article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
u scorpii is one of the well observed recurrent novae , characterized by the shortest recurrence period ( @xmath17 yr ) , the fastest decline of its light curve ( 0.6 mag day@xmath16 ) , and its extremely helium - rich ejecta ( he / h@xmath18 by number ; see , e.g. , @xcite for summary ) .
historically , the outbursts of u sco were observed in 1863 , 1906 , 1936 , 1979 , 1987 , and the latest in 1999 .
especially , the 1999 outburst was well observed from the rising phase to the cooling phase by many observers ( e.g. , @xcite ) , including eclipses ( @xcite ) , thus providing us with a unique opportunity to construct a comprehensive model of u sco during the outburst .
our purpose in this letter is to construct a detailed theoretical model for u sco based on our light - curve analysis .
a part of the method has been described in hachisu & kato ( 1999 ) , in which they explain the second peak of t crb outbursts , but we again briefly summarize it in 2 , and will fully describe it in a separate paper . in 3 , by directly fitting our theoretical light curve to the observational data , we derive various physical parameters of u sco .
discussions follow in 4 , especially in relation to the recently proposed progenitor model of type ia supernovae ( sne ia ) .
schaefer ( 1990 ) and schaefer & ringwald ( 1995 ) observed eclipses of u sco in the quiescent phase and determined the orbital period ( @xmath19 days ) and the ephemeris ( hjd 2,451,235.777 @xmath20 1.23056@xmath21 ) at the epoch of mideclipse .
it is very likely that the companion is a slightly evolved main - sequence star ( ms ) that expands to fill its roche lobe after a large part of the central hydrogen is consumed .
the phase duration ( @xmath22 ) of the primary eclipses in quiescent phase indicates an inclination angle of @xmath23 ( e.g. , @xcite ) .
our model is shown graphically in figure [ usco_fig80_w30 ] .
we have revised kato s ( 1990 ) u sco light - curve models in the following two ways : ( 1 ) the opacity has been changed from the los alamos opacity ( cox , king , & tabor 1973 ) to the opal opacity ( @xcite ) and ( 2 ) the reflection / irradiation effects both of the companion star and the accretion disk are introduced in order to follow the light curve during the entire phase of the outburst .
the visual light curve is calculated from three components of the system : ( 1 ) the white dwarf ( wd ) photosphere , ( 2 ) the ms photosphere , and ( 3 ) the accretion disk surface . in the thermonuclear runaway model ( e.g. , starrfield , sparks , & shaviv 1988 ) , wd envelopes
quickly expand to @xmath24 or more and then the photospheric radius gradually shrinks to the original size of the white dwarfs .
correspondingly , the optical luminosity reaches its maximum at the maximum expansion of the photosphere and then decays toward the level in the quiescent phase , keeping the bolometric luminosity almost constant .
since the wd envelope reaches a steady - state after maximum , we are able to follow the development by a unique sequence of steady - state solutions as shown by kato & hachisu ( 1994 ) .
optically thick winds , blowing from the wd in the decay phase of novae , play a key role in determining the nova duration because a large part of the envelope mass is quickly blown off in the wind . in the decay phase , the envelope structure at any given time
is specified by a unique solution .
the envelope mass @xmath25 is decreasing because of both the wind mass loss at a rate of @xmath26 and the hydrogen shell burning at a rate of @xmath27 , i.e. , @xmath28 where @xmath29 is the mass accretion rate of the wd . by integrating equation ( [ dmdt_envelope_mass ] ) , we follow the development of the envelope and
then obtain the evolutionary changes of physical quantities such as the photospheric temperature @xmath30 , the photospheric radius @xmath31 , the photospheric velocity @xmath32 , the wind mass loss rate @xmath26 , and the nuclear burning rate @xmath27 .
when the envelope mass decreases to below the critical mass , the wind stops , and after that , the envelope mass is decreased only by nuclear burning .
when the envelope mass decreases further , hydrogen shell - burning disappears , and the wd enters a cooling phase .
after the optical peak , the photosphere shrinks with the decreasing envelope mass mainly because of the wind mass loss .
then the photospheric temperature increases and the visual light decreases because the main emitting region moves blueward ( to uv then to soft x - ray ) . we have assumed a blackbody photosphere of the white dwarf envelope and estimated visual magnitude of the wd photosphere with a window function given by allen ( 1973 ) .
the photospheric surface is divided into 32 pieces in the latitudinal angle and into 64 pieces in the longitudinal angle as shown in figure [ usco_fig80_w30 ] . for simplicity
, we do not consider the limb - darkening effect . to construct a light curve , we have also included the contribution of the companion star irradiated by the wd photosphere .
the companion star is assumed to be synchronously rotating on a circular orbit and its surface is also assumed to fill the inner critical roche lobe , as shown in figure [ usco_fig80_w30 ] . dividing the latitudinal angle into 32 pieces and the longitudinal angle into 64 pieces , we have also summed up the contribution of each patch , but , for simplicity , we neglect both the limb - darkening effect and the gravity - darkening effect of the companion star . here
we assume that 50% of the absorbed energy is reemitted from the companion surface with a blackbody spectrum at a local temperature .
the original ( nonirradiated ) photospheric temperature of the companion star is assumed to be @xmath33 k because of @xmath34 in eclipse ( @xcite ; see also @xcite ) .
two other cases of @xmath35 and 4400 k have been examined but do not show any essential differences in the light curves .
we have included the luminosity coming from the accretion disk irradiated by the wd photosphere when the accretion disk reappears several days after the optical maximum , i.e. , when the wd photosphere shrinks to smaller than the size of the inner critical roche lobe .
then we assume that the radius of the accretion disk is gradually increasing / decreasing to @xmath36 in a few orbital periods , where @xmath37 is a numerical factor indicating the size of the accretion disk and @xmath38 is the effective radius of the inner critical roche lobe for the primary wd component .
the surface of the accretion disk absorbs photons and reemits the absorbed energy with a blackbody spectrum at a local temperature . here
, we assume that 50% of the absorbed energy is emitted from the surface while the other is carried into the interior of the accretion disk and eventually brought into the wd .
the original temperature of the disk surface is assumed to be constant at @xmath39 k including the disk rim .
the viscous heating is neglected because it is much smaller than that of the irradiation effects .
we also assume that the accretion disk is axisymmetric and has a thickness given by @xmath40 where @xmath41 is the height of the surface from the equatorial plane , @xmath42 the distance on the equatorial plane from the center of the wd , and @xmath43 is a numerical factor showing the degree of thickness .
we have adopted a @xmath42-squared law simply to mimic the effect of the flaring up of the accretion disk rim ( e.g. , schandl , meyer - hofmeister , & meyer 1997 ) and the exponent does not affect the disk luminosity so much mainly because the central part of the disk is not seen .
the surface of the accretion disk is divided into 32 pieces logarithmically and evenly in the radial direction and into 64 pieces evenly in the azimuthal angle as shown in figure [ usco_fig80_w30 ] .
the outer rim of the accretion disk is also divided into 64 pieces in the azimuthal direction and 8 pieces in the vertical direction by rectangles .
we have reproduced the light curves in the quiescent phase ( @xmath44-magnitude , @xcite ) by adopting a set of parameters such as @xmath45 , @xmath46 , a 50% irradiation efficiency of the accretion disk , and @xmath47 for the wd luminosity of @xmath48 ( @xmath49 and @xmath50 kpc , or @xmath51 and @xmath52 kpc , see below ) , which is roughly corresponding to the mass accretion rate of @xmath53 yr@xmath16 for the @xmath6 wd .
figure [ vmag1370va1_m15 ] shows the observational points and our calculated @xmath54-magnitude light curve ( solid line ) . to maintain an accretion disk , we assume a mass accretion rate @xmath55 yr@xmath16 during the outburst . to fit the early linear decay phase ( @xmath56 days after maximum ) , we have calculated a total of 140 @xmath54-magnitude light curves for five cases of the wd mass : @xmath57 , 1.37 , 1.36 , 1.35 , and @xmath58 , with various hydrogen content of @xmath59 , 0.05 , 0.06 , 0.07 , 0.08 , 0.10 , and 0.15 of the envelope , where the metallicity @xmath60 is fixed , each for four companion mass of @xmath61 , 1.1 , 1.5 , and @xmath62 .
we choose @xmath63 as a limiting mass just before the sn ia explosion in the w7 model ( @xmath64 ) of nomoto , thielemann , & yokoi ( 1984 ) .
we have found that the early 7 day light curve hardly depends on the chemical composition and the companion mass but is mainly determined by the white dwarf mass .
this is because ( 1 ) the early - phase light curve is determined mainly by the wd photosphere ( fig .
