the mean - field free energy .
if the minimum eigenvalue of @xmath193 is positive for all @xmath143 , the mean - field free energy is convex for @xmath143 and @xmath241 .
therefore , @xmath157 corresponds to @xmath182 such that the minimum eigenvalue of @xmath193 becomes zero at some @xmath143 .
next we consider the case of @xmath246 .
we rewrite @xmath247 as @xmath248 . \ ] ] from this expression , l_-(,)-_z+ .
hence , we have _ l_-(,)-_z+ .
[ eq : l-1 ] on the other hand , from eq .
( [ eq : l- ] ) , @xmath249 thus we have _ l_-(,)-_z+ .
[ eq : l-2 ] from eqs .
( [ eq : l-1 ] ) and ( [ eq : l-2 ] ) , we obtain _ l_-(,)=-_z+ . therefore , when @xmath246 , we have @xmath197 . | we study equilibrium states of quantum spin systems with nonadditive long - range interactions by adopting an appropriate scaling of the interaction strength , _
i.e. _ , the so called kac prescription . in classical spin systems
, it is known that the equilibrium free energy is obtained by minimizing the free energy functional over the coarse - grained magnetization . here
we show that it is also true for quantum spin systems . from this observation
, it is found that when the canonical ensemble and the microcanonical ensemble are not equivalent in some parameter region , it is not necessarily justified to replace the actual long - range interaction by the infinite - range interaction ( curie - weiss type interaction ) . on the other hand , in the parameter region where the two ensembles are equivalent ,
this replacement is always justified .
we examine the heisenberg xxz model as an illustrative example and discuss the relation to experiments . |
polarized pulse profiles are the key observations to understand the physical processes within pulsar magnetosphere .
various polarization features have been observed ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the linear polarizations are dominating for whole or some parts of profiles , reaching @xmath1 such as vela and psr b1737 - 30 @xcite .
the position angles ( pa ) of linear polarization often follow s - shaped curves , which has been well explained by the rotating vector model ( rvm ) @xcite .
some pulsars exhibit two orthogonally polarized modes in the pa curves , where a @xmath2 jump is accompanied by low linear polarization @xcite .
the circular polarizations of integrated profiles are usually weak ( less than @xmath3 ) and diverse @xcite .
there exist two main types of circular polarization signature : the antisymmetric type with a sign reversal in the mid - pulse and the symmetric type without sign changes over the whole profile @xcite . to understand these diverse polarization features , there have been a lot of theoretical works on pulsar emission processes and propagation effects in pulsar magnetospheres .
pulsar radio emission is generally believed to be coherent radiation from relativistic particles streaming along the open magnetic field lines in pulsar magnetosphere .
the coherence can be caused either by the stable charge bunches ( ` antenna ' mechanism , e.g. @xcite ) or by the instabilities ( ` maser ' mechanism ) including the curvature plasma instability @xcite , two - stream instability ( e.g. * ? ? ?
* ) , the beam - plasma instability induced by the curvature drift @xcite , etc .
various models have been constructed based on these two coherent manners ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . among these models ,
curvature radiation from charge bunches serves as one of the most probable mechanisms , whose coherent process , luminosity and spectrum have been investigated already ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
its polarization features are the most important base for theoretical considerations .
@xcite pointed out that the curvature radiation can lead to the sign reversal feature for circular polarization of the core component . by incorporating rotation effects ,
@xcite studied the phase lag between the centers of the pa curve and the intensity profile . @xcite considered the detailed geometry for curvature radiation and explained the correlation between the pa and the sign reversal of circular polarization .
recently , @xcite and @xcite independently considered the co - rotation effect and the emission geometry , and demonstrated that the circular polarization can be of a single sign or sign reversals depending on the density gradient of particles along the rotation phase .
in addition to the emission process , propagation effects may also have significant influences on pulsar polarizations .
the plasma within pulsar magnetosphere has two orthogonally polarized transverse modes , the ordinary mode ( o - mode ) and the extraordinary mode ( x - mode ) ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
when emission leaves the magnetosphere , it is coupled to these two modes . as the emitted waves propagate outward , they first experience the adiabatic walking " , i.e. , with their polarization following the direction of the local magnetic field @xcite . then
, with the change of plasma conditions , the wave mode coupling " occurs near the polarization limiting radius where the mode evolution becomes non - adiabatic , which will results in circular polarization @xcite .
as the waves propagate further to the far magnetosphere , the cyclotron absorption " takes place , which will lead to intensity absorption and net circular polarization if the distributions of electrons and positrons are asymmetric @xcite .
moreover , unlike the rectilinear propagation for the x - mode waves , the o - mode waves will undergo refraction additionally in the near regions of their generation @xcite , which makes the two components separated finally .
these physical processes within pulsar magnetosphere have been studied for the emission generation and propagation .
however , the joint researches for emission and propagation to explain observational facts are rare . as the first try ,
@xcite considered the emission of the relativistic particles subjected to parallel and perpendicular ( compare to local magnetic field ) accelerations , and demonstrated the influence of `` adiabatic walking '' on the ordering of the polarization vectors within the @xmath0 emission cones .
they found that the combination for emission in the inner magnetosphere and propagation effects in the outer magnetosphere can naturally explain the observed @xmath1 polarization and the sudden orthogonal mode transitions observed for the individual pulses .
however , they did not consider the rotational effect of the relativistic particles and the refractions for the o - mode waves .
the beamed emissions within the emission cone were improperly treated as coherent radiations . @xcite
depicted the most detailed pictures of curvature radiation , but propagation effects are not included yet . @xcite and @xcite studied the influences of the propagation effects on the mean profiles , but the intensity and polarization of the emitted waves were assumed rather than calculated from the emission mechanisms .
stimulated by these researches , we focus on the general picture for the combined effects of curvature radiation and propagation effects , and tries to tie the theoretical explanation with polarization observations . in this paper , we study the polarization features of the emission generated by curvature radiation and their modifications during wave propagation in pulsar magnetosphere through numerical calculations and simulations .
the paper is organized as follows . in section 2 , we present the theoretical grounds of curvature radiation and propagation effects for our calculations . polarized emission within the @xmath0 cone and the evolution of two separated natural modes in magnetosphere are studied in section 3 . in section 4 ,
the intensities of both modes within the entire pulsar beam are calculated for the density models of uniformity , cone , core and patches .
discussion and conclusions are given in section 5 .
is inclined by an angle of @xmath4 with respect to the rotation axis @xmath5 .
sight line @xmath6 has impact angle @xmath7 and makes an angle @xmath8 with respect to @xmath5 .
the @xmath9 coordinate system is fixed in space and used for the emission analysis , with @xmath10-axis along @xmath11 and @xmath6 in @xmath12-plane .
the rotation phase @xmath13 is the azimuthal angle of @xmath14 in @xmath9-frame .
the @xmath15 system is also a fixed frame and used for the wave propagation analysis , with the @xmath16-axis along @xmath6 , @xmath11 in @xmath17-plane , and @xmath18.,scaledwidth=48.0% ] in general , pulsar magnetosphere has an inclined dipole magnetic field , which rotates uniformly in free space , @xmath19 , \label{eq : staticb}\ ] ] here @xmath20 is the magnetic field on the neutron star surface , @xmath21 is the neutron star radius , @xmath22 is the unit vector along @xmath23 , and @xmath24 represents the unit vector of the magnetic dipole moment , as shown in fig .
[ fig : geometry ] . according to @xcite ,
relativistic particles are generated in the vacuum gaps above the polar caps of pulsar magnetosphere by the sparking process and streaming out along the open magnetic field lines . due to the bending of the field lines
, the relativistic particles will experience perpendicular acceleration and produce curvature radiation .
the coherent bunches of the relativistic particles are simply treated as huge point charges . for a relativistic particle traveling with velocity @xmath25 along an open field line , its radiation field @xmath26 in direction @xmath27 and the corresponding fourier components
@xmath28 have been depicted in detail by @xcite .
the energy radiated per unit frequency per unit solid angle reads @xcite @xmath29 .
\label{eq : energy}\end{aligned}\ ] ] here @xmath30 is the distance between the trajectory center and the observer , @xmath31 is the curvature radius for the particle trajectory , lorentz factor @xmath32 , @xmath33 is the angle between the emission direction @xmath27 and the particle velocity @xmath25 , @xmath34 , and @xmath35 are the modified bessel functions of the second kind .
immediately after the electromagnetic waves were generated through curvature radiation from the relativistic particles , they will be coupled to the local plasma modes to propagate outwards from the emission regions .
for the sake of simplicity , the plasma within the magnetosphere is assumed to be cold and moves with a single velocity @xmath36 .
the plasma is also assumed to have a density of @xmath37 , where @xmath38 is the multiplicity factor , @xmath39 represents the goldreich - julian density @xcite . in general
, there are four wave modes within the plasma of pulsar magnetosphere @xcite , two transverse and two longitudinal waves .
three out of them were commonly investigated ( e.g. * ? ? ?
* ) , which are the x - mode with refraction index @xmath40 , the sub - luminous o - mode with @xmath41 and the super - luminous o - mode wave with @xmath42 .
the subluminous o - mode ( or the alfven o - mode wave ) tends to follow the field lines and suffers serious landau damping @xcite , which will make this mode invisible .
only the x - mode and superluminous o - mode ( hereafter o - mode for simplicity ) can propagate outside the magnetosphere .
these two modes have different evolution behaviors for the trajectories and polarizations .
the x - mode waves are always the transverse ones and propagate rectilinearly .
it is convenient to analyze the propagation process in the fixed @xmath15-frame , as shown in fig .
[ fig : geometry ] . here , the @xmath16-axis is set along the wave vector @xmath43 , e.g. , the direction for the sight line @xmath44 .
let the parameters with subscript `` i '' denote the quantities at the emission point .
at an instant time @xmath45 , one photon has traveled over a distance of @xmath46 along the ray from the emission point .
the position vector @xmath23 and the pulsar rotation phase @xmath13 can be written as @xmath47 @xmath48 here , @xmath49 denotes the light cylinder radius .
the plasma conditions , such as the density @xmath50 , the magnetic field @xmath51 , and hence the dielectric tensor @xmath52 vary along the trajectory .
the o - mode wave evolves in a different manner .
refractive index of the o - mode wave is not so close to the unity as the x - mode , especially near the emission region with higher plasma density , which means that it will be deflected away from the magnetic axis .
when the o - mode wave propagates outside to the low density region , the refraction effect becomes weaker and weaker , and the trajectory gets more close to the straight line just as the x - mode @xcite .
the trajectory for the o - mode wave can be described by the hamilton equations . with the dispersion relation of the mode taken into account ,
the hamilton equations become @xcite , @xmath53 here , @xmath54 is the unit vector aligned with the magnetic field , @xmath55 represents the three dimensional refractive index , @xmath56 is the refractive index along the magnetic field , @xmath57 with the plasma frequency @xmath58 , @xmath59 , @xmath60 , and @xmath61 with @xmath62 . and refraction angle along the ray for the o - mode wave , for different observing frequencies @xmath63 in panels a ) and d ) , for different lorentz factors @xmath64 in ( b ) and ( e ) , for different multiplicities @xmath38 in ( c ) and ( f ) .
the refraction limiting radii , @xmath65 , are indicated by the arrows in panels ( d ) , ( e ) and ( f ) for the default curve ( solid line in each panel ) . the wave is emitted from a height @xmath66 at rotation phase @xmath67 .
the default parameters for the calculation are @xmath68 , @xmath69 and @xmath70 for a pulsar with @xmath71 , @xmath72 and @xmath73 .
, scaledwidth=48.0% ] by solving the hamilton equations in the 3-dimensional magnetosphere , we obtained the refractive index @xmath74 for the o - mode wave and the refraction angle ( the angle between @xmath27 and initial emission direction ) along the ray , as shown in fig .
[ fig : ref ] . near the emission region ,
the refractive index of the o - mode wave @xmath75 , and the wave is deflected away from the emission direction . when the photon propagates outwards , the refraction effect can be neglected when @xmath76 , and then the o - mode trajectory becomes rectilinearly , i.e. , with almost the fixed angle compared to the initial emission direction .
the transition happens at the refraction - limiting radius ( rlr ) , @xmath65 , where @xmath77 according to @xcite .
it is redefined numerically as @xmath78 , as indicated in fig.[fig : ref ] .
furthermore , high frequency emission is less refracted during its propagation , as shown in panels ( a ) and ( b ) .
in addition , if the plasma within pulsar magnetosphere is thin ( a small @xmath38 ) and streams with a relatively larger velocity ( a large lorentz factor @xmath64 ) , the o - mode wave will be less affected by the refraction process , as shown in fig.[fig : ref ] ( b ) , ( c ) , ( e ) and ( f ) . for the reasonable parameters chosen in fig.[fig : ref ] ,
the refraction - limiting radius are about 50200@xmath79 and the final refraction angles are @xmath80@xmath81 . in summary ,
our 3-d refraction calculation confirms the results from the traditional 2-d treatment , in which only the refraction within a fixed field line plane is considered ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . , between the emission points of the o - mode and x - mode waves for various @xmath38 and @xmath64 .
it is measured with respect to the wave lengths of the electromagnetic wave @xmath82 or the plasma wave @xmath83 .
the default parameters for the calculations are the same as those in fig .
[ fig : ref].,scaledwidth=40.0% ] as demonstrated above , if the low frequency o - mode wave ( small @xmath63 ) travels through the dense ( large @xmath38 ) and/or less relativistic ( small @xmath64 ) plasma within pulsar magnetosphere , they will experience sufficient refraction . on the contrary ,
the x - mode wave travels rectilinearly regardless of the plasma conditions .
hence , for a given sight line at a fixed rotation phase , the observed emissions of x - mode and o - mode originate from two discrete regions separated by @xmath84 at a given emission height
. as well known , pulsar radiations are coherent .
to get efficient coherency , it requires the scale of particle bunches to be much smaller than the wave length of the emitting waves .
[ fig : mode_split ] shows how the separation of the emission points for the two modes ( in unit of the curvature - radiation wave - length @xmath82 or the plasma oscillation wave length @xmath83 ) varies with plasma density @xmath38 and lorentz factor @xmath64
. obviously , the separation @xmath85 are much larger than the wave length @xmath82 , especially for larger @xmath38 and smaller @xmath64 .
moreover , the bunching are induced by the plasma oscillation with a wave length @xmath86 .
the separation @xmath84 is also significantly larger than @xmath83 .
therefore , the observed x - mode and o - mode waves at a fixed rotation phase are incoherent , which means the observed radiation should be the incoherent superposition of the two modes from two discrete emission regions . [ cols="^,^,^,^,^,^,^,^ " , ] using equation ( [ eq : intensity ] ) , we can calculate the polarization emissions from the whole open field line region without considering the co - rotation of the relativistic particles .
[ fig : beam_no - rot ] shows the calculated intensity and polarization beams for the four density models of the out - flowing particles in the form of uniformity , cone , core and patches . for the emission at a given point of pulsar beam , the x - mode and o - mode components ( @xmath87 and @xmath88 ) have a similar intensity due to the quasi symmetric geometry of the magnetic field geometry within the @xmath0 emission cone .
furthermore , as mentioned previously , the x - mode waves propagate rectilinearly , while the o - mode waves suffer refraction .
hence , the patterns for @xmath87 and @xmath88 are slightly different due to the refraction . generally speaking ,
the o - mode image is larger than that for the x - mode . for the moderate plasma parameters used in the calculations , the refraction effect is weak and the refraction induced differences are hard to see directly from the patterns for @xmath87 and @xmath88 .
however , it becomes clear from the intensity differences @xmath89 , which are shown in the right panels .
they demonstrate the outward shift of @xmath88 with respect to @xmath87 due to refraction .
if there is no refraction , @xmath89 will almost be zero for the entire beam .
for the uniform density model with @xmath90 in the open field line region at a given emission height , the emission patterns are shown in fig .
[ fig : beam_no - rot](a ) . as we see , @xmath91 and @xmath88 are stronger in the outer parts of pulsar beam where the curvature radii for the particle trajectories are smaller and curvature radiation should be stronger naturally .
in addition , due to the refraction induced outward bending of the o - mode trajectories , the intensity differences , @xmath89 , keep to be positive for almost all parts of the beam .
their largest difference is @xmath92 in fig .
[ fig : beam_no - rot ] for the typical parameters .
hence , the emergent radiation suffers serious depolarization due to the incoherent addition of the two orthogonal modes , as shown in fig .
[ fig : profiles - norot](a ) .
the net polarization within the entire beam is of x - mode .
the out - flowing particles can have a variable distribution in the polar and azimuthal directions in the magnetic axis frame , i.e. , @xmath93 .
the density models for the particles in the form of cone , core and patches have been widely used in the studies of pulsar emissions ( e.g. * ? ? ?
the conal shaped density model can be defined on the neutron star surface and extended to the high magnetosphere , described as , @xmath94,\nonumber\\ g(\phi)&=&1 .
\label{eq : conal}\end{aligned}\ ] ] here , @xmath95 , @xmath96 is the polar angle of a field line footed on the neutron star surface , @xmath97 represents the polar angle maximum , i.e. , the polar angle for the last open field line footed on the neutron star surface . for the conal density model , the intensities of the emitted wave
are shown in fig .
[ fig : beam_no - rot](b ) .
obviously the emissions are peaked at the density cone for both x - mode and o - mode components .
the intensity differences , @xmath89 , exhibit sense reversals due to the refraction of the o - mode component .
therefore , the net polarization exhibits mode jumps twice when a sight line cuts across the beam , as shown in fig .
[ fig : profiles - norot](b ) for @xmath98 , @xmath99 , @xmath100 , @xmath101 , and @xmath102 .
note that the x - mode dominates the central parts of the polarization profiles while o - mode dominates the two profile wings . only for a sight line cutting across the edge part of the cone , for example , @xmath103 , the pure o - mode polarization can be detected .
the core density model can also be described by eq .
[ eq : conal ] except with @xmath104 , the emission patterns and pulse profiles for which are shown in fig .
[ fig : beam_no - rot](c ) and fig .
[ fig : profiles - norot](c ) .
the intensity distribution of both x - mode and o - mode components are similar as those for the conal density model .
the intensity near the beam center is still very weak . because the curvature radiation intensity is still negligible due to the very large curvature radii ( infinity for the central point which stands for the magnetic axis ) ,
although the density for the particles is peaked at the beam center .
the patch density model can also be described by eq .
[ eq : conal ] , except that @xmath105 $ ] . here , @xmath106 represents the position for the density patch in the magnetic azimuth direction and @xmath107 is the characteristic width of the patch . for the eight density patches located at different azimuthal directions around the magnetic axis ,
their emission patterns and pulse profiles are shown in fig .
[ fig : beam_no - rot](d ) and fig .
[ fig : profiles - norot](d ) . the x - mode components are stronger in the central regions near the magnetic axis , while the o - mode components are stronger in the outer parts of pulsar beam , as shown by the pattern for @xmath89 .
the refraction behaves similarly as in the density cone . to summarize
, the refraction will bend the o - mode components to the outer parts of the open field line region .
it leads the intensity differences for the x - mode and o - mode components at a given position and hence causes serious depolarization .
generally the x - mode dominates the central parts of the beam while the o - mode dominates the outer parts .
when the relativistic particles stream out in pulsar magnetosphere , the influences of co - rotation on the emissions can not be ignored . here
, we calculate the emissions of particles in the whole open field line region for the density models of uniformity , cone , core and patches .
the results are shown in fig .
[ fig : beam_rot ] and fig .
[ fig : profiles ] . for the uniform density model ,
the total intensities , @xmath108 , are stronger at the edge parts of the beam , similar as the case without rotation . however , due to the rotation - induced bending of the particle trajectories , the pattern for @xmath108 is not symmetric around the beam center , the leading part is stronger than the trailing part
. the x - mode components , @xmath87 , are stronger at the two sides of the beam in the @xmath109 direction , while the o - mode components are stronger at the two sides of the beam in the @xmath13 direction .
hence , the intensity differences , @xmath89 , show quadruple features . when sight lines of different angles of @xmath109 cut across the beam , various profiles are shown in fig . [
fig : profiles](a ) . for the profiles detected by the sight lines except @xmath98
, the x - mode component appears at the central rotation phase , while the o - mode component emerges at two sides of the profiles . for the conal density model , the emissions
are mainly detected near the density cone ( see fig . [
fig : beam_rot]b ) . when the sight lines cut across the central part of the beam ( e.g. @xmath98 and @xmath99 ) , the two intensity peaks are dominated by o - mode components with a high linear polarization ( see fig .
[ fig : profiles]b ) .
however , when the sight lines cut across the edge parts of the beam , e.g. @xmath110 or @xmath111 , x - mode dominates the single intensity peak .
for a modest @xmath109 , e.g. @xmath112 and @xmath101 , the polarization profile shows changes of modes as `` o x o '' in phases , with the x - mode dominating the central part and the o - mode the two profile wings .
this `` o x o '' structure of the mean profile was also predicted by @xcite , and also comparable to observations @xcite . for the core density model ,
the emission patterns and pulse profiles are shown in fig .
[ fig : beam_rot](c ) and fig .
[ fig : profiles](c ) . unlike the emissions from the particles without rotation , the relativistic particles traveling along the central magnetic field lines also produce considerable radiation .
the emission direction is bent towards the rotation direction with the dominating o - mode .
when sight lines cut across the beam , the x - mode emission is detected in the central part , and the o - mode emissions from the two sides . for eight density patches located on pulsar polar cap ,
the emission beam patterns and pulse profiles are shown in fig .
[ fig : beam_rot](d ) and fig .
[ fig : profiles](d ) .
the distributions of @xmath87 , @xmath88 , @xmath113 , and @xmath89 for the eight density patches are similar as those discrete parts of the density cone . in summary ,
the rotation has significant influences on pulsar emission beam and polarization .
it causes the emissions of the leading components stronger than the trailing ones .
though pa curves are quite similar to these without rotation , the pulse profiles have significant polarization in some parts , different from the case without rotation .
in this paper , we have investigated the emission and the propagation of a single photon , the waves within the @xmath0 emission cone , and the emission within the entire open field line region .
the polarized waves are generated through curvature radiation from relativistic particles traveling along the curved magnetic field lines and co - rotating with a pulsar .
once the polarized waves are generated , they will be coupled to the local plasma modes ( x - mode and/or o - mode ) to propagate outwards in pulsar magnetosphere .
the x - mode component propagates outside in a straight line , while the o - mode component suffers the refraction as described by the hamilton equations and its trajectory bends towards outside , which causes the two components separated from each other . both x - mode and o - mode components should experience the `` adiabatic walking '' with the polarization vectors following the orientation of magnetic fields along the ray .
we calculated the emission and propagation of the polarized waves within @xmath0 around the velocity direction of the relativistic streaming particles for a given rotation phase and emission height .
furthermore , for different density models of particles in the form of the uniformity , cone , core and patches within the open field line region , we calculated the intensities for the x - mode and o - mode components , @xmath87 and @xmath88 , as well as the total intensity , @xmath108 , and the linear polarization , @xmath114 , within the emission beam ( fig .
[ fig : beam_no - rot ] and [ fig : beam_rot ] ) . when sight lines cut across the beam , pulse profiles show various polarizations ( fig .
[ fig : profiles ] ) .
we draw the following conclusions . 1 .
both the x - mode and o - mode wave components can be produced by the curvature radiation process .
their intensities vary a lot for different emission points , but always could be comparable for any emission cone .
2 . the refraction will bend the o - mode emissions towards the outer part of pulsar beam , which is serious for low frequency wave ( low @xmath63 ) if the plasma within pulsar magnetosphere has a high density ( @xmath38 ) and a small velocity ( @xmath64 ) .
thus , the x - mode and o - mode components of initial emission are separated in pulsar magnetosphere , the observed polarization intensity at a given rotation phase should be the incoherent mixture of the x - mode and o - mode emission cones from discrete emission positions . _
the orthogonal mode jumps happen naturally due to the change of the dominance of the two modes .
the co - rotation of particles with magnetosphere has significant influence on pulsar polarized emissions .
if the co - rotation is not considered , the observed magnitude of x - mode and o - modes within the emission beam should be comparable , which means a serious depolarization for pulsar profiles as shown in fig .
[ fig : beam_no - rot ] and [ fig : profiles - norot ] .
however , if the co - rotation is incorporated , the magnitude distributions for the x - mode and o - mode components appear in quite different regions of the pulsar emission beam ( see fig . [
fig : beam_rot ] and [ fig : profiles ] ) , which means the final profiles could have significant polarization with different modes .
since the refraction bend the o - mode emissions towards the outer part of pulsar beam , the final polarization profiles usually show the `` o x o '' modes in phases , with the x - mode dominating the central part and the o - mode at the two wings , which agrees with predictions by @xcite and observations shown by @xcite and @xcite .
note that in some case the central x - mode or the o - mode in out - wings may be weak and only one mode is observable .
the asymmetric intensity distribution of x - mode and o - mode emission caused by the additional co - rotation of particles can be used to understand the high degrees of linear polarization . in this paper , we studied the emission from curvature radiations with or without the rotation involved , and the initial propagation ( i.e. `` adiabatic walking '' and refraction ) of the polarized waves within pulsar magnetosphere . when the waves propagate outwards , other propagation effects , e.g. wave mode coupling , should be considered since it could alter the polarization states .
for example , @xmath115 decreases as the wave propagates outside until @xmath116 . near the region @xmath117
, the wave mode coupling happens , which will lead the generation of circular polarization @xcite . as was demonstrated in @xcite and @xcite , the sense of circular polarization is closely related to the polarization modes and pa curve gradient . for the same pa curve gradient for a pulsar ,
the circular polarization generated by the wave mode coupling should have opposite signs for the x - mode and o - mode in different parts of profiles , which means the sign reversal of circular polarization accompanying the orthogonal mode jump .
our calculations can explain some observation facts .
however , various observation features exist , for example , the partial cone polarization features @xcite , the evolvement of the polarization intensity across frequencies @xcite , high degree of circular polarization @xcite , the polarization evolution for the precession pulsar j1141 - 6545 @xcite , etc , which need further investigations . in our model ,
the plasma within the pulsar magnetosphere is simply assumed to be cold , symmetric for electron and positron and with a density distribution of uniformity , cone , core and patches .
however , the actual energy and density distributions of the particles are not known .
the plasma generated through the cascade process in the polar cap region can be hot and has a wide energy distribution @xcite . meanwhile , the sparking process may also lead to irregular density patches .
nevertheless , only if the o - mode components are coupled to the superluminous o - mode can they leave the magnetosphere , which suggests the intrinsical nonlinear process @xcite .
however , little is known about the nonlinear process up to now . in our calculations for the coherent curvature radiation , the coherency
is simply treated by assuming the coherent bunch to be a huge point charge .
however , the actual coherent manner , both in the bunching form of particles and the maser - type instabilities , will lead to diverse radiation patterns . in our study , the dipole magnetosphere is used to investigate the curvature radiation and propagation processes , while the distortion of the dipole magnetosphere caused either by pulsar rotation or by the polar cap current @xcite is not considered .
these factors will make the final emergent radiation much more complicated @xcite and should be investigated in the future .
the authors thank the anonymous referee for helpful comments .
this work has been supported by the national natural science foundation of china ( 11273029 and 11003023 ) and the young researcher grant of national astronomical observatories , chinese academy of sciences . | the propagation of polarized emission in pulsar magnetosphere is investigated in this paper .
the polarized waves are generated through curvature radiation from the relativistic particles streaming along curved magnetic field lines and co - rotating with the pulsar magnetosphere . within the @xmath0 emission cone ,
the waves can be divided into two natural wave mode components , the ordinary ( o ) mode and the extraordinary ( x ) mode , with comparable intensities .
both components propagate separately in magnetosphere , and are aligned within the cone by adiabatic walking .
the refraction of o - mode makes the two components separated and incoherent .
the detectable emission at a given height and a given rotation phase consists of incoherent x - mode and o - mode components coming from discrete emission regions . for four particle - density models in the form of uniformity , cone , core and patches
, we calculate the intensities for each mode numerically within the entire pulsar beam . if the co - rotation of relativistic particles with magnetosphere is not considered , the intensity distributions for the x - mode and o - mode components are quite similar within the pulsar beam , which causes serious depolarization . however , if the co - rotation of relativistic particles is considered , the intensity distributions of the two modes are very different , and the net polarization of out - coming emission should be significant .
our numerical results are compared with observations , and can naturally explain the orthogonal polarization modes of some pulsars .
strong linear polarizations of some parts of pulsar profile can be reproduced by curvature radiation and subsequent propagation effect .
[ firstpage ] curvature radiation - rotation - relativistic particles - pulsars : general |
mars , one of four terrestrial planets in the solar system , has a small mass of @xmath3 @xmath4 ( @xmath5earth mass ) and an orbital semimajor axis of 1.5au .
terrestrial planets are believed to have formed via collisions among mars - sized planetary embryos in the long - term orbital instability that followed substantial gas dissipation ( _ i.e. _ , chaotic growth ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the @xmath0hf-@xmath0w short - lived chronometer ( @xmath6 9myr ) gives a very short accretion timescale for mars @xcite , similar to the gas dissipation time inferred from disk and meteorite observations @xcite .
this indicates that mars is likely to be a remnant of the early generation of planetary embryos formed via collisional accretion of planetesimals .
collisions between planetesimals produced planetary embryos , which grew further through collisions with surrounding planetesimals @xcite . once embryos became massive , they stirred surrounding planetesimals , producing collisional fragments that contributed to embryo growth @xcite .
when the surrounding bodies were depleted due to the radial drift of small fragments and/or accretion onto embryos , the growth of embryos stalled @xcite . in oligarchic growth , the initial size of planetesimals and disk mass influence the final masses of embryos and their accretion timescales .
small planetesimals are easily fragmented during collisions .
the resulting fragments are small enough to feel some drag from the gas , leading them to spiral towards the sun . as a result , small planetesimals tend to produce embryos that have smaller masses than when large initial planetesimals are considered . on the other hand ,
the interaction of fragments with the gas damps their relative velocities and increases the efficiency with which they can be accreted by embryos through gravitational focusing .
thus , with small planetesimals the timescale of embryo growth is shorter than when large planetesimals are considered . during formation of the terrestrial planets ,
the segregation of metal from silicate produced metallic cores , the timing of which can be estimated using @xmath0hf-@xmath0w systematics .
hafnium , a lithophile ( i.e. , rock - loving ) element , was retained .
in contrast , tungsten , a siderophile ( i.e. , metal - loving ) element , was partitioned into metal cores .
hafnium-182 ( @xmath0hf ) decays to tungsten-182 ( @xmath0w ) with a half life of 9myr .
the decay of @xmath0hf in the martian mantle after removal of w in the core resulted in the production of excess radiogenic @xmath0w compared to other isotopes of w in chondrites .
the @xmath0w excess in the martian mantle is known from measurements of shergottite - nakhlite - chassignite ( snc ) meteorites that are thought to have been ejected from mars by impactors ( _ e.g. _ , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . strictly speaking , @xmath0w variations record the timing of core formation but this process presumably tracked planetary accretion . in this contribution
, we use @xmath0w excess in the martian mantle and the mass of mars to constrain the disk surface density and initial planetesimal size in the mars - forming region .
we use the following disk model for the initial surface mass density of planetesimals @xmath7 and gas @xmath8 : @xmath9 where @xmath10 and @xmath11 are , respectively , the solid and gas surface densities at 1au in the minimum - mass solar nebula ( mmsn ) model @xcite , and @xmath12 is the distance from the sun .
we vary the scaling factor @xmath13 to investigate the conditions for mars formation .
the solid surface density evolves naturally in the simulation , while the gas density is artificially imposed to decrease as a function of time as @xmath14 with a gas dissipation timescale @xmath15 .
disk observations indicate that @xmath15 is on the order of several million years @xcite .
@xcite showed that model outputs of statistical simulations of runaway and oligarchic growth are controlled primarily by the planetesimals that dominate the initial solid surface density , so the results are largely insensitive to the initial size distribution of other minor bodies .
for this reason , we have adopted a fixed initial planetesimal radius @xmath16 rather than using an actual initial size distribution , which is poorly known anyhow @xcite .
we limit our analysis to @xmath17 km because in a turbulent disk , smaller planetesimals would not be sensitive to gravitational focusing , thus preventing runaway growth to proceed ( e.g. , * ? ? ? * ; * ? ? ?
* ; * ? ? ?
we performed numerical simulations of mars formation via collisional evolution of planetesimals . in the model , the disk
is divided into four annuli .
the width of each annulus is given by 0.2 times the annulus characteristic radius defined by the mean of the inner and outer radii of the annulus .
the characteristic radii of the annuli are set to 1.5 , 1.8 , 2.2 , and 2.7au .
@xcite showed that because the timescale for dust removal by gas drag is much shorter than the timescale for embryo growth , considering a larger disk did not change significantly the final embryo mass .
the mass distribution of bodies in each annulus is treated as discrete mass batches with the mass ratio between adjacent mass batches fixed at 1.2 .
we follow the evolution of the mass distribution via collisions between bodies .
the collisional rates depend on the relative velocities between colliders , which depend on their eccentricities and inclinations .
the mass distribution affects eccentricities and inclinations by collisionless interaction between bodies .
we therefore calculate the mass and velocity evolution simultaneously , applying the collision rates @xcite and the velocity evolution rates due to mutual interactions of bodies @xcite and due to gas drag @xcite .
we do not treat collisions and collisionless interactions between bodies in different annuli and therefore adopt an annulus width larger than the feeding zones of planetary embryos . the total fragment mass ejected by a single impact between colliders with masses @xmath18 and @xmath19
is assumed to be @xmath20 using the dimensionless impact energy @xmath21 @xcite , where @xmath22 is the collisional velocity between the colliders and @xmath23 is the specific impact energy ( @xmath24 of impacted material ) needed for the ejection of a half - mass of colliders .
hydrodynamic impact simulations give @xmath25 for 100 km sized or smaller bodies @xcite , while @xmath25 of bodies with radius @xmath26 km are determined by the gravitational binding energy @xcite .
connecting the two regimes ( above and below about 100 km ) , we express @xmath23 value of a body with mass @xmath27 , radius @xmath28 , and density @xmath29 as @xmath30 , where @xmath31 is the gravitational constant and @xmath32 from collisional simulations with distinct element method @xcite . here
we use @xmath33 , @xmath34 , @xmath35 , and @xmath36 , which were obtained from hydrodynamic simulations for basalt with @xmath37 @xcite .
following @xcite and @xcite , when the bodies reach a certain mass @xmath38 such that the sum of their mutual hill radii equals the radial - bin width divided by @xmath39 @xcite , the bodies are regarded as `` runaway bodies '' that do not experience collisions and dynamical interactions due to close encounters among them . using statistical methods
, we can accurately calculate embryo formation and growth until the subsequent stage of chaotic growth , which may not be significant for mars .
the statistical simulation used in the present contribution reproduces the results of @xmath40-body simulations when collisional fragmentation is ignored @xcite .
it also reproduces the analytical solution for mass loss due to a collisional fragmentation cascade @xcite .
embryos grow via collisions with surrounding planetesimals and fragments .
if we assume that the bodies surrounding embryos are not lost at all , then embryos can accrete all the bodies in their feeding zones .
the maximum mass of the embryo is called the isolation mass and is given by ( e.g. , * ? ? ?
* ) @xmath41 where we use eq .
( [ 201002_7dec12 ] ) for the solid surface density .
therefore , a disk with @xmath42 produces mars - size embryos at 1.5au if embryos can collect all solids .
however , since small planetesimals are vulnerable due to low gravity , mass loss can occur by collisional fragmentation of planetesimals with radial drift of the resulting fragments , so the final embryo mass can be smaller than @xmath43 .
destructive collisions between planetesimals produce small bodies , which can in turn be broken apart resulting in still smaller bodies . in the collisional fragmentation cascade ,
the mass distribution of bodies follows a power law ; the surface number density of bodies with mass ranging from @xmath27 to @xmath44 is proportional to @xmath45 , where @xmath46 with @xmath47 @xcite .
the mass distribution is determined by the collisional cascade for bodies with @xmath48 , while runaway growth of planetesimals regulates the mass distribution of bodies with @xmath49 .
with such a mass distribution , planetesimals with radii @xmath50 determine the solid surface density .
the timescale of planetesimal depletion by the collisional fragmentation depends on @xmath23 of planetesimals and is given by @xcite @xmath51 where we assume that the planetesimal eccentricities are in an equilibrium state controlled by stirring by embryos and damping by gas drag and use @xmath52 with @xmath53 for planetesimals .
erosive ( cratering ) collisions with specific impact energies smaller than @xmath23 control this timescale @xcite , which is 45 times shorter than that derived by ignoring erosive collisions as previous studies did ( e.g. , * ? ? ? * ) .
this timescale was obtained by assuming power law mass distribution in the collision cascade .
several effects can lead to departure from such power - law behavior but this would not change significantly the timescale because it is determined by bodies for which the mass distribution does not deviate considerably from a power - law distribution with @xmath54 . since @xmath55 is inversely proportional to the solid surface density @xcite , @xmath55 increases with decreasing @xmath56 and the full - depletion timescale of planetesimals is much longer than @xmath55 . the timescale for embryo growth via planetesimal accretion is given by @xcite , @xmath57 for @xmath58 km , embryos can reach @xmath43 in a disk with @xmath59 because @xmath60 . however , the final embryo mass is smaller than the isolation mass for @xmath61 km .
larger @xmath13 is needed for smaller planetesimals to compensate for loss of mass through gas drift of collision fragments and achieve the same final embryo mass .
note that the necessary @xmath13 to produce mars - size embryos is lower than @xmath13 given by posing @xmath62 , because @xmath55 and @xmath63 increase when the solid surface density decreases @xcite . for small values of @xmath16 ( @xmath6410 km ) ,
embryo growth proceeds to a large extent by fragment accretion .
the collisional fragmentation cascade halts when eccentricities and inclinations of the bodies are damped well by gas drag .
fragments thus accumulate at @xmath65 .
this condition gives a typical fragment size of 110 m depending on the embryo mass @xmath66 .
the radial drift timescale of such bodies is given by ( see the derivation in * ? ? ?
* ) @xmath67 where @xmath68 is the keplarian frequency , we use the value of @xmath23 for @xmath69 m , @xmath70 is the stopping time by gas drag for the fragments under the stokes regime , @xmath71 is the keplarian velocity , @xmath72 is the dimensionless viscous stirring rate for low eccentricity , and @xmath73 .
since small fragments have low eccentricities and inclinations , fragment accretion allows fast growth of embryos .
the embryo growth timescale via fragment accretion is given by @xcite @xmath74 if we ignore the radial drift of fragments , embryos reach @xmath75 in a short timescale of @xmath76 @xcite . however , fragments are removed by radial drift due to gas drag in a shorter timescale of @xmath77 , which stalls embryo growth before they reach isolation masses . in the case of fragment accretion ( small initial planetesimals ) ,
formation of mars - size embryo requires @xmath78 ; @xmath79 . with larger initial planetesimals ,
collisional fragmentation of planetesimals is less effective and therefore embryos can reach the mass of mars for smaller @xmath13 . to summarize , small initial planetesimals are prone to fragmentation and mass loss occurs through gas - induced drift of fragments .
as a result , small planetesimals tend to produce small embryos ( and vice versa ) .
the eccentricities and inclinations of fragments produced by collisions between small planetesimals are damped by interaction with the gas .
the resulting reduction in relative velocities increases the gravitational cross section of embryos for fragments , so small planetesimals allow rapid embryo growth ( and vice versa ) .
we start the investigation for embryo growth in a mmsn disk , where the final embryo mass is about the mass of mars if embryos can fully accrete bodies in their feeding zones ; @xmath80 at 1.5au .
[ fig : embryo_growth]a shows the time evolution of planetary embryo masses at 1.5au for @xmath81 , 3.7 , 27 and @xmath82 km with @xmath83myr .
growth rates and final masses of embryos depend on @xmath16 .
mars - sized embryos can be formed from planetesimals with @xmath84 km in the mmsn disk , while a more massive disk is necessary to form mars from smaller initial planetesimals .
the embryo mass evolution affects the predicted @xmath0w radiogenic excess caused by the partitioning of w into the core and the decay of @xmath0hf in the mantle . in the embryo mantle , the excess abundance of @xmath0w compared to the chondritic reference value ( chur = chondritic uniform reservoir ) is given by ( see the derivation in appendix [ appendix ] ; * ? ? ?
* ; * ? ? ?
* ) @xmath85 where @xmath86 \times 10 ^ 4 $ ] , @xmath87 , @xmath88 , @xmath89 is the embryo mass at time @xmath90 , and @xmath91 is the decay constant of @xmath92 . from previous studies , @xmath93 , @xmath94 , @xmath95 and @xmath96 (
* and references therein ) .
our simulations generate as model output the time evolution of an embryo mass @xmath89 , allowing us to calculate @xmath97 .
[ fig : embryo_growth]b shows the evolution of @xmath98 during embryo growth with different initial planetesimal radii in the mmsn disk .
the predicted @xmath99 values are smaller than the measured one for @xmath81100 km .
small @xmath16 values produce @xmath100 close to the martian mantle value ( fig .
[ fig : embryo_growth]b ) but the final embryo mass is too small and a more massive disk is required ( fig . [
fig : embryo_growth]a ) . the results for massive disks with @xmath101
are shown in fig .
[ fig : embryo_growth_s2.7 ] .
smaller initial planetesimals with @xmath102 km can make mars - size embryos , while larger planetesimals make embryos that are much larger than mars .
the predicted @xmath103w value for @xmath102 km is about 2.84 , which matches the observation .
therefore , a parameter set of @xmath104 and @xmath105 km can account for the mass and @xmath103w value of mars . for @xmath101 and @xmath102 km
, the two inner annuli produce 8 mars mass bodies , while smaller bodies are formed in the outer annuli close to the outer edge of the simulation .
the predicted @xmath103w values of the embryos are similar but slightly lower for embryos formed in the outer disk relative to those formed in the inner disk . in the proposed scenario ,
one of the embryos becomes mars while the others accumulate to make earth and venus size planets in the subsequent stage of chaotic growth ( see [ sc : subsequent ] ) .
to investigate the conditions required for mars formation , we have evaluated systematically the effects of initial planetesimal size and solar disk mass on the final mass and accretion timescale of embryos in the mars - forming region .
[ fig : cont_mass_ew]a summarizes the final embryo masses and the @xmath97 values derived from 88 simulations at 1.5au for a logarithmic grid of @xmath106 , where @xmath107 , 1.4 , 1.9 , 2.7 , 3.7 , 5.2 , 7.2 , 10 and @xmath81 , 1.9 , 3.7 , 7.2 , 14 , 27 , 52 , 100 , 190 , 370 , 720 km .
final embryo masses reach the mass of mars on the thick solid curve in the @xmath13 and @xmath16 space of fig . [
fig : cont_mass_ew]a .
our statistical simulations present a mass distribution of embryos that spans a factor 45 in mass , which is a measure of uncertainty in the predicted embryo mass ( thin solid curves ) .
the tendency that small initial planetesimals require a massive disk for the formation of mars - sized bodies is explained by the analytical solutions for the final embryo mass derived by @xcite .
mass loss due to collisional fragmentation with radial drift of fragments reduces the final mass for small @xmath16 .
the @xmath108 values are calculated from eq .
( [ eq : ew ] ) , where @xmath109 is the error in the predicted @xmath110 value calculated by propagating all uncertainties on model parameters . at 1.5 au , the parameter space that reproduces both the accretion timescale and final mass of mars
is @xmath111 km and @xmath112 . to evaluate the robustness of our conclusion
, we have carefully investigated the parameters that could affect the mass and accretion timescale of mars .
in particular , the influences of embryo atmospheres and gas drag on accretion dynamics , embryo migration , disk dissipation timescales , and disruption strength of bodies are discussed below .
[ fig : other_effects]a shows the results of simulations including the collisional enhancement for capture of small particles by embryos induced by the presence of a planetary atmosphere , with a density profile determined by the atmospheric opacity @xcite .
the parameter space that explains the mass and @xmath113w of mars is similar to the case without the collisional enhancement by planetary atmosphere .
the reason is that embryos have small atmospheres and the atmospheric effect is negligible for an embryo smaller than mars @xcite .
as discussed previously , gas can damp velocities of small particles , thus increasing the efficiency of their capture by embryos .
@xcite further found that gas can directly affect the collisional cross - sections between embryos and small bodies . while the dissipative nature of gas drag promotes collision , strong particle - gas coupling acts against it
by allowing fragments to be carried along with the gas passed the embryo .
[ fig : other_effects]b illustrates the results of simulations including the collisional cross section affected by gas drag that is derived by @xcite .
the collisional enhancement is significant only at @xmath114 for mars - size embryos .
since the typical fragments have radii @xmath115 m corresponding to @xmath116 @xcite , this collisional enhancement is minor for mars formation ( compare figs . [
fig : cont_mass_ew]a with [ fig : other_effects]b ) .
we additionally investigated how the gas dissipation time @xmath15 affected the accretion history of mars .
the final embryo mass and @xmath113w value for @xmath117myr are almost the same as those for @xmath83myr ( fig .
[ fig : other_td]a ) . fig .
[ fig : other_td]b shows the results for an extremely short gas dissipation time of @xmath118myr .
the parameter space where the mass and accretion timescale of mars are reproduced is not very different from that obtained when the dissipation time is set to higher values of 5 or 10 myr .
recent studies suggested that terrestrial planets were constructed from a narrow ring between the current orbits of earth and venus ( 0.7 to 1au ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . in this scenario ,
mars was likely scattered from around the outer edge of the truncated disk .
we thus additionally investigate the possibility of mars formation at 1au by setting the characteristic radii of the innermost and outermost annuli at 1 and 2.7au , respectively .
figure [ fig : cont_mass_ew]b shows the final embryo mass and the @xmath119 value at 1au in the @xmath16 and @xmath13 space .
the parameter space required for the production of mars - sized embryo is similar to the case at 1.5au .
however , larger planetesimals with @xmath120 km could explain the mass and accretion timescale of mars .
a very important development of this work and other recent studies @xcite is the recognition that erosive collisions and fragment accretion dominate embryo growth .
the treatment of fragmentation depends on @xmath23 , which affects final embryo masses @xcite .
the @xmath23 value is uncertain .
for instance , bodies made of rubble piles have @xmath23 smaller than what we used , thus producing smaller embryos @xcite . with @xmath121 km
, low @xmath23 requires larger @xmath13 for mars formation , which may rise the value of plausible initial planetesimal radii only slightly .
terrestrial planet formation was investigated by @xmath40-body simulations @xcite .
our statistical simulation does not follow the orbits of bodies individually , but reproduces well the mass and random - velocity evolution obtained from @xmath40-body simulations under the assumption of perfect accretion @xcite . for terrestrial planet formation ,
@xcite performed statistical simulations similar to ours and obtained embryos of mass @xmath122 g after about 0.1myr for @xmath123 and @xmath124 km , which is slightly larger than embryo masses from our simulations ( @xmath125 g ) . using the fragmentation model of @xcite , @xcite carried out statistical simulations for @xmath126 km and obtained embryos of mass about @xmath127 g at 1myr for @xmath128 . in our simulations ,
the embryo masses are of about @xmath129 g at 1myr for @xmath130 and @xmath81 km .
thus , the final embryo masses predicted in our simulations are systematically lower than those obtained by previous workers for the same sets of conditions .
the difference stems from a different treatment of erosive ( cratering ) collisions . in the fragmentation model of @xcite ,
erosive collisions with specific impact energies smaller than @xmath23 yield less ejecta than those that are expected from impact laboratory experiments , while these are well reproduced by the fragmentation model developed by @xcite , which is used in the present contribution .
@xcite carried out simulations of terrestrial planet formation where planetary embryos were treated as individual entities , while small bodies were followed using mass bins .
these simulations showed that smaller initial planetesimals produced smaller planetary embryos .
this tendency is consistent with ours .
@xcite used a similar model ; starting from planetary embryos and planetesimals .
the simulations in @xcite followed collisional fragmentation cascades using mass bins with a mass ratio between adjacent bins of 10 .
these wide mass bins artificially slowed down planetesimal depletion by the fragmentation cascade by a factor of @xmath131 .
@xcite and @xcite considered the migration of embryos due to tidal interaction with the nebula ( type i migration ; e.g. , * ? ? ?
we briefly discuss the effect of type i migration on terrestrial planet growth below . from our simulations ,
tens of mars size embryos with orbital separation of 10 mutual hill radii are expected to form in the terrestrial planet forming region . for mars to remain small during chaotic growth , @xcite and @xcite suggested that planetary embryos must have been concentrated inside about 1au .
such orbital concentration may be attained by type i migration of embryos caused by gravitational interaction with the surrounding nebula ( e.g. , * ? ? ?
* ) if the torque cancellation due to the unsaturation of corotation torque occurs at @xmath132au where heating by disk accretion determines disk temperature @xcite .
mars may have evaded collisions with other embryos during chaotic growth by being scattered to its current orbit at 1.5au .
the planets resulting from simulations of chaotic growth have orbital eccentricities higher than those of venus and earth @xcite , but these eccentricities were most likely damped by gravitational interactions with a tenuous gas disk and/or a planetesimal swarm @xcite .
this dynamical friction could have been achieved by small ( @xmath133 m ) leftover planetesimals that were accreted by earth and mars after completion of core formation @xcite .
the mass of the late veneer on mars is estimated from abundances of highly siderophile elements in snc meteorites to be @xmath134 g . assuming a chondritic impactor composition ( @xmath135 ; * ? ? ?
* ) , this would have delivered @xmath136 g of w @xcite
. the martian mantle contains @xmath137 g of w ( 62 ppb w ; * ? ? ?
thus , planetesimals responsible for damping the eccentricities and inclinations of the terrestrial planets may have delivered only @xmath138 of the w inventory of the martian mantle and would have had a negligible effect on the w isotopic composition .
@xcite proposed that jupiter migrated inward until saturn was caught in a 2:3 resonance with jupiter , at which point jupiter migrated outwards .
the u - turn migration of jupiter at 1.5 au would have truncated the disk and could explain mars present orbit and mass .
if the jupiter s passing fully cleaned up small bodies , proto - mars could no longer grow after it was scattered to the current orbit .
our simulations can be interpreted in light of this scenario by considering mars isolation time @xmath139 , after which the embryo stops growing .
the u - turn migration of jupiter occurred at @xmath15 @xcite .
proto - mars could be scattered to its current orbit after the migration of jupiter ; @xmath140 . for @xmath141myr
several times longer than the gas dissipation time @xmath15 , mars small mass and rapid accretion is best explained by starting with small planetesimals as was discussed above ( fig .
[ fig : ti]a ) .
however , if one considers a rapid isolation time , which is more likely in the context of the grand tack scenario @xcite , models that start with large planetesimals are allowed ( fig .
[ fig : ti]b ) , because mars avoids late accretion of chondritic material . as discussed in sect .
3.3 , considering the conditions needed to form jupiter and saturn , the scenario of formation of mars from large planetesimals in a mmsn disk is unlikely .
a standing question is whether collisions between mars and other embryos during chaotic growth could have lead to further growth and modified the w isotope evolution of the martian mantle . during such collisions ,
the core of the impactor may partially merge with the target core without complete isotopic equilibration with the target mantle @xcite . therefore mars growth via collisions of sub - mars embryos in the subsequent stage of chaotic growth
might be allowed to some extent . however
, a planetesimal disk leading to small embryos would not have contained enough mass to produce the terrestrial planets , which makes it unlikely that mars grew by the protracted accretion of small embryos .
mars has experienced at least one collision at a later stage .
the northern hemisphere of mars is lower and has thinner crust that the southern hemisphere , a feature that is commonly referred to as the martian hemispheric dichotomy .
several interpretations have been proposed , including one that calls for a large impact that would have obliterated the northern hemisphere less than 100 myr after solar system formation ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
@xcite used smooth particle hydrodynamics simulations to constrain the impactor to a 1,000km - radius body that would have impacted mars at a @xmath142 angle . using the mass - radius relationship @xmath143 ( @xmath144 is earth radius ;
* ) , such an impactor would have delivered @xmath145 of the mass of mars .
most likely , this would have taken place after most @xmath0hf had decayed , possibly @xmath146myr after solar system formation .
we have evaluated the impact of such a late impact on the w isotopic evolution of the martian mantle for one of the simulations that reproduces the mass and @xmath103w values of mars ( @xmath147 and @xmath148 km ) .
the calculated @xmath103w shifts from 2.7 to 2.5 when taking into account this late impactor .
this is well within uncertainties , so our conclusions are not affected by the possibility that the martian dichotomy was produced by a late impactor .
the assumption of a 1000-km radius body is conservative as @xcite considered a head on collision with a smaller body of 320 km radius , which would have a negligible effect on @xmath103w . as discussed in [
sc : subsequent ] , two scenarios can explain the small mass and @xmath103w value of mars : ( i ) small planetesimals in a massive disk formed mars or ( ii ) large planetesimals in an @xmath149mmsn disk formed mars but its growth was arrested early by inward then outward migration of jupiter .
further constraints bearing on this issue can be obtained by examining the conditions necessary for giant planet formation .
if an embryo reaches a critical core mass , it can not sustain an static atmosphere and starts rapid gas accretion to form a gas giant planet , such as jupiter and saturn ( e.g. , * ? ? ?
* ; * ? ? ?
however , before the embryo attains the critical core mass , collisional fragmentation induced by gravitational stirring stalls the growth of the embryo @xcite .
@xcite suggested that fragment accretion enhanced by tenuous atmospheres allowed formation of massive cores .
@xcite meanwhile pointed out that combination of gas drag and stirring by an embryo inhibits the accretion of fragments with radii larger than 30 m onto the embryo .
however , the typical fragment radius is smaller than 30 m and hence the obstacle to fragment accretion is negligible ( see the discussion in * ? ? ? * ) .
since the removal of fragments due to radial drift is more effective than accretion , embryos can not reach the critical core mass starting from small initial planetesimals . although the core formation must be completed before gas depletion ,
the formation timescale of a core is longer than @xmath150 for large planetesimals , for which initial runaway growth or cores is slow .
therefore , jupiter s core is likely to have formed from accretion of intermediate size planetesimals in a massive disk @xcite .
jupiter s gas accretion was truncated by its gap opening . at the edge of the gap ,
a radial pressure maximum was created .
small fragments accumulated at the pressure maximum , which might have lead to the formation of saturn s core in @xmath151yr @xcite .
heat released by solid accretion stabilizes atmospheres of cores and delays the rapid gas accretion to form gas giant planets , and hence high solid accretion rates of cores result in large cores ( e.g. , * ? ? ? * ) .
saturn s core is estimated to be larger than jupiter s core @xcite , which is difficult to explain in classical core formation models .
@xcite recently proposed a core formation scenario by accretion of sub - meter sized pebble , which also produces a higher accretion rate for jupiter unless the pebble surface density has a flat radial profile ( the radial slope @xmath152 ) .
a promising model to explain saturn s large core relative to jupiter is if saturn s core was formed at the gap opened by jupiter @xcite . in the context of the model of @xcite , which uses the same prescriptions as used in the present contribution ,
jupiter and saturn can be formed in limited conditions for planetesimal sizes and disk masses .
large planetesimals @xmath153200 km in a massive disk @xmath154 are required to form a massive core of @xmath155 at 5.2au for jupiter @xcite . on the other hand
, the larger core of saturn requires very rapid formation , which is realized by fragment accretion at the pressure maximum caused by jupiter s gap opening in the solar nebula .
this yields an additional constraint described empirically by the relationship @xmath156 @xcite .
both conditions for formation of the cores of jupiter and saturn are satisfied for @xmath157 and @xmath158100 km ( see fig . [
fig : cont_mass_ew ] ) . as shown in figs .
[ fig : cont_mass_ew][fig : ti ] , the conditions required to explain the accretion of the cores of jupiter and saturn are inconsistent with those inferred for mars .
the condition required for jupiter and saturn formation produces earth - sized or larger embryos in the terrestrial planet region and the predicted @xmath159 adopting @xmath160 for earth s mantle is @xmath161 , which is much larger than @xmath162 in terrestrial rocks @xcite .
one way to try to reconcile the apparently inconsistent conditions inferred for mars on the one hand and jupiter and saturn on the other hand is to modify the disk surface density . in the discussion above and in fig .
[ fig : cont_mass_ew][fig : ti ] , we assumed @xmath163 .
it is possible for instance that the disk surface density had a shallower radial slope of @xmath164 .
the conditions for forming the cores of jupiter and saturn become @xmath165 and @xmath166 km . if we use a massive disk of 0.1 solar mass required by jupiter formation , the radius of initial planetesimals for mars formation becomes less than 5 km for @xmath167 .
changing the disk density profile does not seem to solve the discrepancy that exists between mars and the giant planets .
a way to reconcile the constraints derived from inner and outer solar system bodies is if planets formed in a relatively massive solar nebula ( @xmath168 ) but planetesimals sizes were different in the inner and outer regions . in this scenario , jupiter and saturn cores
would have formed from planetesimals of 50 km radius while mars formed from smaller planetesimals of @xmath133 km radius .
@xcite proposed that 100 km or larger planetesimals were necessary to explain the mass distribution of asteroids in the main belt , but @xcite concluded that planetesimals of 100 m size could also reproduce the asteroid mass distribution .
the sizes of planetesimals are related to their formation mechanisms ( _ e.g. _ , * ? ? ?
* ; * ? ? ?
the heliocentric gradient in planetesimal sizes discussed here may be related to the inferred contrast in solid surface densities and material properties across the snow line .
we thank d. minton , a. morbidelli , and an anonymous reviewer for comments that helped improve the manuscript .
hk gratefully acknowledges the support from grants - in - aid from mext ( 23103005 ) and `` institutional program for young researcher overseas visits '' ( r29 ) .
nd acknowledges the support of nasa ( nnx12ah60 g ) and nsf ( ear-0820807 ) .
the mantle @xmath0w value evolves due to extraterrestrial delivery , removal of metal to core , and radioactive decay of @xmath0hf .
a planetary embryo mainly grows through collisions with small bodies . if the sizes of bodies are much smaller than the thickness of the embryo mantle , the sinking metal breaks up in tiny droplets that can equilibrate isotropically with the surrounding mantle @xcite . on the other hand ,
one third of embryo mass growth is provided by collisions with similar mass bodies @xcite .
according to @xcite , contribution of large bodies to the @xmath103w evolution of the martian mantle is negligible .
therefore , we assume that the bodies accreted by mars are fully equilibrated with the mantle before metal removal into the core . accordingly , the equation governing the @xmath0w isotopic evolution is , @xmath169_{\rm mantle } ) = \ , [ ^{182}{\rm w}]_{\rm chur } \frac{dm}{dt } -d^{\rm w } \ , [ ^{182}{\rm w}]_{\rm mantle } \frac{d m_{\rm core}}{dt } + \lambda \ , [ ^{182}{\rm hf}]_{\rm mantle } m_{\rm mantle},\label{142100_5nov12}\ ] ] where @xmath170 and @xmath171 are , respectively , the mantle and core masses ( @xmath172 ) , @xmath173 is the w metal / silicate partition coefficient .
the mass balance for w in the bulk planet can be written as , @xmath174_{\rm mantle } m_{\rm mantle } + d^{\rm w } [ { \rm w}]_{\rm mantle } m_{\rm core } = [ { \rm w}]_{\rm chur } m.\label{124541_5nov12}\ ] ] on the other hand , hf does not go to the mantle and the mass balance is given by @xmath175_{\rm mantle } m_{\rm mantle } = [ { \rm hf}]_{\rm chur } m.\label{125138_5nov12}\ ] ] we assume that the core mass fraction , @xmath176 , is constant .
we then obtain , from eqs .
( [ 124541_5nov12 ] ) and ( [ 125138_5nov12 ] ) , @xmath177 the variation of @xmath178 and @xmath179 is given by eq .
( [ 142100_5nov12 ] ) divided by @xmath180 using eqs .
( [ 124541_5nov12 ] ) : @xmath181 using eq .
( [ 144436_5nov12 ] ) and @xmath182 and @xmath183 , we then have @xmath184 integration of eq .
( [ 133959_23nov12 ] ) results in @xmath185 applying the definition of @xmath186 , eq .
( [ 153641_5nov12 ] ) can be rewritten as eq .
( [ eq : ew ] ) .
adachi , i. , hayashi , c. , nakazawa , k. 1976 .
the gas drag effect on the elliptical motion of a solid body in the primordial solar nebula .. progress of theoretical physics 56 , 1756 - 1771 .
andrews - hanna , j. c. , zuber , m. t. , banerdt , w. b. 2008 .
the borealis basin and the origin of the martian crustal dichotomy .
nature 453 , 1212 - 1215 .
benz , w. , asphaug , e. 1999 .
catastrophic disruptions revisited .
icarus 142 , 5 - 20 .
bromley , b. c. , kenyon , s. j. 2011 .
migration of planets embedded in a circumstellar disk .
the astrophysical journal 735 , 29 .
chambers , j. e. 2006 .
planet formation with migration .
the astrophysical journal 652 , l133-l136 .
chambers , j. 2008 . oligarchic growth with migration and fragmentation .
icarus 198 , 256 - 273 .
chambers , j. e. , wetherill , g. w. , boss , a. p. 1996 . the stability of multi - planet systems .
icarus 119 , 261 - 268 .
dahl , t. w. , stevenson , d. j. 2010 .
turbulent mixing of metal and silicate during planet accretion and interpretation of the hf - w chronometer .
earth and planetary science letters 295 , 177 - 186 .
dauphas , n. , pourmand , a. 2011 .
hf - w - th evidence for rapid growth of mars and its status as a planetary embryo .
nature 473 , 489 - 492 .
dauphas , n. , chaussidon , m. 2011 . a perspective from extinct radionuclides on a young stellar object : the sun and its accretion disk
. annual review of earth and planetary sciences 39 , 351 - 386 .
foley , c. n. , wadhwa , m. , borg , l. e. , janney , p. e. , hines , r. , grove , t. l. 2005 . the early differentiation history of mars from @xmath0w-@xmath187nd isotope systematics in the snc meteorites .
geochimica et cosmochimica acta 69 , 4557 - 4571 .
hansen , b. m. s. 2009 .
formation of the terrestrial planets from a narrow annulus .
the astrophysical journal 703 , 1131 - 1140 .
hayashi , 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 , 35 - 53 .
haisch , k. e. , jr . ,
lada , e. a. , lada , c. j. 2001 . disk frequencies and lifetimes in young clusters .
the astrophysical journal 553 , l153-l156 .
ikoma , m. , nakazawa , k. , emori , h. 2000 .
formation of giant planets : dependences on core accretion rate and grain opacity .
the astrophysical journal 537 , 1013 - 1025 .
inaba , s. , ikoma , m. 2003 . enhanced collisional growth of a protoplanet that has an atmosphere .
astronomy and astrophysics 410 , 711 - 723 inaba , s. , wetherill , g. w. , ikoma , m. 2003 .
formation of gas giant planets : core accretion models with fragmentation and planetary envelope .
icarus 166 , 46 - 62 .
inaba , s. , tanaka , h. , nakazawa , k. , wetherill , g. w. , kokubo , e. 2001 .
high - accuracy statistical simulation of planetary accretion : ii .
comparison with n - body simulation .
icarus 149 , 235 - 250 .
iwasaki , k. , tanaka , h. , nakazawa , k. , hiroyuki , e. 2001 .
the gas - drag effect on the orbital instability of a protoplanet system .
publications of the astronomical society of japan 53 , 321 - 329 .
jacobsen , s. b. 2005 .
the hf - w isotopic system and the origin of the earth and moon . annual review of earth and planetary sciences 33 , 531 - 570 .
johansen , a. , klahr , h. , henning , t. 2006 .
gravoturbulent formation of planetesimals .
astrophysical journal 636 , 1121 - 1134 .
kleine , t. , mezger , k. , munker , c. , palme , h. , bischoff , a. 2004 .
^1 ^ 8 ^ 2hf-^1 ^ 8 ^ 2w isotope systematics of chondrites , eucrites , and martian meteorites : chronology of core formation and early mantle differentiation in vesta and mars .
geochimica et cosmochimica acta 68 , 2935 - 2946 .
kleine , t. , touboul , m. , bourdon , b. , nimmo , f. , mezger , k. , palme , h. , yin , q .- z . , jacobsen , s.b .
, halliday , a.n .
hf - w chronometry and the accretion and early differentiation of asteroids and terrestrial planets .
geochimica et cosmochimica acta 73 , 5150 .
kenyon , s. j. , bromley , b. c. 2004 . detecting the dusty debris of terrestrial planet formation . the astrophysical journal 602 , l133-l136 .
kenyon , s. j. , bromley , b. c. 2004 .
collisional cascades in planetesimal disks .
ii . embedded planets . the astronomical journal 127 , 513 - 530 .
kenyon , s. j. , bromley , b. c. 2006 . terrestrial planet formation . i. the transition from oligarchic growth to chaotic growth .
the astronomical journal 131 , 1837 - 1850 .
kenyon , s. j. , bromley , b. c. 2009 .
rapid formation of icy super - earths and the cores of gas giant planets .
the astrophysical journal 690 , l140-l143 .
kobayashi , h. , ormel , c. w. , ida , s. 2012 .
rapid formation of saturn after jupiter completion .
the astrophysical journal 756 , 70 .
kobayashi , h. , tanaka , h. 2010 .
fragmentation model dependence of collision cascades .
icarus 206 , 735 - 746 .
kobayashi , h. , tanaka , h. , krivov , a. v. 2011 .
planetary core formation with collisional fragmentation and atmosphere to form gas giant planets .
the astrophysical journal 738 , 35 .
kobayashi , h. , tanaka , h. , krivov , a. v. , inaba , s. 2010 . planetary growth with collisional fragmentation and gas drag .
icarus 209 , 836 - 847 .
kokubo , e. , ida , s. 1998 .
oligarchic growth of protoplanets .
icarus 131 , 171 - 178 .
kokubo , e. , ida , s. 2000 .
formation of protoplanets from planetesimals in the solar nebula .
icarus 143 , 15 - 27 .
kokubo , e. , ida , s. 2002 .
formation of protoplanet systems and diversity of planetary systems .
astrophysical journal 581 , 666 - 680 .
kominami , j. , ida , s. 2002 . the effect of tidal interaction with a gas disk on formation of terrestrial planets .
icarus 157 , 43 - 56 .
kretke , k. a. , lin , d. n. c. 2012 .
the importance of disk structure in stalling type i migration .
the astrophysical journal 755 , 74 .
lambrechts , m. , johansen , a. 2012 .
rapid growth of gas - giant cores by pebble accretion .
astronomy and astrophysics 544 , a32 .
levison , h. f. , thommes , e. , duncan , m. j. 2010 .
modeling the formation of giant planet cores .
i. evaluating key processes . the astronomical journal 139 , 1297 - 1314 .
lodders , k. , fegley , b. 1997 . an oxygen isotope model for the composition of mars .
icarus 126 , 373 - 394 .
marinova , m. m. , aharonson , o. , asphaug , e. 2008 .
mega - impact formation of the mars hemispheric dichotomy .
nature 453 , 1216 - 1219 .
mizuno , h. 1980 .
formation of the giant planets .
progress of theoretical physics 64 , 544 - 557 .
morbidelli , a. , bottke , w. f. , nesvorn , d. , levison , h. f. 2009 .
asteroids were born big .
icarus 204 , 558 - 573 . morishima , r. golabek , g. j. , samuel , h. 2012 .
@xmath40-body simulations of oligarchic growth of mars : implications for hf - w chronology .
earth and planetary science letters 366 , 616 .
morishima , r. , schmidt , m. w. , stadel , j. , moore , b. 2008 .
formation and accretion history of terrestrial planets from runaway growth through to late time : implications for orbital eccentricity . the astrophysical journal 685 , 1247 - 1261 .
nimmo , f. , hart , s. d. , korycansky , d. g. , agnor , c. b. 2008 .
implications of an impact origin for the martian hemispheric dichotomy .
nature 453 , 1220 - 1223 .
obrien , s. , osborne , s. , buckley , k. , fehse , r. , amann , a. , oreilly , e. p. , barry , l. p. , anandarajah , p. , patchell , j. , ogorman , j. 2006
. inverse scattering approach to multiwavelength fabry - prot laser design .
physical review a 74 , 063814 .
ogihara , m. , ida , s. , morbidelli , a. 2007 .
accretion of terrestrial planets from oligarchs in a turbulent disk .
icarus 188 , 522 - 534 .
ohtsuki , k. , stewart , g. r. , ida , s. 2002 .
evolution of planetesimal velocities based on three - body orbital integrations and growth of protoplanets .
icarus 155 , 436 - 453 .
okuzumi , s. , tanaka , h. , kobayashi , h. , wada , k. 2012 .
rapid coagulation of porous dust aggregates outside the snow line : a pathway to successful icy planetesimal formation . the astrophysical journal 752 , 106 .
ormel , c. w. , cuzzi , j. n. 2007 .
closed - form expressions for particle relative velocities induced by turbulence . astronomy and astrophysics 466 , 413 - 420 .
ormel , c. w. , klahr , h. h. 2010 . the effect of gas drag on the growth of protoplanets .
analytical expressions for the accretion of small bodies in laminar disks .
astronomy and astrophysics 520 , a43 .
ormel , c. w. , dullemond , c. p. , spaans , m. 2010 .
accretion among preplanetary bodies : the many faces of runaway growth .
icarus 210 , 507 - 538 .
ormel , c. w. , kobayashi , h. 2012 . understanding how planets become massive .
i. description and validation of a new toy model .
the astrophysical journal 747 , 115 .
paardekooper , s .- j . ,
baruteau , c. , crida , a. , kley , w. 2010 . a torque formula for non - isothermal type i planetary migration - i. unsaturated horseshoe drag . monthly notices of the royal astronomical society 401 , 1950 - 1964 .
paardekooper , s .- j . ,
baruteau , c. , kley , w. 2011 . a torque formula for non - isothermal type i planetary migration - ii .
effects of diffusion .
monthly notices of the royal astronomical society 410 , 293 - 303 .
rafikov , r. r. 2004 . fast accretion of small planetesimals by protoplanetary cores . the astronomical journal 128 , 1348 - 1363 .
rubie , d. c. , melosh , h. j. , reid , j. e. , liebske , c. , righter , k. 2003 . mechanisms of metal - silicate equilibration in the terrestrial magma ocean . earth and planetary science letters 205 , 239 - 255 .
samuel , h. 2012 .
a re - evaluation of metal diapir breakup and equilibration in terrestrial magma oceans . earth and planetary science letters 313 , 105 - 114 .
saumon , d. , guillot , t. 2004 .
shock compression of deuterium and the interiors of jupiter and saturn .
the astrophysical journal 609 , 1170 - 1180 .
schlichting , h. e. , warren , p. h. , yin , q .- z .
the last stages of terrestrial planet formation : dynamical friction and the late veneer . the astrophysical journal 752 , 8 .
sotin , c. , grasset , o. , mocquet , a. 2007 .
mass radius curve for extrasolar earth - like planets and ocean planets .
icarus 191 , 337 - 351 .
stewart , s. t. , leinhardt , z. m. 2009 .
velocity - dependent catastrophic disruption criteria for planetesimals .
astrophysical journal 691 , l133-l137 .
tanaka , h. , takeuchi , t. , ward , w. r. 2002 .
three - dimensional interaction between a planet and an isothermal gaseous disk .
i. corotation and lindblad torques and planet migration . the astrophysical journal 565 , 1257 - 1274 .
youdin , a. n. , lithwick , y. 2007 .
particle stirring in turbulent gas disks : including orbital oscillations .
icarus 192 , 588 - 604 .
walker , r. j. 2009 .
highly siderophile elements in the earth , moon and mars : update and implications for planetary accretion and differentiation . chemie der erde / geochemistry 69 , 101 - 125 .
walsh , k. j. , morbidelli , a. , raymond , s. n. , obrien , d. p. , mandell , a. m. 2011 . a low mass for mars from jupiter s early gas - driven migration .
nature 475 , 206 - 209 .
warren , p. h. , kallemeyn , g. w. , kyte , f. t. 1999 .
origin of planetary cores : evidence from highly siderophile elements in martian meteorites .
geochimica et cosmochimica acta 63 , 2105 - 2122 .
wasson , j. t. , kallemeyn , g. w. 1988 .
compositions of chondrites .
royal society of london philosophical transactions series a 325 , 535 - 544 .
weidenschilling , s. j. 2008 .
accretion of planetary embryos in the inner and outer solar system .
physica scripta volume t 130 , 014021 .
weidenschilling , s. j. 2011 .
initial sizes of planetesimals and accretion of the asteroids .
icarus 214 , 671 - 684 .
weidenschilling , s. j. , spaute , d. , davis , d. r. , marzari , f. , ohtsuki , k. 1997 .
accretional evolution of a planetesimal swarm .
icarus 128 , 429 - 455 .
wetherill , g. w. , stewart , g. r. 1993 .
formation of planetary embryos - effects of fragmentation , low relative velocity , and independent variation of eccentricity and inclination .
icarus 106 , 190 .
wilhelms , d. e. , squyres , s. w. 1984 .
the martian hemispheric dichotomy may be due to a giant impact .
nature 309 , 138 - 140 .
[ fig : embryo_growth ] ( a ) time evolution of embryo mass at 1.5au in the mmsn disk ( @xmath188 ) , starting from different planetesimal radii @xmath16 .
the gray dotted line indicates the mass of mars .
( b ) evolution of @xmath98 obtained from the mass evolution shown in panel ( a ) .
the filled circle and the error bar represent the @xmath0w excess in the martian mantle , @xmath189 @xcite .
[ fig : cont_mass_ew ] constraints from final embryo masses and @xmath190 on the initial radius of planetesimals , @xmath16 , at 1.5au ( a ) and 1au ( b ) in the @xmath191mmsn disk .
embryos reach the mass of mars on the solid curve .
the region between the thin curves produces embryos within a factor 2 of the mass of mars . on the thick dashed curve ,
the predicted @xmath192 values correspond to the mean value obtained from martian meteorites .
the thick dashed curve indicates the parameter space where @xmath193 . below this limit
, the accretion timescale of mars would be too long to explain @xmath98 measured in martian meteorites .
the colored region at small @xmath16 represents the condition satisfying the mass and @xmath98 value of mars and the other at @xmath194100 km indicates the parameters for the formation of jupiter and saturn cores . for reference , the right vertical axis represents the disk mass corresponding to @xmath13 in the left axis , assuming that the inner and outer edges of the disk are at 0.4 and 30au , respectively .
[ fig : other_effects ] same as fig .
[ fig : cont_mass_ew]a , but including the accretion enhancement by planetary atmosphere ( a ) and gas drag ( b ) . fig .
[ fig : other_td ] same as fig .
[ fig : cont_mass_ew]a , but for different gas dissipation timescales @xmath117myr ( a ) and 1myr(b ) . | mars is likely to be a planetary embryo formed through collisions with planetesimals , which can explain its small mass and rapid formation timescale obtained from @xmath0hf-@xmath0w chronometry . in the classical theory of planet formation
, the final embryo mass is determined only by the solid surface density .
however , embryos can stir surrounding planetesimals , leading to fragmentation through erosive ( cratering ) collisions .
we find that radial drift of small fragments can drastically reduce the solid surface density . on the other hand ,
embryo growth is accelerated by fragment accretion . since collisional fragmentation efficiency depends on the initial size of planetesimals , the final embryo mass and its growth time
are determined by the initial planetesimal size and disk surface density .
we have investigated the effect of these two parameters on the mass of mars and the predicted radiogenic excess of @xmath0w in the martian mantle .
two scenarios can explain the rapid formation of small mars : ( i ) it formed by accretion of small planetesimals in a massive disk or ( ii ) it formed from large planetesimals but its growth was arrested by the inward then outward migration of jupiter .
taking into account all constraints , we conclude that mars is likely to have formed in a massive disk of about @xmath1 solar mass from planetesimals smaller than @xmath2 km in radius .
such small planetesimal size can not explain core accretion of jupiter , suggesting that there may have been a heliocentric gradient in planetesimal size in the solar nebula .
key words : mars ; terrestrial planets ; planetary formation ; planetesimals ; origin , solar system |
the properties of the x - ray emission of the powerful radio galaxies have been investigated in the literature ( fabbiano et al .
1984 , crawford & fabian 1995 , 1996 ) . in the framework of the agns unification models ( barthel 1989 , urry & padovani 1995 ) the nuclear x ray emission of the frii radio galaxies
is expected to be heavily absorbed along the line of sight by a circumnuclear molecular torus .
strong absorption has indeed been detected in the x
ray observations of the narrow line frii radio galaxies cyg a ( ueno et al . 1994 ) and 3c 194 ( crawford & fabian 1996 ) ; moreover , intrinsic absorption has been discovered in the x ray spectra of the broad line radio galaxies ( blrgs ) 3c 109 ( allen & fabian 1992 ) and 3c 287.2 ( crawford & fabian 1995 ) .
worrall et al . (
1994 ) have suggested that part of the soft x ray emission of some strong frii radio galaxies is synchrotron self
compton ( ssc ) in the agn jets .
ssc may also explain the x ray emission observed with rosat in two emission regions coincident with the radio hot
spots of cygnus a ( harris et al . 1994 ) , but it is generally too weak to be detected even in the nearby radio galaxies ( hardcastle et al .
1998a ) .
it is well known that non thermal mechanisms can produce extended x ray emission in the lobes of the radio galaxies where the relativistic electrons can interact with the microwave background photons ( cmb ) and radiate via the inverse compton ( ic ) process .
the relativistic particle densities and magnetic field strengths in the extended lobes of radio galaxies are usually estimated on the basis of the minimum energy assumption ( equipartition ) .
the detection of ic scattered x
rays from the radio lobes could provide an invaluable tool to determine the value of these important physical parameters .
unfortunately , this process is not particularly efficient at low redshifts ( the ic emissivity increases as @xmath1(1 + z)@xmath2 ) so that rosat and asca have failed to detect this effect in nearby radio galaxies with a few exceptions ( notably , fornax a : feigelson et al .
1995 , kaneda et al .
1995 ; cen b : tashiro et al .
1998 ) . in a recent paper brunetti et al .
( 1997 ) have shown that in the framework of the unification scheme linking strong frii radio galaxies and radio loud quasars it is possible to predict large x
ray fluxes by the ic scattering of the far / near ir photons from a typical `` hidden '' quasar with the relativistic electrons in the radio lobes . even if this effect is expected to dominate the ic contribution also at high redshifts , and predicts x ray fluxes and spectral shapes consistent with those observed for a number of strong distant frii radio galaxies ( redshift z @xmath1 1 ) , it is difficult to be tested in detail because of the low spatial resolution and sensitivity of the available x ray telescopes .
the far / near ir photons from the `` hidden '' quasar are scattered in the x ray band mainly by mildly relativistic electrons ( @xmath3 ) which are not those responsible for the synchrotron radio emission ( typically @xmath4 ) .
therefore , the detection of the ic emission predicted by brunetti et al.(1997 ) model would not only provide additional evidence in favour of the radio galaxy / quasar unification scheme , but it would also provide an unique handle to probe the spectral and spatial distributions of the relativistic particles over a wider energy range . nevertheless , optical and radio observations have suggested that relatively distant radio galaxies ( z@xmath5 0.3 ) lie at the center of moderately rich clusters of galaxies ( yates et al .
1989 , hill & lilly 1991 ) with a dense hot intracluster medium ( garrington & conway 1991 ) .
in addition , x ray data are consistent with a scenario in which more than two third of the 50 , or so , nearby brightest clusters have cooling flows ( fabian 1994 ) . as a consequence in a number of powerful radio galaxies the thermal emission from the intracluster medium
is expected to dominate the observed soft x ray flux , as in cyg a ( reynolds & fabian 1995 ) , producing extended x ray structures and shadowing any other contribution .
this scenario is further complicated by the fact that deep rosat hri observations of nearby radio galaxies have revealed x ray deficit associated with the radio lobes possibly originating from the hydrodynamical interaction of the jets and radio hot
spots with the intracluster medium ( cyg.a : carilli et al .
1994 ; ngc 1275 : bhringer et al . 1993 ; 3c 449 : hardcastle et al .
1998b ) .
we present here a deep rosat hri observation of 3c 219 , a nearby ( z=0.1744 ) powerful radio source identified with a cd galaxy of magnitude m@xmath6=-21.4 ( taylor et al.1996 ) , belonging to a non abell cluster ( burbidge & crowne 1979 ) . despite of its cluster membership ,
the rosat pspc and asca archive data suggest that thermal emission , if present , is negligible so that this source could be a good candidate to detect possible ic contribution to the x ray flux . the radio structure is well studied ( perley et al .
1980 , bridle et al .
1986 , clarke et al .
1992 ) : it is a classical double lobed frii radio galaxy that spreads over @xmath1 180on the sky plane corresponding to a projected size @xmath1 460 kpc.km s@xmath7mpc@xmath7 and @xmath8 are assumed throughout ] a strong jet extends over 20 ( @xmath1 50 kpc ) through the south western lobe , while a weak counterjet is visible in the north
eastern lobe . the total radio spectrum , largely dominated by the extended radio structure , between 178 and 750 mhz is @xmath9 ( @xmath10 , laing et al . 1983 ) .
fabbiano et al .
( 1986 ) have reported the detection of a broad paschen@xmath11 line in excess of the predictions of case b recombination implying the presence of an absorbed broad line region with @xmath12 mag .
more recent optical studies have shown that the 3c 219 spectrum is dominated by a starlight continuum , but a non stellar component and emission line features are also present ( lawrence et al .
1996 ) ; from the broad line ratio @xmath13 , assuming typical broad line region parameters , we infer @xmath14 . in a h@xmath11+[n ii ] image taken with the 4 m kitt peak telescope the galaxy appears to be point like ( baum et al .
1988 ) with a nearby , possibly interacting companion @xmath15 to the south east .
an image of 3c 219 has been taken with the hst telescope s wfpc2 through a broadband red ( f702w ) filter ( de koff et al .
1996 ) ; the galaxy is well resolved and does not present relevant features ( dusty lane , distortions ) on a scale @xmath16 pc ( i.e. hst resolution ) .
the data analysis is presented in sect . 2 and 3 while the proposed interpretation is discussed in sect.4 .
lllllllll instrument & en . range & n@xmath17 & @xmath18 & n@xmath19 & @xmath20 & @xmath21 & @xmath22 & @xmath23 + rosat & 0.12.4 kev & @xmath24 & @xmath25 & & & & & 7.6/10 + & & 1.55(f ) & @xmath26 & & & & & 7.7/11 + & & 1.55(f ) & 1.80 ( f ) & 18.7@xmath27 & [email protected] & & & 9.9/10 + asca & 0.610 kev & @xmath29 & @xmath30 & & & & & 298.5/283 + & & 1.55(f ) & @xmath31 & & & & & 359.2/284 + & & @xmath32 & @xmath33 & & & 6.4(f ) & @xmath34 224 & 295.8/285 + rosat+asca & 0.110 kev & 1.55(f ) & @xmath35 & @xmath36 & [email protected] & & & 304.8/293 + @xmath37 equivalent hydrogen column density ( units of @xmath38 @xmath39 ) + @xmath40 photon spectral index + @xmath41 column density of the partial covering gas ( @xmath38 @xmath39 ) + @xmath42 covering fraction + @xmath43 iron line energy ( kev ) + @xmath44 iron line ew ( ev ) + @xmath45 total @xmath46 and degrees of freedom 3c 219 has been observed with the rosat pspc ( pfeffermann et al .
1986 ) on may 3 , 1992 .
we have analyzed the archive data finding that the spatial profile in the 0.12.4 kev range is consistent with that of a point source convolved with the pspc point spread function ( psf ) , taking into account the source spectral properties and the background level .
however , the maximum likelihood detection algorithm suggests some possible evidence of extended emission ( see sect.3 ) . there is no evidence of significant time variability .
the source spectrum was extracted from a circular region with 2radius .
different background spectra have been extracted either from annuli centered on the source or from circular regions uncontaminated by nearby sources . in all cases
the background appears stable without any appreciable variations within the statistical errors .
corrections were included for vignetting and pspc dead time .
source spectra were extracted in the pulse invariant ( pi ) channels in the range 11240 ( @xmath47 0.12.4 kev ) .
the photon event files were analysed using the exsas / midas software ( version 94nov , zimmermann et al .
1993 ) and the extracted spectra were analysed using version 9.0 of xspec ( shafer et al .
1991 ) with the appropriate response matrix .
the resulting exposure time is 4349 s and the background subtracted source count rate is 0.127 @xmath28 0.006 cts s@xmath7 .
the source spectrum was rebinned in order to obtain a significant signal to noise ratio ( @xmath48 ) for each bin and fitted either with a power law or with a thermal plasma spectrum ( raymond smith model ) plus absorption arising from cold material with solar abundance ( morrison & mccammon 1983 ) .
the derived spectral parameters are given in table 1 , where the reported errors are at 90% confidence level ( lampton et al .
1976 ) ; the value of the galactic column density towards 3c 219 , @xmath49 @xmath39 , has been retrieved from the 21 cm radio survey of dickey & lockman ( 1990 ) .
a single power law with the absorption either fixed at the galactic value , or free to vary , provides an acceptable description of the observed spectrum ( table 1 ) .
formally , also a raymond
smith thermal model with solar abundance fits well the observed counts , however the derived temperature is extremely high and almost unconstrained .
the best fit spectral photon index is unusually flat in agreement with the results of prieto ( 1996 ) .
the flat power law slope can be due to the effect of absorption on a steeper continuum . in order to test this possibility
we have tried a partial covering model . assuming a typical agn spectrum with photon index @xmath50=1.8 plus galactic absorption , a good description of the data can be obtained if about 75% of the nuclear radiation is absorbed by a column density of the order of @xmath1 2 @xmath51 @xmath39 ( table 1 ) .
the x ray flux in the 0.12.4 kev energy range , corrected for galactic absorption , is 1.8 @xmath52 10@xmath53 ergs @xmath39 s@xmath7 corresponding to an isotropic luminosity of 1.1 @xmath52 10@xmath54 ergs s@xmath7 ( rest frame ) , while the unabsorbed nuclear luminosity would be @xmath55erg s@xmath7 .
3c 219 was observed with asca ( tanaka , et al .
1994 ) on april 12 , 1995 and on november 13 , 1994 with the gas imaging spectrometers ( gis2/gis3 ) and with the solid state imaging spectrometers ( sis0/sis1 ) .
both observations were analyzed by us using standard calibration and data reduction methods ( ftools ) provided by the asca guest observer facility at goddard space flight center .
the net exposure time for the 1995 observation was 18 ks in the gis detectors and 16.5 ks in the sis .
slightly lower exposure times were obtained for the 1994 observation .
source photons were extracted from a circular region centered on the source with 6@xmath56 radius for gis and 3@xmath56 radius for sis .
the background was estimated using both source - free regions and blank sky gis and sis observations available in the calibration area .
different background estimates give consistent results .
data preparation and spectral analysis were performed using version 1.3 of the xselect package and version 9.0 of the xspec program .
the light curves from each instrument do not show any significant flux variation over the whole observation .
gis and sis spectra were binned with more than 20 cts / bin in the 0.710 kev and 0.610 kev energy ranges respectively .
the lowest sis energy channels have been excluded because of the uncertain calibrations ( dotani et al .
1996 ; cappi et al .
since the spectral parameters obtained by fitting the four detectors separately were all consistent within the errors , data from both pair of sis and gis were fitted simultaneously to the same model , but with the normalization of each dataset allowed to vary relative to the others in order to account for the small discrepancies in the absolute flux calibrations of the detectors .
both observations gave very similar results ; in the following we discuss only the results obtained from the 1995 observation characterized by a better counting statistics .
a single power law model clearly provides an acceptable fit to the data ( table 1 ) .
absorption in excess of the galactic value is required .
there is no need of more complex models .
in particular the addition of a narrow emission line at 6.4 kev does not improve the quality of the fit , while a thermal component is not required by the data .
the absorption corrected x ray flux in the 210 kev energy band is @xmath1 3.6 @xmath57 erg @xmath39 s@xmath7 , which corresponds to a luminosity of @xmath1 2.4 @xmath58 erg s@xmath7 in the source frame . in the overlapping 0.62.0 kev energy range the observed pspc flux is about 15 % lower than the asca flux with a weak dependence on the assumed spectral parameters .
moreover the value of the column density derived from the pspc data assuming a partial covering model ( table 1 ) is in good agreement with the absorption observed in the asca data ( table 1 ) suggesting that this model is viable or that two different spectral components are present .
therefore , in order to make full use of the available spectral band ( 0.110 kev ) we have performed joint fits to the rosat pspc and asca data leaving the relative normalizations free to vary to account for the residual absolute flux uncertainties among the different instruments .
not surprisingly a partial covering model provides a good description of the observed 0.110 kev spectrum : the resulting power law spectrum ( @xmath59 ) is partially ( @xmath60 % ) absorbed by cold gas with a column density @xmath61 @xmath39 ( fig.1 , table 1 ) .
this implies that in the soft 0.12.4 kev band about 38% of the observed flux may be due to a scattered component . a description of the soft x
ray spectrum in terms of thermal emission is not viable , while the addition of a thermal component to the best fit partial covering model does not significantly improve the fit .
any thermal component ( if present ) can not contribute more than 810 % ( 90 % confidence limit ) to the observed x ray flux .
we conclude that the 0.110 kev spectrum of the radio galaxy 3c 219 can be explained in terms of an obscured central source characterized by a power law with a slope typical of radio loud quasars in the 210 kev energy range ( lawson & turner 1997 ) .
an additional unabsorbed spectral component is present in the soft x ray band . in principle
this component may be due to scattering of the soft x ray nuclear radiation by circumnuclear clouds or thermal electrons , but we note that the unabsorbed power law ( table 1 ) has a slope close to that of the synchrotron radio spectrum ( @xmath9 ) , so that the possibility that it might originate from ic scattering of the ir optical nuclear and/or cmb radiation can also be taken into account .
a crucial test to decide which one of these two possibilities should be adopted is the study of the spatial distribution of the x ray flux .
3c 219 was observed with the rosat hri for a total of 28.6 ks between april 10 and may 8 , 1997 .
standard procedures have been employed within the midas / exsas software .
the source counts , estimated with a maximum likelyhood ( ml ) detection algorithm as described in cruddace et al . ( 1988 ) , are 847@xmath2830 corresponding to a background subtracted count rate of @xmath1 0.03 cts s@xmath7 in the 0.12.4 kev band . since the ml algorithm has been conceived to study point like sources and is not efficient for extended sources , the source flux has been computed measuring the counts in a circle of 2radius .
we find evidence of emission outside 20from the nucleus accounting for some 15% of the total flux ( see sect.4 ) .
the resulting total count rate ( nuclear and extended emission ) is @xmath1 0.035 cts s@xmath7 .
the total hri flux of @xmath1 1.6 @xmath57 erg @xmath39 s@xmath7 is consistent , within about 10% , with that derived from the pspc observation ( see @xmath62 2.1 ) assuming the spectral parameters reported in table 1 .
it is known that the intrinsic spatial resolution of the rosat hri of @xmath15fwhm is blurred by the errors due to the relatively poor knowledge of the pointing position as a function of time .
we have tried to correct for the residual aspect solution errors following a procedure developed by harris et al.(1998 ) which selects observing periods with the same roll angle and folds the data according to the wobbling period of 402 s. several subimages are created according to the source statistics and the wobble period , then shifted to a common center and coadded .
the radial profile of the innermost region of the wobble corrected image is then compared with an inflight calibrated psf , kindly provided by i. lehmann ( private communication ) , derived from the average profile of 21 bright stars ( fig.2 ) . from the analysis of the azimuthal distribution of the counts within @xmath120(@xmath1 50 kpc ) from the source peak intensity we find that the nuclear source is resolved in the north south direction ( fig.2 ) . a kolmogorov
smirnov test shows that the observed distribution of the counts in the north south direction differs from the psf at the 99% confidence level , while the test is not conclusive in the case of the east west direction . in order to get more information on the structure an x ray image of 3c 219 , smoothed with a circular gaussian of fwhm=3(fig.3 ) , has been produced with the exsas package , while the further image processing has been performed with the astronomical image processing system ( aips ) package .
an extended structure up to @xmath115 in radius is indeed detected in the north south direction ( fig.3 ) .
the smoothed psf appears to be elliptical ( fwhm= [email protected] ) with the major axis positioned at @xmath170@xmath63 clockwise from the north , at variance with the elongation of the extended structure . of course
, the small fwhm used in the previous smoothing procedure does not allow a detection of the outer extended low brightness features .
the x - ray image of fig.4 has been produced with the exsas package by binning the photon event table in pixels of 2and by smoothing the map with a circular gaussian of @xmath64 = 5 . with these values
the signal to noise ratio is good enough to image the low brightness extended emission significantly above the background .
the rosat hri image has been shifted by 1.5 and 0.6 in right ascension and declination , respectively , in order to align the x - ray peak with the radio core position .
the x ray map has been overlaid to the optical poss
ii digitized image of the field in fig.4 . since 3c 219 is in a cluster of galaxies , the optical image is crowded and a number of objects fall within the x ray structure ; no coincidence with relevant optical objects is found .
the structure of the x - ray brightness distribution between @xmath11525 from the core can be enhanced if the central unresolved component is subtracted from the hri image .
the model used in the subtraction is a circular gaussian ( @xmath64 = 6 ) with the peak coincident with the x - ray peak .
this function represents the convolution between the instrument psf and the smoothing function applied to the image .
in fact , although the instrument psf is not exactly a gaussian , the smoothing gaussian dominates the resulting convolved psf . in order to subtract the nuclear component from the total hri image we have made use of the spectral analysis results of sect.2 .
the 0.12.4 kev flux of the nuclear emission has been estimated by assuming the best fit spectral parameters of the high energy absorbed component in the 0.110 kev fits .
taking into account the different responses of the hri , pspc and asca detectors for a given spectrum , as well as the errors on the best fit spectral parameters , the nuclear flux is estimated to account from 55% to 70% of the hri counts .
the subtraction procedure has been performed several times constraining the amplitude of the nuclear component within this interval .
we also note that the largest contribution of the point like nuclear source ( leaving zero counts after the subtraction at the position of the peak ) is @xmath65% of the total net counts ; this limit is implied from the hri data alone .
the residual map resulting from the subtraction of the nuclear source with a representative amplitude of 64% is shown in fig.5 superposed on the vla radio image at 1.4 ghz ( clarke et al .
1992 ) .
it is evident that the residual x
ray isophotes are strongly correlated with the radio extension and appear to be elongated along the radio structure .
we distinguish three main regions : one coincident with the low brightness part of the northern radio lobe ( n ) , one positioned between the nucleus and the southern radio hot
spot ( s ) and the strongest one ( c ) centered on the radio core .
the c component appears like a curious eight shaped figure with the axis inclined by @xmath66 with respect to radio axis , but the isophotes become more aligned with the radio axis with increasing distance from the nucleus .
this basic structure persists even by varying the amplitude of the point like source within the allowed interval and/or by applying a slightly different @xmath64 in the subtraction procedure .
furthermore one can notice that the c component seems to show a remarkable continuity with the extended structure of fig.3 .
the rosat hri observation presented here shows a complex extended x ray structure surrounding a central point like source .
rosat and asca spectral analysis reveals that the nucleus is absorbed by a column density @xmath67@xmath39 similarly to other rosat findings for blrgs ( crawford & fabian 1995 ) .
the unabsorbed luminosity ( @xmath68erg s@xmath7 ) in the 0.12.4 kev band and the photon index @xmath69 are typical of steep spectrum radio loud quasars , thus strengthening the hypothesis that we are dealing with a hidden quasar in the nucleus of the radio galaxy in agreement with the unification model . according to the spectral analysis
, the nuclear source contributes @xmath70% of the total rosat pspc flux while the contribution from a thermal source , if any , can not exceed @xmath71% of the total flux . therefore a large fraction , if not all , of the flux associated with the extended structure should be characterized by a power law spectrum with a slope close to that of the radio emission ( @xmath72 ) . this and the spatial correlation with the overall radio structure suggest that the ic process could be a likely mechanism to generate most of the extended x ray flux .
however , a considerable fraction ( @xmath73% ) of the extended emission flux is confined to the eight shaped structure ( component c in fig.5 ) whose brightness distribution , elongated in a direction at large angle with the radio axis and falling down radially roughly as @xmath74 , looks very much different from that expected by a simple ic with the cmb photons . as a consequence
we have first checked whether this feature could be due to a non spherically symmetric cooling flow .
we find that the addition of a cooling flow to an absorbed power law source may give an adequate representation of the combined rosat
asca spectrum only if a metallicity @xmath75 solar is assumed .
acceptable fits , but worse than those of the partial covering model ( sect.2 ) , have been obtained with an absorbed power law plus constant or non constant ( s=2 , mushotzky & szymkowiak 1988 ) pressure cooling flow models .
we derive intracluster gas densities between @xmath76 and , by applying a king s model , cluster luminosities @xmath77erg s@xmath7 in the 0.12.4 kev band , that is a factor 1560 larger than the upper limit to any thermal contribution within the 2rosat pspc extracted radius ( sect.2 ) .
furthermore , the cluster emission would give out x rays in excess of the observed hri background by a factor of @xmath11.56 at @xmath11distance from the nucleus .
we conclude that a cooling flow can not be responsible for the c component .
we propose that the main features of this component can be explained by the ic scattering of the ir optical radiation from a hidden quasar with the surrounding relativistic electrons of the radio source , according to the model developed by brunetti et al.(1997 ) .
optical emission of the nuclear source , as seen by the relativistic electrons in the radio lobes , is made of two components : the direct radiation from the quasar , for those electrons located within the emission cones ( assumed half opening angle @xmath78 ) , plus the reprocessed radiation from the dusty molecular torus surrounding the quasar . in the model
we have fixed the inclination of the radio axis with respect to the sky plane at @xmath79 , close to that inferred from the radio jet data ( bridle et al.1986 ) , while the direction of the torus axis ( @xmath80 ) is a free parameter constrained by the requirement that the nuclear source is not directly seen by the observer ( @xmath81@xmath39 ) and that the ic brightness distribution closely matches the observations ( fig.6 ) . as already stated in the introduction ,
a relevant parameter for the efficiency of the ic model is the ir radiation from the putative hidden quasar .
unfortunately 3c 219 was not detected in a pointed iras observation by impey & gregorini ( 1993 ) and it has not been observed with iso .
the ir spectral properties of the blrgs observed by iras are poorly known .
heckmann et al .
( 1994 ) found that the weighted iras 1086@xmath82 rest frame spectral index of a sample of 9 blrgs given by the superscampi procedure is @xmath83 .
however , the few detected blrgs show a large dispersion in the 2560@xmath82 spectral indices ranging from -0.47 to + 0.44 ( impey & gregorini 1993 , heckmann et al .
1994 , golombek et al .
1988 ) . by assuming an ir spectral shape consistent with the iras upper limits and with the range of observed blrg spectra and , also , with the predicted spectra of dusty tori models ( pier & krolik 1992 )
, the 6100@xmath82 isotropic ir luminosity of 3c 219 could be as large as @xmath84 erg s@xmath7 . to estimate the ir luminosity of the hidden quasar we first notice that the monochromatic @xmath85 luminosities of a sample of low redshift ( @xmath86 ) radio galaxies are on the average @xmath87 times smaller than those of a sample of quasars in the same redshift interval ( heckmann et al .
1992 ) . by applying pier & krolik ( 1992 ) models to the adopted geometrical configuration of 3c 219 we predict an ir luminosity a factor 37 lower than seen from a face on quasar , depending on the torus parameters .
if these ratios and the upper limits for the observed ir luminosity of 3c 219 are adopted , the 6100@xmath82 rest frame upper limit on the luminosity of the hidden quasar would be @xmath88erg s@xmath7 .
we have also tried to estimate the luminosity of the hidden quasar by following a different approach .
if a mean optical x - ray spectral index @xmath89 ( @xmath90 ; brunner et al . 1994 ) is adopted , then from the 0.12.4 kev rest frame luminosity of @xmath55 erg s@xmath7 ( @xmath91 ) of the unabsorbed nuclear source , and a typical optical spectral index @xmath92 ( richstone & schmidt 1980 ) , we derive a 350650@xmath93 rest frame luminosity @xmath94 erg s@xmath7 .
a similar value ( @xmath95erg s@xmath7 ) is obtained by making use of the correlations between the radio , optical and x ray powers of radio loud quasars ( browne & murphy 1987 , kembhavi 1993 ) . since heckmann et al .
( 1992,94 ) have shown that the 6100@xmath82 rest frame luminosity of radio loud quasars is a factor 68 larger than the 350650@xmath93 rest frame optical luminosity , we find an expected 6100@xmath82 luminosity @xmath96erg s@xmath7 , larger but close to the upper limit derived in the first estimate previously discussed .
as a consequence in the ic model described below we will adopt a 6100@xmath82 luminosity @xmath97erg s@xmath7 for the 3c 219 hidden quasar , which corresponds to an ir optical 1000.35@xmath82 rest frame luminosity of @xmath98erg s@xmath7 if typical ir optical spectral parameters are assumed ( brunetti et al .
1997 ) . in our model computations ,
we assume an emission pattern of the reprocessed ir radiation close to that predicted by the theoretical models of pier & krolik ( 1992 ) . it should be stressed , however , that , due to the smoothing made by the psf of the rosat hri , a precise knowledge of this pattern is not crucial . a possible additional beamed ir emission ( he s et al .
1995 ) has not been taken into account , but being it directional anyway it would affect the calculation only in a small fraction of the radio volume .
in order to get some insight into the spatial distribution of the relativistic particles , we have constructed a 3d contour of the radio
volume by making use of the weakest isophote in the 1.4ghz vla map of clarke et al .
the radio galaxy volume is assumed to be symmetric around the line joining the mid points of the weakest radio isophote at each fixed distance from the nucleus .
we have deprojected the structure with the inclination angle @xmath79 and obtained a virtual 3d model of the radio galaxy .
obviously , this is a rather crude approximation as might be indicated by the complex structure of the radio isophotes . given the spectral properties of the hidden quasar and those of the relativistic electrons , a numerical code computes the total ic soft x
ray luminosity and brightness distribution projected on the plane of the sky . in order to obtain a map to be compared with that derived from the observations
, the contribution from the ic scattering of the cmb photons and the observed background level were added to the theoretical matrix , the brightness distribution was convolved with the rosat hri psf and smoothed with the same gaussian function used for the data ( sect.3 ) . as a first approximation we consider an uniform distribution of the relativistic particles and an electron spectrum extended to lower energies with the slope indicated by the radio spectral index , not modified by radiative and adiabatic losses .
one can distinguish two regions in the x ray brightness distribution of the model ( fig.7 ) : at small distances from the nucleus ( comparable with the minor axis of the radio galaxy ) the x ray emission distribution depends mainly on the nuclear radiation field , the x ray axis being that of the quasar illumination cone ; at larger distances from the nucleus the x ray distribution is mainly determined by the distribution of the relativistic particles and the x ray emission tends to be more and more aligned with the radio structure .
the ic x ray flux from the far lobe can be considerably larger than that from the near one depending on the inclination of the radio axis on the sky plane . as shown by brunetti et al.(1997 ) in the case of ellipsoidal radio galaxies , the predicted ratio for an inclination @xmath79 can range up to @xmath87 , depending on the relative contribution due to the ic scattering of the cmb photons and on the luminosity of the hidden quasar .
this may explain the larger extension of the observed brightness distribution toward the northern lobe . at larger distances from the nucleus , where the nuclear photon energy density becomes lower than that of the cmb ,
the ic with the cmb dominates the x ray flux and the resulting x ray brightness distribution becomes aligned with the radio axis .
the model reproduces fairly well the observed brightness in the extended inner regions where the ic emission is dominated by the scattering of the radiation from the nuclear source ( fig.5 and fig.7 ) .
the required inclination of the dusty torus is such that its axis makes an angle of @xmath99 with the radio axis and @xmath100 with the line of sight .
large tilts between the radio and dusty lane / torus axis for a number of radio galaxies have been discovered with ground based telescopes ( mllenhoff et al . 1992 ) and more recently with hst observations ( ford et al .
1994 , de juan et al .
it also follows that the distribution of the relativistic particles within the c component is uniform , at least on the smoothing scale ( @xmath71 kpc ) .
the required relativistic electron density can be compared with that derived by the minimum energy ( equipartition ) argument . since relativistic electrons with energies much lower ( @xmath101 ) than those of the synchrotron ( radio ) electrons dominate our ic model ,
we apply the equipartition equations given by brunetti et al .
( 1997 ) for a fixed low energy cut off in the electron spectrum ( in this paper @xmath102 and @xmath103 ) .
the equipartition magnetic field strength evaluated by assuming the minimum energy condition over all the radio volume and equal energy density between negatively and positively charged particles is @xmath104 g ( with standard equipartition formulae it would be @xmath105 g ) .
our model requires a density of relativistic particles such that @xmath106 g , i.e. 3.3 times smaller than the equipartition value .
the minimum energy hypothesis is not fulfilled , the energy in the particles being a factor @xmath110 larger than in the equipartition case . in the case of fornax
a , feigelson et al.(1995 ) also found a similar , although smaller , departure from the equipartition condition , while in the case of the radio galaxy pks 1343601 ( cen b ) , tashiro et al.(1998 ) find that the energy density of the relativistic particles ( @xmath107 ) , equally distributed between negative and positive charges , is a factor @xmath108 larger than that of the magnetic field . with the parameters given in tashiro et al .
paper , by including the energy contribution of mildly relativistic particles ( @xmath109 ) , we derive a ratio @xmath110 between particle and magnetic field energy densities , that is a value of the same order as that found in the case of 3c 219 .
the simple model discussed so far can not explain the rather complex structure seen in the external regions where the ic scattering of the quasar s photons becomes more and more negligible .
the ic scattering of the cmb photons may explain the observed features under the assumption that there are deviations from the assumed uniformity and simple 3d geometry of the spatial distribution of the relativistic particles .
in general , variations of a factor 22.5 in the relativistic electron column densities in excess of those of our model would be sufficient to give the observed x ray brightness distribution .
there is some evidence that this might be the case .
let us consider first the n component .
both the radio brightness ( fig.5 ) and the 1.45 ghz spectral index distributions ( clarke et al . 1992 ) indicate the possible presence of a flow of relativistic particles toward the e w direction from the northern hot spot to the n component , which coincides with a region of steeper radio spectral index . since at the position of the n
component our basic ic model predicts a number of counts @xmath1 3/4 that of the background ( of which @xmath1 20% from the ic scattering of the nuclear photons ) , the observed signal can be generated by an increase of a factor @xmath12.5 in the number of relativistic electrons .
we have tested this hypothesis under the assumption that the back flow might provide the required enhancement in the density of relativistic particles leaving the magnetic field strength unchanged , i.e. a factor @xmath13.3 lower than the equipartition value of sect.4.1 . by assuming an electron injection spectrum @xmath111 and radiative losses
, we find a synchrotron radio brightness of the western part of the north lobe and a 1.45 ghz spectral index @xmath112 both consistent with clarke et al.(1992 ) findings .
the implied age of the particles reservoir would be @xmath113 years .
( for these calculations we have used the synage package of murgia & fanti , 1996 ) .
let us consider now the s component .
our basic ic model predicts a number of counts , of which @xmath135% from the ic scattering of the nuclear photons , close to that of the background . here
an enhancement of a factor of 2 in the density of relativistic electrons would be sufficient to account for the observed x ray flux .
we notice that the s component lies in a region between the radio jet and the southern hot
spot where there is evidence of a systematic steepening of the radio spectral index and where the magnetic field lines surround a lobe of brighter radio emission , being perpendicular to the line joining the radio jet with the hot spot itself ( see the polarization map in clarke et al .
this may suggest a strong interaction between two relativistic plasmas , one of which a back flow from the hot spot .
we have also considered the possibility that the s and n components are of thermal origin .
the contribution of these components to the total x ray flux is @xmath115% and we know from the spectral analysis that at most @xmath110% of the total flux can be thermal ( @xmath114 kev ) .
thus it is possible that at least one of these components is of thermal origin . from the rotation measure ( rm ) and
depolarization maps clarke et al.(1992 ) find that an external clumpy medium is responsible for the observed rm and depolarization features .
for the sake of clarity we present in fig.8 the x ray brightness overlayed onto the depolarization map .
the coincidence of the s component with the depolarization structure transverse to the radio axis is striking , while a spotted distribution of moderate depolarization is also observed on the southern part of the n component . on the other hand
, we notice that the x ray isophotes of the northern radio lobe appear to carefully avoid regions of larger depolarization .
this is a somewhat contradictory result .
therefore , we have concentrated our attention on the s component only .
let us suppose that , according to the spectral analysis of sect.2 , the source is surrounded by a magnetized thermal plasma with a temperature @xmath471.5 kev and in pressure equilibrium with the relativistic plasma . in our non equipartition model
the pressure inside the radio lobes is @xmath115dyne @xmath39 and the resulting external gas density would be @xmath116@xmath117 . the strength of the s component entails an emission measure @xmath118 pc @xmath119 , that is a depth of the emission region @xmath16 kpc , much larger than the observed feature . the thermal model may be eased by assuming that the gas is highly clumped .
we find that a structure with an overall size @xmath4750 kpc , a filling factor of a few percentage points and clumps with a mean size @xmath47 1 kpc may account for the observed x ray intensity and depolarization ( @xmath120 ) .
a detailed model would require the knowledge of the rm structure function ( tribble , 1991 ) .
we tentatively conclude that a thermal origin of the s component can not be ruled out .
finally , we notice that the hot spots are not detected at the sensitivity level of our hri observation . while the binning and smoothing procedures used to enhance the hri image statistics give a convolved psf much larger than the southern hot spot dimension , thus depressing any upper limit on the x ray flux from the hot spot itself , the convolved psf is comparable with the northern hot spot extension . the northern hot spot has been resolved at 1.4 ghz with an angular size of 12 arcsec ( clarke et al .
1992 ) . from the analysis of the digitized 1.4 ghz map of 3c 219 we derive a radio flux @xmath47250 mjy contributed by the hot spot within a circular region of 90 arcsec@xmath121 , comparable with our x ray beam size , and
compute an equipartition ( @xmath102 ) magnetic field strength @xmath122 g .
the 2@xmath64 upper limit of the x ray brightness of the hot spot is @xmath123 erg s@xmath7 @xmath39 arcsec@xmath124 .
this allows us to set an upper limit to the density of the relativistic electrons compton scattering the cmb photons and a lower bound to the magnetic field of @xmath125 g , which is a factor of 4.8 lower than the equipartition value .
therefore , a deviation from the equipartition condition in the hot spot as well as in the lobes can not be ruled out .
a deep rosat hri x ray image of the frii radio galaxy 3c 219 shows a point like source , coincident with the nucleus of the galaxy and accounting for @xmath3474 % of the total net counts ( @xmath160 % of the total net counts by taking into account the spectral information ) , and an emission aligned with the radio structure that extends from the innermost region of the radio galaxy to the hundred kpc scale .
the 0.110 kev spectrum is well represented by a partial covering model in which @xmath1 74% of a power law spectrum ( @xmath126 ) is absorbed by a column density @xmath127 @xmath39 .
this spectral slope coincides with that of the synchrotron emission of the radio lobes .
the spectral analysis indicates that at most 10% of the x ray flux could be due to a thermalized gas with a temperature of @xmath11.5 kev .
the point like source , whose de absorbed ( isotropic ) luminosity in the 0.12.4 kev band is @xmath55erg s@xmath7 , can be identified with the emission of a quasar hidden in the nucleus of 3c 219 .
this lends further support to the unification of frii radio galaxies with radio loud quasars .
subtraction of the point like source leaves an extended , unabsorbed circumnuclear component ( named c ) whose luminosity in the 0.12.4 kev band is @xmath128erg s@xmath7 .
it appears unlikely that a cooling flow may explain this component because the x ray emission from the associated intracluster gas would exceed by far the observed counts . as a matter of fact
, we have no evidence of a diffuse hot intracluster gas . in agreement with the unification scheme
, we have constructed a simplified model in which the main features of the c component can be satisfactorily explained as ic scattering of the ir optical radiation from the hidden quasar , and surrounding dusty / molecular torus , with the relativistic electrons uniformly distributed in the radio lobes .
since the c component is significantly inclined with respect to the direction of the radio jet , our model implies a large tilt ( @xmath66 ) between the radio jet and the torus axis . because of the limits one can place on the luminosity of the hidden quasar , we find that the x ray flux of the c component can be accounted for if the equipartition condition is violated , the energy in the relativistic particles being at least a factor 10 larger than that in equipartition conditions .
the electrons mainly involved in the ic production of the soft x
rays have energies much lower ( @xmath129 ) than those of the electrons producing the synchrotron radio emission .
therefore , if our model is correct , it can provide useful constraints on the particle acceleration mechanisms and aging of the radio source .
the x ray structures ( named s and n components ) observed in the outer regions can not be explained by our model in which the relativistic particles are uniformly distributed within the radio lobes ( it should be borne in mind that we have consistently included in our computations the contribution from the ic scattering of the cmb photons , which is of minor importance for the c component , but dominant as one goes further out in the radio lobes ) . accounting for these structures by the ic process requires positive fluctuations of a factor 22.5 in the column densities of the relativistic electrons .
we tentatively associate this possibility with the evidence of back flows from the hot spots .
alternatively , we can not exclude the possibility of a thermal contribution from a hot ( kt @xmath1 1.5 kev ) clumpy gas surrounding the radio lobes in localized regions .
the presence of moderate depolarization and rm gradients from the radio maps supports this hypothesis in the case of the s component .
observations with axaf will provide an invaluable tool to verify our model and the nature of the x ray emission of 3c 219 .
we warmly thank ingo lehmann and gnther hasinger for their help in the wobble correction procedure for the analysis of the hri data , and gianni zamorani for helpful discussions .
we also thanks the referee for halpful comments .
the radio image has been retrieved by the web archive of the dragns by leahy , bridle and strom ( http://www.jb.man.ac.uk/atlas/ ) .
this research has made use of the nasa / ipac extragalactic database ( ned ) which is operating by the jet propulsion laboratory , california institute of technology under contract with the national areonautic and space administration .
this work has been partially supported by the italian space agency ( asi ars9670 ) .
allen w. , fabian a.c .
, 1992 , mnras 258,29p barthel p. d. , 1989 , apj 336 , 606 baum s.a . ,
heckman t.m . ,
bridle a. , van breugel w.j.m . ,
miley g.k . , 1988 , apjs 68 , 643 bhringer h. , voges w. , fabian a.c . ,
edge a.c . ,
neumann d.m . , 1993 , mnras 264l , 25 bridle a.h .
, perley r.a . ,
henriksen r.n . , 1986 , aj 92 , 534 browne i.w.a . ,
murphy d.w . , 1987 ,
mnras 226 , 601 brunetti g. , setti g. , comastri a. , 1997 , a&a 325 , 898 brunner h. , lamer g. , worrall d.m .
, staubert r. , 1994 , a&a 287 , 436 burbidge g. , crowne a.h . , 1979 , apjs 40 , 583 cappi m. , matsuoka m. , comastri a. , brinkmann w. , elvis m. , et al . , 1997 , apj 478 , 492 carilli c.l .
, perley r.a . ,
harris d.e . , 1994 , mnras 270 , 173 clarke d.a .
, bridle a.h . , burns j.o .
, perley r.a .
, norman m.l . , 1992 ,
apj 385 , 173 crawford c.s . , fabian a.c . , 1995 ,
mnras 273 , 827 crawford c.s .
, fabian a.c . , 1996 , mnras 282 , 1483 cruddace r.g . , hasinger g. , schmitt j.h.m.m . , 1988 ,
proceedings of : astronomy for large database : scientific objectives and methodological approaches , garching ( germany ) , 177 de koff s. , baum s.a . , sparks w.b .
, biretta j. , golombek d. , et al . , 1996 ,
apjs 107 , 621 de juan l. , colina l. , golombek d. , 1996 , a&a 305 , 776 dickey j.m .
, lockman f.j . , 1990 , ara&a 28 , 215 dotani t. , et al .
1996 , asca newsl . 4 , 3 fabbiano g. , trincheri g. , elvis m. , miller l. , longair m. , 1984 , apj 277 , 115 fabbiano g. , willner s.p . , carleton n.p .
, elvis m. , 1986 , apj 304 , l37 fabian a.c . , 1994 , ara&a 32 , 277 feigelson e. d. , laurent - muehleisen s. a. , kollgaard r. i. , fomalont e. b. , 1995 , apj 449 , l149 ford h.c . , harms r.j .
, tsvetanov z.i . ,
hartig g.f . ,
dressel l.l . , et al . , 1994 , apj 435l , 27 garrington s.t . , conway r.g . , 1991 ,
mnras 250 , 198 golombek d. , miley g.k .
, neugebauer g. , 1988 , aj 95 , 26 hardcastle m.j . , birkinshaw m. , worrall d.m . , 1998a , mnras 294 , 615 hardcastle m.j .
, worrall d.m . , birkinshaw m. , 1998b , astro
ph/9802175 harris d. e. , carilli c. l. , perley r. a. , 1994 , nature 367 , 713 harris d.e . , silverman j.d . , hasinger g. , lehmann i. , 1998 , submitted to a&a heckman t.m . , chambers k.c . , postman m. , 1992 , apj 391 , 39 heckman t.m .
, odea c. p. , baum s. a. , laurikainen e. , 1994 , apj 428 , 65 he s r. , barthel p.d .
, hoekstra h. , 1995 , a&a 303 , 8 hill g.j .
, lilly s.j . , 1991 , apj 367 , 1 kaneda h. , tashiro m. , ikebe y. , ishisaki y. , kubo h. , et al . , 1995 , apj 453 , l13 kembhavi a. , 1993 , mnras 264 , 683 impey c. , gregorini l. , 1993 , aj 105 , 853 laing r.a . , riley j.m .
, longair m.s . ,
1983 , mnras 204 , 151 lampton m. , margon b. , bowyer s. , 1976 , apj 208l , 177 lawrence c.r .
, zucker j.r . , readhead a.c.s . ,
unwin s.c .
, pearson t.j . , xu w. , 1996 , apjs 107 , 541 lawson a.j . ,
turner m.j.l . , 1997 ,
mnras 288 , 920 mllenhoff c. , hummel e. , bender r. , 1992 , a&a 255 , 35 morrison r. , mccammon d. , 1983 , apj 270 , 119 murgia m. , fanti r. , 1996 , internal report , ira 228/96 mushotzky r.f .
, szymkowiak a.e .
, in cooling flows in cluster and galaxies;proc . of the nato adv .
research workshop , cambridge , uk , june 2226 , 1987 .
dordrecht , netherlands , kluwer academic publishers , 1988 , 53 perley r.a . , bridle a.h . ,
willis a.g . ,
fomalont e.b .
, 1980 , aj 85 , 499 pfeffermann , e. , briel u.g .
, hippmann h. et al . , 1986 , in proc .
spie 733 , soft x - ray optics and technology , p. 519
, krolik j.h . , 1992 ,
apj 401 , 99 prieto m.a . , 1996 , mnras 282 , 421 reynolds c.s .
, fabian a.c . , 1996 ,
mnras 278 , 479 richstone d.o . , schmidt m. , 1980 , apj 235 , 361 shafer , r.a . ,
haberl , f. , arnaud , k.a . ,
tennant , a.f .
1991 , xspec user s guide , esa tm09 , noordwijk , the netherlands tanaka y. , inoue h. , holt s.s . , 1994 ,
pasj 46 , l37 tashiro m. , kaneda h. , makishima n. , iyomoto n. , idesawa e. , et al .
, 1998 , apj 499 , 713 taylor g.l . , dunlop j.s .
, hughes d.h . ,
robson e.i . , 1996 ,
mnras 283 , 930 tribble p.c . , 1991 ,
mnras 250 , 726 ueno s. , koyama k. , nishida m. , yamauchi s. , ward m.j . , 1994 , apj 431 , l1 urry c.m .
, padovani p. , 1995
, pasp 107 , 803 worrall d.m . ,
lawrence c.r . ,
pearson t.j . ,
readhead a.c.s . , 1994 ,
apj 420 , l17 yates m.g .
, miller l. , peacock j.a . , 1989 , mnras 240 , 129 zimmermann , h.u . , belloni , t. , izzo , c. , kahabka , p. , schwentker , o. 1993 , exsas user s guide , mpe report 244 | spectral analysis of rosat pspc and asca archive data and a recent rosat hri observation of the powerful frii radio galaxy 3c 219 reveal an absorbed point like source , coincident with the nucleus , and a non thermal extended component aligned with the radio structure .
the point like source can be readily interpreted as a hidden quasar in the nucleus , giving further support to the unification scheme of frii radio galaxies and radio loud quasars .
the coincidence between the x ray ( 0.110 kev ) and radio spectral slopes suggest that most of the emission associated with the extended component is due to the inverse compton process in the radio lobes .
the extended circumnuclear emission can be understood as compton scattering of the ir optical radiation emitted by the hidden quasar and surrounding dusty / molecular torus .
this is the first observational evidence supporting the existence of this effect , which also probes the relativistic electron spectrum at energies much lower than those involved in the synchrotron radio emission .
the observed x ray flux can be matched by assuming that the energy density of the relativistic particles exceed the equipartition value by about a factor of 10 . at larger distances from the nucleus ( @xmath0kpc ) the inverse compton scattering with the cosmic microwave background photons
becomes more important and may explain the observed x
ray features if positive fluctuations in the column densities of relativistic electrons are present .
alternatively , one can not exclude with the present data a thermal contribution by hot clumpy gas surrounding the radio lobes . |
a wide range of astronomical evidence suggests the existence of a gravitationally - interacting , non - luminous component of the universe which has yet to be identified .
this dark matter ( dm ) component comprises approximately 27% of the energy density of the universe , and is responsible for the galactic structures visible today @xcite .
there are viable arguments that dm could be composed of weakly interacting massive particles ( wimp ) , and the darkside-50 ( ds ) experiment is a search for wimp - nuclear collisions as a fixed detector on earth moves through the wimp field of the milky way .
such interactions are rare , and impart usable recoil energies of @xmath0kev to a nucleus @xcite .
thus to reach the required sensitivities , detectors having large target masses and background suppression are needed .
a number of technologies are now used for direct detection of dark matter wimps .
these include ; cryogenic bolometers with ionization or scintillation detection , sodium / cesium iodide scintillation detectors , bubble chambers , a point contact germanium detector , and liquid noble gas detectors filled with xenon or argon .
+ the heart of the darkside-50 ( ds ) experiment is a 50 kg time projection chamber ( tpc ) which is a two - phase liquid argon ( lar ) detector . the ds system is designed to achieve background - free operation for at least 0.1 tonne - year of exposure .
the tpc exploits the extraordinary pulse shape discrimination ( psd ) of lar @xcite providing powerful rejection of electromagnetic ( electron or gamma induced ) signals .
cosmogenic background is suppressed to very low levels by mounting ds at a 3800 mwe ( water equivalent ) depth at the laboratori nazionali del gran sasso ( lngs ) laboratory @xcite @xcite , and surrounding the tpc with two nested veto shields of high purity water ( ctf ) and liquid scintillator ( lsv ) , figure [ dsexp ] .
in addition , local radioactive background is significantly reduced by efficient chemical and cryogenic purification methods @xcite .
the experiment uses argon obtained from gas wells ( uar ) which is depleted in the radioactive @xmath1ar isotope @xcite . despite the small detector size , ds has a sensitivity to wimp - nucleon cross sections of approximately @xmath2 @xmath3 for a wimp mass of @xmath4 gev / c@xmath5 .
+ this paper describes the ds electronics , the data collection and event - building systems ( daq ) , and the slow control and data monitoring systems of the experiment .
darkside-50 consists of three embedded detector systems .
viewed from the inner to the outer these systems are : a lar tpc , a liquid scintillator system ( lsv ) composed of 50% ppo ( 2,5-diphenyloxazole ) + 50% tmb ( trimethylbenzene ) @xcite , and a water cherenkov veto system ( ctf ) which also acts to shield the lsv and tpc from local radioactive background . an enclosed clean room ( clr ) sits atop the ctf , and is used to insert the tpc , and to provide internal access to the vetoes when the water and scintillator are removed .
the clr isolates the tpc and its components from surface radioactivity , particular radon , during assembly . + electrical connections to the tpc , approximately 5 - 10 m in length , are made through vacuum - tight pipes from the tpc to a feed - through in a flange on the floor of the clr . to minimize noise from the pmt signals ,
the tpc front end electronics are located in the clr racks just above the flange , which is as close as possible to the tpc .
however , readout of the veto signals from the lsv and ctf are patched through the side of the ctf to a control room ( cor ) by cables some 40 m in length .
thus synchronization of these systems is one of the electronic challenges and is addressed below .
the tpc , shown in figure [ dstpc ] , is a cylindrical detector containing 50 kg of lar .
the inner cylindrical surface of the detector is a teflon reflector coated with wavelength shifter ( tetraphenyl butadiene ) @xcite .
the coating absorbs the 128 nm scintillation photons emitted by excitation in the lar , and re - emits visible photons which are viewed by thirty - eight hamamatsu low - background r11065 pmt @xcite .
the pmt are divided equally between the top and bottom surfaces of the cylindrical tpc , and receive the wavelength - shifted light through fused silica windows .
these windows are also coated with the wavelength shifter on the inner surfaces and have transparent conductive layers of indium tin oxide ( ito ) evaporated on both the inner and outer surfaces .
coating both surfaces with ito allows the inner window surface to serve as a grounded anode ( top ) and a high voltage cathode ( bottom ) of the tpc , while maintaining the outer surfaces at the average pmt photocathode potential .
the fused silica anode ( top ) window has a cylindrical rim extending downward into the tpc to form a 1 cm - thick gas pocket above the lar .
high voltage is applied between the tpc cathode and anode to produce , by field shaping rings , a uniform electric field throughout the active volume .
also , light from lar excitations ( s1 ) is collected by the pmt .
electrons from ionization drift upwards in the electric field , eventually reaching the liquid surface where they are extracted , and accelerated in the gas pocket producing secondary ionization light ( s2 ) .
the drift time of the ionization charge in the lar , and the amplitude of the light signal in the gas are adjustable by varying the potentials on the cathode and an extraction grid placed just below the lar surface .
this design not only provides three - dimensional information of the position of the event in the lar , but also allows additional discrimination between the ionization due to ar recoils and that of electromagnetic interactions , ( _ ie _ @xmath6 or @xmath7 events ) .
however , the experimentally applied discrimination mainly uses the ratio of the fast component of the ( s1 ) ionization signal ( the first 90 ns ) to its total integral @xcite .
thus , it is important to minimize electronic noise and use a low frequency cutoff to accurately determine the timing of the detected photo electrons .
+ as explained above , wavelength - shifted light from an event in the lar was viewed through fused silica windows by 38 low - background pmt .
although the pmt were not placed in the sensitive volume of the tpc , they were immersed in lar and operated at a temperature of approximately 87 k. after extensive testing , it was found that charge accumulation on internal components in the pmt induced erratic behavior .
this was mitigated by decreasing the tube voltage , reducing the pmt gain to a few @xmath8 .
however , the reduced gain required a local pre - amplifier operating at lar temperatures to drive the signals through the approximate 5 - 10 m of cable to the amplifiers in the clr without significant contribution from noise and dispersion .
both the tpc amplifiers and digitizers were located in racks in the clr , just above the ctf .
+ the cryogenic pre - amplifier is mounted directly on the pmt base as shown figure [ preamp_fig ] .
it was constructed using discrete components as illustrated in figure [ preamp_circuit ] .
power - loss density is important as the pre - amplifier is mounted on the pmt base which is immersed in lar .
local boiling can cause a multitude of problems in high voltages , electronics , and tpc operations . consequently the pre - amplifier is designed with a total power consumption of less than 90 mw / channel .
it is also constructed to be as radio - pure as possible , which required the use of a cirlex @xcite circuit board , low radioactive components , and the avoidance of copper - beryllium ( cube ) connectors .
its input impedance is determined by the internal stray - capacitance of the pmt and the optimization of signal - to - noise by waveform shaping .
the relatively high - load resistance produces a passive gain with respect to the pmt coupling and is back terminated by 50 @xmath9 .
the active gain of the pre - amplifier is 3 v / v , for a total 24v / v compared to the output without a pre - amplifier .
the maximum output is 3v providing a dynamic range in excess of 1500 photoelectrons ( pe ) .
pole - zero cancellation is applied to the output signals sent to the amplifier .
high quality , double shielded cables @xcite are used to transmit these signals from the tpc to the amplifiers in the clr .
+ special care was taken to maintain high signal integrity during the implementation .
thus , sma rf connectors and double shielded signal cables are used where ever possible .
the cryogenic pre - amplifiers used surface mounted components with a grounded , full - copper backplane . the power supply was carefully filtered to reduce noise components .
impedance matching and signal reflection of the output signals were also adjusted to minimize cross - talk and electromagnetic interference ( emi ) .
+ while the hermetically sealed cryostat containing the pre - amplifiers should provide shielding from external noise sources , penetrations by signal , test , and power cables conveyed external noise to the pre - amplifiers .
this is exacerbated by the requirement that the pmt anodes were maintained at ground potential in order to use dc signal coupling , so that the metal envelope of the pmt remained floating at the cathode potential and could not be easily shielded .
consequently interference pick - up by the pmt dynodes becomes impressed on the photocurrents prior to amplification .
given the above issues , noise abatement required simultaneously considering the full detector system , including the cryogenetic and the slow control components , as well as the tpc electronics .
it was found that pump - drivers , switching power supplies , telecommunication transceivers , and high voltage supplies all provided contribution to the signal noise .
these were addressed in situ after installation .
+ pump motor controllers were especially noisy sources because most sent unfiltered and un - shielded signals to the appliance they controlled .
these sources of noise were mitigated by variac - based sinusoidal motor controls .
noisy switching power supplies were identified and individually treated by filtering or replacing the supplies with linear components .
in particular , noise from the pmt high voltage ( hv ) supplies was reduced by adding filters to the hv output .
+ ground loops were also mitigated by floating all connections through the cryostat wall . at the same time , all cable shields were connected via low - inductance ac coupling to the cryostat surface , to reduce the rf energy which would otherwise enter the cryostat . as a result , the noise at the output of the amplifier chain is dominated by the pre - amplifier electronic noise floor .
thus there are no noise - dominant spectral components within the frequency range of interest .
the output noise is 70 @xmath10v for an active bandwidth of 150 mhz .
+ a picture of the cryogenic pre - amplifier mounted of the pmt base the circuit diagram of the cryogenic pre - amplifier the amplifier in the clr provides further amplification and signal conditioning .
the coupling between the pre - amplifier and the amplifier boards is illustrated in figure [ fe_coupling ] .
one amplifier module handles five channels and provides the following outputs for each channel . + * an amplified output with gain of 10 v / v and an input noise equivalent of 20 @xmath10v over a bandwidth of 200 mhz .
this output is typically used for monitoring the performance of specific channels * a shaped output with a gain of 10 v / v with a bandwidth of 90 mhz and a noise equivalent of 15 @xmath10v .
the bandwidth is matched to the sampling frequency of the caen v1720 digitizer which samples at 250 ms / s * an attenuated output with gain of 0.5 v / v with bandwidth of 40 mhz .
the bandwidth of this output is matched to the sampling frequency of the caen 1724 digitizer sampling at 100 ms / s * discriminated low - voltage - differential - signals ( lvds ) , with a minimum pulse duration of 4 ns .
the discrimination threshold is remotely programmable .
these signals are sent to the trigger unit , and to scalars for monitoring all outputs are tuned to match the -2 to 1 v dynamic range of the digitizers .
the typical output noise is approximately 1 mv , 50% of which is due to the digitizer .
the @xmath1110 gain signals are sent to caen v1720 250 mhz 12 bit waveform digitizers @xcite and the low gain signals sent to caen v1724 100 mhz 14 bit digitizers .
both digitizer types are one unit 6u vme64x vme modules , servicing eight channels .
they have circular memory buffers of 10 ms / ch , and are locally fpga compliant .
the use of two digitizer types extends the dynamic range , providing a linear response between 1 and 10,000 pe , capturing signals without saturation from both lar and ar gas .
the large dynamic range also improves calibration options and provides a more detailed study of backgrounds .
+ the multiple adcs are synchronized with a common 50 mhz clock and external trigger .
they run in parallel and require three common control signals : a `` trigger '' , a `` run - enable '' , and a `` clock '' .
the trigger and run - enable can be either ttl or nim level signals , however the clock signals are differential lvds .
the distribution of all signals requires equal delays in a star - like topology in order to keep timing synchronization . +
as described above , there are two embedded veto systems , a water cherenkov detector ( ctf ) , and a liquid scintillator detector ( lsv ) .
the effectiveness of these veto systems is described in @xcite and included references .
the ctf is an active veto , tagging penetrating cosmic muons and cosmogenically produced charged showers in materials surrounding ds and shielding the tpc and lsv from external radioactivity .
the ctf is a cylindrical , steel tank , 11 m in diameter and 10 m high , filled with 1000 tons of high - purity water .
it is instrumented along the floor and bottom half of the cylindrical wall by 80 etl-9351 8 inch pmt @xcite .
studies revealed little spatial dependence of the pmt placement , so they were more - or - less equally spaced .
the walls are covered with reflecting @xcite tyvek sheets .
the veto electronics for both the lsv and ctf are located in the cor .
+ the lsv is a 4 m diameter , stainless steel sphere which lies inside the ctf and encloses the tpc . it is filled with liquid scintillator and instrumented with 110 hammatsu r5912-hqe - lri 8 inch pmt @xcite .
the interior walls of the sphere are covered with reflecting lumirror sheets @xcite .
the lsv is designed to moderate , capture , and thus provide a veto signal for neutrons which might enter , or exit , the tpc .
neutrons can be thermalized by scattering from protons in the scintillator liquid , and are efficiently captured by @xmath12b nuclei .
capture on @xmath12b proceeds to the @xmath13li ground state ( gs ) and a @xmath13li excited state with branching ratios of 6.4 % and 93.6 % , respectively .
the ground state decays by emitting a @xmath14kev @xmath15 particle , and the excited state decays with emission of a @xmath16kev @xmath15 particle and a 478 kev @xmath7 .
+ the measured lsv light yield is @xmath17pe / mev .
scintillation light produced by the nuclear products of the @xmath13li(gs ) decay is quenched and is expected to have a beta - equivalent energy between 60 to 60 kev corresponding to 25 - 30 pe .
neutrons can also capture on hydrogen with the emission of a @xmath18 mev @xmath7-ray yielding approximately @xmath19 pe .
+ most photons are collected in about 100 ns , but the tail of the distribution extends some 300 ns , and must be considered when designing the appropriate dynamic range of the acquisition electronics .
for example , the average number of photons per channel in an event with energy @xmath20kev is typically less than one , but can increase by more than an order of magnitude when a neutron is captured near a pmt .
in addition , a muon crossing the lsv produces a large signal corresponding to approximately 2 mev per transit - length in the scintillator .
thus to correctly determine the path of a muon , the veto electronics must be able to efficiently detect a signal as small as 0.3 pe , but should gracefully handle signals which saturate the front end electronics .
although the most important events occur at low energy , the system should recover from saturation in less than 2 @xmath10s .
+ good timing is also important in order to obtain track reconstruction .
this requires knowledge and stability of the relative time differences between pmts for both vetoes @xcite .
therefore time alignment and synchronization among channels must have a precision of @xmath21 0.3 ns , and remain stable for at least a 24 hour period .
+ output signals from the veto pmts are ac coupled directly to amplifiers through approximately 40 m of low dispersion , 50 @xmath9 , coaxial cable .
the front end analog boards ( feab ) and front end digital boards ( fedb ) are designed to match the input timing , noise , and impedance of the pmt signals , but more importantly they match the bandwidth of the waveform digitizers .
one pcb ( board ) contains 16 channels arranged in 2 rows of 8 channels , and has 8 layers with dimensions 274@xmath11274 mm .
figure [ veto_circuit ] is a schematic diagram of signal flow from the pmt through the feab to the fedb .
+ the feab configuration uses 3 texas instruments operational amplifiers per channel @xcite , one with a gain of 1 and two with a gain of 10 .
the x10 signals are split providing digitizer inputs , monitors , discriminated outputs , and a summed signal for 16-channels .
the circuit has 230 mhz bandwidth and a total noise of 200 @xmath10v@xmath22 . a linear power supply , able to deliver 15 a with a dual voltage @xmath237.5 v ,
was constructed to mitigate signal noise induced by the power source .
discrimination uses a dual , voltage comparator @xcite with differential output .
the signals are sent by lvds to the digital board .
the discrimination threshold is set using a standard by i@xmath5c bus with an 8 bit dac .
the rear panel of the feab has 16 hv connectors which connect the pmt cables directly to the caen a1536 high voltage boards @xcite , and these are housed in a caen mainframe @xcite .
+ the feabs are controlled by a fedb .
these boards contain a xilinx spartan 6 fpga @xcite which is used to generate a veto trigger .
the trigger firmware is constructed from logical patterns in the discriminator inputs from the feab .
these input signals arrive at a mezzanine board on the fedb where they are duplicated , one sent to a time to digital converter ( tdc ) and the other converted to the single - ended ttl standard of the fpga . a master clock signal is used for global timing .
the fpga is interfaced to a pic-32 micro - controller ( @xmath10c ) @xcite via an i / o bus controlled by the @xmath10c .
both the fpga and @xmath10c firmware were developed for the project .
+ some slow control operations are also handled by the fedb , and these are managed by the @xmath10c using ethernet and fpga connections .
the @xmath10c runs a single process which obtains an ip address from the lan dchp and acts as a command interpreter , waiting for incoming socket connections on a specific port .
in addition to resolving commands , the @xmath10c dialogues via i@xmath5c to set and read the 12 bit channel input offset - compensations and the 8 bit discriminator thresholds .
it also controls an alphanumeric display via an i@xmath5c bus to display data and status of a specific channel .
+ the pmt signals from both the ctf and lsv are amplified by the feab boards and their waveforms are digitized by commercial national instruments pxie-5162 modules @xcite .
the x10 channels are placed in 4 pxie chassis , with two chassis used for the ctf and two for the lsv .
each chassis houses analog inputs for the waveform digitizers ( 56 for the lsv , 40 for the ctf ) and connects to digital ttl lines for the trigger and synchronization interface .
the waveform is sampled at @xmath24 gs / s ( 800 ps / sample ) with 10 bit resolution ( 2 byte ) .
each waveform digitizer has 4 bnc input channels and is sampled asynchronously and continuously .
when a trigger is received , a block of samples within an acquisition window is stored in the digitizer memory buffer .
the timing of the trigger with respect to this window is configurable , so that samples prior to a trigger can be stored .
the input voltage range for each channel can be selected as either @xmath25 , @xmath26 , @xmath27 , @xmath28 , @xmath29 or @xmath30 v peak to peak , with the usual amplitude range set between @xmath25 and @xmath31 volt .
@xcite + a zero suppression algorithm is applied to reduce the quantity of stored data .
this operation has a user specified threshold , minimum width , and number of collected pre and post - samples . when the waveform crosses , and stays above a threshold for a specified number of samples ( minimum width ) , the algorithm returns the entire waveform between the upward and downward threshold crossings including a specified number of pre and post - samples .
if the waveform goes below and then back above threshold before the specified number of post - samples are collected , the algorithm waits for the waveform to drop back below threshold again before counting post - samples .
therefore , if two pulses appear on the same channel with overlapping zero suppression windows , they are combined into one larger pulse .
all the samples outside the accepted sample pulse widths are discarded .
data are usually taken with a zero suppression threshold of @xmath32 or @xmath33 mv , a minimum width of 4 samples ( @xmath34 ns ) , and with pre and post - sample widths set to 25 ns .
+ the ni pxie-8133 controller @xcite is an intel core i7 - 820qm quad - core processor ( @xmath35 ghz frequency ) , with @xmath36ghz ( single - core turbo - boost ) running the windows7 os .
it is housed in a ni pxie-1075 chassis and connects to the pxie modules using the pxi express system .
two gigabit ethernet ports are used to connect the controller to the local area network .
the data acquisition code in the controller , asynchronously collects data from the digitizers and performs zero suppression .
the zero suppressed waveforms are then transmitted in a packet to the veto builder via tcp / ip .
the controller also receives commands from the master run controller .
+ each processor controls up to 14 digitizer boards , one ni pxie-6674 t timing and synchronization module @xcite , and one ni pxie-7961r flexrio fpga module @xcite .
the veto system requires four pxie chassis for the 190 channels , with each chassis housing one pxie controller .
the tpc and veto systems can operate independently , but must be synchronized in order to correlate tpc events with veto signals . although the the two systems are physically displaced , the main difference between them is the width of the data acquisition window , which is @xmath37 350 @xmath10s for a tpc event , while the veto has either a `` short window '' of @xmath38@xmath10s or a `` long window '' of @xmath39@xmath10s .
tpc triggering is obtained by logically processing signals received from the front - end , to determine the number of pe within a set timing window , and the veto trigger is generated in a similar way .
synchronization between the systems is obtained through a high precision time - stamp obtained from a common 50 mhz clock slaved to a 1 pulse per second ( 1pps ) signal received from the lngs gps .
the tpc trigger is always the trigger - master when collecting synchronized data .
however , each system can be locally triggered for calibration and testing .
+ the tpc data acquisition system ( daq ) includes one vme crate containing 5 caen v1720 , 5 caen v1724 modules , a caen v1495 @xcite logic module , and a v976 nim - to - ttl fanout . when an analog signal in any of the 38 pmt channels of the tpc crosses a threshold , a logic pulse of fixed width representing a pe(s ) is generated and sent to the caen v1495 logic module , which is the trigger master .
the tpc trigger in the v1495 is obtained by firmware processing any selected set of these digitized pmt `` hits '' by compiling a running count within a sliding time window .
when the running count crosses a defined threshold , a trigger pulse is initiated and a latch generated to inhibit further triggers until the active trigger is processed .
the veto trigger is generated in the same way using the fpga in pxie-7961r modules .
the parameters used in the fpga trigger firmware are listed in table [ tpcfirmware ] .
parameters marked with [ + ] can be modified by inputs through the vme interface to the v1495 .
the tpc v1495 trigger module is the trigger master for both tpc and veto systems .
it includes appropriate dead time in the data acquisition window for either lsv or ctf veto trigger types in order to prevent trigger overlap . + .
parameters used in the tpc logic unit to determine a trigger .
items market with a [ + ] can be modified by the vme interface to the logic unit [ cols= " < , < " , ] the master v1495 trigger module increments a trigger - number ( trigger - id ) for each processed trigger .
this trigger - id is sent to the veto system so that all veto events within the tpc acquisition window are marked with the same i d .
the lower eight bits of the trigger number are sent to the v1720 as an 8-bit tpc trigger ordinal to further correlate an adc event with the trigger .
there are also different trigger - type identifiers as described in the next sections .
the g2 trigger provides a method to study signals and backgrounds within a selected energy range . as an example
, it can tag events with high - multiplicity s1 signals .
it is based on summing hits from all enabled tpc discriminator channels occurring within a fixed time window which is opened after a pre - trigger is generated by a tpc majority trigger .
the trigger logic uses the same discriminator signals which are used to form the tpc majority trigger logic . a block - diagram of the g2 trigger logic is shown in figure [ g2_trigger ] .
the trigger logic follows the following steps .
+ * the trigger pattern is latched in an internal register after a positive tpc majority logic condition generates a pre - trigger . * a time window of 1 to 5 @xmath10s is immediately generated by a gate generator and used to inhibit other tpc pre - triggers .
* 7-bit counters ( one for each of the 38 channels plus 2 spares ) start counting hits within the window . the counters either stop at the end of the window , or stop when a counter reaches a maximum value of 127 . *
after the window closes , a sequencer scans the counters , summing their values .
the tpc pre - triggers remain inhibited by an adder busy signal for an additional 0.85 @xmath10s . *
the resulting sequencer sum is compared to two thresholds .
if the sum exceeds a high threshold a `` high '' multiplicity g2 pre - trigger is generated
. if the sum is less than the high threshold , but exceeds the low threshold a `` medium '' multiplicity g2 pre - trigger is generated . if the sum is lesser than the low threshold , a `` low '' multiplicity g2 pre - trigger is generated .
* both the high and medium trigger types can be pre - scaled by separate , programmable factors .
the triggers are then provided to the system trigger logic and used in the trigger selection . *
the process re - starts after the pre - scale window closes and the counters are initialized .
the dead time associated with any g2 trigger is equal to the sum of the time window , the 0.85 @xmath10 adder busy , and the g2 acquisition window inhibit .
the g2 acquisition - window - inhibit setting is equal to the v1720 adc acquisition window setting .
+ the veto triggers are obtained in a pxie-7961 module @xcite which is similar to the caen v1495 module , as both allow trigger logic to be programmed into an on - board fpga .
veto trigger signals are combined and sent to the master v1495 located in the clr .
when a veto trigger is received , all inputs to the v1495 are evaluated and if appropriate , a separate trigger signal , using trigger - type bits , is stored in the trigger - id . then a trigger signal and the trigger - id are sent back to a cor fanout and to each of the 4 pxie timing modules in the 4 ni pxie crates .
the signals are finally distributed to all the pxie modules through an internal bus .
the pxie timing modules in the crates also receive the @xmath40 mhz clock , and the 1pps signal .
the veto trigger has the 2 operational modes given below .
* tpc - only - mode*. in this configuration the trigger rate is determined by the internal activity of the tpc .
the length of the data acquisition window is set to @xmath41 @xmath10s to detect prompt and delayed events in the veto which are correlated to tpc activity 2 .
* pass - through - mode * in this configuration , the lsv and ctf triggers are logically combined ( or ) and used as the input trigger to the master v1495 . for this mode ,
the input trigger is determined by the lsv trigger rate .
the length of the data acquisition window is set to @xmath42 @xmath10s in order to acquire the entire scintillation signal and must be adjusted to account for the delay in the hardware trigger .
the veto fpga has internal fifo memory used to store the trigger - id for each incoming veto trigger .
this memory is read for each event to provide a trigger marker .
the veto v1495 decodes the trigger - id and places it in the memory for readout via the vme bus for inclusion in the tpc data stream . a 1pps signal from the lngs gps receiver
is used to synchronize the v1495 and the veto fpga modules .
+ figure [ global_trigger ] shows the generation and flow of trigger signals . unless otherwise noted , signals in the block diagram are active high ttl , ecl or lvds .
there are four fpga - based trigger logic units .
these are ; 1 .
tpc v1495 trigger module located at the tpc site ( clr ) ; 2 .
v1495 trigger module located at the veto site ( cor ) ; 3 .
veto fanout v1495 module located at the veto site ( cor ) ; 4 .
veto fpga pxie-7961 logic module located in the pxie crate ( cor ) a 32-bit trigger - id is generated for each trigger , acquired by the pxie timing module , and routed to the veto fpga for decoding .
the time stamps and the trigger - id are saved in a dma buffer in the veto fpga and are read asynchronously by the veto daq .
thus , the trigger - id is read out via the vme bus as a part of the tpc data stream , and the same trigger - id and the trigger signal are also available for the veto daq readout via a digital adapter module , ni 6583 @xcite , attached to a veto fpga .
a 24-bit `` event - counter '' is programmed to count all incoming triggers for each event .
this allows the same event to be correlated between different adcs as they all have identical time tags .
the event counters are stored in the data header . both the tpc and veto counters are synchronized at the beginning of the run by the external run - enable signal , and all systems are kept in synchronization using a `` star '' distribution with equal time delays of clock , trigger and run - enable signals .
the tpc v1495 trigger output is enabled at the same time as the run - enable signal and the veto system receives run - enable signal from the tpc v1495 trigger . + it is necessary to inhibit triggers when any adc memory is full .
this is accomplished using a chained `` or '' of the adc memory - full signals and firmware in the v1720 tpc adcs . without these signals it would not be possible to guarantee event synchronization between two sub - systems .
generation of any trigger - type is inhibited for an appropriate time after a trigger is issued .
+ three counters are recorded and used to create a time stamp for each trigger .
these are listed below . *
the 1pps counter counts the number of seconds elapsed between the trigger and run - enable signal .
the run - enable signal occurs at run - start so this counter represents the number of seconds between the start of the run and the trigger .
* the gps fine - time counter counts the number of @xmath40 mhz clock cycles elapsed between the trigger and either the run - enable or 1pps signal , as determined by whichever signal occurs last .
the fine - time counter represents the number of 20 ns bins between either the run - enable or 1pps signal depending on the selected signal . *
the gps one - second - counter counts the number of @xmath40 mhz clock cycles between the 1pps signal occurring just prior to the run - enable signal and the run - enable signal itself .
this counter is used to provide the number of @xmath40mhz cycles between a prior 1pps signal and the following run - enable signal .
the counters are stored in the veto fpga memory , and provide the time of a trigger with respect to the run - start with a precision of @xmath43 ns .
this creates a unique time - stamp for a triggered event .
identical time stamps are computed by the tpc and veto daq systems , so that synchronization of events can be maintained .
+ a secondary data stream for the veto system is implemented using tdcs .
one of the discriminated fedb veto outputs from each channel is connected to a caen v1190 multi - hit 128 channel tdc @xcite .
the two 1190 modules are located in the cor and connect to the tpc v1495 trigger unit through the local v1495 fanout logic unit .
this secondary system does not replace the main veto readout but provides a back - up system and solves some tpc integration issues .
consequently the tdc provides an independent data stream for the lsv and ctf vetoes , and is synchronized to the tpc trigger providing a check on the efficiency of the main veto trigger .
thus a tpc event has direct mapping to the tdc data , and is correlated to the veto adc stream , solving a problem when the veto triggers before a tpc trigger is generated .
+ the clock and fan - out , cfm @xcite , is an `` in - house '' , single width 6u vme module .
the input accepts ttl , lvttl , or cmos signals with selective 50 or 10k @xmath9 termination .
in addition a second connector accepts standard lvds signals and is pin compatible with the caen v1720 adc pinouts .
the cfm uses a texas instruments cdce906 programmable 3-pll clock synthesizer / multiplier / divider @xcite and is the source of the output signals .
the cdce906 operates in various modes including `` pll - bypass '' and `` divider '' .
jumper pins allows the input frequency from either from an external source or from an on - board 100 mhz ocxo aocjy2 @xcite quartz crystal oscillator .
the block - diagram of the cfm design is shown in figure [ cfm_fig ] .
a 2x8-channel caen v976 fanout module is used to distribute trigger and run - enable signals .
it is assumed that the tpc and veto systems run independently at selected times , collecting data with their own triggers .
for this reason , two modes of operation are available , which can be selected using a run mode register stored at each site . + 1 .
the system can run in local mode , in which each sub - system receives and records only its own triggers ; or 2 .
the system can run with event synchronization were the tpc daq system is the master .
the daq is designed for high parallelization , as illustrated in figure [ tpc_diagram ] .
there are five , 8-channel caen v1724 digitizers and five , 8-channel v1724 digitizers which are read via optical cable into the pci interface @xcite of three local computers .
one of these clr computers is also optically connected to the vme bus through the caen 2718 crate controller , sharing information with the v1495 logic trigger module and therefore with the caen 1190 multi - hit tdcs in the cor .
digitizer output signals are sent to fragment servers and are multiplexed through a 40 gb optical fiber from the clr to the cor .
all computers run the artdaq software @xcite tools which transfer data from the front - end computers through the daq system to disk , and provide a selection of events for online data quality monitoring , figure [ artdaq ] .
the event builder compresses the data and an `` aggregator - server '' puts the events in sequential order storing them on disk .
the aggregator also provides events for online monitoring , including run parameters from disk files , and closes files at appropriate times during a run .
+ the veto daq architecture is designed to acquire data from 4 different pxie controllers , preserving synchronization between each device and the tpc .
the system software is constructed of two software elements : acquisition - and - readout software and builder software .
software is developed in labview , and conforms to labview coding and documentation standards .
the veto daq uses the parallelism inherent in labview to perform the main operations .
+ the readout software implements the following tasks : * it acts as a server to listen for commands from the run controller ( _ e.g. _ to request the start and stop data acquisition , communicate the status of the system , or provide the number of acquired events ) ; * it checks cpu , and memory usage in the controller ; * it drives a state machine to perform the data acquisition operations ( _ e.g. _ initialize hardware , fetch data , stop , etc . ) ; * it communicates with the fpga module , retrieving the trigger - id and the time stamp of the event ; * it performs zero suppression on the acquired data ; * it bundles zero - suppressed data from different digitizers in a data fragment structure , which contains the time - stamps and the trigger - id of the event ; * it sends data to the veto builder over the network ; each veto readout software is installed in the pxie controllers and is automatically loaded when the controller boots .
the veto daq architecture is based on a producer - consumer data flow driven by a state machine .
communication over the network uses the labview simple messaging reference library ( stm ) .
the content of a message is parsed , evaluated , and a response message broadcast . during data acquisition ,
data fragments from every chassis are collected by the veto builder .
the builder checks the consistency of the trigger - id and timestamps of the data fragments .
finally , the events are entered in sequential order and written on disk in a custom binary format .
the daq sub - systems are handled by a common run controller which is configured to permit differential data acquisition modes .
1 . global runs , where the tpc and veto triggers are correlated , and 2 .
local runs , where sub - detectors can be run independently .
the run controller handles the initialization , start - up , and stopping of data acquisition , and logs to the experimental data base relevant run parameters , such as the tpc and veto configuration .
the run controller also supervises the data acquisition .
when a sub - systems displays unusual behavior , data acquisition is stopped and the daq sub - systems are reset and re - initialized .
the run controller is developed in labview , and conforms to labview coding and documentation standards .
communication over a network with the tpc sub - systems is performed using the xml - rpc protocol , while communication with veto sub - systems is performed using labview stm protocol .
this paper reports the electronics and data acquisition systems developed for the darkside-50 dm search .
the daq involves data collection from a 2-phase lar tpc which is enclosed by liquid scintillator and water cherenkov vetos .
the paper describes the electronics required to obtain digitized pmt waveforms from the various detector systems , transmit this data in parallel to event building computers , and maintain system synchronization .
event - building and run control are also discussed . a dynamic range of over 2000 is obtained by using dual waveform digitizers for each tpc channel , and a novel cryogenic pre - amplifier was implemented which provides extremely low noise and signal distortion .
+ data acquisition is stable with a maximum accepted data rate of over 50 hz , and up to 6 tb of compressed data can be written to disk per day.the full system has been operational several months , and can be controlled remotely .
this work was supported in part by grants from us national science foundation nos . 1004051 , 1242571 , 1314268 , , 1314501 , 1314507 , 1314479 , 1314752 , the us department of energy nos .
de - fg02 - 91er40671 , de - ac02ch11359 the istituto nazionale di fisica nucleare ( italy ) , and the national science center grant no .
umo-2012/05/e / st2/20333 ( poland ) .
the collaboration gratefully acknowledges the support of nicola cescato , fabio cortinovis , thierry debelle and andrea nobile of national instruments .
+ 99 the dark matter scientic assessment group , report on the direct detection and study of dark matter , http://science.energy.gov/hep/hepap/reports/ ( 2007 ) .
m. w. goodman and e. witten , phys .
d31 ( 1985 ) 3059 m.g .
boulay and a. hime , astropart .
25 ( 2006 ) 179 .
g. bellini et al jcap05 ( 2012 ) 015 .
g. bellini et al jcap08 ( 2013 ) 049 .
d. akimov et al .
( darkside collaboration ) , astroparticle physics journal 9 ( 2013 ) 44 .
j. xu et al . , arxiv ( 2012 ) 1204.6011 ; h.o .
back et al . , arxiv ( 2012 ) 1204.6024 .
g. alimonti et al .
( borexino collaboration ) astropart .
phys . 16(2022)205 ( 2002 ) .
d. n. mckinsey , et al , nucl .
inst . and meth .
r. acciam et al .
proceedings of the 2nd international conference on technology and instrumentation in particle physics tipp ( 2011 ) .
multiflex 86 lowloss http://www.koax24.de/en/ coaxial - cable / overview-50-o/24 - 32-mm - gr2/multiflex-86.html http ; //www.caen.it / home / products / mdular pulse processing electronics .
anton empl et al . , arxiv ( 2014 ) 1406.6081 ; jcap 8 ( 2014 ) 64 .
g. alimonti , et al , nucl .
meth . a 406 ( 1998 ) 411
schol / superk / alvaro_thesis.pdf http;//http://www.hamamatsu.com
/ jp / en / product / category / index.html http:://www.toray.jp / films / en / properties / lumirror/. g. bellini et al .
phys rev d 89 ( 2014 ) 112007 ; g. alimonti et al .
( borexino collaboration ) nim a 600 ( 2009 ) 568 ; g. bellini et al .
( borexino collaboration ) jinst 6 ( 2011 ) 05005 arxiv ( 2008 ) 0806.2400 .
http://www.ti.com/product/ths3201 , ths3202 .
http://www.analog.com/static/imported-files/data_sheets/ adcmp567.pdf .
/http://www.caen.it / home / products/. /http://www.caen.it / home / products/. /http://www.xilinx.com / products / programmable devices / fpga . ww1.microchhip.com/downloads/en/devicedoc/30009904p.pdf .
/http:// www.ni.com/pdf/manuals/373858a.pdf .
/http://www.ni.com / pdf / manuals/372870b.pdf .
/http://www.ni.com / pdf / manuals/373089b.pdf .
/http://sine.ni.com / nips / cds / view / p / lang / en / nid/208165 . /http
: www.ni.com / pdf / manuals/372537b.pdf .
/ csite / caenprod.jsp?idmod=484&parent=11 .
/http://www.ni.com / pdf / manuals/375506c.pdf .
/http//www.ni.com / pdf / manuals/373089b.pdf .
/ csite / caenprod.jsp?idmod=787&parent=11 .
/http://www.ni.com / pdf / manuals/323452a .
b. baldin , `` clock fanout module '' , privite communication , november 2012 .
/http://www.microcrystal.com / index.php / en / products/ oscillators / ocxo . /http://www.caen.it / csite / caenprod.jsp?idmod=627&parent=51 .
k. biery , et al . ,
real time conference ( rt ) , 2012 18th ieee - npss . | this paper reports on the electronics and data acquisition system of a dark matter ( dm ) search using a 50 kg dual - phase , liquid argon time projection chamber ( tpc ) which was embedded in water and liquid scintillator veto detectors ( ds-50 ) .
if dm is a subatomic particle , a possible candidate is a weakly interacting massive particle ( wimp ) , and the ds-50 experiment is a direct search for evidence of wimp - nuclear collisions at the laboratori nazionali del gran sasso ( lngs ) .
the light from nuclear excitation in the tpc as a result of a wimp - nuclear collision , is collected by 38 photomultiplier tubes ( pmt ) positioned on the top and bottom of the cylindrical tpc cryostat .
the two veto detectors , instrumented with pmt , shield the tpc and tag background events due to cosmogneic and local radioactivity .
all pmt signals from both the tpc and veto systems are digitized and their waveforms transferred in parallel to event building computers running the artdaq software developed at the fermi national laboratory ( fnal ) .
this paper describes the system triggers , the synchronization of the data flows , and the event building for both the tpc and veto . |
jnosi and gallas @xcite analyzed statistics of the alpine river danube daily water level collected , over the period 1901 - 97 , at nagymaros , hungary .
the authors found , in the one day logarithmic rate of change of the river water level , similar characteristics to those of company growth ( see stanley et al .
@xcite ) which shows that the properties seen in company data are present in a wider class of complex systems .
bramwell et al .
@xcite defined a daily water level mean and variance and computed the daily river water level fluctuations .
they have shown a data collapse of the danube daily water level fluctuations histogram to the ( reversed ) bramwell - holdsworth - pinton ( bhp ) probability density function ( pdf ) .
dahlstedt and jensen @xcite described the statistical properties of several river systems .
they did a careful study of the size of basin areas influence in the data collapse of the rivers water level and runoff , in particular of river negro at manaus , to the reversed bhp and to the gaussian pdf showing that not all rivers have the same statistical behavior . in this paper , we study , again , the south american river negro daily water level at manaus ( 104 years ) . we compute and present a cyclic fit for the negro daily water level period and the negro daily water level standard deviation . we show that the histogram of the danube water level fluctuations is on top of the reversed bhp pdf , which does not happen for the negro daily water level .
we define the _ danube daily water level period _ @xmath0 by @xmath1 where @xmath2 is the number of observed years and @xmath3 is the danube daily water level time series .
of february were eliminated . ] in figure [ fig1 ] , we show a fit to the danube daily water level period @xmath0 .
the mean period fit , using the first harmonic of the fourier series , is given by @xmath4 the percentage of variance explained by the fit is @xmath5 .
we define the _ negro daily water level period _ @xmath6 by @xmath7 where @xmath8 is the number of observed years and @xmath9 is the negro daily water level time series .
of february were eliminated . ] in figure [ fig2 ] , we show a fit @xmath10 of the negro daily water level period @xmath6 .
the mean period fit , using the first four sub - harmonics of the fourier series , is given by @xmath11 where @xmath12 and @xmath13 are given in table [ tab1 ] .
the percentage of variance explained by the fit is @xmath14 .
we define the _ danube daily water level standard deviation _ @xmath0 by @xmath15 in figure [ fig3 ] , we show the chronogram of the danube daily water level standard deviation .
we define the _ negro daily water level standard deviation _
@xmath16 given by @xmath17 in figure [ fig4 ] , we show a fit of the negro daily water level standard deviation .
the fit @xmath18 , using the first ten sub - harmonics of the fourier series , is given by @xmath19 where @xmath12 and @xmath13 are given in table [ tab2 ] .
the percentage of variance explained by the fit is @xmath20 . following bramwell et .
al @xcite , we define the _ danube daily water level fluctuations _ @xmath21 by @xmath22 in figure [ fig5 ] , we show the danube daily water level fluctuations @xmath21 . as shown by bramwell et al .
@xcite , the ( reversed ) bhp pdf falls on top of the histogram of the danube daily water level fluctuations , in the semi - log scale ( see figure [ fig7 ] ) .
we define the _ negro daily water level fluctuations _ @xmath23 by @xmath24 in figure [ fig6 ] , we show the negro daily water level fluctuations @xmath23 . in figure
[ fig8 ] , we show the histogram of the negro daily water level fluctuations . in figure [ fig9 ] , we show the data collapse of the histogram in the semi - log scale to the bhp pdf .
we computed and presented a cyclic fit for the negro daily water level period and for the negro daily water level standard deviation using the first four and ten sub - harmonics , respectively .
we computed the histogram of the negro daily water level fluctuations at manaus , and we compared it with the bhp pdf .
the histogram of the negro daily water level fluctuations is close to the bhp pdf .
we have shown that the histogram of the danube water level fluctuations is on top of the reversed bhp pdf , which does not happen for the river negro daily water level fluctuations .
we thank imre jnosi for providing the river danube data and mrs .
andrelina santos of the agncia nacional de guas of brazil for providing the river negro data .
bramwell , s.t . ,
christensen , k. , fortin , j.y .
, holdsworth , p.c.w . ,
jensen , h.j . ,
lise , s. , lpez , j.m . ,
nicodemi , m. & sellitto , m .
( 2000 ) universal fluctuations in correlated systems , _ phys .
lett . _ * 84 * , 37443747 .
bramwell , s.t . ,
fortin , j.y .
, holdsworth , p.c.w . ,
peysson , s. , pinton , j.f . ,
portelli , b. & sellitto , m .
( 2001 ) magnetic fluctuations in the classical xy model : the origin of an exponential tail in a complex system , _ phys . rev _ *
e 63 * , 041106 .
bramwell , s.t . ,
fennell , t. , holdsworth , p.c.w . , & portelli , b. ( 2002 ) universal fluctuations of the danube water level : a link with turbulence , criticality and company growth , _ europhysics letters _ * 57 * , 310 .
bramwell , s.t .
, holdsworth , p.c.w . , & pinton , j.f .
( 1998 ) universality of rare fluctuations in turbulence and critical phenomena , _ nature _ * 396 * , 552554 .
stanley , m.h.r . ,
amaral , l. a. n. , buildrev , s. v. , havlin , s. , leschhorn , h. maass , p. , salinger , m.a .
& stanley , e. ( 1996 ) scaling behaviour in the growth of companies .
_ letters to nature _ * 379 * , 29 , 804 - 806 . | we study the european river danube and the south american river negro daily water levels .
we present a fit for the negro daily water level period and standard deviation .
unexpectedly , we discover that the river negro and danube are mirror rivers in the sense that the daily water levels fluctuations histograms are close to the bhp and reversed bhp , respectively .
river systems , hydrological statistics , data analysis 92.40.qh , 07.05.kf |
ferromagnetic systems driven by external magnetic fields have been under investigation for many years @xcite
. early work by suhl already showed that there is a threshold of the pump amplitude where the uniform state becomes unstable to homogeneous driving fields @xcite .
when the difference between the pump field and its critical value is small ( the weakly nonlinear case ) , the experimental findings can be analyzed by extensions of the theory of suhl as reviewed in @xcite .
the pattern formation in such a system can be described by amplitude equations @xcite . however , when the probe is _ strongly _ driven , such perturbative procedures break down .
one - dimensional numerical simulations have predicted dynamic domains for a model including a transversely rotating field @xcite .
dynamic domains are stable solutions to the equations of motion . in
the reference frame rotating with the driving field they exhibit a stationary domain structure , like that known from static domains , yet in the lab frame their position is stationary while the magnetization within the domains rotates at different angles .
plefka also obtained dynamic domains in one dimension by numerical simulation and was able to explain characteristic elements of the structure analytically @xcite for a very simplified model .
the experimental difficulties involved with large pump amplitudes were solved already twenty years ago @xcite . renewed
interest in this topic has led to recent work , where , using the faraday effect , dynamic domains were observed in garnet films driven by high power inhomogeneous driving fields @xcite .
the length scales of the patterns we are interested in are large compared to the atomic distance of such substances and so the relevant quantity is a macroscopic variable , the local magnetization @xmath0 . using a model that includes a saturating static field and a strong transverse pump field @xcite
, we consider the structures occurring in a ferromagnetic film of side @xmath1 with normal in the @xmath2-direction .
the paper is organized as following : in sec.ii the physical model is described which enters in the landau - lifshitz equation via the effective fields .
the numerical results for the various dynamical domains are presented in sec .
iii . employing simplifications of the original model the linear stability of the structures found
is considered in sec .
iv leading to partial description of the numerical results .
conclusions are presented in sec.v .
the dynamics of @xmath0 are described by the landau - lifshitz equation @xcite : @xmath3 @xmath4 is a dimensionless damping coefficient .
the effective magnetic field @xmath5 is made up of both external and internal magnetic fields : @xmath6 these terms are , in order of appearance : a static field of magnitude @xmath7 in the @xmath2-direction ( perpendicular to the plane of the film ) ; an in - plane pump field with amplitude @xmath8 and frequency @xmath9 ; an isotropic exchange field ( @xmath10 ) ; the demagnetizing part of the dipolar field , describing the geometry of the film ; and a uniaxial anisotropy , also in the @xmath2-direction .
the explicit time dependence of ( [ ll1 ] ) is removed by a transformation to the rotating reference frame , yielding @xmath11 while the effective magnetic field now has the form @xmath12 an alternative , frequently used form of the landau - lifshitz equation is the gilbert form @xmath13 \,\ , .\nonumber\end{aligned}\ ] ] this is obtained by multiplying equation ( [ eq1 ] ) from the left by @xmath14 and recalling that @xmath15 .
the damping coefficient @xmath4 is the same in both the gilbert form of the landau - lifshitz equation ( [ ll5 ] ) and the common form ( [ eq1 ] ) . throughout this work we present results in the rotating reference frame .
we recall that static results in this frame of reference transform to structures rotating with frequency @xmath9 in the @xmath16-@xmath17 plane in the laboratory frame .
hence a state that is _ statically _ stable in the rotating reference frame is _ dynamically _ stable in the lab frame .
, @xmath18 , @xmath19 , @xmath20 , @xmath21 . a ) `` spin - up '' bubble , period @xmath1 ; b ) horizontal stripes , period @xmath22 ; c ) horizontal and vertical stripes , period @xmath1 ; d ) diagonal stripes , period @xmath23 ; e ) `` spin - down '' bubble , period @xmath1 ;
f ) `` spin - down '' bubble , period @xmath22,scaledwidth=40.0% ] we simulate equation ( [ eq1 ] ) with ( [ eq2 ] ) on a vector - parallel computer and workstations , using a spectral approach to deal with the internal field contributions . the number of mesh points for the two - dimensional system was varied up to @xmath24 and it is seen that a minimum spatial grid size of @xmath25 is necessary to recognize to good accuracy all magnetic structures that are also visible with a larger number of mesh points .
periodic boundary conditions are used . for @xmath26 mesh points ,
the boundary conditions also have no effect on the structures within the system .
this is confirmed by also performing the simulations with unpinned spins , and observing that the same results are obtained.a euler integration scheme is sufficient and therefore used for the time integration of the system .
indeed more advanced methods , like runge - kutta and ndsolve of mathematica , give essentially identical numerical results . throughout the course of a simulation we require that the magnitude of the magnetization remain constant @xmath27 .
however , the accumulation of numerical errors means that this is not the case , and the magnetization tends to drift away from the value one .
therefore we add an additional bloch - like damping term to the equation of motion ( [ eq1 ] ) that recalls the length of the magnetization vector back to one at each time step .
this bloch damping , although a reaction to a numerical artifact , is also justified physically @xcite . to investigate systematically the presence of dynamically stable structures , all numerical and physical parameters except the amplitude of the driving field are kept constant ( @xmath28 , @xmath19 , @xmath20 , @xmath21 ) , while @xmath8 is varied from @xmath29 upwards .
depending on the initial states and on the value of @xmath8 , different coexisting dynamically stable solutions are found .
the solution found most frequently in numerical simulations is the homogeneous or uniform solution . in this case , all the spins throughout the system rotate at the same frequency , in phase , and with the same constant @xmath30 .
the dynamically stable homogeneous state is found numerically for values of @xmath8 below a certain critical value @xmath31 and above another critical value @xmath32 .
some of the two - domain dynamically stable structures also found are shown in figure [ ndd1 ] .
this is a shaded contour plot of the @xmath30 component of the magnetization throughout the system .
the dark areas depicted are those where @xmath33 while the light areas imply @xmath34 .
figure [ ndd1 ] makes no statement about the @xmath16 and @xmath17 components of the magnetization , which ( apart from at the wall ) are spatially homogeneous and vary periodically in time with period @xmath9 .
this implies that the spins within the domains rotate in the @xmath16-@xmath17 plane in phase and at the same frequency .
is the amplitude of the driving field , scaledwidth=35.0% ] the simulations were performed with a great many different initial conditions .
the results shown in figure [ ndd1 ] are characteristic of _ all _ the dynamic domain results obtained , i.e. dynamically stable domain states either have either translational symmetry ( * b , c * and * d * ) or cylindrical symmetry ( * a , e * and * f * ) in the @xmath2-component of the magnetization . it is observed that dynamically stable domain states only exist between two values of the driving field , called @xmath35 and @xmath36 .
as we approach the upper bound @xmath37 the domain states become ever more flattened out , until eventually at @xmath37 the domain state merges into the homogeneous state .
this continuous transition is in direct contrast to the lower bound @xmath38 where a well - defined domain state suddenly no longer exists .
the values of @xmath38 and @xmath37 are seen to be the same whether bubbles or stripes are under investigation .
the spatially distinct uniform regions are separated by a domain wall . here
the magnetization does not jump from the value in one domain to the next , but rather varies continuously across the finite wall width .
the spins inside the wall rotate at the same frequency with the rest of the system , but with a phase shift compared to the spins within the domains . in the domain wall
all the spins are parallel . in particular , this is also true for the bubble domain wall and so the whole structure including wall is _ not _ rotationally symmetric .
the rotational symmetry of the system is broken by the presence of the driving field @xmath8 .
such bubble domains are always obtained , even if rotationally symmetric initial conditions are used .
the structure of the wall differs depending on @xmath8 . for small @xmath8 there is a large negative @xmath39 component of the magnetization . as @xmath8 increases , @xmath40 dominates ever more at the center of the wall .
in addition , the width of the wall increases with @xmath8 .
we noted above that the domains themselves become flattened out .
this , together with the ever widening wall , means that it is difficult to say precisely when the domain state merges into the homogeneous state at the upper bound @xmath37 .