[ vmag1370va1_m15 ] , dotted line ) and therefore by the wind mass loss rate and ( 2 ) the optically thick wind is driven by the strong peak of opal opacity , which is due not to helium or hydrogen lines but to iron lines ( @xcite ) . therefore , the determination of the wd mass is almost independent of the hydrogen content , the companion mass , or the disk configuration .
the @xmath6 light curve is in much better agreement with the observations than are the other wd masses .
the distance to u sco is estimated to be 5.48.0 kpc , as shown in figure [ vmag1370va1_m15 ] , if we assume no absorption ( @xmath65 ) . here
, we obtain 5.4 kpc for the fit to the upper bound and 8.0 kpc for the fit to the lower bound of the observational points . for an absorption of @xmath66 ( @xcite ) , we have the distance of 4.16.1 kpc , and then u sco is located 1.52.3 kpc above the galactic plane ( @xmath67 ) .
these ranges of the distance are reasonable compared with the old estimates of the distance to u sco such as 14 kpc ( e.g. , @xcite ) . to fit the cooling phase ( @xmath68 3040 days after maximum )
, we must adopt the hydrogen content of @xmath69 among @xmath59 , 0.05 , 0.06 , 0.07 , 0.08 , 0.10 , and 0.15 for @xmath70 .
this is because the hydrogen content @xmath71 is equivalent to the mass of hydrogen burning in the envelope .
therefore , @xmath71 determines the duration of hydrogen shell burning , i.e. , the duration of the midplateau phase . for @xmath69 , the optically thick wind stops at @xmath72 days , and the steady hydrogen shell - burning ends at @xmath73 days .
this duration of the strong wind phase is very consistent with the bepposax supersoft x - ray detection 1920 days after the optical peak ( @xcite ) because supersoft x - rays are self - absorbed by the wind itself during the strong wind phase .
it should be noted that we do not use this detection of the supersoft x - rays to constrain any physical parameters .
hydrogen shell - burning begins to decay from @xmath73 days but still continues to supply a part of the luminosity ; the rest comes from the thermal energy of the hydrogen envelope and the hot ash ( helium ) below the hydrogen layer .
the thermal energy amounts to several times 10@xmath74 ergs , which can supply a bolometric luminosity of 10@xmath75 erg s@xmath16 for ten days or so until @xmath76 days , as seen in figures [ vmag1370va1_m15 ] and [ vmag1370ecl_m15 ] .
the envelope mass at the peak is estimated to be @xmath77 for @xmath78 , @xmath69 , and @xmath60 ; thus , the average mass accretion rate of the wd becomes @xmath15 yr@xmath16 in the quiescent phase between 1987 and 1999 , if no wd matter is dredged up into the envelope .
such high mass accretion rates strongly indicate that the mass transfer is driven by a thermally unstable mass transfer under the constraint that the companion star is a slightly evolved main - sequence star ( e.g. , @xcite ) .
the wind carries away about 60% of the envelope mass , i.e. , @xmath79 , which is much more massive than the observational indication of @xmath80 in the 1979 outburst by williams et al .
the residual , @xmath81 , can accumulate in the wd .
therefore , the wd can grow in mass at an average rate of @xmath82 yr@xmath16 .
our model fully reproduces the observational light curve if we choose @xmath83 and @xmath84 during the wind phase and @xmath85 and @xmath86 during the cooling phase for @xmath6 wd @xmath20 @xmath87 ms with @xmath88 .
since we obtain similar light curves for four companion masses , i.e. , 0.8 , 1.1 , 1.5 , and 2.0 @xmath89 , we show here only the results of @xmath90 .
it is almost certain that the luminosity exceeds the eddington limit during the first day of the outburst , because our model can not reproduce the super eddington phase .
eclipses in the 1999 outburst were observed by matsumoto & kato ( 1999 ) . the depth of the eclipses is @xmath91 , and it is much shallower than that in quiescent phase ( @xmath92 , @xcite ) .
their observation also indicates almost no sign of the reflection effect by the companion star .
thus , we are forced to choose a relatively large disk radius that exceeds the roche lobe ( @xmath93 ) and a large flaring - up edge of the disk ( @xmath94 ) . if we adopt a size of the accretion disk that is smaller than the roche lobe , on the other hand , the companion star occults completely both the accretion disk and the white dwarf surface . as a result
, we obtain a deep primary minimum as large as 1.52.0 or more .
we have tested a total of 140 cases of the set ( @xmath95 ) , i.e. , 14 cases of @xmath96 by @xmath97 step each for 10 cases of @xmath98 by @xmath99 step .
the best fit light curve is obtained for @xmath100 and @xmath46 .
then a large part of the light from the wd photosphere is blocked by the disk edge , as shown in figure [ usco_fig80_w30 ] . the large disk size and flaring - up edge
may be driven by the kelvin - helmholtz instability because of the velocity difference between the wind and the disk surface .
after the optically thick wind stops , photon pressure may drive the surface flow on the accretion disk ( @xcite ) , and we may expect an effect on the accretion disk similar to what the wind has , but it may be much weaker .
thus , we may have a smaller radius of @xmath101 but still have @xmath102 .
a much more detailed analysis of the results will be presented elsewhere .
it has been suggested that u sco is a progenitor of sne ia ( e.g. , starrfield , sparks , & truran 1985 ) because its wd mass is very close to the critical mass ( @xmath103 ) for the sn ia explosion ( @xcite ) .
hachisu et al .
( 1999 ) proposed a new evolutionary path to sne ia , in which they clarified the reason why the companion star becomes helium - rich even though it is only slightly evolved from the zero - age main sequence .
a typical progenitor of sne ia in their wd+ms model , @xmath104 , @xmath105 , and @xmath106 yr@xmath16 just before the sn ia explosion , has exactly the same system as u sco .
thus , it is very likely that the wd in u sco is now growing in mass at an average rate of @xmath80 yr@xmath16 toward the critical mass for the sn ia explosion and will soon explode as an sn ia if the wd core consists of carbon and oxygen .
we are very grateful to many amateur astronomers who observed u sco and sent their valuable data to vs - net .
we also thank the anonymous referee for many critical comments to improve the manuscript .
this research has been supported in part by the grant - in - aid for scientific research ( 08640321 , 09640325 , 11640226 , 20283591 ) of the japanese ministry of education , science , culture , and sports .
k. m. has been supported financially as a research fellow for young scientists by the japan society for the promotion of science . | a theoretical light curve for the 1999 outburst of u scorpii is presented in order to obtain various physical parameters of the recurrent nova .
our u sco model consists of a very massive white dwarf ( wd ) with an accretion disk and a lobe - filling , slightly evolved , main - sequence star ( ms ) .
the model includes a reflection effect by the companion and the accretion disk together with a shadowing effect on the companion by the accretion disk .
the early visual light curve ( with a linear phase of @xmath0 days after maximum ) is well reproduced by a thermonuclear runaway model on a very massive wd close to the chandrasekhar limit ( @xmath1 ) , in which optically thick winds blowing from the wd play a key role in determining the nova duration . the ensuing plateau phase ( @xmath2 days ) is also reproduced by the combination of a slightly irradiated ms and a fully irradiated flaring - up disk with a radius @xmath3 times the roche lobe size .
the cooling phase ( @xmath4 days ) is consistent with a low hydrogen content @xmath5 of the envelope for the @xmath6 wd .
the best fit parameters are the wd mass @xmath7 , the companion mass @xmath8 ( @xmath9 is acceptable ) , the inclination angle of the orbit @xmath10 , and the flaring - up edge , the vertical height of which is @xmath11 times the accretion disk radius .
the duration of the strong wind phase ( @xmath12 days ) is very consistent with the bepposax supersoft x - ray detection at @xmath13 days because supersoft x - rays are self - absorbed by the massive wind .
the envelope mass at the peak is estimated to be @xmath14 , which is indicating an average mass accretion rate @xmath15 yr@xmath16 during the quiescent phase between 1987 and 1999 .
these quantities are exactly the same as those predicted in a new progenitor model of type ia supernovae . |
an important consequence of quantum gravity scenarios is the existence of a minimum length .
since the heisenberg uncertainty principle does not exert any restriction on the measurement precision of the particle s positions or momenta , there is no minimum measurable length in the usual heisenberg picture . in the past few years
, quantum mechanics with modification of the usual canonical commutation relations has been investigated intensively@xcite-@xcite .