, scaledwidth=40.0% ] as indicated in figure [ ndd3a ] , there is tristability in the system , and which pattern actually occurs depends on the initial conditions of the system .
two hysteresis loops occurring in the system are shown in figure [ ndd3 ] . in particular
, the behavior of the system at @xmath38 is of interest
. we select as an initial state a domain state which would be stable just above @xmath38 . reducing @xmath8 a little below the critical value @xmath38 , the two walls on either side of the `` spin - down '' domain begin to move towards each other , and the magnetization within the walls to rotate faster and faster than the angular frequency of the driving field .
eventually the walls collide and leave the system in a homogeneous `` spin - up '' state .
this process is shown in figure [ bubmovie ] .
the velocity of walls in this transient state is smaller the closer we are to the value @xmath41 , and it can take up to @xmath42 , where @xmath43 is the period of the driving field , for the wall collision to occur . , @xmath19 , @xmath20 for different values of @xmath8 .
the solid lines show the @xmath40 , @xmath39 and @xmath30 components of the magnetization determined analytically from eq .
( [ hfp ] ) , while the dots represent results from numerical simulation .
the critical values @xmath44 and
@xmath45 as shown are both determined analytically from eq .
( [ homoeqcrit ] ) and observed in numerical simulations , scaledwidth=40.0% ]
homogeneous solutions must satisfy the following conditions @xmath46 and so the effective magnetic field becomes : @xmath47 we insert ( [ hheff ] ) into equation ( [ ll5 ] ) and note that for steady state solutions ( @xmath48 ) the expression in square brackets in ( [ ll5 ] ) must be parallel to @xmath49 , with some unknown proportionality factor @xmath50 : @xmath51 where we have introduced the following definitions : @xmath52 the parameter @xmath53 is a measure of the deviation of the system from resonance and is called the detuning , while @xmath54 is the rescaled frequency .
taking the scalar product of equation ( [ h3 ] ) with @xmath55 , @xmath56 and @xmath57 respectively and eliminating the unknown factor @xmath50 yields two equations for the 3 components of @xmath49 .
a third equation is obtained from the requirement that @xmath27 .
we obtain a fourth order polynomial in @xmath30 : @xmath58 m_z^4 + 2 \delta ( k-1 ) m_z^3+\left[h^2+\delta^2\right .
\nonumber\\ & & \left.-(k-1)^2-\gamma^2\right]m_z^2 -2 \delta ( k-1 ) m_z-\delta^2=0\quad .\end{aligned}\ ] ] either 2 or 4 solutions to equation ( [ hfp ] ) are found , depending on the parameter regime .
a saddle - node bifurcation separates the region of two solutions from that with four solutions .
the stability of these four fixpoint solutions is discussed by means of a linear stability analysis ( see appendix ) .
figure [ abb9.2 ] shows the components of the magnetization for that fixpoint where @xmath59 , and the positions of @xmath60 and @xmath61 , between which this fixpoint is unstable to nonuniform perturbations with wavenumber @xmath62 . for comparison ,
the dynamically stable homogeneous states found numerically are also shown .
we make a simple ansatz to describe the magnetization in a two - domain state : @xmath63 i.e. we consider only the magnetization within each domain @xmath64 , and imagine ourselves to be a considerable distance away from the domain walls .
the quantity @xmath65 describes the proportion of the system in each domain .
-component of the magnetization within dynamic domains as obtained from numerical simulations and analytical calculation for different values of @xmath8 .
each pair of dots for a given value of @xmath8 corresponds to the @xmath30 values far from the domain wall for a single numerically stable simulation ; the solid lines show the analytical result eq .
( [ eqnd.4 ] ) .
the upper bound @xmath66 is both calculated from eq .
( [ upbound ] ) and observed numerically , while the lower bound @xmath67 is only observed numerically , scaledwidth=40.0% ] again we look for steady state solutions ( @xmath68 ) , and demand that the spatial inhomogeneity in each domain vanish ( @xmath69 ) .
the demand that the term in square brackets in ( [ ll5 ] ) be parallel to @xmath49 yields : @xmath70 with @xmath71 these two equations for @xmath72 and @xmath73 are not independent , but rather are coupled via the demagnetizing term @xmath74 which holds for the entire system and appears in each effective field @xmath75 . using the ansatz ( [ dan ] ) we write down the demagnetizing term as @xmath76 a
further ansatz still needs to be made to obtain domain solutions , and motivated by the symmetry of the numerical results we set : @xmath77 although this symmetry is not inherent in the equations of motion , it is observed in all numerical simulations without exception and , more importantly , it fixes @xmath65 and permits a simple closed solution form . by taking the scalar product of ( [ dgll ] ) for @xmath64 with @xmath55 , @xmath56 and @xmath57 and eliminating the unknown @xmath78 , we find consistency only if @xmath79 this is an equilibrium condition for dynamic domains that fixes the proportion of the system in each domain . the effective magnetic field ( [ heffi ] ) in each domain @xmath80
is now @xmath81 comparing equation ( [ dheff ] ) with ( [ hheff ] ) , we see that all results from the homogeneous calculation in section [ homosec ] can now be applied , providing @xmath7 is replaced by @xmath82 ( thus @xmath53 is replaced by @xmath83 ) and @xmath84 by @xmath85 . equation ( [ hfp ] ) then becomes @xmath86 with solutions @xmath87 the other components of the magnetization are also simply expressed as @xmath88 figure [ abb9.3 ] shows the @xmath2-component of the magnetization within each domain , as given in ( [ eqnd.4 ] ) and compared to numerical simulations . we note that the domain solution can only exist for @xmath89 a linear stability analysis ( see appendix ) shows that the simple domain state ( [ dan ] ) is stable for _ all _ values of @xmath8 where the solution exists .
this is in contrast to the numerical evidence , which indicates that there is a lower bound ( shown as @xmath90 in figure [ abb9.3 ] ) below which no numerical domain solutions exist .
there is _ no _ indication of such a lower bound in any calculations carried out using the ansatz ( [ dan ] ) .
we recall that the ansatz ( [ dan ] ) does _ not _ describe an entire dynamic domain , only that part far from the domain wall
. therefore any criteria for stability determined in performing a linear stability analysis on ( [ dan ] ) will be only _ necessary _ conditions for stability , and not _ sufficient _ conditions .
critically , we have neglected any description of the shape or dynamics of the domain wall , and merely assumed a discontinuous transition from one domain to the next . in order to obtain a full description of dynamic domains , and , in particular , to determine the position of the lower bound @xmath38 , it will be necessary to include the wall and consider its effect on the stability of the system .
this work presents a model for a ferromagnetic film in which the exchange , dipolar and anisotropy effects are taken into account .
numerical simulation shows , depending on the amplitude of the transverse pump field applied , different coexisting homogeneous states and dynamic domain structures .
the homogeneous state and its stability is wholly understood , and , using a simple model , certain features of the dynamic domain state are also clarified .
the magnetization within the domains is determined , as is the proportion of the system in each dynamic domain .
nevertheless a lower bound for the existence of dynamic domains is observed in the numerical simulations that is not found with the simple domain - state model .
an extension to this work has to include the effect of the domain wall , both its structure and its dynamics .
such an extension has been worked out @xcite and will be published separately @xcite leading to an understanding of the numerical results in their entirety .
we wish to acknowledge interesting discussions with h. dtsch and t. wbbeking
we denote the basic solution computed above as @xmath91 .
we then add a small time and space dependent perturbation to @xmath91 , in particular we consider one fourier mode of such a perturbation : @xmath92 now the magnetization @xmath93 must have magnitude 1 . thus @xmath94 must lie in a plane perpendicular to @xmath91 : @xmath95 we insert ansatz ( [ han ] ) into the landau - lifshitz equation in the rotating reference frame ( [ eq1 ] ) and sort for powers of @xmath96 . to order @xmath96 : @xmath97 \quad , \end{aligned}\ ] ] where we have introduced
@xmath98 the field @xmath99 depends on whether the case @xmath100 or @xmath101 is being considered , i.e. whether the perturbation @xmath102 is uniform or non - uniform . if @xmath103 , @xmath104 , whereas if @xmath105 , the fluctuations averaged over the entire system are zero and so @xmath106 .
therefore @xmath107 we introduce a coordinate system defined as @xmath108 in this coordinate system @xmath94 is written as @xmath109 and equation ( [ ho1 ] ) assumes the form
@xmath110 \quad .\end{aligned}\ ] ] we examine the four fixpoint solutions of ( [ hfp ] ) for stability with respect to uniform ( @xmath100 ) and nonuniform ( @xmath101 ) perturbations by considering the eigenvalues of the linear operator @xmath111 defined by the right - hand side of equation ( [ eq3 ] ) .
the eigenvalues of the @xmath112 matrix @xmath111 for @xmath100 are positive for two of the four fixpoints . for the other two fixpoints ,
the trace of @xmath111 for @xmath105 is negative for all values of @xmath113 , so we only need consider its determinant , given by @xmath114 where @xmath115 varying @xmath8 , the determinant first becomes negative ( and hence the fixpoint unstable ) for @xmath116 at critical values of the driving field called @xmath60 and @xmath61 given by @xmath117 the stability analysis of the domain solution is performed analogously to that of the homogeneous solution in section [ hstabsect ] .
a small perturbation is added to the magnetization in each domain @xmath64 .
as the fourier modes in the first domain would couple to all of those in the second domain , we instead perform the calculation in real space , writing @xmath118 we obtain a system of 2 equations for each @xmath119 , equivalent to ( [ eq3 ] ) @xmath120(-\beta\ui \vec{e}_3\ui + \gamma\ui \vec{e}_2\ui)\nonumber \\ & & \quad+ \gamma(\vmo\ui \cdot \vec{h}_{\rm eff,0}\ui ) ( \beta\ui \vec{e}_2\ui+\gamma\ui \vec{e}_3\ui)\\ & & \quad + \left [ k ( m_{z,0}^2 - 1)\gamma\ui - ( \overline{m}_z)_1 \right ] ( \vec{e}_2\ui+ \gamma \vec{e}_3\ui ) \nonumber\\ & & \quad + j \left[\beta{''}\ui ( \vec{e}_3\ui-\gamma \vec{e}_2\ui)+\gamma{''}\ui ( -\vec{e}_2\ui - \gamma \vec{e}_3\ui)\nonumber \right]\end{aligned}\ ] ] where @xmath121 is defined in ( [ dheff ] ) . in the case of _
uniform _ perturbations , equation ( [ dll2 ] ) reduces to a system of four coupled ordinary differential equations . numerically computing the eigenvalues of this @xmath122 matrix
shows that all eigenvalues are negative for @xmath123 and so the domain state is stable to ( local ) uniform perturbations . for _ non - uniform _
perturbations , we descritize equation ( [ dll2 ] ) by replacing the spatial derivatives with finite difference expressions . for all values of @xmath123 , the real part of the eigenvalue spectrum is strictly negative for both solutions @xmath124 in ( [ eqnd.4 ] ) .
hence the domain state is also stable to ( local ) non - uniform perturbations .
finally , we may consider what happens when the _ position of the wall is displaced from its equilibrium position _ given by ( [ d79 ] ) , i.e. , additionally letting @xmath125 .
again , we obtain a linearized equation of motion . integrating this equation over a small section of the containing the wall , and then letting the size of this section go to zero @xmath126
we obtain a single equation for @xmath127 : @xmath128 q_1 \quad .\ ] ] the expression in square brackets in ( [ qeqn ] ) is strictly positive , and so again the domain solution is stable to displacement of the wall from its equilibrium position . | the spatiotemporal structure formation problem is investigated in the region far above the transverse ferromagnetic resonance instability .
the investigations are based on the dissipative landau - lifshitz equation and have been performed on a model which takes external fields , isotropic exchange fields , anisotropy fields and the demagnetizing part of the dipolar field into consideration .
the numerical simulations for these models exhibit stationary domain structure in the rotating frame . employing analytical methods and simplifying the model , certain features , such as the magnetization within the domains and the proportion of the system in each domain ,
are described analytically . |
reliable particle identification ( pid ) is essential for any high energy physics experiment , and in particular , it is crucial for the success of @xmath0-factories @xcite . in the @xmath0-factories , pid is required to tag @xmath0-meson flavour for cp violation studies in the neutral @xmath0-meson system , and to suppress backgrounds in precision measurements of rare @xmath0 and @xmath1 decays .
based on the successful operation of the belle experiment , the kekb @xcite is being upgraded to the superkekb collider , which is designed to give an instantaneous luminosity of @xmath2 .
this is about 40 times larger than its predecessor .
the belle detector @xcite is also being upgraded to the belle ii detector @xcite in order to be able to perform in the upgraded accelerator conditions and to achieve the desired physics goals . to cope with high rates as well as with more stringent requirements on separation capabilities for rare decay channels ,
more efficient pid is needed for the momentum range up to 4 @xmath3 . in belle
ii , pid will be performed by the time - of - propagation ( top ) counter in the central ( barrel ) region and the aerogel ring imaging cherenkov ( arich ) counter in the forward endcap region .
figure [ fig : pid ] shows a schematic of the upper half vertical cross - section of the belle ii detector in which both the pid systems are shown .
the working principle of both the top and arich counters is based on imaging the cherenkov rings .
the cherenkov photons are emitted by a charged particle moving with a velocity @xmath4 exceeding the speed of light @xmath5 in a dielectric medium ( refractive index @xmath6 ) at a characteristic angle ( @xmath7 ) given by , @xmath8 .
charged particles of the same momentum but having different masses will have different velocities , and hence the angle of cherenkov photons will be different .
in this proceedings , we report the design , method and present status of the pid systems in both the regions of the belle ii detector .
a top counter module primarily consists of a quartz radiator bar , micro - channel plate photomultipliers ( mcppmts ) and a front - end readout .
the cherenkov photons emitted from a charged particle in the quartz bar are internally reflected toward the detection plane .
the obtained cherenkov ring image is characterized by its time of arrival and impact position in the detector plane @xcite .
two quartz bars each having dimension @xmath9 , a mirror with dimension @xmath10 , and a small expansion prism of dimension @xmath11 are epoxied with an optical glue to form the quartz radiator , as shown in the figure [ fig : module ] ( left ) .
the radiator is enclosed in a box made of aluminum honeycomb panels and is supported by peek polymer buttons to leave an air gap between the radiator and the panels .
the mirror has a radius of curvature of about 6500 mm .
as such , it focuses parallel rays in the cherenkov cone to a single point and thus removes the effect of the bar thickness ( see figure [ fig : module ] ( right ) ) .
non - parallel rays are focused to different points of the detector plane , which gives a better charged particle separation and also allows to correct for chromatic dispersion .
optical components of the radiator should be of high quality to prevent any photon loss during propagation or reflections . to ensure this ,
several quality acceptance requirements are needed @xcite .
cherenkov photons , emitted in the radiator due to passage of a charged particle , go through total internal reflections and are finally register at the expansion volume end by an array of mcppmts .
different path lengths are traversed by the photons depending on @xmath7 , which results in a different time of propagation .
the arrival time , including the time of flight of the charged particle and the time of propagation of emitted photons , and position in the detector plane of each detected cherenkov photon are used to compute a likelihood for a given particle mass hypothesis .
a square - shaped mcppmt with an effective area @xmath12 has been developed in collaboration with hamamatsu photonics kk for the top counter @xcite@xcite .
it has a multi - alkali photocathode whose average quantum efficiency is about 28% for wavelengths around 380 nm . in the mcppmt , there are two multi - channel plates ( mcp ) of thickness @xmath13 with a bore size diameter of @xmath14 .
the bias angle of the pore is @xmath15 .
the readout from each mcppmt anode is divided into @xmath16 channels of area @xmath17 .
each module contains two rows of 16 mcppmts , giving in total 32 mcppmts or 512 anode channels per module .
the photo - electron emitted by the photo - cathode is multiplied in the mcp pores and then detected at one of the anode channel .
the transit time spread is about 35 ps . by applying the nominal operating bias , a gain of @xmath18
is achieved .
the mcppmts are physically mounted on the front - end electronics modules .
the main components of the front - end electronics are : front board to host the mcppmt array ; high voltage board to provide high voltages to the mcppmts ; application - specific integrated circuits ( asics ) ; and standard control , read - out , and data ( scrod ) board .
the asics consist of switched - capacitor array memory for the fast sampling and storage of time - sampled analog waveforms .
the scrod is the main controller of the entire module and acts as an entry / exit command / data point to the unit @xcite@xcite .
a small prototype of the top counter was tested at the @xmath19 positron beam at leps ( laser electron photon beam line at spring-8 ) in june 2013 . for the beam test
, the positron beam was tracked by the leps spectrometer and the beam timing was obtained from the accelerator rf .
data were taken with the beam hitting the prototype at normal incidence and also at some angle . in the latter case ,
the emitted cherenkov photons will mostly be reflected from the mirror and then reach the detector plane . for both cases of beam incidence ,
results in data and simulations were found to be in agreement . in a single event ,
the number of detected photons was 20 - 30 .
an integrated distribution in data and simulations over many events for the normal incidence case is shown in figure [ fig : beamtest ] .
the arich counter is basically a proximity focusing rich .
its main components are : aerogel tiles as a radiator , an array of position sensitive photon detectors , and a readout system @xcite@xcite@xcite .
the aerogel radiator consists of two layers with increasing refractive indices along the particle path so that the cherenkov photons emitted by each layer overlap at the photo - detection plane , as shown in figure [ fig : focconf ] ( left ) .
the two aerogel tiles are of thickness 20 mm each and have refractive indices of 1.045 and 1.055 for upstream and downstream tiles , respectively .
this arrangement of two layers with different refractive indices gives better performance than a single aerogel layer for the entire thickness ( 40 mm ) , as the latter arrangement leads to degradation of the single photon resolution because of an increase in emission point uncertainty .
the single photon resolution ( @xmath20 ) for the dual radiator is significantly better than for the case of single radiator ( @xmath21 ) without affecting the number of photons detected ( figure [ fig : focconf ] ) . for the focusing configuration ( center ) and for a 40 mm single homogeneous aerogel layer
( right ) @xcite.,width=566 ] the photo - detector for the arich should be sensitive to single photon detection , able to provide position information , be immune to the 1.5 t magnetic field perpendicular to the photon detection plane , and be tolerant of the high radiation environment .
a hybrid avalance photo - detector ( hapd ) was developed in a joint effort with hamamatsu .
the hapd consists of a bialkali photocathode , a vacuum region with a high electric field , and four avalanche photo - diode ( apd ) chips that are further pixelated in @xmath22 pixels giving in total 144 pixels of size ( @xmath23 ) .
an incident photon is coverted into a photo - electron by the photocathode , which has a peak quantum efficiency of about 28% at 400 nm .
the photo - electron subsequently accelerates towards the pixelated apds via a high electric field and gets multiplied due to bombardment gain by a factor of about 1800 .
an additional gain of about 40 is achieved from the avalanche process of the apds , resulting an overall gain of 70,000 .
the arich covers @xmath24 in the forward endcap of the belle ii detector ; the aerogel radiator plane and the photo - detection plane are separated by a distance of 200 mm .
the radiator plane consists of 124 pairs ( dual - layer , in total 248 ) of aerogel tiles while the photo - detection plane contains 420 hapds . at the outermost edge of these two planes , planar mirrors ( in total , 18 ) are placed to redirect the outside - going photons towards the detection plane .
dedicated high gain and low noise electronics were developed for the readout . to each hapd a front - end board with four asics and a field programmable gate array is attached .
the digitized hit information is collected by a merger board from front - end boards , and then communicated to further stages of the data acquisition system @xcite .
a prototype arich was tested at the desy test beam in may 2013 .
a photograph of the setup is shown in figure [ fig : arich_bt ] ( left ) .
the accumulated hit distribution with respect to the track position is shown in figure [ fig : arich_bt ] ( right ) , in which a cherenkov ring is clearly visible .
the cherenkov angle resolution is found to be 15.8 mrad , and on average 9 photons per track are detected .
the top and arich detectors will provide pid information in the barrel and forward endcap region , respectively , of the belle ii detector .
the working principle of both the detectors is based on imaging the cherenkov ring created by the passage of charged particles .
the top utilizes the impact position and time of arrival of the cherenkov photons at the detection plane after total internal reflections in the quartz radiator bar to obtain pid information , whereas the arich is a proximity focusing rich detector .
prototypes of both the detectors demonstrated their expected performance in test - beams .
all 16 top counters have been successfully installed in the belle ii detector .
installation of the hapds and the aerogel tiles has already started and will be finished by the summer of 2016 .
detailed simulations have been performed with the belle ii software framework , based on which we expect an excellent charged kaon efficiency ( @xmath25 ) with a very small ( @xmath26 ) charged pion misidentification probability .
9 krizan p , 2009 _ journal of instrumentation _ * 4 * , p11015 .
kurokawa s and kikutani e 2003 _ nucl .
methods phys .
res . , sect .
_ a * 499 * 1 and other papers included in this volume ; abe t
_ 2013 _ prog .
phys . _ 03a001 and following articles up to 03a011 . | we report on the charged particle identification ( pid ) systems for the upcoming belle ii experiment .
the time of propagation counter in the central region and the proximity focusing ring imaging cherenkov counters with aerogel radiator in the forward region will be used as the pid devices . they are expected to provide a kaon identification efficiency of more than 94% at a low pion misidentification probability of 4% . the motivation for the upgrade , method and status of both systems are discussed . |
tycho s supernova remnant ( snr ; hereafter tycho ) , is the remnant of the sn observed in 1572 @xcite , first characterized by @xcite as a `` type i '' sn . @xcite identified light echoes from the event , and spectroscopy of these echoes by @xcite matched the spectrum to a normal sn ia .
many authors have adopted a distance of 2.3 kpc , suggested by @xcite based on an analysis of optical observations of nonradiative shocks , and by @xcite , based on h i absorption .
distances of over 4 kpc have also been reported , based on higher - resolution h i data @xcite . in an earlier work ( @xcite , hereafter w13 ) , we favored a distance of 3.5 kpc , based on comparisons of hydrodynamic simulations with data .
there , we examined the infrared ( ir ) colors of the remnant , fitting models of warm dust to the _ spitzer _ broad - band fluxes .
these models are sensitive to the post - shock gas density , and we found a variation in density as a function of azimuthal angle around the shell , with densities in the east and northeast several times higher than in the west and southwest .
sne ia are believed to result from a thermonuclear explosion of a white dwarf destabilized by mass transfer in a binary system , but the nature of the binary is unknown .
the two leading models are the single - degenerate ( sd ) and double - degenerate ( dd ) scenarios . in the sd channel @xcite , a white dwarf accretes matter from a non - degenerate companion and explodes when it reaches the chandrasekhar limit . in this scenario , the companion star should survive , though evidence for this via searches for this companion star in type ia snrs is scant @xcite .
the dd channel @xcite , on the other hand , involves an explosion triggered by the merger of two white dwarfs , and no surviving companion is expected . while tycho is clearly round in shape , a departure from spherical symmetry
is suggested by measurements of the proper motion of the shell .
@xcite used radio images from 1983/84 and 1994/95 and found that the shock velocity varies by a factor of three as a function of position around the shell .
@xcite confirmed these azimuthal variations from _
rosat _ images .
@xcite examined _ chandra _ images from 2000 , 2003 , and 2007 to measure the expansion of the remnant at 39 positions around the periphery and determined that the x - ray proper motions of the forward shock vary by about a factor of two .
the radio and x - ray emission from the forward shock in tycho both arise from the same physical process , nonthermal synchrotron emission .
electrons are accelerated by turbulent magnetic fields , amplified at the shock front .
this acceleration produces a nonthermal `` tail '' to the particle energy distribution , where electrons reach relativistic energies up to tens of tev .
these highest energy electrons , spiraling in the magnetic field , produce x - ray synchrotron emission @xcite , while the radio emission results from particles with energies in the gev range .
while most young remnants exhibit thermal x - ray emission at their forward shock , @xcite show that in tycho , the emission at the immediate edge is nonthermal . in this work
, we report on proper motion measurements of the forward shock , made from newly - obtained x - ray and radio data from _ chandra _ and the _
karl g. jansky very large array ( vla ) _ in 2015 and 2013/14 , respectively .
while the measurements made by @xcite and @xcite follow the same general trend , there are discrepancies between them .
this is not unexpected , as these were independent groups of authors using different techniques with relatively short time baselines .
with much longer time baselines ( twice as long in the x - ray band and three times longer in the radio ) , as well as a consistent approach to the measurements in both , we can greatly reduce the uncertainties and disparities in the previous measurements .
accurate determinations of shock velocities in tycho are important for the physics of particle acceleration , as electron maximum energies and efficiencies of acceleration and magnetic - field amplification depend on high powers of the velocity .
the x - ray and radio profiles have different shapes , owing to the physics of synchrotron emission .
x - ray synchrotron emission is confined to a thin rim that rises sharply at the shock front , only to fall again fairly quickly behind it as a result of synchrotron losses on the electrons , damping of the magnetic field , or both .
this phenomenon is discussed at length in @xcite and @xcite ( for tycho in particular ) .
radio synchrotron emission results from lower energy electrons with a longer life , and persists well behind the shock front , creating a plateau emission profile .
_ chandra _ has observed tycho four times prior to our 2015 observations : 50 ks in 2000 , 150 ks in 2003 and 2007 , and 750 ks in 2009 .
the 2000 observation used the acis - s3 chip , and the positioning of tycho on the chip resulted in about 25% of it along the southern shell being cut off by the chip edge @xcite .
subsequent observations used the i - array and cover the entire remnant . for this work ,
we use the longest time baselines . for most of the remnant
, our measurements are made between the 2000 and 2015 data .
for the portions in the south that were not covered by the 2000 observation ( approximately one third of our regions ) , we substitute the 2003 epoch . on a side .
[ images ] , width=604 ] we observed tycho for 150 ks on 2015 apr 22 - 24 using the acis i - array .
the time intervals from the 2000 and 2003 observations to the 2015 are 14.6 and 12.0 year , respectively .
we follow a similar data reduction procedure to @xcite , using version 4.7 of ciao and version 4.6.5 of caldb to process all epochs .
we examined the light curves and found no significant background flares . to align all epochs to a common reference frame
, we first used the ciao task ` wavdetect ` to detect point sources in the field .
the tasks ` wcs_match ` and ` wcs_update ` then reprojected each events file to identical wcs coordinates .
we use the deep 2009 observation as the relative `` reference '' frame to which all other epochs are aligned .
we smooth the x - ray images slightly , using a 2-pixel gaussian .
this has virtually no effect on the profile shapes , but significantly decreases the pixel - to - pixel poisson noise level .
we imaged two epochs of l - band ( @xmath11.4 ghz ) observations , both obtained with the vla in its a , b , c , and d configurations .
the first epoch ( pi w. van breugel ) , from 19831984 , is published in @xcite , but is re - imaged here .
the second was obtained by our team as part of project vla/13a-426 ( pi j.w .
hewitt ) , on 2013 feb 9 ( d configuration ) , 2013 jun 8 ( c configuration ) , 2013 sep 28 ( b configuration ) and 2014 feb 20 ( a configuration ) . for simplicity
, we will refer to these as the 2013 " observations ( and the first epoch as 1983 " ) .
two other vla epochs ( 1994 and 2002 ) have also been obtained , but we desire the longest time baseline , and only use the 1983 and 2013 data .
data are edited and calibrated using standard routines in aips , and data from all configurations were concatenated in the uv plane using ` vbglu ` for the 1983 epoch and ` dbcon ` for the 2013 epoch . in both epochs , the source j0217
+ 7349 was used for complex gain calibration , while 3c48 was used for absolute flux calibration and , in the case of the 2013 epoch , bandpass calibration .
we imaged in aips using multi - scale clean in ` imagr ` .
data in the a configuration were obtained over 1983 nov 1314 , providing 5.4 hours on source with bandwidth of 3.1 mhz at each of two intermediate frequencies ( ifs ) ; ifs were tuned to 1375 and 1385 mhz .
b configuration data were observed 1984 jan 57 , yielding 4.6 hours on source with bandwidth of 6.2 mhz @xmath2 2 ifs , and ifs are positioned at 1365 and 1442 mhz .
c configuration data were obtained on 1983 april 1718 , with one if at 1375 mhz and 25 mhz of bandwidth ; 5.1 hours of on - source time were obtained .
d configuration data were obtained on 1983 jun 18 , with 1.8 hours on source and 25 mhz of bandwidth tuned to 1375 mhz .
the narrow bandwidths in a and b configurations minimize the effects of bandwidth smearing .
we imaged this epoch using a briggs robust value of @xmath3 ( nearly uniform weighting ) , yielding a fwhm of the synthesized beam , @xmath4 , at a pa = 28.2@xmath5 . in each of our 2013 configurations ,
observations yielded 0.5 hour on source and made use of the full 1 ghz bandwidth of the l band .
however , to effectively compare with the 1983 epoch , we only made use of the sixth and seventh spectral windows , which are 64 mhz wide and centered at 1346 and 1410 mhz .
we imaged these data with a briggs robust value of @xmath6 ( uniform weighting ) to produce an image with resolution @xmath7 at a pa = 274.7@xmath5 . because the observations from the two epochs had different point spread functions , and because both of these were elliptical , we degrade both images to a circular beam of @xmath8 using the aips task ` convl ` . both the x - ray and radio images
are shown in figure 1 .
in choosing our measurement regions , we attempt to best match our previous regions in w13 , with some modifications necessitated by the data , such as choosing locations where the shock front is `` sharpest '' ( little diffuse emission ) .
we use x - ray data only in the 3 - 8 kev range for two reasons .
first , this eliminates virtually all thermal emission , selecting only the nonthermal synchrotron radiation that defines the shock front .
secondly , this allows for a more direct comparison with the radio data , where the emission is also synchrotron .
our regions are `` projection '' boxes in ds9 , and range in width from 20 - 60@xmath9 , chosen to lie along the shock front .
we made our boxes large enough to provide good statistics within the region of interest , but small enough so that the curved shock front remains mostly straight .
there are two gaps in our coverage : one at a pa ( e of n ) of @xmath10 ( between boxes 2 and 3 ) and the other at @xmath11 ( between boxes 17 and 1 ) . in these locations ,
the shock front is too diffuse to get a good profile and a robust proper motion measurement . our procedure for measuring the proper motion mirrors that found in @xcite , where a fuller description can be found .
other snr works , such as @xcite and @xcite , have used this technique as well .
we extract the 1d radial profiles from both epochs , with uncertainties on each data point , then shift epoch 1 relative to epoch 2 , minimizing the value of @xmath12 .
profiles are extracted in pixel space ( x - ray pixel : 0.492@xmath9 , radio pixel : 0.4@xmath9 ) , with shifts calculated on a grid of 2000 points with a size of 0.025 pixels .
we only fit for the shift within an area ( specific to each region ) containing the filament edge , as shown in figure [ goodprofile ] .
while we use identical regions in both the x - ray and radio , there are slight differences in the way that we calculate the uncertainties on the profiles . for the x - ray band , we work in photon counts , and take the square root of the number of counts in each pixel as the uncertainty .
the radio images are in units of jy beam@xmath13 , and we consider the `` noisiness '' of the image to be the main source of uncertainty .
we calculate the rms dispersion within the off - source background , effectively a flat - field .
we obtain uncertainties of 8.5 @xmath14 and 4.0 @xmath14 jy beam@xmath13 for the 1984 and 2013 images , respectively , and apply these constant uncertainty values to all points along their respective radio profiles .
for both the x - ray and radio profiles , reduced @xmath12 values were generally close to 1 .
for the uncertainties on the proper motion measurements themselves , we include estimates of both the statistical and systematic uncertainties .
the statistical errors are the 90% confidence limits resulting from a @xmath12 increase of 2.706 .
the 90% uncertainties in each direction are virtually identical , so we average them for a single number . for the systematic errors , the uncertainties arising from different wcs alignments are negligible .
however , what is not trivial is the angle of the projection box to the shock front . following a procedure similar to @xcite , we vary the angle of the projection box by 1 and 2@xmath5 in both directions with respect to the shock front for each region and re - measure the proper motion . to be conservative ,
we adopt the 2@xmath5 variation as the measure of our systematic uncertainty . as with the statistical uncertainties , because the values in each direction are similar , we average them together for a single value .
the vast majority of our regions are `` well - behaved , '' in that the shape of the profile is virtually identical in both epochs .
we show an example of a well - behaved profile in x - ray and radio in figure [ goodprofile ] . however , there are a few regions where this is not the case .
we show the worst example from each wavelength regime in figure [ badprofile ] . in the x - ray band , region 3 ( p.a .
62@xmath5 ) is shown , where the profile clearly changes shape between the 2000 and 2015 data .
we do report a proper motion here , defining it as the shift in the leading edge of the shock front , when the profile rises above the level of the background .
this region is even worse in the radio ( not shown ) , where no clear shock front is present and the profile simply gently fades into the background . we do not report a radio proper motion for this region .
region 15 ( pa 295@xmath5 ) is shown as our `` worst - case '' ( with the exception of region 3 , discussed above ) radio profile .
the profile becomes more elongated between the two epochs , for reasons that are unknown .
nonetheless , we apply the same technique as above by fitting the shift of the leading edge of the shock .
this leads to a value of the proper motion , as reported in table [ measurements ] , but this value should be used with caution .
interestingly , the x - ray profiles from this region do not show this behavior . .
[ all ] , width=566 ] the results of our measurements are given in table [ measurements ] and figure [ all ] .
we confirm the existence of a clear velocity gradient in the shock speeds .
even if we disregard the potentially problematic x - ray measurement of region 3 ( discussed above ) , we still see a difference of approximately a factor of 2 in the expansion velocities between the e / ne portions of the remnant and the w / sw .
furthermore , with a few exceptions , the x - ray and radio proper motions agree within errors .
the biggest discrepancy occurs in region 15 , which we have already pointed out is problematic .
while we report all proper motions in units of arcsec yr@xmath13 , a conversion to velocity relies upon the distance to tycho , which may lie from 2.3 to 4 kpc .
if we adopt a fiducial distance of 3 kpc , the conversion factor is 0.1@xmath9 yr@xmath13 @xmath15 1422 km s@xmath13 .
this distance would imply maximum shock velocities of about 5300 km s@xmath13 , consistent with the low pre - shock densities reported in w13 , @xcite , and @xcite .
it is interesting to note that our region 4 corresponds to the well - known optical `` knot - g '' of @xcite , and the proper motion we measure is fully consistent with their value of 0.20 @xmath16 yr@xmath13 .
in w13 , we described hydrodynamical simulations of a spherically - symmetric sn explosion into a density gradient in the ism , which is the simplest explanation for the different densities and shock velocities found around the remnant s periphery .
this leads to a remnant where the geometric center _ can not _ be the site of the original explosion .
we showed that the remnant can still be remarkably round , despite higher expansion velocities on one side .
an important result of these simulations , discussed in w13 , is that the ratio of the velocity semi - amplitude ( v@xmath17 - v@xmath18)/(v@xmath17 + v@xmath18 ) to the radial offset from the center of the explosion ( r@xmath17 - r@xmath18)/(r@xmath17 + r@xmath18 ) is roughly constant at a value of 2.2 @xmath19 for ages between about 300 and 700 year . with our refined proper motion measurements presented here
, we can more accurately calculate the radial offset necessary to explain the observations .
we use the average of the x - ray and radio proper motions given in table [ measurements ] , omitting region 3 , as this has an unusually low x - ray proper motion and no measureable radio value .
we use a @xmath12 minimization algorithm to fit a sinusoidal function , @xmath20 to the data shown in figure [ all ] , for both the amplitude , @xmath21 , and the vertical offset , @xmath22 , in units of arcsec yr@xmath13 , as well as for the phase , @xmath23 . in this formalism ,
@xmath24 and @xmath25 , and the velocity semi - amplitude reduces to @xmath26 .
we obtain best - fit values of a = 0.062 @xmath27 0.0045 and y = 0.293 @xmath27 0.002 , leading to a velocity semi - amplitude of 0.212 @xmath27 0.015 .
dividing this by the ratio of 2.2 gives an offset of ( 9.6 @xmath28)% of the radius of the remnant , or ( 23 @xmath29)@xmath9 , with the offset in the @xmath30 direction , which occurs in the ne at a position angle of 51@xmath31 with respect to the geometric center of @xcite .
this is comparable to the search radius of @xcite ( who found no viable candidate for a remaining donor star under the single - degenerate type ia sn scenario ) , and that of @xcite , who report a possible candidate .
this candidate , `` star g , '' is 31.4@xmath9 from our explosion center .
this is in contrast to @xcite , who find an offset in the nw direction .
this results in a large ( @xmath32 ) displacement between our best - fit explosion center , located at @xmath33 = 0@[email protected]@xmath36 and @xmath37 = 64@[email protected]@xmath9 , and their results .
we note , though , that their paper uses the previously reported x - ray and radio proper motion measurements of r97 and k10 .
also , our results are based on 2d hydrodynamical simulations instead of a thin - shell approximation that is strictly valid only for spherically - symmetric remnants .
whatever the status of the companion star in tycho , we stress that a circular morphology does not guarantee an explosion site in the center of the remnant .
new observations of tycho with _
chandra _ in 2015 and the vla in 2013/14 have stretched the baselines for proper motion measurements of the forward shock to 12 - 15 year and @xmath39 year , respectively .
we applied a self - consistent approach to both data sets , refining previous proper motion measurements .
the shock velocity in tycho varies by approximately a factor of two from one side of the remnant to the other , consistent with previous measurements , and with density variations inferred from _ spitzer _ observations .
this leads to an offset of about 10% of the remnant s radius between the geometric center of the remnant and the site of the explosion . despite the circular appearance of tycho ,
offsets such as this could exist in remnants of other sne as well , impacting the search for surviving companion stars .
support for this work was provided through chandra award go4 - 15074z issued by the chandra x - ray observatory center , which is operated by the smithsonian astrophysical observatory for and on behalf of nasa under contract nas8 - 03060 .
nrao is a facility of the nsf operated under cooperative agreement by associated universities , inc .
lc acknowledges nsf ast-1412980 . 1 & 15 & 64 & 0.265 & 7.48 ( 9 ) & 0.0104 , 0.0127 & 0.318 & 13.9 ( 13 ) & 4.59 @xmath40 , 9.85 @xmath40 + 2 & 30 & 48 & 0.218 & 8.64 ( 13 ) & 0.0224 , 6.54 @xmath40 & 0.317 & 16.5 ( 18 ) & 4.52 @xmath40 , 8.25 @xmath40 + 3 & 62 & 26 & 0.063 & 36.2 ( 18 ) & 0.0371 , 8.87 @xmath40 & & & + 4 & 80 & 40 & 0.193 & 12.9 ( 12 ) & 0.0104 , 0.0184 & 0.205 & 20.7 ( 22 ) & 8.33 @xmath40 , 0.0145 + 5 & 101 & 30 & 0.243 & 21.3 ( 28 ) & 0.0300 , 0.0187 & 0.273 & 41.1 ( 38 ) & 5.70 @xmath40 , 8.46 @xmath40 + 6 & 118 & 47 & 0.281 & 14.8 ( 15 ) & 0.0194 , 0.0126 & 0.345 & 44.3 ( 38 ) & 5.38 @xmath40 , 4.84 @xmath40 + 7 & 143 & 32 & 0.306 & 17.8 ( 22 ) & 0.0461 , 0.0321 & 0.319 & 33.9 ( 36 ) & 0.0106 , 9.90 @xmath40 + 8 & 153 & 20 & 0.319 & 24.3 ( 30 ) & 0.0463 , 0.0112 & 0.300 & 26.9 ( 28 ) & 6.56 @xmath40 , 0.0138 + 9 & 173 & 35 & 0.333 & 15.7 ( 20 ) & 0.0168 , 7.33 @xmath40 & 0.328 & 36.8 ( 38 ) & 4.26 @xmath40 , 4.59 @xmath40 + 10 & 185 & 30 & 0.337 & 16.1 ( 18 ) & 0.0187 , 3.07 @xmath40 & 0.347 & 16.7 ( 20 ) & 6.23 @xmath40 , 4.85 @xmath40 + 11 & 213 & 44 & 0.336 & 18.3 ( 23 ) & 0.0137 , 9.74 @xmath40 & 0.359 & 39.7 ( 33 ) & 4.92 @xmath40 , 7.54 @xmath40 + 12 & 236 & 47 & 0.334 & 20.2 ( 25 ) & 9.84 @xmath40 , 2.47 @xmath40 & 0.364 & 27.2 ( 24 ) & 7.08 @xmath40 , 0.0120 + 13 & 255 & 42 & 0.298 & 20.6 ( 18 ) & 8.26 @xmath40 , 5.07 @xmath40 & 0.340 & 19.0 ( 22 ) & 9.90 @xmath40 , 0.0122 + 14 & 276 & 20 & 0.337 & 12.2 ( 16 ) & 9.77 @xmath40 , 0.0216 & 0.408 & 21.0 ( 18 ) & 0.0130 , 0.0306 + 15 & 295 & 37 & 0.229 & 17.2 ( 19 ) & 0.0133 , 0.0124 & 0.366 & 121.2 ( 38 ) & 0.0220 , 0.0275 + 16 & 312 & 42 & 0.305 & 28.7 ( 24 ) & 7.63 @xmath40 , 4.34 @xmath41 & 0.311 & 23.6 ( 20 ) & 3.93 @xmath40 , 6.52 @xmath40 + 17 & 333 & 42 & 0.170 & 16.2 ( 12 ) & 0.0147 , 8.48 @xmath40 & 0.242 & 17.7 ( 22 ) & 3.61 @xmath40 , 3.39 @xmath40 + | we present newly obtained x - ray and radio observations of tycho s supernova remnant using _ chandra _ and the karl g. jansky very large array in 2015 and 2013/14 , respectively . when combined with earlier epoch observations by these instruments , we now have time baselines for expansion measurements of the remnant of 12 - 15 year in the x - rays and 30 year in the radio . the remnant s large angular size allows for proper motion measurements at many locations around the periphery of the blast wave .
we find , consistent with earlier measurements , a clear gradient in the expansion velocity of the remnant , despite its round shape .
the proper motions on the western and southwestern sides of the remnant are about a factor of two higher than those in the east and northeast .
we showed in an earlier work that this is related to an offset of the explosion site from the geometric center of the remnant due to a density gradient in the ism , and using our refined measurements reported here , we find that this offset is @xmath0 towards the northeast .
an explosion center offset in such a circular remnant has implications for searches for progenitor companions in other remnants . |
my assignment at the 1999 icfa instrumentation school is to survey some current developments in particle physics , and to describe the kinds of experiments we would like to do in the near future and illustrate the demands our desires place on detectors and data analysis . like any active science ,
particle physics is in a state of continual renewal .
many of the subjects that seem most fascinating and most promising today simply did not exist as recently as twenty - five years ago .
other topics that have preoccupied physicists for many years have been reshaped by recent discoveries and insights , and transformed by new techniques in accelerator science and detector technology . to provide some context for the courses and laboratories at this school ,
i have chosen three topics that are of high scientific interest , and that place very different demands on instrumental techniques .
i hope that you will begin to see the breadth of opportunities in particle physics , and that you will also look beyond the domain of particle physics for opportunities to apply the lessons you learn here in istanbul .
i begin with the remarkable neutrino , a subatomic particle that our instruments must be able to detect both by its presence and by its absence , depending on the circumstances .
in particular , i note the interest in observing and characterizing neutrino oscillations in order to determine the properties of the neutrinos , and describe some of the new instruments contemplated to do that
. then i will talk about physics at the high - energy frontier , focusing on what we hope to learn from the top quark in the next set of experiments at fermilab s tevatron collider .
third , i will tell you some of the ways we hope to explore the landscape of spacetime , and explore some of the signs we might find that the three - plus - one dimensions of ordinary experience are not the whole story .
my treatment of all of these will be schematic ; in a single lecture , my intent is to raise questions and present a wide range of challenges and opportunities .
neutrinos are tiny subatomic particles that carry no electric charge , have ( almost ) no mass , move ( nearly ) at the speed of light , and hardly interact at all .
they are among the most abundant particles in the universe . as you listen to this lecture , inside your body are more than 10 million ( @xmath3 ) relic neutrinos left over from the big bang . each second
, some @xmath4 neutrinos made in the sun pass through you . in one tick of the clock , about a thousand neutrinos made by cosmic - ray interactions in earth s atmosphere
traverse your body .
other neutrinos reach us from natural sources , including radioactive decays of elements inside the earth , and artificial sources , such as nuclear reactors .
our awareness of neutrinos started with a puzzle in 1914 that led to an idea in 1930 that was confirmed by an experiment in 1956 .
today , neutrinos have become an important tool for particle physics and astrophysics , and fascinating objects of study that may yield important new clues about the basic laws of nature .
there was a neutrino puzzle even before physicists knew there was a neutrino .
natural and artificial radioactivity includes nuclear beta ( @xmath5 ) decay , observed as @xmath6 where @xmath7 is the old - fashioned name for an electron and @xmath8 stands for the nucleus with @xmath9 protons and @xmath10 neutrons .
examples are tritium @xmath5 decay , @xmath11 neutron @xmath5 decay , @xmath12 and @xmath5 decay of lead-214 , @xmath13 for two - body decays , the principle of conservation of energy & momentum says that the @xmath5 particle , or electron , should have a definite energy , indicated by the spike in figure [ fig : betaspec ] .
what was observed was very different : in 1914 , james chadwick ( later to discover the neutron ) showed conclusively @xcite that in the decay of radium b and c ( @xmath14pb and @xmath14bi ) , the @xmath5 energy follows a continuous spectrum , as shown in figure [ fig : betaspec ] .
how could we account for this completely unexpected behavior ?
might it mean that energy and momentum are not uniformly conserved in subatomic events ?
although chadwick did not discover the neutron until 1932 , we can use a little chronological license to sharpen our puzzle by considering a cartoon of neutron @xmath5 decay .
the continuous @xmath5 spectrum means that , in general , the products of the decay of a stationary neutron will not have balanced momenta ( zero net momentum ) , as shown in the left - hand frame in figure [ fig : enocons ] . for the products of a system at rest to drift off in some direction flies in the face of physical intuition , though we have to concede that at the time , physicists intuition was still largely derived from macroscopic experience
the @xmath5-decay energy crisis tormented physicists for years . on december 4 , 1930 , wolfgang pauli addressed an open letter - spectrum
'' pauli concluded , `` for the moment i dare not publish anything about this idea and address myself confidentially first to you .
i admit that my way out may seem rather improbable _ a priori _ .
nevertheless , if you do nt play you ca nt win .
therefore , dear radioactives , test and judge . ''
pauli s neutrino , together with the discovery of the neutron , also resolved a vexing nuclear spin - and - statistics problem . ] to a meeting on radioactivity in tbingen .
pauli could not attend in person because his presence at a student ball in zurich was `` indispensable . '' in his letter , pauli advanced the outlandish idea of a new , very penetrating , neutral particle of vanishingly small mass . because pauli s new particle interacted very feebly with matter , it would escape undetected from any known apparatus , taking with it some energy , which would seemingly be lost . the balance of energy and momentum would be restored , as shown in the right - hand frame of figure [ fig : enocons ] , by the particle we now know as the electron s antineutrino . the proper decay scheme for the neutron is thus @xmath15 pauli s new particle was indeed a `` desperate remedy , '' but it was , in its way , very conservative , for it preserved the principle of energy and momentum conservation and with it the notion that the laws of physics are invariant under translations in space and time .
the hypothesis fit the facts ; after the discovery of the neutron in 1932 , fermi named the new particle the neutrino , to distinguish it from the neutron , and constructed his four - fermion theory of the weak interaction .
experimental confirmation of pauli s neutrino had to wait for dramatic advances in technology . detecting a particle as penetrating as the neutrino required a large target and a copious source of neutrinos . in 1953 , clyde cowan and fred reines @xcite used the intense beam of antineutrinos from a fission reactor @xmath16 and a heavy target ( @xmath17 of liquid scintillator ) containing about @xmath18 protons to detect the reaction @xmath19 initial runs at the hanford engineering works were suggestive but inconclusive .
moving their apparatus to the stronger fission neutrino source at the savannah river nuclear plant , cowan and reines and their team made the definitive observation of inverse @xmath5 decay in 1956 @xcite .
in addition to the electron ( @xmath20 ) and its neutrino ( @xmath21 ) , we now recognize two other pairs of pointlike , spin-@xmath22 particles that are not affected by the strong interaction : @xmath23 these particles are known as leptons , from the greek @xmath24 = thin , inspired by the small mass of the electron , muon , and their neutrinos compared with the mass of the proton and neutron , the lightest of the baryons ( from the greek @xmath25 = heavy ) .
the muon neutrino is created in charged pion decay , @xmath26 and neutrino beams produced at accelerators are overwhelmingly muon neutrinos .
the two - neutrino experiment carried out at brookhaven national lab in the early 1960s @xcite demonstrated that the neutrinos produced in pion decays do not initiate inverse @xmath5 decay , so that @xmath27 is distinct from @xmath21 .
the donut ( direct observation of nu - tau ) experiment @xcite under analysis at fermilab is a three - neutrino experiment .
using a prompt neutrino beam in which decays of the charmed - strange meson provide the @xmath28 source , the experimenters aim to observe the reaction @xmath29 i expect to see results from donut in the year 2000 .
neutrinos traverse vast amounts of material .
the fission antineutrinos detected by cowan , reines , and their collaborators have an inverse @xmath5-decay cross section @xmath30 .
accordingly , their interaction length , @xmath31 where @xmath32 is avogadro s number , @xmath33 is the specific gravity of the target , and @xmath34 is the proton fraction of the target nucleons , is very long .
it corresponds to about @xmath35 of water , or a column density of @xmath36 kilotonnes per square centimeter . on average ,
a fission neutrino would traverse more than four light - years of lead the distance from earth to @xmath37-centauri before interacting . at higher energies
, the cross section for the inclusive reaction @xmath38 grows at first @xmath39 , then more slowly
so the interaction length decreases .
the high - energy dependence of the @xmath40 interaction length is shown in figure [ fig : lint ] .
the interaction length of a 100-gev neutrino is approximately 25 million kilometers of water , or about 230 earth diameters .
if you should stand in fermilab s neutrino beam , only one neutrino in @xmath41 will interact in your body .
what are the consequences of the great interaction length ?
first , there is the missing - energy signature for neutrinos .
second , there is the difficulty of detecting neutrino interactions .
third because the other known particles all interact more than neutrinos do we have the possibility of preparing filtered neutrino beams , which we shall discuss in a moment .
fourth , because neutrinos can penetrate great columns of matter , whereas electromagnetic radiation is blocked by a few hundred grams of material , it is appealing to consider the promise of neutrino astronomy for peering into the hearts of dense structures or looking back at the state of the universe before recombination of ions and electrons into neutral atoms . at high - energy accelerators ,
neutrinos are tertiary products of collisions @xmath42 followed by the decays @xmath43 in an evacuated decay space up to a kilometer long .
the decay region is followed by an earthen shield , perhaps made denser by the inclusion of blocks of iron or steel , that absorbs photons , surviving mesons , and other hadrons , and ranges out many of the muons . at fermilab ,
the total length of the neutrino beam line is about @xmath44 . in the most recent study of deeply inelastic neutrino scattering at fermilab , 800-gev protons from the tevatron
deliver @xmath45 neutrinos in 5 pings over 2.5 seconds each minute , spread over the @xmath46 face of the nutev detector .
the detector is made up of 690 t of iron , scintillator , and drift chambers .
events are studied from a `` fiducial volume '' of 390 t. the @xmath45 neutrinos per minute produce about 10 20 events that are recorded for off - line analysis .
experiments of this kind give us our best look at the interior of the proton , and reveal its quark structure in exquisite detail @xcite .
_ neutrinos are very light . _
no one has ever weighed a neutrino .
the best kinematical determinations set upper bounds @xcite on the dominant neutrino species emitted in nuclear beta decay ( @xmath47 ) , charged - pion decay ( @xmath48 at 90% cl ) , and @xmath49-lepton decay ( @xmath50 at 95% cl ) . although there are prospectsfact 99 in lyon by alvaro de rjula , ref .
] for improving these bounds and the measurement of a nonzero mass would constitute a real discovery
they are sufficiently large that it is of interest to consider indirect ( nonkinematic ) constraints from other quarters .
if neutrino lifetimes are greater than the age of the universe , the requirement that neutrino relics from the big bang not overclose the universe leads to a constraint on the sum of neutrino masses . for relatively light neutrinos ( @xmath51 ) , the total mass in neutrinos , @xmath52 where @xmath53 is the number of spin degrees of freedom of @xmath54 plus @xmath55 , sets the scale of the neutrino contribution to the mass density of the universe , @xmath56 .
if we measure @xmath57 as a fraction of the critical density to close the universe , @xmath58 , where @xmath59 is the reduced hubble parameter , then @xmath60 an assumed bound on @xmath61 then implies a bound on @xmath62 . a very conservative bound results from the assumption that @xmath63 : it is that @xmath64 .
recent observations suggest that the total matter density is considerably smaller than the critical density , so that @xmath65 . if we fix @xmath66 and choose the plausible value @xmath67 , then we arrive at the still generous upper bound @xmath68 . taking into account the best ( and model - dependent ) information about the hot- and cold - dark - matter cocktail @xcite needed to reproduce the observed fluctuations in the cosmic microwave background
, it seems likely that cosmology limits @xmath69 .
it is worth remarking that the cosmological desire for hot dark matter has been on the wane .
if neutrinos were exactly massless , neutrino physics would be simple
. there would be no pattern of masses to explain , no neutrino decays , no mixing among lepton generations , and no neutrino oscillations
. the only question would be why ?
if neutrinos do have masses , the electroweak theory has more unexplained parameters : the neutrino masses , three mixing angles , and a -violating phase .
neutrinos may decay , and neutrinos can oscillate from one electroweak species to another . in the quantum world
, particles are waves .
if neutrinos @xmath70 have different masses @xmath71 , each neutrino flavor may be a mixture of different masses .
let us consider two species for simplicity , and take @xmath72 if neutrinos are emitted with a definite momentum @xmath73 , the wave functions corresponding to the two mass eigenstates evolve with different frequencies .
as a consequence , a beam born as pure @xmath27 may evolve a @xmath21 component with time .
if the neutrino momentum is large compared with the neutrino masses , @xmath74 , then the probability for a @xmath21 component to develop in a @xmath27 beam after a time @xmath75 is @xmath76 measuring the propagation distance @xmath77 , approximating the neutrino energy as @xmath78 , and using the conversion factor @xmath79 , we can re - express @xmath80 arising as it does from a slight frequency mismatch , neutrino flavor oscillation is analogous to the beat - frequency phenomenon observed when two tuning forks with _ almost _ the same pitch
are sounded at the same time .
if the forks are sounded individually , producing waves or , we might find it difficult to distinguish the pitches .
but when we sound them together , the sound intensity @xmath81 swells and fades periodically because the two sound waves are different , reflecting the physical difference between the two tuning forks .
the probability that a neutrino born as @xmath27 remain a @xmath27 at distance @xmath82 is @xmath83 the probability for a @xmath27 to metamorphose into a @xmath21 , @xmath84 depends on two parameters related to experimental conditions : @xmath82 , the distance from the neutrino source to the detector , and @xmath85 , the neutrino energy .
it also depends on two fundamental neutrino parameters : the difference of masses squared , @xmath86 , and the neutrino mixing parameter , @xmath87 .
the amplitude of the probability oscillations is given by @xmath87 , as shown in figure [ fig : oscparam ] ; the wavelength of the oscillations is @xmath88 and the oscillation probability is greatest at a distance @xmath89 , where @xmath90 is an integer .
many experiments have now used natural sources of neutrinos , neutrino radiation from fission reactors , and neutrino beams generated in particle accelerators to look for evidence of neutrino oscillation .
the nuclear burning that powers the sun produces neutrinos as well as light and heat @xcite .
overall , a network of reactions we may summarize as @xmath91 leads to the spectrum of neutrinos shown in figure [ fig : sunjb ] .
because solar neutrinos interact so feebly , they can only be detected in a very massive target and detector .
the super - kamiokande detector in japan consists of @xmath92 tonnes of pure water viewed by @xmath93 photomultiplier tubes to detect cherenkov light .
it is sited @xmath94 under a mountain , under @xmath95 .
the great advantage of the super - k detector is that it detects neutrino interactions @xmath96 in real time , and determines the neutrino direction from the electron direction .
super - k has demonstrated that the brightest object in the neutrino sky is the sun , which proves that nuclear fusion powers our star .
the disadvantage of the water - cherenkov technique ( see figure [ fig : sunjb ] ) is that it is only sensitive to the highest - energy solar neutrinos .
five solar - neutrino experiments report deficits with respect to the predictions of the standard solar model : kamiokande and super - kamiokande using water - cherenkov techniques , sage and gallex using chemical recovery of germanium produced in neutrino interactions with gallium , and homestake using radiochemical separation of argon produced in neutrino interactions with chlorine .
these results suggest the oscillation @xmath97 .
cosmic rays that interact in earth s atmosphere produce neutrinos in the decays of pions , kaons , and muons , in the approximate proportions @xmath98 .
five atmospheric - neutrino experiments report anomalies in the arrival of muon neutrinos : kamiokande , imb , and super - kamiokande using water - cherenkov techniques , and soudan ii and macro using sampling calorimetry .
the most striking result is the zenith - angle dependence of the @xmath27 rate reported last year by super - k @xcite , which is shown in figure [ fig : atmo ] .
the electron - like events follow the monte carlo simulation , but the muon - like events exhibit a deficit that is most pronounced for upward - coming neutrinos , _
i.e. _ , those for which the flight path is longest , up to @xmath99 .
these results suggest the oscillation @xmath100 .
auxiliary information disfavors the sterile - neutrino ( @xmath101 ) interpretation .
the atmospheric- and solar - neutrino anomalies are both disappearance phenomena .
a single experiment reports the appearance of neutrinos that would not be seen in the absence of neutrino oscillations .
the lsnd experiment @xcite reports the observation of @xmath102-like events is what should be an essentially pure @xmath103 beam produced at the los alamos meson physics facility , suggesting the oscillation @xmath104 .
this result has not yet been reproduced by any other experiment .
a host of experiments have failed to turn up evidence for neutrino oscillations in the regimes of their sensitivity .
these results limit neutrino mass - squared differences and mixing angles . in more than a few cases ,
positive and negative claims are in conflict , or at least face off against each other . over the next five years , many experiments will seek to verify , further quantify , and extend these claims .
our colleagues working to assess the feasibility of very - high - energy muon colliders have given us the courage to think that it may be possible , not too many years in the future , to accumulate @xmath105 ( or even @xmath106 ) muons per year .
it is very exciting to think of the possibilities that millimoles of muons would raise for studies in fundamental physics . from the perspective of a muon collider
, the 2.2-@xmath107s lifetime of the muon presents a formidable challenge .
but if the challenge of producing , capturing , storing , and replenishing many unstable muons can be met , the decays @xmath108 offer delicious possibilities for the study of neutrino interactions and neutrino properties @xcite .
we would have at our disposal for the first time not only neutrino beams of unprecedented intensity , but also controlled beams rich in high - energy electron neutrinos . in a _ neutrino factory _
, the composition and spectra of intense neutrino beams will be determined by the charge , momentum , and polarization of the stored muons .
the beam from a @xmath109 storage ring contains @xmath103 and @xmath21 , but no @xmath27 , @xmath102 , @xmath28 , or @xmath110 .
the neutrino spectra are given by @xmath111\;,\quad \cos\theta = \hat{p}_{\nu}\cdot \hat{s}_{\mu}\;,\ ] ] and @xmath112\ ; , \ ] ] where @xmath113 measures the neutrino energy and @xmath114 specifies the muon s spin direction . at the energies best suited for the study of neutrino oscillations tens of gev , by our current estimates
the muon storage ring is compact .
we could build it at one laboratory , pitched at a deep angle , to illuminate a laboratory on the other side of the globe with a neutrino beam whose properties we can control with great precision . by choosing the right combination of energy and destination , we can tune future neutrino - oscillation experiments to the physics questions we will need to answer , by specifying the ratio of path length to neutrino energy and determining the amount of matter the neutrinos traverse . although we can use each muon decay only once , and we will not be able to select many destinations , we may be able to illuminate two or three well - chosen sites from a muon - storage - ring neutrino source .
that possibility added to the ability to vary the muon charge , polarization , and energy may give us just the degree of experimental control it will take to resolve the outstanding questions about neutrino oscillations .
_ the detector challenge . _ to distinguish oscillations among @xmath21 , @xmath27 , @xmath28 , and a possible fourth , `` sterile , '' neutrino @xmath101 that does not experience weak interactions , we require a target _ cum _ detector of several kilotonnes perhaps several tens of kilotonnes that can distinguish electrons , muons , and taus , and measure their charges .
this is a straightforward requirement for the muons , but the short - lived ( @xmath115 ) @xmath49 and the eager - to - shower electron are more difficult to deal with .
top is a most remarkable particle , even for a quark .
a single top quark weighs @xmath116 , about as much as an atom of gold .
but unlike the gold atom , which can be disassembled into 79 protons , 79 electrons , and 118 neutrons , top seems indivisible , for we discern no structure at a resolution approaching @xmath117 .
top s expected lifetime of about 0.4 yoctosecond ( @xmath118 ) makes it by far the most ephemeral of the quarks .
the compensation for this exceedingly brief life is a measure of freedom : top decays before it experiences the confining influence of the strong interaction . in spite of its fleeting existence ,
the top quark helps shape the character of the everyday world .
the discovery experiments were carried out at fermilab s tevatron , in which a beam of 900-gev protons collides with a beam of 900-gev antiprotons .
creating top - antitop pairs in sufficient numbers to claim discovery demanded exceptional performance from the tevatron , for only one interaction in ten billion results in a top - antitop pair , through the reaction observing traces of the disintegration of top into a @xmath119-quark and a @xmath120-boson , the agent of the weak interaction , required highly capable detectors and extraordinary attention to experimental detail .
both the @xmath119-quark and the @xmath120-boson are themselves unstable , with many multibody decay modes .
the @xmath119-quark s mean lifetime is about @xmath121 .
it can be identified by a decay vertex displaced by a fraction of a millimeter from the production point , or by the low - momentum electron or muon from the semileptonic decays @xmath122 , @xmath123 , each with branching fraction about 10% .
the @xmath120 boson decays after only @xmath124 on average into @xmath125 , @xmath126 , @xmath127 , or a quark and antiquark ( observed as two jets of hadrons ) , with probabilities 1/9 , 1/9 , 1/9 , and 2/3 . the characteristic modes in which @xmath128 production can be sought are shown with their relative weights in figure [ fig : tarte ] .
dilepton events ( @xmath129 @xmath130 , and @xmath131 ) are produced primarily when both @xmath120 bosons decay into @xmath132 or @xmath133 .
events in the lepton + jets channels @xmath134 occur when one @xmath120 boson decays into leptons and the other decays through quarks into hadrons .
another challenge to experiment is the complexity of events in high - energy @xmath135 collisions .
the top and antitop are typically accompanied by scores of other particles .
the discovery experiments scanned @xmath136 events per second .
each detector is an intricate apparatus operated by an international collaboration of about 450 physicists .
the tracking devices , calorimeters , and surrounding iron for muon identification occupy a volume about three stories high and weigh about 5000 tons .
the collider detector at fermilab ( cdf ) , a magnetic detector with solenoidal geometry , profited from its high - resolution silicon vertex detector ( svx ) to tag @xmath119-quarks with good efficiency .
the d detector ( d - zero ) had no central magnetic field , emphasizing instead calorimetric measurement of the energies of produced particles .
a top event from the cdf detector , shown in figure [ fig : cdftop ] , displays the power of the silicon vertex detector to resolve secondary @xmath119 decays at only a small remove from the production vertex .
the d event pictured in figure [ fig : dzerotop ] shows one @xmath120 boson reconstructed from a muon plus missing energy , and a second @xmath120 reconstructed from two jets . _ the detector challenge . _ to select top - quark events with high efficiency , and to measure them effectively , require the ability to decide quickly which few of the @xmath3 events per second in tevatron run 2 ( @xmath137 per second at the lhc ) are interesting ; excellent @xmath119 tagging , resolving secondary vertices within a few tenths of a millimeter of the collision point in a hostile high - radiation environment ; efficient lepton identification and measurement of both electron and muon momenta , even for leptons in jets ; calorimetry that provides good jet - jet invariant mass resolution and a reliable measurement of missing transverse energy ; a hermetic detector . for the moment
, the direct study of the top quark belongs to the tevatron .
early in the next century , samples twenty times greater than the current samples should be in hand , thanks to the increased event rate made possible by fermilab s main injector and upgrades to cdf and d .
boosting the tevatron s energy to @xmath138 per beam will increase the top yield by nearly 40% .
further enhancements to fermilab s accelerator complex are under study . a decade from now , the large hadron collider at cern will produce tops at more than ten thousand times the rate of the discovery experiments . electron - positron linear colliders or muon colliders may add new opportunities for the study of top - quark properties and dynamics .
in the meantime , the network of understanding known as the standard model of particle physics links the properties of top to many phenomena to be explored in other experiments . how can we expect new experiments to extend our knowledge of the top quark ?
tevatron run 2 , to begin in march 2001 , is to accumulate @xmath139 of integrated luminosity .
we anticipate a determination of the top mass to @xmath140 in run 2 , @xmath141 with @xmath142 in run 2@xmath143 or in lhc experiments . combined with a measurement of the @xmath120-boson mass to @xmath144 ,
this measurement will make it possible to infer the higgs - boson mass with increased precision .
it is now possible to begin asking how precisely top fits the profile of anticipated properties in its production and decay . it should be possible to determine the @xmath145 coupling in single - top production : @xmath146 in run 2 , @xmath147 in run 2@xmath143 .
one of the exploratory goals of the new round of experiments will be to search for @xmath128 resonances , rare decays , and deviations from the expected pattern of top decays .
finally , it may prove possible to begin to probe the ephemeral lifetime of top .
on the theoretical front , the large mass of top encourages us to think that the two problems of mass may be linked at the electroweak scale .
it is popular to say that top quarks were produced in great numbers in the fiery cauldron of the big bang some fifteen billion years ago , disintegrated in the merest fraction of a second , and vanished from the scene until my colleagues learned to create them in the tevatron .
that would be reason enough to care about top : to learn how it helped sow the seeds for the primordial universe that evolved into our world of diversity and change .
but it is not the whole story ; it invests the top quark with a remoteness that veils its importance for the everyday world .
the real wonder is that here and now , every minute of every day , the top quark affects the world around us . through the uncertainty principle of quantum mechanics , top quarks and antiquarks
wink in and out of an ephemeral presence in our world .
though they appear virtually , fleetingly , on borrowed time , top quarks have real effects .
quantum effects make the coupling strengths of the fundamental interactions
appropriately normalized analogues of the fine - structure constant @xmath37vary with the energy scale on which the coupling is measured .
the fine - structure constant itself has the familiar value @xmath148 in the low - energy ( or long - wavelength ) limit , but grows to about @xmath149 at the mass of the @xmath150 boson , about @xmath151 .
vacuum - polarization effects make the effective electric charge increase at short distances or high energies .
in unified theories of the strong , weak , and electromagnetic interactions , all the coupling `` constants '' take on a common value , @xmath152 , at some high energy , @xmath153 .
if we adopt the point of view that @xmath154 is fixed at the unification scale , then the mass of the top quark is encoded in the value of the strong coupling @xmath155 that we experience at low energies .
unified theory that contains the standard - model gauge group @xmath156 .
the final result is more general . ] assuming three generations of quarks and leptons , we evolve @xmath155 downwards in energy from the unification scale .
the leading - logarithmic behavior is given by @xmath157 for @xmath158 .
the positive coefficient @xmath159 means that the strong coupling constant @xmath160 is smaller at high energies than at low energies .
this behavior opposite to the familiar behavior of the electric charge is the celebrated property of asymptotic freedom . in the interval between @xmath161 and @xmath162 ,
the slope @xmath163 ( where @xmath164 is the number of active quark flavors ) steepens to @xmath165 , and then increases by another @xmath166 at every quark threshold . at the boundary @xmath167 between effective field theories with @xmath168 and @xmath169 active flavors , the coupling constants @xmath170 and @xmath171 must match .
this behavior is shown by the solid line in figure [ fig : courant ] .
the dotted line in figure [ fig : courant ] shows how the evolution of @xmath172 changes if the top - quark mass is reduced .
a smaller top mass means a larger low - energy value of @xmath172 , so a smaller value of @xmath155 .
to discover the dependence of @xmath173 upon the top - quark mass , we calculate @xmath174 evolving up from low energies and down from the unification scale , and match : @xmath175 identifying @xmath176 we find that @xmath177 thanks to qcd , we have learned that the dominant contribution to the light - hadron masses is not the masses of the quarks of which they are constituted , but the energy stored up in confining the quarks in a tiny volume .
our most useful tool in the strong - coupling regime is lattice qcd . calculating the light hadron spectrum from first principles
has been one of the main objectives of the lattice program , and important strides have been made recently . in 1994 ,
the gf11 collaboration @xcitecarried out a quenched calculation of the spectrum ( no dynamical fermions ) that yielded masses that agree with experiment within 510% , with good understanding of the residual systematic uncertainties .
the cp - pacs collaboration centered in tsukuba has embarked on an ambitious program that will soon lead to a full ( unquenched ) calculation @xcite . neglecting the tiny `` current - quark '' masses of the up and down quarks
, the scale parameter @xmath178 is the only mass parameter in qcd .
it determines the scale of the confinement energy that is the dominant contribution to the proton mass . to a good first approximation , @xmath179 where the constant of proportionality @xmath180 is calculable using techniques of lattice field theory .
we conclude that , in a simple unified theory , @xmath181 this is a wonderful result .
now , we ca nt use it to compute the mass of the top quark , because we do nt know the values of @xmath182 and @xmath154 , and havent yet calculated precisely the constant of proportionality between the proton mass and the qcd scale parameter . never mind ! the important lesson
no surprise to any twentieth - century physicist is that the microworld does determine the behavior of the quotidian .
we will fully understand the origin of one of the most important parameters in the everyday world
the mass of the proton only by knowing the properties of the top quark .
ordinary experience tells us that spacetime has @xmath183 dimensions .
could our perceptions be limited ?
that is the question raised allegorically in _ flatland , _ a victorian fable published in 1880 by a british schoolmaster , edwin abbott abbott ( 1839 1926 ) , and still widely available @xcite . like any question that tests our preconceptions and unspoken assumptions , `` what is the dimensionality of spacetime ? ''
is a legitimate scientific question , to which we should return from time to time .
it is given immediacy by recent theoretical work .
for its internal consistency , string theory requires an additional six or seven space dimensions , beyond the @xmath183 dimensions of everyday experience . until recently it has been presumed that the extra dimensions must be compactified on the planck scale , with a compactification radius @xmath184 part of the vision of string theory
is that what goes on in the small curled - up dimensions does affect the everyday world : excitations of the calabi
yau manifolds determine the fermion spectrum , for example .
the great gap between the electroweak scale of about @xmath185 and the planck scale of about @xmath186 gives rise to the hierarchy problem of the electroweak theory @xcite .
the conventional approach to new physics has been to extend the standard model to understand why the electroweak scale ( and the mass of the higgs boson ) is so much smaller than the planck scale .
a novel approach that has evolved over the past two years is instead to _ change gravity _ to understand why the planck scale is so much greater than the electroweak scale @xcite .
now , experiment tells us that gravitation closely follows the newtonian force law down to distances on the order of @xmath187 .
let us parameterize deviations from a @xmath188 gravitational potential in terms of a relative strength @xmath189 and a range @xmath190 , so that @xmath191\ ; , \ ] ] where @xmath192 is the mass density of object @xmath193 and @xmath194 is the separation between body 1 and body 2 .
elegant experiments that study details of casimir and van der waals forces imply bounds on anomalous gravitational interactions , as shown in figure [ fig : nonngrav ] .
below about a millimeter , the constraints on deviations from newton s inverse - square force law deteriorate rapidly , so we are free to consider changes to gravity even on a small but macroscopic scale .
that is precisely the possibility raised by the interesting new idea that some extra dimensions of spacetime might be relatively speaking large .
the new idea is to consider that the @xmath195 standard - model gauge fields , plus needed extensions , reside on @xmath183-dimensional branes , not in the extra dimensions , but that gravity can propagate into the extra dimensions .
how does this hypothesis change the picture ?
the dimensional analysis ( gauss s law , if you like ) that relates newton s constant to the planck scale changes .
if gravity propagates not only in the @xmath183 dimensions of minkowski space , but also in @xmath169 extra dimensions with radius @xmath196 , then @xmath197 where @xmath198 is gravity s true scale . the correlation between @xmath198 and the size @xmath196 of the @xmath169 large extra dimensions is given in table [ tbl : ed ] for some representative cases . .radius
@xmath196 of @xmath169 large extra dimensions for low values of the scale @xmath198 of gravity . [ cols="^,^,^,^,^",options="header " , ] [ tbl : mpp ] if a missing - energy signal is found , it will be a challenge to distinguish an extradimensional signal from the classic @xmath196-parity conserving signature for supersymmetry .
we should be so lucky ! we are only beginning to explore the possible implications of almost - accessible extra dimensions . among the fascinating new worlds to explore are the randall sundrum mechanism for localizing gravity in an extra dimension ( in the vicinity of a brane or domain wall ) @xcite , and the speculation of arkani - hamed and schmaltz that the fermion mass hierarchy reflects fermion wave packets separated in an extra dimension @xcite . the characteristic missing - energy signature of a kauza klein excitation will be hard to distinguish from other new physics .
we need to think in more detail about the backgrounds and optimal search techniques .
the search for large extra dimensions reinforces the informative metaphor of a collider and its detectors as an ultramicroscope .
are extra dimensions large enough to see ?
it is also interesting to consider what might we observe above the threshold for exciting extra dimensions .
although the basic concepts that underlie these speculations have a sound basis in string theory , most of the scenarios put forward so far have more to do with storytelling than with theoretical rigor . however
, it is plain that our inability to disprove at once the outlandish idea of large extra dimensions is a measure of the potential for experimental surprises , and an indication of nature s capacity to amaze us !
to a great degree , the progress of particle physics has followed from progress in accelerator science and instrumentation .
there is no substitute for experiment , and experiment requires inventions in both hardware and software and continuous innovation in analysis .
the slogan , `` yesterday s sensation is today s calibration and tomorrow s background , '' embodies both the challenge and the opportunity of advances in experimental technique . in the middle of the revolution we are experiencing indeed , making in our conception of nature , when we deal with fundamental questions about our world , including _ _ _ _
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ what are the symmetries of nature , and how are they hidden from us ? _ _ are the quarks and leptons composite ? _ _ are there new forms of matter , like the superpartners suggested by supersymmetry ? _ _ are there more fundamental forces ? _ _ what makes an electron an electron , a neutrino a neutrino , and a top quark a top quark ? _ _ what is the dimensionality of spacetime ? _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ we can not advance without new instruments that extend our senses and allow us to create and understand new experience far beyond the realm of everyday human experience .
i wish you a productive two weeks in istanbul , and hope that what you learn about instrumentation , the spirit of experimentation , and the habits of mind necessary to make detectors act as reliable extensions of our senses will open new horizons for you and for science .
it is a pleasure to thank our istanbul hosts for their energetic and delightful hospitality .
i commend the panel on instrumentation innovation and development of the international committee for future accelerators for their sponsorship of this series of instrumentation schools , and salute the leaders of the instrumentation courses for their infectious enthusiasm . on behalf of the students ,
i thank the great laboratories of particle physics , particularly cern and fermilab , for providing much of the apparatus that makes possible the hands - on experience that gives this school its special character .
finally , i would like to express my thanks to the students who came to istanbul from many interesting parts of the globe and whose lively curiosity made my visit memorable .
i hope i may have the opportunity to welcome many of you to fermilab in the future .
pais , a. , `` introducing atoms and their nuclei , '' in _ twentieth century physics , _ edited by brown , l. m. , pais , a. , and pippard , b. ( institute of physics publishing , bristol & philadelphia ; american institute of physics , new york , 1955 ) , vol .
ii , p. 43 .
turner , m. s. `` dark matter and energy in the universe '' astro - ph/9901109 , to be published in _
physica scripta _ ( proceedings of the nobel symposium , particle physics and the universe ; enkoping ,
sweden , august 20 - 25 , 1998 ) , offers @xmath199 .
conrad , j. m. , in _ proceedings of the 29th international conference on high energy physics , _
vancouver , edited by a. astbury , d. axen , and j. robinson ( world scientific , singapore , 1999 ) , p. 25
( hep - ex/9811009 ) , and `` where in the world is the oscillating neutrino ? ''
talk at inner space / outer space 1999 and panic99 , available at ; fisher , p. , kayser , b. , and mcfarland , k. s. , `` neutrino mass and oscillation , '' _ annu .
nucl . part .
sci . _ * 49 , * 481 ( 1999 ) , hep - ph/9906244 ; dilella , l. , `` accelerator and reactor neutrino experiments , '' hep - ex/9912010 .
simmons , e. h. , `` thinking about top : looking outside the standard model , '' hep - ph/9908511 .
willenbrock , s. , `` thinking about top within the standard model , '' hep - ph/9905498 .
georgi , h. , quinn , h. r. , and weinberg , s. , .
antoniadis , i. , `` a possible new dimension at a few tev , '' .
lykken , j. d. , `` weak scale superstrings , '' , hep - th/9603133 .
arkani - hamed , n. , dimopoulos , s. , and dvali , g. , , `` the hierarchy problem and new dimensions at a millimeter,''hep - ph/9803315 .
ignatios antoniadis , i. , arkani - hamed , n. , dimopoulos , s. , and dvali , g. , `` new dimensions at a millimeter to a fermi and superstrings at a tev , '' , hep - ph/9804398 .
antoniadis , i. , benakli , k. , quirs , m. , `` production of kaluza
klein states at future colliders , '' hep - ph/9403290 .
dienes , k. r. , dudas , e. , and gherghetta , t. `` extra space - time dimensions and unification , '' hep - ph/9803466 .
giudice , g. , rattazzi , r. , and wells , j. , `` quantum gravity and extra dimensions at high - energy colliders , '' , hep - ph/9811291 .
nussinov , s. , and shrock , r. , `` some remarks on theories with large compact dimensions and tev scale quantum gravity , '' hep - ph/9811323 .
hewett , j. l. , `` indirect collider signals for extra dimensions , '' hep - ph/9811356 .
arkani - hamed , n. and schmaltz , m. , `` hierarchies without symmetries from extra dimensions , '' hep - ph/9903417 ; eugene a. mirabelli , e. a. and schmaltz , m. , `` yukawa hierarchies from split fermions in extra dimensions , '' hep - ph/9912265 . | i sketch some pressing questions in several active areas of particle physics and outline the challenges they present for the design and operation of detectors . boxedeps # 10= 0=0 1= 1=1 0>1 # 1 # 1#1| # 1| # 1 # 1#1| # 1| # 1 # 1#1_0 # 1| # 1| # 1#2#3 @xmath0 #
1#2#3#4 @xmath1 # 1#2#3#4#5#6 @xmath2 # 1#2#3 , # 2 ( 19#3 ) # 1#2#3 , # 2 ( # 3 ) # 1#2#3 , # 2 ( 19#3 ) # 1#2#3 , # 2 ( # 3 ) # 1#2#3 , # 2 ( 19#3 ) # 1#2#3 , # 2 ( 19#3 ) # 1#2#3 , # 2 ( 19#3 ) # 1#2#3 , # 2 ( 19#3 ) # 1#2#3 , # 2 ( 19#3 ) # 1#2#3 , # 2 ( 19#3 ) # 1#2#3 _ annals of physics ( new york ) _ * # 1 * , # 2 ( 19#3 ) # 1#2#3 , # 2 ( 19#3 ) # 1#2#3 , # 2 ( 19#3 ) # 1#2#3*#1 * , # 2 ( 19#3 ) # 1#2#3 , # 2 ( 19#3 ) # 1#2#3_nucl .
phys .
b ( proc . supp . ) _ * # 1 * ( 19#3 ) # 2 # 1#2#3#4 , # 2 ( # 3january februarymarchaprilmayjunejulyaugust septemberoctobernovemberdecember , 19#4 ) |
magnetization reversal in a single - domain ferromagnetic nanostructure is an important phenomenon for both fundamental physics and applications .
the conventional method for reversing magnetization is to apply a direct magnetic field to a ferromagnet along the reversed direction , where the field magnitude @xmath0 should be larger than the anisotropy field ( or coercivity ) @xmath1 to energetically stabilize the reversed state @xcite .
however , this method requires a large field @xmath0 anti - parallel to the magnetization , as well as large power consumption , for the reversal because ferromagnets with large @xmath1 are used in practical applications to keep the high thermal stability @xmath2 , where @xmath3 , @xmath4 , and @xmath5 are the magnetization , the volume of the ferromagnet , and the temperature , respectively .
recently alternative methods , such as spin - torque - induced magnetization reversal @xcite and microwave - assisted magnetization reversal ( mamr ) @xcite , have been proposed to reduce the reversal field magnitude .
the optical magnetization reversal with circularly polarized light @xcite is another possibility , where the combination of the ultrafast heating and the magnetic field , both of which are generated by the circularly polarized laser , enables the ultrafast magnetization reversal without an external field . in mamr ,
the microwave produces a circularly rotating magnetic field applied to the ferromagnet , in which the field direction lies in a plane perpendicular to the easy axis .
a rotating frame is conventionally used to understand why the reversal field becomes smaller than @xmath1 in mamr @xcite . in the rotating frame ,
the field acting on the magnetization is independent of time .
the effect of the rotating field is converted to an additional dc magnetic field @xmath6 pointing in the reversed direction @xcite , where @xmath7 and @xmath8 are the frequency of the rotating field and the gyromagnetic ratio , respectively , i.e. , the total dc magnetic field pointing in the reversed direction is @xmath9 .
the additional field @xmath6 energetically stabilizes the reversed state , and reduces the reversal field magnitude . in a low - frequency region ,
the reversal field linearly decreases as the frequency increases , which is qualitatively consistent with this conventional picture .
however , both the experiments and the numerical simulations of the landau - lifshitz - gilbert ( llg ) equation have revealed that such a conventional picture can not explain the dependence of the reversal field on the frequency in a high - frequency region @xcite . in the high - frequency region ,
the reversal field slightly increases as the frequency increases ; see , for example , fig .
[ fig : fig5 ] below .
moreover , the magnitude of the total dc magnetic field for the reversal , @xmath9 , in the high - frequency region is much larger than @xmath1 .
this result seems like in contradiction with the stoner - wohlfarth theory @xcite , in which the magnetization reversal should occur when @xmath9 becomes slightly larger than @xmath1 because only the reversed state is energetically stable .
okamoto _ et al .
_ studied the dependence of the reversal field on the frequency of the rotating field for a co / pt nanodot with 120 nm diameter both experimentally and numerically based on the micromagnetic model @xcite .
they found that the excitation of the spin wave , which arises from the difference in the local demagnetization field between the end and center of the dots , leads to a reduction of the reversal field .
this result is of great advance in understanding the reversal mechanism of mamr .
however , it is still unclear why the reversal field jumps to a high value at a certain frequency .
moreover , the numerical simulations based on the macrospin ( single domain ) model also show the jump of the reversal field @xcite , indicating that not only the excitation of the spin wave but also other mechanisms lead to this jump . a fabrication of the ferromagnet smaller than the exchange length ( typically @xcite , on the order of 10 nm ) is an unavoidable and indispensable challenge for both fundamental physics and practical applications .
in such nanostructure , the magnetization dynamics is well described by the macrospin model .
therefore , it is important to clarify the magnetization reversal mechanism , such as the origin of the jump of the reversal field , using the macrospin model .
the purposes of this paper are to explain why a large field is required to reverse the magnetization in the high - frequency region in mamr and to derive equations that determine the reversal field in both the low- and high - frequency regions without the time - dependent solution of the macrospin llg equation .
it is pointed out that the rotating field not only energetically stabilizes the reversed state but also produces a torque acting on the magnetization .
we find that this torque , whose strength is proportional to the frequency of the rotating field , prevents the reversal , causing the reversal field to become relatively large in the high - frequency region .
the direction of this preventing torque is expressed by the triple vector product , analogous to the spin torque @xcite .
this fact motivates us to use the energy balance equation for the estimation of the reversal field of mamr , which was recently pointed out to be useful for estimating the reversal current in the spin - torque reversal problem @xcite but has been never applied to the mamr problem .
the equations determining the reversal field , eqs .
( [ eq : condition_low ] ) and ( [ eq : condition_high ] ) , are derived for both the low- and high - frequency regions from the energy balance equation .
these formulas show that the reversal field in the low - frequency region converges to eq .
( [ eq : condition_low_zero_damping ] ) as the damping constant decreases , while the reversal field in the high - frequency region is independent of the damping constant .
the boundary between the low- and high - frequency regions is also estimated from the energy balance equation .
the validities of these formulas are quantitatively confirmed by comparison with the numerical simulation of the llg equation .
the paper is organized as follows . in sec .
[ sec : landau - lifshitz - gilbert equation in rotating frame ] , the energy balance equation is derived from the llg equation in the rotating frame . the equations determining the reversal fields in the low- and high - frequency regions
are derived in secs [ sec : reversal in low frequency region ] and [ sec : reversal in high frequency region ] , respectively .
the validities of these formulas over a wide frequency range are confirmed in sec .
[ sec : comparison with numerical simulation ] by comparison with the numerical simulation of the llg equation . in sec .
[ sec : comparison with other work ] , the present result is compared with the previous work in refs . @xcite . the conclusion appears in sec .
[ sec : conclusion ] .
the external magnetic field pointing in the negative @xmath10 direction and the rotating field with the frequency @xmath7 are denoted as @xmath0 and @xmath11 , respectively . in the rotating frame ,
@xmath12 and @xmath13 axes are parallel to the rotating field and the @xmath10 axis , respectively . ] the system we consider is schematically shown in fig .
[ fig : fig1 ] , where the unit vector pointing in the magnetization direction is denoted as @xmath14 .
the ferromagnet has a uniaxial easy axis with the anisotropy field @xmath1 along the @xmath10 axis . throughout this paper ,
the initial state is taken to be @xmath15 , although the following formulas are applicable to an arbitrary initial condition .
the external field @xmath0 is applied to the negative @xmath10 direction .
the rotating field with the magnitude @xmath11 and the frequency @xmath7 is applied in the @xmath16 plane , where the @xmath17 axis is parallel to the rotating field at the initial time @xmath18 .
the magnetization dynamics under the effect of the magnetic field , @xmath19 are described by the llg equation @xcite , @xmath20 where the gilbert damping constant is denoted as @xmath21 . because the llg equation conserves the magnitude of the magnetization ,
the magnetization dynamics are described by the trajectory on the surface of the unit sphere .
it is convenient to use the rotating frame @xmath22 , in which the @xmath13 axis is parallel to the @xmath10 axis , and @xmath12 axis is parallel to the rotating field @xmath11 , as shown in fig . [
fig : fig1 ] .
we denote @xmath14 in the rotating frame as @xmath23 .
it should be noted that the value of @xmath24 is invariant by this transformation . because the value of @xmath21 in the conventional ferromagnet is small @xcite , higher order terms of @xmath21
are neglected in the following ; i.e. , we use the approximation that @xmath25 .
then , the llg equation in the rotating frame is given by @xmath26 where @xmath27 \mathbf{e}_{z^{\prime}}$ ] can be regarded as the magnetic field in the rotating frame . the transformation procedure from eq .
( [ eq : llg_orig ] ) to eq .
( [ eq : llg ] ) is shown in appendix a. it can be understood from eq .
( [ eq : llg ] ) that the rotating field plays two roles for the reversal .
first , the magnetization dynamics can be regarded as a motion of a point particle in the potential @xmath28 , @xmath29 the second term on the right - hand side of eq .
( [ eq : potential ] ) indicates that the rotating field produces the dc magnetic field @xmath6 pointing in the negative @xmath13 direction , and energetically stabilizes the reversed state @xcite .
second , the rotating field produces a torque proportional to the frequency @xmath7 , which appears in the third term on the right - hand side of eq .
( [ eq : llg ] ) .
the important point is that this torque points to the positive @xmath13 direction , and therefore , prevents the reversal .
it should also be emphasized that the torque direction is expressed by the triple vector product , as is similar to the spin torque @xcite .
therefore , in the following calculations , let us conventionally call this torque spin torque .
it was shown in the spin - torque reversal problem @xcite that the magnetization reversal condition can be derived from the energy balance equation between the work done by spin torque and the energy dissipation due to the damping . in the following sections , we apply this method to investigate the reversal field in mamr . to this end ,
the derivative of @xmath30 with respect to time on the constant energy curve is necessary . from eq .
( [ eq : llg ] ) , we find that @xmath31 \\ & - \alpha \gamma m \left [ \bm{\mathcal{b}}^{2 } - \left ( \mathbf{m}^{\prime } \cdot \bm{\mathcal{b } } \right)^{2 } \right ] .
\label{eq : dedt } \end{split}\ ] ] the integral of eq .
( [ eq : dedt ] ) over a precession period of the magnetization on the constant energy curve of @xmath30 is @xmath32 , where @xmath33 , \label{eq : w_s}\ ] ] @xmath34 , \label{eq : w_alpha}\ ] ] are the work done by spin torque and the energy dissipation due to the damping during the precession , respectively @xcite .
while @xmath35 can be both positive and negative depending on the field and the frequency , @xmath36 is always negative
. the calculation procedures of eqs .
( [ eq : w_s ] ) and ( [ eq : w_alpha ] ) without the time - dependent solution of eq .
( [ eq : llg ] ) are shown in appendix b. the damping constant @xmath21 is assumed to be scalar in the above formulation .
on the other hand , safonov studied the magnetization relaxation near equilibrium with a tensor damping @xcite .
the presence of the tensor damping was also suggested in the spin torque problem @xcite . in appendix c ,
we discuss how the formulas derived in the following sections are modified when the scalar damping @xmath21 is replaced by the tensor damping .
in this section , we study the reversal field in the low - frequency region .
let us first show in fig .
[ fig : fig2 ] ( a ) the trajectory of typical magnetization dynamics in the low - frequency region obtained by numerically solving eq .
( [ eq : llg ] ) .
the time evolution of @xmath37 is shown in fig .
[ fig : fig2 ] ( b ) .
the values of the parameters are @xmath38 emu / c.c . , @xmath39 koe , @xmath40 oe , @xmath41 rad/(oe@xmath42s ) , @xmath43 ghz , and @xmath44 , which are typical values used in the experiments and the numerical simulations @xcite .
we judged that the magnetization is reversed when the condition @xmath45 is satisfied .
the minimum field satisfying this condition is @xmath46 koe .
starting from the initial state @xmath47 , the magnetization precesses around an axis lying in the positive @xmath48 plane .
after a half period of precession , the magnetization changes the precession direction , and falls into the reversed state .
koe , where the frequency of the rotating field is @xmath43 ghz .
the trajectory is described on the unit sphere .
( b ) the time evolution of @xmath37 .
( c ) schematic view of the potential @xmath30 ( normalized by @xmath49 in the @xmath48 plane .
the horizontal axis represents the angle @xmath50 of the magnetization from the @xmath13 axis .
( d ) the enlarged view of panel ( c ) around @xmath51 $ ] .
the directions of the damping and spin torque are indicated by arrows . ]
next , let us analytically derive the equation determining the reversal field .
figure [ fig : fig2 ] ( c ) shows the map of the potential @xmath30 in the @xmath48 plane , where the horizontal axis is the angle @xmath50 of the magnetization from the @xmath13 axis . the potential with @xmath46 koe has metastable , saddle , and stable points at @xmath52 , @xmath53 , and @xmath54 , respectively .
the directions of the damping and the spin torque between the initial state ( @xmath55 ) and the saddle point in the potential @xmath30 are indicated by the arrows in fig .
[ fig : fig2 ] ( d ) .
the spin torque supplies the energy to the ferromagnet when the magnetization is in @xmath56 because it is anti - parallel to the damping , while the spin torque dissipates the energy when the magnetization is in @xmath57 because it is parallel to the damping .
the function @xmath35 in @xmath58 is negative ( @xmath59 ) because the spin torque magnitude ( @xmath60 ) increases as the angle @xmath61 increases .
therefore , the spin torque totally dissipates the energy during the precession .
the magnetization reversal occurs when the magnitude of the energy @xmath62 dissipated during the dynamics from @xmath63 to @xmath64 is smaller than the energy difference between the initial state and the saddle point , @xmath65 , i.e. , the reversal condition is @xmath66 strictly speaking , the time - dependent solution of @xmath67 is necessary to calculate @xmath68 .
however , in the low - frequency region , because the difference between @xmath69 and @xmath70 is small , @xmath68 can be approximated to @xmath71/2 $ ] .
the numerical factor @xmath72 appears because the reversal occurs after the half period of the precession .
therefore , the reversal field @xmath73 can be defined as the field @xmath0 satisfying the condition @xmath74 .
\label{eq : condition_low}\ ] ] equation ( [ eq : condition_low ] ) is the main result in this section . the reversal field estimated from eq .
( [ eq : condition_low ] ) is 4.708 koe , which is almost identical to that ( 4.709 koe ) obtained from the numerical solution of the llg equation ( [ eq : llg ] ) .
it should be noted that the approximation @xmath75/2 $ ] works well for small @xmath21 because the magnetization dynamics occurs almost on the constant energy curve when @xmath76 . in the zero - damping limit
, the reversal field is estimated from the condition @xmath77 , and is given by @xmath78 where @xmath64 depends on @xmath73 through the condition @xmath79\sin\theta_{\rm saddle}-h_{\rm k } \sin\theta_{\rm saddle}\cos\theta_{\rm saddle}=0 $ ] . with @xmath80 .
the arrows indicate the directions of the spin torque and the damping . ]
let us also elaborate on the frequency range in which eq .
( [ eq : condition_low ] ) obtains good agreement with the numerical solution of the llg equation ( [ eq : llg ] ) . at a certain field magnitude
@xmath81 , which is larger than @xmath73 estimated by eq . (
[ eq : condition_low ] ) , the metastable state disappears , and the potential has only one stable point and a saddle point .
we denote the potential @xmath30 with @xmath82 as @xmath83 , which satisfies @xmath84 at the saddle point @xmath85 .
according to the conventional stoner - wohlfarth theory @xcite , the magnetization reversal should occur because the potential has only one minimum .
however , in the present case , the work done by spin torque , @xmath86 , on the constant energy curve of @xmath87 , is positive because the direction of the spin torque is always opposite that of the damping , as shown in fig .
[ fig : fig3 ] .
then , the condition @xmath88 should also be satisfied to reverse the magnetization : if eq .
( [ eq : condition_boundary ] ) is not satisfied , the spin torque preventing the reversal overcomes the damping , and thus the magnetization can not reverse its direction .
it is found that eq .
( [ eq : condition_boundary ] ) is satisfied for @xmath89 ghz for the above parameters .
therefore , we define the low - frequency region in which eq .
( [ eq : condition_low ] ) is valid as @xmath89 ghz . because the reversal field discontinuously becomes large above this frequency
, we call this frequency the jump frequency .
it should be emphasized that the jump frequency is independent of @xmath90 .
in this section , we study the magnetization reversal in the high - frequency region .
as mentioned in the previous section , for @xmath91 ghz , the spin torque preventing the reversal becomes sufficiently large .
then , a large field is required to reverse the magnetization .
figure [ fig : fig4 ] ( a ) shows the trajectory of typical magnetization dynamics in the high - frequency region obtained by numerically solving eq .
( [ eq : llg ] ) .
the time evolution of @xmath37 is shown in fig .
[ fig : fig4 ] ( b ) .
the values of the parameters are those used in sec .
[ sec : reversal in low frequency region ] except @xmath92 ghz .
the minimum field satisfying the condition @xmath45 is @xmath93 koe .
starting from the initial state , the magnetization precesses on the constant energy curves near the @xmath13 axis many times .
the precession amplitude slightly increases with time , and finally the magnetization reverses to @xmath94 .
the reversal trajectory covers almost all of the unit sphere , as shown in fig .
[ fig : fig4 ] ( a ) .
koe , where the frequency of the rotating field is @xmath92 ghz .
the trajectory is described on the unit sphere .
( b ) the time evolution of @xmath37 .
( c ) schematic view of the potential @xmath30 ( normalized by @xmath49 in the @xmath48 plane .
the horizontal axis represents the angle @xmath50 of the magnetization from the @xmath13 axis .
( d ) the enlarged view of @xmath30 around @xmath95 .
the directions of the damping and spin torque are indicated by arrows . ]
figure [ fig : fig4 ] ( c ) shows the potential @xmath30 in the @xmath48-plane at the reversal field . according to the stoner - wohlfarth condition @xcite @xmath96^{2/3}+h_{\rm ac}^{2/3}$ ] , a field @xmath0 larger than @xmath97 koe is enough to reverse the magnetization , above which the potential @xmath30 has only one minimum .
nevertheless , a large field ( @xmath98 koe ) compared with the stoner - wohlfarth condition is required for the reversal in the present case because the spin torque prevents the reversal .
figure [ fig : fig4 ] ( d ) shows the enlarged view of @xmath30 near the initial state , where the arrows indicate the directions of the damping and the spin torque .
the maximum of @xmath30 is located at @xmath99 while the angle satisfying @xmath100 is located at @xmath101 . for a reason similar to that discussed in sec .
[ sec : reversal in low frequency region ] , the function @xmath102 is positive . when @xmath103 , the magnetization loses the energy , and falls into the reversed state . on the other hand , when @xmath104 , the magnetization climbs @xmath30 from @xmath63 to @xmath105 .
therefore , the reversal field can be defined as the field satisfying the condition @xmath106 equation ( [ eq : condition_high ] ) is the main result in this section .
the reversal field estimated from eq .
( [ eq : condition_high ] ) is 7.109 koe , which is identical to that obtained from the numerical solution of the llg equation ( [ eq : llg ] ) .
another important conclusion from eq .
( [ eq : condition_high ] ) is that the reversal field is independent of @xmath21 because both @xmath35 and @xmath36 are proportional to @xmath21 . the validity of this conclusion is investigated in the next section .
in this section , we confirm the validities of eqs .
( [ eq : condition_low ] ) and ( [ eq : condition_high ] ) over a wide range of the frequency @xmath7 . the circles in fig .
[ fig : fig5 ] show the reversal field estimated numerically solving the llg equation ( [ eq : llg ] ) , where the frequency range is @xmath107 ghz .
the gilbert damping constant is @xmath108 .
the reversal field magnitude linearly decreases as the frequency increases for @xmath109 ghz . above @xmath110 ghz
, the reversal field jumps to a high value at which @xmath111 , and slightly increases as the frequency increases .
the reversal fields obtained from eqs .
( [ eq : condition_low ] ) and ( [ eq : condition_high ] ) are also shown in fig .
[ fig : fig5 ] by the solid lines . as
mentioned in sec .
[ sec : reversal in low frequency region ] , eq .
( [ eq : condition_low ] ) is valid for @xmath89 ghz .
therefore , we use eq .
( [ eq : condition_low ] ) for @xmath89 ghz while eq .
( [ eq : condition_high ] ) is used for @xmath91 ghz .
equations ( [ eq : condition_low ] ) and ( [ eq : condition_high ] ) show good agreement with the numerical solution of the llg equation ( [ eq : llg ] ) , indicating the validities of these formulas . ) while the solid lines are obtained from eqs .
( [ eq : condition_low ] ) and ( [ eq : condition_high ] ) .
the value of @xmath21 is @xmath108 .
the dashed line is the reversal field estimated from ref .
@xcite , and is discussed in sec . [
sec : comparison with other work ] . ] , @xmath108 , and @xmath112 . the symbols ( square , circle , and triangle ) are obtained from eq .
( [ eq : llg ] ) while the lines ( dotted , solid , and dashed ) are obtained from eqs .
( [ eq : condition_low ] ) and ( [ eq : condition_high ] ) .
( b ) the enlarged view of panel ( a ) in the low - frequency region . ]
figure [ fig : fig6 ] ( a ) shows the dependences of the reversal fields on the frequency for @xmath113 , @xmath108 , and @xmath112 , where the numerical solution of the llg equation ( [ eq : llg ] ) is represented by the symbols ( square , circle , and triangle , respectively ) , while the reversal fields obtained from eqs .
( [ eq : condition_low ] ) and ( [ eq : condition_high ] ) are represented by the lines ( dotted , solid , and dashed , respectively ) .
the frequency ( @xmath114 ghz ) at which the reversal field of the llg equation ( [ eq : llg ] ) jumps to a high value is independent of @xmath21 , which is consistent with the discussion in sec .
[ sec : reversal in low frequency region ] .
the enlarged view in the low - frequency region is shown in fig . [
fig : fig6 ] ( b ) . in the low - frequency region ,
the difference between the solutions of the llg equation ( [ eq : llg ] ) and the energy balance equation ( [ eq : condition_low ] ) becomes small as @xmath21 decreases , because the approximation @xmath115/2 $ ] used in the derivation of eq .
( [ eq : condition_low ] ) is valid for a sufficiently small @xmath21 .
also , the reversal field becomes independent of @xmath21 with decreasing @xmath21 , which is consistent with eq .
( [ eq : condition_low_zero_damping ] ) . in the high - frequency region ,
the solution of the llg equation ( [ eq : llg ] ) is also independent of @xmath21 , which is consistent with eq .
( [ eq : condition_high ] ) .
these results also imply the validity of eqs .
( [ eq : condition_low ] ) and ( [ eq : condition_high ] ) . .
the circles are obtained from eq .
( [ eq : llg ] ) while the solid line is obtained from eq .
( [ eq : condition_boundary ] ) . ]
as we end this section , we study the relation between the rotating field magnitude @xmath11 and the jump frequency , i.e. , the frequency defining the boundary between the low- and high - frequency regions .
figure [ fig : fig7 ] shows the dependences of the jump frequency on @xmath11 obtained from eq .
( [ eq : condition_boundary ] ) ( solid line ) and the numerical solution of the llg equation ( circle ) .
as shown , the jump frequency monotonically increases with increasing @xmath11 . although the clarification of the relation between the jump frequency and the other parameters such as @xmath11 is desirable , it is difficult to analytically solve eq .
( [ eq : condition_boundary ] ) with respect to the jump frequency .
we consider that the jump frequency does not necessarily relate to , for example , the ferromagnetic resonance ( fmr ) frequency because the jump frequency is determined by the energy balance of the magnetization at the saddle point of the potential @xmath30 while the fmr frequency is the frequency of the harmonic oscillation around the stable point .
however , a further investigation on the jump frequency is beyond the scope of this paper .
in this section , we compare the above result with the previous work of bertotti _ et al .
they expanded the llg equation around its steady - state point , @xmath116 , where @xmath50 and @xmath117 are the zenith and azimuth angles characterizing the magnetization direction , and satisfy the following equations : @xmath118 @xmath119 small deviation of the magnetization , @xmath120 , from the steady points satisfy @xmath121 and @xmath122 , where components of a @xmath123 matrix @xmath124 is obtained from the llg equation . the trace and determinant of @xmath124
are given by @xmath125 = -\frac{2 \alpha}{1+\alpha^{2 } } \left [ v - \frac{\sin^{2}\theta}{2 } \gamma h_{\rm k } + 2\pi f \cos\theta \right ] , \label{eq : trace_bertotti}\ ] ] @xmath126 = \frac{v^{2}-\gamma h_{\rm k } v \sin^{2}\theta + ( \alpha
2\pi f \cos\theta)^{2}}{1+\alpha^{2 } } , \label{eq : determinant_bertotti}\ ] ] where @xmath127 . according to ref .
@xcite , the reversal field is estimated from eqs .
( [ eq : condition_bertotti_1 ] ) , ( [ eq : condition_bertotti_2 ] ) and ( [ eq : determinant_bertotti ] ) with the condition @xmath128=0 $ ] . the dashed line in fig .
[ fig : fig5 ] is the reversal field estimated from this method . as shown , the method of bertotti _ et al .
_ reveals larger reversal field in our simulation .
the difference between our and their results arises from the following reason . in our analytical and numerical calculations , both the microwave and external field
are applied from @xmath18 with the constant magnitudes .
the initial state of the magnetization , @xmath129 , locates above the stable or saddle point , as shown in fig .
[ fig : fig2 ] ( d ) .
on the other hand , bertotti _ et al .
_ considered the instability of the magnetization around the steady point corresponding to the stable or saddle point . in this case , a relatively large energy compared with our situation is required to overcome the potential barrier and reverse the magnetization direction .
therefore , the reversal field estimated from the method of bertotti _ et al . _ becomes larger than our estimation .
we emphasize that both the results of bertotti _ et al . _ and
our method are useful to estimate the reversal field .
for example , numerical simulation of the llg equation @xcite showed good agreement with the theory of bertotti _ et al .
_ @xcite . in ref .
@xcite , the magnitude of the dc magnetic field is linearly increased with time until it reaches a certain value . in this case
, the magnetization first relaxes to a stable point of the potential , and after that the magnetization reverses its direction when the saturated value of the dc magnetic field is larger than the reversal field estimated by the method of bertotti _ et al .
_ on the other hand , our approach is applicable when the magnitude of the dc magnetic field is fixed from @xmath18 , as mentioned above . to clarify the applicability of the theory of bertotti _ et al .
_ more precisely , an estimation of relaxation time from the initial state to the steady point , which should be shorter than the time to saturate the dc magnetic field magnitude , will be necessary .
in conclusion , we studied the dependence of the reversal field in microwave - assisted magnetization reversal on the frequency of the rotating field theoretically .
the microwave produced a dc magnetic field pointing in the reversed direction , which energetically stabilized the reversed state .
the microwave simultaneously produced a torque proportional to the frequency of the rotating field .
because this torque prevented the reversal , a large field was required to reverse the magnetization in the high - frequency region .
the equations determining the reversal fields in both the low- and high - frequency regions were derived from the energy balance equation .
the formulas showed that the reversal field in the low - frequency region became converged to eq .
( [ eq : condition_low_zero_damping ] ) as the damping constant decreased , while the reversal field in the high - frequency region was independent of the damping constant .
the boundary between the low- and high - frequency regions , which was independent of the damping constant , was also estimated from the energy balance equation .
the comparison with the numerical solution of the landau - lifshitz - gilbert equation showed quantitatively good agreement , guaranteeing the validities of the formula .
the author would like to acknowledge h. imamura , t. yorozu , h. kubota , h. maehara , and s. yuasa for their valuable discussions .
this work was supported by jsps kakenhi grant - in - aid for young scientists ( b ) 25790044 .
the transformation from the laboratory frame to the rotating frame is described by the rotation matrix , @xmath130 for example , the relation between @xmath14 and @xmath67 is given by @xmath131 .
similarly , the field @xmath132 is transformed as @xmath133 , where @xmath134 relates to @xmath135 in eq .
( [ eq : llg ] ) as @xmath136 .
also , @xmath137 should be replaced by @xmath138 .
then , eq . ( [ eq : llg_orig ] ) is transformed as @xmath139 equation ( [ eq : llg_rotate ] ) is equivalent to eq .
( [ eq : llg ] ) . however , for the following reason , we use eq .
( [ eq : llg ] ) instead of eq .
( [ eq : llg_rotate ] ) . as shown in secs .
[ sec : reversal in low frequency region ] and [ sec : reversal in high frequency region ] , a potential map is useful to investigate the magnetization dynamics .
the potential is usually defined as the integral of the field with respect to the magnetization @xcite , where the field appears in both the conservative and the damping torques of the llg equation .
when we use eq .
( [ eq : llg_rotate ] ) , the definition of the potential is not clear because the fields that appeared in the conservative torque , @xmath140 , and in the damping torque , @xmath134 , are different . on the other hand , when we use eq .
( [ eq : llg ] ) , the potential can be well defined as @xmath141 .
therefore , we express the llg equation in the rotating frame in the form of eq .
( [ eq : llg ] ) .
equations ( [ eq : w_s ] ) and ( [ eq : w_alpha ] ) can be calculated without the time - dependent solution of @xmath142 obtained from eq .
( [ eq : llg ] ) . using the conservative torque term of the llg equation
, the integration variable can be transformed from the time @xmath143 to @xmath37 , i.e. , from @xmath144 to @xmath145 , where the numerical factor @xmath146 appears by restricting the integral range to @xmath147 and due to the symmetry of the system with respect to the @xmath48 plane . because the llg equation conserves the magnetization magnitude , @xmath148 appearing in eqs .
( [ eq : w_s ] ) and ( [ eq : w_alpha ] ) can be replaced by @xmath149 . also , from eq .
( [ eq : potential ] ) , @xmath150 can be expressed in terms of @xmath37 as @xmath151.\ ] ] therefore , the integrands in eqs .
( [ eq : w_s ] ) and ( [ eq : w_alpha ] ) can be expressed in terms of @xmath37 only .
the integral range can be determined from eq .
( [ eq : potential ] ) by fixing the value of @xmath30 .
when the damping constant @xmath21 is replaced by the tensor damping , @xmath35 and @xmath36 in eqs .
( [ eq : w_s ] ) and ( [ eq : w_alpha ] ) should be redefined as @xmath152 , \label{eq : w_s_general}\ ] ] @xmath153 , \label{eq : w_alpha_general}\ ] ] where @xmath154 is the tensor damping in the rotating frame , and has nine components @xmath155 ( @xmath156 ) , in general .
the tensor product is defined as , for example , @xmath157 . according to the discussions in secs .
[ sec : reversal in low frequency region ] and [ sec : reversal in high frequency region ] , and using eqs .
( [ eq : w_s_general ] ) and ( [ eq : w_alpha_general ] ) , two conclusions are obtained for the case of the tensor damping .
first , eqs .
( [ eq : condition_low ] ) and ( [ eq : condition_high ] ) are still applicable to estimate the reversal fields in the low- and high - frequency regions , respectively , because the explicit forms of @xmath35 and @xmath36 do not affect the derivation of these equations .
equation ( [ eq : condition_boundary ] ) is also applicable to determine the boundary between the low- and high - frequency regions .
however , the argument that eqs .
( [ eq : condition_boundary ] ) and ( [ eq : condition_high ] ) are independent of the damping constant does not necessarily hold .
this is because the integrals of eqs .
( [ eq : w_s ] ) and ( [ eq : w_alpha ] ) are independent of the scalar damping @xmath21 , and therefore , the frequency or field satisfying eqs .
( [ eq : condition_boundary ] ) or ( [ eq : condition_high ] ) is also independent of @xmath21 , while in the case of the tensor damping , the integrals of eqs .
( [ eq : w_s_general ] ) and ( [ eq : w_alpha_general ] ) depend on the components of @xmath154 , in general .
second , eq . ( [ eq : condition_low_zero_damping ] ) is also valid because this equation is derived by the condition @xmath77 , which is independent of the damping . | the reversal condition of magnetization in a nanomagnet under the effect of rotating magnetic field generated by a microwave is theoretically studied based on the landau - lifshitz - gilbert equation . in a rotating frame ,
the microwave produces a dc magnetic field pointing in the reversed direction , which energetically stabilizes the reversed state .
we find that the microwave simultaneously produces a torque preventing the reversal .
it is pointed out that this torque leads to a jump in the reversal field with respect to the frequency .
we derive the equations determining the reversal fields in both the low- and high - frequency regions from the energy balance equation .
the validities of the formulas are confirmed by a comparison with the numerical simulation of the landau - lifshitz - gilbert equation . |
in a recent article @xcite(intended for an audience of experimentalists ) , the author has reviewed arguments that many aspects of the physics of the cuprate superconductors can be understood by using their proximity to paramagnetic mott insulators .
further , a distinction was made between mott insulators with collinear and non - collinear spin correlations , and it was argued that current experimental evidence suggests that we need only consider the collinear class .
a phenomenological description of the ground states and excitations of these classes of mott insulators was provided , along with a discussion of their experimental implications .
a more technical discussion ( intended for theorists ) of such insulators , along with a description of the effective field theories which describe their low energy properties appears in ref .
@xcite . here
, we briefly recall the derivation and properties of the effective field theory of mott insulators with _ collinear _ spin correlations , which is expressed in terms of a compact u(1 ) gauge field .
the non - collinear class leads naturally to a @xmath0 gauge theory , but we will not consider it here . the reader is referred to these previous reviews @xcite for complete citations to the literature .
we focus on mott insulators on a @xmath1 dimensional bipartite lattice of sites @xmath2 .
the su(2 ) spin operator @xmath3 on site @xmath2 at imaginary time @xmath4 can be written as @xmath5 here @xmath6 on the two sublattices , @xmath7 is the spatial co - ordinate of site @xmath2 , @xmath8 is a unit length vector in spin space , and @xmath9 is the ( integer or half - odd - integer ) angular momentum of each spin .
the antiferromagnetic exchange interaction between near neighbor spins implies that @xmath10 will be a slowly varying function of its spacetime arguments . a standard analysis of the coherent state path integral of over the su(2 )
@xmath3 spins shows that the low energy quantum fluctuations are described by the following partition function @xmath11 , \label{zl}\end{aligned}\ ] ] where @xmath12 is the spin - wave velocity , and @xmath13 is a coupling constant which controls the strength of the quantum fluctuations .
excluding the first berry phase term , this is the action of the o(3 ) non - linear sigma model in @xmath14 spacetime dimensions .
here we are primarily interested in the consequences of the berry phases : @xmath15 is defined to be the oriented area of the spherical triangle defined by @xmath16 , @xmath17 , and an arbitrary reference point @xmath18 ( which is usually chosen to be the north pole ) . the theory ( [ zl ] ) can be considered to be the `` minimal model '' of antiferromagnets . in dimensions @xmath19 it has at least two phases : at small @xmath13 there is the conventional magnetically ordered `` nel '' phase with @xmath20 , while at large @xmath13 there is a `` quantum disordered '' paramagnetic phase which preserves spin rotation invariance with @xmath21 .
we are especially interested here in the nature of this paramagnetic state . in this section
, we will manipulate @xmath22 in this large @xmath13 regime , and derive an alternative formulation which allows easier computation of the integral over the berry phases .
the key to an analysis of the large @xmath13 regime is a better understanding of the nature of @xmath23 .
we will see that @xmath23 behaves in many respects like the time - component of a compact u(1 ) gauge field , and indeed , this accounts for the suggestive notation .
all physical results should be independent of the choice of the reference point @xmath18 , and it is easy to see by drawing triangles on the surface of a sphere that changes in @xmath18 amount to gauge transformations of @xmath23 . if we change @xmath18 to @xmath24 , then the resulting @xmath25 is related to @xmath23 by @xmath26 where @xmath27 measures the oriented area of the spherical triangle defined by @xmath28 , @xmath18 , and @xmath24 .
furthermore , as we will discuss more completely below , the area of any spherical triangle is uncertain modulo @xmath29 , and this accounts for the ` compactness ' of the u(1 ) gauge theory .
we proceed with our analysis of @xmath22 .
first , we discretize the gradient terms of the o(3 ) sigma model
. we will limit our considerations here to antiferromagnets on @xmath1 dimensional cubic lattices , but similar considerations apply to other bipartite lattices .
we also discretize the imaginary time direction , and ( by a slight abuse of notation ) use the same index @xmath2 to refer to the sites of a @xmath14 dimensional cubic lattice in spacetime .
on such a lattice we can rewrite ( [ zl ] ) as @xmath30 where the sum over @xmath31 extends over the @xmath14 spacetime directions .
we have also dropped unimportant factors of the lattice spacing and the spin - wave velocity in ( [ f1 ] ) .
as noted above , we are especially interested here in the large @xmath13 regime where there are strong fluctuations of the @xmath32 .
there are strong cancellations from the berry phases between different spin configurations in this regime , and so the second term in @xmath33 has to be treated with great care .
we will do this by promoting the field @xmath34 to an independent degree of freedom , while integrating out the @xmath32 .
notice that we have now introduced all @xmath14 components of the compact u(1 ) gauge field with the index @xmath31 , while only the @xmath35 component appears explicitly in ( [ f1 ] ) .
the remaining components appear naturally as suitable degrees of freedom when we integrate the @xmath32 out .
formally , the integration over the @xmath32 can be done by introducing new ` dummy ' variables @xmath36 and rewriting ( [ f1 ] ) by introducing factors of unity on each link ; this leads to @xmath37 in the first expression , if the integral over the @xmath36 is performed first , we trivially return to ( [ f1 ] ) ; however , in the second expression we perform the integral over the @xmath32 variables first , at the cost of introducing an unknown effective action @xmath38 for the @xmath36 . in principle ,
evaluation of @xmath38 may be performed order - by - order in a `` high temperature '' expansion in @xmath39 : we match correlators of the @xmath36 flux with those of the @xmath40 flux evaluated in the integral over the @xmath32 with positive weights determined only by the @xmath39 term in ( [ f1 ] ) .
rather than undertaking this laborious calculation , we can guess essential features of the effective action @xmath38 from some general constraints .
first , correlations in the @xmath32 decay exponentially rapidly for large @xmath13 ( with a correlation length @xmath41 ) , and so @xmath38 should be local .
second , it should be invariant under the lattice form of the gauge transformation ( [ gauge ] ) @xmath42 associated with the change in the reference point on the unit sphere from @xmath18 to @xmath43 , with @xmath44 equal to the area of the spherical triangle formed by @xmath32 , @xmath18 and @xmath43 .
finally the area of any triangle on the sphere is uncertain modulo @xmath45 and so the effective action should be invariant under @xmath46 the simplest local action which is invariant under ( [ f3 ] ) and ( [ f4 ] ) is that of _ compact u(1 ) quantum electrodynamics _ and so we have @xmath47 for large @xmath13 ; comparison with the large @xmath13 expansion shows that the coupling @xmath48 . in ( [ f5 ] ) , @xmath49 is the discrete lattice derivative along the @xmath31 direction , and the sum over @xmath50 extends over all plaquettes of the @xmath14 dimensional cubic lattice
both notations are standard in the lattice gauge theory literature .
the first term in the action ( [ f5 ] ) is , of course , the standard ` maxwell ' term of a compact u(1 ) gauge field . in this language
, the berry phase has the interpretation of a @xmath51 coupling to a fixed matter field with ` current ' @xmath52 .
this corresponds to static matter with charges @xmath53 on the two sublattices .
it is this matter field which will crucially control the nature of the ground state .
the remaining analysis of @xmath33 depends upon the spatial dimensionality @xmath1 . in @xmath54 , a dual model of ( [ f5 ] ) is solvable , and the results are in complete accord with those obtained earlier by bethe ansatz and bosonization analyses of spin chains .
we will consider the @xmath55 case in the section below .
there has been relatively little discussion of the @xmath56 case ( which exhibits both confining and deconfining phases of the gauge theory ) , and this remains an important avenue for future research .
as is standard in duality mappings , we first rewrite the partition function in @xmath57 spacetime dimensions by replacing the cosine interaction in ( [ f5 ] ) by a villain sum over periodic gaussians : @xmath58 where @xmath59 is the total antisymmetric tensor in three dimensions , and the @xmath60 are integers on the links of the _ dual _ cubic lattice , which pierce the plaquettes of the direct lattice . throughout this subsection
we will use the index @xmath61 to refer to sites of this dual lattice , while @xmath2 refers to the direct lattice on sites on which the spins are located .
we will now perform a series of exact manipulations on ( [ f6 ] ) which will lead to a dual _ interface _ model @xcite .
this dual model has only positive weights this fact , of course , makes it much more amenable to a standard statistical analysis .
this first step in the duality transformation is to rewrite ( [ f6 ] ) by the poisson summation formula : @xmath62 where @xmath63 ( like @xmath60 ) is an integer - valued vector field on the links of the dual lattice ( here , and below , we drop overall normalization factors in front of the partition function ) .
next , we write the berry phase in a form more amenable to duality transformations . choose a ` background ' @xmath64 flux which satisfies @xmath65 where @xmath2 is the direct lattice site in the center of the plaquette defined by the curl on the left - hand - side .
any integer - valued solution of ( [ d2 ] ) is an acceptable choice for @xmath66 , and a convenient choice is shown in fig [ fig1 ] . .
the circles are the sites of the direct lattice , @xmath2 , while the crosses are the sites of the dual lattice , @xmath61 ; the latter are also offset by half a lattice spacing in the direction out of the paper ( the @xmath67 direction ) .
the @xmath66 are all zero for @xmath68 , while the only non - zero values of @xmath69 are shown above .
notice that the @xmath70 flux obeys ( [ d2]).,width=240 ] using ( [ d2 ] ) to rewrite the berry phase in ( [ f6 ] ) , applying ( [ d1 ] ) , and shifting @xmath63 by the integer @xmath71 , we obtain a new exact representation of @xmath33 in ( [ f6 ] ) : @xmath72 the integral over the @xmath36 can be performed independently on each link , and its only consequence is the imposition of the constraint @xmath73 .
we solve this constraint by writing @xmath74 as the gradient of a integer - valued ` height ' @xmath75 on the sites of the dual lattice , and so obtain @xmath76 this is the promised 2 + 1 dimensional interface , or height , model in almost its final form .
the physical properties of ( [ d5 ] ) become clearer by converting the `` frustration '' @xmath66 in ( [ d5 ] ) into offsets for the allowed height values .
this is done by decomposing @xmath66 into curl and divergence free parts and writing it in terms of new fixed fields , @xmath77 and @xmath78 as follows : @xmath79 the values of these new fields are shown in fig [ fig2 ] . and ( b ) @xmath80 introduced in ( [ xy ] ) .
the notational conventions are as in fig [ fig1 ] .
only the @xmath35 components of @xmath80 are non - zero , and these are shown in ( b ) .
, width=384 ] inserting ( [ xy ] ) into ( [ d5 ] ) , we can now write the height model in its simplest form @xcite @xmath81 where @xmath82 is the new height variable we shall work with .
notice that the @xmath80 have dropped out , while the @xmath77 act only as fractional offsets ( for @xmath9 not an even integer ) to the integer heights . from ( [ he2 ] ) we see that for half - odd - integer @xmath9 the height is restricted to be an integer on one of the four sublattices , an integer plus 1/4 on the second , an integer plus 1/2 on the third , and an integer plus 3/4 on the fourth ; the fractional parts of these heights are as shown in fig [ fig2]a ; the steps between neighboring heights are always an integer plus 1/4 , or an integer plus 3/4 . for @xmath9 an odd integer ,
the heights are integers on one square sublattice , and half - odd - integers on the second sublattice .
finally for even integer
@xmath9 the offset has no effect and the height is an integer on all sites .
we discuss these classes of @xmath9 values in turn in the following subsections . in this case
the offsets @xmath83 are all integers , and ( [ he1 ] ) is just an ordinary three dimensional height model which has been much studied in the literature .
unlike the two - dimensional case , three - dimensional height models generically have no roughening transition , and the interface is always smooth . with all heights integers ,
the smooth phase breaks no lattice symmetries .
so square lattice antiferromagnets with @xmath9 even integer can have a paramagnetic ground state with a spin gap and no broken symmetries .
this is in accord with the exact ground state for a @xmath84 antiferromagnet on the square lattice found by affleck _
_ , the aklt state @xcite .
now the heights of the interface model can take four possible values , which are integers plus the offsets on the four square sublattices shown in fig [ fig2]a . as in section
[ s2 ] , the interface is always smooth _ i.e. _ any state of ( [ he1 ] ) has a fixed average interface height @xmath85 , and _ any _ well - defined value for this average height breaks the uniform shift symmetry of the height model under which @xmath86 .
after accounting for the height offsets , we will see below that any smooth interface must also break a lattice symmetry with the development of _ bond order _ : this allows a number of distinct spin gap ground states of the lattice antiferromagnet .
it is useful , first , to obtain a simple physical interpretation of the interface model in the language of the @xmath87 antiferromagnet @xcite . from fig
[ fig2]a it is clear that nearest neighbor heights can differ either by 1/4 or 3/4 ( modulo integers ) . to minimize the action in ( [ he1 ] ) , we should choose the interface with the largest possible number of steps of @xmath88 .
however , the interface is frustrated , and it is not possible to make all steps @xmath89 and at least a quarter of the steps must be @xmath90 .
indeed , there is a precise one - to - one mapping between interfaces with the minimal number of @xmath90 steps ( we regard interfaces differing by a uniform integer shift in all heights as equivalent ) and the dimer coverings of the square lattice : the proof of this claim is illustrated in fig [ fig3 ] . in ( [ he1 ] ) .
each dimer on the direct lattice is associated with a step in height of @xmath90 on the link of the dual lattice that crosses it .
all other height steps are @xmath88 .
each dimer represents a singlet valence between the sites , as in fig [ fig2].,width=220 ] we identify each dimer with a singlet valence bond between the spins ( the ellipses in fig [ fig2 ] ) , and so each interface corresponds to a quantum state with each spin locked in the a singlet valence bond with a particular nearest neighbor .
fluctuations of the interface in imaginary time between such configurations correspond to quantum tunneling events between such dimer states , and an effective hamiltonian for this is provided by the quantum dimer model @xcite .
the nature of the possible smooth phases of the interface model are easy to determine from the above picture and by standard techniques from statistical theory @xcite .
interfaces with average height @xmath91 ( modulo integers ) correspond to the four - fold degenerate bond - ordered states in fig [ fig4]a , while those with @xmath92 ( modulo integers ) correspond to the four - fold degenerate plaquette bond - ordered states in fig [ fig4]b . on the square lattice
: ( a ) the columnar spin - peierls states , and ( b ) plaquette state .
the different values of the @xmath93 on the links are encoded by the different line styles .
both states are 4-fold degenerate ; an 8-fold degenerate state , with superposition of the above orders , also appears as a possible ground state of the generalized interface model.,width=259 ] all other values of @xmath94 are associated with eight - fold degenerate bond - ordered states with a superposition of the orders in fig [ fig4]a and b. support for the class of bond - ordered states described above has appeared in a number of numerical studies of @xmath87 antiferromagnets in @xmath55 which have succeeded in moving from the small @xmath13 nel phase to the large @xmath13 paramagnet .
these include studies on the honeycomb lattice @xcite , on the planar pyrochlore lattice @xcite , on square lattice models with ring - exchange and easy - plane spin symmetry @xcite , and square lattice models with su(@xmath95 ) symmetry @xcite .
this case is similar to that @xmath9 half - odd - integer , and we will not consider it in detail .
the berry phases again induce bond order in the spin gap state , but this order need only lead to a two - fold degeneracy .
the primary topic discussed in this paper has been the effective field theory of paramagnetic mott insulators with collinear spin correlations .
this field theory is the compact u(1 ) gauge theory in ( [ f5 ] ) , and applies in all spatial dimensions .
we also reviewed duality mappings of ( [ f5 ] ) which are special to @xmath55 spatial dimensions , and mapped the theory onto the interface model ( [ he1 ] ) .
finally , we reiterate that paramagnetic mott insulators with non - collinear spin correlations are described by a @xmath0 gauge theory which has not been presented here .
fouet , m. mambrini , p. sindzingre , and c. lhuillier , phys .
b * 67 * , 054411 ( 2003 ) ; duality mapping for a lattice with the symmetry of the planar pyrochore is in ref .
@xcite , with a prediction for the bond order observed . | this is a summary of a central argument in recent review articles by the author ( physica a * 313 * , 252 ( 2002 ) , annals of physics * 303 * , 226 ( 2003 ) , and rev .
mod .
phys , july 2003 ) .
an effective field theory is derived for the low energy spin singlet excitations in a paramagnetic mott insulator with collinear spin correlations . |
letting @xmath10 be the function that takes a positive integer to the sum of the squares of its ( base 10 ) digits , a positive integer @xmath11 is said to be a _
happy number _ if @xmath12 for some @xmath13 @xcite .
these ideas were generalized in @xcite as follows : fix an integer @xmath9 , and let @xmath14 , where @xmath15 are integers . for each integer @xmath16 , define the function @xmath17 by @xmath18 if @xmath19 for some @xmath20 , then @xmath11 is called an _
@xmath21-power @xmath1-happy number_. we further generalize these functions by allowing the addition of a constant after taking the sum of the powers of the digits .
( throughout this work , all parameters are assumed to be integers . )
[ d : gen ] fix integers @xmath8 and @xmath9 .
let @xmath14 , where @xmath15 are integers .
for each integer @xmath16 , define the _ augmented generalized happy function _
@xmath22 , by @xmath23 in section [ s : properties ] , we examine various properties of the function @xmath24 , which , for ease of notation , we denote by @xmath0}}}$ ] . in section [ s :
cbgoodness ] , we state and prove theorem [ t : happysequence ] , an analogue to the existence of arbitrarily long sequences of consecutive happy numbers .
although this result is quite general , it leaves a particular case unresolved , which we make explicit in conjecture [ c : happysequencebetter ] , and in section [ s : conj ] , we prove that the conjecture holds for small values of @xmath2 and @xmath1 .
in this section , we consider the function @xmath0}}}= s_{2,b , c}$ ] .
note that @xmath26 } = s_2 $ ] , as defined in section [ s : introduction ] .
fix @xmath8 , @xmath9 , and @xmath27 .
we say that @xmath11 is a _
fixed point _ of @xmath0}}}$ ] if @xmath0}}}(a ) = a$ ] , and that @xmath11 is in a _ cycle _ of @xmath0}}}$ ] if @xmath0}}}^k(a ) = a$ ] for some @xmath13 .
the smallest such @xmath4 is called the _ length _ of the cycle . as is well - known ( see , for example , @xcite ) , @xmath10 has exactly one fixed point and one nontrivial cycle .
the standard proof of this uses a lemma similar to the following , which is a generalization for the function @xmath0}}}$ ] .
@xmath38}}}(a ) & = \sum_{i=0}^n a_i b^i - \left(c + \sum_{i=0}^n a_i^2\right ) \\ & = a_n\left(b^n - a_n\right ) + \sum_{i=1}^{n-1 } a_i \left(b^i - a_i\right ) + a_0(1-a_0)-c\\ & \geq 1\left(b^m-1\right ) + 0 + ( b-1)(1 - ( b-1))-c\\ & = b^m - b^2 + 3b - 3 - c \\ & > 0.\end{aligned}\ ] ] by lemma [ l : blemma4cycle ] , to prove that @xmath41 is @xmath42$]-cycle - good , it suffices to prove that each of the @xmath43 is @xmath42$]-cycle - good . to each @xmath44 , we associate a pair @xmath45 , as given in table [ t : conjpf3 ] and to each @xmath43 , we associate a pair @xmath46 , as in table [ t : conjpf4 ] . to see that @xmath47 is [ c , b]-cycle - good , fix a cycle , @xmath44 , of @xmath0}}}$ ] and note that for each @xmath48 , @xmath0}}}^{k_i}(t + n_i ) \in c_i$ ] .
thus @xmath47 is @xmath42$]-cycle - good .
now , fix @xmath49 such that @xmath50 .
a direct calculation shows that for each @xmath51 , @xmath0}}}^{k_j^\prime}(t + n_j^\prime ) \in v_1 $ ] . since @xmath47 is @xmath42$]-cycle - good , so is @xmath43 .
thus , by lemma [ l : blemma4cycle ] , @xmath41 is @xmath42$]-cycle - good .
h. g. grundman and e. a. teeple , _ sequences of consecutive happy numbers _ , rocky mountain j. math .
* 37 * ( 2007 ) , no . 6 , 19051916 . h. g. grundman and e. a. teeple , _ sequences of generalized happy numbers with small bases _ , j. integer sequences * 10 * ( 2007 ) , article 07.1.8 . | an augmented generalized happy function , @xmath0}}}$ ] maps a positive integer to the sum of the squares of its base-@xmath1 digits and a non - negative integer @xmath2 .
a positive integer @xmath3 is in a _
cycle _ of @xmath0}}}$ ] if , for some positive integer @xmath4 , @xmath0}}}^k(u ) = u$ ] and for positive integers @xmath5 and @xmath6 , @xmath5 is _ @xmath6-attracted _ for @xmath0}}}$ ] if , for some non - negative integer @xmath7 , @xmath0}}}^\ell(v ) = w$ ] . in this paper
, we prove that for each @xmath8 and @xmath9 , and for any @xmath3 in a cycle of @xmath0}}}$ ] , ( 1 ) if @xmath1 is even , then there exist arbitrarily long sequences of consecutive @xmath3-attracted integers and ( 2 ) if @xmath1 is odd , then there exist arbitrarily long sequences of 2-consecutive @xmath3-attracted integ |
the theory of oscillations of rotating fluids of variable density [ rayleigh- taylor ( r - t ) effect ] was developed in detail by chandrasekhar ( 1961 ) , hereafter c61 .
a large variety of the magnitohydrodynamic ( mhd ) problems , including the stability of inviscid and viscous fluids in the case of two uniform layers separated by a horizontal boundary with and without rotation as well as the magnetic field effects were analyzed using perturbation technique .
the simplest case of a one - dimensional gravitational force was studied .
a similar analysis was also implemented for the case of an exponentially varying density .
it follows from c61 that the quasiperiodic oscillations ( qpo ) with the twin `` kilohertz '' frequencies represent the main frequencies of the stable gravity waves in the rotational frame of reference .
it also as follows from c61 that the twin `` kilohertz '' frequencies should be on the order of the keplerian frequency . in this
_ paper _ i present the results of the study of the r - t effect for a particular case of fluid oscillations in an accretion flow under influence of a central gravitational force .
this particular r - t analysis is important in view of the high and low frequency detection by the rossi x - ray timing explorer ( rxte ) in a number of low mass x - ray binaries ( strohmayer et al .
1996 , van der klis et al .
1996 ) , black hole candidate sources ( morgan , remillard & greiner 1997 ; strohmayer 2001a , b ; remillard 2002 ) and by extreme ultraviolet explorer , chandra x - ray observatory and optical observations in white dwarfs ( mauche 2002 and woudt & warner 2002 ) .
the presence of two observed peaks with frequencies @xmath0 and @xmath1 in the upper part of the power spectrum became a natural starting point in modeling the phenomena . in ns binaries , for example sco x-1 , the lower frequency part of the power spectrum , contains two horizontal branch oscillation ( hbo ) frequencies @xmath2 hz and @xmath3 hz ( probably the second harmonic of @xmath4 ) which slowly increase with the increase of @xmath0 and @xmath1 ( van der klis et al .
_ any plausible model faces the challenging task of describing the dependences of the peak separation @xmath5 on @xmath0 and @xmath1 _ .
attempts have been made to relate @xmath0 and @xmath1 and the peak separation @xmath6 with the neutron star ( ns ) spin . in the sonic point
beat frequency model by miller , lamb & psaltis ( 1998 ) the khz peak separation @xmath7 is considered to be close to the ns spin frequency and thus @xmath7 is predicted to be constant .
however observations of khz qpos in a number of binaries ( sco x-1 , 4u 1728 - 34 , 4u 1608 - 52 , 4u 1702 - 429 and etc ) show that the peak separation decreases systematically when the high [ kilohertz ( khz ) ] frequencies increase ( for a recent review see van der klis 2000 , hereafter vdk ) . for sco x-1 vdk
found that the peak separation of khz qpo frequencies changes from 320 hz to 220 hz when the lower khz peak @xmath0 changes from 500 hz to 850 hz .
the correlation between high frequency ( lower khz frequency ) and low frequency ( broad noise component ) qpos previously found by psaltis , belloni & van der klis ( 1999 ) for black hole ( bh ) and neutron star ( ns ) systems has been recently extended over two orders of magnitude by mauche ( 2002 ) to white dwarf ( wd ) binaries . _ accepting the reasonable assumption that the same mechanism produces the qpo in wd , ns and bh binaries , one can argue that the data exclude relativistic models and , beat frequency models as well as any model requiring either the presence or absence of a stellar surface or a strong magnetic field . _ the transition layer model ( tlm )
was introduced by titarchuk , lapidus & muslimov ( 1998 ) , hereafter tlm98 , to explain the dynamical adjustment of a keplerian disk to the innermost sub - keplerian boundary conditions ( it is at the star surface for nss and wds ) .
tlm98 argued that a shock should occur where the keplerian disk adjusts to the sub - keplerian flow .
thus the transition layer bounded between the sub - keplerian boundary and the adjustment radius can undergo various type of oscillations under the influence of the gas , radiation , magnetic pressure and gravitational force .
osherovich & titarchuk ( 1999 ) , hereafter ot99 , suggested that the phenomological model of a one - dimensional classical oscillator in the rotational frame of reference could explain the observed correlations between twin khz frequencies @xmath0 , @xmath1 and the hbo frequencies .
they further suggested that the oscillations of the fluid element that bounced from the disk shock region ( at the adjustment radius ) would be seen as two independent oscillations parallel and perpendicular to the disk plane respectively .
this is due to the presence of a coriolis force in the magnetospheric rotational frame of reference . in this paper
i show that when the pressure gradients can be neglected , the problem of the rayleigh - taylor wave oscillations ( buoynancy effect ) in the rotational frame of reference is reduced to the ot99 formulation ( see 3 ) .
this result provides a solid basis for application of the ot99 model for interpretation of the qpo phenomena observed in x - ray binaries . the main goal of this _ paper _ is to demonstrate that there is an inevitable effect of the gravity wave oscillations in the heterogeneous fluid of the accretion flow near compact objects . in
2 , i formulate the problem of the gravity wave propagation in the bounded medium in the rotational frame of reference . in
3 i present analysis and solutions of this problem using a perturbation method in the context of three dimensional periodic waves with assigned wave numbers .
applications of the gravity modes and their relations with the qpo observations is presented in 4 .
summary and conclusions are drawn in 5 .
lord rayleigh ( 1883 ) has treated the non - rotating inviscid case of the present problem .
he developed a general theory for any density configuration @xmath8 ( the z axis being the upward drawn vertical ) .
rayleigh s treatment of inviscid superposed fluids was extended by bjerkness et al .
( 1933 ) to include the influence of rotation . with an assumption of the gravitational force directed vertically , hide ( 1956 ) developed the theory for any density @xmath8 and viscosity @xmath9 profiles and any @xmath10 ( where @xmath10 is an angle between the rotational axis and the vertical ) . in this study
we consider a case of an inviscid and incompressible fluid .
the equation of relative motion appropriate to the problem is @xmath11 where the fluid is supposed to rotate uniformly about an axis whose direction is specified by the unit vector @xmath12 .
the tensor notation follows the summation convention and the unit vector @xmath13 in the radial direction is introduced . in equation
( 1 ) @xmath14 denotes the density , @xmath15 the angular velocity of rotation , @xmath16 the @xmath17th component of the ( eulerian ) velocity vector , p is the pressure and @xmath18 is the acceleration due to gravity .
this term @xmath18 can represent the effective gravity which includes the centrifugal and radiation pressure forces .
for an incompressible fluid the continuity equation is @xmath19 because ( similar to c61)diffusion effects are ignored in this analysis , an individual fluid element retains the same density throughout its motion .
hence it is required that @xmath20 because the equilibrium situation in the comoving frame is a static one it is characterized by @xmath21 .
we now assume the equilibrium situation to be slightly disturbed , so that @xmath22 so we shall write @xmath23 and treat @xmath16 , @xmath24 , and @xmath25 as quantities of the first order of smallness so that products of such quantities can be ignored .
we are interested in the study of the gravity wave oscillations in the disk transition layer where the accretion flow is sub - keplerian and the temperature of the flow is of order 5 kev or more ( see cartoon diagram of the system in titarchuk , osherovich & kuznetsov 1999 , hereafter tok , fig .
1 ) . these temperatures of order of 5 kev are a representative values for plasma temperatures inferred from x - ray spectra during khz qpo events ( see more in titarchuk , bradshaw & wood 2001 and titarchuk & wood 2002 ) . at this stage we introduce the cartesian coordinate system . we take the @xmath26 axis to be in the direction of the ( upward ) vertical and the @xmath27 axis to be such that the @xmath28 plane contains the angular velocity vector @xmath29 which then has components @xmath30 . in the general case
the gravitational force has components @xmath31 where @xmath32 . if @xmath33 are the components of the velocity @xmath34 , then on substituting in equations ( 1 ) , ( 2 ) and ( 3 ) we find @xmath35 one can see that the variation of @xmath14 is ignored in all terms except for the ones representing the buoyancy force ( see c61 for details ) .
in order to illustrate the buoyancy ( r - t ) effect we consider the simplest solutions of the problem including those which already exist in the literature . _ case a _ in the case
when we allow the vector @xmath29 to rotate around the @xmath26 axis at angle @xmath10 , we introduce the cylindrical coordinate system , namely taking the @xmath26 axis in the direction of the ( upward ) vertical , the @xmath27 axis as the radial axis and the @xmath36 axis as the azimuthal axis in the horizontal plane
. then the vector @xmath29 has components @xmath37 and gravitational force vector has components @xmath38 where @xmath39 .
( i ) with assumptions that @xmath40 is only a function of @xmath41 the scale height of @xmath40 is order of @xmath41 we obtain that @xmath42 $ ] where @xmath43 , @xmath44 are the radial and vertical components of displacement respectively and consequently the radial and vertical components of the perturbation velocity are @xmath45 , @xmath46 . thus the vector @xmath47 is transformed into @xmath48 $ ] .
furthermore if we neglect the effects of pressure gradients , the set of equations ( 5 - 7 ) can be reduced to the system where the right hand side consists of only the vector @xmath49 .
then , for radial , azimuthal and vertical displacements @xmath50 , @xmath51 , @xmath52 this system takes the form : @xmath53 ot99 have already analyzed the solution of eqs .
( 10 - 12 ) ( see eqs . 2 - 4 , in ot99 ) in the case of @xmath54 and @xmath55 .
they found that in the rotational frame of reference , the radial oscillation with the main frequency @xmath56 is split to the oscillations taking place near the horizontal ( disk ) plane ( @xmath57 ) with the hybrid frequency @xmath58 and to oscillations taking place near the vertical plane ( @xmath59 ) with the low branch frequency @xmath60 .
for @xmath61 the dispersion equation for the frequency @xmath62 @xmath63=0\ ] ] besides the nonoscillatory mode ( @xmath64 ) , describes two oscillatory eigenmodes . for @xmath55
they are @xmath65 and @xmath66 .
the relation of the model eigenfrequencies @xmath67 , @xmath68 and @xmath69 with the qpo observed frequencies : horizontal branch oscillation freqiencies @xmath4 , khz frequencies were studied by ot99 ; tok , kuznetsov & titarchuk ( 2002 ) ; and titarchuk ( 2002 ) , hereafter t02 .
it is worth noting that the relation between @xmath70 , @xmath15 , @xmath71 and @xmath72 predicts the existence of the invariant quantity @xmath10 .
( e.g. titarchuk & osherovich 2001 ) .
t02 calculated @xmath10 and its uncertainty of @xmath10 finding that the inferred @xmath73values are consistent with being constant at least for four z sources , sco x-1 , gx 340 + 0 , gx 5 - 1 , gx 17 + 2 ( see more on this issue in section 4 ) .
( ii ) . with assumptions that @xmath74 and @xmath75
the set of equations for @xmath50 , @xmath51 , @xmath52 is similar to eqs ( 10 - 12 ) where the right hand side vector is the zero vector , @xmath76 . in this case
the dispersion equation for @xmath62 ( for small @xmath77 ) @xmath78 has the only one nontrivial root @xmath79 which is related to the eigenmode oscillations taking place parallel to the disk ( @xmath57)-plane .
we discuss the application of this solution to the observation in section 4 .
_ case b _ if we do not neglect the pressure gradient and assume the vertical gravitational force and the vector @xmath80 directed along the vertical ( @xmath81 ) , then the entire problem ( eqs . 5 - 9 ) is reduced to set of equations which have already been analyzed by chandrasekhar in c61 .
he studied two cases : ( 1 ) in the case of two uniform fluids separated by a horizontal boundary he showed that for two adjacent , hydrostatic , inviscid fluids , the low fluid having density @xmath82 and the upper layer having density @xmath83 the eigenfrequency @xmath84 was @xmath85^{1/2}\ ] ] where @xmath86 is the frequency in the absence of rotation .
the frequency @xmath86 is of order of @xmath87 if the k - wave number ( @xmath86 depends on k and density difference @xmath88 ) of order @xmath89 and @xmath90 for a stratified medium of density @xmath91 , where @xmath92 is the scale height , the r - t instability occurs for positive @xmath93 , whereas stable gravity waves occur for negative @xmath93 .
the r - t frequency @xmath94^{1/2}$ ] if one assumes that the wave number @xmath95 is of order of @xmath96 ( where @xmath97 is a layer size ) .
it is worthwhile to emphasize that even in the simplest case of the vertical gravitational force ( c61 ) the hybrid frequency @xmath71 is an eigenfrequency ( which is also true for the case with no pressure gradient effects , see above ) .
formula ( 10 ) for @xmath84 can also be reduced to @xmath71 if one assumes that @xmath98 .
_ case c _ now we consider the case when the vector @xmath99 rotates uniformly around the vertical at angle @xmath10 , then the gravitational vector @xmath100 , also taking into account the pressure effects . the general case of @xmath101 can be analyzed in a similar way and it will be presented elsewhere . following the usual practice in problems of this kind ,
we seek solutions of equations ( 5 - 9 ) which are of form ( see e.g. c61 ) @xmath102 where @xmath103 , @xmath104 and @xmath105 are the horizontal and vertical wave numbers of the harmonic perturbations respectively .
we also assume that @xmath106 is a function of @xmath27 ( or @xmath107 ) and @xmath26 which the scale heights are @xmath108 and @xmath92 for radial and vertical density profiles respectively , @xmath109 .
upon substituting for @xmath110 in the form ( 16 ) equations ( 5 ) to ( 8) become @xmath111 we also use the relation @xmath112 which follows from eqs .
( 9 ) and ( 16 ) in order to express @xmath113 through @xmath40 , @xmath114 and @xmath115 .
the set of equation ( 17 - 20 ) assumes a nontrivial solution only if the determinant of the system @xmath116 .
this equation provides the dispersion relation for the determination of @xmath117 : @xmath118 where @xmath119 , @xmath120 , @xmath121,$ ] @xmath122 .
the specific wave values @xmath123 are determined by the conditions imposed on the oscillatory domain boundary .
thus for a given set of boundary conditions the analysis of the r - t instability is reduced to the analysis of the roots of of fourth order algebraic equation ( 21 ) , which depends on the main parameters of the atmosphere , @xmath93 , @xmath124 and ratio of the wave numbers @xmath125 .
we assume that @xmath126 and @xmath127 is the case of interest .
figure 1 illustrates the specific behavior of the polynomial p(n ) which should help one to understand the presence ( or absence ) of its roots for given coefficient @xmath128 , @xmath129 , @xmath130 ( @xmath131 , @xmath132 ) .
for example , if @xmath133 , @xmath134 and @xmath135 ( see case @xmath17 below ) then @xmath136 has two complex conjugate roots @xmath137 and two real roots @xmath138 , @xmath139 ( in fact , @xmath140 ) . because @xmath134 , the absolute value of the positive root @xmath139 is larger than that of the negative one @xmath138 .
but @xmath141 is a sum of the polynomial ( real and complex conjugate ) roots , and because of @xmath142 one can come to conclusion that @xmath143 . therefore ( in this case ) the stable oscillatory mode exists and @xmath144 are related to these damped oscillations .
below we present all cases with the oscillatory stable solutions : _ case i _ : @xmath145 @xmath146 , @xmath147 and @xmath148 ( see fig .
1 , curve i ) . for such conditions @xmath133 and @xmath134 .
then equation ( 21 ) has two real roots which relate to the unstable ( growing ) and stable ( decaying ) modes with one pair of complex conjugate roots corresponding to the stable oscillatory mode : @xmath149 where @xmath150^{1/2}$ ] and @xmath151 .
these roots of equation ( 21 ) are found using the sequential approximation method : first we solve equation ( 21 ) with @xmath152 , @xmath153 and then in the next stage we look for roots of equation ( 21 ) as @xmath154 . _ case ia _ : @xmath145 @xmath146 , @xmath147 and @xmath155 .
for such conditions @xmath133 and @xmath134 .
equation ( 21 ) has just one pair of the complex conjugate roots and two real roots which relate to the unstable ( growing ) and stable ( decaying ) modes ( see fig .
1 , curve ia ) .
the oscillatory mode @xmath144 is stable for which we have @xmath156 where @xmath157^{1/2}$ ] .
_ case ib _ : @xmath158 @xmath159 , @xmath147 . for such conditions @xmath133 and @xmath134 and @xmath148
this case is similar to case i ( see fig 1 , curve i ) when equation ( 21 ) has two real roots which relate to the unstable ( growing ) and stable ( decaying ) modes and complex conjugate roots which correspond to the stable oscillatory mode : @xmath160 _ case ii _ : @xmath161 , @xmath162 . for such conditions @xmath163 and @xmath134 and @xmath148 .
equation ( 21 ) has a pair of complex conjugate roots ( see fig .
1 , curve ii ) and one of them @xmath144 is related to the stable oscillatory mode : @xmath160 _ case iii _ : @xmath164 , @xmath162 , and @xmath165 . for such conditions @xmath163 and @xmath166 .
equation ( 21 ) has a pair of complex conjugate roots ( see fig .
1 , curve iii ) and one of them @xmath167 is related to the stable oscillatory mode : @xmath168 where @xmath169^{1/2}$ ] . finally , we single out the case when the effective gravitational force goes to zero .
this can happen during a burst event , when the gravitational forces are compensated by the radiation pressure forces ( e.g. titarchuk 1994 ) .
for such conditions , @xmath170 and @xmath171 ^ 2 $ ] .
equation ( 21 ) has zero roots @xmath172 and conjugate complex roots @xmath173\ ] ] which are related to a pure harmonic mode .
in fact , this result also follows from treatments detailed in c61 and ot99 .
the hybrid frequency then becomes simply @xmath174 for @xmath175 .
this case of the perfect coherent oscillations would also occur if the density profile is quasi - uniform i.e. when @xmath176 but @xmath177 .
these two conditions ( on either the effective gravity force or the density profile ) can be realized during the burst event . in this section
the main goal is to reveal all cases when stable gravity modes exist and when the stability breaks down .
this analysis is particularly important in a view of the transient nature of qpo features ( see e.g. zhang et al .
1998 ) . with an increase of bolometric luminosity ( presumably in mass accretion rate ) the khz qpo frequency increases and
then entirely disappears ! at low rates it appears once again !
the stability analysis presented here leads to conditions for the existence and the destruction of gravity modes in terms of density profiles scale heights @xmath178 and boundary conditions ( @xmath179 ) .
in fact , the strong dependence of the gravity wave stability on the density profile was a central point of chandrasekhar s analysis ( c61 ) . for the accretion disk cases
, we can give an example where stable gravity modes are followed by instabilities ( the gravity mode destruction ) . in case
( ib ) we have a stable mode with @xmath180 as a qpo frequency .
when the density profile changes , over @xmath181coordinate from @xmath182 to @xmath183 then the stable g - mode with @xmath180 can still be sustaned [ see case ( ia ) ] , but if the density profile over z coordinate is stabilized ( @xmath164 ) and that over r coordinate is inverted ( i.e. from @xmath184 to @xmath183 ) then the qpo oscillations are no longer stable .
as we have seen in the previous section , there are two stable oscillatory gravity modes : one is associated with the hybrid frequency @xmath180 ( see cases @xmath185 ) and another with the low branch frequency @xmath186 ( cases @xmath187 and @xmath188 ) .
we also estimate the decay rate for oscillations as @xmath189 and the qpo quality value @xmath190 . the presence and absence of these modes depend on the atmospheric structure ( scale height inverses @xmath93 and @xmath124 ) and on the imposed boundary conditions ( wave numbers @xmath103 , @xmath104 and @xmath105 ) .
furthermore , it is possible to restore the related boundary conditions if one compares the observed qpo features with the calculated mode frequencies and q - values .
the analysis made in section 3 is also necessary in order to control an accuracy of numerical calculations of the set of hydrodynamical equations ( 5 - 9 ) .
to illustrate the results obtained in 3 we should specify the orders of the introduced quantities @xmath191 .
namely we assume that @xmath192 ( where @xmath193 ) and @xmath194 if we also assume that @xmath195 then @xmath196 . with these assumptions expressions for the polynomial coefficients @xmath197 ( see eq .
21 ) are significantly simplified : @xmath198 , @xmath199 , @xmath200 . because @xmath193 and @xmath201 we can calculate @xmath186 as @xmath202 and @xmath203 tok introduced the classification scheme for the qpo features and related the observed high and low khz frequencies to the hybrid and keplerian frequencies , @xmath69 and @xmath68 respectively .
they also attributed the hectohertz frequencies detected in the atoll source 4u 1728 - 34 ( ford & van der klis 1999 ) to the low branch .
the angle @xmath10 was found to be almost twice that found in sco x-1 ( see also t02 for details of @xmath73 determination ) .
the low frequencies in atoll sources ( as a rule ) are three times higher than those in the z - sources .
the hectohertz frequencies have been identified in several other neutron star lmxbs ( 4u 0614 + 09 ; van straaten et al .
2000 , 2002 ; sax j1808.4 - 3658 , 4u 1705 - 44 ; wijnands & van der klis 1998 ) .
they weakly depend on the khz frequencies , their ratio to low khz frequency being about ( 4 - 5 ) ( van straaten et al .
2002 , hereafter s02 ) .
the hectohertz frequency lorentzian profile are very broad with @xmath204-values around 1 or even less .
we suggest that these observed frequencies can be identified as the low branch frequencies ( see formulas 28 - 29 ) . taking into account resonance effects ( t02 )
, we correct the error bars of the hectohertz frequencies presented in s02 . in figure 2 the best fit to the data ( s02 ) is presented using formula ( 28 ) which includes one fit parameter q. the best - fit value of @xmath205 for which @xmath206 . the estimated @xmath207 are also in agreement with the observed values of @xmath204 ( s02 , table 2 ) .
the harmonic modes @xmath208 can be related to a coherent oscillation frequency of 582 hz which is observed in 4u 1636 - 53 during the superburst ( strohmayer & markwardt 2002 ) . if @xmath209 ( which is our case ) @xmath210 is a double frequency of the ns spin frequency .
furthermore , we have already found the same eigenmode with the frequency @xmath211 in case ( aii ) when effects of the buoynant forces and pressure gradients are neglected ( see 3 ) .
in fact , miller ( 1999 ) has reported the detection of the coherent pulsations of frequency @xmath212 hz during the burst development in 4u 1636 - 53 .
_ thus the ns spin frequency @xmath213 hz and the double ns spin frequency @xmath214 hz have probably been detected in 4u 1636 - 53 .
_ it is also worth noting that the rxte observations of neutron binaries establish that the khz qpo frequencies do not seem to exceed beyond a certain upper limit ( zhang et al .
this observational effect may be a result of the density profile .
our stability analysis presented in 3 clearly indicates such a possibility ( see also c61 ) .
i have presented a detailed study of the rayleigh - taylor ( r - t ) instability in the accretion flow . to summarize i have : ( 1 ) put forth arguments to explain the qpo phenomena , as a result of the r - t effect in the rotational frame of reference .
( 2 ) formulated and solved the mathematical problem of the gravity wave propagation ( r - t effect ) in the accretion flow . ( 3 ) concluded that the stable gravity modes in the rotational frame of reference are related to the hybrid and low branch frequencies .
( 4 ) demonstrated that the particular problem of the gravity waves in the rotational frame of reference , in the approximation of very small pressure gradients , is reduced to the problem of the classical oscillator in the rotational frame of reference which was previously introduced and applied for the interpretation of khz qpo observation by ot99 .
i demonstrate that these frequencies are intrinsic features of the r - t effect .
they appear in various configurations of the accretion flow depending on assumptions regarding the density profiles , the boundary conditions and the effects of the pressure forces .
it is not by chance the high and low frequencies phenomenon has common observational appearances for a wide range of objects classes from black hole sources down to white dwarfs ( mauche 2002 ) .
( 5 ) investigated the conditions for the density profile and the wave numbers ( boundary conditions ) when the gravity modes are stable .
( 6 ) identified the observed qpo frequencies seen in the power density spectra of ns lmxbs using the inferred gravity mode frequencies .
in particular , i found that the inferred low branch frequencies and their q - values are consistent with the qpo hectohertz frequencies observed in the atoll sources 4u 1728 - 34 and 4u 0614 + 19 .
( 7 ) during the ns long ( super ) burst event , i find that the observer should see oscillations at double ns spin frequency .
the coriolis force is the only force which acts in the rotational frame of references and its presence causes perfect coherent pulsations with a frequency twice of the ns spin frequency .
finally one can conclude that the r - t gravity wave oscullations must be present and the related qpos should be detected in any system where the gravity , buoynancy and coriolis force effects can not be excluded ( even in the earth and solar environments ) .
bjerkness , v. bjerkness , j. , solberberg , h. & bergron , t. 1933 , _ physikalish hydrodynamik _ , berlin : springer chandrasekhar , s. 1961 , _ hydrodynamics and hydromagnetic stability _ , oxford : oxford at the caredon press ( c61 ) | accretion onto compact objects in x - ray binaries ( black hole , neutron star ( ns ) , white dwarf ) is characterized by non - uniform flow density profiles .
such an effect of heterogeneity in presence of gravitational forces and pressure gradients exhibits raylegh - taylor gravity waves ( rtgw ) .
they should be seen as quasioperiodic wave oscillations ( qpo ) of the accretion flow in the transition ( boundary ) layer between the keplerian disk and the central object . in this paper
i show that the main qpo frequency , which is very close to the keplerian frequency , is split into separate frequencies ( hybrid and low branch ) under the influence of the gravitational forces in the rotational frame of reference
. the rtgws must be present and the related qpos should be detected in any system where the gravity , buoynancy and coriolis force effects can not be excluded ( even in the earth and solar environments ) .
the observed low and high qpo frequencies are an intrinsic signature of the rtgw .
i elaborate the conditions for the density profile when the rtgw oscillations are stable .
a comparison of the inferred qpo frequencies with qpo observations is presented .
i find that hectohertz frequencies detected from ns binaries can be identified as the rtgw low branch frequencies .
i also predict that an observer can see the double ns spin frequency during the ns long ( super ) burst events when the pressure gradients and buoyant forces are suppressed .
the coriolis force is the only force which acts in the rotational frame of reference and its presence causes perfect coherent pulsations with a frequency twice of the ns spin .
the qpo observations of neutron binaries have established that the high qpo frequencies do not go beyond of the certain upper limit .
i explain this observational effect as a result of the density profile inversions .
also i demonstrate that a particular problem of the gravity waves in the rotational frame of reference in the approximation of very small pressure gradients is reduced to the problem of the classical oscillator in the rotational frame of reference which was previously introduced and applied for the interpretation of khz qpo observation by osherovich & titarchuk . |
structures of nanometric dimensions are strongly dependent of size and shape mainly due to the fact that the characteristic scales of many physical phenomena are comparable to the dimensions of current nanostructures .
magnetism makes no exception and fundamental research of nanomagnets is further fuelled by their prospective applications into technological devices @xcite and biomedical applications @xcite .
the implementation of nanomagnets into potential devices requires a deep control of their physical properties ; hence it is very important to control the shape and size of the nanoparticles which are crucial to determine their magnetic properties . within currently available nanostructures , nanotubes made from 3d transition metals and their alloys
are of particular importance , since the nanotube topologies provide us of two surfaces for modification , allowing the generation of multifunctional magnetic nanoparticles . clearly , a technology capable to modify different surfaces would be highly desirable @xcite .
more recently hollow tubular nanostructures have been synthesized srh+05,ncm+05,wll+05,tgj+06 and may be useful for applications in biotechnology , because low density magnetic nanotubes ( mns ) can float in solutions and are more suitable for in vivo applications @xcite . in recent years
, the magnetism of planar and circular nanowires has been intensely investigated , while mns have received less attention , although mns do not exhibit magnetic vortex cores or bloch points @xcite , leading to a different behaviour with a more controllable reversal process .
it is well know that changes in the cross section of the mns strongly affect the reversal mechanism and the overall magnetic behaviour wll+05,dkg+07,ela+07 - 1,ela+07 - 2,lsc+09 .
recent articles have shown that mns present three main equilibrium states : a uniform state ( u ) , a mixed state ( m ) and a vortex state ( v ) .
with basis in theoretical models it has been argued that for nanotubes with outer radius less than @xmath0 ( where @xmath1 is the exchange length , @xmath2 the saturation magnetization and @xmath3 the stiffness constant of the magnetic material ) the magnetization is almost uniform and oriented principally along the cylindrical axis @xcite .
we call this configuration uniform state . also for nanotubes with @xmath4
the magnetic equilibrium state correspond to the so - called mixed state , which is a mixture of u and v states @xcite .
this state presents a uniform magnetization along the middle region of the tube and near the lower and upper surfaces the magnetization deviates from the uniformity in order to reduce the magnetostatic field @xcite .
finally , there is a transition from the m state to the vortex state as can be seen from figure 7 of lsc+09 . on the other hand
, it has been argued that domain wall propagation in nanostructures , which can be achieved by external fields or spin - currents , is of basic scientific and technological interest @xcite .
more recently , it has been shown that the size - dependent reversal process in mns occurs via dw nucleation and subsequent dw propagation .
the dw microstructure depends on the tube cross section and can be a transverse wall ( tw ) for tube radius ( @xmath5 ) smaller than a critical radius ( @xmath6 ) , or a vortex wall ( vw ) for @xmath7 lae+07,ebj+08,aea+08,bep+09 , where @xmath8 is the ratio between the inner ( @xmath9 ) and outer tube radius .
the critical radius defined above depends on the magnetic material and on the shape factor @xmath10 and then it can be tailored manipulating its geometrical parameters . in polycrystalline systems ,
the critical radius , @xmath11 , ranges from a few nanometers to 20 nm approximately , and therefore , since experimentally fabricated mns satisfy @xmath12 , we can expect that the propagation of a vortex dw be the dominant magnetization reversal mechanism for ferromagnetic nanotubes @xcite .
although the magnetostatic field of mns with uniform magnetization has been already investigated @xcite , an analytical model for the magnetostatic field produced by a mn with the above mentioned configurations , either equilibrium or transient states , has not yet been performed .
our intention is to fill this gap by using simple magnetization models @xcite . since nanostructures are usually polycrystalline , the crystallographic orientations of the crystallites are random and , as a consequence , the average magnetic anisotropy of the particle is very small . in view of that
, it will be neglected in our calculations @xcite .
this paper is organized as follows : in sec .
[ model ] we present the magnetization models , including in sec .
modela the equilibrium states m and u and the vw , and then in sec .
modelb we discuss the case of the tw . results and discussions are summarized in sec .
[ results ] and conclusions in sec .
[ remarks ] .
in this section we present explicit expressions to evaluate the magnetostatic field for nanotubes with the magnetization states described above . within the continuum theory of ferromagnetism @xcite ,
the magnetostatic field is given by @xmath13 , where @xmath14 is the magnetostatic potential , which depends on the magnetization itself .
the basic assumption we address in this paper is that the nanotubes we are considering are thin enough to avoid any dependence of the magnetization in the radial coordinate . in what follows we present the magnetization and the related magnetostatic field for the equilibrium states ( u and m ) and both dws structures aforementioned .
first , we consider nanotubes with vortex - like configurations , as the mixed state and the transient state with a vw confined within the tube .
the magnetization of such magnetic configurations can be generally written as @xmath15 , where @xmath16 , and @xmath17 , with @xmath18 being the angle between the local magnetization vector and the cylindrical @xmath19 axis . with this rather general form for the magnetization we can compute the magnetostatic field produced by @xmath20 . following the calculations of @xcite
, the magnetostatic potential can be written as@xmath21 dq,\]]where @xmath22 and @xmath23 come from the surface and volumetric magnetic charges and can be written as @xmath24@xmath25 in the potential we have also defined : @xmath26 , \]]where @xmath27 are bessel functions . the magnetostatic field associated with the above potential can be written as @xmath28 , and then , the cylindrical components of the field are @xmath29 ,
@xmath30 dq,\]]@xmath31 dq.\]]at this point we are in position to calculate the magnetostatic field through the reversal process for mns , as well as to investigate the changes in the magnetostatic field when the equilibrium state is properly considered .
in what follows we will use the equations above to calculate the magnetostatic field for the m state and the vortex dw .
the arrangement of magnetic moments in a mn with small radius can be seen as a uniform state in the middle region of the tube , but with vortex - like deviations of the magnetization located at the tube ends wll+05,lsc+09,lss+07,cub+07 .
this mixed state has been observed by micromagnetic simulations @xcite and studied theoretically @xcite .
it has been shown that two equilibrium states can be studied with the model for the m state .
in other words , the u state is a special case of the former , as we will see later .
the model we use for the m state is given by the function @xcite @xmath32where @xmath33 and @xmath34 are the dimensions of the regions where the magnetization deviates from the @xmath19-axis . in the model , @xmath35 and
@xmath36 correspond to angles between the magnetization and the @xmath19-axis evaluated at the tube ends .
note that we obtain the uniform magnetization state in the limit @xmath37 ( @xmath38 ) which occurs approximately for a radius smaller than 4@xmath39 @xcite .
the parameters of the model are such that they minimize the total energy for each set of geometrical parameters .
for simplicity we consider symmetric nanotubes , that is , tubes without difference between both tube ends . in those cases the four parameters ( @xmath33 , @xmath34 , @xmath35 , @xmath40 ) variational problem
is reduced to a two parameters variational problem , when @xmath41 and @xmath42 @xcite .
then , @xmath22 and @xmath23 become @xmath43@xmath44dy.\]]with these quantities in hand we can compute the magnetostatic field @xmath45 at any point .
also , the uniform magnetization state is recovered with @xmath46 , where the functions reduces to @xmath47 @xmath48 .
the magnetization of the vortex wall in a nanotube can be expressed in a similar way to the magnetization of the mixed state , that is @xmath49 , but now @xmath50as presented in @xcite . here
@xmath51 is the location of the center of the vortex wall , and @xmath52 is the wall width . within the above model
, the magnetostatic field can be also expressed with equations ( 7 ) and ( 8) , but now the functions @xmath22 and @xmath23 become@xmath53@xmath54 the magnetization of the transverse wall in a nanotube can be expressed in a similar way to the vortex wall , but with @xmath55 @xcite , and @xmath56 being the same of the vortex wall . it can be noted that the main difference between vortex and transverse walls ( in nanotubes ) is the direction of the in - plane magnetization , which makes the nature of the wall properties to be different .
the model that we use for the transverse wall @xcite leads us to understand some basic properties , like the dw width or the energy barrier for the reversal process .
here we are interested in the visualization of the magnetostatic field produced by a static domain wall .
it can be shown that this field can be written as @xmath57 , without the usual axial symmetry of vortex - like states .
the case of a vortex wall presents a magnetostatic potential independent of the angular coordinate @xmath58 , which secures angular symmetry . following @xcite ,
the field can be written as @xmath59 .
provided the magnetostatic potential of the tw can be written as @xmath60the components of the dipole field created for the tw can then be expressed as @xmath61@xmath62@xmath63where @xmath64and@xmath65here , @xmath66 , with @xmath67 .
the function @xmath68 possesses the information about the dw microstructure and can be straightforwardly integrated .
finally , notice that the widths of the vortex and transverse walls are essentially different , and should be evaluated within the framework presented in @xcite .
in the above sections we have obtained explicit expressions to calculate the magnetostatic field for different magnetic states in nanotubes . in what follows we present plots of the magnetostatic field produced by the magnetic states relevant to nanotubes .
we begin the discussion with the equilibrium states ( u and m ) and then we analyze the case of vortex and transverse walls propagating along the nanostructure . as mentioned earlier
, the uniform magnetization state ( @xmath46 ) is a special case of the mixed state ( @xmath69 ) , and occurs when the tube radius is approximately less than 4@xmath39 , provided the exchange energy dominates over the dipolar one @xcite . when @xmath70 the magnetization at the tube ends is not entirely parallel to the cylindrical @xmath71 axis , instead it develops vortex domains at the ends in order to reduce the dipolar energy @xcite .
the long range character of the dipolar interaction makes that the dipolar energy grows as @xmath72 , whereas the exchange energy grows as @xmath73 , with @xmath73 the number of atomic magnetic moments in the tube .
therefore , the edge vortex domains increases in size with the radius . in the language of the above equations , the size of the vortex domains
is given by the parameters @xmath35 and @xmath33 .
as @xmath5 increases over 4@xmath39 , @xmath35 and @xmath33 increases , as the reader can see from figures 5 and 6 in @xcite . to calculate the magnetostatic field we need to know the set of parameters ( @xmath74 ) which minimize the total energy of the mixed state . using the expressions for the total energy of @xcite , we obtain for nanotubes with @xmath75 and @xmath76 different values of ( @xmath74 ) depending on their radius .
these results are summarized in table [ table1 ] .
note that for @xmath76 the transition to a flux - closure vortex state occurs approximately at @xmath77 @xcite .