such works , which are motivated by string theory and quantum gravity , suggest a minimal length as @xmath0 .
this means that there is no possibility to measure coordinate @xmath1 with accuracy smaller than @xmath2 , where @xmath3 is a positive parameter and depends on the expectation value of the position and momentum operators @xcite . since in string theory
the minimum observable distance is the string length , @xmath4 is proportional to this length .
if we set @xmath5 , the usual heisenberg algebra is recovered . in low energy limit , these quantum gravity effects can be neglected , but in circumstances such as the very early universe or in the strong gravitational field of a black hole one has to consider these effects .
the modification induced by the generalized uncertainty principle ( gup ) on the classical orbits of particles in a central force potential has been considered in @xcite . alternatively , it could be argued that a fundamental length in nature may have dramatic impacts on our universe .
also , it could be seen that this length may be related to vacuum energy , i.e. dark energy @xcite .
the introduction of @xmath6 in the formalism of gup has been studied in the literature @xcite-@xcite .
most of different analyses are performed by introducing an ultraviolet ( uv ) cut - off on the planck scale@xcite .
but , other works introduce an infrared ( ir ) cut - off by using the so - called q - bargmann fock formalism @xcite . in this formalism
, if we impose the @xmath7 as uv cut - off and the cosmological constant @xmath6 as an ir one , the uv scale is dual to the ir one . in this paper , we are going to proceed to study the effects of gup on the kepler problem .
the main consequence of this work is the constraint imposed on the minimal observable length and the parameter @xmath3 in comparison with observational data of mercury .
also , by comparing the value of the perihelion shift of mercury in schwarzschild - de sitter spacetime with the resulted value of gup , we obtain a relation between the minimal length and the cosmological constant .
now if the cosmological constant is time dependent , we will have a minimal length to vary with time .
this matter could be important in quantum gravity scenarios for studying the early universe and black hole physics .
finally , we investigate the stability of circular orbits of planets in this framework .
we show when the angular momentum of the planet is large , the condition for stability of circular orbits differs considerably from classical mechanics .
we consider the commutation relation with the gup as @xmath8 = i \hbar ( 1+\beta p^{2})\delta_{ij},\label{eq1}\ ] ] this commutation relation implies the minimal length uncertainty relation @xmath9 which appears in perturbative string theory @xcite . if the components of the momentum @xmath10 are assumed to commute with each other @xmath11=0 , \label{eq3}\ ] ] then the commutation relation among the coordinate are determined as @xmath12=2 i\hbar\beta ( p_{i}x_{j}-p_{j}x_{i } ) .
\label{eq4}\ ] ] since the commutation relations ( [ eq1 ] ) , ( [ eq3 ] ) and ( [ eq4 ] ) do not break the rotational symmetry , the generators of rotations can be expressed in terms of the position and momentum operators as follows @xmath13 in the classical limit , using the poisson bracket , we can write equations ( [ eq1 ] ) , ( [ eq3 ] ) and ( [ eq4 ] ) as follow @xmath14 the poison bracket must possess the same properties as the quantum mechanics commutators , namely , it must be anti - symmetric , bilinear and satisfy the leibniz rules and the jacobi identity .
therefore , we can derive the general form of the poisson bracket for any functions of the coordinates and momenta as @xmath15 also , we can obtain the time evolutions of the coordinates and momenta as follows @xmath16 now we can use these equations to study the motion of a macroscopic object . in this paper , we consider hamiltonian of a particle in a central potential @xmath17 the time evolution of the coordinates and momentums became @xmath18 where we have neglected terms of the order of @xmath19 . for motion in central force potential , the @xmath20 s are conserved due to rotational symmetry @xmath21 .
the conservation of the @xmath22 s imply that the motion of the particle will be confined to a 2-dimensional plane spanned by the coordinate and momentum vectors at any point in time .
now , by using equations ( [ eq9 ] ) we can drive the equation of motion of a particle @xmath23 in the next section , we study the kepler problem in this framework .
as we know that the kepler potential is @xmath24 , so the equation ( [ eq11 ] ) became @xmath25 since @xmath26 is a very small parameter and @xmath27 is very large , we can ignore the second term in comparison with the first term .
in spherical coordinates , the equations above can be written as @xmath28 @xmath29 @xmath30 we assume the case of the plane orbit that is a valid solution , and for simplicity we consider the equatorial orbits i.e @xmath31 . then the above equations become as follows @xmath32 @xmath33 @xmath34 according to equation ( [ eq18 ] ) , @xmath35 is a constant of motion so equation ( [ eq16 ] ) becomes @xmath36 now , for the right hand side of the equation above , we consider a new potential as @xmath37 the second term in parenthesis is very small compared to one .
so , by neglecting it we have @xmath38 as can be seen , the contribution of the radial components of the kinetic energy is less than ordinary classical mechanics by one .
here we shall solve equation ( [ eq19 ] ) to obtain the orbits of planets in the solar system .
first , we define @xmath39 then @xmath40 , @xmath41 and the radial equation ( [ eq19 ] ) changes to @xmath42 on the other hand , in the classical kepler problem we have @xmath43 solving this equation we obtain @xmath44 where @xmath45 and @xmath46 . now for solving equation ( [ eq23 ] ) , to first order in @xmath3 , we suggest @xmath47 where @xmath48 must satisfy @xmath49.,\label{eq27}\ ] ] by solving the equation above we obtain @xmath50,\label{eq28}\ ] ] so we have @xmath51+\frac{\delta}{b^{2}}(1 + e^{2}),\label{eq29}\ ] ] that @xmath52
. then we can obtain the perihelion shift per revolution as @xmath53 using the variables @xmath54 and @xmath55 , where @xmath56 is the mass of the sun , we have @xmath57 in the equation above , @xmath26 the same as @xmath58 and @xmath56 is constant . in the case of the mercury planet , using the data @xmath59 , @xmath60 , @xmath61 , @xmath62 , @xmath63 and @xmath64 , we found @xmath65 as @xmath66 for consistency with the observational results , the @xmath67 must be small .
then we conclude that the parameter @xmath68 must be very small and the planetary system is very sensible in this parameter .
so , small changes in @xmath3 implies sensible changes at very large scales .
in other words , there is a connection between the physics at small scales and the physics at large scales .
now , we consider the observed perihelion shift of mercury as @xcite @xmath69 if it is assumed that @xmath70 , we can obtain @xmath71 and @xmath72 note that this limit is @xmath73 orders of magnitude below the planck length . however , for a best comparison it will be necessary to study the shift to the perihelion in a curved space . as we know ,
in the case of general relativity with the schwarzschild - de sitter metric , the shift to the perihelion is @xcite @xmath74 so , for the perihelion shift of mercury @xmath75(\frac{rad}{rev}),\label{eq37}\ ] ] now , if @xmath6 is of the order of @xmath76 @xcite , the second term above is the of same order as the observational error .
since the effect of spacetime quantization for a composite macroscopic body is much weaker than one of the constituent particles @xcite , we should consider the effect of gup on the perihelion shift as a perturbation effect which is very small . using this assumption , we can find a upper bound for @xmath3 as @xmath77 then we obtain @xmath78 and @xmath79 by defining @xmath80 the equation ( [ eq40 ] ) can been written as @xmath81 if @xmath82 then @xmath83 and vice versa .
it means that the cosmological constant of the spacetime is related to the minimal observable length . without a the minimal length ,
there is no cosmological constant and vice versa .
it can been seen that the minimal length @xmath84 and the cosmological constant scale @xmath85 could be dual of each other@xcite-@xcite .
on the other hand , there is an extensive literature suggesting that the relation @xmath86 plays a fundamental role in cosmology ( here @xmath87 is the hubble parameter ) @xcite .
this relation has been obtained in a number of empirical models @xcite.the dependence @xmath86 explains the current observations successfully and provides a more accurate age for the universe .
according to this relation , the cosmological constant varies with cosmic time and therefore the minimal length of spacetime decays with time .
this matter could be very important in quantum gravity scenarios such as string theory .
for studying the stability of the planet circular orbits , we start from equation ( [ eq16 ] ) . using the general form of a central force @xmath88 we have @xmath89 then , by considering @xmath90 ,
the effective potential can be written to first order in @xmath3 as @xmath91 where @xmath92 is the mass of the planet and in a planet orbit @xmath93 .
now , the conditions of a stable circular orbit are @xmath94 where @xmath95 is the radius of orbit . therefore , we should have @xmath96 and @xmath97 from the equation ( [ eq47 ] ) we have @xmath98 using this result , the second condition for stability of circular orbits in central force can be written as @xmath99 this condition for stability differs from the one in general classical mechanics .
when @xmath100 this results in @xmath101 which is a known condition in general classical mechanics .
however , when @xmath102 , the condition for stability of circular orbits depends on the angular momentum .
since @xmath26 is too small , the second term in equation ( [ eq49 ] ) is effective when the angular momentum of the planet is very large .
also , @xmath103 could have non - integer values .
in this paper , we have investigate the effects of gup on the kepler problem .
in other words we found the equation of motion and the perihelion shift of the planets in this framework .
also , by comparing the value of the perihelion shift of mercury in schwarzschild - de sitter spacetime with the resulted value of gup , we showed that the minimal length @xmath84 and the cosmological constant scale @xmath104 could be dual of each other .
since the cosmological constant could depend on the hubble parameter , we concluded that the minimal length of the spacetime varies with cosmic time . in the end
, we studied the stability of circular orbits of planets and showed that the condition for stability is different from general classical mechanics , if the angular momentum of the planet is very large .
we wish to thank h. r. sepangi for his careful reading of the article and for his constructive comments .
99 d. amati , m. ciafaloni and g. veneziano , phys .
b 216 , 41 ( 1989 ) .
m. maggiore , phys . rev .
d 49 , 5182 ( 1994 ) , arxiv : hep - th/9305163 . l. j. garay , int . j. mod
. phys . a 10 , 145 ( 1995 ) .
a. kempf , g. mangano , and r. b. mann , phys .
d 52 , 1108 ( 1995 ) .
j. magueijo , l. smolin , phys .
88 ( 2002 ) 190403 , arxiv : hep - th/0112090
. j. magueijo , l. smolin , phys . rev .
d 71 ( 2005 ) 026010 , arxiv : hep - th/0401087 . s. das and e. c. vagenas , phys .
lett . 101 , 221301 ( 2008 ) .
a. f. ali , s. das and e. c. vagenas , phys .
b 678 , 497 ( 2009 ) .
p. pedram , int .
d 19 , 2003 ( 2010 ) .
s. benczik , l. n. chang , d. minic , n. okamura , s. rayyan and t. takeuchi , phys .
d 66 ( 2002 ) 026003 .
x. calmet , modern physics letters a , 22:2027 - 2034 ( 2007 ) .
l. n. chang , z. lewis , d. minic and t. takeuchi , adv .
high energy phys .
2011 , 493514 ( 2011 ) .
p. chen and r. j. adler , nucl .
124 , 103 - 106 ( 2003 ) .
r. j. adler , p. chen and d. i. santiago , gen .
33 , 2101 - 2108 ( 2001 ) .
a. kempf , j. math .
38 , 1347 - 1372 ( 1997 ) . s. pireaux , j. p. rozelot and s. godier , astrophys.space sci .
284 , 1159 - 1194 ( 2003 ) , astro - ph/0109032 .
j. m. romero and j. d. vergara , mod.phys.lett .
a , 18 , 1673 - 1680 ( 2003 ) , arxiv : hep - th/0303064 .
wickramasinghe.t , international symposium on astrophysics research and science education , vatican observatory ( 1999 ) .
v. kagramanova , j. kunz and c. lammerzahl , phys.lett .
b 634 , 465 - 470 ( 2006 ) , arxiv : gr - qc/0602002 . g. amelino - camelia , prl 111 , 101301 ( 2013 ) . m.s . berman , phys .
d 43 , 1075 ( 1991 ) .
i. chakrabarty and a. pradhan , gravitation and cosmology 7 , 55 ( 2001 ) .
d. kalligas , p.s .
wesson and c.w .
everitt , gen .
24 , 351 ( 1992 ) . | various approaches to quantum gravity , such as string theory , predict a minimal measurable length and a modification of the heisenberg uncertainty principle near the plank scale , known as the generalized uncertainty principle ( gup ) . here
we study the effects of gup which preserves the rotational symmetry of the spacetime , on the kepler problem . by comparing the value of the perihelion shift of the planet mercury in schwarzschild - de sitter spacetime with the resulted value of gup
, we find a relation between the minimal measurable length and the cosmological constant of the spacetime . now ,
if the cosmological constant varies with time , we have a variable minimal length in the spacetime .
finally , we investigate the effects of gup on the stability of circular orbits .
+ keywords : generalized uncertainty principle , kepler problem , variable minimal length .
+ pacs numbers : 04.60.-m , 04.60.bc , 04.80.cc -2.5 cm |
the hot diffuse x - ray emitting gas ( intracluster medium , icm for short ) provides a powerful tool to inspect the internal dynamics of galaxy clusters . for the typical density and temperature of the intracluster gas ,
the main emission mechanism is the bremsstrahlung and , for a large amount of clusters , the radiative cooling time in the central regions is significantly shorter than the hubble time . as a consequence ,
if no additional heating mechanism is present , the gas cools and is expected to flow inwards , forming a _ cooling flow_. the standard model of cooling flows ( see * ? ? ? * for a review ) predicted the gas to be a multiphase medium in which there is a broad range of temperatures and densities present at all radii .
mass deposition rates were estimated to be as large as hundreds of solar masses per year @xcite .
this model was strengthened by the general thought that in presence of magnetic fields the thermal conduction must be highly suppressed @xcite , which is a necessary condition for the multiphase cooling to operate .
in fact , no heating exchange between the different phases must occur in order that they may coexist .
there is some observational evidence that modest magnetic fields are present throughout the intracluster medium .
the current measurements of intracluster magnetic fields are based on faraday rotation measure ( rm ) in radio sources seen through clusters ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) ; direct evidence also comes from measurements of extended regions of radio synchrotron emission in clusters ( see e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
both the excess rm values and the radio halo data suggest modest magnetic fields , at a few microgauss levels , throughout the cluster .
recent and _ chandra _ observations have shown that in the central regions , the temperature drops to about one third of its overall mean value and there is no evidence of temperatures smaller than @xmath2 kev @xcite , suggesting that the gas does not cool below these cutoff temperatures . moreover , the new estimated mass deposition rates are significantly smaller than those evaluated by using previous x - ray satellites data @xcite .
lastly , the new data show that clusters spectra are better represented by a single ( or double ) temperature model rather than the standard multiphase ( multi - temperature ) cooling flow model @xcite .
these new results clearly show that the standard cooling flow model is not a satisfying description of the internal dynamics of the icm .
some source of heat which stops the cooling flow and balances radiative losses must be sought .
the nature of this source and the origin of the heat mechanism is still unclear .
one possible candidate is thermal conduction .
recent works by @xcite and @xcite show that in the presence of turbulent magnetic fields , the conductivity can be as large as a fraction of the spitzer value and thus can play a significant role in balancing cooling flows . as a consequence
, thermal conduction has been recently re - introduced as a possible heat source to balance the energy losses ( see e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
however , as we will discuss more in detail in [ sub : cond ] , thermal conduction fails in supplying the needed heat in the central regions ( see also * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
heating from a central active galactic nucleus is another possibility .
the idea is supported by the fact that most of `` cooling clusters '' host a central active galactic nucleus with strong radio activity @xcite .
several models in the literature explore agn heating processes to assess if they can balance the radiative losses .
one of the most appealing mechanisms involves buoyant gas bubbles , inflated by the agn , that subsequently rise through the cluster icm heating it up @xcite . however , all the models predicting that radiative cooling is balanced only by energy input from the central agn fail @xcite .
in particular , @xcite analyzed several heating mechanisms induced by the central agn and concluded that no simple mechanism is able to quench the cooling flow .
moreover , the required mechanism needs a finely tuned heating source .
indeed , the heat source must provide sufficient energy to stop the cooling flow , but not enough to trigger strong convection or the metallicity gradients observed in all cooling flow clusters @xcite would be destroyed . as a consequence
, an agn can be an efficient mechanism in the very center of the cluster but it is unlikely to be strong enough to provide energy to the outer parts of the `` cooling region '' .
so , it may be viewed as complementary to thermal conduction which fails in quenching the cooling flow in the innermost regions .
recently , @xcite and @xcite concluded that both thermal conduction and heating from a central agn can play an important role in balancing the cooling .
in particular , @xcite ( rb02 hereafter ) developed a model where both thermal conduction and heating from a central agn co - operate in balancing the radiative losses .