@xmath78 ( @xmath74 ) for magnetic nanotubes with @xmath75 and @xmath76 as a function of their radius .
the arrows in figure [ ferro ] show the normalized magnetostatic field for a uniformly magnetized mn with @xmath79 .
the colour code indicates the strength of the field at each point ; white colour indicates that the magnetostatic field is stronger at the tube ends because magnetic charges ( @xmath80 ) resides only at the surfaces such that @xmath81 , with @xmath82 a unit vector normal to the surface of the sample @xcite .
provided that the u and m states ( as well as the vortex wall ) can be represented with @xmath83 , it follows that the surface magnetic charges are given by @xmath84 , where @xmath85 or @xmath86 . in the upper surface of the tube ( @xmath87 )
@xmath88 and then @xmath89 , whereas in the lower surface @xmath90 , and @xmath91 .
if the actual equilibrium state is the u or the m state , we find positive magnetic charges ( @xmath92 ) located at the upper surface , and this region behaves like a source of charges , as in figure [ ferro ] ; in elementary electromagnetism the field lines emanate from positive charges .
conversely , in the lower surface @xmath93 and the region behaves like a sink of charge ( the field lines converge to these points ) .
it can be shown that for nanotubes in the mixed state , the dipole field looks very similar to the one plotted in figure [ ferro ] . as a consequence , we do not present the same plots for tubes in the mixed state , instead we show in figure [ hdvsrho ] the dependence of the magnitude of the dipole field ( @xmath94 ) with the radial coordinate for different tube radius and @xmath95 ( just @xmath96 above the upper surface of the tube ) .
we have normalized all the curves to @xmath97 , and in order to remark the difference between the u state and the m state , we present groups of curves for both states . .
the left panel shows curves for the uniform state and the right panel shows the magnetostatic field for the mixed state .
different curves represent different tube radius , as depicted in the figure , and parameters have been taken from table 1 .
, width=302 ] we observe that a proper consideration of the equilibrium state reduces considerably the dipole field .
this effect becomes critical as the tube radius is increased , which increment the size of the vortex domains of the mixed state @xcite .
it is worth to mention about the importance of a proper consideration of the actual magnetic state in tubular nanostructures . in the sake of simplicity , it is a very common practice in magnetism to approximate the magnetization as a uniform field , in spite that uniform magnetization can be achieved ( at zero applied field ) only in nanostructures with dimensions of the order of the exchange length kravchuk , bookchapter , or when the nanoparticle is an ellipsoid with the appropriate size .
moreover , cylindrical magnetic nanowires and nanotubes are frequently consider as infinite structures , and only in this limit the approximation of a uniform dipole field becomes plausible , because the cylindrical nanostructures can be considered as ellipsoids only if they are infinitely long .
however , actual magnetic nanostructures are not infinite and finite size effects can be relevant .
it is well known that the dipole field produced by a nanoparticle with an ellipsoidal shape is uniform if its size is below the single domain limit ( in absence of an external field ) aharoni96 . in this limit
one can safely write the dipole field in terms of the corresponding demagnetizing factor : @xmath98 . .
the fields have been normalized to @xmath99 , where @xmath100 is the corresponding demagnetizing factor .
the long dashed line corresponds to the u state for @xmath101 .
the short dashed line corresponds to the u state and @xmath102 .
the thick solid line correspond to u state and @xmath103 whereas the circles correspond to the m state for @xmath103 .
the horizontal line illustrates the strength of the dipole field calculated in the usual ellipsoid - like approximation for uniform magnetization : @xmath104 .
, width=264 ] the difference between uniform and non - uniform dipolar fields is illustrated in figure [ nz ] , where we have plotted @xmath105 , the component of the magnetostatic field along the tube axis , as a function of @xmath19 for different tube radius and magnetic states .
the horizontal solid line depicts the strength of the ( uniform ) dipole field ( @xmath99 ) produced by an infinite mn with uniform magnetization , whereas the long dash , short dash and solid lines correspond to the field produced by a mn with a uniform magnetization state with radius @xmath101 , 10 and 20 , respectively .
the circles correspond to the dipole field produced by a mn with a mixed state rather the incorrect uniform magnetization state , and with @xmath103 .
the fields have been normalized to @xmath99 , where @xmath106 is the corresponding demagnetizing factor for a nanotube ela+07 - 1,beleggia06 . in order to illustrate the basic characteristics of the magnetostatic field produced by a mn with a vortex domain wall
, we have chosen nanotubes with @xmath107 , @xmath76 and @xmath108 . following a previous work by some of us @xcite
, we can calculate numerically the dw width ( @xmath52 ) by minimizing the total energy including exchange .
it is found that @xmath109 , and for the values mentioned above we obtain a wall width @xmath110 . in figure [ vw750 ]
we show the magnetostatic field normalized to his strength for a vortex wall located at @xmath111 , whereas figure [ vw500 ] shows the same plot for the vw confined at the middle of the tube ( @xmath112 ) . .
the colours indicate the strength of the magnetostatic field .
parameters are presented in the main text.,width=302 ] .
the colours indicate the strength of the magnetostatic field .
parameters are presented in the text.,width=302 ] both figures illustrate three regions where there are magnetic charge ; upper and lower tube ends , and the region of the dw around the wall position @xmath113 , which has been highlighted with red colour .
there are negative surface charges at the tubes ends , because @xmath114 and @xmath115 .
also , there is a positive volume charge localized in the domain wall region . a volume magnetic charge @xcite is given by @xmath116 , and thus in our case @xmath117 .
besides , @xmath118 , and therefore , @xmath119 , just in the dw region .
therefore , the dw acts as a source of magnetic charge . note
that we are considering here head - to - head dws , and in order to analyze the case of tail - to - tail walls , we should replace @xmath120 by @xmath121 and the corresponding magnetic charges must have the opposite sign as discussed here for head - to - head walls . .
parameters are the same of figure [ vw500 ] while the dashed region represent the region occupied by the magnetic matter . , width=302 ] finally , in figure [ hdstrength ] we illustrate the strength of the dipolar field as a function of the distance to the z axis , and for different @xmath19 values .
we have considered the same parameters as in figure [ vw500 ] which means that we have magnetic matter just for @xmath122 and @xmath123 . to visualize the dipole field produced by a transverse wall , we have chosen mns with the same parameters of the above section , that is @xmath75 , @xmath124 and @xmath108 . following a previous work
@xcite we can calculate the transverse dw width ( @xmath52 ) by minimizing the total energy , as stated before .
again , it is found that @xmath125 , and for the values mentioned above we obtain @xmath126 , a rather small value as compared to the corresponding dw width for the vw of the above section . the magnetic field for the transverse wall configuration is given by equations ( htrho-[htz ] ) .
note that the angular component of the field @xmath127 is not zero ( unless @xmath128 ) and therefore the field does not present azimuthal symmetry as the magnetostatic field for the vortex dw .
it can be shown that for the geometry discussed here , the vector field along the plane defined by @xmath129 lies completely in that plane , where @xmath130 . along this plane
it is found that the field is very similar to the field for the vw ( see figure [ vw500 ] ) .
this can be understood by noting that , mathematically , the field for the tw is very similar to the case of the vw .
the main differences are the wall width and the term dependent on the angle @xmath58 , and the function @xmath131 defined in ( [ a ] ) . in figures
[ tw1 ] and [ tw2 ] we show the magnetostatic field produced by the magnetization of a transverse dw located at @xmath112 . figure [ tw1 ] depicts the top view of the field along the plane @xmath132 , and figure [ tw2 ] shows the plane @xmath133 .
both plots represents the in - plane component of the field ( @xmath134 ) normalized to their strengths given by the colour code .
the central ring illustrates the top view of the nanotube .
we have also included information about @xmath135 which is given by the length of the arrows .
it can be noted that near the ferromagnetic tube , the z component of the field is more important , and thus the arrows are smaller . ) .
the plot represent the top view of the field , at the plane @xmath132 .
the colour indicates the strength of the magnetostatic field , and along the contours the value of the field is constant .
parameters are presented in the text.,width=302 ] .
the colours indicate the strength of the magnetostatic field . along the contours
the value of the field is constant .
parameters are presented in the text.,width=302 ]
in summary , we have investigated the dipole field produced by equilibrium and non - equilibrium domain wall states of magnetic nanotubes . using a continuous model we have obtained simple expressions to evaluate the dipole fields . on one hand
, we can conclude that the consideration of magnetic states , as the non - uniform mixed state instead the idealized uniform state , reduces considerably the magnetostatic field .
thus , it is important for researchers to have in mind this approach when they compare with experimental results , in order to avoid the overestimation of stray fields . on the other side , magnetic domain walls , which can be manipulated by external fields or spin - currents ,
introduce volumetric magnetic charges into the nanostructure .
these charges move in the direction of the applied field if the dw is a head - to - head wall , and in the opposite direction if the wall is a tail - to - tail one .
finally , we have concluded that the magnetostatic field strongly depends on the mechanisms of magnetization reversal .
our results are intended to provide guidelines to use the magnetostatic field generated by tubular nanostructures in prospective applications such as the generation of a magnetic trap .
this work was partially supported by fondecyt grants no . 11080246 and 11070010 , financiamiento basal para centros cientficos y tecnolgicos de excelencia , millennium science initiative under project p06 - 022-f and the program bicentenario en ciencia y tecnologa pbct under project psd-031 .
landeros pedro , escrig juan and altbir dora 2009 in _ electromagnetic , magnetostatic , and exchange - interaction vortices in confined magnetic structures _ edited by e. o. kamenetskii ( research signpost : kerala ) . | the non - uniform magnetostatic field produced by the equilibrium and non equilibrium magnetic states of magnetic nanotubes has been investigated theoretically .
we consider magnetic fields produced by actual equilibrium states and transverse and vortex domain walls confined within the nanostructure .
our calculations allow us to understand the importance of the magnetostatic field in nanomagnetism , which is frequently considered as a uniform field .
moreover , our results can be used as a basis for future research of others properties , such as the investigation of spin waves when domain walls are present , or the motion of a magnetic particle near a magnetic field . |
large fluctuations are a distinguishing feature of soft particulate flows , like flows of granular media @xcite , bubbles and foams @xcite , and in living matter such as biological tissues @xcite .
very dense systems are in a jammed state .
they only move in response to a strong external force .
less packed systems are in a fluid state .
they flow in response to any finite force .
the flows are highly intermittent and involve rare , very large fluctuations @xcite that can trigger transitions between the jammed and the fluid state @xcite .
landslides @xcite and avalanches @xcite are transitions from a jammed to a fluid state .
clogging of hoppers @xcite and breakdown of silos @xcite involve the transition from a fluid to a jammed state .
predicting the frequency of appearance of such fluctuations is a question of great practical and theoretical interest .
fluctuation relations ( frs ) compare the probability of the forward progression of a dynamics and its reverse ; akin of watching a movie played in forward and reverse direction .
they provide an exact symmetry property of the probability distribution function characterizing the likelihood to encounter a given course of states in an observation of the dynamics .
close to equilibrium this symmetry entails linear response .
far from equilibrium frs have been adopted in micro - biological systems to determine the free energy of a folding rna @xcite and thermodynamic properties of other biomolecules @xcite .
in contrast to the dynamics of the microscopic biological systems the dynamics of most macroscopic systems do not move against an exerted force @xcite .
the strongly fluctuating and intermittent flows of soft particulate matter are a noticeable exception to this rule . here
, we analyze the statistics of those very large fluctuations where the flow is moving against a driving force .
we discuss rare fluctuations in flows of soft particulate matter , where the injected power , @xmath0 takes negative values in a finite domain that is subjected to a velocity gradient @xmath1 and that resists flow by a shear stress @xmath2 ( the shear stress is the force resisting the flow , see supplementary notes 1 , 2 , and 3 ) . in a steady state the injected energy balances the energy dissipated by the viscosity of the fluid .
hence , on average @xmath3 takes a positive value , and in the thermodynamic limit it does not fluctuate . when there is a finite number of particles in the considered domain there is a small chance to encounter rare fluctuations where @xmath3 takes a negative value
this can either be due to the reversal of the shear stress @xmath2 or to the velocity gradient @xmath1 . while one might naively expect that fluctuations in @xmath2 and @xmath1 would equally contribute to such violations ,
our numerical simulations show an unexpected interplay of these two mechanisms of rare fluctuations .
moreover , we establish a variation of fluctuation relation ( fr ) for the statistics of the injected power driving the flow and use it to define an effective temperature for far - from - equilibrium soft particulate flows .
our approach can be easily generalized to study negative power fluctuations and effective temperatures both in simulations and experiments in a wide range of problems such as in the sheared foams , vibrated granular media , particles down an inclined plane , emulsions and other soft particulate media .
* probability distribution function of injected power @xmath3 . * in the fig .
[ fig : prob_total]-a , we show a typical example of the probability distribution function ( pdf ) , @xmath4 , of the local power flux rescaled by the mean power , i.e. , @xmath5 .
the pdf exhibits several remarkable features
. the power flux can take negative values with a rather high probability .
the distribution is strongly skewed towards positive events . at the both sides the pdf decays exponentially to a good approximations .
it is very different from a gaussian distribution .
still , the negative part of the pdf ( the shaded area in fig . [ fig : prob_total]-a ) decreases rapidly with system size . in the following this area will be denoted as @xmath6 .
the slopes of the exponential decay are roughly proportional to the number of particles in the considered volume such that @xmath7 decays exponentially to zero with system size .
* probability of rare fluctuations . * in fig . [ fig : prob_total]-b we show @xmath7 as a function of packing fraction @xmath8 .
the lines in different color indicate data for different shear rates , @xmath9 . for shear rates , @xmath10 ( olive ) ,
this probability is very small , and it grows upon decreasing the shear rate .
there is pronounced growth in the fluid and in the jammed states .
however , close to the jamming point @xmath11 ( marked by the vertical dashed line ) it remains small for all shear rates .
the critical point lies to the right of the minima .
however , the minima converge towards @xmath11 in the limit @xmath12 and system size @xmath13 .
hence , the minimum will eventually approach @xmath11 ; the discrepancy is due to finite - size effects .
the emergence of the global minimum is remarkable because fluctuations are expected to diverge close to a critical point @xcite , i.e. , one expects a larger probability to encounter fluctuations close to the critical point .
the minimum is a distinctive feature of the jamming transition .
it has no counterpart in equilibrium thermodynamics .
* decomposition of the probability of rare fluctuations . * in order to understand the origins of the negative power injection , we recall that the power flux has two contributions : the local shear stress @xmath14 , and the local velocity gradient @xmath1 .
negative power injection arises whenever either @xmath15 and @xmath16 , or @xmath17 and @xmath18 .
the former events will be denoted as @xmath19 . in this case
the velocity profile remains monotonic and negative power injection arises from fluctuations of the shear stress .
the latter events will be denoted as @xmath20 . in this case
the negative power injection is connected to rare fluctuations where the velocity profile is no longer monotonic ( see supplementary fig . 1 ) .
the joint probability of these events sum up to the probability to encounter negative power injection : @xmath21 ( in supplementary note 5 and supplementary fig .
3 we numerically prove this equality ) .
figure [ fig : prob_joint ] presents the numerical results for these two joint probabilities , i.e. , @xmath22 ( left axis , filled symbols ) and @xmath23 ( right axis , hollow symbols ) as function of packing fraction @xmath8 for various shear rates . for a given shear rate ( a given color ) , the intersection point of the joint probabilities is very close to the jamming point .
hence , @xmath11 splits the probability space into two disjoint regions .
accordingly , the two types of mechanisms , shear - stress and velocity - gradient fluctuations , are mutually exclusive .
the reason can be sketched as follows .
the positive shear stress corresponds to head - to - head collisions while the negative shear stress is associated to backup collisions during which the local angular momentum is in the same and in the opposite direction of the global induced angular momentum of the flow , respectively ( see supplementary fig . 2 and supplementary note 4 ) . for @xmath24 ,
since the average coordination number is relatively small , the negative contacts can be observed with high probability although they are suppressed by increasing the shearing flow .
this justifies the decreasing dependence of the probability of negative stress as function of both shear rate and packing fraction . for @xmath25 , since the coordination number jumps to a value @xmath26 , the negative contacts can in average be suppressed by the positive ones since the global symmetries of the flow favor the positive contacts .
this elucidates the behavior seen in fig .
[ fig : prob_joint ] for the negative shear stress and since the negative power includes exclusive contributions from the negative shear stress and negative velocity gradient , our argument naturally explains the observation of the mutually exclusive fluctuations .
* fluctuation relation . * at this point we identified qualitatively different physics underlying the fluctuations of fluid and jammed systems . in order to gain more insight into the parameter dependence of the strength of fluctuations
we establish now a fluctuation relation for the flows .
it will characterize the width of the pdfs @xmath27 by an effective temperature @xmath28 .
we conjecture that the relation @xmath29=\beta p$ ] holds where @xmath30 has inverse dimension of power .
here , the constant @xmath31 is the relevant elastic time scale which represents the typical time scale of a single collision .
it is approximately independent of @xmath8 and @xmath9 ( see supplementary note 6 ) .
figure [ fig : ft ] presents the numerical verification of our conjecture for two packing fractions @xmath32 and @xmath33 that lie below and above the critical point , @xmath34 , respectively .
there is a linear dependence between @xmath29 $ ] and @xmath3 whose slope is a decreasing function of the shear rate @xmath35 .
this implies that @xmath28 is an increasing function of the shear rate in accordance with the shear - rate dependence of the average kinetic temperature of particles in the flow .
surprisingly , the correspondence is not only qualitative .
it even holds quantitatively .
all data shown in fig . [
fig : ft]-a collapse to a straight line when @xmath29 $ ] is multiplied by the granular temperature , @xmath36 ( fig .
[ fig : t_eff_master]-b ) .
the resulting straight line has a slope @xmath37 with @xmath38 .
( in supplementary note 8 and supplementary fig . 8 , we numerically prove that our proposed fr is also satisfied for highly damped systems corresponding to non - brownian suspensions . ) * scaling collapse of effective and kinetic temperatures .
* in the jammed state the effective temperature , @xmath28 , and the granular temperature , @xmath36 , differ : @xmath28 is always larger than @xmath36 .
we explore the parameter dependence of the two temperatures by exploring their scaling properties @xcite . in fig .
[ fig : t_eff_master]-a we demonstrate that the full @xmath9 and @xmath8 dependence of the temperatures for different system sizes can be represented in terms of a master plot where @xmath39 with appropriate choice of the exponents @xmath40 , @xmath41 and @xmath42 . for the fluid state
we thus find the well - known bagnoldian scaling with exponent @xmath43 . in the jammed state , we find that @xmath28 and @xmath36 still collapse uniformly in the critical region , @xmath44 .
for jammed flows , however , they segregate into two different branches in accord with our earlier statement . in this limit @xmath28
approaches a constant yield stress emerges .
in contrast , @xmath36 shows a power - law behavior with exponent @xmath45 .
we have checked consistency of all our exponents , i.e. , exponents of granular temperature and components of stress tensor , in supplementary note 7 and supplementary fig .
we also show in supplementary note 8 that these exponents are universal in a sense that the same scaling collapse is achieved with the same critical exponents for highly dissipative regime in connection with the non - brownian suspensions .
+ heussinger et al .
@xcite have studied fluctuations of some observables in the flow of an assembly of frictionless , soft discs at zero temperature , in the vicinity of and slightly above @xmath11 .
they have found that the contact - number fluctuations and relative fluctuations of the shear stress diverge upon approaching @xmath11 from above .
they also report on strong finite - size effects when using @xmath8 as control parameter . however , the effective temperature @xmath28 in our study , is the product of thermodynamically forbidden fluctuations ( negative stress below @xmath11 and negative velocity gradient above @xmath11 ) which are specific to small - size systems and vanish in the vicinity of @xmath11 .
our observation indicates that the scaling behavior of @xmath28 is not significantly altered by the finite - size effects below and above @xmath11 for @xmath46 see fig .
[ fig : t_eff_master]-a .
we find that to a very good extent , @xmath28 is independent of the system size .
we only see small deviations for @xmath47 .
these deviations vanish as @xmath48 .
the flow of particulate matter is similar to classical fluids in so far as it involves the motion of many particles that interact by short - range forces . as function of packing fraction
the flows undergo a phase transition from a fluid into a jammed state .
close to the critical point the materials obey scaling relations @xcite , reminiscent of critical phenomena .
the data collapse of the granular temperature , i.e. , the kinetic energy per degree of freedom , is shown here in fig .
[ fig : t_eff_master]-a . in the fluid state and in the critical region
this temperature agrees with an effective temperature that characterizes the probability to encounter different power injections .
the proposed effective temperature is sensitive to the inherent properties of the systems , and it potentially qualifies as the effective temperature that has been searched for recently with great urgency @xcite .
the effective temperatures proposed in the past @xcite are based on fluctuation - dissipation relations , i.e. , they assume linear - response .
our study goes beyond linear response by introducing a fluctuation relation in order to define a shear - rate dependent effective temperature .
this effective temperature is valid for packing fractions far from the jamming point , in contrast to the previous ones that are meaningful measure only near the transition point @xcite .
+ various types of fluctuation theorems have been extensively studied over the last two decades @xcite .
motivated by molecular dynamics simulations , evans et al .
@xcite proposed an empirical fluctuation relation for entropy production rate in a @xmath43-dimensional sheared lennard - jones fluid .
later , this empirical relation was rigorously proved by gallavotti and cohen @xcite .
this is now known as steady - state fluctuation theorem ( ssft ) . in a steady - state fluctuation theorem ,
the entropy production rate is time averaged over a single , randomly sampled interval of duration @xmath31 .
in contrast , the transient fluctuation theorem ( tft ) of evans and searles @xcite applies to a system that evolves from an initial equilibrium state to a nonequilibrium steady state .
tfts are different from ssfts from a practical point of view . whereas tfrs rely on ensemble averaging all starting from the same initial macro - state , ssfrs
may be verified from steady - state evolution of a system over a sufficiently long time @xcite . as we have already stressed out ,
properties of soft particulate flows are predominated by the rare fluctuations which result to intermittent behavior of these flows .
it can be seen that fluctuation relations of type ssfts are not suitable for soft particulate flows .
the reason is that as a consequence of averaging process during the sampling time , the rare fluctuations can be washed out .
therefore , we use an instantaneous , time - independent fluctuation relation to characterize strength of rare fluctuations . whether our postulated fluctuation relation would enjoy a rigorous treatment ,
will remain a theoretical challenge .
in addition , we have shown here that fluctuations in soft particulate flows differ essentially from those of classical fluids .
first of all , they are very strong as demonstrated by the exponential decay of the pdfs of negative power injection ( fig . [
fig : prob_total]-a ) rather than the much faster decay of gaussian distributions .
even for shear rates as large as @xmath49 we observe negative power injection ( lowermost curve in fig .
[ fig : prob_total]-b ) .
even more surprising , rare fluctuations are strongly suppressed close to the critical point ( minima of the curves in fig .
[ fig : prob_total]-b ) .
they behave exactly contrary to the strength of critical fluctuations that diverge at the critical point and die out rapidly outside the critical region @xcite .
it will be challenging numerically , but extremely interesting from a conceptual point of view to explore how classical fluids behave in this respect .
finally , we have shown that there are different physical mechanisms underlying rare fluctuations in fluid and jammed states : in fluid states negative power injection originates from fluctuations where the shear stress takes a negative sign in jammed states they arise in regions with negative shear rates .
this dichonometry of mutually exclusive mechanisms of rare fluctuations above and below a critical state is a distinctive feature of soft particulate flows that has no counterpart in equilibrium thermodynamics .
rheological properties of particulate flows are commonly characterized in terms of hydrodynamic equations and constitutive relations @xcite .
fluctuations are not a part of the modeling .
the present study takes a different approach to characterize the systems : we focus on the fluctuations as an inherent property of the dynamics .
remarkably , the fluctuations obey a local , time - independent fluctuation relation , and this relation can be used to define an effective temperature of the system .
in contrast to the hydrodynamic approaches the temperature is not a field variable in this setting . rather it characterizes the variability of snapshots of the flow .
it is a scalar quantity that characterizes the ensemble of observations of the flows , taking full note of fluctuations .
it neither requires to find appropriate heuristic constitutive equations , nor does it rely on the scale separation at mesoscopic scales that is implicit to the definition of thermodynamic fields .
hence , it is less prone to pitfalls arising from inapt choices of constitutive relations and applicable to a larger class of far - from - equilibrium flows .
thus , the present study opens a qualitatively new road to the description of far - from - equilibrium particulate flows .
+ in a recent study , maloney @xcite investigated distribution of dissipation power @xmath50 in durian s bubble model . whereas we find that @xmath27 has always exponential tails , it is shown that above jamming density @xmath50 becomes power law for small shear rates .
this implies that distributions of injection and dissipation powers might not be equivalent .
stationarity condition implies that first moments of these distributions must be equal .
but as one can see , higher moments of these distributions , which refer to the characteristic of the tails , might be different .
this suggests a new avenue of research for investigation of steady state properties of non - equilibrium states . in this context ,
maloney s work @xcite together with our approach to calculate @xmath27 in shear flows provide a solid framework for investigation of similarities and differences between distributions of injection and dissipation powers .
we perform molecular - dynamics simulations of two - dimensional frictionless bidisperse disks .
particles interact via short range repulsive and dissipative forces .
two particles @xmath51 and @xmath52 of radii @xmath53 and @xmath54 ( where @xmath55 , @xmath37 stand for two different radius of bidisperse particles ) at positions @xmath56 and @xmath57 interact when @xmath58 . here
@xmath59 is called the mutual compression of particles @xmath51 and @xmath52 , @xmath60 .
the particles interact via a linear dashpot model , @xmath61 , where @xmath62 and @xmath63 are denoted as elastic and dissipative constant , respectively . throughout the study
we adopt the values @xmath64 and @xmath65 , respectively . in order to prevent crystallization we use a @xmath66 binary mixture of particles where the ratio of the radii of large and small particles is set to @xmath67 .
the equations of motion are non - dimensionalized by choosing the unit of the length to be @xmath68 , and setting the mass of each particle equal to its area , @xmath69 ^ 2 $ ] .
finally , the ratio of @xmath62 and @xmath63 provides the time scale @xmath70 .
lees - edwards boundary conditions are applied along @xmath71-direction .
they create a uniform overall shear rate , @xmath9 .
these equations of motion are integrated with a @xmath72-order predictor - corrector gear algorithm with time step , @xmath73 .
* acknowledgments . * fruitful discussions with j. nagler and m. schrter are highly acknowledged .
s. h. e. r. thanks the korea institute for advanced study for providing computing resources ( kias center for advanced computation - linux cluster system ) for this work , and specially consultations from hoyoung kim .
a.a.s . would like to acknowledge supports from the alexander von humboldt foundation , and partial financial supports by the research council of the university of tehran .
we also would like to thank referees of nat .
comm . for fruitful comments which eventually increased quality of our manuscript
. + * data availability . *
the authors declare that the data supporting the findings of this study are available from the authors on request .
s.h.e.r . performed the simulations .
all authors discussed the data , participated in the data analysis , and in writing the manuscript .
the first two authors had equal contributions in the development of the study .
the authors declare no competing financial interests . for @xmath32 , @xmath74 , and @xmath75 .
the solid straight lines show the exponential decay of the pdf for large and small arguments .
the area of the shaded region gives the probability , @xmath7 , that the local power takes a negative value , i.e. , to encounter a negative power injection .
( * b * ) the probability to observe a negative power injection @xmath7 as a function of packing fraction @xmath8 for different shear rates @xmath76 and @xmath77 from top to bottom , respectively , and system size @xmath75 .
the vertical dashed line marks the critical packing fraction , @xmath11 , where the jamming transition occurs in the static limit .
error bars correspond to square root of variance .
[ fig : prob_total],title="fig:",scaledwidth=47.5% ] + ( left axis , filled symbols ) and @xmath78 ( right axis , hollow symbols ) as functions of packing fraction , @xmath8 , for various shear rates , @xmath9 . in the fluid state , @xmath24 , the dominant mechanism of negative power injection is the reversal of the shear stress . in the jammed state ,
@xmath25 , it is due to the reversion of the velocity gradient .
error bars correspond to square root of variance .
[ fig : prob_joint],title="fig:",scaledwidth=50.0% ] + $ ] vs @xmath3 for two packing fractions ( * a * ) @xmath32 and , ( * b * ) @xmath79 .
the solid lines are linear fits of slope @xmath80 of the data for different shear rates
. the slope decreases by increasing the shear rate @xmath35 , implying that the effective temperature @xmath28 increases by @xmath35 .
the slope has a weak dependence on @xmath35 in the jammed state . for @xmath81 and @xmath82 representing number of positive @xmath83 and negative @xmath84 cases , the corresponding error bar of @xmath85 is equal to @xmath86 . [
fig : ft],title="fig:",scaledwidth=50.0% ] + and @xmath87 the effective and granular temperatures collapse onto a scaling function .
however , to achieve a data collapse we had to adopt slightly different critical densities , @xmath88 and @xmath89 for @xmath28 and @xmath36 , respectively . in the fluid state and in the critical state the temperatures match . for the fluid state they exhibit bagnoldian scaling with exponent @xmath43 . in the critical state they still share same non - trivial scaling for @xmath90 . in the jammed state the temperatures segregate into two different branches ; @xmath28 approaches a constant and @xmath36 follows a power - law behavior with exponent @xmath91 .
different system sizes are given by different symbols in which filled and hollow symbols refer to @xmath36 and @xmath28 , respectively .
the color code corresponds to different shear rates @xmath92 ( purple ) , @xmath93 ( magenta ) , @xmath94 ( blue ) , @xmath95 ( golden ) , and @xmath96 ( yellow ) .
( * b * ) the collapse of all data presented in fig .
[ fig : ft]-a when the vertical axis is multiplied by a factor of @xmath97 with @xmath38 . in these data
, we cover a large range of packing fractions around jamming , @xmath98 .
error bars correspond to square root of variance .
[ fig : t_eff_master],title="fig:",scaledwidth=45.0% ] + | soft particulate media include a wide range of systems involving athermal dissipative particles both in non - living and biological materials .
characterization of flows of particulate media is of great practical and theoretical importance .
a fascinating feature of these systems is the existence of a critical rigidity transition in the dense regime dominated by highly intermittent fluctuations that severely affects the flow properties . here
, we unveil the underlying mechanisms of rare fluctuations in soft particulate flows .
we find that rare fluctuations have different origins above and below the critical jamming density and become suppressed near the jamming transition .
we then conjecture a time - independent local fluctuation relation , which we verify numerically , and that gives rise to an effective temperature .
we discuss similarities and differences between our proposed effective temperature with the conventional kinetic temperature in the system by means of a universal scaling collapse . |
language models play an important role in many applications like speech recognition , machine translation , information retrieval and nature language understanding .
traditionally , the back - off n - gram models @xcite are the standard approach to language modeling .
recently , neural networks have been successfully applied to language modeling and have achieved the state - of - the - art performance in many tasks . in neural network language models ( nnlm ) , the feedforward neural networks ( fnn ) and recurrent neural networks ( rnn ) @xcite are two popular architectures .
the basic idea of nnlms is to use a projection layer to project discrete words into a continuous space and estimate word conditional probabilities in this space , which may be smoother to better generalize to unseen contexts .
fnn language models ( fnn - lm ) @xcite usually use a limited history within a fixed - size context window to predict the next word .
rnn language models ( rnn - lm ) @xcite adopt a time - delayed recursive architecture for the hidden layers to memorize the long - term dependency in language .
therefore , it is widely reported that rnn - lms usually outperform fnn - lms in language modeling . while rnns are theoretically powerful , the learning of rnns needs to use the so - called back - propagation through time ( bptt ) @xcite due to the internal recurrent feedback cycles .
the bptt significantly increases the computational complexity of the learning algorithms and it may cause many problems in learning , such as gradient vanishing and exploding @xcite .
more recently , some new architectures have been proposed to solve these problems .
for example , the long short term memory ( lstm ) rnn @xcite is an enhanced architecture to implement the recurrent feedbacks using various learnable gates , and it has obtained promising results on handwriting recognition @xcite and sequence modeling @xcite . moreover ,
the so - called temporal - kernel recurrent neural networks ( tkrnn ) @xcite have been proposed to handle the gradient vanishing problem .
the main idea of tkrnn is to add direct connections between units in all time steps and every unit is implemented as an efficient leaky integrator , which makes it easier to learn the long - term dependency . along this line
, a temporal - kernel model has been successfully used for language modeling in @xcite . comparing with rnn - lms , fnn - lms
can be learned in a simpler and more efficient way .
however , fnn - lms can not model the long - term dependency in language due to the fixed - size input window . in this paper
, we propose a novel encoding method for discrete sequences , named _ fixed - size ordinally - forgetting encoding _ ( fofe ) , which can almost uniquely encode any variable - length word sequence into a fixed - size code . relying on a constant
forgetting factor , fofe can model the word order in a sequence based on a simple ordinally - forgetting mechanism , which uses the position of each word in the sequence .
both the theoretical analysis and the experimental simulation have shown that fofe can provide _ almost _ unique codes for variable - length word sequences as long as the forgetting factor is properly selected . in this work ,
we apply fofe to neural network language models , where the fixed - size fofe codes are fed to fnns as input to predict next word , enabling fnn - lms to model long - term dependency in language .
experiments on two benchmark tasks , penn treebank corpus ( ptb ) and large text compression benchmark ( ltcb ) , have shown that fofe - based fnn - lms can not only significantly outperform the standard fixed - input fnn - lms but also achieve better performance than the popular rnn - lms with or without using lstm . moreover
, our implementation also shows that fofe based fnn - lms can be learned very efficiently on gpus without the complex bptt procedure .
assume vocabulary size is @xmath0 , nnlms adopt the 1-of - k encoding vectors as input . in this case , each word in vocabulary is represented as a one - hot vector @xmath1 .
the 1-of - k representation is a context independent encoding method .
when the 1-of - k representation is used to model a word in a sequence , it can not model its history or context .
we propose a simple context - dependent encoding method for any sequence consisting of discrete symbols , namely _ fixed - size ordinally - forgetting encoding _ ( fofe ) . given a sequence of words ( or any discrete symbols ) , @xmath2 , each word @xmath3 is first represented by a 1-of - k representation @xmath4 , from the first word @xmath5 to the end of the sequence @xmath6 , fofe encodes each partial sequence ( history ) based on a simple recursive formula ( with @xmath7 ) as : @xmath8 where @xmath9 denotes the fofe code for the partial sequence up to @xmath3 , and @xmath10 ( @xmath11 ) is a constant forgetting factor to control the influence of the history on the current position .
let s take a simple example here , assume we have three symbols in vocabulary , e.g. , _ a _ , _ b _ , _ c _ , whose 1-of - k codes are @xmath12 $ ] , @xmath13 $ ] and @xmath14 $ ] respectively . in this case , the fofe code for the sequence _
\{abc } _ is @xmath15 $ ] , and that of _ \{abcbc } _ is @xmath16 $ ] .
obviously , fofe can encode any variable - length discrete sequence into a fixed - size code .
moreover , it is a recursive context dependent encoding method that smartly models the order information by various powers of the forgetting factor .
furthermore , fofe has an appealing property in modeling natural languages that the far - away context will be gradually forgotten due to @xmath17 and the nearby contexts play much larger role in the resultant fofe codes .
given the vocabulary ( of @xmath0 symbols ) , for any sequence @xmath18 with a length of @xmath19 , based on the fofe code @xmath20 computed as above , if we can always decode the original sequence @xmath18 unambiguously ( perfectly recovering @xmath18 from @xmath20 ) , we say fofe is unique .
[ theorem - fofe - alpha - less - half ] if the forgetting factor @xmath10 satisfies @xmath21 , _ fofe _ is unique for any @xmath0 and @xmath19 .
the proof is simple because if the fofe code has a value @xmath22 in its @xmath23-th element , we may determine the word @xmath24 occurs in the position @xmath25 of @xmath18 without ambiguity since no matter how many times @xmath24 occurs in the far - away contexts ( @xmath26 ) , they do not sum to @xmath22 ( due to @xmath27 ) .
if @xmath24 appears in any closer context ( @xmath28 ) , the @xmath23-th element must be larger than @xmath22 . for @xmath29
, we have the following theorem : [ theorem - fofe - alpha - less - one ] for @xmath29 , given any finite values of @xmath0 and @xmath19 , _ fofe _ is almost unique everywhere for @xmath30 , except only a finite set of countable choices of @xmath10 .
the complete proof @xcite is given in appendix a. based on theorem [ theorem - fofe - alpha - less - one ] , fofe is unique almost everywhere between @xmath31 only except a countable set of isolated choices of @xmath10 . in practice ,
the chance to exactly choose these isolated values between @xmath31 is extremely slim , realistically almost impossible due to quantization errors in the system . to verify this
, we have run simulation experiments for all possible sequences up to @xmath32 symbols to count the number of collisions .
each collision is defined as the maximum element - wise difference between two fofe codes ( generated from two different sequences ) is less than a small threshold @xmath33 . in figure
[ fig : collisions ] , we have shown the number of collisions ( out of the total @xmath34 tested cases ) for various @xmath10 values when @xmath35 , @xmath36 and @xmath37 . , the magnitudes of the resultant fofe codes become much larger . as a result ,
the number of collisions ( as measured by a fixed absolute threshold @xmath33 ) becomes smaller .
] the simulation experiments have shown that the chance of collision is extremely small even when we allow a word to appear any times in the context . obviously , in a natural language , a word normally does not appear repeatedly within a near context .
moreover , we have run the simulation to examine whether collisions actually occur in two real text corpora , namely ptb ( 1 m words ) and ltcb ( 160 m words ) , using @xmath35 , we have not observed a single collision for nine different @xmath10 values between @xmath38 $ ] ( incremental @xmath39 ) .
the architecture of a fofe based neural network language model ( fofe - fnnlm ) is as shown in figure [ fig : fofe_bigram ] .
it is similar to standard bigram fnn - lms except that it uses a fofe code to feed into neural network lm at each time instance .
moreover , the fofe can be easily scaled to other n - gram based neural network lms . for example , figure [ fig : fofe_trigram ] is an illustration of fixed - size ordinally forgetting encoding based tri - gram neural network language model .
fofe is a simple recursive encoding method but a direct sequential implementation may not be efficient for the parallel computation platform like gpus .
here , we will show that the fofe computation can be efficiently implemented as sentence - by - sentence matrix multiplications , which are particularly suitable for the mini - batch based stochastic gradient descent ( sgd ) method running on gpus . given a sentence , @xmath40 , where each word is represented by a 1-of - k code as @xmath4 @xmath41 .
the fofe codes for all partial sequences in @xmath18 can be computed based on the following matrix multiplication : @xmath42 \left [ \begin{gathered } { \bf e}_1 \hfill \\ { \bf e}_2 \hfill \\ { \bf e}_3 \hfill \\
\hspace{0.1 cm } \vdots \hfill \\ { \bf e}_t \hfill \\
\end{gathered } \right ] = { \bf m } { \bf v}\ ] ] where @xmath43 is a matrix arranging all 1-of - k codes of the words in the sentence row by row , and @xmath44 is a @xmath19-th order lower triangular matrix . each row vector of @xmath45 represents a fofe code of the partial sequence up to each position in the sentence .
this matrix formulation can be easily extended to a mini - batch consisting of several sentences .
assume that a mini - batch is composed of n sequences , @xmath46 , we can compute the fofe codes for all sentences in the mini - batch as follows : @xmath47 \left [ \begin{gathered } { \bf v}_1 \hfill \\ { \bf v}_2\hfill \\ \quad \vdots \hfill \\ { \bf v}_n \hfill \\
\end{gathered } \right ] = { \bf \bar{m } } { \bf \bar{v}}\ ] ] when feeding the fofe codes to fnn as shown in figure [ fig : fofe_bigram ] , we can compute the activation signals ( assume @xmath48 is the activation function ) in the first hidden layer for all histories in @xmath18 as follows : @xmath49 where @xmath50 denotes the word embedding matrix that projects the word indices onto a continuous low - dimensional continuous space . as
above , @xmath51 can be done efficiently by looking up the embedding matrix .
therefore , for the computational efficiency purpose , we may apply fofe to the word embedding vectors instead of the original high - dimensional one - hot vectors . in the backward pass
, we can calculate the gradients with the standard back - propagation ( bp ) algorithm rather than bptt . as a result ,
fofe based fnn - lms are the same as the standard fnn - lms in terms of computational complexity in training , which is much more efficient than rnn - lms .
we have evaluated the fofe method for nnlms on two benchmark tasks : i ) the penn treebank ( ptb ) corpus of about 1 m words , following the same setup as @xcite .
the vocabulary size is limited to 10k .
the preprocessing method and the way to split data into training / validation / test sets are the same as @xcite .
ii ) the large text compression benchmark ( ltcb ) @xcite . in ltcb
, we use the _ enwik9 _
dataset , which is composed of the first @xmath52 bytes of enwiki-20060303-pages-articles.xml .
we split it into three parts : training ( 153 m ) , validation ( 8.9 m ) and testing ( 8.9 m ) sets .
we limit the vocabulary size to 80k for ltcb and replace all out - of - vocabulary words by a @xmath53unk@xmath54 token .
details of the two datasets can be found in table [ tab : datasets ] .
.the size of ptb and ltcb corpora in words . [ cols="^,^,^,^",options="header " , ] [ tab : wiki_summary ] we have further examined the fofe based fnn - lms on a much larger text corpus , i.e. ltcb , which contains articles from wikipedia .
we have trained several baseline systems : i ) two n - gram lms ( 3-gram and 5-gram ) using the modified kneser - ney smoothing without count cutoffs ; ii ) several traditional fnn - lms with different model sizes and input context windows ( bigram , trigram , 4-gram and 5-gram ones ) ; iii ) an rnn - lm with one hidden layer of 600 nodes using the toolkit in @xcite , in which we have further used a spliced sentence bunch in @xcite to speed up the training on gpus .
moreover , we have examined four fofe based fnn - lms with various model sizes and input window sizes ( two 1st - order fofe models and two 2nd - order ones ) . for all nnlms , we have used an output layer of the full vocabulary ( 80k words ) . in these experiments ,
we have used an initial learning rate of 0.01 , and a bigger mini - batch of 500 for fnn - lmms and of 256 sentences for the rnn and fofe models .
experimental results in table [ tab : wiki_summary ] have shown that the fofe - based fnn - lms can significantly outperform the baseline fnn - lms ( including some larger higher - order models ) and also slightly overtake the popular rnn - based lm , yielding the best result ( perplexity of 107 ) on the test set .
in this paper , we propose the fixed - size ordinally - forgetting encoding ( fofe ) method to _ almost _ uniquely encode any variable - length sequence into a fixed - size code . in this work
, fofe has been successfully applied to neural network language modeling .
next , fofe may be combined with neural networks @xcite for other nlp tasks , such as sentence modeling / matching , paraphrase detection , machine translation , question and answer and etc .
this work was supported in part by the science and technology development of anhui province , china ( grants no .
2014z02006 ) and the fundamental research funds for the central universities from china , as well as an nserc discovery grant from canadian federal govenment .
we appreciate dr .
barlas oguz at microsoft for his insightful comments and constructive suggestions on theorem [ theorem - fofe - alpha - less - one ] .
* theorem 2 : * _ for @xmath29 , given any finite values of @xmath0 and @xmath19 , _ fofe _ is almost unique everywhere for @xmath30 , except only a finite set of countable choices of @xmath10 . _
* proof : * when we decode a given fofe code of an unknown sequence @xmath18 ( assume the length of @xmath18 is not more than @xmath19 ) , for any single value @xmath55 in the @xmath23-th position of the fofe code , there are only two possible cases that may lead to ambiguity in decoding : ( i ) word @xmath24 appears in the current location of @xmath18 ; or ( ii ) word @xmath24 appears multiple times in the history of @xmath18 and the total contribution of them happens to be @xmath55 .
for case ( ii ) to happen , the forgetting factor @xmath10 needs to satisfy at least one of the following polynomial equations : @xmath56 where the above coefficients , @xmath57 , are equal to either @xmath58 or @xmath59 .
if the word @xmath24 appears in the @xmath25-th location ahead in the history , we have @xmath60 . otherwise , @xmath61 . we know , each equation in eq.([eq - fofe - unique - formula ] ) is a @xmath19-th ( or lower ) order polynomial equation .
it can have at most @xmath19 real roots for @xmath10 .
moreover , since @xmath62 , we can only have a finite set of equations in eq.([eq - fofe - unique - formula ] ) .
the total number is not more than @xmath63 .
therefore , in total , we can only have a finite number of @xmath10 values that may satisfy at least one equation in eq.([eq - fofe - unique - formula ] ) , i.e. , at most @xmath64 possible roots @xcite . among them , only a fraction of these roots lies between @xmath31 . except these countable choices of @xmath10 values , eq.([eq - fofe - unique - formula ] ) never holds for any other @xmath10 values between @xmath31 . as a result ,
case ( ii ) never happens in decoding except some isolated points of @xmath10 .
this proves that the resultant fofe code is _ almost _ unique between @xmath31 .
@xmath65 slava katz .
estimation of probabilities from sparse data for the language model component of a speech recognizer .
_ ieee transactions on acoustics , speech and signal processing ( assp ) _ , volume 35 , no 3 , pages 400 - 401 .
tomas mikolov , stefan kombrink , lukas burget , jan cernocky and sanjeev khudanpur .
extensions of recurrent neural network language model . in _ proc . of international conference on acoustics , speech and signal processing ( icassp ) _ , pages 5528 - 5531 .
yong - zhe shi , wei - qiang zhang , meng cai and jia liu .
temporal kernel neural network language model . in _ proc . of international conference on acoustics , speech and signal processing ( icassp)_. pages 8247 - 8251 . | in this paper , we propose the new fixed - size ordinally - forgetting encoding ( fofe ) method , which can almost uniquely encode any variable - length sequence of words into a fixed - size representation .
fofe can model the word order in a sequence using a simple ordinally - forgetting mechanism according to the positions of words . in this work ,
we have applied fofe to feedforward neural network language models ( fnn - lms ) .
experimental results have shown that without using any recurrent feedbacks , fofe based fnn - lms can significantly outperform not only the standard fixed - input fnn - lms but also the popular recurrent neural network ( rnn ) lms . |
metastable molecules are formed due to collision of two atoms , when the energy of a bound molecular state in a closed channel is close to the energy of two atoms in the center of mass in an open channel .
the weak coupling between the channels leads to the strong mixing of them , and the resulting metastable molecular states , which called the feshbach resonances @xcite , will have a finite lifetime and will decay both in the initial channel and in the other channels . the process of formation if intermediate metastable molecular states can be controlled , in particular , by the external magnetic field by the change of the resonance detuning , when the bound state in a closed channel has the different hyperfine state then the incoming atoms in the open channel . in this case
, the difference between the hyperfine states , which is caused by the difference in the zeeman shifts , can be controlled by the detuning of the feshbach resonance by means of the magnetic field @xcite . in case where magnetically tunable resonances are absent , an alternative method for control of feshbach resonance is the optical method [ 5 ] ( optical feshbach resonance ) .
optical control of feshbach resonances by means of quantum interference was proposed in work @xcite . in the theory of resonant collision , in addition to feshbach method @xcite , which has been developed for studies of nuclear reactions and is successfully applied for collision atoms in bec , there is an alternative fano approach @xcite exists exploiting the configuration interaction in multielectron atoms .
both approaches assume appearance of resonance phenomena when discrete states are coupled with continuum .
fano technique is usually associated with asymmetry of shapes of resonance lines which is known in atomic physics as `` fano profile '' .
similar interference phenomena of asymmetry of resonance line shape are also observed in nuclear reactions @xcite .
fano technique is , however , used not only in atomic physics for , e.g. , studies of autoionization and rydberg states [ 9 ] , resonance ionization of atoms @xcite , and laser induced continuum structures ( lics ) @xcite .
the fano technique is also used for considerations of resonant collision [ 9 ] including those of electrons with atoms with formation of negative ions and interference phenomena in the field of laser radiation @xcite .
fano technique is widely used in also other fields of physics .
it is , e.g. , used for explanation of asymmetry in the absorption impurity ions in crystals , which is caused by for formation of excitonic resonances @xcite . with use of these resonances works @xcite , study the phenomenon of storage and reconstruction of quantum information in solids .
the asymmetric form of the fano resonance important in considerations of nanoscale structures of interacting quantum systems @xcite . in the present work we consider collision of atoms with formation of the metastable molecules ( feshbach resonances ) and
the subsequent stimulated transition to the lower unbound molecular electronic states with emission of photon of the laser radiation .
expressions for the cross sections of the elastic and inelastic resonance scattering and the intensity of the stimulated emission of the photons have been obtained .
consider elastic and inelastic collision of two atoms with formation of feshbach resonance ( fig.1 ) .
@xmath2 the interaction which couples electronic states in the open and closed channel .
laser radiation with frequency @xmath3 couples the upper molecular quasi bound state with the lower uncoupled molecular state with interaction @xmath4 .
is excited atom b. ] the hamiltonian for the considered process ( fig.1 ) has the following form : @xmath5 we represent the solution of the schrdinger equation with the hamiltonian ( [ eq : hamiltonian ] ) as : @xmath6 ) for the wave vector and using that @xmath7 , where @xmath3 is the frequency of laser radiation , into the shrdinger equation we obtain the following system of differential equations for the coefficients of the expansion ( [ eq : shrodinger ] ) : @xmath8 after the fourier transform for the coefficient in the formula ( [ eq : shrodinger ] ) @xmath9 for the fourier components we obtain for the fourier components of the expansion coefficient the following system equations : @xmath10 @xmath11 @xmath12 now we obtain from the equation ( [ eq:5b]),([eq:5c ] ) @xcite : @xmath13u^*_ec_e(\lambda)\end{aligned}\ ] ] @xmath14\omega^*_{e-\omega}c_e(\lambda),\end{aligned}\ ] ] where @xmath15 in expression ( [ eq7 ] ) @xmath16 and @xmath17 are the full resonant shifts and width and @xmath18 denotes the principle value : @xmath19 by substituting expressions ( [ eq:8]),([eq:9 ] ) into formula ( [ eq : shrodinger ] ) from the ortonormalization condition we obtain the following expression for the @xmath20 : @xmath21 for the first solution obtain the following expression : @xmath22 , we have the following expression @xmath23 ) and ( [ eq10 ] ) are provide the ortonormalization condition for quasienergy functions : @xmath24
asymptotic ( @xmath25 ) expressions for continuous - spectrum wave functions with orbital angular momentum l are known to have the following appearance : @xmath26 with @xmath27 begin the legendre polynomials .
taking into account that the wave functions of bound states of atoms vanish asymptotically ( @xmath28 ) at large distance ( @xmath25 ) we can write the quasienergy wave functions ( [ eq9])([eq10 ] ) as follows : @xmath29 @xmath30 where @xmath31 is the phase of non - resonant scattering and @xmath32 is the phase caused by resonant feshbach scattering .
@xmath33 we now represent the scattering state vector @xmath34 at @xmath35 as superposition of quasienergy function ( [ eq13]),([eq14 ] ) @xmath36 then , if we require the presents of incoming and outgoing waves in the first , elastic , channel and the absence of incoming wave in the second , inelastic , channel we can write for the expansion coefficients in ( [ eq16 ] ) @xmath37 the following : @xmath38 from expressions ( [ eq18 ] ) we can obtain for scattering state vector ( [ eq16 ] ) : @xmath39 p_{l_1}(\cos\theta_1)-e^{i\omega t } \frac{\sin\eta}{k(\lambda-\omega)r_2 } \frac{\sqrt{\gamma_f(\lambda)\gamma_l(\lambda)}}{\gamma(\lambda ) } \nonumber \\ e^{i\delta_2}e^{-i(k(\lambda-\omega))r_2-\frac{l_2\pi}{2}}p_{l_2}(\cos\theta_2)\end{aligned}\ ] ] from expression ( [ eq19 ] ) we can obtain the formulas for corresponding cross sections @xmath40 let as separate in ( [ eq20a ] ) for elastic scattering the resonant state and write the potential part of scattering cross sections in the form @xmath41 the total cross section of elastic scattering we write as : @xmath42 where @xmath43 from expressions ( [ eq7 ] ) and ( [ eq:15 ] ) for the resonant elastic and inelastic cross sections , we obtain : @xmath44 ] ] distribution for the spectral intensity of the emission radiation at the stimulated transition to the lower unbound molecular state has the following form @xcite : @xmath45 where @xmath46 is a full intensity of incident laser radiation field .
@xmath47 the spectral intensity of the stimulated emission of photons coming from unity volume of the gas are obtained : @xmath48 ] in figures [ fig : elastic],[fig : inelastic],[fig : inten ] , shows the curves corresponding to the formulas ( 29,30,32 ) , for the values @xmath49 and @xmath50 , where @xmath51 is the density of excited atoms in the gase .
it is seen from expression ( [ eq27 ] ) at hight concentration of a excited atoms in the gas we have an intense stimulated radiation .
it should be noted that in the case of dense gases had to consider the cooperative effects , and the collision of atoms .
these phenomena , we will consider in future studies .
we consider collision of two atoms with formation of the metastable molecules ( feshbach resonance ) and the subsequent stimulated transition under the influence of the laser radiation to a lower unbound molecular electronic state with emission of photon of the laser radiation field .
this a situation can develop realized , in particular , for @xmath0 molecules due to resonance scattering of two @xmath1 atoms in states @xmath52 and @xmath53 with formation of metastable molecular state @xmath54 , with subsequent stimulated transition to the lower @xmath55 unbound molecular electronic state , with emission of photons of the laser radiation .
expressions for the cross sections of the elastic and inelastic resonance scattering and the intensity of the stimulated emission of photons coming from unity volume of the gas are obtained .
the typical atomic gas density is @xmath56 .
if the density of excited atoms is @xmath57 , then the intensity of emitted photons during the simulated transition is @xmath58 times larger . as a result
, we have a source of high radiation , which serves as an example of an excimer laser . from the expression ( 32 ) ,
the peak of the spectral intensity in the graph at @xmath59 is @xmath60
we are very grateful to professor a.v .
papoyan for fruitful discussion .
work was supported by the ministry of education and science of armenia ( mesa ) project 15t-1c066
10 h. feshbach _ ann .
_ 19 287 1962 h. eddy timmermans and paolo tommasini and mahir hussein and arthur kerman _ physics reports _ * 315 * 199 1999 r.a .
duine and h.t.c .
physics reports _ * 396 * 115 2004 immanuel bloch , and jean dalibard , and wilhelm zwerger , _ rev .
phys . _ * 80 * 885 2008 cheng chin , and rudolf grimm , and paul julienne , and eite tiesinga , _ rev .
phys . _ * 82 * 1225 2010 s. e. harris , _ phys . rev .
a _ * 66 * 010701 2002 dominik m. bauer , matthias lettner , christoph vo , gerhard rempe & stephan drr , _ nature physics _ * 5 * 339 2009 yong liu , and jian li , and gao - ren wang , and shu - lin cong , _ phys . lett . _ * 378 * 43 2014 u. fano , _ phys .
rev . _ * 124 * 1866 1961 john m. blatt and victor f. weisskopf , new york : wiley ; london : chapman & hall , pp 864 1952 a.e .
kazakov , v.p .
makarov , m.v .
fedorov , _ jetp _ * 43 * 20 1976 a.d .
gazazian and m.a .
sarkissian , _ physics letters a _ * 78 * 133 1980 p.l knight and m.a lauder and b.j dalton , _ physics reports _ * 190 * 1 1990 a.d .
gazazyan , r.g .
jetp _ * 58 * 5 p. 903
gazazyan , m.l .
ter - mikaelyan , r.g . unanyan _ laser physics _ * 5 * 1 199 1995 j. c. phillips , _ phys .
* 12 * 142 1964 akio shibatani , yutaka toyozawa , _ journal of the physical society of japan _ * 25 * n2 335 1968 a. d. gazazyan , e. a. gazazyan , _ laser physics _ 16 2006 a. d. gazazyan , e. a. gazazyan , a. g. margaryan _ euro .
j. d _ 53 243 2009 andrey e. miroshnichenko , sergey flach , yuri s. kivshar , _ rev .
mod . phys _ * 82 * 2257 2010 boris lukyanchuk , nikolay i. zheludev , stefan a. maier , naomi j. halas , peter nordlander , harald giessen & chong tow chong , _ nature materials _ * 9 * 707 2010 w. heithler , `` the quantum theory of radiation : third edition '' _ published by oxford university press , london _ 1954 | the formation of metastable molecules ( feshbach resonances ) at the collision of two atoms and subsequent stimulated transition to a lower unbound electronic molecular state , with emission of a photon of the laser radiation has been investigated .
this can develop , in particular , for @xmath0 molecules due to resonance scattering of two @xmath1 atoms .
the considered process is a basis for the creation of excimer lasers .
expressions for the cross sections of elastic and inelastic resonance scattering and the intensity of the stimulated emission of the photons have been obtained .
november 2016 |
systems of particles interacting through two - body non - integrable potentials , also called long - range interactions , abound in nature .
common examples are plasmas interacting through repulsive or attractive coulomb potential , self - gravitating systems ( globular clusters , galaxies ) involving interaction through attractive newton potential , two - dimensional turbulence , and many others . in addition , there are several model systems with non - integrable interactions which have been studied extensively in recent years , such as spins , vortices in two dimensions , etc @xcite . very often , these systems are acted upon by external stochastic forces that drive them out of equilibrium . unlike systems with short - range interactions , stochastic forces in long - range interacting systems act coherently on all particles , and not independently on each particle .
consider , e.g. , globular clusters being influenced by the gravitational potential of their galaxy , which produces a force that fluctuates along their physical trajectories . in addition , galaxies themselves feel the random potential of other surrounding galaxies , and their halos are subjected to transient and periodic perturbations , which may be due to the passing of dwarfs or to orbital decaying @xcite .
dynamics of plasmas are also strongly influenced by fluctuating electric and magnetic fields due to the ever - changing ambiance @xcite . in situations of stochastic
driving , the systems at long times often reach a nonequilibrium stationary state that violates detailed balance . in such a state , the power injected by the external random fields balances on average the dissipation , and there is a steady flux of conserved quantities through the system . study of nonequilibrium stationary states ( ness ) is an active area of research of modern day statistical mechanics .
one of the primary challenges in this field is to formulate a tractable framework to analyze nonequilibrium systems on a common footing , similar to the one due to gibbs and boltzmann that has been established for equilibrium systems @xcite .
this paper provides , to our knowledge , the first study of ness in long - range systems with statistical mechanical perspectives .
common theoretical approaches to study isolated systems with long - range interactions include the kinetic theory description of relaxation towards equilibrium . in plasma physics and astrophysics
, this approach leads to the lenard - balescu equation or to the approximate landau equation @xcite .
one of the main theoretical results of this paper is a detailed development of a generalization of this kinetic theory approach to describe nonequilibrium stationary states in systems with long - range interactions driven by external stochastic forces , valid in the limit of small external stochastic fields .
the nonequilibrium kinetic equation that we obtain describes the temporal evolution of the one - particle distribution function . in the limit of small external forcing ,
the system settles into a stationary state , in which we find the one - particle momentum distribution to be non - gaussian .
the predictions of our kinetic equation for spatially homogeneous stationary states compare very well with results of our extensive @xmath0-particle numerical simulations on a paradigmatic model of long - range interacting systems .
our numerical simulations also exhibit a nonequilibrium phase transition between homogeneous and inhomogeneous states .
close to the phase transition , we demonstrate the occurrence of bistability between these two types of states , with a mean residence time that diverges as an exponential in the inverse of the strength of the external forcing , in the limit of low values of such forcing .
similar bistable behavior has recently been observed in two - dimensional turbulence with stochastic forcing @xcite .
we believe that such phase transitions are essential phenomena for geophysical flows and climate , for which the two - dimensional euler equations are a simplified paradigmatic model .
there exists a very strong analogy between the two - dimensional euler equations and the vlasov equation relevant for leading order dynamics of the model we discuss in this paper @xcite .
one of the motivations of the present work is to be able to study analogous phenomena in a setup for which the theory can be more easily worked out . in a recent letter , we reported on some of the above findings , in one of the first studies of non - equilibrium stationary states in systems with non - integrable potentials and driven by external stochastic fields @xcite . the aim of this paper is to present a detailed derivation of the results given in ref .
@xcite , as well as to report on additional empirical results , more specifically non - equilibrium phase transitions .
we note that ref .
@xcite presents a computation of the same kinetic equation as the one we describe in this work and in @xcite .
nevertheless , the result is different .
the structure of the paper is as follows . in the following section ,
we define the dynamics we are going to consider of a long - range interacting system driven by external stochastic forces .
we also discuss the paradigmatic example of the hamiltonian mean - field ( hmf ) model . in section [ methods ] , we discuss the methods we adopt to analyze the dynamics .
in particular , we give a detailed derivation of the kinetic theory to study spatially homogeneous stationary states of the dynamics .
we describe the numerical simulation scheme that we employ to study the dynamics , specifically , to check the predictions of our kinetic theory .
this is followed by a discussion in section [ results1 ] of the results obtained from the kinetic theory , and their comparison with numerical simulation results for spatially homogeneous stationary states . in section [ results2 ] , we discuss the results of numerical simulations of spatially inhomogeneous stationary states
. we report on the very interesting bistable behavior in which the system in the course of its temporal evolution switches back and forth between homogeneous and inhomogeneous states , with a mean residence time that we show to be diverging as an exponential in the inverse of the strength of the external stochastic forcing , in the limit of low values of such forcing .
we close the paper with concluding remarks .
some of the technical details of our computation are collected in the four appendices .
consider a system of @xmath0 particles interacting through a long - range pair potential , and described by the hamiltonian @xmath1 here , @xmath2 and @xmath3 are , respectively , the coordinate and the momentum of the @xmath4-th particle , and @xmath5 is the two - body interaction potential .
we take the particles to be of unit mass . in this paper , for simplicity , we regard @xmath2 s as scalar periodic variables of period @xmath6 ; generalization to @xmath7 , with @xmath8 or @xmath9 , is straightforward . in plasma physics , the typical number of particles interacting with one particle is given by the coupling parameter @xmath10 , where @xmath11 is the number density , and @xmath12 is the debye length .
it is then usual to rescale time such that the inverse of @xmath13 multiplies the interaction term @xcite . in self - gravitating systems ,
the dynamics is dominated by collective effects , so that it is natural to rescale time in such a way that the parameter @xmath14 multiplies the interaction potential @xcite .
these facts explain the rescaling of the potential energy by @xmath14 in eq .
( [ hamiltonian ] ) , called the kac scaling in systems with long - range interactions @xcite .
we emphasize that no generality is lost in adopting the kac prescription .
we perturb the system ( [ hamiltonian ] ) by a statistically homogeneous gaussian stochastic field @xmath15 with zero mean , and variance given by @xmath16 the resulting equations of motion for the @xmath4-th particle are @xmath17 the property that the gaussian fields @xmath18 are statistically homogeneous , i.e. , the correlation function @xmath19 depends solely on @xmath20 , is consistent with any perturbation that respects space homogeneity .
such a property is necessary for the discussions later in the paper on the kinetic theory approach to describe spatially homogeneous stationary states of the dynamics ( [ equations_of_motion ] ) .
note that @xmath21 is the correlation , so that it is a positive - definite function @xcite , and its fourier components are positive : @xmath22 we find it convenient to use the equivalent fourier representation of the gaussian field @xmath15 as follows : @xmath23 , \label{noise - form}\end{aligned}\ ] ] where @xmath24 , @xmath25 and @xmath26 are independent scalar gaussian white noises satisfying @xmath27 for stochastic dynamics of the type eq .
( [ equations_of_motion ] ) , general mathematical results allow to prove that the dynamics is ergodic ( see , for instance , @xcite ) . using the it formula @xcite to compute the time derivative of the energy density @xmath28 , and averaging over noise realizations give @xmath29 where @xmath30 is the kinetic energy density . on integration , we get for homogeneous states for which @xmath31 that @xmath32 the average kinetic energy density in the stationary state is thus @xmath33 .
we define the kinetic temperature of the system as @xmath34 as a result , we have @xmath35 let us note that in the dynamics ( [ equations_of_motion ] ) , fluctuations of intensive observables due to stochastic forcing are of order @xmath36 , while those due to finite - size effects are of order @xmath37 . moreover , the typical timescale associated with the effect of stochastic forces is @xmath38 , as is evident from eq .
( [ kinetic - energy-1 ] ) , while the one associated with relaxation to equilibrium due to finite - size effects is of order @xmath0 , see @xcite .
our theoretical analysis to study the dynamics ( [ equations_of_motion ] ) by means of kinetic theory is valid for any general two - particle interaction potential @xmath5 .
however , in order to perform simple numerical simulations with which we may check the predictions of the kinetic theory , we specifically make the choice @xmath39 , that defines the stochastically - forced attractive hamiltonian mean - field ( hmf ) model , as detailed below . the hamiltonian mean - field ( hmf ) model is a paradigmatic model to study long - range interacting systems .
the model describes particles moving on a circle under deterministic hamiltonian dynamics , and interacting through the interparticle potential @xmath39 @xcite .
it displays many features of generic long - range interacting systems , e.g. , existence of quasistationary states @xcite . in equilibrium
, the model has a second - order phase transition from a high - energy spatially homogeneous phase to a low - energy inhomogeneous phase at the energy density @xmath40 , corresponding to the critical temperature @xmath41 . in a system of @xmath0 particles , the degree of spatial inhomogeneity at time
@xmath42 is measured by the magnetization variable @xmath43 , defined as @xmath44 in the thermodynamic limit @xmath45 , the magnetization in the steady state decreases continuously as a function of energy from one to zero at the transition energy @xmath46 , and remains zero at all higher energies .
when forced by the stochastic forces @xmath18 resulting in the dynamics ( [ equations_of_motion ] ) , we call the corresponding model the stochastically - forced hmf model .
we note for later purpose that the fourier transform of the hmf interparticle potential is , for @xmath47 , @xmath48/2 $ ] , where @xmath49 is the kronecker delta .
here , we develop a suitable kinetic theory description to study the dynamics ( [ equations_of_motion ] ) in the joint limit @xmath45 and @xmath50 . while the first limit is physically motivated on grounds that most long - range systems indeed contain a large number of particles , the second one allows us to study stationary states for small external forcing . moreover , for small @xmath51 , we will be able to develop a complete kinetic theory for the dynamics . for simplicity
, we discuss here the continuum limit @xmath52 , when stochastic effects are predominant with respect to finite - size effects .
the generalization of the following discussion to the cases @xmath53 of order one and @xmath54 is straightforward , as pointed out at the end of this subsection . for the development of the kinetic theory
, we assume the system to be spatially homogeneous ; a possible generalization to the non - homogeneous case will be discussed in the conclusions of the paper . as a starting point to develop the theory
, we consider the fokker - planck equation associated with the equations of motion ( [ equations_of_motion ] ) .
this equation describes the evolution of the @xmath0-particle distribution function @xmath55 , which is the probability density ( after averaging over noise realizations ) to observe the system with coordinates and momenta around the values @xmath56 at time @xmath42 .
this equation can be derived by standard methods @xcite ; we have @xmath57+\frac{1}{2n}\sum_{i , j=1}^{n}\frac{\partial v(q_{i}-q_{j})}{\partial q_{i}}\left[\frac{\partial}{\partial p_{i}}-\frac{\partial}{\partial p_{j}}\right]f_{n } \nonumber \\ & & + \frac{\alpha}{2}\sum_{i , j=1}^{n}c(|q_{i}-q_{j}|)\frac{\partial^{2}f_{n}}{\partial p_{i}\partial p_{j}}. \label{eq : n - particle - fokker - planck}\end{aligned}\ ] ] in [ detailed - balance - proof ] , by applying the so - called potential conditions @xcite for the above fokker - planck equation , we prove that a necessary and sufficient condition for the stochastic process ( [ equations_of_motion ] ) to verify detailed balance is that the gaussian noise is white in space , that is , @xmath58 for all @xmath59 .
this condition is not satisfied for a generic correlation function @xmath21 , in which case , the steady states of the dynamics are true nonequilibrium ones , characterized by non - vanishing probability currents in configuration space , and a balance between external forces and dissipation .
similar to the liouville equation for hamiltonian systems , the @xmath0-particle fokker - planck equation ( [ eq : n - particle - fokker - planck ] ) is a very detailed description of the system .
using kinetic theory , we want to describe the evolution of the one - particle distribution function @xmath60 where we have used the notation @xmath61 .
we note that with this definition , the normalization is @xmath62 .
substituting in the fokker - planck equation ( [ eq : n - particle - fokker - planck ] ) the reduced distribution functions @xmath63 and using standard techniques @xcite , we get a hierarchy of equations , similar to those of the bogoliubov - born - green - kirkwood - yvon ( bbgky ) hierarchy , as follows : @xmath64\nonumber \\ & & -\frac{\alpha}{2}\sum_{i , j=1}^{s}c(|q_{i}-q_{j}|)\frac{\partial^{2}f_{s}}{\partial p_{i}\partial p_{j}}=\sum_{i=1}^{s}\int dz_{s+1}\frac{\partial}{\partial q_{i}}v(q_{i}-q_{s+1})\frac{\partial f_{s+1}}{\partial p_{i}}\end{aligned}\ ] ] for @xmath65 . in this paper , we use both the notations @xmath66 and @xmath67 to denote the derivative of a function @xmath68 . with a slight abuse of the standard vocabulary , we will refer to eq .
( [ bbgky - stochastic ] ) as the bbgky hierarchy equation .
now , as is usual in kinetic theory , we split the reduced distribution functions into connected and non - connected parts , e.g. , @xmath69 and similarly , for other @xmath70 s with @xmath71 . in [ bbgkyclosure ] , we show that the connected part @xmath72 of the two - particle correlation is of order @xmath51 , so that we may write @xmath73 where @xmath74 is of order unity ; more generally , the connected part of the @xmath59-particle correlation is of higher order , with respect to @xmath51 , in the small parameters @xmath51 and @xmath14 .
then , to close the bbgky hierarchy , we neglect the effect of the connected part of the three - particle correlation on the evolution of the two - particle correlation function .
this scheme is justified at leading order in the small parameter @xmath51 , and is the simplest self - consistent closure scheme for the hierarchy while taking into account the effects of the stochastic forcing . with our assumption that the system is homogeneous , i.e. , @xmath75 depends on @xmath76 , and @xmath74 depends on @xmath77 , @xmath78 and @xmath79 only , the first two equations of the hierarchy are then @xmath80-\frac{\alpha}{2}c(0)\frac{\partial^{2}f}{\partial p^{2}}=\alpha\frac{\partial}{\partial p}\int{\rm d}q_{1}{\rm d}p_{1}v'(q_{1})g(q_{1},p , p_{1},t),\ ] ] and @xmath81 where @xmath82 and @xmath83 are the vlasov operators linearized about the one - particle distribution @xmath75 , and acting , respectively , on the first pair @xmath84 and on the second pair @xmath85 of the function @xmath86 .
explicitly , for a function @xmath68 of @xmath87 , the expression for the linear vlasov operator @xmath88 is @xmath89 so that we have for @xmath90 , @xmath91 @xmath92 is obtained from eq .
( [ eq : linear - vlasov - operator ] ) by exchanging the subscripts @xmath93 and @xmath94 . to obtain from eqs .
( [ eq : stochastic - bbgky-1-eq ] ) and ( [ eq : stochastic - bbgky-2-eq ] ) a single kinetic equation for the distribution function @xmath75 , we have to solve eq .
( [ eq : stochastic - bbgky-2-eq ] ) for @xmath74 as a function of @xmath75 and plug the result into the right hand side of eq .
( [ eq : stochastic - bbgky-1-eq ] ) . because the two equations are coupled
, this program is not achievable without making further simplifying assumption .
nevertheless , we readily see from these equations that the two - particle correlation @xmath74 evolves over a timescale of order one , whereas the one - particle distribution function @xmath95 evolves over a timescale of order @xmath38 .
we may then use this timescale separation and compute the long - time limit of @xmath74 from eq .
( [ eq : stochastic - bbgky-2-eq ] ) by assuming @xmath75 to be steady in time ; this is the equivalent of the bogoliubov s hypothesis in the kinetic theory for isolated systems with long - range interactions .
note that for this timescale separation to be valid , we must also suppose that the one - particle distribution function @xmath95 is a stable solution of the vlasov equation at all times .
indeed , if this is not the case , it can be shown that @xmath74 diverges in the limit @xmath96 @xcite .
the physical content of this hypothesis is that the system slowly evolves from the initial condition through a sequence of quasistationary states to the final stationary state . because we assume the system to be homogeneous in space , it is useful to fourier transform eqs .
( [ eq : stochastic - bbgky-1-eq ] ) and ( [ eq : stochastic - bbgky-2-eq ] ) with respect to the spatial variable ; we get @xmath97-\frac{\alpha}{2}c(0)\frac{\partial^{2}f}{\partial p^{2}}=-2\pi i\alpha\sum_{k=-\infty}^{\infty}kv_{k}\frac{\partial}{\partial p}\int\mathrm{d}p'\ , g_{k}(p , p',t),\ ] ] and @xmath98 where @xmath99 is the fourier transform of @xmath100 with respect to the spatial variable , and @xmath101 is the @xmath59-th fourier coefficient of the pair potential @xmath5 .
the explicit expression for the @xmath59-th fourier component of the linear vlasov operator @xmath102 acting on a function @xmath103 is @xmath104 one has analogous expressions for @xmath105 and for @xmath106 .
we readily see that @xmath107 . from the right hand side of eq .
( [ eq : stochastic - bbgky-1-eq - fourier - transform ] ) , we see that to obtain a single kinetic equation , we need only the fourier transform @xmath108 , more specifically , its integral with respect to the second momentum variable @xmath109 .
actually , it can be shown that @xmath108 does not have a well - defined time - asymptotic ( it converges only in the sense of distribution ) , while its integral with respect to @xmath109 does have ; this is connected to the mechanism of landau damping @xcite .
the structure of eqs .
( [ eq : stochastic - bbgky-1-eq ] ) and ( [ eq : stochastic - bbgky-2-eq ] ) , or , equivalently , of eqs .
( [ eq : stochastic - bbgky-1-eq - fourier - transform ] ) and ( [ eq : stochastic - bbgky-2-eq - fourier - transform ] ) is very familiar in kinetic theories ; we refer the reader to @xcite for a general discussion .
equation ( [ eq : stochastic - bbgky-2-eq ] ) , or , equivalently , eq . ( [ eq : stochastic - bbgky-2-eq - fourier - transform ] ) , is called the lyapunov equation for the two - point correlation of a stochastic variable described by an ornstein uhlenbeck process .
however , there is a difference from the standard finite - dimensional case @xcite in that in our case , @xmath110 is a linear infinite - dimensional operator acting on a functional space , instead of being a finite - dimensional one , i.e. , a matrix .
this makes it non - trivial to compute the long - time asymptotic of the right hand side of eq .
( [ eq : stochastic - bbgky-1-eq ] ) , where @xmath74 is the solution of eq .
( [ eq : stochastic - bbgky-2-eq ] ) with @xmath75 steady in time .
a possible way to achieve this goal is to follow the derivation of the lenard
balescu equation from the bbgky hierarchy , as may be found in the appendix a of ref .
for explicit technical details , see @xcite , in which the method to solve the lyapunov equation in a general manner is discussed , and , subsequently , applied to the derivation of kinetic theories for long - range interacting systems and two - dimensional turbulence models . in the present case , the linear transform of the stationary solution of the lyapunov equation , eq .
( [ eq : stochastic - bbgky-2-eq - fourier - transform ] ) , which is needed to compute the right hand side of eq .
( [ eq : stochastic - bbgky-1-eq - fourier - transform ] ) , can be written ( see ref .
@xcite ) in the frequency space as @xmath111(p , p_{1})&\equiv & \lim_{t\to \infty}\int dp_{1}\ , g_{k}(p , p_{1},t ) \\ & = & \frac{1}{\pi}\int_{\gamma}\mathrm{d}\omega\,\bigl(r_{f , k}(\omega)b\bigr)(p)\,\int \mathrm{d}p'\,\bigl(r_{f ,- k}(-\omega)b^{*}\bigr)(p'),\end{aligned}\ ] ] where @xmath13 is a contour which passes above all the poles of @xmath112 , and @xmath113 is the resolvent operator , defined as @xmath114 while @xmath115 .
the discussion of the explicit form of the resolvent operator is a standard topic in plasma theory , and involves the phenomenon of landau damping ; we refer the reader to classical references for this result , for example , @xcite .
its action on a function @xmath68 , defined for @xmath116 such that @xmath117 , is @xmath118,\ ] ] where @xmath119 is the dielectric function , which for @xmath117 is given by @xmath120,\ ] ] and by its analytic continuation for @xmath116 when @xmath121 . both the resolvent operator and the dielectric function are defined for @xmath122 by their analytic continuation , which will still be denoted by the same symbols .
now , inserting eq .
( [ resolvent_operator ] ) into eq .
( [ solution_lyapunov_equation ] ) , and with some calculations whose details will be reported in @xcite , we get the kinetic equation @xmath123\frac{\partial f}{\partial p}\right]=0 , \label{kinetic_equation}\ ] ] where @xmath124(p)&=&\frac{1}{2}c(0)\nonumber \\ & & + 2\pi\sum_{k=1}^{\infty}v_{k}c_{k}\int^{*}{\rm d}p_{1}\left[\frac{1}{|\epsilon(k , kp)|^{2}}+\frac{1}{|\epsilon(k , kp_{1})|^{2}}\right]\frac{1}{p_{1}-p}f'(p_{1},t).\end{aligned}\ ] ] we recall that @xmath101 is the @xmath59-th fourier coefficient of the pair potential @xmath5 , the quantity @xmath125 is defined in eq . ( [ ckdefinition ] ) , @xmath126 is the dielectric function defined in eq .
( [ eq : stochastic - dielectric - function ] ) , and @xmath127 indicates the cauchy integral or principal value . the kinetic equation ( [ eq : stochastic - kinetic - equation ] ) is the central result of the kinetic theory developed in this paper .
it has the form of a non - linear fokker - planck equation @xcite because the diffusion coefficient @xmath128(p)$ ] is itself a functional of the one - particle distribution function @xmath75 .
the linear part of the diffusion coefficient @xmath129 is the mean - field effect of the stochastic forces , whereas the effect of two - particle correlation is encoded in the non - linear part . in the next section ,
we describe how we use this kinetic equation to get information about the nonequilibrium stationary states of the dynamics . in the foregoing
, we discussed the kinetic theory in the limit @xmath130 .
the extension to the general case is straightforward : because of the linearity of the equations of the hierarchy ( [ eq : stochastic - bbgky-1-eq ] ) and ( [ eq : stochastic - bbgky-2-eq ] ) , the finite-@xmath0 and stochastic effects give independent contributions .
the kinetic equation at leading order of both stochastic and finite - size effects is @xmath131+q_{n}[f],\ ] ] where @xmath132 is the operator described in eq .
( [ eq : stochastic - kinetic - equation ] ) , and @xmath133 ( of order @xmath14 ) is the lenard - balescu operator .
for instance , in the case @xmath134 and in dimensions greater than one , the operator @xmath133 is responsible for the relaxation to boltzmann equilibrium after a timescale of order @xmath0 , whereas the smaller effect of @xmath132 selects the actual temperature after a longer timescale of order @xmath38 . at kinetic temperature @xmath135 , and with @xmath136 , where @xmath137 .
the values of the integration step size @xmath138 used are marked in the figure .
the data are obtained by using the integration algorithm described in section [ numerical - scheme ] . that the magnetization plots collapse onto one curve shows the stability of our algorithm with respect to variation in @xmath138 .
we have checked that the final value of the magnetization matches with the prediction from equilibrium statistical mechanics.,width=340 ] here we describe how we may simulate the dynamics ( [ equations_of_motion ] ) by means of a numerical integration scheme . to simulate the dynamics over a given time interval @xmath139 $ ] , choose a time step size @xmath138 , and set @xmath140 as the @xmath11-th time step of the dynamics .
here , @xmath141 , where @xmath142 . in our numerical scheme , at every time step , we first discard the effect of the noise and employ a fourth - order symplectic algorithm to integrate the deterministic hamiltonian part of the dynamics @xcite .
subsequently , we add the effect of noise and implement an euler - like first - order algorithm to update the dynamical variables .
, in the sense that one obtains different magnetization profiles as a function of time @xmath143 .
the situation gets worse for small @xmath51 , when one needs to use very small @xmath138 to obtain consistent results .
therefore , for faster and efficient simulation , we adopted the mixed " scheme described in the text . ]
specifically , one step of the scheme from @xmath144 to @xmath145 involves the following updates of the dynamical variables for @xmath146 : for the symplectic part , we have , for @xmath147 , @xmath148 , \nonumber \\ \label{formalintegration3 } \\ & & q_i\big(t_{n}+\frac{m\delta t}{4}\big)=q_i\big(t_n+\frac{(m-1)\delta t}{4}\big)+a(m)\delta t ~p_i\big(t_n+\frac{m\delta t}{4}\big),\nonumber \end{aligned}\ ] ] where the constants @xmath149 s and @xmath150 s are given in ref .
@xcite . at the end of the update ( [ formalintegration3 ] ) , we have the set @xmath151 .
next , one includes the effect of the stochastic noise by leaving @xmath152 s unchanged , but by updating @xmath153 s as @xmath154+\sqrt{\alpha}\big[\sqrt{c_0}\delta x^{(0)}(t_{n+1})\nonumber \\ & & + \sum_{k=1}^{n_r } \sqrt{2c_k } \big\{\delta x^{(k)}(t_{n+1})\cos\big(kq_i(t_{n+1})\big)+\delta y^{(k)}(t_{n+1})\sin\big(kq_i(t_{n+1})\big)\big\}\big ] .
\label{formalintegration4}\end{aligned}\ ] ] here @xmath155 and @xmath156 are gaussian distributed random numbers with zero mean and unit variance .
the outcome of implementing this mixed scheme for the stochastically - forced hmf model is shown in fig .
[ algo - check - fig ] , where one may observe consistent results with respect to change of @xmath138 over a wide range of values . in numerical simulations reported later in the paper , we exclusively used this mixed scheme to simulate the dynamics ( [ equations_of_motion ] ) .
, and ( b ) @xmath157 as a function of @xmath158 , at kinetic temperature @xmath159 , with modes @xmath160 excited with amplitudes satisfying @xmath161 , where @xmath137 . the data for different @xmath0 and @xmath51 values are obtained from numerical simulations of the stochastically - forced hmf model with @xmath162 , and involve averaging over @xmath163 histories for @xmath164 and @xmath165 histories for @xmath166 .
the data collapse implies that @xmath51 is the timescale of relaxation to the stationary state .
the inset shows the data without time rescaling by @xmath51 .
similar plots for different parameter values were reported in ref .
@xcite.,title="fig:",width=302 ] , and ( b ) @xmath157 as a function of @xmath158 , at kinetic temperature @xmath159 , with modes @xmath160 excited with amplitudes satisfying @xmath161 , where @xmath137 . the data for different @xmath0 and @xmath51 values are obtained from numerical simulations of the stochastically - forced hmf model with @xmath162 , and involve averaging over @xmath163 histories for @xmath164 and @xmath165 histories for @xmath166 .
the data collapse implies that @xmath51 is the timescale of relaxation to the stationary state .
the inset shows the data without time rescaling by @xmath51 .
similar plots for different parameter values were reported in ref .
@xcite.,title="fig:",width=302 ] , on ( a ) linear , and ( b ) semi - log scales , for @xmath167 , and @xmath168 at kinetic temperature @xmath159 .
the plots correspond to modes @xmath93 and @xmath94 excited with amplitudes satisfying @xmath161 , where @xmath137 .
the data denoted by crosses and squares are results of @xmath0-body simulations of the stochastically - forced hmf model with @xmath169 and @xmath170 independent realizations of the dynamics , while the red continuous lines refer to the theoretical prediction from the kinetic theory . for comparison
, the black broken line shows the gaussian distribution with the same kinetic energy ( stationary state of the stochastically - forced hmf model at @xmath159 , @xmath171).,title="fig:",width=302 ] , on ( a ) linear , and ( b ) semi - log scales , for @xmath167 , and @xmath168 at kinetic temperature @xmath159 .
the plots correspond to modes @xmath93 and @xmath94 excited with amplitudes satisfying @xmath161 , where @xmath137 .
the data denoted by crosses and squares are results of @xmath0-body simulations of the stochastically - forced hmf model with @xmath169 and @xmath170 independent realizations of the dynamics , while the red continuous lines refer to the theoretical prediction from the kinetic theory . for comparison
, the black broken line shows the gaussian distribution with the same kinetic energy ( stationary state of the stochastically - forced hmf model at @xmath159 , @xmath171).,title="fig:",width=302 ] ( p)$ ] for the stationary momentum distribution @xmath172 at kinetic temperature @xmath159 , with @xmath173 , and either ( i ) @xmath174 , or , ( ii ) @xmath175.,title="fig:",width=302 ] + we now focus on how to obtain from the kinetic equation ( [ eq : stochastic - kinetic - equation ] ) predictions for the nonequilibrium stationary states of the system . according to eq .