one of the main advantages of this model is that it reaches a stable final equilibrium state and it is able to reproduce the main observed quantities , such as the temperature profile ( with a minimum temperature @xmath3 kev ) . in this paper
, we use m87/virgo observations of three satellites ( namely , _
chandra _ and _ beppo - sax _ ) to test various heating models on this cluster . to this end
, we apply to the m87 data the deprojection technique to recover some physical quantities of the icm such as the gravitational mass , the entropy and the heating required to balance the cooling flow .
the paper is organized as follows : in [ sec : sp2d ] we report details about the analysis of the three ( _ chandra _ , and _ beppo - sax _ ) m87 datasets ; in [ sec : depro ] , we revise briefly the spectral deprojection technique that we adopt for our analysis ; in [ sec : m87 ] , we deproject the m87 data , we test that the spherical symmetry hypothesis holds and we derive the gravitational mass for m87 ; in [ sec : heat ] , we determine the amount of extra - heating needed to balance the cooling flow when thermal conductivity is assumed to operate at a fraction of the spitzer value .
lastly , in [ sec : concl ] we summarize our results .
thanks to its proximity , the virgo cluster and its giant elliptical central galaxy m87 , represent an incomparable target to inspect the internal properties of the icm . aiming to a precise characterization of the icm , we use observations of the three satellites _
chandra _ , and _ beppo - sax_. these satellites views are complementary : the sharp psf ( @xmath4 ) of _ chandra _ can provide a precise analysis of the innermost ( say @xmath5 kpc ) regions of m87 ; the large collecting area and its wide field of view allow a good inspection of the intermediate regions ( up to @xmath6 kpc ) ; in the outermost regions of the cluster , where the angular resolution is less critical and data are highly contaminated by the background , _ beppo - sax _ is a better choice and data can be collected out to a radius of @xmath7 kpc .
m87 has been observed during the pv phase of .
the details of this observation have been widely discussed in several publications ( see e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we make use of the results of the spectral analysis described and discussed in ( * ? ? ?
* m02 hereafter ) .
the cluster is divided in 139 regions for 12 concentric annuli , centered on the emission peak .
the regions are the same as those presented in m02 , apart from the annuli in the @xmath8 arcmin range which have been taken 0.5 arcmin wide instead of 1 arcmin wide . unlike @xcite , we decided to use annuli at least @xmath9 wide in order to avoid possible psf contaminations ( see also * ? ? ?
for a detailed description of the mos psf see @xcite .
we refer the reader to m02 for all the details of the spectral analysis procedure and remind the reader that the accumulated spectra in all the regions are fitted with two different models : ( i ) a single temperature ( 1 t ) model ( _ vmekal _ in ) and ( ii ) a two temperature ( 2 t ) model ( _ vmekal _ + _ vmekal _ in ) .
the 1 t fits of the accumulated spectra provide for each region the emission - weighted temperature @xmath10 of the gas and the emission integral @xmath11 where @xmath12 and @xmath13 are the electron and proton density .
the 2 t fits of the spectra provide for each region the temperatures and the emission integrals of the two different components of the gas .
these quantities will be used to derive the density and the temperature profile of the cluster .
we have analyzed the _
_ acis - s3 observation ( obs .
352 ; see also * ? ? ?
* ) centered on m87 ( @xmath14:@xmath15:@xmath16 ; @xmath17:@xmath18:@xmath19 ) using ciao 2.1.1 and caldb 2.15 .
we have followed the procedures described in the science threads available at the chandra x - ray center on - line pages .
the light curve was filtered for high background events obtaining an effective exposure time of 35.2 ks . from our analysis
we excluded the agn and the associated jet cutting off a narrow rectangular region centered in @xmath14:@xmath15:@xmath20 , @xmath17:@xmath18:@xmath21 ( j2000 ) with lengths @xmath22 , @xmath23 and rotated by @xmath24 from e to n. we have extracted spectra from regions of concentric annuli centered on the emission peak ( @xmath25 , @xmath26 , @xmath27 , @xmath28 , @xmath29 , @xmath30 and @xmath31 ) ; the @xmath26 annulus was divided into four 90@xmath32 regions starting from a position angle of 45@xmath32 , all the other annuli have been divided into 8 regions 45@xmath32 wide starting from position angle 0@xmath32 .
the background used in the spectral fits was extracted from blank - sky observations using the _ acis - bkgrnd - lookup _ script .
we have fitted each spectrum with the same models ( 1 t and 2 t ) used for the analysis of the spectra and using the effective areas and response matrix derived with the routines _ mkwarf _ and _ mkwrmf _ for extended sources given in ciao .
the energy range is the same as the one used for the spectral analysis .
we have analyzed the pointed _ beppo - sax _ observation of the virgo cluster ( obs . i d .
60010001 ) , adding the mecs2 and mecs3 data and obtaining an effective exposure time of 25.1 ks .
the analysis of data follows the procedure described in details in @xcite .
we have extracted spectra for 7 concentric annuli centered on the emission peak ( @xmath33 deg , @xmath34 deg j2000 ) , each annulus is 2@xmath35 wide up to 10@xmath35 from the peak and 4@xmath35 wide from 10@xmath35 to the maximum 20@xmath35 radius .
we have fitted each spectrum with a _ mekal _ model absorbed for the galactic @xmath36 using the appropriate effective area computed for extended source as described in @xcite and the background spectrum extracted from blank - sky files for the same annular regions .
the energy range considered in the spectral analysis is @xmath37 kev in all cases out of the @xmath38 annulus for which we have considered the @xmath39 kev range to avoid spectral distortions from the supporting structure of the instrument entrance windows .
note that , as we will discuss in the next paragraph , for _ beppo - sax _ data , only the 1 t model has been used to fit the accumulated spectra
. for each region observed with or _
chandra _ , an f - test is used to establish whether the 2 t model provides a better description of the data with respect to the 1 t model .
we used two different criteria for and _ chandra _ data to select the regions represented by a 2 t model . as far as data are concerned , for those regions
whose f - test provides a probability @xmath40 , the 2 t description is retained , whereas for those regions whose f - test provides a probability @xmath41 , the 1 t description is adopted .
the regions having an f - test probability within the [ 0.90 - 0.95 ] range have been rejected and excluded from the analysis .
a somewhat more stringent criterion has been adopted for the _ chandra _ data , regions with f - test probability within the [ 0.75 - 0.98 ] range have been rejected , because of the lower statistical quality of the latter dataset .
it is worth noting that m02 shows ( using data ) that the regions which are better represented by a 2 t model match the radio `` arms '' which are visible in the m87 map at 90 cm @xcite . at large radii , where we have no evidence of a second temperature component from the data , we will use the results of 1 t fits to the _ beppo - sax _ data , which , because of its limited spectral coverage , is insensitive to the cooler component . while the cool component is related to the radio `` arms '' , the hot component in the regions described by the 2 t model , is very similar to the 1 t gas of the other regions which do not feature strong radio emission and are located at similar radial distances from the cluster core . in fig .
[ fig : cfrt_s ] we plot the emission weighted temperatures for 1 t and 2 t regions .
the open circles represent the temperatures of the 1 t regions , while the triangles plot the temperature of the hot component in those regions which are fitted by a 2 t model .
we plot error bars for a few representative points .
all the other error bars are not reported for a clear reading of the plot . the fig .
[ fig : cfrt_s ] shows that for those annuli where we have 1 t and 2 t regions , the temperature of the hot component of the 2 t regions is not separate from the temperature of the single phase gas of the 1 t regions .
so , the hot component and the single phase gas are distributed in a regular and symmetric fashion . as far as only the hot component
is considered for the 2 t regions , the cluster appears to be approximatively spherically symmetric , which is an important condition for the application of the deprojection technique .
as already outlined , for each region we consider the emission integral @xmath42 and the emission - weighted temperature @xmath10 .
for those regions fitted by the 2 t model , we retain the @xmath42 and @xmath10 related to the hot component .
the radially averaged profile for each physical quantity is determined starting from the region - by - region description .
we assign to each annulus , a mean @xmath10 by averaging on all the @xmath10 of the regions belonging to the annulus .
the @xmath42 is the sum of the @xmath42 of the regions along the ring .
some attention must be paid in this averaging ( or summation for @xmath42 ) procedure as some portions of the observed regions can be masked .
in fact , there are some pixels of the field of view of the different instruments which must be rejected for different reasons : ( i ) they are ccd hot or dead pixels ; ( ii ) they correspond to the gaps between nearby ccds ; ( iii ) they correspond to regions which are never observed by the instruments ( for mecs ) or which are partially outside the field of view ; ( iv ) they correspond to some point source contaminating the x - ray cluster emission ; ( v ) they correspond to the excluded agn rectangular regions ( _ chandra _ ) . as a consequence , even if the geometrical area ( or equivalently the emitting volume ) of the regions belonging to the same ring is the same , the effective emitting area ( or volume ) is a fraction of the geometrical area depending on the number of rejected pixels .
the measured @xmath42 in each region accounts for photons coming from the effective emitting area ( or volume ) ; hence , each @xmath42 must be renormalized .
the factor of normalization is given by the ratio @xmath43 where @xmath44 is the geometric area of the region and @xmath45 is the total area of the rejected pixels in the region .
the same normalization factors are used as weights for the determination of the averaged @xmath10 . in fig .