( [ eq : stochastic - kinetic - equation ] ) , @xmath38 is only a timescale ; thus , at leading order in @xmath51 and except for a time rescaling , the parameter @xmath51 does not affect the time evolution of the system .
this statement holds also beyond the leading order for what concerns the evolution of the kinetic energy ; its evolution may be obtained directly from the equations of motion ( [ equations_of_motion ] ) , as discussed in section [ model ] , and can also be obtained from the kinetic equation ( [ eq : stochastic - kinetic - equation ] ) , as detailed in [ kinetic - energy - evolution ] . for the evolution of other observables , there will be corrections at higher orders in @xmath51 . as previously discussed , the dynamics of the system does not respect detailed balance if the forcing is not white in space . at the level of the kinetic equation , by inspecting the definition of the diffusion coefficient ,
( [ diffusion_coefficient ] ) , we see that the effect of correlations induced by the stochastic forces is modulated by the fourier component @xmath176 of the interparticle potential .
then , taking the forcing spectral amplitudes @xmath177 different from zero if and only if @xmath178 , the non - linear part of the diffusion coefficient vanishes .
on the other hand , taking @xmath179 for the modes for which @xmath180 leads to a diffusion constant which has a non - vanishing non - linear part . to be concrete ,
let us discuss these two scenarios in the context of the stochastically - forced hmf model .
since the fourier transform of the hmf interparticle potential is , for @xmath47 , @xmath48/2 $ ] , it follows that only the stochastic force mode with wave number @xmath181 contributes to the non - linear part of the diffusion coefficient ; all the other stochastic force modes result in only a mean - field contribution through the term @xmath182 .
thus , for the case @xmath183 , the two relevant parameters that dictate the evolution of the stochastically - forced hmf model by the kinetic equation ( [ eq : stochastic - kinetic - equation ] ) with a non - linear diffusion coefficient are @xmath182 and @xmath184 . from eq .
( [ kinetic - temp - defn ] ) , since @xmath182 is related to the kinetic temperature @xmath185 , we take @xmath185 and @xmath184 to be the two relevant parameters . from eq .
( [ kinetic - energy - steady - state ] ) , we know that @xmath186 equals the kinetic energy in the final stationary state . also , eq . ( [ eq : g ] ) implies that @xmath187 .
if however @xmath188 , then , at leading order in @xmath51 , the dynamics of the system is described by a linear fokker - planck equation ; this equation is the same as the one which describes the hmf system when coupled to a langevin thermostat , studied in @xcite .
this means that for this particular choice of the parameters , the detailed balance is broken for the dynamics , but this feature can not be seen in the kinetic theory , being an effect at a higher order in @xmath51 . in this case , the homogeneous stationary states of the kinetic equation have gaussian momentum distribution @xmath172 .
as has been studied thoroughly in the context of canonical equilibrium of the hmf model , these states are stable for kinetic energies greater than @xmath189 , i.e. , for @xmath190 . except for the special case of @xmath191 , the stationary velocity distribution of the kinetic equation ( [ eq : stochastic - kinetic - equation ] )
is in general not gaussian .
this can be seen semi - analytically by observing that the gaussian distribution function @xmath192 with @xmath193 chosen such that the value of the kinetic energy is the one selected by @xmath185 , solves the linear fokker - planck equation with the diffusion coefficient given by @xmath194 to prove that the gaussian distribution function is not a stationary solution of eq .
( [ eq : stochastic - kinetic - equation ] ) , we have to prove that the contribution to @xmath195 from the non - linear part of @xmath128(p)$ ] in eq .
( [ diffusion_coefficient ] ) does not vanish .
this result can be proven with an asymptotic expansion @xcite for large momenta of the integrals which appear in the diffusion coefficient .
we report the straightforward computation in [ nongaussian ] . from the same analysis
, one can deduce that , even though the distribution function is not gaussian , its tails are gaussian . on the basis of the above discussions
, we expect that for values of @xmath185 and @xmath184 such that @xmath196 and @xmath197 , the stationary states will be close to homogeneous states with gaussian momentum . in order to locate the actual stationary states of the kinetic equation
, we have devised a simple numerical scheme , based on the observation that a linear fokker - planck equation whose diffusion coefficient @xmath198 is strictly positive admits a unique stationary state @xmath199.\ ] ] for a given distribution @xmath200 , we compute the diffusion coefficient @xmath201 through eq .
( [ diffusion_coefficient ] ) , and then @xmath202 using @xmath203 and eq . ( [ stationary - fokkerplanck ] ) .
this procedure defines an iterative scheme . whenever convergent , this scheme leads to a stationary state of eq .
( [ eq : stochastic - kinetic - equation ] ) .
each iteration involves integrations , so that we expect the method to be robust enough when starting not too far from an actual stationary state .
however , we have no detailed mathematical analysis yet . implementing this iterative scheme
, we observed that the distribution @xmath204 to which the scheme converges is independent of the initial distribution @xmath205 .
moreover , the convergence time is exponential in the number of steps @xmath11 whenever @xmath185 is not too close to loss of stability of @xmath204 with respect to the linear vlasov dynamics ; in practice , we are able to get reliable results for @xmath206 . in order to check the theoretical predictions discussed above , we performed numerical simulations of the stochastically forced hmf model .
figure [ p2p4 ] shows the evolution of the kinetic energy and @xmath207 , where they have been compared with our theoretical predictions . in the case of @xmath208 ,
we have compared the long - time asymptotic value with the kinetic theory prediction for the stationary state , computed numerically by using the iterative solution for the stationary distribution .
the figure illustrates a very good agreement between the theory and simulations . for a more accurate check of the agreement ,
we have obtained the stationary momentum distribution from both @xmath0-body simulations and the numerical iterative scheme discussed above .
a comparison between the two , shown in fig . [ fp ] , both on linear and semi - log scales , shows a very good agreement between theory and simulations . in this figure
, we also show the gaussian distribution with the same kinetic energy , to illustrate the point that the stationary momentum distribution of the system is far from being gaussian .
the diffusion coefficient @xmath128(p)$ ] is shown in figure [ dp ] . in passing ,
let us remark that , with an iterative scheme analogous to the one described above , one could have also obtained the full time evolution @xmath95 that obeys the kinetic equation ( [ eq : stochastic - kinetic - equation ] ) .
however , we will not address this point here .
we also note that while a linear fokker - planck equation with non - degenerate diffusion coefficient can be proven to converge to a unique stationary distribution @xcite , this is not true in general for non - linear fokker - planck equations like eq .
( [ eq : stochastic - kinetic - equation ] ) .
we expect that if the dynamics is not too far from detailed balance , the kinetic equation will have a unique stationary state .
far from equilibrium , the kinetic equation could lead to very interesting dynamical phenomena , like bistability , limit cycle or more complex behaviors .
the main issue is then the analysis of the evolution of the kinetic equation .
although some methods to study this type of equation exist @xcite , we have only the numerical iterative scheme described above to provide some preliminary answers .
a more rigorous mathematical analysis is left for future studies .
until now , we have considered homogeneous stationary states of the dynamics ( [ dynamics ] ) , and have discussed a kinetic theory to analyze them .
although our theory can in principle be extended to include inhomogeneous stationary states , its actual implementation to get , e.g. , the single - particle distribution , would require more involved computations than the one we encountered for homogeneous states . in order to get preliminary answers , we have performed extensive numerical simulations of the dynamics in the context of the stochastically - forced hmf model .
our specific interest is to know about how the magnetization behaves as the kinetic temperature is reduced from high values . ; the different plots correspond to different number of modes excited in the spectrum with amplitudes satisfying @xmath209 , where @xmath137 , the index @xmath210 denotes the mode number , while @xmath211 is the total number of excited modes with @xmath212 . in all cases , the modes excited were chosen to have equal amplitudes , with @xmath213 , while the tuning rate for @xmath185 is @xmath214 . it may be noted that forcing equally a large number of modes ( @xmath215 ) reproduces the equilibrium magnetization profile as illustrated by the match with the analytical equilibrium solution in the panel ( a ) . in panel
( b ) , the first - order nonequilibrium phase transition is marked by the vertical dashed line . in panel
( c ) , besides the first - order transition , we also show the dynamical transition to the collapsed state by the vertical dashed dotted line . in panel ( d ) , the nonequilibrium phase transition and the dynamical transition almost coincide , and we just show the latter one by the vertical dashed dotted line.,title="fig:",width=302 ] ; the different plots correspond to different number of modes excited in the spectrum with amplitudes satisfying @xmath209 , where @xmath137 , the index @xmath210 denotes the mode number , while @xmath211 is the total number of excited modes with @xmath212 . in all cases , the modes excited were chosen to have equal amplitudes , with @xmath213 , while the tuning rate for @xmath185 is @xmath214 .
it may be noted that forcing equally a large number of modes ( @xmath215 ) reproduces the equilibrium magnetization profile as illustrated by the match with the analytical equilibrium solution in the panel ( a ) . in panel
( b ) , the first - order nonequilibrium phase transition is marked by the vertical dashed line . in panel
( c ) , besides the first - order transition , we also show the dynamical transition to the collapsed state by the vertical dashed dotted line . in panel ( d ) , the nonequilibrium phase transition and the dynamical transition almost coincide , and we just show the latter one by the vertical dashed dotted line.,title="fig:",width=302 ] ; the different plots correspond to different number of modes excited in the spectrum with amplitudes satisfying @xmath209 , where @xmath137 , the index @xmath210 denotes the mode number , while @xmath211 is the total number of excited modes with @xmath212 . in all cases , the modes excited were chosen to have equal amplitudes , with @xmath213 , while the tuning rate for @xmath185 is @xmath214 .
it may be noted that forcing equally a large number of modes ( @xmath215 ) reproduces the equilibrium magnetization profile as illustrated by the match with the analytical equilibrium solution in the panel ( a ) . in panel
( b ) , the first - order nonequilibrium phase transition is marked by the vertical dashed line . in panel ( c ) , besides the first - order transition , we also show the dynamical transition to the collapsed state by the vertical dashed dotted line . in panel ( d ) , the nonequilibrium phase transition and the dynamical transition almost coincide , and we just show the latter one by the vertical dashed dotted line.,title="fig:",width=302 ] ; the different plots correspond to different number of modes excited in the spectrum with amplitudes satisfying @xmath209 , where @xmath137 , the index @xmath210 denotes the mode number , while @xmath211 is the total number of excited modes with @xmath212 . in all cases , the modes excited were chosen to have equal amplitudes , with @xmath213 , while the tuning rate for @xmath185 is @xmath214 . it may be noted that forcing equally a large number of modes ( @xmath215 ) reproduces the equilibrium magnetization profile as illustrated by the match with the analytical equilibrium solution in the panel ( a ) . in panel
( b ) , the first - order nonequilibrium phase transition is marked by the vertical dashed line . in panel
( c ) , besides the first - order transition , we also show the dynamical transition to the collapsed state by the vertical dashed dotted line . in panel
( d ) , the nonequilibrium phase transition and the dynamical transition almost coincide , and we just show the latter one by the vertical dashed dotted line.,title="fig:",width=302 ] in the case when the stochastic forcing respects detailed balance ( i.e. , when the noise spectrum is flat and all modes are excited ) , the stochastically - forced hmf model reduces to the brownian mean - field ( bmf ) model studied previously @xcite . here , we know that the system settles into an equilibrium state in which it exhibits a second - order phase transition at the kinetic temperature @xmath216 : on increasing @xmath185 from low values , the magnetization decreases continuously to zero at @xmath217 and remains zero at higher temperatures . in the following ,
we excite only a limited number of modes @xmath211 , but the amplitudes of all excited modes are equal ( @xmath177 equals @xmath218 for all @xmath219 , and is zero otherwise , where the constant @xmath218 is related to the temperature ) .
figure [ inhom-1](a ) shows that with @xmath220 , one reproduces very well the equilibrium profile of the magnetization as a function of temperature . on reducing the value of @xmath211 ,
the system is driven more and more out of equilibrium . indeed , fig .
[ inhom-1](b ) shows that with @xmath221 , the magnetization profile changes ; in particular , it develops a discontinuity around a temperature @xmath222 , reminiscent of a first - order phase transition .
the transition temperature is denoted by the vertical dashed line . with @xmath223 , fig .
[ inhom-1](c ) shows that the discontinuity gets more pronounced , and @xmath224 is now shifted to a higher value ( denoted again by the vertical dashed line ) .
a new feature appears in this plot , namely , at a temperature @xmath225 , the magnetization attains the maximal value of unity , which it retains for all lower temperatures .
this value of unity corresponds to a state in which the particles are very close to one another on the circle , thus defining a collapsed " state .
we found that this state , as well as the transition to it , persist on changing the system size @xmath0 .
now , it is known that trajectories of ensembles of dissipative dynamical systems forced by the same realization of a stochastic noise converge to a single one @xcite .
these attracting trajectories are referred to as the ones due to the so - called stochastic attractor .
although we did not perform a detailed characterization of the collapse in our model , we believe that the phenomenon is related to stochastic attractors . coming back to fig .
[ inhom-1](c ) , we see that for temperatures @xmath226 , the magnetization shows strong fluctuations . reducing the number of excited modes to a single one , namely , to the one that coincides with the fourier mode of the hmf potential
, it seems from fig .
[ inhom-1](d ) that only the dynamical transition to the collapsed state at a temperature @xmath227 persists .
the hint that the nature of the phase transition at @xmath224 is of first - order comes from the hysteresis plots of fig .
[ hys ] . to obtain these plots ,
one monitors the magnetization while tuning adiabatically the kinetic temperature across @xmath224 from higher to lower values and back to complete a full cycle . as is evident from fig .
[ hys ] , the observed hysteresis is between the collapsed state and the zero - magnetization state . in principle , it should be possible to observe a hysteretical behavior between the magnetized and the zero - magnetization state . to achieve this in simulations involving adiabatic tuning of temperature
, one should not allow the system to make the transition to the collapsed state , which requires conditions close to those that ensure detailed balance .
however , a possible drawback of this method is that closeness to detailed balance might lead to narrow hysteresis loops .
moreover , the adiabatic tuning of temperature should not be very slow , as otherwise one observes bistability instead of the hysteresis .
all these factors make the observation of hysteresis between the magnetized and the zero - magnetization state difficult to observe numerically ; further explorations of this will be the subject of future investigations .
for two different values of @xmath51 . in each case , the modes @xmath228 are excited with amplitudes satisfying @xmath229 , where @xmath230 . in all cases , @xmath231 , while the tuning rate for @xmath185 is @xmath214 .
the grey points correspond to the case when the temperature is decreased from high values , while the black points correspond to the case when the temperature is increased from low values.,title="fig:",width=302 ] for two different values of @xmath51 . in each case , the modes @xmath228 are excited with amplitudes satisfying @xmath229 , where @xmath230 . in all cases , @xmath231 , while the tuning rate for @xmath185 is @xmath214 .
the grey points correspond to the case when the temperature is decreased from high values , while the black points correspond to the case when the temperature is increased from low values.,title="fig:",width=302 ] in order to explore further the region in fig .
[ inhom-1](c ) close to @xmath224 , and to ascertain the nature of the phase transition at @xmath224 , we fix the value of the temperature to be @xmath232 , and monitor the magnetization as a function of time .
the time series of the magnetization is shown in fig .
[ inhom-2](a ) , in which one observes clear signatures of bistability , whereby the system switches back and forth between homogeneous ( @xmath233 ) and inhomogeneous ( @xmath234 ) states .
in addition , we show in fig . [
inhom-2](b ) the distribution of the magnetization around the phase transition temperature : the distribution is bimodal with a peak around a zero value and another one around a positive value . when decreasing the temperature across the phase transition region , we clearly see that the peak heights of the distributions of the magnetization at the zero and non - zero values interchange . these two features , together with the hysteresis plots of fig .
[ hys ] , support the first - order nature of the transition around @xmath224 which can be estimated from fig .
[ inhom-2](b ) to be @xmath235 . , with @xmath236 , and with modes @xmath228 excited , whose amplitudes satisfy @xmath237 , where @xmath137 .
the figure shows clear signatures of bistability in which the system during the course of evolution switches back and forth between spatially homogeneous ( @xmath238 ) and inhomogeneous ( @xmath239 ) states .
( b ) distribution @xmath240 of the magnetization @xmath241 as a function of @xmath185 at a fixed value of @xmath242 .
the data are obtained from numerical simulation results similar to ( a ) for magnetization in the stochastically - forced hmf model , with @xmath243 , and with modes @xmath228 excited , whose amplitudes satisfy @xmath237 , where @xmath137.,title="fig:",width=302 ] , with @xmath236 , and with modes @xmath228 excited , whose amplitudes satisfy @xmath237 , where @xmath137 .
the figure shows clear signatures of bistability in which the system during the course of evolution switches back and forth between spatially homogeneous ( @xmath238 ) and inhomogeneous ( @xmath239 ) states .
( b ) distribution @xmath240 of the magnetization @xmath241 as a function of @xmath185 at a fixed value of @xmath242 .
the data are obtained from numerical simulation results similar to ( a ) for magnetization in the stochastically - forced hmf model , with @xmath243 , and with modes @xmath228 excited , whose amplitudes satisfy @xmath237 , where @xmath137.,title="fig:",width=302 ] in the inhomogeneous state , for two values of @xmath51 .
the data are obtained from simulations with modes @xmath228 excited , whose amplitudes satisfy @xmath237 , where @xmath137 . here , the kinetic temperature @xmath232 , while @xmath244.,width=302 ] ptau from fig .
[ inhom-2 ] , it is clear that the system has two well separated attractors , corresponding to the homogeneous and inhomogeneous states .
a question of immediate interest is : how long does the system stay in one state before switching to the other ?
let us define the residence time as the time the system stays in one state before it switches to the other . in the limit of low noise level @xmath51 ,
there is a clear separation between the natural dynamical time and the typical residence time , as is evident from fig .
[ inhom-2](a ) . as a result
, one may conjecture that two successive switching events are statistically independent of one another . in case
such a conjecture holds for our model , the residence time statistics will be a poisson process , characterized solely by the probability per unit time , @xmath245 , of switching from the inhomogeneous state to the homogeneous state , and the probability per unit time , @xmath246 , for the reverse switch .
the distribution of residence time @xmath247 in each phase is then exponential : p_()=(- _ ) , so that the average residence times in the two states are @xmath248 .
such an exponential form of the residence time distribution is verified from our simulation data displayed in fig .
note that generating such a plot requires running simulations of the dynamics for long enough times so that the magnetization switches back and forth between the two states a sufficient number of times , and one has good statistics for the residence times . for low values of @xmath51 , such as those used in fig .
[ ptau ] , this was often not feasible due to very long simulation times .
this results in bad statistics , and hence , the form of the plot displayed in fig . [ ptau ] , which , though good , may be improved upon by running longer simulations . we conclude that our conjecture of two successive jumps being independent holds for our model , and that the average residence time fully characterizes the switching process for small enough @xmath51 .
we now discuss how the residence times depend on the system parameters , in particular , on @xmath51 . for an equilibrium system
, the type of switching process described above is an activation process with a residence time described by the arrhenius law .
a simple model of such an activation process is the langevin dynamics of a hamiltonian system in a potential @xmath249 .
the noise level is then related to the temperature , and the arrhenius law takes the form @xcite @xmath250 here , @xmath251 and @xmath252 are respectively the potential energy barrier as observed from the inhomogeneous and the homogeneous state . in a non - equilibrium context such as ours , there is no obvious equivalent of a potential , but the law given by eqs .
( [ tres - expect1 ] ) and ( [ tres - expect2 ] ) is expected to hold on a fairly general basis , in the limit of small noise .
this may be established from the instanton theory , or , from the freidlin - wentzell theory , which allows to compute @xmath249 explicitly for a given model @xcite .
our system does not fulfill the hypothesis of freidlin - wentzell theory , nevertheless , it is interesting to check if the law given by eqs .
( [ tres - expect1 ] ) and ( [ tres - expect2 ] ) holds .
our simulation data shown in fig .
[ mean - tau ] show that the dependence of @xmath253 on @xmath51 , as in eqs .
( [ tres - expect1 ] ) and ( [ tres - expect2 ] ) , holds also for our model , thereby suggesting that in the limit of low noise , the system behaves as one with transitions activated by a weak noise .
the @xmath254 of the mean residence time @xmath255 in the two bistable states , namely , the inhomogeneous ( @xmath256 ) state and the homogeneous state ( @xmath233 ) .
the plot is based on data obtained from simulations with modes @xmath228 excited , whose amplitudes satisfy @xmath237 , where @xmath137 . here , the kinetic temperature @xmath232 , while @xmath244 . the straight line fits imply that @xmath257 , in accordance with eqs .
( [ tres - expect1 ] ) and ( [ tres - expect2 ] ) . that the slopes of the two straight lines in the plot are different could be due to the fact that the height of the barrier is different
when observed from the inhomogeneous and the homogeneous state.,width=302 ] mean - tau we conclude this section by describing briefly the algorithm to find @xmath258 to produce fig .
[ ptau ] , and @xmath253 to produce fig .
[ mean - tau ] . to this end
, one has to identify from the time series data of the magnetization ( see fig .
[ inhom-2](a ) for an example ) the switching time instants between the two states . in the limit of very small @xmath51 ,
the distinction between two states should be obvious .
however , we could not reach such a limit in our numerical simulations because the simulation time grows exponentially with @xmath38 ( see fig .
[ mean - tau ] ) . for intermediate values of @xmath51 ,
it is then a challenge to define precisely the two states . indeed , as may be seen in fig .
[ inhom-2](a ) , the data show strong fluctuations and hence , one needs to filter out spurious " switching events and retain only the genuine ones .
this may be done efficiently as we now discuss .
we first obtain from the data a rough estimate of the mean of the magnetization when the system is in the two states , the homogeneous and the inhomogeneous one .
let us denote by @xmath259 and @xmath260 these estimates when the system is in the inhomogeneous and the homogeneous state , respectively .
let us define a
threshold " value of the magnetization @xmath261 as the average of @xmath259 and @xmath260 ; magnetization crossing this threshold to switch from one state to another however is not a precise enough criterion to define a switching event , as is obvious from fig .
[ inhom-2 ] .
we thus resort to our algorithm which we now illustrate for the case when the system is in the homogeneous state ; when the system is in the inhomogeneous state , the algorithm may be defined in a manner similar to the one below . in our algorithm , we identify a switching event as the one for which the following two conditions are satisfied , namely , ( i ) that the magnetization crosses the threshold to switch from the homogeneous to the inhomogeneous state , ( ii ) the magnetization after the switching reaches the value @xmath259 before reaching the value @xmath260 . when a switching event occurs , the switching time is defined as the time at which the magnetization crossed the threshold .
this algorithm allows us to precisely define the switching times , from which we compute the switching time statistics @xmath262 , and hence , the mean residence time @xmath253 .
in this work , we considered long - range interacting systems driven by external stochastic fields , thereby leading to generic nonequilibrium stationary states . to study spatially homogeneous stationary states , we developed a kinetic theory approach by generalizing the known results for isolated long - range systems .
our theoretical approach is quite general , being applicable to any long - range inter - particle potential , space dimensions and boundary conditions .
our extensive numerical simulations on a paradigmatic model of long - range interacting systems demonstrated a very good agreement with the theory . besides
, our simulations for this representative case illustrated very interesting bistable behavior between homogeneous and inhomogeneous states , with a mean residence time that diverges as an exponential in the inverse of the strength of the external stochastic forces in the limit of low values of such forces .
let us note that another route to deriving the kinetic theory studied in this paper is to adopt an approach similar to the one due to klimontovich for isolated systems , by writing down the time evolution equation for the noise - averaged empirical measure @xmath263 . in the resulting equation
, the noise appears as a multiplicative term , which can be treated perturbatively , leading to the kinetic equation ( [ kinetic_equation ] ) .
this work leaves open some interesting issues , e.g. , for technical simplicity , we assumed a homogeneous stationary state for the development of the kinetic theory .
it would be of interest to generalize the theory to inhomogeneous states ; in this regard , the method due to heyvaerts reported recently may come of help @xcite .
another issue is to study the dynamics of the kinetic equation ( [ kinetic_equation ] ) , both analytically and numerically , which may unveil very interesting behaviors , such as limit cycles .
one may also hope to develop a kinetic theory similar to the one analyzed here for related systems , for example , the point vortex model and the euler equations in two - dimensional turbulence @xcite .
c. n. acknowledges the egide scholarship funded by ministre des affaires trangres .
s. g. and s. r. acknowledge the contract loris ( anr-10-cexc-010 - 01 ) .
f. b. acknowledges the anr program statocean ( anr-09-sysc-014 ) .
numerical simulations were done at psmn , ens - lyon .
s. g. is grateful to the korea institute for advanced study ( kias ) , seoul for the kind hospitality during the time part of this work was being done and written up .
we prove here that the dynamics defined by the equations of motion ( [ equations_of_motion ] ) satisfies detailed balance if and only if @xmath58 for all @xmath59 , that is , if the stochastic forcing has a white spectrum in space .
we start from the @xmath0-particle fokker - planck equation ( [ eq : n - particle - fokker - planck ] ) associated with the equations of motion ( [ equations_of_motion ] )
. it will be useful to rewrite it in the following way : @xmath264 + \frac{1}{2}\sum_{i , j=1}^{2n}\frac{\p^2}{\p x_i\p x_j}[b_{i , j}(\bold{x } ) f_n(\bold{x})]\,,\ ] ] where @xmath265 for @xmath266 , @xmath267 for @xmath268 , and we use the notation @xmath269 .
the drift vector @xmath270 is a function of the @xmath271 s , and is given by @xmath272 similarly , the expression for the ( symmetric ) diffusion matrix @xmath273 is : @xmath274 and @xmath275 otherwise .
we moreover introduce the constants @xmath276 , which denote the parity with respect to time inversion of the variables @xmath271 , and the notation @xmath277 . it can be shown ( see @xcite , sect . 5.3.5 , or @xcite , sect .
6.4 ) that the dynamics described by a fokker - planck equation of the form ( [ fokker - planck - appendix ] ) satisfies detailed balance if and only if the following two conditions are satisfied ( @xmath278 ) : @xmath279 and @xmath280 where @xmath281 is the stationary solution of ( [ fokker - planck - appendix ] ) . in our case ,
in which the drift and the diffusion terms are given by eqs .
( [ drift - appendix ] ) and ( [ diffusion - appendix ] ) , respectively , the condition ( [ condition-1-appendix - db ] ) is trivially satisfied .
our proof goes as follows : we solve formally eq .
( [ condition-2-appendix - db ] ) , and show that @xmath281 is a stationary solution of eq .
( [ fokker - planck - appendix ] ) if and only if the non - vanishing part of @xmath273 is proportional to the identity matrix .
then , it is simple to show that this implies that the spectrum of the forcing has to be white in space .
equation ( [ condition-2-appendix - db ] ) for @xmath266 is also trivially satisfied . on the other hand , for what concerns @xmath268
, we have @xmath282 where @xmath283 .
we introduce the @xmath284 matrix @xmath285 whose components are given by @xmath286 , and observe that , for generic values of the @xmath2 s , @xmath285 admits an inverse @xmath287 . integrating eq .
( [ condition-2-appendix - db-2 ] ) , we thus have @xmath288,\ ] ] where @xmath289 is an undetermined function . inserting eq .
( [ stationary - solution - appendix - db ] ) into the fokker - planck equation ( [ fokker - planck - appendix ] ) , imposing that it is a stationary solution , and with some calculations , we get @xmath290=0\,.\ ] ] then , @xmath291 is a function of the hamiltonian @xmath292 , that is @xmath293 for some function @xmath294 . on the other hand , because @xmath291 is given by the formula in eq .
( [ stationary - solution - appendix - db ] ) , we can also deduce that @xmath294 is an exponential , and thus , that @xmath291 is gaussian in the velocities .
we conclude that @xmath287 ( and hence , @xmath285 ) has to be independent of the @xmath2 s and proportional to the identity . finally , from the form of @xmath295 in eq .
( [ ckdefinition ] ) , we see that this condition on @xmath285 is satisfied if and only if the spectrum of the forcing is white in space .
[ bbgkyclosure ] we analyze here in detail the closure of the bbgky hierarchy discussed in the text , in particular , the reasons for which the connected part of the two - particle correlation is of order @xmath51 , while higher correlations are negligible at leading order in @xmath51 , so that this closure is self - consistent . in the following ,
we expand the functions @xmath296 and @xmath297 as @xmath298 and @xmath299 and similarly , for other @xmath300 s for @xmath301 .
now , let us write explicitly the first two equations of the bbgky hierarchy ( [ bbgky - stochastic ] ) . the first one , obtained from eqs .
( [ bbgky - stochastic ] ) and ( [ 2-particle - connected - correlations - appendix ] ) , is @xmath302}{\p q}-\alpha\frac{\partial}{\partial p}[pf]-\frac{\alpha}{2}c(0)\frac{\partial^{2}f}{\partial p^{2}}=\frac{\partial}{\partial p}\int{\rm d}q_{1}{\rm d}p_{1}\,v'(q - q_1)\tilde{g}(z , z_1,t),\ ] ] where @xmath303(q)=\int { \rm d}q_{1}{\rm d}p_{1}\,v(q - q_1)f(q_1,p_1,t)\ ] ] is the mean - field potential . for the second equation of the hierarchy , we use eqs .
( [ 3-particle - connected - correlations - appendix ] ) and ( [ eq : stochastic - bbgky-1-eq - appendix ] ) to get @xmath304}{\p q_1}+\frac{f(z_2)}{n}\frac{\p v(q_1-q_2)}{\p q_1}\frac{\p f}{\p p_1}+\frac{1}{n}\frac{\p v(q_1-q_2)}{\p q_1}\frac{\p \tilde{g}}{\p p_1}\nonumber\\ & & + \frac{\p f}{\p p_1}\int { \rm d}z_3 \frac{\p v(q_1-q_3)}{\p q_1}\tilde{g}(z_2,z_3 ) + \frac{\p}{\p p_1}[\alpha p_1 \tilde{g}]+\frac{\alpha}{2}c(|q_1-q_2|)\frac{\p f}{\p p_1}\frac{\p f}{\p p_2}\nonumber\\ & & + \frac{\alpha}{2}c(0)\frac{\p^2 \tilde{g}}{\p p_1 ^ 2}+\frac{\alpha}{2}c(|q_1-q_2|)\frac{\p^2 \tilde{g}}{\p p_1 \p p_2}+\int { \rm d}z_3\ , \frac{\p v(q_1-q_3)}{\p
q_1}\frac{\p h}{\p p_1}\big]\nonumber \\ & & + \{1\leftrightarrow 2\},\end{aligned}\ ] ] where the symbol @xmath305 stands for an expression obtained from the bracketed one on the right hand side by exchanging the subscripts @xmath93 and @xmath94 .
let us analyze the order of magnitude of various terms in eq .
( [ eq : stochastic - bbgky-2-eq - appendix ] ) .
first of all , we have @xmath306 , as it is normalized to unity . however , we do not know a priori the order of magnitude of @xmath307 and @xmath68 .
thus , the order of magnitude of all but the terms @xmath308 and @xmath309 is unknown . in the continuum limit @xmath310
, we have @xmath311 so that it is natural to guess that @xmath312 .
let us also observe that in the limit @xmath313 , we have @xmath314 so that we obtain @xmath315 . in the limit @xmath54 , the kinetic theory leads to the lenard - balescu equation .
once we have established that @xmath312 , one can write down the equation of the hierarchy for @xmath68 and , with similar reasoning as above , one then finds that @xmath68 is at least of order @xmath316 ( or , @xmath317 depending on whether @xmath318 or the reverse ) , so that the term @xmath319 is negligible in eq .
( [ eq : stochastic - bbgky-2-eq - appendix ] ) , as may be straightforwardly checked .
the iterative procedure can be repeated at all orders of the hierarchy .
discarding three - particle and higher - order correlations is thus a self - consistent procedure .
moreover , note that in eq .
( [ eq : stochastic - bbgky-2-eq - appendix ] ) , some of the terms are of higher orders ( @xmath317 , @xmath320 , ... ) with respect to @xmath51 , and thus , can be discarded .
the final form of the second equation of the bbgky hierarchy is thus @xmath321}{\p q_1 } + \frac{\p f}{\p p_1}\int { \rm d}z_3 v'(q_1-q_3)\tilde{g}(z_2,z_3 ) \nonumber \\ & & + \frac{\alpha}{2}c(|q_1-q_2|)\frac{\p f}{\p p_1}\frac{\p f}{\p p_2}\big]+\{1\leftrightarrow 2\}.\end{aligned}\ ] ] note that @xmath312 implies , see eq .
( [ eq : stochastic - bbgky-1-eq - appendix ] ) , that the mean - field effect of the stochastic forces gives a contribution at the same order to the two - particle correlation induced by them .
we derive here the evolution of the kinetic energy as obtained from the kinetic equation ( [ eq : stochastic - kinetic - equation ] ) .
let us recall that the average kinetic energy density at time @xmath42 in the continuous limit can be written as @xmath322 the starting point to obtain its time evolution is to multiply the kinetic equation ( [ eq : stochastic - kinetic - equation ] ) by @xmath323 , and then , to integrate over @xmath76 . neglecting for the moment the non - linear part of the diffusion coefficient , and integrating by parts
, we get @xmath324 which gives @xmath325 the kinetic energy density in the stationary state is thus @xmath33 .
we now have to prove that the non - linear part of the diffusion coefficient ( [ diffusion_coefficient ] ) does not contribute to the time evolution of the kinetic energy .
such a result is expected and is usually valid for collisional terms ( i.e. , those terms in the kinetic equations which are given by two - particle correlation ) , for example , in the boltzmann equation or in the lenard - balescu equation @xcite .
the contribution to @xmath326 from the non - linear part of the diffusion coefficient is a sum of terms proportional to @xmath327\right\ } ; \nonumber \\\end{aligned}\ ] ] we will show that each of such terms vanishes independently . indeed , integrating the last expression over @xmath76 by parts , we get that @xmath328.\ ] ] exchanging now the variables @xmath329 and @xmath76 and the order of integration , we get that the above equation may be rewritten as @xmath330\,.\ ] ] summing up the last two equations , we therefore have @xmath331,\ ] ] which vanishes on integrating by parts both with respect to @xmath329 and @xmath76 .
we prove here that for a general forcing spectra , the gaussian distribution function in eq .
( [ gaussian ] ) is not a stationary solution of the kinetic equation ( [ eq : stochastic - kinetic - equation ] ) , and that the tails of any stationary state are gaussian . for the first point
, we have to prove that the contribution to @xmath332 from the non - linear part of @xmath128(p)$ ] in eq .
( [ diffusion_coefficient ] ) is not vanishing .
this result can be proven with an asymptotic expansion @xcite for large momenta of the integrals which appear in the diffusion coefficient . given any function @xmath333 , we approximate integrals of the form @xmath334 by expanding @xmath335 in taylor series .
we get , for example , @xmath336\,\simeq\,\frac{1}{p^2},\ ] ] where , in the last equality , we have taken into account the fact that the gaussian distribution being even , the terms containing @xmath337 with @xmath59 odd do not contribute . in a similar way , we have @xmath338 and @xmath339\,\simeq\,\frac{2\beta ^{3/2}}{\sqrt{\pi}\,p^2}\,\int dp_1\,\frac{p_1 ^ 2\,e^{-\beta p_1 ^ 2}}{|\epsilon(k , k p_1)|^2},\ ] ] where we have used the fact that @xmath340 is an even function of @xmath76 . with these results , we can evaluate the non - linear part of the kinetic equation : for large @xmath329 , the non - linear contribution to @xmath341 is @xmath342\left[\frac{4\beta^{5/2}}{\sqrt{\pi}}\,e^{-\beta p_1 ^ 2}\right].\ ] ] it can be shown that such a term is a non - vanishing function of @xmath329 .
this completes the proof : for a generic forcing spectra , the stationary state , when exists , is not gaussian . using the same asymptotic expansion as before
, it can be checked that the diffusion coefficient @xmath128(p)$ ] converges to @xmath343 for any distribution @xmath75 . from this observation and eq .
( [ stationary - fokkerplanck ] ) , it follows that any stationary solution of the kinetic equation ( [ eq : stochastic - kinetic - equation ] ) has gaussian tails .
10 campa a , dauxois t and ruffo s , _ statistical mechanics and dynamics of solvable models with long - range interactions _ , 2009 _ phys .
* 480 * 57 bouchet f , gupta s and mukamel d , _ thermodynamics and dynamics of systems with long - range interactions _ , 2010
_ physica a _ * 389 * 4389 chavanis p h , _ phase transitions in self - gravitating systems _ , 2006 _ int .
b _ * 20 * 3113 topical issue : long - range interacting systems bouchet f and venaille a , _ statistical mechanics of two - dimensional and geophysical flows _ , 2012 _ phys . rep . _
* 515 * 227 .
weinberg m d , _ noise driven evolution in stellar systems i. theory _ ,
2001 _ mon . not .
r. astron .
* 328 * 311 liewer p c , _ measurements of microturbulence in tokamaks and comparisons with theories of turbulence and anomalous transport _ , 1985 _ nucl . fusion _ * 25 * 543 dhar a , _ heat transport in low - dimensional systems _ , 2008 _ adv .
* 57 * 457 derrida b , _ non - equilibrium steady states : fluctuations and large deviations of the density and of the current _ , 2007 _ j. stat .
_ p07023 jarzynski c , _ nonequilibrium work relations : foundations and applications _
, 2008 _ eur .
j. b _ * 64 * 331 nicholson d r , _ introduction to plasma physics _ , 1992
( krieger , malabar , florida ) lifshitz e m and pitaevski l p , _ physical kinetics _ , 2002 ( butterworth - heinemann , london ) bouchet f and simonnet e , _ random changes of flow topology in two - dimensional and geophysical turbulence _ , 2009 _ phys .
lett . _ * 102 * 94504 briggs r j , daugherty j d and levy r h , _ role of landau damping in crossed - field electron beams and inviscid shear flow _ , 1970 _ phys
. fluids _ * 13 * 421 dikii l a , _ the stability of plane - parallel flows of an ideal fluid _ , 1960 _ dokl . akad .
nauk sssr _ * 135 * 1068 , translated in _ sov .
phys . doklady _ * 5 * , 1179 nardini c , gupta s , ruffo s , dauxois t and bouchet f , _ kinetic theory for non - equilibrium stationary states in long - range interacting systems _ , 2012 _ j. stat .
_ l01002 chavanis p h , _ kinetic theory of spatially homogeneous systems with long - range interactions : i. general results _ , 2012 _ eur .
j plus _ * 127 * 19 heggie d and hut p , _ the gravitational million - body problem _ , 2003 ( cambridge university press , cambridge , uk ) kac m , uhlenbeck g e and hemmer p c , _ on the van der waals theory of the vapor - liquid equilibrium i. discussion of a one - dimensional model _ , 1963 _ j. math .
* 4 * 216 papoulis a , _ probability , random variables and stochastic processes _ , 1965 ( tokyo : mcgraw - hill kogakusha ) rey - bellet l , _ ergodic properties of markov processes _ , in _
open quantum systems ii .
the markovian approach , lecture notes in mathematics 1881 _ , 2006 ( springer , berlin ) gardiner c w , _ handbook of stochastic methods for physics , chemistry and the natural sciences _
, 1983 ( springer - verlag , berlin ) antoni m and ruffo s , _ clustering and relaxation in hamiltonian long - range dynamics _ , 1995 _ phys . rev
* 52 * 2361 yamaguchi y y , barr j , bouchet f , dauxois t and ruffo s , _ stability criteria of the vlasov equation and quasi - stationary states of the hmf model _ , 2004 _ physica a _ * 337 * 36 risken h , _ the fokker - planck equation : methods of solutions and applications _ , 1989 ( springer - verlag , berlin ) huang k , _ statistical mechanics _ ,
1987 ( wiley , new york ) nardini c and bouchet f , 2013 ( in preparation ) balescu r , _ statistical dynamics : matter out of equilibrium _ , 1987
( imperial college press , london ) mclachlan r i and atela p , 1992 _ nonlinearity _ * 5 * 541 baldovin f and orlandini e , _ nos - hoover and langevin thermostats do not reproduce the nonequilibrium behavior of long - range hamiltonians _
, 2007 _ int .
b _ * 21 * 4000 chavanis p h , baldovin f and orlandini e , _ noise - induced dynamical phase transitions in long - range systems _ , 2011 _ phys . rev
. e _ * 83 * 040101(r ) frank t d , _ nonlinear fokker - planck equations : fundamentals and applications _
, 2005 ( springer , berlin ) arnold
l , _ random dynamical systems _ , 2003 ( springer - verlag , berlin ) maritan a and banavar j r , _ chaos , noise , and synchronization _ , 1994 phys
lett . * 72 * 1451 ; pikovsky a s , _ comment on chaos , noise , and synchronization " _ 1994 _ phys .
lett . _ * 73 * 2931 ; maritan a and banavar j r , _ maritan and banavar reply _ 1994 _ phys . rev .
lett . _ * 73 * 2932 kramers h a , _ brownian motion in a field of force and the diffusion model of chemical kinetics _ ,
1940 physica * 7 * 284 van kampen n g , _ stochastic processes in physics and chemistry _ , 1992 ( north holland , amsterdam ) hnggi p , talkner p and borkovec m , _ reaction rate theory : fifty years after kramers _ 1990 _ rev .
_ * 62 * 251 freidlin m i and wentzell a d _ random perturbations of dynamical systems _ , 1998
( springer , berlin ) touchette h , _ the large deviation approach to statistical mechanics _
, 2009 _ phys .
rep . _ * 478 * 1 heyvaerts j , _ a balescu - lenard - type kinetic equation for the collisional evolution of stable self - gravitating systems _ ,
2010 _ mon . not .
. soc . _ * 407 * 355 | we study long - range interacting systems driven by external stochastic forces that act collectively on all the particles constituting the system .
such a scenario is frequently encountered in the context of plasmas , self - gravitating systems , two - dimensional turbulence , and also in a broad class of other systems . under the effect of stochastic driving ,
the system reaches a stationary state where external forces balance dissipation on average .
these states have the invariant probability that does not respect detailed balance , and are characterized by non - vanishing currents of conserved quantities . in order to analyze spatially homogeneous stationary states
, we develop a kinetic approach that generalizes the one known for deterministic long - range systems ; we obtain a very good agreement between predictions from kinetic theory and extensive numerical simulations .
our approach may also be generalized to describe spatially inhomogeneous stationary states .
we also report on numerical simulations exhibiting a first - order nonequilibrium phase transition from homogeneous to inhomogeneous states .
close to the phase transition , the system shows bistable behavior between the two states , with a mean residence time that diverges as an exponential in the inverse of the strength of the external stochastic forces , in the limit of low values of such forces . |
in the local universe more than half of the stellar mass resides in passively evolving galaxies ( peg ) , and above @xmath2 pegs outnumber starforming galaxies by a factor of 10 or more ( e.g. baldry et al . 2004 ) . as most pegs are morphologically classified as ellipticals , and most ellipticals are pegs ( e.g. , renzini 2006 , scarlata et al .
2007 ) sometimes pegs and ellipticals , or early type galaxies ( etg ) , are used as synonyms , though there is no perfect overlap , neither at low nor high redshift ( e.g. scarlata et al . 2007 ) , this makes pegs an especially important class of cosmic objects , and understanding their formation and evolution is central to the broader issue of galaxy formation in general . only in recent years
it has become possible to observationally map the evolution of the peg population all the way to fairly high redshifts ( @xmath3 ) , as a result of combining space and ground observations taken at the most powerful facilities .
the ultimate aim of these efforts is that of tracing the evolution with redshift of the number density and mass function of pegs , all the way to their disappearance , i.e. to the epoch when their precursors were still actively star forming , or were not assembled yet
. thus , to this end three main issues have to be attacked , namely : identify suitable peg candidates at high redshift , measure their redshift , and estimate age , metallicity and mass of their stellar populations . in this paper
i will briefly report the main results from projects such as the k20 ( cimatti et al .
2002 ; mignoli et al . 2005 ) , gmass ( cimatti et al . , in preparation ; kurk et al . , in preparation ) , and a joint eso / subaru project ( kong et al .
2006 ; daddi et al . , in preparation ) .
multiobject spectroscopy in the rest - frame ultraviolet has been so far the main tool that has allowed us to precisely locate high-@xmath4 pegs in space ( and time ) .
quite similar results are being obtained by the gdds team , and those are reported by pat mccarthy at this meeting .
the spectroscopic identification of pegs is relatively easy up to redshift @xmath6 . as shown in figure 1 , the caii h&k lines together with the 4000 break are the strongest features , and are very effective for measuring accurate redshifts .
large samples of pegs have then been culled up to this redshift ( e.g. zucca et al . 2005 ; faber et al .
2007 ; scarlata et al . 2007 ) .
however , at higher redshifts these features first become contaminated by atmospheric oh emission , and then move beyond the sensitivity limit of ccd detectors .
hence , other features need to be used to measure redshifts with optical spectrographs .
an inspection of the peg spectrum in figure 1 shows that to the blue of the caii doublet the strongest features are an isolated fei line at @xmath7 , and a complex feature extending over the range @xmath8 , due to various fe , mgi and mgii lines , that we call the mg - uv feature .
it is basically thanks to this feature that all redshifts of pegs at @xmath9 have so far been measured .
the fist such case was for a peg at @xmath10 selected for being a radiogalaxy ( dunlop et al 1996 ; spinrad et al .
then almost a decade passed before other pegs at @xmath9 could be spectroscopically identified ( 5 such objects by glazebrook et al .
2004 , and 4 objects by cimatti et al .
the most exciting aspect of these 2004 discoveries was that the 9 pegs were found over a combined field of only @xmath11 arcmin@xmath12 , while the dunlop et al .
object was selected from a catalog covering virtually the whole sky .
the clear implication was that massive pegs at high redshift ought to be much more numerous that many had suspected .
figure 2 shows the coadded spectrum of the 4 cimatti et al .
objects , along with a synthetic spectrum of a 1.1 gyr old simple stellar population ( ssp ) .
the excellent match between the observed and the synthetic spectrum implies that the bulk of stars in the four galaxies had to form at @xmath13 , given that they lie at @xmath14 .
indeed , the power of the mg - uv feature is that it allows us to get at once the redshift _ and _ an age estimate , hence to set a formation redshift .
this is illustrated in figure 3 , showing that indeed the feature appears only in populations that have evolved passively since at least @xmath15 gyr , while the feature is absent for younger ages or for populations with ongoing star formation .
therefore , this feature being observable up to high redshifts is very useful for age - dating pegs ( see e.g. , maraston et al .
2006 ) .
thanks to the mg - uv feature soon more high-@xmath4 pegs were discovered ( mccarthy et al .
2004 ; daddi et al .
2005 ) , but still within relatively small fields and capitalizing on spectroscopic surveys not specifically targeting such objects .
for a more focused spectroscopic study of high-@xmath4 pegs an effective criterion for selecting candidates was clearly required .
an efficient photometric criterion for the selection of pegs at high redshift was proposed by daddi et al .
( 2004 ) , based on the expected location of @xmath9 pegs on the @xmath17 vs @xmath18 two - color diagram .
moreover , the criterion was validated on existing spectroscopic data ( k20 , goods , etc . ) . the @xmath16 plot from kong et al .
( 2006 ) is shown in figure 4 .
kong et al .
then estimated that the comoving volume density of `` passive bzks '' ( pbzks ) with @xmath19 and @xmath20 is only @xmath21 of the corresponding space density at @xmath22 .
since the majority of such massive pegs appears to be already in place at @xmath23 , this implies that the epoch between @xmath24 and @xmath23 ( @xmath25 gyr ) must be that during which most of massive pegs are either finally assembled , and/or any residual star formation in their precursors is shut off .
besides high-@xmath4 pegs , the bzk criterion is also very efficient at selecting actively star - forming galaxies at @xmath26 ( hence called sbzks ) , even if rather heavily reddened .
the early zcosmos spectroscopic results have demonstarted that indeed the contamination by lower redshift interlopers is extremely small for sbzk - selected targets ( lilly et al .
thus , a sizable fraction of sbzks must turn into pbzks and lower redshift pegs between @xmath27 and @xmath28 .
how this transformation is taking place , in terms of star - formation quenching , merging , disk instability , and morphological transformation is perhaps the major open issue concerning the evolution of galaxies over the last @xmath29 gyr of cosmic time . of course , this must include the co - evolution of hosting galaxies and their central supermassive black holes ( smbh ) , and a first grasp at the agn activity among sbzk galaxies and the concomitant growth of their stellar mass and smbh can be found in daddi et al .
( 2007 ) . , reproduced from fig .
while extensive multi - object spectroscopy of star - forming galaxies at @xmath30 is relatively easy nowadays ( e.g. with vimos and fors2 at the vlt and lris at keck ) , an extensive spectroscopic survey of high-@xmath4 pegs is still lacking .
i like to report here some preliminary result from a couple of _ pilot _ projects , aiming primarily at understanding how to better plan for future massive surveys .
figure 5 shows the coadded spectrum for 17 pbzks selected from the sample of kong et al .
( 2006 ) , and observed with the vimos spectrograph at the vlt with an integration time of 2.5h ( daddi et al . in preparation ) .
the spectra were taken with the intermediate resolution ( @xmath31 ) red grism , while targeting sbzks and other galaxies as well , so to fully exploit the multiplex of the instrument .
overplotted to the coadded spectrum of the pbzks is the template spectrum for the @xmath32 pegs from the k20 survey ( from mignoli et al .
2005 ) .
figure 6 shows the coadded spectrum for 10 pbzks selected from the gmass sample , and observed with the fors2 spectrograph at the vlt with an integration time of 30h ( sic ! ) ( cimatti et al . in preparation ; kurk et al . in preparation ) .
the spectra were taken with a low resolution ( @xmath33 ) red grism .
the same lower redshift peg spectrum is also overplotted .
note that the coadded spectrum in this figure has the same s / n an individual galaxy would have if integrated for 300 hours . individual spectra taken with fors2 have far better s / n than those taken with vimos , given the much longer integration time , and the superior performance of fors2 for @xmath34 . in practice , we are fairly secure that all 10 fors2 redshifts are correct , while a few ( perhaps 3 or 4 , or @xmath35 ) among the vimos redshifts may not be correct .
optical spectroscopic observations in the rest - frame uv have demonstrated that pegs exist in sizable number up to at least @xmath36 .
the surface density of @xmath16-selected pegs is @xmath37 arcmin@xmath38 down to @xmath39 , substantially lower than the slit - packing capability of multiobject spectrographs such as vimos or fors2 .
moreover , the rest - frame uv continuum is quite naturally very faint in pegs , even in high - redshift ones , given that for being passive these galaxies lack ob stars in sizable number . actually , the absence of these stars is a requisite for detecting the mg - uv feature .
therefore , none of the optical instruments currently at 8 - 10 m telescopes is optimally suited to exploit the mg - uv feature for mapping the population of high-@xmath4 pegs .
very long integration times would be required , coupled with only a marginal exploitation for the specific targets of the instrument multiplex .
an alternative option to optical spectroscopy is to go to the near - ir , following the caii doublet and the 4000 break as they redshift beyond @xmath40 .
however , the secure detection even of the strongest absorption lines such as caii h&k would need extremely long integration times , which could be afforded only if very many objects could be observed simultaneously .
the only near - ir mos currently in operation is moircs at the subaru telescope ( tokoku et al .
several bzk - selected galaxies from the sample of kong et al .
( 2006 ) were ready to be observed with moircs during a 5-night run in april 2007 .
unfortunately the spectrographic channel of moircs had a technical problem when just inserting the first cryogenic mask into the instrument , and only the imaging mode could be used in the following nights .
as well know , cryogenics is critical for operations in the @xmath41 band , but a warm instrument working just in the @xmath42 band would be perfectly adequate for observing the caii doublet and the 4000 break up to @xmath24 and beyond . a first opportunity in this sense
will possibly be offered once more by the subaru telescope , using the fmos multifiber instrument ( kimura et al .
2003 ) . with its operation over the large fov at the prime focus of subaru
, fmos will have a very good match between its multiplex and the surface density of pbzks , and therefore promises to be perhaps a better option for mapping the population of high-@xmath4 pegs before jwst .
thus , spectroscopy in the rest - frame ultraviolet has been instrumental for demonstrating the existence of passively evolving galaxies at high redshifts , as well as for validating an efficient photometric criterion for selecting them
. however , it appears that near - ir multiobject spectroscopy targeting classical features in the rest - frame optical such as the caii doublet and the 4000 break may be better suited for a massive survey of these objects .
baldry ik , galzebrook k , brinkmann j , ivezi z , lupton lh , et al .
apj 600:681 bruzual g , charlot s 2003 .
mnras 344:1000 bundy k , ellis rs , conselice cj , taylor
je , cooper mc , et al . 2006 .
651:120 cimatti a , mignoli m , daddi e , pozzetti l , fontana a , et al . 2002b .
392:395 cimatti a , daddi e , renzini a , cassata p , vanzella e , et al . 2004 .
430:184 cimatti a , daddi e , renzini a 2006 .
453:l29 daddi e , cimatti a , renzini a , fontana a , mignoli m , et al .
617:747 daddi e , renzini a , pirzkal n , cimatti a , malhotra s , et al . 2005b .
626:680 daddi e , alexander dm , dickinson m , gilli r , renzini a , et al . 2007 . in press ( arxiv:0705.2832 )
dunlop j , peacock j , spinrad h , dey a , jimenez r , et al . 1996 .
381:581 faber sm , willmer cna , wolf c , koo dc , weiner bj , et al . 2007 .
665:265 glazebrook k , abraham rg , mccarthy pj , savaglio s , chen h - s , et al .
430:181 kimura m , maihara t , ohta k , iwamuro f , eto s , et al .
spie 4841:974 kong x , daddi e , arimoto n , renzini a , broadhurst t , et al . 2006 .
638:72 lilly sj , le fevre o , renzini a , zamorani g , scodeggio m , et al .
apjs cosmos special issue , in press ( astro - ph/0612291 ) maraston c , daddi e , renzini a , cimatti a , dickinson m , et al .
652:85 mccarthy pj , le borgne d , crampton d , chen h - w , abraham r , et al . 2004 .
614:l9 mignoli m , cimatti a , zamorani g , pozzetti l , daddi e , et al .
437 : 883 renzini a 2006 .
ara&a , 44 , 141 scarlata , c. ; carollo , c. m. ; lilly , s. j. ; feldmann , r. ; kampczyk , p. et al .
apjs , cosmos special issue , in press ( astro - ph/0701746 ) spinrad h , dey a , stern d , dunlop j , peacock j , et al . 1997 .
484:581 tokoku c , suzuki r , omata k , konishi m , yoshikawa t , et al . 2006 .
spie 6269 , 62694 zucca e , ilbert o , bardelli s , tresse l , zamorani g , et al . | beyond redshift @xmath0 the only spectral feature that allows one to get the redshift of passively evolving galaxies ( peg ) with optical spectrographs is a characteristic structure due to a set of iron and magnesium lines lines at @xmath1 in the rest frame .
the same feature permits also to estimate the time elapsed since the cessation of star formation .
current efforts at observing high redshift pegs at the vlt and subaru telescopes are briefly reviewed . |
despite its great experimental success ( see _ e.g. _ refs .
@xcite ) , it is well known that general relativity ( gr ) does not exhibit the most general form to couple matter with curvature .
indeed , these can be coupled , for instance , in a nonminimal way @xcite ( see also refs .
@xcite for early proposals in cosmology ) : this can have a bearing on the dark matter @xcite and dark energy @xcite problems ( see ref .
@xcite for a discussion in the context of @xmath1 theories , and ref .
@xcite for encompassing reviews ) , as well as inflation @xcite and structure formation @xcite . this putative nonminimal coupling ( nmc )
modifies the well - known energy conditions @xcite and can give rise to several implications , from solar system @xcite and stellar dynamics @xcite to close time - like curves @xcite and wormholes @xcite .
following the argument that @xmath1 theories should be derived from a more complete theory as low - energy phenomenological models @xcite , one also finds strong fundamental motivation for the presence of a nonminimal coupling ( nmc ) , as it arises from , for instance , one - loop vacuum - polarization effects in the formulation of quantum electrodynamics in a curved spacetime @xcite , as well as in the context of multi - scalar - tensor theories , when considering matter scalar fields @xcite ( as explicitly shown in ref .
@xcite ) . furthermore
, a nmc was put forward in an earlier proposal @xcite , developed in the context of riemann - cartan geometry , with another study showing that it clearly affects the features of the ground state @xcite . from a phenomenological standpoint
, one can consider that a natural way of extending the einstein - hilbert action implies substituting the linear curvature term and minimal coupling with holomorphic functions @xmath2 and @xmath3 .
as such , one considers the action functional @xcite , s = d^4 x , [ model]where @xmath4 ( @xmath5 ) are arbitrary functions of the scalar curvature , @xmath6 , @xmath7 is the determinant of the metric and @xmath8 . by setting @xmath9 and @xmath10 ,
the above encompasses the well - known @xmath1 theories , which are widely used to study the effect of modifications of gravity in a plethora of scenarios , _
e.g. _ the starobinsky inflationary model @xmath11 @xcite , the accelerated expansion of the universe @xcite , solar system tests @xcite , amongst many other studies ( see ref .
@xcite for a review ) . assuming that the functions @xmath12 are holomorphic
, one may express them as f_i(r ) = _
j=-^a_ij r^j , where including negative powers of @xmath6 allows for the earliest @xmath13 dark energy models @xcite .
inverse powers of @xmath6 clearly require that the background cosmological value of the scalar curvature is non - vanishing , as is characteristic of _
e.g. _ an exponential or power - law evolution of the scale factor .
the description of the non - trivial functions @xmath12 as a combination of infinite terms allows in principle the probing of individual terms of the summation by ascertaining the dynamical impact of the action eq .
( [ model ] ) in a specific setting where a particular term @xmath14 is dominant from the astrophysical @xcite and solar system @xcite to galactic @xcite and cluster @xcite scenarios , up to a cosmological scale @xcite ; this justifies the assumption of simple power laws for the generic functions @xmath15 .
interestingly , a recent study @xcite has shown that the choice @xmath16 and @xmath17 appears to negate the cosmological impact of the additional dynamics found in eq .
( [ field0 ] ) ; indeed , this choice enables the rewriting of the action functional eq .
( [ model ] ) as s = ( rm)^n d^4 x , [ model2]where @xmath18 is a characteristic mass scale to be determined .
this shows that such a subset of models can be viewed in terms of a generalised measure replacing the usual invariant @xmath19 .
variation with respect to the metric yields the modified field equations , [ field0 ] ( f_1 + f_2 ) g_&= & 1 2 f_2 t_+ _
( f_1 + f_2 ) + & & + 1 2 g _ , with @xmath20 and @xmath21 .
as expected , gr is recovered by setting @xmath22 and @xmath9 , while the usual field equations of @xmath1 theories are obtained from @xmath23 @xcite . the trace of reads ( f_1 + f_2 ) r = 1 2 f_2 t -3 ( f_1 + f_2 ) + 2 f_1 .
[ trace0]resorting to the bianchi identities , one concludes that the energy - momentum tensor of matter may not be ( covariantly ) conserved , since _
t^=f_2 f_2(g^-t^)_r , [ cov ] can be non - vanishing ( see refs .
@xcite for thorough discussions ) .
in ref . @xcite ( see also refs .
@xcite ) , it was shown that is equivalent to the action involving two additional scalar fields @xmath24 ( with dimensions of mass - squared ) , s = d^4 x .
[ model3]indeed , the variation of the above with respect to the scalar fields leads to = r , = f_1 ( ) + f_2 ( ) , which , upon replacing in eq .
( [ model3 ] ) , yields the original action functional eq .
( [ model ] ) .
the above action is in the jordan frame , as the scalar field @xmath25 appears coupled to the scalar curvature @xmath6 .
one may perform the conformal transformation ( while maintaining the original coordinate system ) @xmath26 , so that = ^-2 , r = , where @xmath27 denotes the dalembertian operator constructed from the metric @xmath28 . introducing the above relations into eq . ( [ model3 ] ) leads to a decoupling of the rewritten scalar curvature @xmath29 from the scalar field @xmath25 , so that the action functional is now in the einstein frame : [ equivaction ] s = d^4x ( 2 + l ) , where @xmath30 and @xmath31 are two scalar fields , related with the scalar curvature and the non - trivial @xmath2 and @xmath3 functions through [ deffields ] ^1 = 2 , ^2 = r , where the potential is given by [ potential ] u(^1,^2 ) = 1 4 ( -2 3 ^1 ) , and a modified lagrangian density is used when constructing the lagrangian density of the latter , [ clstar ] l ( ^1,^2,g _ , ) = f_2(^2 ) ( g _ , ) , where @xmath25 denotes existing matter fields .
notice that the non - minimal coupling present in the definition above is retained throughout the derivation , and that both @xmath32 and @xmath3 add an additional coupling through both the exponential term depending on @xmath30 and the `` physical metric '' [ physicalmetric ] g_= g _ , which is used to construct the original lagrangian density @xmath33 of matter ( _ e.g. _ through the contraction of indices ) .
the above can be better understood through the use of the specific example : for this , one adds next - to - leading order terms for both @xmath32 and @xmath3 , as previously explored in ref .
@xcite : f_1(r ) = r + r^2 6m^2 , f_2(r ) = 1 + 2 , where @xmath34 is the characteristic mass scale required for starobinsky inflation @xcite .
this is equivalent to the formulation eq .
( [ equivaction ] ) in the einstein frame , with the scalar fields ^1 = 2 ( 1 + r 3m^2 + 2 ) , ^2 = r , driven by the potential u(^1,^2 ) = 1 4 ^2 ( -2 3 ^1 ) .
in this section one approaches the issue of directly computing the gravity propagator of the nmc model under consideration . the improper use of the equivalence with a multi - scalar - tensor theory @xcite has led to misinterpretations reported in ref .
@xcite and discussed in the following section . for robustness , the computation of the propagator is discussed without resorting to the aforementioned equivalence multi - scalar - tensor theory ( see ref .
@xcite for an analysis of alternative @xmath35 theories ) .
the ( inverse ) gravity propagator around minkowski spacetime is given by _
^-1^h_=1 2 _ , [ invprop ] where @xmath36 is the energy - momentum tensor of matter and @xmath37 is a small perturbation to the minkowski background metric @xmath38 , _
i.e. _ @xmath39 . to compute it , the ( 4-rank tensor ) spin projector operators
are introduced @xcite , @xmath40 with the transversal and longitudinal projectors in momentum space given by @xmath41 the set @xmath42 constitute a complete set of projection operators ^i_ap^j_b=^ij_abp^i_a , p^2+p^1+p_s^0+p_w^0=1 , as one can easily verify .
@xmath43 and @xmath44 represent the four degrees of freedom of transverse and traceless spin-2 and spin-1 degrees , while @xmath45 , @xmath46 represent spin-0 scalar multiplets . furthermore , @xmath47 and @xmath48 mix the two scalar multiplets , according to ^0_ijp^0_k=_jkp^0_ij , p^0_ijp^0_kl = _
il_jkp^0_k , p^0_k p^0_ij = _
ikp^0_ij . as shown in refs .
@xcite , the gravity propagator of @xmath49 theories in a flat spacetime is given by _
f = _ gr + 1 2p^0_s k^2+m^2 , where @xmath50 and _ gr = 1 k^2 ( p^2 -1 2 p^0_s ) , is the gravity propagator of gr .
the above clearly depicts the additional scalar degree of freedom introduced by @xmath1 theories , as known from the equivalence between the latter and a scalar - tensor theory @xcite . in order to obtain the above propagator , the modified field eqs .
( [ field0 ] ) ( with @xmath9 ) were expanded around a flat spacetime , together with the leading order expansion @xmath51 . to extend this result to a non - trivial nmc @xmath52
, it suffices to perform the same procedure to the additional terms introduced by the latter , assuming the leading - order expansion below : f_2 ( r ) = 1 + r 6m^2 f_2 = 1 6m^2 .
notice that one could alternatively redefine the energy - momentum tensor @xmath53 as the variation of the nmc lagrangian density term , _
= -2 g^ = f_2 t_+ 2 _ ( f_2 ) - 2 f_2 r _ , where @xmath54 is the usually defined energy - momentum tensor of matter .
strikingly , the above shows that the linearized field equations can not be written in the form , as the matter lagrangian density @xmath33 and the metric perturbation and its derivatives @xmath55 appear coupled : instead , applying the spin projector formalism would eventually yield an irreducible equation of the form _ ^h_=1 2 g[(p_i^j)^ _ h _ , t _ ] , which , even if one could write the lagrangian density @xmath33 as a function of the energy - momentum tensor @xmath54 and the metric @xmath39 , would still possess cross - terms between @xmath56 and @xmath57 . as such , one finds that the direct derivation of a gravity propagator is unfeasible , since the nmc directly implies that one can not write the metric perturbation as given by the application of spin projectors to the energy - momentum tensor .
in a recent study @xcite , tamanini and koivisto claim that a nmc model is not viable , due to the appearance of ghost degrees of freedom . to derive this , ref .
@xcite starts by integrating out the auxiliary scalar field @xmath31 , thus rewriting the action , , as [ koivistoaction ] s = d^4x , where @xmath58 is a function obtained from the substitution of the solution @xmath59 of the field equation for this scalar field .
the authors then argue that a viable nmc model should allow the coupling to any lagrangian density for matter @xmath33 , and choose = - 12g^ _ , _ , x , _
i.e. _ a purely kinetic scalar field with no potential .
the above action with this choice is then compared with purely kinetic k - essence models @xcite , given by the action [ kessenceaction ] s = d^4x .this is done with little regard for the differences between k - essence and a nmc model , as no in depth justification is offered : instead , it is simply stated that such comparison is warranted because the kinetic term of @xmath30 is canonical , so that its contribution ( both the canonical kinetic term and the effect on @xmath60 can be disregarded .
conditions for positivity of the energy density @xmath61 of the matter scalar field @xmath62 and the speed of sound @xmath63 at which its perturbations propagate are then introduced @xcite , [ kenergy ] ( ) = 2 x_0 p_,x(x_0 ) - p(x_0 ) > 0 , and [ kspeed ] c_s^2 ( ) = p_,x(x_0 ) 2x_0 p_,x x(x_0 ) + p_,x(x_0)0 , and used to argue severe constraints on the form of the nmc function @xmath3 .
notice that the above expression is evaluated at the background value @xmath64 @xcite .
thus , the case @xmath65 does not necessarily imply that the pressure is independent of the kinetic term @xmath66 defined above ( as would be the case for a cosmological constant , where the associated pressure and density are constant ) : it may simply happen that the evaluation of a complex form at the background value vanishes , _
i.e. _ @xmath67 but @xmath68 @xcite . inspection shows that the comparison between and ( [ kessenceaction ] ) leaves the dynamical effect of the scalar field @xmath30 out of the ensuing discussion : this additional degree of freedom should have some impact on what otherwise bears a resemblance with k - essence models , specially since it is not an independent matter field , but is dynamically related to both the curvature and the lagrangian density of matter through . as such
, the purpose of this work is not to claim that no pathological behaviour can arise in nmc models as indeed one knows that the condition @xmath69 is required for the absence of ghosts in @xmath1 theories @xcite , while in nmc models it is known that dolgov - kawasaki instabilities appear if @xmath70 is negative @xcite : notwithstanding the need for future work , this comment aims at showing that the recent study by tamanini and koivisto does not address the problem at hand in a rigorous fashion .
several remarks are in order , as discussed below . as pointed out following , the nmc manifests itself in by both the original nmc ( now as a function of @xmath31 ) , a factor arising from the conformal transformation from the original jordan frame of to the einstein frame , and the physical metric resulting from the latter .
in particular , ref .
@xcite has shown that modifications to cosmological perturbations arise both from the effect of the modified field eqs .
( [ field0 ] ) as well as the non - conservation of the energy - momentum tensor , something which is not captured in the direct comparison with k - essence .
thus , if one aims at studying a nmc purely kinetic scalar field , @xmath71 can not be introduced directly in , as one should recall that the kinetic term of @xmath62 can not be constructed from the metric @xmath72 , but instead from the physical metric , so that one obtains [ diffconf ] x = 1 2 g^ _ , _ , = 1 2 ( 23 ^ 1 ) g^ _ , _ , = ( 23 ^ 1)x , where one recalls that the solution @xmath59 depends on both @xmath73 and @xmath30 . thus the k - essence function is of the form @xmath74x , \varphi^1)$ ] , with a dependence on @xmath30 that is much more evolved than assumed in ref .
@xcite which implies that neglecting these terms is improper .
conditions and ( [ kspeed ] ) are based upon the study of perturbations of the k - essence field @xmath62 around the cosmological frw background , so that the kinetic term @xmath66 is evaluated at its background cosmological value @xcite which is therefore constrained to be positive , @xmath75 , and can not assume any arbitrary value , as also claimed in ref .
@xcite .
as discussed in the previous paragraph , the correct kinetic term should be computed using the physical metric @xmath76 , leading to . since the derivation of and ( [ kspeed ] ) in the context of k - essence has no such distinction , the underlying computation of perturbations
can not be applied to the present context .
furthermore , the calculations leading to and ( [ kspeed ] ) assume that the cosmological evolution is driven by k - essence ( the dominant contribution to the dynamics , which also include minimally coupled baryonic matter ) @xcite , so that the analysis of ref .
@xcite , if correct , would only apply to a proposal for dark energy based upon a nmc scalar field with no potential term : as such , it would only be useful in ruling out such a model but not the existing proposal for a phase of accelerated expansion of the universe due to a nmc between curvature and baryonic matter @xcite , which is described by @xmath77 @xcite , its energy density , and requires no additional scalar field .
given the above , it is clear that a correct treatment of cosmological perturbations of a nmc purely kinetic scalar field requires the inclusion of the dynamical contribution of the additional degree of freedom embodied in @xmath30 . indeed , considering the perturbed metric ds^2 = ( 1 + 2)dt^2 - ( 1 - 2 ) a^2(t ) _ ij dx^i dx^j . and assuming small perturbations @xmath78 , @xmath79 yields , to first order in all perturbations , the field equation for @xmath62 @xcite , & & 0 = p_,x k^2 a^2 + p_,x _ 0 - 3 p_,x _ 0 + + & & ( p_,x + p_,xx _ 0 ^ 2 ) ( + 3 h - _ 0 ) + + & & [ ( 3 p_,xx + p_,xxx _ 0 ^ 2 ) _ 0 _ 0 + ( p_,x + p_,xx _ 0 ) _ 0 ^ 2 ] + + & & [ ( p_,x + p_,xx ^2_0 ) _ 0 + ( 3h p_,x + p_,x _ 0 ) _ 0 ] + & & + [ ( 3h p_,xx + p_,x _ 0 ) _ 0 ^ 3 - 2 ( 3h p_,x + p_,x _ 0 ) _ 0 - + & & ( 2 p_,x + 5 p_,xx _ 0 ^ 2 + p_,xxx _ 0 ^ 4 ) _ 0 ] . if one simply takes the ratio between the second time derivative and the gradient terms , ensues . as shown below , inclusion of the additional degree of freedom @xmath30 yields different results , showing that the above computation ignores the effect of other terms ( such as the gravitational potentials @xmath80 and @xmath81 ) , which are related to @xmath82 through the perturbed einstein equations . the authors of this work do not know of any study of cosmological perturbations of a interacting two - field k - essence model , precisely due to the intricate nature of the ensuing equations of motion ( and its perturbations ) ; existing works on interacting @xcite , non -
interacting @xcite and purely kinetic multifield k - essence models @xcite do show that the relevant criteria for model stability is that the speed of sound of curvature perturbations should be positive defined , and that the latter is related to the speed of sound of all the scalar fields considered .
notice that , since the scalar field @xmath30 has a canonical kinetic term , one can not use the matrix formalism of ref .
@xcite to assess the propagation of curvature perturbations in an interacting multifield k - essence model , which is only valid for non - linear kinetic terms . for the sake of argument we study what would happen if the contribution of @xmath30 to the k - essence function could somehow be disregarded ( as assumed in ref .
@xcite ) and forego this interaction term of , @xmath83 , but keep the kinetic term of latter thus considering instead the action [ multifieldaction ] s = d^4x , where @xmath84 and @xmath85 .
one may now resort to ref .
@xcite , where it was shown that the speed of sound of curvature perturbations is given by c_s^2 = ( _ i + p_i ) ( ) ^-1 , i = x , y .using eqs .
( [ kenergy ] ) and ( [ kspeed ] ) for the contribution of @xmath66 and @xmath86 , so that @xmath87 ( echoing the well - known result that perturbations of a canonical scalar field propagate at the speed of light ) , one gets c_s^2 = x p_,x + y ) + y , which collapses into when @xmath88 .
this highlights the relevance of the contribution of @xmath30 , and shows that the analysis of ref .
@xcite is only correct if the latter can be disregarded , as argued before .
again , evaluating this expression would require a knowledge of both @xmath66 and @xmath89 derived from a cosmological model for dark energy arising from a nmc purely kinetic scalar field , which has nothing in common with the proposal put forward in ref .
@xcite , and as such is not pursued : instead , the aim of the above discussion is to show the incorrect use of an analogy with a single scalar field pure kinetic k - essence model . finally , the constraints on nmc models obtained in ref .
@xcite are derived from the imposition that the speed of sound is lesser than that of light , @xmath90 , due to the usual argument for causality .
however , following ref .
@xcite , one may instead defend that obtaining a faster - than - light propagation of perturbations for a particular matter species does not invalidate the underlying theory , but instead shows that the adopted matter content is unphysical ( in the context of the particular model under scrutiny ) .
thus , the conclusions of ref .
@xcite refer to the rather unnatural assumption of nmc k - essence and does not have any bearing on models for dark matter based on nmc baryonic matter .
furthermore , this prohibition of a superluminal speed of sound is not as obvious as it may appear at first sight : indeed , lorentz invariant theories are allowed with @xmath91 , since the speed of sound can then act as a causality - imposing upper limit for the velocity albeit at the cost of possibly ill - defined microscopic properties of matter , and with the caveat that @xmath92 is then fundamentally unbounded itself @xcite . a more thorough analysis of the lorentzian signature of the spacetime manifold relies on global hyperbolicity as a sufficient requirement for causality .
while this translates into the condition @xmath93 in a static spacetime , non - trivial backgrounds can suitably evade this inequality , thus rendering the constraints obtained in ref .
@xcite devoid of meaning ( see ref .
@xcite for an extended discussion in the context of k - essence and ref .
@xcite for a survey of the issue in alternative gravity models ) .
in this work we discuss the viability of nmc models , and argue that the recent claims of pathological behaviour of non - minimally coupled models are not conclusive , due to the improper comparison between the latter and k - essence : our argument is based upon four major , independent points : \1 .
the use of results from single scalar field purely kinetic k - essence implies that the background cosmological dynamics ( namely the current accelerated expansion of the universe , _ i.e. _ dark energy ) is described by the latter or , as ref .
@xcite argues , from a nmc scalar field with @xmath94 : as such scenario has never been put forward ( and is highly unnatural to begin with ) , but the existing proposals rely instead on nmc baryonic matter ( a much more plausible assumption ) @xcite , the comparison with k - essence essentially does not refer to the considered models .
the conformal transformation used to arrive at the equivalent action relies on a conformal transformation from the jordan to the einstein frame , which implies that the lagrangian density of matter should be constructed with a physical metric @xmath95 related to @xmath30 : this enables a more complex dependency of the putative k - essence function @xmath96 on @xmath30 , which should not be dismissed .
even if this interaction between the purely kinetic lagrangian density @xmath97 and the additional degree of freedom @xmath31 could somehow be disregarded , the effect of the canonical term of the latter leads to a speed of sound for curvature perturbations different from the single field k - essence form argued in @xcite again highlighting the relevance of @xmath31 .
finally , the use of the constraint @xmath98 is not completely clear or universally accepted , as the structure of spacetime in an evolving universe implies that the usual causality argument established in minkowski spacetime may no longer be valid . given the above , we believe that the analysis of tamanini and koivisto @xcite is unfounded , as it presents an incorrect analogy with k - essence which , even if correct , would not apply to models in which there is only baryonic matter .
we naturally concede that the interesting questions raised in that work prompt further investigation namely how the analysis of the propagation of perturbations in nmc cosmological models can lead to further constraints on the model , and how an nmc is related to causality and the global structure of spacetime .
the authors would like to thank f. s. n. lobo for fruitful discussions , and the referee for his / her valuable remarks and criticism .
the work of o.b . and j.p .
was partially supported by fct ( fundao para a cincia e a tecnologia , portugal ) under the project ptdc / fis/111362/2009 .
s. capozziello , v. cardone , s. carloni and a. troisi , , 292 ( 2004 ) ; j. mbelek , , 761 ( 2004 ) ; s. capozziello , v. cardone and a. troisi , , 1423 ( 2007 ) ; s. capozziello , a. stabile and a. troisi , , 104019 ( 2007 ) ; t. chiba , t. l. smith and a. l. erickcek , , 124014 ( 2007 ) . a. nunez and s. solganik , hep - th/0403159 ; , 189 ( 2005 ) ; a. de felice , m. hindmarsh and m. trodden , , 005 ( 2006 ) ; t. biswas , e. gerwick , t. koivisto and a. mazumdar , , 031101 ( 2012 ) ; t. s. koivisto and n. tamanini , , 104030 ( 2013 ) ; t. biswas , t. koivisto and a. mazumdar , arxiv:1302.0532 [ gr - qc ] . | in this work we explore the viability of nonminimally coupled matter - curvature gravity theories , namely the conditions required for the absence of tachyon instabilities and ghost degrees of freedom .
we contrast our finds with recent claims of a pathological behaviour of this class of models , which resorted to , in our view , an incorrect analogy with @xmath0-essence . |
the electron density in 2-dimensional electron gases ( 2degs ) at oxide interfaces can be very effectively modulated by electric field effect @xcite .
such possibility has been successfully exploited to tune the properties of the laalo@xmath0/srtio@xmath0 ( lao / sto ) system and of its variants , e.g. for spanning the superconducting @xcite magnetic @xcite , and spin - orbital @xcite phase diagrams .
the effect of an applied electric field on the transport properties has been typically interpreted in terms of a standard field effect transistor model , in which charge is transferred between the 2d electron gas ( 2deg ) and the gate electrode as in a capacitor charge / discharge process .
further measurements highlighted a quite complex response of the transport properties of the 2deg as a function of gate voltage .
bell et al . ,
@xcite and more recently biscaras et al .
@xcite demonstrated that anomalous effects , well beyond the capacitive charge - discharge of a field - effect device , occur during and after the application of a positive back - gate ( bg ) voltage
. such bg bias perturbs in fact the properties of the 2deg , bringing it into a metastable state with a conductivity reduction up to about a factor two .
the effect is qualitatively interpreted in @xcite in terms of charges flowing away from the quantum well during a positive gate polarization . beyond this , no attempt of microscopic and electrostatic analysis has been proposed so far and no method to switch the system back from the metastable to the pristine state has been found . in this paper ,
the persistent effect of the first positive polarization is on three suitably chosen characteristic samples with different initial carrier density is analyzed and discussed .
we show that a colossal " switch with a resistive ratio @xmath1 can be obtained in a specific class of oxide interfaces ( represented by sample * c * in table [ tab1 ] ) .
the nature of the metastable highly resistive state is addressed in detail , exploiting a strong analogy with the retention state of floating - gate ( fg ) mosfet device .
we argue that the formation of such insulating retention state is related to a negative space charge region formed by electrons that are localized in shallow traps in proximity of the interface , that electrostatically gate the quantum well .
we show that , in full analogy with the case of the semiconducting fg - based eprom devices , exposure to light can be employed to detrap the electrons destabilizing the metastable retention state .
illumination causes the instantaneous collapse of the electron system back to the metallic ground state , with a resistivity drop exceeding in sample * c * four orders of magnitude . [ cols="^,^,^,^,^,^,^,^ " , ] on the base of the self - consistent calculation of the electrostatic potential reported in @xcite , it was shown that the energy defined as @xmath2 in fig .
[ fig3]b should be zero in absence of a population of localized charges @xmath3 near the interface .
this observation is a quite straightforward consequence of the one - dimensional poisson equation .
this is readily recognized by observing that the cb profile in sto , outside the 2deg ( i.e. in absence of a space charge ) , can only have a constant slope , fixed by the external gating @xmath4 .
we will show in the following that @xmath5 is pretty small in the pristine samples with respect to the amount of charge @xmath6 that is transferred by field effect .
we will also show that , due to the above described difference between the band diagrams reported in fig .
[ fig3]a and fig .
[ fig3]b , and to the presence of localized acceptor states near the interface , the correct analogous of our system is given not by the mos capacitor shown in fig .
[ fig3]a , but by the floating - gate mosfet device described in fig .
[ fig4]a , c , e. in appendix [ sec : model ] we describe the behavior of each sample during the quarter - cycles , within the general framework set above , resorting to a number of very reasonable and straightforward assumptions .
furthermore , a semi - quantitative evaluation of all the above described scds will be provided for each sample in each step of the @xmath7 waveform . as shown in fig .
[ fig4]a , c the functioning mechanism of fg mosfets is based on the possibility to transfer carriers from the channel to the fg states ( fig . [ fig4]c ) through the thin separating oxide barrier @xmath8 , under positive bias . the electric field caused by
the negatively charged fg keeps the mosfet in an off state , the so - called retention state , even when the control gate is unbiased ( fig . [
it is easily anticipated that carriers bubbling out of the 2deg can have the same effect , if they can accumulate in localized acceptor states ( las ) present in the sto barrier , such to create a space - charge region .
we postulate the presence of such las now ( their origin will be discussed later ) and depict them graphically in figs .
[ fig4]b , d , f as a series of shallow potential wells .
their energy depth can be estimated of the order of @xmath9 ev .
let s now exploit the analogy with fg devices to address our data .
_ phase 1 _ + during phase 1 , no bias is applied .
the sheet resistance value is 550@xmath10/@xmath11 .
+ _ phase 2 _ + as soon as @xmath4 is increased to 200v , this fpp pushes an extra sheet density of capacitively induced charges @xmath6 , adding to the native sheet density of the 2deg @xmath12 , towards the qw through the control electronics .
@xmath6 can be estimated of the order of 2 - 3 x 10@xmath13 @xmath14 at 200v , 20k@xcite .
the flow of @xmath6 towards the qw during the positive ramp causes the resistance of the sample first to decrease and then to saturate quickly , as soon as , at a critical gate bias , the sheet density of states in the qw ( @xmath15 ) is filled .
all extra charges induced above such bias , not finding available states in the qw , are pulled away from the interface by the electric field present in the barrier .
[ fig4]d shows that such excess electrons , trapped in the las , act similarly to the trapped electrons in a fg device , persistently changing the resistance of the sample .
in the course of this second step , at constant positive @xmath4 , a small damped upturn of the resistance of about 5% takes place .
as previously observed in @xcite , the qw , previously brought to complete filling by the fpp ( fig .
[ fig4]d ) , slowly looses the electrons lying , energywise , in proximity of the qw edge , for a thickness " of the order of k@xmath16 t . such electrons drift along the potential gradient and are eventually also trapped in the las .
when a similar analysis as the one performed in @xcite was applied , similar values for the characteristic escape time @xmath17 were found .
+ _ phase 3 _ + when the bg bias is set back to zero , our interface is left in a high resistive retention state , in analogy to a fg mosfet device ( fig . [ fig4]e ) .
such state is metastable , since at @xmath4=0 the empty quantum well states lye again below the trap states filled during the fpp .
the sample is unable to recover the pristine ground state , unless special procedures are applied , as discussed later in this work .
the sample resistance is increased by about a factor two .
the trapped electrons perturb the electrostatics of the system .
trapped electrons effectively gate the 2deg by inducing in it an opposite positive surface charge density , as can be inferred by charge conservation . the semi - quantitative model reported in appendix a provides , for sample a in the third step , the estimate , @xmath18 .
+ _ phase 4 _ + the sample resistance is further increased by about an extra factor two . in this configuration ,
the electrostatic effect of the negatively charged fg and bg add to each other , doubling the effect of the a single negative gate .
hence we have : @xmath19 .
+ again , this result will be more formally derived in appendix a. + during the successive cycles , no further irreversible effect takes place .
once the excess electrons have been trimmed " during the fpp , in fact , the fermi level never exceeds again the edge of the qw , even under positive bias . during the following steps
, the electrons are reversibly transferred between the back gate and the 2deg through the control electronics without further transfer to / from the las .
let s now analyze one by one the four quarter - cycles of the bg square wave on samples b and c , and in particular the fpp effect . for sample * b * at 20k ( upper panel ) modulated by the applied gate voltage ( v@xmath20)(lower panel ) .
the dark - gray area indicates the instrument compliance limit .
the dark - gray area indicates the instrument compliance limit .
the brighter area indicates the temporal region under analysis in this work , for sake of clarity split in four phases according with the v@xmath20 modulation and labeled with the blue circles .
( b ) a detailed view of the temporal behavior in the region indicated by the red circle in the upper panel of ( a).,scaledwidth=48.0% ] during phase 1 , at 20k , the sample has an r@xmath21 .
the decrease of sheet resistance taking place during phase 2 ( fig .
[ fig5 ] ) is pretty small .
this suggests that the qw is , as in the previous case , very close to complete filling in the virgin state .
once again , the second zero - bias state , in retention mode ( phase 3 ) , has a resistance that is about a factor 2 higher than the virgin state . a quantitative difference with the case of sample * a * is found in the negative polarization state ( phase 4 ) .
the negative bg bias here completely depletes the qw well of mobile charges , bringing the sample to an insulating state . for sample *
c * at 20k ( upper panel ) modulated by the applied gate voltage ( v@xmath20)(lower panel ) .
the dark - gray area indicates the instrument compliance limit .
the dark - gray area indicates the instrument compliance limit .
the brighter area indicates the temporal region under analysis in this work , for sake of clarity split in four phases according with the v@xmath20 modulation and labeled with the blue circles .