[ fig : t2d - ei ] , we plot the averaged profiles for @xmath10 and @xmath46 , respectively in panels ( a ) and ( b ) .
circles refer to data , squares to _ chandra _ data and triangles refer to _ beppo - sax _ data .
the values for @xmath10 and @xmath46 are also reported in table [ tab : t2d - ei ] .
the emission integral is reported and plotted in units , @xmath47ei $ ] where @xmath48 is the angular distance of the source in cm , @xmath49 is the redshift and @xmath42 is in @xmath50 .
@xmath51 is the area of the ring in arcmin@xmath52 .
note that error bars for the @xmath42 are quite small , especially for the data thanks to its high effective area which allows a very precise measure of the emission integral up to @xmath53 kpc .
the @xmath46 profiles from the three different data sets , match each other in the common ranges . on the contrary ,
[ fig : t2d - ei](a ) shows a systematic difference between the three temperature profiles . the discrepancy between and _ beppo - sax _
is probably related to the use of different energy bands . for the observation of the virgo cluster which has a temperature of @xmath54 kev ,
the energy range ( @xmath55 kev ) is more suitable than the _ beppo - sax _ energy range ( @xmath56 kev ) and , consequently , estimations are probably more reliable . for what concerns the discrepancy between and _ chandra _ temperature profiles , the differences are of the order of a few percent and well within the cross - calibration uncertainties between _ chandra _ and . to make full use of the combined
_ chandra _ , and _ beppo - sax _ data and at the same time avoid unphysical jumps in deprojected quantities when moving from one dataset to the next , we decided to shift , through a scale renormalization , the _ chandra _ and the _ beppo - sax _ datasets . was chosen as reference dataset because of its higher statistics .
the scale factors have been derived by imposing that the three temperature profiles match in the common ranges .
we find a renormalization factor @xmath57 for the _ chandra _ temperature profile , and @xmath58 for the _ beppo - sax _ temperature profile .
the temperature profiles , corrected for the scale factor , are plotted in fig .
[ fig : tew3r ] .
the final joint data set used for the analysis is given by ( i ) the _ chandra _ data in the @xmath59 range ( 4 points ) ; ( ii ) the data in the @xmath60 range ( 8 points ) ; ( iii ) the _ beppo - sax _ data in the @xmath61 range ( 3 points ) .
the deprojection technique has become very popular to investigate the intracluster medium properties @xcite . under the assumption of spherical symmetry ,
the different 3-d quantities describing the icm are derived from the 2-d projected ones , starting from the outer shell and moving inwards following an onion - peeling technique . among the different prescriptions available for the deprojection
, we decided to adopt the _ spectral deprojection
_ introduced by @xcite ( see also * ? ? ?
the physical quantities to be deprojected are those obtained from the spectral analysis described in the previous section .
each 3-d variable @xmath62 is related to the projected one @xmath63 according to the relation @xmath64 where @xmath65 is the inverse of the transposed matrix @xmath66 whose elements @xmath67 are the volumes of the @xmath68-@xmath69 shell projected on the @xmath70-@xmath69 ring .
the detailed evaluation of this matrix can be found in @xcite . by replacing @xmath63 with ( i ) the emission - weighted measured temperature @xmath71 ; ( ii ) the ring luminosity @xmath72 and ( iii ) @xmath73 we can derive respectively @xmath74 , the emissivity @xmath75 and the electron density @xmath12 , where the relation @xmath76 has been used .
the main advantage in using this technique is that deriving the electron density @xmath12 from @xmath42 is straightforward without any assumption on its functional shape .
moreover , since our measurements of the @xmath42 are very accurate , an immediate and precise determination of the electron density profile can be obtained .
it is worth noting that eq .
( [ eq : depro ] ) is derived using the onion peeling procedure where the contribution to the emission in each shell is obtained from the projected quantity by subtracting off the emission contribution of the outer shells starting from the edge of the cluster and moving inwards .
in addition to this basic prescription , a correction factor accounting for the cluster emission beyond the maximum radius @xmath77 to infinity must be included .
the procedure to evaluate this correction factor is presented in appendix [ appen ] . in practice , in eq .
( [ eq : depro ] ) the 2-d variable to be deprojected ( @xmath63 ) is replaced by an effective one @xmath78 ( see eq .
[ eq : feff ] ) . from now on , in order to avoid confusion , we will use @xmath10 for the 3-d deprojected temperature and @xmath79 for the 2-d emission - weighted temperature .
following the prescription described in the previous section , we derive @xmath75 , @xmath12 and @xmath10 profiles for m87 .
all the extracted values are reported in table [ tab : tneps ] . in fig .
[ fig : nel_tempfit ] we plot ( filled circles ) the deprojected electron density @xmath12 and temperature @xmath10 , respectively in panel ( a ) and ( b ) . for comparison , in fig .
[ fig : nel_tempfit ] we overplot ( open triangles ) the deprojected profiles obtained by @xcite who used a different choice of regions and a somewhat different technique for spectral deprojection .
the @xcite results plotted here are those obtained by fitting the mos data with a 2 t model .
our profiles are in reasonable agreement with those derived by @xcite .
particular attention must be paid in the computation of the error bars for @xmath10 and @xmath12 . in evaluating errors
, we want to consider that , even if the deprojection technique assumes that the spherical symmetry condition is fulfilled , the dispersion of the @xmath42 and @xmath79 measurements along each ring around the averaged value is often significantly larger than the error of each measure . in particular , this occurs for the data , since the large effective area allows a very good statistics for m87 and provides @xmath42 measurements with very small error bars .
the scatter of the data around the averaged value of the ring is a measure of the data displacement from the spherical symmetry . in order to account for this displacement in the final error evaluation
, we decided to assign to the @xmath42 and @xmath79 of each region , an error ( @xmath80 and @xmath81 respectively ) which is the linear sum of two different contributions .
the first contribution is simply the error derived from the spectral fit with the 1t/2 t model .
the second contribution is given by the dispersion of the measurements along the ring around the averaged value .
error bars for the deprojected quantities @xmath12 and @xmath10 reported in fig . [
fig : nel_tempfit ] have been obtained running 1000 monte - carlo simulations , sampling the @xmath42 and @xmath79 of each region around their mean value assuming gaussian distributions for the errors @xmath80 and @xmath81 .
the profiles plotted in fig .
[ fig : nel_tempfit ] will be used as starting points to derive some other quantities ( such as mass and conductivity ) . in most cases , gradients of @xmath12 and @xmath10
are involved .
some smoothing procedure will be required to manage these derivatives .
consequently , we apply a smoothing algorithm which replaces each point with the average value obtained using boxes of 3 points , i.e. : @xmath82 . in smoothing the temperature profile
, we excluded from the smoothing procedure the two last ( _ beppo - sax _ ) points , in order to preserve the final decreasing behavior of the @xmath10 profile .
the final temperature profile is obtained by applying the smoothing procedure twice : firstly , we smooth the starting deprojected profile and then the obtained values are smoothed again . for the density profile , we applied the smoothing procedure separately to the first 6 points and the others in order to preserve the 2-@xmath83 behavior .
again the two last points have been excluded from the smoothing operation and the smoothing has been applied twice .
the open diamonds in fig . [
fig : nel_tempfit ] represent the @xmath12 and @xmath10 profile after the smoothing operation . an alternative solution to smoothing
is provided by the use of analytical functions fitting the profiles .
the solid lines in the figures are the best fits to the data . for the electron density profile we use the fitting function : @xmath84^{\alpha_1 } } } + { { n_2 } \over { \left[1+\left({r \over { r_2}}\right)^2\right]^{\alpha_2 } } } , \label{eq:2beta}\ ] ] which corresponds to a 2-@xmath83 model . for the temperature profile we find that the function @xmath85 provides a good description of the data .
for the temperature profile , we find the following best - fit values : @xmath86 kev , @xmath87 kev , @xmath88 . for what concerns the electron density , we find @xmath89 , @xmath90 , @xmath91 @xmath50 , @xmath92 , @xmath93 and @xmath94 @xmath50 .
the inferred best fit values have large statistical errors .
this is due to the large number of free parameters adopted .
so , we decided to fix two parameters , namely , the core radius and the slope of the second component ; we fix the slope at large radii @xmath95 to the value obtained from the rass measurements @xcite setting @xmath96 ( @xmath97 ) .
for the radius @xmath98 we decided to use the value of the @xmath99 inferred from the best fit of the temperature profile , which defines the scale radius for the rise of the temperature . having fixed the values of @xmath98 and @xmath95 , we find : @xmath100 , @xmath101 @xmath50 , @xmath102 and @xmath103 @xmath50 .
note that the inferred value of @xmath104 ( @xmath105 kpc ) roughly corresponds to the agn jet extension ( e.g. * ? ? ?