( b ) a detailed view of the temporal behavior in the region indicated by the red circle in the upper panel of ( a).,scaledwidth=48.0% ] this is under many aspects the most interesting sample ( fig . [ fig6 ] ) . even for this low - carrier density sample ,
the qw is close to complete filling before the fpp , as shown by the very minor resistance decrease at positive bias .
this suggests a general statement , that samples with lower carrier densities are not characterized by an emptier qw with respect to higher carrier density samples , but by a smaller " qw , hosting a smaller number of bound states .
we observe that in the retention state , the system is insulating both at zero and at negative @xmath4 .
this means that the fpp persistently changes a highly metallic sample into an insulator , which can only temporarily recover its pristine metallic character during a positive bias . as a peculiar feature of sample * c * , at difference to other samples , at every successive positive pulse of the waveform ( phase 2 " ) , the same peculiar exponentially damped resistivity increase typical of the fpp takes place
this implies that the qw is filled until the edge , and above , each time a pulse of capacitively induced electrons with surface density @xmath6 is provided .
let s now switch to the second part of this paper , by addressing the issue of how can the retention state be erased .
this is not only crucial for future hypothetical applications .
it is a condition for performing further experimental work , since a phenomenon that can be seen only once in the life of a sample would certainly not allow one any systematic investigation .
biscaras et al.@xcite reported a full recovery of the pristine state during a warm - up of the sample towards room temperature .
our experiments lead to slightly different results , suggesting that the behavior reported in @xcite , at least , does not represent the general case .
( t ) curves for the sample * c * in different conditions : pristine state , retention state and during the erasing procedure .
( a ) the erasing procedure is performed applying a positive bg voltage of 200v and heating up the system to room temperature . in the inset the temporal evolution of the r@xmath22 once at 300k .
( b ) a different erasing procedure is carried out by illuminating for few seconds the sample with a 625 nm led at @xmath23 .
the system is then warmed up to room temperature .
, scaledwidth=48.0% ] a number of subsequent experiments are summarized in fig .
[ fig7]a for sample c. the sample was first cooled to 20k ( black r(t ) curve ) .
the fpp performed at 20k brought the sample into its immeasurably resistive retention state , as shown by the blue r(t ) curve . during the warm - up phase , the sample attempted " to recover the pristine resistance , through a number of steps taking place at specific activation temperatures . in spite of such partial recovery , it preserved a room temperature resistance an order of magnitude higher than the pristine value . in order to recover the pristine state , we attempted a field - annealing procedure , by keeping the sample under positive voltage for about 30 minutes at room temperature .
a slow resistance decrease was obtained , as shown in the inset . at the end of such procedure
, the samples indeed recovered the pristine properties , as shown by the red r(t ) curve .
the same procedure successfully applied to all samples , although a much longer time ( above 10 h ) was needed to reset sample * a * and sample * b * , as shown in @xcite . on the base of the above - described model of the oxide interface retention state
, we assume that the positive voltage , coupled to the room temperature thermal energy , induces a slow creep of trapped charges from the fg states into the control gate , across the whole thickness of the substrate .
this procedure is reminiscent of the reset procedure of current electrically erasable programmable read - only memories ( e@xmath24prom ) devices , also based on fg mosfet technology . at difference to our samples , characterized by a macroscopically thick control gate barrier , e@xmath24proms can be reset instantaneously by a control gate polarization cycle , thanks to their ultra - thin oxide gate layer .
an instantaneous resetting method for our samples was finally mutuated from previous fg eprom technology(fig .
[ fig7]b ) . for eproms , which did not allow for an instantaneous electrical reset , uv light was used to bring the device out of the metastable retention state and back to the pristine state .
the high energy of uv photons was necessary for the electrons in the fg to exceed the oxide barrier energy , as sketched in fig .
[ fig8 ] . due to the small energy barriers confining our trapped electrons in their localized states , as suggested by our sketch reported in fig .
[ fig8]b , visible red light ( 625 nm ) was employed in our experiment at @xmath23 .
the sample was first switched to the highly resistive state by a positive polarization pulse and then switched back by a total photon dose of about @xmath25 photons / cm@xmath24 delivered by a led source in about 3 - 5 seconds ( fig . [ fig8]b ) .
this colossal and unprecedented , light - induced insulator - to - metal transition results in a sudden switch with a resistive ratio @xmath1 at 20k .
the physics of with phenomenon and the intensity of the effect are totally new with respect to standard photoconductivity effects , @xcite as demonstrated in detail in @xcite .
after illumination , a r(t ) curve was taken during warm - up .
the sample reproduced the curve of the pristine samples , except for two minor features : a photoconductivity effect in the first part of the warm - up curve , and a typical feature at about 200 k , often seen in lao / sto warm - up plots @xcite .
the reported data and their interpretation in terms of band diagrams pose strong limits to our capability to truly probe the enriched portion of the lao / sto phase diagrams by a back - gate bias .
data collected on three different samples show that the sheet conductance in the nominal carrier - enriched " state is equal ( sample * c * ) or smaller ( samples * a * , * b * ) than in the pristine state , for all cycles of our bg square waveform , except for the first . even during the first cycle ,
i.e. soon after the fpp , the sheet conductivity exceeds the pristine one by no more that 10% ( sample * a * ) or by much less ( samples * b * and * c * ) .
this is related to the fact that the available , empty , confined 2deg states , marked in gray in fig .
[ fig3]b , are very few in pristine samples with respect to the occupied ones . as shown in @xcite and discussed in appendix a , such states should not exist at all in an ideal " system where no localized states are present .
the presence of a space charge accumulated in the las will allow instead for a non - null second derivative of the potential profile , thus effectively increasing the quantum well edge @xmath27 shown in fig .
the space charge in the las , induced during the fpp , plays therefore an essential role in making the 2deg tunable in enrichment mode during the following cycles .
these considerations also suggest that a limited occupancy of the las might be present also in pristine samples and play a role , by inducing the formation of a weak space charge , in determining the actual lao / sto band diagram and thus in confining the 2deg .
our data support a scenario in which shallow localized acceptor states are present in proximity of the 2deg , as theoretically suggested in @xcite , and in which a dynamical exchange of carriers between the mobile and trapped states controllably takes place under field effect and light .
any experiment trying to identify the chemical origin of the trap states would go beyond the scope of this work .
our present data only allow to state that the trap states lie in the @xmath28 gap and that their depth is in - between several times room temperature kt values ( i.e. about 0.1 ev ) and our red light photon energy ( i.e. 2 ev ) .
interestingly , the presence of trap states with energy lying about 0.3 ev below the sto conduction band has been demonstrated in sto single crystals by optical spectroscopy@xcite .
the origin of such trap states might be possibly related to the presence of cation vacancies , or of interstitial hydrogen@xcite .
we have investigated the low - temperature transport properties of back - gated 2degs at oxide interfaces .
our batch of analyzed samples included metallic amorphous - lgo / sto interfaces , a system for which the formation of an interfacial 2deg was never proved before .
we show that our samples can be brought by a positive polarization pulse into a persistent insulating state .
the magnitude of the effect is analyzed as a function of the initial carrier density . in low - carrier - density samples , a non - volatile four - order - of - magnitude resistive switching to the insulating state
is obtained .
this work introduces a new element of complexity to our understanding of lao / sto - like interfaces , addressing them as systems in which two distinct electron reservoirs , mobile carriers and localized charges , interact electrostatically , while dynamically exchanging electrons under external perturbations .
we provided evidence that the colossal resistive switching taking place under field effect can be understood in terms of the formation of a retention state that mimics the analogous state of floating - gate mosfet devices . by suggesting the use of light as a resetting technique ,
this analogy finally lead us to the fabrication of a non - volatile 3-terminal device prototype , where a 4-order - of - magnitude resistance switch can be induced by a voltage pulse and erased by a pulse of light .
the behavior of the three different samples in each of the successive voltage steps can be properly described by a simple electrostatic model , which partially neglects the quantum effects of the backgating potential on the confining qw ( as shown by scopigno et al .
@xcite ) but still provides a clear and reliable understanding of the physical system , consistent with the results already found by biscaras et al . @xcite .
the model is based on few reasonable and straightforward assumptions : 1 . in agreement with the qualitative analysis reported above
, the charge - neutrality conditions for our samples should be written as @xmath15= -@xmath29=@xmath12+@xmath5=@xmath12+@xmath3 .
such equation implies , in particular , that in case all localized acceptor states were empty , all the bound states in the quantum well would be occupied , realizing the condition @xmath12=@xmath15=-@xmath29 . in case a nonzero + @xmath3 density due the occupation of the localized states is present , an equivalent amount of empty bound states @xmath5=@xmath3 , is also present at the interface .
no field - effect - induced @xmath30 doping of the interface is possible .
therefore the scds @xmath12 and @xmath3 can never become positive .
as already first observed by @xcite , the fpp effect , i.e. the loss of carriers from the qw carriers taking place for a high enough positive @xmath4 is an irreversible process at the low temperatures ( 20k ) of our experiment .
reformulated in the language of this paper , this means that the system is unable to transfer charge back from @xmath3 to @xmath12 , even when backgate - induced changes of the potential profile would make this situation energetically more favorable ; an exception to the former assumption takes place , nevertheless , if the system , during a backgate voltage step , has no way to satisfy charge neutrality conditions by keeping @xmath3 constant . in this case
only , due to the large electrostatic monopole energy involved , carriers will be detrapped .
@xmath3 will be consequently decreased to such a level that allows the sample neutrality condition to be satisfied .
this assumption will be crucial in understanding the response of sample c , as shown in the following ; 4 .
the parallel - plate capacitor model , used to describe the variations of scd , breaks down when the 2deg is completely depleted .
in such a case , in fact , no potential can be defined on the upper face of the sto substrate .
the system misses one plate " and can simply not be modeled anymore as a capacitor . during any attempt to push the sample in further depletion by an extra negative voltage ,
the sample reacts to the negative pulse as an open circuit , and remains basically unperturbed , with no further charge transfer taking place . in order to describe the different scds of our different samples ( * a
* , * b * , * c * ) in the relevant quarter - cycles ( 1 , ... , 6 ) we will adopt a couple of indexes .
we start by writing the charge neutrality conditions of the overall system ( including donor states , interface , las and back - gate ) that , for a generic sample x = a , b or c , in each of the four quarter - cycles of the @xmath7 period , read : @xmath31 @xmath32 @xmath33 @xmath34 the sample is initially in a pristine state , which depends both on the deposition conditions and , possibly , on previous interaction with the environment ( e.g. , exposure to light ) .
field - effect measurement estimates of the low - temperature carrier densities of our sample were employed @xcite .
as shown in @xcite on @xmath35 , field - effect carrier density measurements can yield good agreement with hall measurement also in oxides .
we estimate : where the p index stay for pristine .
the low - temperature hall carrier density of sample c , our most important sample , was directly measured as a double - check @xcite , yielding an independently estimation @xmath37 in perfect agreement with previous data .
the numerical considerations reported below should not be intended as a claim of real quantitive estimation of the carrier densities of the different samples during the cycle .
still , they clearly allow to capture the phenomenological response of the three samples to the successive phases of the voltage cycle .
@xmath5 can be estimated by first measuring the maximum positive @xmath4 value leading to a reversible r(@xmath4 ) plot , similarly to the analysis made in @xcite , and then estimating the induced scd according to @xcite . in the case of sample * a * , the filling of the qw took place for @xmath4=40v , similarly to the sample analyzed by @xcite . in the case of sample * c * ,
the filling of the qw took place at about @xmath4=15v .
this allows us making the order of magnitude estimates on the base of the former assumptions and definitions , let s now address the response of all the samples in all the first 6 steps , or quarter - cycles .
the response of the system is determined , for any sample @xmath41 and any step @xmath42 , by the occupation of the qw , @xmath43 and of the las , @xmath44 .
when a positive back - gate voltage is applied , a scd = @xmath6 will be induced at the interface .
the electrons will first fill the empty states , adding a scd = @xmath5 into the quantum well and will then populate the las
. the new scd will be : when @xmath4 is set back to 0v , the electrons trapped in the las are not released back to the 2deg , according to assumption iii , and the population @xmath3 remains unchanged . by combining this assumption with eq .
[ eq3 ] , we obtain the following equations , valid only for sample * a * and * b * , for which @xmath50 in reasonable agreement with the previous estimates for @xmath52 , @xmath53 , @xmath54 , @xmath55 and @xmath6 , the sheet resistance of samples * a * and * b * increases by about a factor 2 . for the case of sample * c * , @xmath56 .
equation [ eq10a ] can not apply , since it would imply a positive charge density at the interface , in disagreement with assumption iv .
it is deduced that , during the decrease of @xmath4 from 200v to 0 , the 2deg will be completely depleted at some point ( @xmath57 ) . below such point , according to assumption ii , the carrier density can not change sign and remains constant , while the las will be depleted , according to assumption iv , to preserve charge neutrality ( eq .
[ eq3 ] ) . as a result : sample * a * has increased its resistivity by a factor 4 with respect to the pristine state , but it is still conducting .
this suggests that @xmath62 2 @xmath63 , in reasonable agreement with the previous estimates for @xmath52 , @xmath54 , and @xmath6 . sample * b * is in an intermediate situation . during the beginning of the negative pulse , is it fully conducting . at @xmath65
it is insulating , while @xmath66 still applies . on the base of the previous estimates for @xmath53 , and @xmath6 .
it can be deduced that the sample becomes insulating at some voltage value @xmath4=@xmath67 pretty close to @xmath65 . at such voltage ,
the condition -@xmath68
= @xmath69 - 2@xmath70 , applies . for increasing @xmath4 ,
the charges remain frozen , similarly to the previous case of sample * c*. the conditions are : when @xmath4 is set back to 0v , step 5 , @xmath3 remains unchanged , maintaining the same value as in step 4 .
[ eq1 ] and from the previous values @xmath72 , we can deduce the state of the different samples : * sample * a * + @xmath73 + similarly to the third quarter - cycle .
* sample * b * + the population of the las remains the same , and under eq .
[ eq1 ] the same conditions as sample * a * apply : + @xmath74 * sample * c * + it remains in depletion mode , in the same conditions as in step 3 and 4 .
+ @xmath75 at the end of this 5@xmath59 quarter - cycle we see that the resistivity measured in all the three samples does not recover the value attained in the 3@xmath49 one , despite our model predicts @xmath12 and @xmath3 the be the same .
this is particularly relevant for samples * a * and * b*. we expect this difference to be due to a residual polarization of sto substrate , in agreement with @xcite .
such hypothesis that will be explored in a future work and can not be described by our present simplified model . during the next positive polarization ( step 6 and the following second quarter - cycles : 10 , etc . )
samples * a * and * b * show two unexpected behaviors , not shared by sample * c * : the r(t ) curve has both a different ( positive ) second derivative and a much higher value than the one in step 2 . the first phenomenon can be explained observing that the charges induced by the positive @xmath4 in samples * a * and * b * will grossly fill the available state in the qw , since @xmath76 = @xmath77 @xmath78 , but are not expected to flow out of the qw . on the other hand ,
the latter phenomenon is not easily described within our framework and its origin will not be discussed in this work .
sample * c * instead replicates at every successive positive polarization the fpp effect shown in fig.6b , i.e. , an exponentially damped recovery towards higher resistance values .
it is in fact recognized that in the case of sample * c * , @xmath79 = @xmath80 @xmath81 @xmath78 , therefore the induced charges will exceed , at every time , the available states in the qw , thus replicating , at every time , the fpp effect .
32ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1126/science.1131091 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1038/nmat4491 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.104.126802 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevlett.103.226802 [ * * , ( ) ] http://dx.doi.org/10.1038/srep06788 [ * * , ( ) ] link:\doibase 10.1038/nature02308 [ * * , ( ) ] link:\doibase 10.1038/srep00758 [ * * , ( ) ] link:\doibase 10.1063/1.3496440 [ * * , ( ) ] link:\doibase 10.1021/nl201821j [ * * , ( ) ] @noop link:\doibase 10.1038/srep14506 [ * * ( ) , 10.1038/srep14506 ] link:\doibase 10.1103/physrevlett.109.246807 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevlett.116.026804 [ * * , ( ) ] @noop * * , ( ) http://link.aps.org/doi/10.1103/physrevb.83.205405 [ * * , ( ) ] link:\doibase 10.1038/ncomms6118 [ * * , ( ) ] link:\doibase 10.1038/nmat4303 [ * * , ( ) ] link:\doibase 10.1103/physrevb.86.045127 [ * * , ( ) ] link:\doibase 10.1063/1.4921092 [ * * , ( ) ] link:\doibase 10.1038/srep08393 [ * * , ( ) ] link:\doibase 10.1002/adom.201300150 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevb.76.155111 [ * * , ( ) ] link:\doibase 10.1103/physrevb.87.125110 [ * * , ( ) ] link:\doibase 10.1021/jp510647b [ * * , ( ) ] link:\doibase 10.1103/physrevlett.113.156801 [ * * , ( ) ]
@noop link:\doibase 10.1103/physrevb.73.165103 [ * * , ( ) ] @noop * * ( ) | the field - effect - induced modulation of transport properties of 2-dimensional electron gases residing at the laalo@xmath0/srtio@xmath0 and lagao@xmath0/srtio@xmath0 interfaces has been investigated in a back - gate configuration . both samples with crystalline and with amorphous overlayers
have been considered .
we show that the nave " standard scenario , in which the back electrode and the 2-dimensional electron gas are simply modeled as capacitor plates , dramatically fails in describing the observed phenomenology .
anomalies appearing after the first low - temperature application of a positive gate bias , and causing a non - volatile perturbation of sample properties , are observed in all our samples .
such anomalies are shown to drive low - carrier density samples to a persistent insulating state .
recovery of the pristine metallic state can be either obtained by a long room - temperature field annealing , or , instantaneously , by a relatively modest dose of visible - range photons .
illumination causes a sudden collapse of the electron system back to the metallic ground state , with a resistivity drop exceeding four orders of magnitude .
the data are discussed and interpreted on the base of the analogy with floating - gate mosfet devices , which sheds a new light on the effects of back - gating on oxide - based 2-dimensional electron gases .
a more formal approach , allowing for a semi - quantitative estimate of the relevant surface carrier densities for different samples and under different back - gate voltages , is proposed in the appendix . |
mrk 766 is a rather bright ( typical flux of about one millicrab ) narrow line seyfert 1 galaxy at redshift 0.012 , showing complex and puzzling x ray variability ( molendi , maccacaro & schaeidt 1993 ; molendi & maccacaro 1994 ; page et al .
1999 ) on time scales as short as @xmath01000 sec ( leighly et al .
the 1992 june rosat pspc observations , discussed by molendi & maccacaro ( 1994 ) , indicated large amplitude variations in the softest part of the spectrum ( 0.10.9 kev ) , while the hardest part ( 0.9 - 2 kev ) was consistent with being constant .
the authors suggested accretion disc emission as the origin of the variable component .
recently , page et al . ( 1999 ) analysed spectral variability in rosat pspc data collected from 1991 to 1994 , reaching different conclusions : a highly variable ( on time scales as short as 5000 s ) hard power law component , and a constant soft excess . the asca
rosat observation reported by leighly et al .
( 1996 ) shows a complex and variable spectrum , with the primary power law photon index changing from 1.6 to 2 within the observation , and variable warm absorber too .
the soft excess variability was anti correlated with that of the hard component .
a narrow iron k@xmath1 line was detected only when the source was in a high flux level .
we observed mrk 766 with bepposax to take advantage of the wide energy band of the satellite and try to better understand the origin of the spectral complexity in this source .
the paper is organized as follows : sec .
2 describes the observation and data reduction ; sec . 3 presents the temporal , and sec.4 the spectral , analysis ; the results are compared with the asca
rosat ones in sec . 5 , discussed in sec . 6 and summarized in sec .
bepposax ( boella et al . 1997 ) observed mrk 766 on 17 - 18 may , 1997 as part of a core program devoted to a spectral survey of bright seyfert 1 galaxies .
the exposure time was 79 ksec for the mecs , the hpgspc and the pds and 42 ksec for the lecs .
we discuss here data from the lecs , mecs and pds only , because the source is too faint for the hpgspc to be profitably used .
spectra and light curves were extracted from the imaging instruments within circles of 8 ( lecs ) and 4 ( mecs ) radii centred on the source .
background spectra taken in the same regions from blank sky observations have been used for subtraction . regarding the pds , background subtracted spectra automatically generated at the sax scientific data centre ( sdc ) have been used .
fixed rise time thresholds have been adopted .
averaged count rates are : [email protected] ( lecs , 0.1 - 10 kev ) ; [email protected] ( mecs , 1.5 - 10 kev , 2 units ) ; [email protected] ( pds , 15 - 200 kev ) .
the 210 kev mean flux ( obtained with model 5 in tables 1 and 2 ) is 2.05@xmath310@xmath4 erg @xmath5 s@xmath6 . in the field of view of the pds
there are two possible confusing sources , both of them bl lac objects . at an angular distance of about 20 arcminutes
there is on 325 .
the source is clearly visible in the mecs , with a count rate of about 1/20 that of mrk 766 .
the faintness of the source and its off axis position do not permit a detailed spectral analysis , but the mecs spectrum appears to be rather steep , and therefore its contribution to the pds is likely to be very small .
more serious is the case of the other source , 2a 1219 + 305 ( pg 1218 + 304 ) , a member of the piccinotti sample ( piccinotti et al . 1982 ) with a mean 210 kev flux of 3@xmath7 erg @xmath5 s@xmath6 and a 2.5 @xmath8 detection with batse ( malizia et al .
the source is about 44 arcminutes away from mrk 766 , i.e. outside the mecs field of view , and it is therefore not possible to estimate its contribution to the pds , which may be substantial .
for this reason we preferred to exclude the pds from the spectral fitting , using the data _ a posteriori _ only as an upper limit to the hard x ray emission of mrk 766 .
the lecs ( 0.110 kev ) and mecs ( 1.5 - 10 kev ) light curves are shown in fig . [ lc ] . the source varies as much as a factor @xmath02 on time scales of a few thousands of seconds . to search for spectral variability
, we have also plotted the ( 1.5 - 4)/(0.1 - 1.5 ) lecs hardness ratio , and the ( 4 - 10)/(1.5 - 4 ) mecs hardness ratio .
clearly , there is significant spectral variability , especially in the second part of the observation . to investigate the energy dependence of this variability , we calculated the normalized excess variance ( nev ; see nandra et al .
1997 ) ( fig . [ nev ] ) for different energy intervals , using a 128 s binning .
the source is variable at all energies , but the amplitude is the highest around 2 kev . the pds light curve and the pds / mecs ratio are also shown in fig .
the light curve is very variable , the background subtracted pds count rates sometimes going down to values consistent with zero .
the light curve is different from those of the lecs and mecs , suggesting either a strong spectral variability or a significant contamination by the nearby bl lac , 2a 1219 + 305 .
the spectral analysis was performed on the lecs and mecs instruments only , because of the possible confusion problems in the pds outlined above .
the pds data have been used only to check _ a posteriori _ that a lecs+mecs best fit spectrum , when extrapolated to higher energies , would not exceed the observed data .
all fits described below have been performed with the xspec software package .
quoted errors refer to 90% confident level for two interesting parameters ( i.e. @xmath9=4.61 ) . because of the significant spectral variability , it is not wise to use the spectrum averaged over the whole observation . on the other hand ,
it is important to collect as many photons as possible to search for spectral details . as a trade off we have divided the observation in two parts , and extracted spectra from : a ) the beginning of the observation till 8@xmath10 sec ( see fig .
[ lc ] ) , because within this interval variations in the hardness ratio , especially in the mecs , are not dramatic ; b ) from this time till the end of the observation . as it will be seen in the following ( tables 1 and 2 ) , no models will give a fully acceptable fit , the null hypothesis probability being 0.12 at most .
this is likely due to the spectral variability discussed in the previous section , as our time selection makes this problem alleviated but not completely cured .
let us start discussing the first half of the observation .
a simple power law absorbed by the galactic column ( i.e. 1.77@xmath310@xmath11 @xmath5 ; elvis , wilkes & lockman 1989 ) is a very poor fit to the data ( see fig .
[ fit_po ] and table 1 , model 1 ) , with an unacceptable reduced @xmath12 of 1.70 . from the figure ,
it is clear that most of the contribution to the @xmath12 comes from the softest part of the spectrum , i.e. below @xmath01 kev , while no evidence for typical seyfert 1 components like the iron line or the reflection continuum is present .
a deficit of counts above 7 kev is also apparent . leaving the column density free to vary ( model 2 ) , the quality of the fit significantly improves , but the fit is still unacceptable ( @xmath13=1.50 ) .
inspection of the residuals suggests the presence of warm absorption features .
we therefore added the absori model ( model 3 in table 1 ) , obtaining a significant improvement in the quality of the fit .
the warm absorber material has been assumed to be ionized by the observed power law .
the best fit value of the ionization parameter , @xmath14 , is about 240 erg cm s@xmath6 ; with this value , the most prominent edge is the o viii one .
the column density of the warm absorber is @xmath08@xmath310@xmath15 @xmath5 .
a further inspection of the residuals reveals an excess around 0.6 kev , which can be well fitted ( model 4 ) by a narrow gaussian line at 0.59(@xmath20.05 ) kev ( source rest frame ) , with an equivalent width of 68(@xmath16 ) ev .
the statistical significance of this line is 97.3% ( f - test ) .
the line energy is consistent with he like oxygen ( 0.57 kev ) ; a fit with the line fixed at this energy , plus a line at 0.65 kev ( h like oxygen ) does not provide a statistical improvement ; the best fit ews are 60 and 30 ev , respectively .
the value of @xmath14 is now lower .
the best fit ionization structures of the absorber and the emitter are somewhat different ( the most relevant ion being o vii in the emitter , and o viii in the absorber ) , but still consistent each other within the ( fairly large ) errors .
finally , a relativistic line ( diskline model in xspec ) fits the line equally well as a narrow gaussian line .
fixing the inner radius at 6 @xmath17 , i.e. the last stable orbit for a schwarzschild black hole , we obtain an outer radius of thousands of @xmath17 and an inclination angle consistent with zero ( i.e. face on disc ) .
while the above fit is satisfactory from a statistical point of view , a deficit of counts above 7 kev is still apparent . adding an absorption edge to model 4 the fit actually improves , the best fitting parameters being an edge energy of 7.55 kev , and @xmath18=0.26 .
this edge may be related to reflection from circumnuclear matter , and therefore we added a compton reflection component .
we allowed the matter to be ionized ( pexriv model in xspec ) , both because the edge is at energies larger than 7.1 kev , the value of neutral iron , and because the lack of any observed iron k@xmath1 line ( see next paragraph ) is only possible , in presence of a compton reflection component , if the iron is mildly ionized and therefore resonant destruction possible ( e.g. matt , fabian & ross 1993 , 1996 ) .
the results are reported in table 1 as model 5 .
interestingly , the value of the ionization parameter of the reflecting matter , even if poorly determined , is consistent with what is expected in order to have the iron line significantly destroyed .
it is worth noticing that such an ionized disc could also account for the observed oxygen emission line , even if the equivalent width is larger than expected by a factor of a few , pointing to a possible oxygen overabundance .
moreover , the best fit parameters of the warm absorber material are now such that the most important oxygen ion is the helium like .
it is therefore possible that both the accretion disc and the warm absorber contribute to the line flux .
alternatively , the iron edge may be due to absorbing material .
we therefore added a second ionized absorber ( model 6 ) , instead of the reflector .
the fit is as good as the one with the reflector .
this second absorber results to be more thick and ionized than the other one ( but the parameters are loosely constrained )
. it may be interesting to note that a similar double
absorber solution has been found in the ( broad line ) seyfert 1 ngc 3516 ( costantini et al . 2000 ) .
leighly et al . (
1996 ) detected a narrow 6.4 kev iron line with an equivalent width of about 100 ev in the asca spectrum of mrk 766 , but only when the source was in a high state .
we searched for an iron line in the bepposax data , but could find only upper limits .
the upper limits to a narrow iron line are 110 , 43 and 36 ev if the line energy is fixed to 6.4 kev ( neutral iron ) , 6.7 kev ( h like iron ) or 6.97 kev ( h like iron ) , respectively .
therefore , our result is marginally consistent with the line found by leighly et al .
of course , the upper limits would increase if the line is broad . as explained in the previous section ,
the lack of an observable line is still consistent with the reflection scenario , because the best fit ionization state is such to have the line destroyed by resonant re absorption . as shown in fig .
[ extr_pds ] , the extrapolation of model 5 to the pds data is rather good , which may be an indication ( but unfortunately not a proof ) that the contribution to the pds count rate from 2a 1219 + 305 is low .
also the extrapolation of model 6 does not exceed the observed pds data .
we then searched for a soft excess ( modeled as a black body ) , a component which has been already observed in this source ( see sec . 1 ) , and
whose presence may be expected by analogy with other narrow line seyfert 1 galaxies ( e.g. boller , brandt & fink 1996 ) . the fit with the soft excess instead of the warm absorber ( other components as in model 5 ) gives a worse @xmath19 ( 1.23/119 ; best fit temperature of about 30 ev ) , while the addition of the black body to model 5 does not provide a significant improvement ( @xmath19=1.13/117 ; best fit temperature of 200 ev ) .
we therefore conclude that such a component is not required by the data , and confirm that the warm absorber is real , and not an artifact of not having included the soft excess .
however , it should be noted that while the fits presented in table 1 have been obtained with a column density of the cold absorber in excess of the galactic value , the column density turns out to be close to the galactic one when the black body component is included .
therefore , the presence of the soft excess remains rather ambiguous .
it is worth noting that the bepposax power law index is significantly steeper than the one observed by asca , while the average 2 - 10 kev flux is about the same in the two observations , making any soft excess in the bepposax data difficult to observe .
cccccccccc & & & & & & & & & # & n@xmath20 & @xmath21 & r / n@xmath22 & @xmath14 & n@xmath23 & @xmath24 & @xmath25 & e.w .
& @xmath12/d.o.f . &
( 10@xmath11 @xmath5 ) & & & ( erg cm s@xmath6 ) & ( 10@xmath11 @xmath5 ) & ( erg cm s@xmath6 ) & ( kev ) & ( ev ) & @xmath13 & & & & & & & & & & & & & & & & & & 1 & 1.77 & 2.01@xmath26 & & & & & & & 215/126 & & & & & & & & & 1.70 2 & 2.29@xmath27 & 2.07@xmath28 & & & & & & & 188/125 & & & & & & & & & 1.50 3 & 3.06@xmath29 & 2.17@xmath30 & & & 80@xmath31 & 240@xmath32 & & & 155/123 & & & & & & & & & 1.26 4 & 2.77@xmath33 & 2.13@xmath34 & & & 43@xmath35 & 120@xmath36 & 0.59@xmath37 & 68@xmath16 & 146/121 & & & & & & & & & 1.20 5 & 4.29@xmath38 & 2.25@xmath39 & 0.84@xmath40 & 94@xmath41 & 38@xmath42 & 78@xmath43 & 0.61@xmath44 & 57@xmath45 & 137/119 & & & & & & & & & 1.15 6 & 3.19@xmath46 & 2.21@xmath47 & 380@xmath48 & 1040@xmath49 & 28@xmath50 & 23@xmath51 & 0.63@xmath52 & 51@xmath53 & 139/119 & & & & & & & & & 1.17 & & & & & & & & & [ cols="^ " , ]
the behaviour of mrk 766 during the bepposax observation is rather different than during the asca
rosat observation , performed on december 1993 and discussed by leighly et al .
the source was at an average flux level very similar in the two observations . while in both cases
there is significant flux and spectral variability on time scales of thousands of seconds , the origin of the spectral variability appears to be different . in the asca
rosat observation , it is driven by significant changes in the spectral index of the power law , while in our observation it seems to be mostly related to changes in the warm absorber properties . indeed , in our observation the power law index is consistent with being constant , but at a value significantly steeper ( i.e. @xmath542.2 ) than those found by leighly et al .
( @xmath541.62 ) .
moreover , in the asca rosat observations a soft excess is strongly required by the data , while in our observation this component may be absent .
( it is worth noting than even in the asca
rosat observations the soft excess is marginally visible when the power law is the steepest , because the soft component is almost constant during the observation ; it may therefore be not surprising that in our observation the soft component is hardly detectable ) .
a further difference is the presence , in the asca data and during the high flux phase , of an iron line with @xmath0100 ev equivalent width , while in the bepposax data such a line is not visible ( but the upper limit on the equivalent width is marginally consistent with the abovementioned value ) . to summarize , in both observations the source underwent large and complex variability , but the pattern and origin of the variations seem to be very different in the two cases .
while the narrow line seyfert 1 galaxy mrk 766 shares with other members of its class the large amplitude and short time scale variability , it does not share ( at least during the observation described here ) another characteristic which is common in narrow line seyfert 1 galaxies , i.e. a prominent soft x
ray emission .
the power law component is , actually , somewhat steeper than the typical value for classical seyfert 1 galaxies , but the soft excess is modest , if present at all .
the warm absorber is also typical of classical , ` broad lines ' seyfert 1s , being dominated by oxygen and neon edges .
no @xmath01 kev feature ( besides those related to the warm absorber and cured by the inclusion of this component ) is present , differently from what observed in other sources of this class ( e.g. leighly et al .
1997 ; fiore et al . 1998 ; leighly 1999 ; vaughan et al .
this is not surprising , as these features have been observed so far only in very steep spectrum sources , a fact naturally explained if these features are due to a blend of resonant absorption lines , mainly from iron l shell ( nicastro , fiore & matt 1999 ) .
the column density and ionized parameter of the warm absorber appear to have been changed between the two halves of the observations , in agreement with the fact that the variability is the largest around 1 kev .
a so dramatic change of the column density , however , is unplausible , and it may be an artifact of having used pure photoionization equilibrium , single zone models , while in reality it is possible that : the absorber is out of equilibrium most of the time ( not surprisingly , given the flux variability ; see nicastro et al .
1999 for non equilibrium models ) ; there is some contribution from collisional ionization ; the absorbing medium is geometrically or physically complex .
similar to other narrow line seyfert 1s ( ton s 180 : comastri et al .
1998 , turner , george & nandra 1998 ; ark 564 : vaughan et al .
1999b ) is , instead , the possible presence of reprocessing from ionized matter .
ionized accretion discs are expected when @xmath55 is high ( ross & fabian 1992 ; matt , fabian & ross 1993 ) , and accretion rates close to the eddington one have been indeed invoked to explain the nls1 phenomenon ( e.g. pounds , done & osborne 1995 ; ) . the ( mild ) ionization of the reprecessor may also account for the lack of any observed iron line ( matt et al .
1993 ; 1996 ) , as the line photons may be resonantly trapped and then destroyed by the auger effect .
reflection from a mildly ionized disc may also explain at least part of the o vii emission line observed in the first half of the observation ( in the second half only an upper limit is obtained , which however is consistent , within the error , with the value measured in the first half ) .
in this case , the line is expected to be broadened by relativistic effects ; the quality of our data is not good enough to distinguish between a narrow and a relativistic line , provided that the inclination angle of the disc is low .
an oxygen line possibly from an ionized disc was observed by piro et al .
( 1997 ) in the asca data of e1615 + 061 , while the same line observed by bepposax in the spectrum of ngc 5548 ( nicastro et al .
2000 ) clearly originates from outflowing material , as demonstrated by the chandra / letg observation ( kaastra et al .
2000 ) . in mrk766
, it is possible that both the accretion disc and the warm absorber contribute to the observed emission , as the best fit ionization structure for both materials includes a significant fraction of o vii .
high energy resolution observations are needed to definitely settle this issue .
the main results of the bepposax observation of mrk 766 may be summarized as follows : @xmath56 the spectrum of mrk 766 is well fitted by an absorbed power law ( with cold absorption in excess of the galactic one ) , a warm absorber , an oxygen line and a further component ( mainly accounting for the iron edge ) which may be either a second and more ionized warm absorber or a mildly ionized disc .
a soft excess is not required by the data , but its inclusion makes the cold absorber closer to the galactic value .
@xmath56 the spectral variability during the observation may be entirely due to variations in the properties of the warm absorber(s ) , even if the behaviour of the absorber is complex and no obvious correlation with the flux is apparent .
@xmath56 the power law is significantly steeper than during the asca
rosat observation ( leighly et al . 1996 ) .
the spectral variability pattern is also very different , indicating that we have observed the source in a different state .
we thank the referee , thomas boller , for useful comments .
we acknowledge the bepposax sdc team for providing pre
processed event files and for their constant support in data reduction .
gm , gcp , ff and fn acknowledge financial support from asi and murst ( grant cofin980232 ) . this research has made use of the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration | the narrow line seyfert 1 galaxy mrk 766 has been observed by bepposax on may 1997 .
the source was fairly variable , both in flux and spectrum , during the observation .
the variability is the largest around 2 kev , possibly due to variations in the warm absorber properties .
an oxygen line is clearly detected , at least in the first part of the observation .
an edge around @xmath07.5 kev , coupled with the lack of any detectable iron line , suggests either reprocessing of the primary x
rays by a mildly ionized disc or absorption by a further , thicker and more ionized material . |
the majority of the information that we have obtained during the last years about the magnetism of the sun and other astrophysical objects is based on the analysis of the polarization of the light .
the polarization state of a light beam is typically described , using the stokes formalism , by the so - called stokes vector : @xmath0 where @xmath1 refers to the total intensity of the beam , @xmath2 describes its circular polarization properties while @xmath3 and @xmath4 are used to described linear polarization .
when the light beam passes through optical elements or , in general , any medium that modifies its polarization properties , the emergent stokes vector can be related to the input one by the following linear relation : @xmath5 the @xmath6 matrix , @xmath7 , is the so - called mueller matrix and it can be used to unambiguously represent any optically passive medium .
it is important to note that these matrices have to obey certain properties so that they have physical meaning @xcite .
the analysis of polarization of a given beam is usually carried out with the aid of modulation schemes .
this is a requisite in the short - wavelength ( optical ) domain because no detectors that present the required polarization sensitivity are still available .
furthermore , the majority of the detectors that are used in the optical are only sensitive to the total amount of light or , in other words , to the total intensity given by stokes @xmath1 .
this is not the case in the long - wavelength ( microwave ) domain , where it is relatively easy to build detectors that analyze directly the polarization properties of the light beam .
modulation schemes have been built to overcome the difficulties when measuring the polarization state of light beams in the short - wavelength domain .
any modulation scheme consists of a train of optical devices that produce a known modification of the input beam so that the observed intensity in the detector is a linear combination of the elements of the input stokes vector .
carrying out several of these measurements , each one with a different combination of optical devices , it is possible to _ infer _ the input stokes vector using a demodulation procedure . assuming that the mueller matrix of the @xmath8-th combination of optical devices is given by @xmath9 , the detected intensity in each case is given by : @xmath10_{00}s^\mathrm{in}_0+[m_j]_{01}s^\mathrm{in}_1+[m_j]_{02}s^\mathrm{in}_2+[m_j]_{03}s^\mathrm{in}_3,\ ] ] where we have used the standard notation @xmath11 for the stokes vector , which will be used extensively in the rest of the paper . by putting several measurements together , the modulation scheme can be expressed as the following linear system : @xmath12 where each row of the matrix @xmath13 is built with the first rows of the different mueller matrices of the different combinations of optical elements used in the modulation scheme .
therefore , the matrix @xmath13 has dimensions @xmath14 , with @xmath15 the number of measurements .
note that , neither the vector @xmath16 is a stokes vector nor the matrix @xmath13 is a mueller matrix .
the vector @xmath16 consists of different intensity measurements , while the matrix @xmath13 does not fulfill , in general , the conditions described by @xcite and @xcite to be considered a mueller matrix .
several different modulation techniques have been developed .
one of the most widespread methods is to use a temporal modulation with a quarter - wave plate and a linear polarizer , with the fast axis of the quarter - wave plate being set to preselected angles @xcite .
this approach has the advantage that the different polarizations states are observed on the same pixel . however , modulation has to be carried out very fast in order not to be limited by atmospheric fluctuations .
another possibility is to use spatial modulation in which the measurements are carried out simultaneously by using a beamsplitter instead of a linear polarizer @xcite .
the advantage is that atmospheric fluctuations are avoided but potential problems appear because the optical path and the detector response are not the same for all the polarization states .
the most advanced polarimeters now use a combination of both approaches in the so - called spatio - temporal modulation @xcite . starting from eq .
( [ eq : modulation ] ) , the input stokes vector can be obtained by solving the previous linear system of equations , so that : @xmath17 where the matrix @xmath18 is the inverse ( or moore - penrose pseudoinverse in the general case that more than four measurements are carried out to infer the input stokes vector ) of the matrix @xmath13 and has dimensions @xmath19 . in order to measure the four stokes parameters ,
the minimum number of measurements is 4 .
in such a case , the matrix @xmath18 is unique and can be calculated as the standard inverse of @xmath13 . however , it is also possible to carry out more measurements than stokes parameters . in this case , the linear system of eq .
( [ eq : modulation ] ) is overdetermined and , in general , no solution that satisfies simultaneously all the equations exist .
recently , it has been shown that it is still possible to choose a solution if we seek for the one that maximizes the efficiency of the modulation - demodulation scheme @xcite .
such efficiency for each stokes parameter , represented with the vector , is obtained from the demodulation matrix @xcite : @xmath20 in this case , it can be shown that the optimal demodulation matrix is given by the moore - penrose pseudoinverse @xcite : @xmath21 the pseudoinverse can also be found with the singular value decomposition of the modulation matrix . note that
, while @xmath22 in the general case , @xmath23 is only valid when @xmath24 ( @xmath25 is the identity matrix ) .
we present a theoretical calculation of how the error propagates in the demodulation process .
this is an important issue that has been partially treated in the past , either considering how the uncertainty in the measurement of @xmath16 propagates to the demodulated stokes vector when the modulation matrix is perfectly known @xcite or taking into account uncertainties in the knowledge of the modulation matrix , though the full covariance matrix is not obtained @xcite .
many papers also deal with how error is propagated in non - ideal mueller matrix polarimeters @xcite while others consider the optimization of such polarimeters @xcite .
in practical situations , the modulation matrix has to be measured once the optics and the modulation scheme have been defined @xcite . as a consequence ,
one expects the elements of the modulation matrix to have some uncertainty produced by the measurement procedure .
additionally , uncertainties can also be found when the rotation of the fast axis of the retarders or of the linear polarizers has some repeatability problems @xcite . in this case , although the average modulation matrix can be known with great precision , random deviations from the average modulation matrix can occur .
these random deviations induce a modification on the modulation scheme that directly affects the inferred polarization properties of the input light beam . for these reasons ,
it is fundamental to characterize the error propagation through the demodulation process and how they affect the measured stokes profiles . in general , since the matrix inversion ( or pseudoinverse ) is a nonlinear process , one should expect to find a full covariance matrix for the elements of the matrix @xmath18 , with non - zero non - diagonal elements .
these non - diagonal elements account for the statistical correlations that appear between elements of the demodulation matrix after the inversion process .
more importantly , a full covariance matrix has to be expected even in the case that the elements of the modulation matrix are not correlated ( i.e. , @xmath26 , with @xmath27 the kronecker delta ) .
note that a diagonal covariance matrix for the matrix @xmath13 comes out when the measurements of the elements of the modulation matrix are statistically independent , something that can be assumed in some cases if the measurements of the elements of the matrix are carried out with proper calibration .
however , this is an issue that has to be carefully analyzed for each case .
the full covariance matrix for the demodulation matrix can be obtained analytically following the standard error propagation formulae .
the starting point is to consider that elements of the demodulation matrix are known nonlinear functions of the elements of the modulation matrix : @xmath28 the most general error propagation formula for such operation reads @xcite : @xmath29 this covariance matrix gives information about the variances of the elements of the demodulation matrix ( diagonal elements ) , as well as covariances between those elements ( non - diagonal elements ) .
note that , for a @xmath19 matrix , the covariance matrix is @xmath30 . if the covariance matrix of the modulation matrix is diagonal ( i.e. , @xmath31 , with @xmath27 the kronecker delta ) , the previous equation simplifies to : @xmath32 as already presented above , this turns out to be a good approximation if the elements of the modulation matrix are carefully measured .
however , our approach is general and can cope with non - diagonal covariance matrices as will be shown below .
the derivatives that appear in eqs .
( [ eq : covarianced_general ] ) and ( [ eq : covarianced_diagonal ] ) can be calculated , in the most general case , starting from eq .
( [ eq : optimal_demodulation ] ) .
however , we distinguish the cases of a square modulation matrix from the general case , because the analytical procedure can be largely simplified . in any case
, we will show that the general expressions are equivalent to those of the particular case of a square matrix @xmath13 .
although it is possible to calculate directly the derivatives from eq .
( [ eq : optimal_demodulation ] ) , it is more advantageous to use the fact that the matrix product between @xmath13 and @xmath18 commute , so that @xmath33 . in this case , it is easy to show that @xcite : @xmath34 therefore , the full covariance matrix of the matrix elements of the demodulation matrix can be calculated by substituting eq .
( [ eq : derivative_square ] ) into eq .
( [ eq : covarianced_general ] ) : @xmath35 in the particular case of a diagonal covariance matrix for @xmath13 , the previous equation simplifies to : @xmath36 when @xmath15 is larger than four , it is not possible to follow the previous approach because @xmath13 and @xmath18 do not commute .
however , inserting the intermediate matrix @xmath37 , using the fact that @xmath38 and that the matrix @xmath39 is always invertible @xcite , the derivative can be expressed , after some algebra , as : @xmath40 the derivative of the elements of the @xmath41 matrix with respect to the elements of @xmath13 can be calculated using the chain rule : @xmath42 the definition of the matrix @xmath39 allows us to obtain the following derivative easily : @xmath43 the derivative of the elements of @xmath41 with respect to the elements of @xmath13 can also be calculated easily from eq .
( [ eq : derivative_square ] ) because @xmath39 is a square matrix , that commutes with its inverse .
therefore : @xmath44 substituting eq .
( [ eq : der_ainvfinal ] ) into eq .
( [ eq : deriv_d_o ] ) and after some algebra , we end up with the following expression for the derivative : @xmath45 this expression reduces to eq .
( [ eq : derivative_square ] ) if @xmath13 is square because the product of @xmath18 and @xmath13 commutes , so that : @xmath46 after substitution in eq .
( [ eq : derivative_nonsquare ] ) , we recover eq .
( [ eq : derivative_square ] ) .
the final expression for the covariance of the demodulation matrix is obtained by plugging eq .
( [ eq : derivative_nonsquare ] ) into eq .
( [ eq : covarianced_general ] ) [ or eq . ( [ eq : covarianced_diagonal ] ) ] : @xmath47 \nonumber \\ & \times & \left [ a_{ak}^{-1 } \delta_{bk } - a_{ak}^{-1 } \sum_m d_{mb } o_{km } - d_{ak } d_{lb } \right].\end{aligned}\ ] ] when both the modulation matrix @xmath13 and the measured intensities @xmath16 present uncertainties , the resulting uncertainty in the demodulated stokes vector has contribution coming from both origins . using standard error propagation formulae applied to eq .
( [ eq : demodulation ] ) , the covariance matrix can be written as : @xmath48 the first contribution takes into account the uncertainty in the knowledge of the modulation matrix while the second contribution takes into account the uncertainty in the measurement of the intensities arriving to the detector @xcite . in the field of optimization of polarimeters , a suitable norm of this very last term ( often neglecting the non - diagonal covariances in the vector @xmath49 that we include here ) is the chosen one to measure the efficiency of a polarimeter @xcite .
particularizing to our problem and calculating the partial derivatives , we end up with : @xmath50 the covariance matrix @xmath51 is typically assumed to be diagonal , since no statistical dependence is assumed between consecutive intensity measurements in the detector , so that @xmath52 .
since this is not generally the case for the quantities @xmath53 , a non - zero ( in general ) covariance between different stokes profiles appears as a consequence of some degree of correlation in the demodulation process . in principle , it is possible to diagonalize the covariance matrix given by eq .
( [ eq : covariance_intensity ] ) to obtain its principal components .
such principal components define the directions along which the correlation between different stokes parameters is minimized . since the principal components define a new reference system , it is possible to rotate the original reference system in which the four - dimensional stokes vector is defined .
this rotation minimizes in a statistical sense the cross - talk .
this could be of interest when synthetic calculations have to be compared with polarimetric observations . in this case , in order to minimize the effect of cross - talk , it could be advantageous to compare the observations with projections along the eigenvectors of the synthetic calculations .
a quantity that is also affected by uncertainties in the knowledge of @xmath13 is the efficiency defined in eq .
( [ eq : efficiency ] ) . applying the error propagation formula ,
we get : @xmath54 if we assume that the covariance matrix of the elements of the modulation matrix is diagonal , we get : @xmath55 the derivatives can be calculated using the definition of the efficiency and the chain rule : @xmath56 where @xmath57 plugging this expression into eq .
( [ eq : derivative_epsilon ] ) , we can simplify it to read : @xmath58 the derivative @xmath59 is obtained from eq .
( [ eq : derivative_nonsquare ] ) in the general case and from eq .
( [ eq : derivative_square ] ) in the case of a square modulation matrix .
according to the previous results , when the covariance matrix of the modulation matrix is diagonal and the variance is the same for all the elements , all the covariance matrices calculated in this section are proportional to the variance of the elements of @xmath13 . as a consequence ,
reducing an order of magnitude the uncertainty in the elements of the modulation matrix reduces an order of magnitude the uncertainty in the demodulation matrix and in the inferred stokes vector .
the previous analytical approach assumes that the error propagates normally .
this is typically a good approximation when the modulation matrix is far from singular . in other words ,
the rows of the modulation matrix have to be as linearly independent as possible .
however , since the inversion is a nonlinear process , it would be possible to obtain distributions in the demodulation matrix ( and as a consequence , in the inferred stokes parameters ) that are far from gaussians .
in order to verify this issue and also to test that the previous analytical approach gives the correct answer , we have carried out a monte carlo experiment . for simplicity , we assume that the measurement uncertainty is the same for all the matrix elements of the modulation matrix and that it is characterized by the variance @xmath60 , so that the full variance matrix is given by : @xmath61 this selection makes the presentation of the results easier , but the previous approach is general and can cope with non - diagonal covariance matrices .
the monte carlo experiment , given an initial modulation matrix , generates @xmath15 instances of such a matrix with added uncertainties : @xmath62 where @xmath63 is a normally distributed random constant with zero mean and unit variance . for each modulation matrix
, we obtain the optimal demodulation matrix following eq .
( [ eq : optimal_demodulation ] ) .
finally , the statistical properties of the demodulation matrix are characterized by the full covariance matrix , given by : @xmath64 - \mathrm{e}[d_{ij } ] \mathrm{e}[d_{kl}],\ ] ] where @xmath65 $ ] stands for the expected value of the @xmath66 random variable .
in such a monte carlo approach , we estimate it using the sample mean , so that @xmath65 = \langle x \rangle$ ] .
the previous analytical approach is applied to several modulation matrices obtained from the literature and that represent different typical modulation schemes .
we analyze the uncertainties in the demodulation matrix and in the inferred stokes parameters using the error propagation formulation presented in the previous section . for illustration purposes ,
the results are also accompanied in some cases with plots obtained using the monte carlo approach .
one of the reasons for that is to demonstrate whether the resulting distributions can be correctly assumed to be gaussians .
the monte carlo results have been obtained using @xmath67 .
we assume that the variance that characterizes the uncertainty in the measurement of the modulation matrix is @xmath68 ( the standard deviation is equal to @xmath69 ) , which is a reasonable value for a standard calibration of the modulation .
the first example is representative of a square modulation matrix in which four measurements are carried out to infer the four stokes parameters .
the matrix is that used by the tenerife infrared polarimeter ( tip ; @xcite ) as presented in @xcite .
although this matrix is not the experimental one for tip , its presents the desired structure and can be used to gain some insight on the properties of the error propagation .
the chosen modulation matrix is : @xmath70\ ] ] the optimal demodulation matrix , obtained from the application of eq .
( [ eq : optimal_demodulation ] ) , is just the inverse of @xmath13 : @xmath71\ ] ] as shown in fig .
[ fig : d00_tip ] , the distribution of the elements of the demodulation matrix are gaussians whose standard deviation can be obtained from the diagonal elements of the matrix defined in eq .
( [ eq : covariance_square ] ) : @xmath72 \times 10^{-3}.\ ] ] although we only present the diagonal elements of the covariance matrix , note that the covariance matrix is a full matrix in which all the elements are different from zero . as a direct consequence of the uncertainties in the modulation matrix
, the efficiency of the proposed scheme presents uncertainties .
they can be calculated using eq .
( [ eq : covariance_efficiency_diagonal ] ) , so that the average value and their standard deviations are : @xmath73 where non - diagonal elements in @xmath74 are non - zero , but typically much smaller than the elements in the diagonal . now we investigate the propagation of uncertainties to the inferred stokes parameters . to this end , we choose an initial light beam defined by @xmath75 .
such stokes vector is representative of what one would observe in relatively low magnetic flux regions of the solar surface .
the application of eqs .
( [ eq : covariance_intensity ] ) and ( [ eq : covariance_square ] ) gives the following covariance matrix : @xmath76 \times 10^{-6 } \label{eq : cov_stokes_tip}\ ] ] this result has been obtained assuming that there is no uncertainty in the measurement of the intensity arriving to the detector , so that @xmath77 . in so doing , we isolate the effect of the modulation on the inferred stokes vectors .
the results of the monte carlo simulation are shown in fig .
[ fig : inferred_stokes_tip ] , demonstrating that the values are quasi - normally distributed around the original value with a dispersion that is given in each panel of the plot . note that the standard deviations are large , almost of the order of the standard deviation of the precision in the measurement of the modulation matrix .
note also that the largest value of @xmath78 are on the diagonal of the matrix , but sizable non - diagonal elements also appear , indicating a certain degree of correlation between the inferred stokes parameters .
for instance , the value @xmath79 indicates that when the inferred stokes @xmath1 ( @xmath80 ) increases , the inferred stokes @xmath2 ( @xmath81 ) tends to systematically decrease .
this can be understood as a cross - talk between all the stokes parameters induced by the demodulation process due to the special structure of the modulation matrix .
the two - dimensional distribution of the inferred stokes @xmath1 and @xmath2 is shown in the left panel of fig .
[ fig : correlationiv ] . in specific cases ,
part of this cross - talk can be corrected for easily based on physical arguments .
this happens , for instance , when observing the stokes profiles induced by the zeeman effect in magnetized regions of the solar surface with the aid of spectropolarimeters .
in such a case , one can assume that stokes @xmath3 , @xmath4 and @xmath2 are zero away from the spectral line and carry out a correction of the cross - talk from stokes @xmath1 to stokes @xmath3 , @xmath4 and @xmath2 .
the diagonalization of the @xmath82 matrix gives the following eigenvector matrix ordered in row format : @xmath83 .
\label{eq : eigenvectors_tip}\ ] ] each row represents the linear combination of stokes parameters that one infers due to the induced cross - talk contamination . note that the weight of one of the stokes parameters is larger in each row , but the contamination from the other parameters is still large , a consequence of the chosen modulation matrix .
we present here results for a typical non - square modulation matrix .
we have chosen the one belonging to the advanced stokes polarimeter ( asp ; @xcite ) as presented in @xcite . as it happens for the tip matrix
, this is probably not the exact matrix used in the polarimeter but is representative of what happens in a scheme in which more than four measurements are used to obtain the four stokes parameters .
the modulation matrix we choose is : @xmath84.\ ] ] after calculating the demodulation matrix and its covariance matrix , we can also calculate the diagonal of the covariance matrix for the efficiency which , transformed into standard deviations , give : @xmath85 and the full covariance matrix for the inferred stokes parameters using @xmath75 as the input stokes vector : @xmath86 \times 10^{-6}. \label{eq : cov_stokes_asp}\ ] ] note that the elements of the diagonal of this covariance matrix are smaller than for the case of only four measurements ( except for the case of @xmath87 ) , probably induced by the larger number of measurements carried out .
furthermore , it is important to point out that this modulation scheme induces no correlations between @xmath88 ( stokes @xmath4 ) and @xmath81 ( stokes @xmath2 ) and any other stokes parameter . on the contrary
, there is a large correlation between stokes @xmath1 and stokes @xmath3 , something that can be clearly seen in the right panel of fig .
[ fig : correlationiv ] .
the diagonalization of the previous matrix gives the following eigenvector matrix ( in row order ) : @xmath89 .
\label{eq : eigenvectors_asp}\ ] ] the demodulation process has induced a cross - talk between stokes @xmath1 and @xmath3 which can be represented by a rotation of @xmath90 .
it is instructive to present results for the following ideal modulation matrix : @xmath91.\ ] ] the efficiency in this case is the maximum that can be reached , while the diagonal of the covariance matrix for the efficiency given as standard deviations is : @xmath92 and the full covariance matrix for the inferred stokes parameters using @xmath75 as the input stokes vector is : @xmath93 \times 10^{-6}. \label{eq : cov_stokes_ideal}\ ] ] the covariance matrix is diagonal with equal uncertainties in stokes @xmath3 , @xmath4 and @xmath2 .
these values are smaller than for the asp example except for stokes @xmath2 , where the uncertainty for the ideal modulation matrix is slightly larger .
more important is the fact that , since the covariance matrix is diagonal , no correlation is found between the inferred stokes parameters , so that no residual cross - talk induced by the modulation is induced in the demodulation .
we have presented analytical expressions for the calculation of the propagation of errors in the demodulation process when the modulation matrix is not known with infinite precision .
this can happen when the modulation system has not been calibrated with enough precision or when the repeatability of the modulation system induces uncertainties in the modulation scheme .
the formulae that we have presented allows polarimeter designers to calculate the errors in the demodulation matrix , the efficiency of the modulation scheme and the inferred stokes parameters .
they are simple to calculate and require only the knowledge of the modulation matrix , together with its covariance matrix .
we have pointed out the fact that , in general , since matrix inversion ( or moore - penrose pseudoinversion ) is a nonlinear operation , non - zero non - diagonal covariances have to be expected in the demodulation matrix even if such non - diagonal correlations are not present in the modulation matrix .
this has the important consequence of generating spurious correlations ( cross - talk ) between the inferred stokes parameters .
we calculate the induced cross - talk by diagonalizing the covariance matrix .
the matrix of eigenvectors represent the reference system in which the cross - talk is minimized .
we have illustrated these points with three different modulation matrices representing three different ways of measuring the polarization state of light beams .
each method has its own advantages and disadvantages .
it is up to the polarimeter designer to choose the modulation scheme depending on the desired precision .
we hope that the formulae present in this paper are of interest for improving the quality of modulation based polarimeters .
routines in fortran 90 and idl for the calculation of the formulae presented in the paper can be obtained after an e - mail request to the authors of this paper .
finantial support by the spanish ministry of education and science through project aya2007 - 63881 is gratefully acknowledged .
v. martnez pillet , m. collados , j. snchez almeida , v. gonzlez , a. cruz - lopez , a. manescau , e. joven , e. paez , j. diaz , o. feeney , v. snchez , g. scharmer , and d. soltau , `` lpsp & tip : full stokes polarimeters for the canary islands observatories , '' in `` high resolution solar physics : theory , observations , and techniques , '' , vol .
183 of _ astronomical society of the pacific conference series _ , t. r. rimmele , k. s. balasubramaniam , and r. r. radick , eds . ( 1999 ) ,
183 of _ astronomical society of the pacific conference series _ , p. 264 .
m. collados , `` high resolution spectropolarimetry and magnetography , '' in `` third advances in solar physics euroconference : magnetic fields and oscillations , '' , vol .
184 of _ astronomical society of the pacific conference series _ ,
b. schmieder , a. hofmann , and j. staude , eds .
( 1999 ) , vol .
184 of _ astronomical society of the pacific conference series _ , p.
3 . e. l. dereniak , d. s. sabatke , m. r. locke , m. r. descour , w. c. sweatt , j. p. garca , d. sass , t. hamilton , s. a. kemme , and g. s. phipps , `` design and optimization of a complete stokes polarimeter for the mwir , '' osti ( 2000 ) .
d. f. elmore , b. w. lites , s. tomczyk , a. p. skumanich , r. b. dunn , j. a. schuenke , k. v. streander , t. w. leach , c. w. chambellan , and h. k. hull , `` the advanced stokes polarimeter - a new instrument for solar magnetic field research , '' in `` polarization analysis and measurement , '' , d. h. goldstein and r. a. chipman , eds .
( 1992 ) , p. 22 . | the polarization analysis of the light is typically carried out using modulation schemes .
the light of unknown polarization state is passed through a set of known modulation optics and a detector is used to measure the total intensity passing the system .
the modulation optics is modified several times and , with the aid of such several measurements , the unknown polarization state of the light can be inferred . how to find the optimal demodulation process has been investigated in the past .
however , since the modulation matrix has to be measured for a given instrument and the optical elements can present problems of repeatability , some uncertainty is present in the elements of the modulation matrix and/or covariances between these elements .
we analyze in detail this issue , presenting analytical formulae for calculating the covariance matrix produced by the propagation of such uncertainties on the demodulation matrix , on the inferred stokes parameters and on the efficiency of the modulation process .
we demonstrate that , even if the covariance matrix of the modulation matrix is diagonal , the covariance matrix of the demodulation matrix is , in general , non - diagonal because matrix inversion is a nonlinear operation .
this propagates through the demodulation process and induces correlations on the inferred stokes parameters . |
the physical boundary between a human being and his or her environment is the skin , but in the space of behaviours the same boundary is less strict . taking decisions , we are not completely selfish ; we are to some extent bound by the social norms .
the way of enforcing norms varies between a direct control and a deep internalization ; in the latter case we treat our conformity to norms in the same way as our payoff .
norms create the society , where we are formed @xcite , and norms are modified by the society members .
as it was formulated by a leading polish psychiatrist antoni kpiski , to decide where to put limits of our own rebel is one of most difficult problem in human life @xcite . solving this problem in our individual scale emerges in the social scale as a time evolution of norms . to search for laws which rule
this process is a worthwhile challenge for the agent - based simulations .
+ a serious advance in this path was done by robert axelrod who formulated the norm game : an algorithm to simulate the conditions of persistence and fall of a social norm @xcite .
the simulations done by axelrod have been questioned @xcite , but his paper has been cited hundreds times and it triggered a cascade of research ; for a recent review of simulations of norms see @xcite .
further , the subject of norms overlaps with the theory of cooperation ; to cooperate is an example of a social norm .
an overview on the latter might provide insight into current trends ; still the research in this field seems to be at its intensively rising stage @xcite .
as norms are beliefs , there is also some overlap with the simulations of opinion dynamics ; for a review of sociophysical simulations on this matter see @xcite .
+ direct motivation of this research comes from statistical data on norm breaking .
perhaps most striking change we have seen deals with the data on divorces in portugal during the carnival revolution .
there , the number of divorces increased from 777 in 1974 to 7773 in 1977 @xcite .
the plots on crime in countries in central europe are more conventional . in fig .
1 we show the data on germany , poland , czech republic , austria and hungary @xcite . in accordance with warnings by eurostat ,
our aim is not to compare the amount of crimes in these countries , but rather to show the changes in some of them .
note that opinion shifts were classified into continuous and abrupt by michard and bouchaud in 2005 @xcite within a theory of imitation . in the axelrod model ,
the driving social mechanism is punishment ; the interaction inhibits the change rather than releases it . + in ng @xcite agents defect a norm with a given probability .
once an agent defects , other agents punish the defector , also with some probability .
the defection is gratified with some payoff , but those who are punished lose . also , those who punish incur some cost .
axelrod considered also a metagame : the possibility of punishment those who do not punish .
the overall success of an agent was measured by his income , but the agent himself - represented by his strategy - was not modified . after some number of games , the genetic algoritm was used to select strategies which yielded the best income .
as indicated in @xcite , this kind of modeling has an advantage to deal with the dynamics of the process .
we should add that neumann criticizes the approach of axelrod for disregarding the functional character of norms @xcite . + recently @xcite we developed a new realization of the axelrod model with two new ingredients .
first is that the model is freed from the payoff parameters ; what is left is just the influence of agents decisions on decisions of other agents .
further , the probabilities of decisions of individual agents ( to defect or not , to punish or not ) are not constant , but they are dynamically modified in each game they play .
second modification is less technical : once an agent decides to defect , his ability to punish in his future games vanishes , and his probability to defect ( boldness ) in future games is kept one until he is punished ; then it is multiplied by @xmath1 .
then , the constant @xmath2 describes the severity of the punishment . on the contrary ,
once an agent punishes , he will never defect the norm , and his probability to punish ( vengeance ) is set to @xmath3 , where @xmath4 is due to the punishment cost .
this vengeance can be further reduced if the agent punishes also in his future games . in this way
, a kind of social labeling takes place : first decision is irreversible , and the whole process can be seen as a social contagion @xcite . as a consequence , a sharp transition of the final boldness as dependent of the initial boldness is found @xcite .
the threshold value varies with the model parameters @xmath5 .
these results depend only quantitatively on the assumed topology of the social network , which determines the probability distribution of the number of punishers , i.e. of the node degree .
+ the aim of the present paper is twofold .
first , we are going to investigate the above mentioned sharp transition in the space of parameters @xmath5 . on the contrary to @xcite , here we are going to assume a given probability distribution of the initial boldness between agents .
this modification makes the calculation closer to a social reality , where different agents present different willingnesses to break the norm and to punish .
we should precise that in our model the parameters @xmath5 represent not the boldness and the vengeance , but rather the modifications of agents boldness and vengeance due to the decisions of other agents . then
, our model parameters describe the interactions between agents and not their actual states .
our goal here is to calculate the critical line on the plane @xmath5 between the final state all defect and nobody is punished and the final state nobody defects. this critical line is a kind of generalization of the critical concentration , calculated in the problems of directed percolation @xcite .
our second aim is to investigate the character of the transition when some biased modifications of the structure of the social network .
we use two kinds of these modifications : _
i ) _ agents ( guards ) at some amount of nodes always punish and never defect , _ ii ) _ agents ( sinners ) at some amount of nodes always defects and never punish . + in our model , agents are placed at nodes of the directed erds - renyi network .
this network is selected for its generic topology .
there is much work on the structures of social networks , mostly done by sociologists @xcite ; for an early list of references see @xcite . with the outburst of scale - free networks
a common opinion appeared that social networks are scale - free .
some of them can indeed be classified as scale - free , in particular those where direct face - to - face contact is not needed ; as citation networks or telephone - calls networks @xcite .
still , the actual structure varies from one social network to another @xcite , and often the network is simply too small to be classified to any type @xcite . to end ,
we have checked @xcite that the investigated threshold appears also in the scale - free growing networks , except the case when the direction of links is determined by the sequence of attaching new nodes .
last but not least , here we are not going to discuss the role of hubs , which could complicate the results .
+ the paper is organized as follows . in the next section
the model assumptions and the details of the calculations are listed . in section 3
we describe the numerical results .
these are two : the critical line on the plane of the parameters , and the transition dependence of the amount of guards / sinners .
section 4 is devoted to discussion of the results in the context of a recent classification of contagion processes @xcite , and of some statistical data on dynamics of crime , presented in fig .
the network size is @xmath6 nodes , the mean number of in - going links ( punishers ) is @xmath7 , and their distribution is poissonian . on the contrary to our former calculations
@xcite , here we assume a homogeneous distribution of the initial probability to defect the norm , i.e. the initial boldness @xmath8 , where @xmath9 is the node index .
as a rule , the initial vengeance @xmath10 , i.e. the initial probability of punishing , is @xmath11 .
this condition is not maintained during the simulation ; however , @xmath12 .
third option is to obey the norm and not punish , with the probability @xmath13 .
+ at each time step , an agent @xmath9 is selected and he breaks the norm with the probability equal to his boldness @xmath8 . if actually he does , his boldness @xmath8 is set to 1 and his vengeance , i.e. the probability of punishing - is set to 0
. then his neighbours @xmath14 are asked , one by one , if they punish @xmath9 .
if one of them punishes , the boldness of @xmath9 is multiplied by a factor @xmath15 , and the vengeance of the punisher @xmath16 is multiplied by @xmath17 .
the defector can be punished only by one neighbour . on the contrary , if a neighbour does not punish , his boldness is set to one and his vengeance is set to zero . in this way
, the process is accompanied by kind of social labeling @xcite : those who break the norm and those who refrain from punishing can not punish in their future games , and those who punish can not break the norm . + as a rule , we calculate the values of the parameters @xmath18 where the threshold appears , i.e. the final state changes from the bold state all defect to the vengeant state all punish. additionally , as a new variant of the game , some amount of sites of the network is selected randomly .
agents at these nodes got special roles of guards or sinners. if an agent is a guard , he always punishes and never defects ; sinners do the opposite , i.e. always defect and never punish .
the value and character of the threshold is observed against the ratio of the number of those special nodes to the whole population @xmath19 .
all these special nodes are either all guards or all sinners.
calculating the final boldness as dependent on @xmath2 and @xmath4 we observe a sharp change of the result , as in fig .
3 a , at some threshold values of @xmath4 and @xmath2 .
this means that we got a critical line @xmath20 in the plane of the parameters @xmath5 - see fig .
4 . this line divides the plane @xmath5 into two areas ; the plane can be treated as a phase diagram , then we can talk about two phases .
abobe the line , we have a bold phase , where @xmath4 is large ; there punishment costs too much and is not effective . below the line
, we have a vengeant phase where the cost of punishment is low .
then , everybody punishes and there is no interest in defection .
+ having added a small amount of sinners or guards to the system we observe that the threshold changes differently in these two cases .
basically the character of the threshold remains the same , just the threshold value is more susceptible to the admixture of guards than to the one of sinners. this can be seen at the positions of two additional curves in fig .
while adding of five percent of sinners apparently produces no effect , twice smaller admixture of guards shifts the critical line upwards , reducing the area of the bold phase .
when the number of modified nodes increases more , the character of the plot @xmath21 gradually changes from an abrupt to a more continuous one .
examples of these curves are shown in fig .
3 b. this change prevents us to investigate the critical line for higher concentration of special nodes ; the transition becomes fuzzy .
the crossover from the sharp to the fuzzy character of the transition between the bold phase and the vengeant phase , observed at the @xmath5 plane , fits into a recent classification of contagion processes @xcite . according to that scheme
, models of contagion processes can be divided into three classes : @xmath22 independent interaction models , @xmath23 stochastic threshold models , and @xmath24 deterministic threshold models .
sharp transitions , like those found in our results , are characteristic for class @xmath25 .
models in class @xmath26 give fuzzy curves , similar to those obtained here for larger amounts of admixtured guards or sinners. these models are also called critical mass models. in our terms , a critical amount of sinners should not be punished to get the transition to the bold phase ; this could mean that local concentration of guards should be small .
if this is so , our results do fit into the classification proposed in @xcite . as we noted in the introduction
, a similar classification was developed in @xcite .
note that the basis of @xcite was the random - field ising model , which is far from the picture of contagion .
+ coming back to the sociological reality , let us add a few words on the data presented in fig . 1 in terms of punishment and its cost .
we stress that the parameters @xmath2 and @xmath4 do not mean directly the amount of units which a punished and punishing agent should pay , but just the measure of the decrease of the probability that he will defect and punish again . in these terms ,
large numbers on the statistics of crimes in a country can be interpreted as an indication that the punishment cost are large or the punishment is weak .
accordingly , an increase of the data can mean that the punishment @xmath2 decreased below some critical value or the punishment cost @xmath4 increased .
then , new generations faced with an issue , to break a given norm or to preserve it , decide in a collective way , and these decisions are visible in the statistical data . if this point of view is accepted , we should admit in particular that law in germany , hungary and poland is broken much more frequently after 1993 .
however , the data presented in fig . 1 can be seen also as a demonstration , that police in these countries is less punished for detecting crimes after 1993 , than before .
obviously , this punishment is not open and intended ; still it can appear as a consequence of burdensome bureaucratic procedures , faulty organization , unclear rights and shifts of political aims .
99 r. i. m. dunbar , _ the social brain : mind , language , and society in evolutionary perspective _ , annual review of anthropology * 32 * ( 2003 ) 163 - 181 .
a. kpiski , _ self - portrait of a man _
( in polish ) , wyd .
literackie , krakw 2003 .
r. axelrod , _ an evolutionary aproach to norms _ , amer .
political sci .
80 ( 1986 ) 1095 - 1111 .
r. axelrod , _ complexity of cooperation .
player - based models of competition and collaboration _ , princeton up , princeton 1997 .
m. galan and l. izquierdo , ( 2005 ) _ appearances can be deceiving : lessons learned re - implementing axelrod s evolutionary approach to norms_. jasss * 8 * ( 3 ) 2 ( 2005 ) .
m. neumann , _ homo socionicus : a case study of simulation models of norms _ , jasss * 11 * , no .
4/6 ( 2008 ) .
r. axelrod , _ on six advances in cooperation theory _ , analyse und kritik , * 22 * ( 2000 ) 130 - 151 .
g. deffuant , s. moss and w. jager , _ dialogues concerning a ( possibly ) new science_. jasss 9 ( 1 ) 1 ( 2006 ) . c. castellano , s. fortunato and v. loreto , _ statistical physics of social dynamics
, in press ( arxiv:0710.3256 )
. data of eurostat ( http://epp.eurostat.ec.europa.eu/ ) q. michard and j .- p .
bouchaud , _ theory of collective opinion shifts : from smooth trends to abrupt swings _ , eur .
j. b * 47 * ( 2005 ) 151 - 159 .
k. kuakowski , _ cops or robbers - a bistable society _
c * 19 * ( 2008 ) 1105 - 1111 .
k. kuakowski , a. dydejczyk and m. rybak , _ the norm game : how a norm fails _ , lncs * 5545 * ( 2009 ) 835 - 844 .
d. stauffer and a. aharony , _ introduction to percolation theory _ ,
routledge , london 2003 . b. bollobs and o. riordan , _ percolation _ , cambridge up , cambridge 2006 . s. wasserman and k. faust , _ social network analysis : methods and applications _ , cambridge up , cambridge 1994 .. j. scott , _ social network analysis : a handbook _ , sage publications , newbury park 1991 .
r. a. hanneman and m. riddle , _ introduction to social network methods_. riverside , ca : university of california , riverside 2005 ( published in digital form at http://faculty.ucr.edu/ hanneman/ ) m. e. j. newman , _ the structure of scientific collaboration networks _ , pnas * 98 * ( 2001 ) 404 - 409 .
m. c. gonzlez , h.j .
herrmann , j. kertsz , and t. vicsek , _ community structure and ethnic preferences in school friendship networks _ , physica a * 379 * ( 2007 ) 307 - 316 . j .-
onnela , j. saramki , j. hyvonen , g. szab , m. argollo de menezes , k. kaski , a .-
barabsi and j. kertsz , _ analysis of a large - scale weighted network of one - to - one human communication _ , new j. phys .
* 9 * ( 2007 ) 179 .
m. e. j. newman , d. j. watts and s. h. strogatz , _ random graph models of social networks _ , pnas * 99 * ( 2002 ) 2566 - 2572 . l. a. n. amaral , a. scala , m. barthlmy and h. e. stanley , _ classes of small - world networks _ ,
pnas * 97 * ( 2000 ) 11149 - 11152 .
m. schnegg , _ reciprocity and the emergence of power laws in social networks _
c * 17 * ( 2006 ) 1067 - 1076 .
p. s. dodds and d. j. watts , _ universal behavior in a generalized model of contagion _ , phys .
lett . * 92 * ( 2004 ) 218701 . g. marshall ( ed . ) , _ the concise oxford dictionary of sociology _ , oxford up , oxford 1994 . | the norm game ( ng ) introduced by robert axelrod is a convenient frame to disccuss the time evolution of the level of preserving norms in social systems .
recently ng was formulated in terms of a social contagion on a model social network with two stable states : defectors or punishers . here
we calculate the critical line between these states on the plane of parameters , which measure the severities of punishing and of being punished .
we show also that the position of this line is more susceptible to the amount of agents who always punish and never defect , than to those who always defect and never punish .
the process is discussed in the context of the statistical data on crimes in some european countries close to wrocaw - the place of this conference - around 1990 . * the norm game on a model network : a critical line * + m. rybak , a. dydejczyk and k. kuakowski + _ _ faculty of physics and applied computer science , agh university of science and technology , al .
mickiewicza 30 , pl-30059 krakw , poland @[email protected] , [email protected] , [email protected] _ keywords : _ social networks ; multiagent systems |
measurements of nuclear magnetic dipole and electric quadrupole moments provide valuable insights into the nuclear structure based on the intriguing interplay of single particle and collective degrees of freedom . in particular ,
nuclei in the @xmath12- and @xmath13-shell model space have recently received considerable attention through many new experiments inspired by the unique possibility to compare extensive data with results from large - scale shell model calculations .
this progress is also related with the observation that general features of the nuclear medium can be studied in lighter nuclei as well .
for instance , superdeformation which had been exclusively investigated for a long time on nuclei in the rare earth mass region and beyond @xcite has recently been observed in @xmath14ar @xcite , @xmath15ca @xcite , @xmath16zn @xcite and @xmath17zn @xcite .
the obvious advantage in all these cases is the possibility to directly compare data with structure calculations based on highly developed nuclear shell model codes thus providing new features of this phenomenon and collectivity in general on a microscopic non - phenomenological level .
+ in the same context measurements of @xmath0 factors and b(e2 ) values have been carried out for even - a zn isotopes with @xmath18 @xcite .
two sets of large - scale shell model calculations were applied , based either on a @xmath15ca core and an @xmath13-shell model space or a @xmath7ni core and an @xmath19-shell configuration space with the inclusion of the @xmath3 orbital ; the latter was considered to be particularly important for the heavier zn isotopes . in these measurements
coulomb excitation of 160 mev zn projectiles was achieved in collisions with a carbon target employing inverse kinematics . at this beam energy
essentially the first @xmath20 states were strongly excited whereas higher - lying states were only weakly populated .
+ this deficiency was overcome in a succeeding experiment aiming at higher excitation energies of both @xmath21zn and @xmath2zn by increasing the projectile energies to 180 mev close to the coulomb barrier @xcite . in these new measurements
the @xmath0 factors and the b(e2 ) values of the now accessible @xmath1 states were determined for the first time . whereas the g(@xmath1 ) value of @xmath21zn was found to be consistent with predictions of both the collective and the spherical shell model , the value for @xmath2zn of g(@xmath1 ) = @xmath22 turned out to be in severe conflict with both model predictions .
the negative sign of the @xmath0 factor was a clear indication for a dominant @xmath3 neutron component in the wave function , whereby the schmidt value is @xmath23 which , however , was not verified by the calculations .
on the other hand , the newly determined b(e2 ) values were very well explained by these shell model calculations based on a @xmath7ni core plus @xmath3 neutrons .
evidently , in this case the explicit inclusion of the @xmath3 orbital seems to be crucial for @xmath2zn , as the alternative calculations , assuming an inert @xmath15ca core but excluding the @xmath3 orbital , underestimate the experimental e2 strength . in order to set the @xmath0 factor result of the @xmath2zn(@xmath1 ) state on more solid grounds new measurements have been performed . as emphasized in @xcite such an effort
is required as in the previous measurements the energy resolution of the nai(tl ) scintillators used for @xmath24 detection did not allow to fully separate the relevant ( @xmath25 ) @xmath24 line from a neighbouring ( @xmath26 ) 1261 kev line . as the latter was also strongly doppler - shifted due to the short nuclear lifetime , the separation of the two @xmath24 lines was particularly crucial for the detector pair placed in the forward hemisphere . on the other hand , estimates of an eventual admixture of the @xmath27 state to the measured precession of the @xmath1 state , based on the observed line intensities and the angular correlations of the corresponding @xmath24 transitions , could not explain the deduced @xmath0 factor , even under the assumption of a negative @xmath0 value for the @xmath27 state .
in the present experiment a beam of isotopically pure @xmath2zn ions was accelerated to an energy of 180 mev at the munich tandem accelerator providing intensities of 20 @xmath28na on a multilayered target .
the latter consisted of 0.44 mg/@xmath29 natural carbon deposited on a 3.34 mg/@xmath29 gd layer , which was evaporated on a 1.4 mg/@xmath29 ta foil , backed by a 4.49 mg/@xmath29 cu layer . between c and gd as well as between ta and cu thin layers of natural titanium ( @xmath30 mg/@xmath29 )
provided good adherence being very crucial for the precession experiments .
the same target had been used in former measurements under almost identical conditions @xcite .
the target was cooled to liquid nitrogen temperature and magnetized to saturation by an external field of 0.06 t. the relevant level scheme of @xmath2zn for coulomb excitation is shown in fig
. 1 @xcite .
the excited zn nuclei move through the gd layer at mean velocities of @xmath31 ( @xmath32 ) experiencing spin precessions in the transient field and are ultimately stopped in the hyperfine - interaction - free environment of the cu backing .
the de - excitation @xmath24 rays were measured in coincidence with the forward scattered carbon ions detected in a 100 @xmath33 m si detector at @xmath34 .
a 5 @xmath33 m thick ta foil between target and particle detector served as a beam stopper which , however , was transparent to the carbon recoils . as in the previous experiments the detector was operated at a very low bias of @xmath35 v to establish a thin depletion layer for separating the energies of carbon ions from those of light particles as protons and @xmath36 particles resulting from sub - coulomb fusion and transfer reactions . under these conditions
very clean @xmath24-coincidence spectra were obtained .
intrinsic ge detectors of @xmath37 relative efficiency were used for @xmath24 detection .
2 shows a typical coincidence spectrum with emphasis on the ( @xmath25 ) @xmath24 line .
evidently , all relevant @xmath24 lines were well resolved allowing a rigorous determination of their intensities required for the angular correlations as well as for the precessions of the nuclear states in question .
particle-@xmath24 angular correlations @xmath38 have been measured for determining the slope @xmath39 = [ 1/w(@xmath40)]@xmath41[dw(@xmath40)/d@xmath40 ] in the rest frame of the @xmath24 emitting nuclei at @xmath43 = @xmath44 and @xmath45 , where the sensitivity to the precessions was optimal for all transitions of interest .
precession angles , @xmath46 , were derived from counting - rate ratios r for up and down directions of the external magnetizing field which can be expressed as @xcite , @xmath47 where @xmath0 is the @xmath0 factor of the nuclear state and @xmath48 the transient field acting on the nucleus during the time interval @xmath49 ) which the ions spend in the gadolinium layer of the target ; the exponential accounts for nuclear decay with lifetime @xmath50 in the gd layer .
simultaneously with the precessions the lifetimes of several excited states have been redetermined using the doppler - shift - attenuation - method ( dsam ) . for the analysis of the doppler - broadened lineshapes , which were observed with a ge detector placed at @xmath34 to the beam direction ,
the computer code lineshape @xcite has been used .
specific details of the analysis procedure are given in @xcite .
the @xmath0 factors have been determined from the measured precession angles by calculating the effective transient field @xmath48 on the basis of the empirical linear parametrization ( see @xcite ) : @xmath51 with @xmath52 where the strength parameter @xmath53 @xcite , and @xmath54 is the attenuation factor accounting for the demagnetization of the gd layer induced by the zn beam ( see @xcite ) .
the same scheme has been successfully applied in many former measurements . [ cols="^,^,^,^,^,^,^,^,^ " , ] [ tb : table2 ] in order to study the importance of the @xmath3 orbit we have performed shell model calculations using @xmath7ni as closed shell core , with a model space defined by protons and neutrons occupying the single - particle orbitals @xmath55 , @xmath56 , @xmath57 and @xmath3 ( @xmath8 ) . to determine the effective interaction we use the recent charge - dependent potential model of machleidt , the so - called cd - bonn interaction @xcite .
the final effective two - body interaction is obtained via many - body perturbation theory to third order , employing a renormalized nucleon - nucleon interaction defined for @xmath7ni as closed shell core and including folded - diagrams to infinite order . for details , see for example ref .
a harmonic oscillator basis was used , with an oscillator energy determined via @xmath58 = 10.1 mev , a = 56 being the mass number . for the single - particle energies we employ values adapted from grawe in ref .
@xcite , resulting in the energy differences @xmath59 mev , @xmath60 mev , @xmath61 mev for neutrons and @xmath62 mev , @xmath63 mev , @xmath61 mev for protons .
the effective interaction has not been corrected for any eventual monopole changes .
this means that the only parameters which enter our calculations are those defining the nucleon - nucleon interaction fitted to reproduce the scattering data , the experimental single - particle energies and the oscillator basis .
furthermore , for the computation of the b(e2 ) s and @xmath0 factors , we have used unrenormalized magnetic moments and the canonical unrenormalized charges for protons and neutrons .
the latter entails charges of @xmath64 and @xmath65 for protons and neutrons , respectively .
the free - nucleon operator for the magnetic moment is defined as @xmath66 with @xmath67 , @xmath68 , @xmath69 , @xmath70 .
note , however , that the magnetic moment operator in finite nuclei is modified from the free - nucleon operator due to core - polarization and meson - exchange current ( mec ) corrections @xcite .
the effective operator is defined as @xmath71 , \ ] ] where @xmath72 , @xmath73 , @xmath74 or @xmath75 , with @xmath76 the free - nucleon , single - particle @xmath77 factors ( @xmath78 ) and @xmath79 the calculated correction to it .
note the presence of a new term @xmath80 $ ] , absent from the free - nucleon operator , which is a spherical harmonic of rank @xmath81 coupled to a spin operator to form a spherical tensor of multipolarity @xmath82 . in the calculations below we limit ourselves to a calculation with free operators only .
an obvious modification of the latter is to use renormalized moments , as done in the recent work of ref .