* ; * ? ? ?
the deprojection method is based on the assumption of spherical symmetry . as discussed in [ sec : sp2d ] , the hot component in those regions which are described by the 2 temperature ( 2 t ) model behaves very similarly to the gas in the regions described by a single temperature ( 1 t ) model .
however , we should like to verify if any correlation of the hot component with the radio emission exists , invalidating our assumption of spherical symmetry .
to this aim , we divided the cluster in sectors and deprojected separately each sector .
again we consider only the hot component for the 2 t regions .
the sectors have been chosen according to the radio emission regions : the @xmath106 $ ] , @xmath107 $ ] and @xmath108 $ ] sectors have been cumulated together to form the non - radio sector .
furthermore , we analyzed separately each of the following sectors : @xmath109 $ ] , @xmath110 $ ] , @xmath111 $ ] , @xmath112 $ ] which correspond or are close to the radio emission arms .
the _ chandra _ data are not suitable to perform such a study , because the statistics is not very high . in this case
, possible azimuthal variations can be hidden by errors .
therefore , in order to verify the assumption of spherical symmetry , we consider only the data ( circles in fig .
[ fig : t2d - ei ] ) . nevertheless , also with data , it is quite difficult to use small sectors to perform sector - by - sector comparisons since their statistics is not very high .
a significant comparison can be made between the whole cluster and the non - radio selected sector . in fig .
[ fig : ne_tslice ] ( in panel ( a ) and ( b ) respectively ) we compare the @xmath12 and @xmath10 profiles for the whole cluster ( filled dots ) with the non - radio sector ( open diamonds ) . in general , no significant differences are evident .
the profiles of the electron density are almost identical whereas there is a slight tendency of the temperature calculated on the whole cluster to be smaller than the temperature of the non - radio sector .
the difference is due to the fact that in the non - radio sector only 1 t regions contribute , while in the whole cluster profile also the contribution of the hot component temperature of the regions described by a 2 t model is accounted .
this temperature is slightly smaller than the overall temperature and produces a mild decrease of the whole cluster temperature profile with respect to the non - radio sector temperature profiles .
however , differences are well within 1@xmath113 ; thus , excluding the radio emission sectors ( or the regions described by the 2 t model ) does not affect significantly the results and no evident azimuthal asymmetry can be highlighted .
we can conclude that the spherical symmetry assumption , which is an important condition for our analysis , is tenable .
it is also worth noting that , in order to derive most of the other physical quantities ( mass , conduction , entropy , etc . )
a smoothing operation on @xmath10 and @xmath12 is necessary , so that the small differences reported above would not anyhow affect our results .
once the basic quantities @xmath12 and @xmath10 are obtained through the deprojection , under suitable assumptions , other related quantities describing the icm can be derived .
one of the most important is the gravitational mass .
supposing that the gas is in hydrostatic equilibrium within the potential well of the dark matter , the gravitational mass @xmath114 within a radius @xmath115 can be derived via the hydrostatic equilibrium equation : @xmath116 \label{eq : hee}\ ] ] where @xmath117 is the mean molecular weight , @xmath118 the gravitational constant and @xmath119 the proton mass . even if eq .
( [ eq : hee ] ) provides a direct method to derive the gravitational mass , it is a differential equation and the temperature and the electron density are involved through their gradients .
irregular features in the profile induce jumps on the evaluated mass . a classic solution
consists in smoothing the data and replacing the @xmath12 and the @xmath10 profiles with their smoothed counterparts plotted in fig .
[ fig : nel_tempfit ] ( open diamonds ) .
as far as the temperature is concerned , a smoothing procedure is viable since errors are rather large and the general shape of the profile is quite smooth .
correspondingly , the smoothed profile is compatible ( always within 1@xmath113 ) with the original one . on the contrary , for the electron density where errors are small
, the smoothing procedure could hide some features which are physical . in order to assess whether the smoothing affects results , we compare the gravitational mass obtained using the temperature and the density smoothed profiles with the gravitational mass derived using the _ unsmoothed _ profiles where
errors have been evaluated using the standard monte carlo technique . as we show in fig .
[ fig : mass ] , the smoothed profile ( filled circles ) agrees with the non - smoothed one ( open diamonds ) within @xmath120 and no significant difference can be highlighted .
it is worth noting that the mass derived without smoothing provides three mass values ( at @xmath121 and @xmath122 kpc ) which are negative and compatible with zero : @xmath123 , @xmath124 and @xmath125 . for these points , in fig .
[ fig : mass ] , we show only the upper limit of the error bar . alternatively to the smoothing procedure , the analytical expressions for @xmath12 and @xmath10 ( eqs .
[ eq:2beta ] and [ eq : tfit ] ) can be used .
the curve in fig .
[ fig : mass ] is the analytical mass obtained using these two best fit profiles .
this mass profile has a plateaux at a radius of about @xmath126 kpc ( @xmath127 arcmin ) .
this flattening behavior is the consequence of the flattening of the 2-@xmath83 profile , at the same radius , which could correspond to the edge of the central cd .
the analytical gravitational mass is very similar to the mass obtained both with the smoothing procedure and with the unsmoothed data .
the differences with respect to the latter curve are limited to a few points ( @xmath128 and @xmath129 kpc ) .
for these points , some `` holes '' appear in the shape of the non - analytical profiles .
it is worth noting that the `` hole '' in a ( integrated ) mass profile is not physical .
it may indicate that in this region the hydrostatic equilibrium hypothesis breaks down ( e.g. * ? ? ? * ) and that an outflow providing additional pressure to support gravity occurs there ( see next section for equations and further details ) . in any case , the error bars of all these points are large enough to make `` holes '' compatible with the analytical profile and no strong evidence is present to claim that an outflow is present . for comparison , the grey - shaded regions in fig .
[ fig : mass ] report the gravitational mass derived by @xcite using rosat pspc data .
our profile , in the common radial regions is in agreement .
it is interesting to note that the point at @xmath130 kpc has a very large error bar in both the estimations .
as previously outlined , finding a heating mechanism able to balance the cooling is not an easy task .
this section will be devoted to the inspection of some heating sources using the m87 data set .
the heating contribution required to balance the radiative cooling can be estimated starting from the thermodynamic equations describing a spherically symmetric cluster : @xmath131 = -\varepsilon + \varepsilon_{cond } + \cal{h}\end{aligned}\ ] ] where @xmath132 is the gravitational potential , @xmath114 the gravitational mass within @xmath115 , @xmath10 the gas temperature , @xmath75 the emissivity and @xmath133 the gas mass density which is related to the electron density according to @xmath134 ( @xmath135 is the mean molecular weight ) .
the equations take into account the possible presence of an inflow or outflow and the flow velocity @xmath136 is taken positive outwards .
the three equations ( [ eq : thermo ] ) are respectively the mass , momentum and energy conservation equations .
they have been derived ( see * ? ? ?
* ) assuming a steady state ; the second equation reduces to hydrostatic equilibrium , eq .
( [ eq : hee ] ) , for @xmath137 .
we include in the _ right - hand - side _ of the last equation , the radiative cooling @xmath75 , and a heat contribution which includes two parts : the thermal conduction @xmath138 which will be widely discussed in the next paragraph , and a generic extra - heating term @xmath139 , which will be studied in detail in section [ sub : rbmod ] .
one obvious heating source is the thermal conduction which operates when temperature gradients occur , and which can have a relatively large efficiency ( a fraction of the spitzer conduction ) even in presence of magnetic fields @xcite . neglecting any extra - heating source ( @xmath140 @xmath141 in eq .