@xcite for nuclei in the @xmath83sn mass region .
the results of the calculations for the energies of the low - lying excited states for both @xmath21zn and @xmath2zn in table [ tb : table2 ] , while @xmath84 ) , @xmath85 ) and @xmath86 ) and the @xmath0 factors @xmath87 , @xmath88 and @xmath89 are presented and compared with the available data .
two types of shell - model results are discussed and shown in this table , sm-1 and sm-2 . in sm-1
we limit the number of neutrons which can be excited to the @xmath3 orbit to two , whereas sm-2 is the full shell - model calculation .
the latter basis is , however , unnecessarily large since the results are converged with four neutrons at most in the @xmath3 orbit .
one sees clearly that with at most two neutrons in the @xmath3 orbit the spectrum is rather poorly reproduced .
however , even for the fully converged calculation the first excited @xmath90 state at 1.655 mev is badly reproduced ( @xmath91=2.406 mev ) , indicating most likely the need for particle - hole excitations , especially from the @xmath92 orbit .
the typical occupation probabilities in the sm-2 calculation of the @xmath3 neutron single - particle orbit span from 2.2 to 2.4 .
the @xmath56 neutron orbit has an average occupancy of 3.5 particles whereas the low - lying @xmath93 neutron orbit has occupancies around three .
we note that for the @xmath94 , @xmath95 and the @xmath96 states there is a satisfactory agreement with the data
. for @xmath2zn the b(e2;2@xmath970@xmath98 ) exhibits a good agreement with data using unrenormalized effective charges while the b(e2;4@xmath98 @xmath992@xmath98 ) is overestimated .
this could be ascribed to deficiencies in the two - body hamiltonian and/or omitted degrees of freedom in the model space . for @xmath21zn
we get 170.4 @xmath100@xmath101 for the b(e2;2@xmath970@xmath98 ) , to be compared with the experimental value of 307 @xmath100@xmath101 , hinting at the need of larger effective charges .
this means in turn that particle - hole excitations involving the @xmath92 orbit may be more important for @xmath21zn than for @xmath2zn , in good agreement with previous shell model calculations including this degree of freedom ( see for example the discussions in refs .
this is also reflected in the @xmath0 factors for @xmath21zn , which tend to be larger than the experimental values .
if we , however , introduce effective magnetic moments by assuming a renormalization factor of @xmath102 for @xmath103 of protons and neutrons , we obtain @xmath0 factors closer to experiment .
however , the @xmath92 orbit , if coupled with the @xmath56 orbit can yield a negative contribution to the @xmath0 factors .
the latter analysis is obviously performed at the level of a simple two - body configuration ( see for example discussions in refs .
@xcite ) , however , together with the @xmath104 configuration these are the only two - body configurations of interest here which can yield a negative @xmath0 factor . for @xmath21zn
the occupancy of the @xmath3 orbit is less than one , and plays therefore a negligible role in the calculation of @xmath0 factors .
since excitations from the @xmath7ni closed shell core are not included in the present model space , this reduction can not be accounted for .
thus , the theoretical values should be larger than the experimental ones . for @xmath2zn and its @xmath0 factors
we note that for the @xmath20 states there is a fair agreement with data , confirming the previous shell - model analysis presented in refs .
for @xmath89 we see that there is again a change from the calculation with only two neutrons in the @xmath3 orbit to the full calculation .
this displays the role played by the @xmath3 orbit , which for the @xmath96 has an average occupation probability of 2.4 in the full sm-2 calculation , much larger than we have for @xmath21zn . for sm-1
the corresponding occupation probability is 1.99 and this difference is clearly reflected in the change of @xmath89 from 0.066 to 0.008 demonstrating thereby the dominating role played by the @xmath104 admixture in the wave function .
however , for sm-2 the @xmath0 factor although very small , is still positive . introducing effective magnetic moments with a scaling factor of @xmath105 reduces the theoretical sm-2 value to @xmath106 .
we speculate again whether particle - hole excitations involving the @xmath92 orbit could yield further reductions and eventually a negative contribution .
furthermore , another possibility is to include the effect of meson - exchange currents , as done in ref .
meson - exchange current corrections arise because nucleons in nuclei are interacting through the exchange of mesons , which can be disturbed by the electromagnetic field . in terms of effective one - body operators
this leads to a correction term @xmath80 $ ] in eq .
( [ eq1 ] ) . in spite of these omitted degrees of freedom ,
we see that our model space displays the important role played by the @xmath104 neutron configuration when going from the sm-1 to the full sm-2 shell - model calculation , lending thereby support to the experimental analysis .
the role of the neutron @xmath107 orbit is also seen in the @xmath0 factors of low - lying @xmath108 states in odd @xmath109 isotopes with a large negative value , although smaller in absolute value than the corresponding schmidt value ( see ref .
we have remeasured the @xmath0 factor of the @xmath1 state in @xmath2zn , with an improved energy resolution of the detectors used .
the experimental value is @xmath10 consistent with our previous measurements .
in addition , the accuracy of the @xmath0 factors of the @xmath4 , @xmath5 and @xmath6 states has been improved and their lifetimes were well reproduced . the experimental results for @xmath21zn and @xmath2zn have been compared with large - scale shell model calculations using @xmath7ni as a closed shell core and a model space consisting of protons and neutrons occupying the @xmath8 single - particle orbits .
the agreement between theory and experiment is rather good and the calculations reproduce well the experimental trends for the @xmath0 factors from @xmath21zn to @xmath2zn , although our theoretical approach is not capable of reproducing the negative sign of the @xmath110 value in @xmath2zn .
this deficiency may be ascribed to particle - hole excitations , with the @xmath92 orbit playing a major role .
the effect of meson - exchange currents may also play a role and will be investigated in future works , together with the inclusion of the @xmath92 orbit . in view of the present results similar measurements for the ge isotones @xmath111ge and @xmath112ge
are highly desirable to search for corresponding effects in the nuclear wave functions .
the authors are thankful to the operating staff of the munich tandem accelerator .
support by the bmbf and the deutsche forschungsgemeinschaft is acknowledged .
the work of te and mhj has been supported by the research council of norway ( program for supercomputing ) through a grant of computing time .
discussions with alex brown ( msu ) and georgi georgiev ( cern ) are gratefully acknowledged .
h. grawe , m. gorska , m. rejmund , m. pftzner , m. lipoglavsek , c. fahlander , m. hellstrm , d. kast , a. jungclaus , k. p. lieb , r. grzywacz , k. rykaczewski , b. a. brown , m. hjorth - jensen , and k. h. maier , in _ highlights of modern nuclear structure _ , ed .
a. covello , ( world scientific , singapore , 1999 ) , p. 137 . | + the @xmath0 factor of the @xmath1 state in @xmath2zn has been remeasured with improved energy resolution of the detectors used .
the value obtained is consistent with the previous result of a negative @xmath0 factor thus confirming the dominant @xmath3 neutron nature of the @xmath1 state .
in addition , the accuracy of the @xmath0 factors of the @xmath4 , @xmath5 and @xmath6 states has been improved and their lifetimes were well reproduced .
new large - scale shell model calculations based on a @xmath7ni core and an @xmath8 model space yield a theoretical value , @xmath9 .
although the calculated value is small , it can not fully explain the experimental value , @xmath10 .
the magnitude of the deduced b(e2 ) of the @xmath1 and @xmath4 transition is , however , rather well described .
these results demonstrate again the importance of @xmath0 factor measurements for nuclear structure determinations due to their specific sensitivity to detailed proton and neutron components in the nuclear wave functions . , , , , ,
, @xmath0 factors and lifetimes ; @xmath11 , projectile coulomb excitation ; inverse kinematics ; transient field and dsam 21.10.ky ; 25.70.de ; 27.50.+e |
the physics of transition metal perovskites with general chemical formula abo@xmath1 ( where a is a large cation , similar in size to o@xmath6 and b is a small transition metal ( tm ) cation ) has attracted and challenged the interest and curiosity of the material science community for many decades , due to huge variety of complex phenomena arising from the subtle coupling between structural , electronic and magnetic degrees of freedom .
the high degree of chemical flexibility and the localized ( i.e. not spatially homogeneous ) character of the dominant tm partially filled @xmath5 states lead to the coexistence of several physical interactions spin , charge , lattice and orbital which are all simultaneously active .
the occurrence of strong lattice - electron , electron - spin and spin - orbit couplings causes several fascinating phenomena , including metal - insulator transitions@xcite , superconductivity@xcite , colossal magnetoresistance@xcite , multiferroicity@xcite , bandgaps spanning the visible and ultraviolet@xcite , and surface chemical reactivity from active to inert@xcite .
when the additional degrees of freedom afforded by the combinatorial assemblage of perovskite building blocks in superlattices , heterointerfaces and thin films are introduced the range of properties increases all the more , as demonstrated by the recent several remarkable discoveries in the field of oxide heterostructures@xcite .
tunability and control of these intermingled effects can be further achieved by means of external stimuli such as doping@xcite , pressure@xcite , temperature and magnetic or electric fields@xcite , thereby enhancing the tailoring capability of perovskites for a wide range of functionalities .
this rich array of behaviors uniquely suit perovskites for novel solutions in different sectors of modern technology ( optoelectronics , spintronics , piezoelectric devices and ( photo)catalysis ) , for which conventional semiconductors can not be used@xcite .
theoretical studies of tm perovskites , aiming to describe and understand the underlying physical mechanisms determining their complex electronic structures have been mainly developed within two historically distinct solid state communities , model hamiltonians@xcite and _ first principles_@xcite , which in the recent years have initiated to fruitfully cross - connect each other methodologies towards more general schemes such as dft+dmft ( density functional theory@xcite + dynamical mean field theory@xcite ) , with the aim to overtake the individual limitations and to improve the applicability and predictive power of electronic structure theory@xcite .
model hamiltonians approaches adopt simplified lattice fermion models , typically the celebrated _ hubbard model _ , inspired by the seminal works of anderson@xcite , hubbard@xcite and kanamori@xcite in which the many - body problem is solved using a small number of _ relevant _ bands and short - ranged electron interactions .
these effective models can solve the many - body problem very accurately , also including ordering and quantum fluctuations , but critically depends on an large number of adjustable parameters ( which can be in principle derivable by _ first principles _
schemes@xcite ) , and its applicability is restricted to finite - size systems@xcite . in dft the intractable many - body problem of interacting electrons is mapped into a simplified problem of non - interacting electrons moving in an effective potential throughout the kohn - sahm scheme@xcite , and electron exchange - correlation ( xc ) effects are accounted by the xc potential which is approximated using xc functionals such as the local density approximation ( lda ) , the generalized gradient approximation ( gga ) _ et similaria_@xcite .
as the name suggests , in dft the ground state properties are obtained only from the charge density , and this makes dft fundamentally different from wavefunction - based approaches as the hartee - fock method , the simplest approximation to the many - body problem which includes the exact exchange but no correlation@xcite . though dft has been widely and successfully used in the last 40 years in solid - state physics and quantum - chemistry to calculate structural data , energetics and , to a lesser extent , electronic and magnetic properties , it suffers of fundamental difficulties mostly due to the approximate treatment of xc effects .
this drawback is particularly severe when dft is applied to the so called strongly correlated systems ( scss ) , whose prototypical examples are transition metal oxides ( tmos ) .
a systematic improvement of these xc - related deficiencies in dft is essentially impossible , but several beyond - dft approaches have been proposed which deliver much more satisfying results .
the most renewed ones are the dft+u@xcite , sic@xcite , hybrid functionals@xcite , and gw@xcite .
for a recent review on dft and beyond applied to transition metal oxides see . [ cols="^,^,^,^,^,^,^,^,^,^ " , ] [ tab:19 ]
it is instructive at this point to compare the set of parameters @xmath7 with the optimum values of @xmath2 obtained throughout the inverse dielectric constant relation @xmath8 introduced in the computational section ( eq . [ eq : alpha ] ) and derived by mapping hybrid dft with gw .
the measured dielectric constant @xmath9@xcite and the corresponding @xmath10 values are also listed in table [ tab:19 ] , along with the pead hse values of @xmath9 obtained for @xmath2=@xmath7 .
remarkably , the agreement between the measured and calculated @xmath9 is very good . the nice correlation between @xmath7 and @xmath11
can be appreciated graphically in fig .
[ fig : gapeps ] .
theses two curves follow a very similar behavior characterized by an initial large value of @xmath2 for the poorly screened d@xmath12 band insulator lasco@xmath1 followed by a plateau of similar values in the range d@xmath13 ( latio@xmath1 ) @xmath14 d@xmath15 ( lafeo@xmath1 ) and finally a sharp decrease towards the more strongly screened metallic compounds characterized by a completely filled @xmath3 manifold . for lanio@xmath1 and lacuo@xmath1
the optimum value of @xmath2 is zero ( not shown ) .
thus , the @xmath16 curve derived from the hse fitting procedure conducted by computing a wide set of structural ( volume , cell shape , jt distortions , atomic positions ) electronic ( band gap and dos ) and magnetic ( spin ordering , magnetic moment ) properties as a function of @xmath2 reflects well the evolution of the screening properties across the la@xmath0o@xmath1 series represented by the dielectric function @xmath17 . however , from a quantitative point of view the two sets of value differ by about 0.07 , as clarified graphically by the open squares in fig .
[ fig : gapeps ] . in order to achieve a good quantitative match between the @xmath7 and @xmath11
curves it is necessary to shift downwards the latter by about 0.07 .
this behavior is attributable to two main reasons : ( i ) the relation @xmath18 identifies a proportionality between @xmath19 and @xmath10 , not an exact equality ( the factor of proportionality is not exactly 1 , as discussed in refs . ) ; ( ii ) as already mentioned before , eq . [ eq : alpha ] holds for standard _ unscreened _ hybrid functionals such as pbe0 .
hse is a range - separated screened hybrid functional which contains already a certain degree of screening ( controlled by the screening factor @xmath20 ) .
therefore it is expected that the optimum @xmath2 derived for pbe0 ( @xmath21 ) will be smaller than the corresponding @xmath20-dependent hse value ( @xmath7 ) @xcite . needless to say ,
that in absence of a systematic study of the role of @xmath20 it is very difficult to quantify its effect on @xmath7 .
we leave this issue open for future studies .
recently , alkauskas _ et al .
_ has proposed that an optimal mixing coefficient can generally be found for any material using the formula@xcite : @xmath22 where @xmath23 and @xmath24 indicate the experimental and semilocal ( gga / lda ) band gap , and @xmath25 ( @xmath26(@xmath2 ) represents the variation of the band gap as a function of @xmath2)@xcite .
this practical relation takes advantage of the linear relation between @xmath26 and @xmath2 , which holds true as long as the valence band maxima and conduction band minimum ( and their associated wave functions ) do not change much by changing @xmath2@xcite . in practice ,
if the experimental bang gap is known , it is sufficient to perform only one hybrid functional calculation for an arbitrary value of @xmath2 plus one semilocal calculation to derive the optimum value of @xmath2 .
clearly , this empirical procedure does not guarantee that the so obtained optimum @xmath2 is the best choice for what concerns the other ground state properties .
we have already addressed this issue for latio@xmath1 for which @xmath2=0.1 give the best band gap but leads to the incorrect magnetic ordering .
the changes of the bang gap as a function of @xmath2 for the series lasco@xmath1-lacoo@xmath1 are reported in fig .
[ fig : gap ] .
the linearity is well preserved for most of the materials with the exception of the small band gap compounds latio@xmath1 and lacoo@xmath1 for which a sudden change of @xmath25 is observed for a critical value of @xmath2 . a departure from linearity
is also found for the jt / mh insulator lamno@xmath1 if we adopt the fully relaxed structure ( full line ) .
this is due to the fact that the cooperative jt distortions , which contribute to the opening of the band gap , do not change linearly with @xmath2 ( as such , the associated wave function will change with @xmath2 ) . indeed , by keeping the atomic coordinates fixed to the experimental positions the gap grows linearly by increasing @xmath2 ( dashed line ) .
the values of @xmath27 obtained from this set of curves are indicated with empty ( red ) circles and included in table [ tab:19 ] .
for some materials we provide more than one optimum mixing parameter since different experimental gaps are reported in literature ( see previous section ) . not surprisingly , we find that the values of @xmath27 are very similar to the corresponding @xmath7 data , with the exception of latio@xmath1 and to a lesser extent lamno@xmath1 , and correlates well with @xmath11 , as visualized in fig .
[ fig : gapeps ] . listed in table [ tab:19 ]
showing the correlation between the hse fitted parameters ( hse fit ) , the inverse dielectric constant relation ( @xmath28 ) , and @xmath29 ( eq . [ eq : gap ] ) .
the light - gray squares represent the @xmath28 values shifted by 0.07 .
this shift roughly reflects the amount of screening incorporated in hse via the screening factor @xmath20 as compared to fully unscreened pbe0 ( see text ) .
, scaledwidth=48.0% ] now , with the rough - and - ready set of optimum hse - fitted values @xmath7 we conclude this paper by providing a general picture of the variation of the properties of la@xmath0o@xmath1 from @xmath0=sc to @xmath0=cu , by comparing our computed results with the available experimental data .
figure [ fig : trend ] shows the remarkably good agreement between the calculated and measured values of the volume ( v ) , tilting angle ( @xmath30 ) , jt distortion , band gap ( @xmath26 ) , magnetic moment ( @xmath31 ) and dielectric constant ( @xmath9 ) . the correlation between v and @xmath32 , as well as between @xmath30 and @xmath33
was already discussed at the beginning of sec .
[ sec : struc ] .
the variation of the magnetic moment as a function of @xmath0 can be easily understood in terms of the progressive @xmath3 and @xmath4 band filling in the high - spin compounds latio@xmath1 ( @xmath34 , @xmath31=0.51 @xmath35 ) , lavo@xmath1 ( @xmath36 , @xmath31=1.3 @xmath35 ) , lacro@xmath1 ( @xmath37 , @xmath31=2.63 @xmath35 ) , lamno@xmath1 ( @xmath38 , @xmath31=3.66 @xmath35 ) , and lafeo@xmath1 ( @xmath39 , @xmath31=3.9 - 4.6 @xmath40 ) .
as already specified the experimental and computed magnetic moments of lacoo@xmath1 should be taken with a certain caution . lanio@xmath1 and
lacuo@xmath1 shown a non - magnetic behavior at pbe level only ( we have already reported that the a small magnetic moment of 0.169 @xmath35 for lanio@xmath1 is found for the fully relaxed structure ) .
( think lines ) and optimum values of @xmath2 ( circles ) obtained throughout the practical formula @xmath29=@xmath41 ( eq . [ eq : gap ] ) .
, scaledwidth=40.0% ] , and jt distortions q@xmath42 and q@xmath1 ) , electronic ( bandgap @xmath26 ) , magnetic ( magnetic moment @xmath31 ) , and dielectric constant ( @xmath9 ) quantities along the la@xmath43o@xmath1 series from @xmath0=sc to @xmath0=cu .
we also show the trend of the tolerance factor @xmath33=(@xmath44+@xmath45)/@xmath46(@xmath32+@xmath47 ) , where @xmath44 , @xmath32 and @xmath47 indicate the ionic radius for la , @xmath0=sc - cu and o , respectively , as well as @xmath32 . for latio@xmath1 we used @xmath2=0.1 .
the character of the insulating gap is also indicated ( bi = band insulator , ct = charge transfer , mh = mott - hubbard , ct / mh = mixed ct and mh character ) . ,
scaledwidth=50.0% ] the variation of the band gaps with the @xmath0 species shown in fig .
[ fig : trend ] is rather complex and in line with the earlier observation of arima@xcite : we observe a general tendency of the mh gap to increase as the tm atomic number increases , whereas the ct gaps follow an opposite behavior .
as expected , there is an apparent correlation between the trend of the band gaps and the optimum fraction of exact exchange displayed in fig .
[ fig : gapeps ] , especially @xmath27 . in lda+u language
this behavior is interpreted as a correlation between the strength of the effective coulomb interaction u and the band gap ( this become particularly evident by comparing the @xmath26 curve in fig.[fig : trend ] with fig . 2 in ref . showing the changes of the effective u with respect to @xmath0 ) . in fig .
[ fig : bands ] we collect the bandstructures of la@xmath0o@xmath1 obtained using the optimum @xmath7 values showing the variation of the electronic dispersion across the whole series . starting from the @xmath48 band insulator lasco@xmath1 the addition of one @xmath5 electron creates an highly localized @xmath3 state right below @xmath49 in latio@xmath1 .
the gradual filling of this @xmath3 manifold leads to a continuous increase of the band width from @xmath50 ( latio@xmath1 ) to @xmath51 ( lacro@xmath1 ) , connected with a gradual increase of the crystal field splitting . in lamno@xmath1 the fully occupied @xmath3 band is pushed down in energy and the valence band maxima is dominated by the half - filled @xmath52 subbands .
the @xmath53 orbital gets completely filled in lafeo@xmath1 which is the last member of the series having a predominantly mh gap .
the inclusion of one additional electron yields a sudden change of the bandstructure manifested by a high increase of @xmath54-@xmath5 hybridization and bandwidth around e@xmath55 . which finally lead to the onset of a metallic state in lanio@xmath1 and lacuo@xmath1 .
three members of the la@xmath0o@xmath1 family ( latio@xmath1 , lavo@xmath1 , and lamno@xmath1 ) are known to display orbital - ordering ( oo ) associated with the partially filled @xmath3 and @xmath56 orbitals located on top of the valence band ( these states are demarcated by thick lines in fig .
[ fig : bands ] ) .
a visual representation of the oo states derived from the optimum hse wavefunctions is shown in fig.[fig : oo ] in terms of charge density isosurfaces of the highest occupied @xmath5 states . in the following
we describe briefly the most important characteristics of the observed oo states .
\(i ) in latio@xmath1 , where the oo originates from the single @xmath3 electron , the lobes have a quasi cigar - like shape with asymmetric contributions along the two main directions , indicating an almost identical occupation of the three @xmath57 , @xmath58 , and @xmath59 @xmath3 shells .
co - planar lobes are arranged in a chessboard - like way with a sign alternation along z , in good agreement with previously reported theoretical@xcite and experimental works @xcite .
there is a clear connection between this chessboard - like ti @xmath60 ordering and the jt structural instability , which is manifested by the tendency of the occupied @xmath3 state to lie along the longer ti - o bond .
this also explains why the chessboard - like oo in latio@xmath1 is not as much evident as in lamno@xmath1 : in latio@xmath1 the difference between the distinct ti - o bondlengths ti - o@xmath61 ti - o@xmath62 ti - o@xmath63 , quantified by the jt parameters q@xmath42 and q@xmath1 , is about one order of magnitude smaller than in lamno@xmath1 ( see table [ tab:2 ] and table [ tab:5 ] ) .
\(ii ) the v@xmath64 ions in lavo@xmath1 can accommodate two electrons in the three possible orbital states @xmath65 , @xmath66 , and @xmath67 .
the spins are arranged according to the c - type ordering , whereas the oo state is found to be g - type , in accordance with the goodenough - kanamori rules@xcite and in agreement with x - ray diffraction@xcite and previous gga@xcite and hf@xcite calculations .
the distribution of the @xmath3 orbitals in the g - type oo state follows the cooperative jt - induced v - o bond - alternation in the xy plane and along the z axis , i.e. the @xmath3 charge density in one specific v site is rotated by 90@xmath68 with respect to that in the 6 neighboring v sites ( four in - plane and two in the adjacent vertical planes ) . as already observed for latio@xmath1 , the @xmath3 orbitals are preferentially occupied along the long - bond direction .
\(iii ) the c - type oo in lamno@xmath1 , originating from the singly occupied @xmath4 state of the mn@xmath69 3@xmath5 electrons in the high - spin configuration @xmath70 has been extensively studied both experimentally@xcite , and theoretically@xcite .
we have also recently addressed this issue throughout a maximally localized wannier functions representation of the @xmath4 states@xcite .
this c - type oo state can be written in the form @xmath71 with the sign of @xmath72 alternating along @xmath73 and @xmath74 and repeating along @xmath75 , as correctly represented by our hse charge density plots . for comparison
we show in fig .
[ fig : oofecr ] the similar charge density isosurfaces calculated for lacro@xmath1 and lafeo@xmath1 , in which the half - filling of the @xmath3 and @xmath4 orbitals inhibit any orbital flexibility .
as expected there is no sign of orbital ordering .
we conclude this section with a comparison of the calculated density of valence and conduction states with available x - ray photoemission and x - ray adsorption spectra .
this is summarized in fig .
[ fig : spectra ] .
the overall picture is satisfactory in terms of bandwidth , intensity and peaks position , though some sizable deviations are visible for lacro@xmath1 , lafeo@xmath1 and for the metallic compounds lacuo@xmath1 and lanio@xmath1 .
these differences between theory and experiment should be attributable to the approximations included in the adopted computational scheme , as mentioned in the previous section .
( g - type ) and lafeo@xmath1 ( a - type ) associated with the topmost occupied bands ( see fig.[fig : bands ] ) .
light ( yellow ) and dark ( blue ) areas indicate spin - down and spin - up , respectively .
, scaledwidth=50.0% ]
in summary , we have studied the ground state properties of the perovskite series la@xmath0o@xmath1 by means of screened hybrid dft following the hse formulation , based on the inclusion of a fraction of exact hf exchange in the short range coulomb kernel of the underlying semilocal pbe exchange - correlation functional . in particular , we have investigated the role of the hse mixing parameter @xmath2 on the ( i ) structural parameters ( volume , jt / gfo distortions , lattice parameters and unit cell symmetry ) ( ii ) electronic character ( band gap , dos , and bandstructure ) , and ( iii ) spin orderings and magnetic moment , as a function of the gradual filling of the @xmath5 manifold from lasco@xmath1 ( @xmath48 ) to lacuo@xmath1 ( @xmath76 : @xmath77 ) . on the basis of a computational fitting of the most relevant experimentally available data
we have derived a set of mixing factors which leads to an accurate qualitative and quantitative description of the physical behavior of all members of the representative la@xmath0o@xmath1 family .
it is found that , apart from lasco@xmath1 , the optimum values of @xmath2 ( @xmath7 ) are significantly smaller than the routinely used standard choice @xmath2=0.25 , especially for the end members ( lasco@xmath1 : @xmath7=0.25 ; latio@xmath1 & lavo@xmath1 : @xmath7=0.10 - 0.15 ; lacro@xmath1 , lamno@xmath1 , and lafeo@xmath1 : @xmath7=0.15 ; lacoo@xmath1 : @xmath7=0.05 ; lanio@xmath1 & lacuo@xmath1 : @xmath7=0.0 , i.e. for these two cases pbe is better than hse ) . this can be understood by correlating the changes of @xmath7 from sc to cu with the corresponding values of the inverse dielectric constant @xmath28 , and by considering that a certain degree of screening is already included by construction in the hse functional throughout the screening length @xmath20 , at variance with the unscreened parent hybrid functional pbe0 ( for which @xmath78 ) .
this suggests that the optimum value of @xmath2 in hse should be smaller than the corresponding pbe0 one : in our specific case it is proposed that the difference between @xmath7 and @xmath21 should about 0.05 - 0.07 , but a more detailed analysis on the influence of @xmath20 is required to achieve more accurate and comprehensive conclusions .
parameter : ( i ) lasco@xmath1 : @xmath79 ; ( ii ) latio@xmath1 : @xmath80 ; ( iii ) lavo@xmath1 : @xmath80 ; ( iv ) lacro@xmath1 : @xmath80 ; ( v ) lamno@xmath1 : @xmath80 ; ( vi ) lafeo@xmath1 : @xmath80 ; ( vii ) lacoo@xmath1 : @xmath81 ; ( viii ) lanio@xmath1 : @xmath82 ; ( ix ) lacuo@xmath1 : @xmath82 .
the calculated and measured spectra have been aligned by overlapping the valence band maxima and conduction band minima .
the experimental data are taken from the collection of spectra presented in ref .
, originally published in separated articles : ( i ) lasco@xmath1 : ref . ;
( ii ) latio@xmath1 : ref . ; ( iii ) lavo@xmath1 : ref . ;
( iv ) lacro@xmath1 : ref . ;
( v ) lamno@xmath1 : ref . ;
( vi ) lafeo@xmath1 : ref . ;
( vii ) lacoo@xmath1 : ref . ; ( viii ) lanio@xmath1 : ref . ; ( ix ) lacuo@xmath1 : ref .. , scaledwidth=40.0% ] an alternative way to obtain a set of optimum @xmath2 is the fitting of the bandgap only , by applying the practical recipe represented by eq .
[ eq : gap ] .
however , though this procedure has the clear advantage of reducing considerably the computational cost , it can lead to an erroneous description of other properties ( for example the best - bandgap @xmath2 in latio@xmath1 stabilizes the wrong spin ordering ) and can only be applied under the assumption that the wavefunction does not change with @xmath2 . for what concerns the description of the modulation of the electronic and magnetic properties across the la@xmath0o@xmath1 series
, we found that for all insulating compounds ( @xmath0=sc to co ) hse is capable to capture the localization of the @xmath83/@xmath4 orbitals and to provide a consistent and quantitatively satisfactory description of all considered quantities , thereby improving the deficient dft based predictions .
for the structural properties , on the other hand , pbe performs rather well , delivering optimized geometry within 1% .
the only exception are the jt parameters in lamno@xmath1 , which pbe finds 60% smaller than experiment .
hse cures this limitation , reproducing quite well the critical jt and gfo structural instabilities , and , in general , its application improves even further the overall agreement with experiment as compared to pbe . the complex nature of the pm correlated metals lanio@xmath1 and lacuo@xmath1 , end members of the la@xmath0o@xmath1 series , is only marginally accounted for by pbe and rather poorly treated at hse level .
this is mostly due to underlying dynamical correlation effects which can not be easily treated at dft / hf level .
for these compounds , pbe might be considered to be a good starting point for more elaborated many - body approaches .
this research was sponsored by the fp7 european community grant athena .
all calculations have been performed on the vienna scientific cluster ( vsc ) .
f. moussa , m. hennion , j. rodrguez - carvajal , h. moudden , l. pinsard , and a. revcolevschi , phys .
b * 54 * , 15149 ( 1996 ) .
b. c. hauback , h. fjellvg , and n. sakai , j. solid state chem . * 124 * , 43 ( 1996 ) .
x. d. zhou , l. r. pederson , q. cai , j. yang , b. j. scarfino , m. kim , w. b. yelon , w. j. james , h. u. anderson , and c. wang , j. appl .
phys . * 99 * , 08m918 ( 2006 ) .
h. wadati , d. kobayashi , h. kumigashira , k. okazaki , t. mizokawa , a. fujimori , k. horiba , m. oshima , n. hamada , m. lippmaa , m. kawasaki , and h. koinuma , phys . rev .
b * 71 * , 035108 ( 2005 ) . | we assess the performance of the heyd - scuseria - ernzerhof ( hse ) screened hybrid density functional scheme applied to the perovskite family la@xmath0o@xmath1 ( @xmath0=sc - cu ) and discuss the role of the mixing parameter @xmath2 ( which determines the fraction of exact hartree - fock exchange included in the density functional theory ( dft ) exchange - correlation functional ) on the structural , electronic , and magnetic properties .
the physical complexity of this class of compounds , manifested by the largely varying electronic characters ( band / mott - hubbard / charge - transfer insulators and metals ) , magnetic orderings , structural distortions ( cooperative jahn - teller like instabilities ) , as well as by the strong competition between localization / delocalization effects associated with the gradual filling of the @xmath3 and @xmath4 orbitals , symbolize a critical and challenging case for theory .
our results indicates that hse is able to provide a consistent picture of the complex physical scenario encountered across the la@xmath0o@xmath1 series and significantly improve the standard dft description .
the only exceptions are the correlated paramagnetic metals lanio@xmath1 and lacuo@xmath1 , which are found to be treated better within dft .
by fitting the ground state properties with respect to @xmath2 we have constructed a set of optimum values of @xmath2 from lasco@xmath1 to lacuo@xmath1 : it is found that the optimum mixing parameter decreases with increasing filling of the @xmath5 manifold ( lasco@xmath1 : 0.25 ; latio@xmath1 & lavo@xmath1 : 0.10 - 0.15 ; lacro@xmath1 , lamno@xmath1 , and lafeo@xmath1 : 0.15 ; lacoo@xmath1 : 0.05 ; lanio@xmath1 & lacuo@xmath1 : 0 ) .
this trend can be nicely correlated with the modulation of the screening and dielectric properties across the la@xmath0o@xmath1 series , thus providing a physical justification to the empirical fitting procedure .
finally , we show that by using this set of optimum mixing parameter hse predict dielectric constants in very good agreement with the experimental ones . |
in traditional cellular networks , the base stations ( bss ) in different cells independently control the transmission with their associated users .
the inter - cell interference is avoided or minimized by adopting different frequency reuse patterns , which only allow non - adjacent cells to reuse the same frequency band .
the frequency reuse factor is assigned to specify the rate at which the same frequency band can be used in the network . due to emerging high - rate wireless multimedia applications , traditional cellular systems have been pushed towards their throughput limits . as a result , it has been proposed to increase the frequency reuse factor such that each cell can be assigned with more frequency bands to increase the attainable throughput . in the special case where all cells can share the same frequency band for simultaneous transmission
, this corresponds to the factor - one or _
universal _ frequency reuse .
however , with more flexible frequency reuse , the inter - cell interference control becomes an essential problem in cellular systems , which has recently drawn significant research attentions ( see , e.g. , @xcite@xmath0@xcite ) . for multicell systems with a universal frequency reuse
, two promising approaches have been proposed to resolve the inter - cell interference problem ( see , e.g. , @xcite and the references therein ) : _ interference coordination _ and _ network mimo ( multiple - input multiple - output)_. in the former approach , the performance of a multicell system is optimized via joint resource allocation among all cells , based on their shared channel state information ( csi ) of all direct and interfering links across different cells .
furthermore , if the baseband signal synchronization among the bss of different cells is available and the transmit messages of different cells are shared by their bss , a more powerful cooperation can be achieved in the downlink via jointly encoding the transmit messages of all bss . in this so - called network mimo approach
, the combined use of antennas at different bss for joint signal transmission resembles the conventional single - cell multiantenna broadcast channel ( bc ) @xcite . in this paper , the former interference coordination approach is adopted due to its relatively easier implementation in practical systems .
more specifically , we study the inter - cell interference coordination for a two - cell ofdma downlink system with universal frequency reuse . all bss and user terminals
are assumed to be each equipped with a single antenna , and thus the system of interest can be modeled as a _
parallel interfering siso ( single - input single - output ) bc_. promising applications of this two - cell system model are illustrated in fig .
[ fig.1 ] , which shows a geographically symmetric setup with two adjacent macrocells , as well as a non - symmetric setup with one macrocell and one inside femtocell .
this paper investigates the joint power and subcarrier allocation over the two cells to maximize their sum throughput , for both centralized and decentralized implementations .
specifically , for the centralized allocation , with the assumption of a global knowledge of all channels in the network , we propose a scheme to jointly optimize power and subcarrier allocation over the two cells by applying the _ lagrange duality _ method from convex optimization @xcite .
this centralized scheme provides a performance benchmark for the decentralized schemes studied subsequently . for the decentralized resource allocation ,
this paper proposes a new _ cooperative interference control _ approach , whereby the two cells independently optimize resource allocation to maximize individual throughput subject to a set of preassigned mutual interference power constraints , in an iterative manner until the resource allocation in both cells converges .
two types of interference power constraints are further examined : one is to constrain the total interference power across all subcarriers from each cell to the active users in its adjacent cell , termed _
joint subcarrier protection _ ( jsp ) ; and the other is to limit the interference power over each individual subcarrier , termed _ individual subcarrier protection _ ( isp ) .
also , the optimal resource allocation rules for each cell to maximize individual throughput with jsp or isp are derived .
the rest of this paper is organized as follows .
section ii introduces the two - cell downlink ofdma system , and formulates the optimization problem for resource allocation .
section iii presents the centralized resource allocation scheme .
section iv proposes two decentralized schemes via the cooperative interference control approach with jsp and isp , respectively .
section v presents simulation results and pertinent discussions .
finally , section vi concludes the paper .
+ as shown in fig .
[ fig.1 ] , we consider a two - cell system sharing the same frequency band with each cell having a downlink ofdma transmission .
we use @xmath1 to denote each of the two cells , which are referred to as the 1st and 2nd cells in this paper , respectively . for convenience ,
let the 1st cell refer to the macrocell and the 2nd cell refer to either the macrocell or the femtocell in fig .
[ fig.1 ] . the total system bandwidth shared by the two cells
is assumed to be @xmath2 hz , which is equally divided into @xmath3 subcarriers ( scs ) indexed by @xmath4 .
each sc is assumed to be used by at most one user inside each cell and could be shared between two users individually selected from the two cells .
in addition , the users in the network are indexed by @xmath5 in the 1st cell and @xmath6 in the 2nd cell , where @xmath7 and @xmath8 are the total numbers of users in each corresponding cell .
furthermore , we denote the channel power gains ( amplitude squares ) from the two bss to their respective users , saying users @xmath9 , in each cell as @xmath10 and @xmath11 , respectively . the inter - cell interference channel gain from @xmath12 to @xmath13
is denoted by @xmath14 , while that from @xmath15 to @xmath16 is by @xmath17 .
we assume that the noise at each user s receiver has independent circularly symmetric complex gaussian ( cscg ) distribution over scs with zero mean and variance @xmath18 , denoted by @xmath19 , where @xmath20 is the noise power spectral density .
in addition , the transmit power allocated to user @xmath16 at sc @xmath21 is denoted by @xmath22 .
thus , over all users and scs in the 1st cell , we can define a power allocation matrix @xmath23 ( @xmath7-by-@xmath3 ) with the non - negative elements denoted by @xmath24 .
@xmath23 is assumed to satisfy an ofdma - based power allocation ( opa ) , in which there exists at most one element in each column being larger than zero and all the other elements are equal to zero .
this opa constraint can be expressed as @xmath25 .
similarly , we can define the power allocation matrix for the 2nd cell as @xmath26 ( @xmath8-by-@xmath3 matrix ) under a similar opa constraint .
the @xmath21th columns of @xmath23 and @xmath27 are denoted by two vectors @xmath28 and @xmath29 , where @xmath30 ( @xmath31 ) is drawn from the @xmath21th column of @xmath32 ( @xmath33 ) . with the above system model and assuming that the inter - cell interference is treated as additional gaussian noise at each user s receiver , the signal - to - interference - plus - noise - ratio ( sinr ) of user @xmath16 at sc @xmath21 in the 1st cell is given by @xmath34 similarly , the sinr of user @xmath13 at sc @xmath21 in the 2nd cell is denoted by @xmath35 .
thus , the achievable sum - rate of user @xmath36 is given by @xmath37 we consider the weighted - sum - rate ( wsr ) in each cell i.e. , @xmath38 where @xmath39 is the ( non - negative ) rate weight of user @xmath40 in the @xmath41th cell .
with individual transmission power constraint at each bs , the following optimization problem can be formulated to maximize the _ system throughput _ defined as @xmath42 @xmath43 where @xmath44 is the given power constraint at @xmath45 , and @xmath46 is the opa constraint for the @xmath41th cell .
in this section , we study the centralized optimization for jointly allocating resources in the two cells so as to maximize the system throughput , which corresponds to solving problem ( [ equ.6 ] ) globally with constraints ( [ equ.7 ] ) and ( [ equ.8 ] ) . for the centralized allocation , it is assumed that all channel gains in the network are collected by a central controller , which is capable of performing a centralized resource allocation and informing the allocation results to each cell for data transmission . due to the non - convex opa constraint and the non - concave objective function over @xmath23 and @xmath27 ,
the optimization problem in ( [ equ.6 ] ) is _ non - convex _ and thus can not be solved efficiently for the global optimum .
nevertheless , the lagrange duality method @xcite can be applied to this problem to obtain a suboptimal solution .
interestingly , according to @xcite , it has been shown that a so - called `` time - sharing '' condition usually holds for resource allocation problems in ofdma , and the duality gap for such problems solved by the lagrange duality method becomes asymptotically zero as the number of subcarriers in the system becomes large . accordingly , in the sequel , we apply the lagrange duality method to solve problem ( [ equ.6 ] ) .
first , we express the _ partial _ lagrangian of problem ( [ equ.6 ] ) as @xmath47 where , for each sc @xmath48 , @xmath49 and @xmath50 are non - negative dual variables associated with the power constraints in ( [ equ.7 ] ) with @xmath51 and @xmath52 , respectively .
the lagrange dual function is then given by @xmath53 hence , the dual problem can be defined as @xmath54 for a given pairs of @xmath55 and @xmath56 , we have @xmath57 where @xmath58 , is obtained by solving the following per - sc maximization problem @xmath59 the maximization problem in ( [ equ.11 ] ) is thus decoupled into @xmath3 per - sc resource allocation problems given by ( [ equ.14 ] ) . due to the opa constraints , for one particular sc @xmath21 , it can be simultaneously assigned to one pair of users @xmath60 from the two cells when the resultant @xmath61 in ( [ equ.10 ] ) attains its maximum value ( with the optimized @xmath62 and @xmath63 ) .
this user pair can be obtained by searching over all possible combinations from users @xmath64 .
thus , the optimal sc and power allocation that solves the problem in ( [ equ.14 ] ) is @xmath65 where @xmath66 is the selected user pair to share sc @xmath21 , and @xmath67 is obtained from ( [ equ.10 ] ) . for a given pair of @xmath68 , the optimal @xmath22 and @xmath69 to maximize @xmath70 in ( [ equ.16 ] )
have no closed - form solutions due to the non - convexity of this problem .
however , an iterative search based on , e.g. , newton s method @xcite can be utilized to find a pair of local optimal solutions for @xmath22 and @xmath69 .
then , we can check all possible user combinations to determine the optimal sc allocation according to ( [ equ.15 ] ) with optimized power allocation . after solving the per - sc problems in ( [ equ.14 ] ) for all @xmath21 s ,
a subgradient - based method , e.g. , the ellipsoid method , can be adopted to solve the dual problem in ( [ equ.12 ] ) so that the power constraints in ( [ equ.7 ] ) at both bss are satisfied .
the details are thus omitted for brevity .
note that problem ( [ equ.6 ] ) can be solved in polynomial time with an overall complexity with order @xmath71 in its dual domain .
specifically , for one particular sc , we search for @xmath72 combinations of user pairs and determine the power allocation for each user pair with @xmath73 iterations .
in addition , @xmath74 is the number of iterations for solving the dual problem in ( [ equ.12 ] ) .
however , this centralized allocation needs a system level coordination with all channel conditions in the two cells , which is a demanding requirement for practical applications . in the next section ,
we propose decentralized schemes for resource allocation , which can be implemented by each cell independently .
in this section , a new cooperative interference control approach is applied to design decentralized resource allocation schemes for the two - cell ofdma downlink system . in this approach
, each cell independently optimizes its resource allocation to maximize individual wsr under its bs s own transmit power constraint , as well as a set of newly imposed constraints to regulate the leakage interference power levels to the active users in its adjacent cell .
the above operation iterates between the two cells , until both cells obtain a converged resource allocation under their mutual inter - cell interferences .
specifically , two decentralized allocation schemes are studied in this section corresponding to two different types of interference power constraints , namely jsp and isp . in this subsection
, we solve the optimal resource allocation problem of maximizing the 1st cell s wsr subject to its bs s power constraint and a given jsp constraint to the active users in the 2nd cell .
similar problem formulation and solution apply to the resource allocation in the 2nd cell and are thus omitted .
consider the resource allocation problem in the 1st cell subject to the leakage interference constraint for the 2nd cell . in order to characterize the leakage interference to the 2nd cell
, @xmath12 needs to know the interference channel gains from it to all active users over different scs in the 2nd cell .
let @xmath75 denote the active user at sc @xmath21 in the 2nd cell , with the corresponding interference channel gain from @xmath12 to @xmath76 being @xmath77 .
it is then assumed that @xmath77 has been perfectly estimated by user @xmath78 in the 2nd cell and fed back to @xmath15 . after collecting @xmath77 for all @xmath21 s from its active users
, @xmath15 sends these channel gain values to @xmath12 ( via a backhaul link connecting these two bss ) . note that if a particular sc @xmath21 is not used by any user in the 2nd cell , the corresponding interference channel gain @xmath77 sent from @xmath15 to @xmath12 is set to be zero regardless of its actual value , so that this sc can be used by the 1st cell without any interference constraint . to maximize the wsr of the 1st cell , the following problem is formulated as @xmath79 subject to @xmath80 @xmath81 where @xmath82 is the given jsp power constraint for protecting all the active users in the 2nd cell .
note that @xmath82 limits the interference power averaged over all the scs ; thus , the corresponding resource allocation scheme is refereed to as the * average * scheme for convenience .
we assume the non - negative dual variables associated with ( [ equ.21 ] ) and ( [ equ.22 ] ) are @xmath83 .
similarly as in the case of centralized allocation , for a given pair of @xmath84 , problem ( [ equ.20 ] ) can be decoupled into @xmath3 per - sc problems in its dual domain , and the optimal allocation for sc @xmath21 in the 1st cell is derived as @xmath85 where @xmath86 with @xmath87 being the interference - plus - noise power at sc @xmath21 .
( [ equ.27 ] ) means sc @xmath21 should be assigned to the user , denoted by @xmath88 , giving the highest value of @xmath89 with the optimized @xmath22 . by letting @xmath90 be zero and considering non - negative power allocation , the power allocation in ( [ equ.27 ] ) should be optimized according to @xmath91 thus , ( [ equ.27 ] ) and ( [ equ.28 ] ) together provide the optimal resource allocation rules at all scs with fixed @xmath92 and @xmath93 .
then , the ellipsoid method can be adopted to iteratively search over @xmath94 and @xmath95 so that the constraints in ( [ equ.21 ] ) and ( [ equ.22 ] ) are simultaneously satisfied .
this algorithm also bears a linear complexity order i.e. , @xmath96 , where @xmath97 denotes the number of iterations for updating @xmath92 and @xmath93 .
similarly as ( [ equ.20 ] ) , the resource allocation problem for the 2nd cell can be formulated to maximize @xmath98 subject to the transmit power constraint @xmath99 of @xmath100 and the jsp constraint @xmath101 to protect the active users in the 1st cell . for a given pair of @xmath102 and @xmath103 , and @xmath82
can be found in the journal version of this paper @xcite . ]
the per - cell resource allocation described above can be iteratively implemented between @xmath104 and @xmath100 until the sc and power allocation in both cells converges . in this subsection
, we study the decentralized resource allocation with isp . similarly to the previous case of jsp , we merely present the solution to the optimization problem for the 1st cel . with the same objective function
as ( [ equ.20 ] ) and bs transmit power constraint as ( [ equ.21 ] ) , we formulate the current problem via replacing the jsp constraint in ( [ equ.22 ] ) by the following isp constraint over each individual sc : @xmath105 where @xmath106 is the interference power constraint for protecting the active user at sc @xmath21 in the 2nd cell .
again , we apply the lagrange duality to solve the per - cell resource allocation problem with isp . following a similar procedure as in jsp
, we can derive the following optimal sc and power allocation rules : @xmath107 @xmath108 where @xmath109 , with @xmath92 being the non - negative dual variable associated with ( [ equ.30 ] ) . according to ( [ equ.35 ] ) and ( [ equ.37 ] ) ,
the optimal sc and power allocation can be determined for all @xmath21 s with any given @xmath94 .
then , the bisection method @xcite can be used to adjust @xmath92 so that the bs transmit power constraint ( [ equ.21 ] ) is satisfied . nevertheless , it is not computationally efficient to individually optimize @xmath106 ( @xmath110 ) for each sc , thus two special schemes are further identified .
one scheme is to set @xmath111 , which means that each cell is not aware of its interference to the adjacent cell , named as the * no protection * scheme .
the other scheme is to set uniform peak interference power constraints over all scs , i.e. , @xmath112 , named as the * peak * scheme .
in this section , simulation results are presented to evaluate the performance of the proposed schemes for the two - cell downlink ofdma system .
it is assumed that @xmath113 mhz and @xmath114 .
in addition , all users rate weights are assumed to be one , and the noise power spectral density @xmath20 is set to be @xmath115 dbm / hz . assuming independent ( time - domain ) rayleigh fading with six independent , equal - energy multipath taps , the frequency - domain channel gains \{@xmath10 } , \{@xmath11 } , \{@xmath17 } , and \{@xmath14 } are modeled as independent cscg random variables distributed as @xmath116 , @xmath117 , @xmath118 and @xmath119 , respectively . for convenience , we normalize @xmath120 , and adjust @xmath121 and @xmath122 to generate different channel models .
[ fig.3 ] and [ fig.6 ] show the results for two macrocells with @xmath123 and fig .
[ fig.8 ] for the case with one macrocell and one femtocell with @xmath124 and @xmath125 ( cf .
1 ) . fig .
[ fig.3 ] shows the system throughput , @xmath126 , achieved by different interference power constraints @xmath101 and @xmath82 using the proposed decentralized scheme with jsp ( i.e. , the average scheme ) in section iv.a for one particular channel realization .
the channel gains are obtained by setting @xmath127 and @xmath128 , while the transmit power limits at two bss are set equally to be @xmath129 watt . in this figure , we have marked one local maximum point obtained by the iterative search method in @xcite .
also , we have marked the system throughput obtained by the centralized scheme proposed in section iii ( * optimal * scheme ) .
it is observed that the system throughput achieved by the decentralized average scheme is suboptimal as compared to that by the centralized optimal scheme .
[ fig.6 ] shows the system throughput against the average inter - cell interference channel gain for various schemes .
the channels are generated via @xmath127 and @xmath130 , with @xmath131 being the average interference channel gain ranging from @xmath132 to 1 .
the proposed decentralized average scheme achieves the system throughput close to that by the centralized optimal scheme for all values of @xmath131 , when the searched optimized values of @xmath101 and @xmath82 are applied @xcite .
if instead the preassigned values for @xmath101 and @xmath82 are applied , throughput degradations are observed to be negligible in the case of @xmath133 for the low inter - cell interference regime with small values of @xmath131 , and in the case of @xmath134 for the high inter - cell interference regime with large values of @xmath131 .
in addition , the * half * scheme ( each cell orthogonally uses half of the overall frequency band ) and the no protection scheme are observed to perform poorly for small and large values of @xmath131 .
moreover , the average scheme with jsp performs superior over the peak scheme with isp , especially when @xmath131 becomes large .
+ finally , fig .
[ fig.8 ] shows the system throughput for a macrocell with a femtocell inside it .
the channel gains are @xmath135 , @xmath136 , and @xmath137 .
the transmit power constraint at the macrocell s bs is assumed to be 1 watt , while that at the femtocell s bs is changed from 0.02 to 2 watts.it is observed that all proposed centralized and decentralized resource allocation schemes outperform the no protection scheme in the achievable system throughput , which eventually becomes saturated with the increased inter - cell interference . at low femtocell snr ,
there exists a noticeable throughput gap between the average and peak schemes , which is due to the fact that when the femtocell suffers detrimental interference from the macrocell , the average scheme can opportunistically allocate the femtocell transmit power to a small portion of scs with best channel conditions . on the other hand , at high femtocell snr ,
both average and peak schemes tend to perform close to the optimal scheme .
in this paper , the downlink cooperative interference control in a two - cell ofdma system is investigated with centralized and decentralized implementations for joint power and subcarrier allocation to maximize the system throughput .
it is shown that the proposed decentralized recourse allocation schemes via the new approach of inter - cell interference power protection achieve a performance close to that of the centralized scheme in various system settings .
in addition , the joint subcarrier protection ( jsp ) with average interference power constraint is shown to achieve a larger system throughput than the more stringent individual subcarrier protection ( isp ) counterpart with peak interference power constraint .
d. gesbert , s. hanly , h. huang , s. shamai shitz , o. simeone , and w. yu , `` multi - cell mimo cooperative networks : a new look at interference , '' _ ieee j. select .
areas communications _ , vol .
9 , pp . 1380 - 1408 , dec .
l. venturino , n. prasad , and x. wang , `` coordinated scheduling and power allocation in downlink multicell ofdma networks , '' _ ieee trans .
vehicular technology _
58 , no . 6 , pp . 2835 - 2848 , july 2009 . w.
yu , t. kwon , and c. shin , `` joint scheduling and dynamic power spectrum optimization for wireless multicell networks , '' in _ proc .
44th conf .
science sys .
_ ( _ ciss _ ) , princeton , nj , march 2010 , pp . 1 - 6 . | this paper studies cooperative schemes for the inter - cell interference control in orthogonal - frequency - division - multiple - access ( ofdma ) cellular systems .
the downlink transmission in a simplified two - cell system is examined , where both cells simultaneously access the same frequency band using ofdma . the joint power and subcarrier allocation over the two cells
is investigated for maximizing their sum throughput with both centralized and decentralized implementations .
particularly , the decentralized allocation is achieved via a new _ cooperative interference control _ approach , whereby the two cells independently implement resource allocation to maximize individual throughput in an iterative manner , subject to a set of mutual interference power constraints .
simulation results show that the proposed decentralized resource allocation schemes achieve the system throughput close to that by the centralized scheme , and provide substantial throughput gains over existing schemes . |
directed brownian motion induced by zero - mean non - equilibrium fluctuations in the absence of macroscopic forces and potential gradients is presently under intense investigation @xcite .
this comes from the desire to understand unidirectional transport in biological systems@xcite , as well as their potential technological applications ranging from classical non - equilibrium models@xcite to quantum systems@xcite .
a ratchet system is generally defined as a system that is able to transport particles in a periodic structure with nonzero macroscopic velocity in the absence of macroscopic force on average . broadly speaking , ratchet devices fall into three categories depending on how the applied perturbation couples to the substrate asymmetry : rocking ratchets@xcite
, flashing ratchets@xcite , and correlation ratchets @xcite .
additionally , entropic ratchets , in which brownian particles move in a confined structure , instead of a potential , were also extensively investigated @xcite .
these studies on directed transport of the brownian particles focused on the normal diffusion . + however
, anomalous diffusion has attracted growing attention , being observed in various fields of physics and related sciences@xcite .
description of physical models in terms of lvy flights and subdiffusion becomes more and more popular @xcite . in the complex systems the distinct class of subdiffusion processes
was reported in condensed phases@xcite , ecology@xcite , and biology@xcite .
superdiffusion driven by levy flights is actually observed in various real systems and is used to model a variety of processes such as bulk mediated surface diffusion @xcite , exciton and charge transport in polymers under conformational motion @xcite , transport in micelle systems or heterogeneous rocks @xcite , two - dimensional rotating flow @xcite , and many others@xcite .
goychuk and coworkers @xcite studied the subdiffusive transport in tilted periodic potentials and established a universal scaling relation for diffusive transport .
dybiec and coworkers @xcite studied the minimal setup for a lvy ratchet and found that due to the nonthermal character of the lvy noise , the net current can be obtained even in the absence of whatever additional time - dependent forces .
del - castillo - negrete and coworkers @xcite also found the similar results in constant force - driven lvy ratchet .
rosa and beims @xcite studied the optimal transport and its relation to superdiffusive transport and lvy walks for brownian particles in ratchet potential in the presence of modulated environment and external oscillating forces .
we also studied the transport of brownian particles in the presence of ac - driving forces and lvy flights and multiple current reversals were observed@xcite recently , much attention has been devoted to the competition between subdiffusion and lvy flights .
the competition is conveniently described by the fractional fokker - planck equation with temporal and spatial fractional derivatives @xcite .
it is very difficult to see this competition in the framework of the fractional fokker - planck dynamics .
magdziarz and coworkers @xcite proposed a equivalent approach based on the subordinated langevin method to visualize the competition on the level of sample paths as well as on the level of probability density functions . based on this approach dybiec and coworkers
@xcite found that due to the competition between lvy flights and subdiffusion , the standard measure used to discriminate between anomalous and normal behavior can not be applied straightforwardly .
koren and coworkers@xcite have investigated the first passage times in one - dimensional system displaying a competition between subdiffusion and lvy flights and found some peculiar phenomena .
what will happen when the particles move in a ratchet potential subjected to subdiffusion and lvy flights ? in order to answer this question we use the subordinated langevin method proposed by the magdziarz and coworkers @xcite to investigate this competition in a minimal lvy ratchet without any external driving forces .
we emphasize on visualizing the competition on the level of the group velocity and diffusion and finding how this competition affects the directed transport .
we consider the transport of the brownian particles driven by lvy flights and subdiffusion in the absence of whatever additional time - dependent forces .
the competition between lvy flights and subdiffusion in a ratchet potential @xmath0 can be described by the fractional fokker - planck equation with temporal and spatial fractional derivatives @xcite @xmath1 p(x , t),\ ] ] where @xmath2 is the probability density for particles at position @xmath3 and time @xmath4 .
the prime stands for the derivative with respect to the space variable @xmath3 .
@xmath5 is the anomalous diffusion coefficient which describes the noise intensity in the subordinated process .
@xmath6 denotes the generalized friction constant . here
@xmath7 is the fractional of the riemann - liouville operator ( @xmath8 ) defined through @xcite @xmath9 where @xmath10 is the gamma function . from the definition ,
it becomes apparent that subdiffusion corresponds to a slowly decaying memory integral in the dynamical equation for @xmath2 .
the operator @xmath11 , @xmath12 , stands for the riesz fractional derivative @xcite with the fourier transform @xmath13 .
the occurrences of the operator @xmath7 and @xmath11 are induced by the heavy - tailed waiting times between successive jumps and the heavy - tailed distributions of the jumps , respectively , in the underlying continuous - time random walk scheme .
the case of @xmath14 , @xmath15 corresponds to the standard fokker - planck equation .
@xmath0 is an asymmetric periodic potential @xmath16,\ ] ] where @xmath17 and @xmath18 are the amplitude and the asymmetric parameter of the potential , respectively . because it is very difficult to solve eq .
( 1 ) analytically and numerically , we used the subordinated langevin method proposed by magdziarz and coworkers @xcite to investigate the transport . in their method , the solution @xmath2 of eq .
( 1 ) is equal to the probability density function of the subordinated process @xmath19 where the parent process @xmath20 is defined as the solution the stochastic differential equation @xmath21 where @xmath22 is the symmetric @xmath23-stable lvy motion with the fourier transform @xmath24 . employing the euler scheme to eq .
( 5 ) , one can obtain @xmath25 @xmath26 where @xmath27 and @xmath28 are the random variables with standard symmetric @xmath23-stable distribution .
the procedure of generating realizations @xmath28 is the following @xcite @xmath29v)}{w}\right]^{\frac{1-\mu}{\mu}},\ ] ] where the random variable @xmath30 is uniformly distributed on@xmath31 , @xmath32 has exponential distribution with mean one . the inverse - time @xmath33-stable subordinator @xmath34 , which is assumed to be independent of @xmath20 ,
is defined as @xmath35 where @xmath36 is the strictly increasing @xmath33-stable lvy motion with laplace transform @xmath37 . using the standard method of summing increments of the process
@xmath36 one can get @xmath38 @xmath39 where @xmath40 and @xmath41 are the skewed positive @xmath33-stable random variables @xcite .
the method to generate the random variables is @xmath42^{\frac{1}{\alpha}}}\left[\frac{\cos(v-\alpha(v+\frac{\pi}{2}))}{w}\right]^\frac{1-\alpha}{\alpha},\ ] ] where @xmath30 and @xmath32 have the same definitions as that in eq .
( 8) . from the above procedures ,
one can obtain the subordinated process @xmath43 and its probability distribution function is equal to the solution of eq .
( 1 ) . for more detailed information on the algorithm
, please see the ref .
( 22 ) . in the classical ratchets
, one can use the average velocity and effective diffusion coefficient to describe the transport .
however , for the noise with distribution of a lvy - stable law , the mean of the noise and the second moment may do not exist . as a consequence ,
the classical stochastic theory ( average velocity and effective diffusion coefficient ) , which is based on the ordinary central limit theorem , is no longer valid . to overcome this problem , dybiec and coworkers @xcite recently proposed a different approach to the lvy ratchet problem based on the quantile line analysis for @xmath44 .
quantile line is a very useful tool for investigation of the overall motion of the probability density of finding a particle in the vicinity of @xmath43 @xcite .
a median line for a stochastic process @xmath43 is a function of @xmath45 given by the relationship @xmath46 .
therefore , one can use the derivative of the median to define the group velocity of the particle packet@xcite , @xmath47 and this definition is valid even for the case of lacking average velocity .
in order to investigate the competition between lvy flights and subdiffusion in a ratchet potential , we carried out extensively numerical simulations based on the subordinated langevin method @xcite .
for simplicity we set @xmath48 and @xmath49 throughout the work . in our simulations , we have considered more than @xmath50 realizations to obtain the accurate median . in order to provide the requested accuracy of the system the dynamics time step was chosen to be smaller than @xmath51 .
we have checked that these are sufficient for the system to obtain consistent results .
firstly , we will investigate the diffusive properties of the brownian particles .
usually , the types of the diffusion processes are determined by the spread of the distance traveled by a random walker .
the diffusion is characterized through the power law form of the mean - square displacement @xmath52 . according to the value of the index @xmath53 , one can distinguish subdiffusion ( @xmath54 ) , normal diffusion ( @xmath55 ) and superdiffusion ( @xmath56 ) . here
, we use the median of square displacement @xmath57 , instead of mean - square displacement , to characterize the diffusion .
1 ( a ) shows the time dependence of @xmath58 for different combinations of @xmath23 and @xmath33 without any external potential .
it is found that the linear time dependence of the median of square displacement , @xmath59 , will occur for the case of @xmath60=1 , which indicates the normal diffusion . however , this is not true , for example @xmath61 and @xmath62 , the process is still non - markov and non - gaussian .
this pseudo - normal diffusion is due to the competition between lvy flights and subdiffusion .
dybiec and coworkers @xcite have presented discussions in detail on this paradoxical diffusion .
1 ( b ) presents the time dependence of @xmath58 in the presence of a ratchet potential .
interestingly , the pseudo - normal diffusion for @xmath61 and @xmath62 with external potentials is not normal .
next , we will study the rectified mechanism of the lvy ratchets .
usually , the ratchet mechanism demands three key ingredients @xcite which are ( a ) nonlinear periodic potential : it is necessary since the system will produce a zero mean output from zero - mean input in a linear system ; ( b)asymmetry of the potential , it can violate the symmetry of the response ; ( c)fluctuating : lvy flights can break thermodynamical equilibrium . in fig .
2 ( a ) , we studied the time dependence of the median for different values of the asymmetry parameter @xmath18 at @xmath63 and @xmath64 .
the median is positive for @xmath65 , zero at @xmath66 , and negative for @xmath67 .
therefore , the asymmetry of the potential will determine the direction of the transport and no directed transport occurs in a symmetric potential .
now we will give the physical interpretation of the directed transport for the case of @xmath68 .
firstly , the particles stay in the minima of the potential awaiting large noise pulse to be catapulted out .
the particles will be thrown out to the left and the right with the equal probabilities . in this case , the distance from minima to maxima is shorter from the right side than that from the left side . consequently , most of the particles are thrown out from the right side , resulting in positive transport .
this gives rise to the overall preferred motion to the right .
figure 2 ( b ) gives the time dependence of the median for different combinations of @xmath23 and @xmath33 .
we find that lvy flights are necessary to obtain the directed transport . for gaussian case ( @xmath69 ) , directed transport disappears .
this is due to the nonthermal character of the lvy flights that can break thermodynamical equilibrium . from fig .
2(a ) and ( b ) we can see that the asymmetry of the potential and the non - equilibrium character of the lvy flights are the two necessary conditions for directed transport .
the direction of the transport is determined by the direction of the steeper slope of of the potential and the lvy flights can break thermaldynamical equilibrium .
these two key ingredients can realize the ratchet effects .
figure 3 illustrates the dependence of the group velocity @xmath70 on the subdiffusive index @xmath33 for different values of the lvy index @xmath23 .
one can see that group velocity @xmath70 increases monotonically with the subdiffusive index @xmath33 . for small values of @xmath33 ,
the waiting time between successive jumps is very long and it is not easy for particles to pass across the barrier .
thus , most particles will stay in their original minima of the potential and the group velocity becomes very small .
especially , we also find that there exists a threshold value of @xmath33 below which no directed transport can be obtained .
the subdiffusion dominates the transport for small values of @xmath33 ( @xmath71 ) , while the effects of the lvy flights become preponderant for large values of @xmath33 .
figure 4 shows the dependence of the group velocity @xmath70 on the lvy index @xmath23 for different values of @xmath33 .
when @xmath72 , the system is under thermodynamical equilibrium and no directed transport appears . for small values of @xmath23 , lvy flights are longer and the outliers in the lvy noise are larger . in this case , the effects of the asymmetry of the potential become very small , resulting in small group velocity . therefore , there exists a optimal value of @xmath23 at which the group velocity takes its maxima .
this can also be confirmed by fig .
3 . for very small values of @xmath33 , for example @xmath73 ,
the group velocity is zero for all values of @xmath23 and the transport is absolutely dominated by subdiffusion .
the group velocity @xmath70 as a function of noise intensity @xmath5 is shown in fig . 5 for different combinations of @xmath23 and @xmath33 .
the curve is observed to be bell shaped which shows the feature of resonance .
when @xmath74 , the particles can not pass across the barrier and there is no directed current .
when @xmath75 so that the noise is very large , the effect of the potential disappears and the group velocity tends to zero , also .
there is an optimal value of @xmath5 at which the group velocity is maximal .
there are two intersections ( @xmath76 and @xmath77 ) between the line of @xmath78 and @xmath64 and the line of @xmath63 and @xmath62 . for simplicity , we define @xmath79 as the group velocity for different values of @xmath23 and @xmath33 .
when @xmath80 , @xmath81 , lvy flights dominates the transport . when @xmath82 , @xmath83 , the transport is governed by subdiffusion .
for the case of @xmath84 , all particles can easily pass across the barrier and lvy flights will mainly contribute to the transport and @xmath81 in fig .
6 , we plot the dependence of the group velocity @xmath70 on the amplitude @xmath17 of the potential for different combinations of @xmath23 and @xmath33 .
when @xmath85 , the effects of the potential disappear and the group velocity tends to zero .
when @xmath86 , the particles can not pass across the barrier and the group velocity goes to zero , also .
thus , the curve shows a peak .
remarkably , there is an intersection between the line of @xmath63 and @xmath62 and the line of @xmath78 and @xmath64 .
this is due to the competition between lvy flights and subdiffusion .
when @xmath87 , lvy flights are predominant and @xmath88 .
when @xmath89 , subdiffusion dominates the transport and @xmath90 . in this case , the height of the barrier is very high and few particles driven by lvy flights can pass across the barrier , the effects of the lvy flights will disappear and subdiffusion will play a major role .
in this paper , we have investigated the directed transport of the brownian particles in a ratchet potential driven by lvy flights coexisting with subdiffusion .
we used recently developed framework of monte carlo simulation @xcite which is equal to the solution of the fractional fokker - planck equation . the group velocity proposed by dybiec and coworkers @xcite
is used to measure the transport .
it is found that the group velocity increases monotonically with the subdiffusive index , while the group velocity as a function of the lvy index is nonmonotonic .
the former is caused by the increase of the waiting time between successive jumps and the latter is owing to the interplay between lvy flights and the height of the barriers .
there is a threshold value of @xmath33 below which the transport is absolutely dominated by subdiffusion and the directed transport disappears .
the dependences of the group velocity on the noise intensity and the amplitude of the potential are also investigated .
there is an optimal value of the noise intensity ( the amplitude of the potential ) at which the group velocity is maximal .
the competition between levy fights and subdiffusion in the ratchet potential is observed on the level of the group velocity as well as the median of square displacement .
the nonthermal character of the lvy flights and the asymmetry of the potential are the necessary conditions for directed transport when the system is in the absence of any external driving forces . because of this competition
, we also found the pseudo - normal diffusion reported by dybiec and coworkers @xcite , in which time dependence of the median of square displacement is linear , @xmath59 , while the process is still non - markov and non - gaussian .
anomalous transport is becoming widely recognized in a variety of the fields . beyond its intrinsic theoretical interest , the results we have presented may have wide applications in some complex systems , such as diffusive transport in plasmas , particles separation with non - gaussian diffusion , and ratchet transport in biology systems that are intrinsically out of equilibrium .
we would like to thank dr .
magdziarz for enthusiastic help on numerical algorithm .
this work was supported in part by national natural science foundation of china with grant nos .
30600122 and 10947166 and guangdong provincial natural science foundation with grant no . 06025073 .
y. f. he also acknowledges the research foundation of education bureau of hebei province , china ( grant no .
2009108 ) 0.4 in p. reimann , phys .
361 , 57 ( 2002 ) ; p. hnggi and f. marchesoni , rev .
81 , 387 ( 2009 ) .
f. julicher , a. adjari , and j. prost , rev .
69 , 1269(1997 ) .
j. rousselet , l. salome , a. adjari , and j. prost , nature 370 , 446 ( 1994 ) ; l. p. faucheux , l. s. bourdieu , p. d. kaplan , and a. j. libchaber , phys .
74 , 1504 ( 1995 ) .
i. dernyi , c. lee , and a. l. barabasi , phys .
lett . 80,1473 ( 1998 ) ; c. lee and b. jank , i. dernyi , a. l. barabsi , nature 400 , 337 ( 1999 ) .
m. o. magnasco , phys .
71 , 1477 ( 1993 ) ; r. bartussek , p. hnggi , and j. g. kissner , europhys .
lett . 28 , 459(1994 ) .
p. reimann , phys .
290 , 149(1997 ) ; j. d. bao and y. z. zhuo , phys .
lett . a 239 , 228 ( 1998 ) ; b. q. ai , l. q. wang , and l. g. liu , chaos , solitons fractals 34 , 1265 ( 2007 ) ; p. reimann , r. bartussek , r. haussler , and p. hnggi , phys . lett .
a 215 , 26 ( 1996 ) . c. r. doering , w. horsthemke , and j. riordan , phys .
72 , 2984 ( 1994 ) ; r. bartussek , p. reimann , and p. hnggi , phys .
76 , 1166(1996 ) .
b. q. ai and l. g. liu , phys .
e 74 , 051114 ( 2006 ) ; b. q. ai , phys . rev .
e 80 , 011113 ( 2009 ) ; f. marchesoni , s. savelev , phys .
e 80 , 011120 ( 2009 ) ; b. q. ai , j. chem . phys .
131 , 054111 ( 2009 ) .
r. metzler and j. klafter , phys .
339 , 1(2000 ) ; a. v. chechkin , o. y. sliusarenko , r. metzler , and j. klafter , phys .
e 75 , 041101 ( 2007 ) .
r. metzler , e. barkai , and j. klafter , phys . rev . lett .
82 , 3563 ( 1999 ) .
e. barkai , r. metzler , and j. klafter , phys .
e 61 , 132 ( 2000 ) .
e. barkai , phys .
e 63 , 046118 ( 2001 ) .
o. v. bychuk and b. oshaughnessy , phys .
74 , 1795 ( 1995 ) ; j. chem .
phys . 101 , 772 ( 1994 ) .
m. a. lomholt , t. ambjornsson , and r. metzler , phys .
95 , 260603 ( 2005 ) .
j. klafter , a. blumen , g. zumofen , and m. f. shlesinger , physica a 168 , 637 ( 1990 ) ; a. ott , j. p. bouchaud , d. langevin , and w. urbach , phys .
65 , 2201 ( 1990 ) .
t. h. solomon , e. r. weeks , and h. l. swinney , phys .
71 , 3975 ( 1993 ) .
i. goychuk , e. heinsalu , m. patriarca , g. schmid , and p. hnggi , phys .
e 73 , 020101(r ) ( 2006 ) ; e. heinsalu , m. patriarca , i. goychuk , g. schmid , and p. hnggi , phys . rev .
e 73 , 046133 ( 2006 ) . b. dybiec , e. gudowska - nowak , and i. m. sokolov , phys .
e 78 , 011117 ( 2008 ) ; b. dybiec , phys .
e 78 , 061120 ( 2008 ) .
d. del - castillo - negrete , v. yu .
gonchar , and a. v. chechkin , phys .
lett . a 387 , 6693 ( 2008 ) .
j. rosa and m. w. beims , physica a 386 , 54 ( 2007 ) .
b. q. ai and y. f. he , unplishied .
m. magdziarz , and a. weron , phys .
e 75 , 056702 ( 2007 ) ; m. magdziarz , a. weron , and k. weron , phys .
e 75 , 016708 ( 2007 ) ; m. magdziarz , a. weron , k. burnecki , and j. klafter , phys . rev .
lett 103 , 180602 ( 2009 ) .
b. dybiec and e. gudowska - nowak , phys .
e 80 , 061122(2009 ) .
t. koren , j. klafter , and m. magdziarz , phys .
e 76 , 031129 ( 2007 ) .
a. janick and a. weron , stat .
9 , 109 ( 1994 ) ; r. weron , stat .
28 , 165 ( 1996 ) ; j. m. chambers , c. mallows , and b. w. stuck .
71 , 340 ( 1976 ) .
s. denisov , p. hnggi , and j. l. mateos , am .
77 , 602 ( 2009 ) .
0.4 in fig . 1 .
time dependence of @xmath58 for different combinations of @xmath23 and @xmath33 : ( a)without external potential at @xmath91 , solid lines present @xmath92 scaling ; ( b)with external potential at @xmath91 , @xmath93 , and @xmath94 . + fig . 2 .
time dependence of the median : ( a ) for different values of the asymmetry parameter @xmath18 at @xmath91 , @xmath93 , @xmath63 , and @xmath64 , the inset shows the potential profile ; ( b)for different combinations of @xmath23 and @xmath33 at @xmath91 , @xmath93 , and @xmath94 .
group velocity @xmath70 versus subdiffusive index @xmath33 for different values of @xmath23 at @xmath91 , @xmath93 , and @xmath94 .
group velocity @xmath70 versus lvy index @xmath23 for different values of @xmath33 at @xmath91 , @xmath93 , and @xmath94 .
group velocity @xmath70 as a function of noise intensity @xmath5 for different combinations of @xmath23 and @xmath33 at @xmath93 and @xmath94 .
group velocity @xmath70 versus the amplitude @xmath17 of the potential for different combinations of @xmath23 and @xmath33 at @xmath91 and @xmath94 . | 0.4 in transport of the brownian particles driven by lvy flights coexisting with subdiffusion in asymmetric periodic potentials is investigated in the absence of any external driving forces . using the langevin - type dynamics with subordination techniques ,
we obtain the group velocity which can measure the transport .
it is found that the group velocity increases monotonically with the subdiffusive index and there exists an optimal value of the lvy index at which the group velocity takes its maximal value .
there is a threshold value of the subdiffusive index below which the ratchet effects will disappear .
the nonthermal character of the lvy flights and the asymmetry of the potential are necessary to obtain the directed transport . some peculiar phenomena induced by the competition between lvy flights and subdiffusion are also observed .
the pseudo - normal diffusion will appear on the level of the median . 0.4 in |
in this paper we consider the applications of a new numerical - analytical technique which is based on the methods of local nonlinear harmonic analysis or wavelet analysis to the nonlinear root - mean - square ( rms ) envelope dynamics [ 1 ] .
such approach may be useful in all models in which it is possible and reasonable to reduce all complicated problems related with statistical distributions to the problems described by systems of nonlinear ordinary / partial differential equations . in this paper
we consider an approach based on the second moments of the distribution functions for the calculation of evolution of rms envelope of a beam .
the rms envelope equations are the most useful for analysis of the beam self forces ( space charge ) effects and also allow to consider both transverse and longitudinal dynamics of space - charge - dominated relativistic high brightness axisymmetric / asymmetric beams , which under short laser pulse driven radio - frequency photoinjectors have fast transition from nonrelativistic to relativistic regime [ 1 ] .
analysis of halo growth in beams , appeared as result of bunch oscillations in the particle - core model , also are based on three - dimensional envelope equations [ 2 ] . from the formal point of view
we may consider rms envelope equations after straightforward transformations to standard cauchy form as a system of nonlinear differential equations which are not more than rational ( in dynamical variables ) .
because of rational type of nonlinearities we need to consider some extension of our results from [ 3]-[10 ] , which are based on application of wavelet analysis technique to variational formulation of initial nonlinear problems .
wavelet analysis is a relatively novel set of mathematical methods , which gives us a possibility to work with well - localized bases in functional spaces and give for the general type of operators ( differential , integral , pseudodifferential ) in such bases the maximum sparse forms .
our approach in this paper is based on the generalization [ 11 ] of variational - wavelet approach from [ 3]-[10 ] , which allows us to consider not only polynomial but rational type of nonlinearities .
our representation for solution has the following form @xmath0 which corresponds to the full multiresolution expansion in all time scales .
formula ( [ eq : z ] ) gives us expansion into a slow part @xmath1 and fast oscillating parts for arbitrary n. so , we may move from coarse scales of resolution to the finest one for obtaining more detailed information about our dynamical process .
the first term in the rhs of equation ( 1 ) corresponds on the global level of function space decomposition to resolution space and the second one to detail space . in this way we give contribution to our full solution from each scale of resolution or each time scale . the same is correct for the contribution to power spectral density ( energy spectrum ) : we can take into account contributions from each level / scale of resolution . in part 2
we describe the different forms of rms equations .
in part 3 we present explicit analytical construction for solutions of rms equations from part 2 , which are based on our variational formulation of initial dynamical problems and on multiresolution representation [ 11 ] .
we give explicit representation for all dynamical variables in the base of compactly supported wavelets .
our solutions are parametrized by solutions of a number of reduced algebraical problems from which one is nonlinear with the same degree of nonlinearity and the rest are the linear problems which correspond to particular method of calculation of scalar products of functions from wavelet bases and their derivatives .
below we consider a number of different forms of rms envelope equations , which are from the formal point of view not more than nonlinear differential equations with rational nonlinearities and variable coefficients .
let @xmath2 be the distribution function which gives full information about noninteracting ensemble of beam particles regarding to trace space or transverse phase coordinates @xmath3 .
then we may extract the first nontrivial bit of ` dynamical information ' from the second moments @xmath4 rms emittance ellipse is given by @xmath5 .
expressions for twiss parameters are also based on the second moments .
we will consider the following particular cases of rms envelope equations , which described evolution of the moments ( 1 ) ( [ 1],[2 ] for full designation ) : for asymmetric beams we have the system of two envelope equations of the second order for @xmath6 and @xmath7 : @xmath8 the envelope equation for an axisymmetric beam is a particular case of preceding equations .
also we have related lawson s equation for evolution of the rms envelope in the paraxial limit , which governs evolution of cylindrical symmetric envelope under external linear focusing channel of strenghts @xmath9 : @xmath10 where @xmath11 according [ 2 ] we have the following form for envelope equations in the model of halo formation by bunch oscillations : @xmath12 where x(s ) , y(s ) , z(s ) are bunch envelopes , @xmath13 , @xmath14 .
after transformations to cauchy form we can see that all this equations from the formal point of view are not more than ordinary differential equations with rational nonlinearities and variable coefficients ( also , b we may consider regimes in which @xmath15 , @xmath16 are not fixed functions / constants but satisfy some additional differential constraint / equations , but this case does not change our general approach ) .
our problems may be formulated as the systems of ordinary differential equations @xmath17 with fixed initial conditions @xmath18 , where @xmath19 are not more than polynomial functions of dynamical variables @xmath20 and have arbitrary dependence of time . because of time dilation we can consider only next time interval : @xmath21 .
let us consider a set of functions @xmath22 and a set of functionals @xmath23 where @xmath24 are dual ( variational ) variables .
it is obvious that the initial system and the system @xmath25 are equivalent .
of course , we consider such @xmath26 which do not lead to the singular problem with @xmath26 , when @xmath27 or @xmath28 , i.e. @xmath29 .
now we consider formal expansions for @xmath30 : @xmath31 where @xmath32 are useful basis functions of some functional space ( @xmath33 , sobolev , etc ) corresponding to concrete problem and because of initial conditions we need only @xmath34 , @xmath35 @xmath36 where the lower index i corresponds to expansion of dynamical variable with index i , i.e. @xmath37 and the upper index @xmath38 corresponds to the numbers of terms in the expansion of dynamical variables in the formal series .
then we put ( [ eq : pol1 ] ) into the functional equations ( [ eq : veq ] ) and as result we have the following reduced algebraical system of equations on the set of unknown coefficients @xmath39 of expansions ( [ eq : pol1 ] ) : @xmath40 where operators l and m are algebraization of rhs and lhs of initial problem ( [ eq : pol0 ] ) , where @xmath41 ( [ eq : lambda ] ) are unknowns of reduced system of algebraical equations ( rsae)([eq : pol2 ] ) .
@xmath43 are coefficients ( with possible time dependence ) of rhs of initial system of differential equations ( [ eq : pol0 ] ) and as consequence are coefficients of rsae .
@xmath44 , @xmath45 are multiindexes , by which are labelled @xmath46 and @xmath47 other coefficients of rsae ( [ eq : pol2 ] ) : @xmath48 where p is the degree of polinomial operator p ( [ eq : pol0 ] ) @xmath49 where q is the degree of polynomial operator q ( [ eq : pol0 ] ) , @xmath50 , @xmath51 .
now , when we solve rsae ( [ eq : pol2 ] ) and determine unknown coefficients from formal expansion ( [ eq : pol1 ] ) we therefore obtain the solution of our initial problem .
it should be noted if we consider only truncated expansion ( [ eq : pol1 ] ) with n terms then we have from ( [ eq : pol2 ] ) the system of @xmath52 algebraical equations with degree @xmath53 and the degree of this algebraical system coincides with degree of initial differential system .
so , we have the solution of the initial nonlinear ( rational ) problem in the form @xmath54 where coefficients @xmath39 are roots of the corresponding reduced algebraical ( polynomial ) problem rsae ( [ eq : pol2 ] ) .
consequently , we have a parametrization of solution of initial problem by solution of reduced algebraical problem ( [ eq : pol2 ] ) .
the first main problem is a problem of computations of coefficients @xmath46 ( [ eq : alpha ] ) , @xmath55 ( [ eq : beta ] ) of reduced algebraical system .
these problems may be explicitly solved in wavelet approach .
the obtained solutions are given in the form ( [ eq : pol3 ] ) , where @xmath56 are basis functions and @xmath57 are roots of reduced system of equations . in our case @xmath56
are obtained via multiresolution expansions and represented by compactly supported wavelets and @xmath57 are the roots of corresponding general polynomial system ( [ eq : pol2 ] ) .
our constructions are based on multiresolution approach .
because affine group of translation and dilations is inside the approach , this method resembles the action of a microscope .
we have contribution to final result from each scale of resolution from the whole infinite scale of spaces .
more exactly , the closed subspace @xmath58 corresponds to level j of resolution , or to scale j. we consider a multiresolution analysis of @xmath59 ( of course , we may consider any different functional space ) which is a sequence of increasing closed subspaces @xmath60 : @xmath61 satisfying the following properties : @xmath62 so , on fig.1 we present contributions to bunch oscillations from first 5 scales or levels of resolution . it should be noted that such representations ( 1 ) , ( 15 ) for solutions of equations ( 3)-(5 ) give the best possible localization properties in corresponding phase space .
this is especially important because our dynamical variables corresponds to moments of ensemble of beam particles .
in contrast with different approaches formulae ( 1 ) , ( 15 ) do not use perturbation technique or linearization procedures and represent bunch oscillations via generalized nonlinear localized eigenmodes expansion .
rosenzweig , fundamentals of beam physics , e - ver- + sion : http://www.physics.ucla.edu/class/99f/250rosenzweig/notes/ l. serafini and j.b .
rosenzweig , _ phys .
e _ * 55 * , 7565 , 1997 . c. allen , t. wangler , papers in ucla icfa proc .
, nov . , 1999 , world sci . , 2000 .
a.n . fedorova and m.g .
zeitlin , wavelet approach to mechanical problems .
symplectic group , symplectic topology and symplectic scales , _ new applications of nonlinear and chaotic dynamics in mechanics _ , 31,101 ( kluwer , 1998 ) .
fedorova , m.g .
zeitlin and z. parsa , variational approach in wavelet framework to polynomial approximations of nonlinear accelerator problems .
* cp468 * , 48 ( american institute of physics , 1999 ) .
+ los alamos preprint , physics/990262 a.n . fedorova and m.g .
zeitlin , nonlinear accelerator problems via wavelets , parts 1 - 8 , proc .
pac99 , 1614 , 1617 , 1620 , 2900 , 2903 , 2906 , 2909 , 2912 ( ieee / aps , new york , 1999 ) .
+ los alamos preprints : physics/9904039 , physics/9904040 , physics/9904041 , physics/9904042 , physics/9904043 , physics/9904045 , physics/9904046 , physics/9904047 . | we present applications of variational wavelet approach to different forms of nonlinear ( rational ) rms envelope equations .
we have the representation for beam bunch oscillations as a multiresolution ( multiscales ) expansion in the base of compactly supported wavelet bases . |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.