[ eq : thermo ] ) , and under the assumption of a spherical , steady state , isobaric cooling flow , the last equation of ( [ eq : thermo ] ) can be rewritten in the form : @xmath142 where @xmath75 is the emissivity , @xmath117 is the mean molecular weight and @xmath143 is the usual mass deposition rate of the cooling flow .
the heating due to thermal conduction @xmath138 is given by : @xmath144 where @xmath145 is the conductivity . for a highly ionized plasma , @xmath145
is given by the @xcite formula : @xmath146 where @xmath147 is the usual coulomb logarithm . starting from these equations we can derive the conductivity @xmath145 required in m87 to stop the cooling flow ( @xmath148 ) .
the inferred values for @xmath145 are plotted in fig .
[ fig : cond ] ( filled circles ) .
we plot @xmath145 as a function of the temperature in order to compare our results with @xcite and @xcite who have performed a similar calculation on a set of clusters .
we recall that the temperature grows with the radius .
hence , the behavior of @xmath145 as a function of the temperature is similar to the behavior of the profile of @xmath145 as a function of the radius . from fig .
[ fig : cond ] we can see that the required conductivity has large values for small temperatures ( i.e. in the central part of the cluster ) and becomes smaller when the temperature increases , i.e. moving towards the outskirts of the cluster . note that the temperature profile in the innermost regions of the cluster is consistent with being constant .
correspondingly , no conduction should be present and the required conductivity is consistent with being as high as infinity .
hence , for these points , error bars will extent to infinity . in order to show in fig .
[ fig : cond ] that the error bar for these points should extent to infinity , we plot their error bars with an arrow . the solid line in fig .
[ fig : cond ] represents @xmath149 given in eq .
( [ eq : spitzer ] ) and dashed line corresponds to 0.3@xmath149 which could be the effective conductivity in presence of turbulent magnetic fields @xcite .
[ fig : cond ] shows that in m87 , the thermal conduction is able to balance radiative cooling only in the outer part of the cluster . for @xmath150 kpc the conductivity required for conduction to balance the cooling is between 0.3@xmath149 and @xmath149 . in the inner @xmath151 kpc in m87
the heating supplied by thermal conduction is not enough and an extra - heating , whatever its source might be , is needed .
the failure of the thermal conduction in the core of the cluster is due to the fact that in these regions the temperature profile flattens ( see fig .
[ fig : nel_tempfit]b ) and the conduction decreases substantially . at the same time ,
the innermost regions are those where the x - ray emissivity is highest and which mostly require a heating source to compensate the radiated energy . in a recent paper ,
@xcite determined the conductivity required for the conduction to balance the radiation losses for a2199 whose temperature is similar to m87 temperature ranging from @xmath152 to @xmath153 kev . for comparison ,
we plot in fig .
[ fig : cond ] the @xcite data for a2199 derived by modeling the temperature profile with two different prescriptions : a power law ( triangles ) and a more complex functional form ( squares ) which flattens at small and large radii ( see eq .
( 6 ) in * ? ? ?
because of the larger distance of a2199 ( @xmath154 ) only a couple of points are within the central 10 kpc .
nevertheless , in agreement with our results , they find that for these two central bins some extra - heating is needed .
the quality of the m87 data set allows to highlight the problem of conduction in the core and to analyze it in greater detail than for a2199 .
clearly , m87 is a good object to test heating models . in fig .
[ fig : cond ] we also plot ( open circles ) the @xmath145 values obtained for m87 by @xcite .
their values are in reasonable agreement with ours , although their analysis procedure is quite different from ours . in @xcite
only _ chandra _ data are included limiting the extension of the deprojected region to the inner 10 kpc and only the 1 t xspec - mekal model is used in fitting spectra extracted from annuli . in this section
we consider eqs .
( [ eq : thermo ] ) in their generic form in order to determine the extra - heating @xmath139 required to balance the radiative cooling in presence of thermal conduction for m87 . eqs .
( [ eq : thermo ] ) can be recasted in the form : @xmath155 \\
\nonumber & & { \cal{h } } = \varepsilon - \varepsilon_{cond}+\varepsilon^\star\end{aligned}\ ] ] where we set : @xmath156 \ , .
\label{eq : epsstar}\ ] ] this term includes the variation of the energy ( per unit volume ) due to the outflow / inflow and the work ( per unit volume ) done by the system during the outflow / inflow . the mass flow rate @xmath143 is positive for an outflow and negative for an inflow .
the last equation of ( [ eq : heating ] ) provides the heating @xmath139 necessary to balance the radiative cooling @xmath75 , in presence of thermal conduction and steady outflow .
the deprojected data @xmath10 , @xmath12 ( or @xmath133 ) , of m87 can be used to solve numerically eqs .
( [ eq : heating ] ) for m87 , deriving @xmath114 , @xmath136 and @xmath139 , once we have fixed the fraction @xmath157 of the spitzer conductivity ( eq . [ eq : spitzer ] ) and some assumption has been made on @xmath143 .
we fix an @xmath158 efficiency ( see * ? ? ?
* ) and we set the outflow mass rate @xmath159 .
this @xmath143 value is similar to the asymptotic value that rb02 obtained from their simulations for the stable final state of the cluster .
we will discuss further on different values for @xmath143 and @xmath157 .
as far as the velocity is concerned , for the assumed values of @xmath157 and @xmath143 , @xmath136 is smaller than a few tens of km / s , for radii larger than @xmath2 kpc .
correspondingly , the term including the velocity in the momentum conservation equation ( the second equation of [ eq : heating ] ) is significantly smaller than the total mass being of the order of @xmath160 versus the @xmath161 of the total mass , so its contribution is negligible .
hence , we can state that the cluster is almost in hydrostatic equilibrium and the mass estimates reported in fig .
[ fig : mass ] , where the term related to the outflow is neglected , are not affected . in order to have a significant contribution from the velocity term and to alter substantially the hydrostatic equilibrium , @xmath143 values as large as several tens
few hundreds of @xmath162 are required .
for the considered values of @xmath143 , the quantity @xmath163 in eq .
( [ eq : epsstar ] ) is negligible with respect to the emissivity @xmath75 .
thus , the extra - heating @xmath139 and the conduction term @xmath138 are completely used to balance the radiative cooling . for what
concerns @xmath139 , the heating required in m87 is plotted in fig .
[ fig : heating ] ( filled circles ) .
as expected , most of the heating is required in the central part of the cluster ( say in the inner @xmath164 kpc ) , where conduction is not efficient .
the heating due to thermal conduction is plotted in fig .
[ fig : heating ] ( dot - dashed line ) ; in the central @xmath165 kpc , where the temperature profile becomes flatter , the thermal conduction drops to zero , apart from the innermost bin ( @xmath166 kpc ) where @xmath10 falls to very small values ( see fig . [
fig : nel_tempfit]b ) , with a large error bar .
the conduction in this bin is @xmath167 , and is in agreement with zero within 1@xmath113 . the heating model developed in ( * ?
* rb02 hereafter ) and @xcite includes a mechanism for heat injection from the central agn .
the mechanism has been called `` effervescent heating '' .
the radio source is supposed to deposit some buoyant gas bubbles in the icm , which do not mix and rise through the icm microscopically .
the bubbles should expand doing work on the surrounding plasma and converting their internal energy in heat .
the buoyant outflow contribution in the energy conservation equation is accounted for in the @xmath163 term , while @xmath139 describes the heat injection due to the adiabatic expansion of the bubbles .
according to the rb02 model , the heating @xmath139 is a function of the pressure ( and its gradient ) and can be expressed according to : @xmath168 where , @xmath169 and @xmath170 @xmath171 is the pressure , @xmath172 is the central pressure , @xmath173 the time - averaged luminosity of the central source , @xmath174 is the adiabatic index of the buoyant gas and @xmath175 the inner heating cutoff radius .
the term @xmath176 introduces an inner cutoff which fixes the scale radius where the bubbles start rising buoyantly in the ambient plasma .
@xmath139 is normalized in such a way that , when integrated over the whole cluster , the total injected power corresponds to the time - averaged power output of the agn .
@xmath139 has been derived in eq .
( [ eq : hbr ] ) assuming a steady state for the bubble flux .
in order to assess if this assumption is reasonable we must compare the different timescales involved in the effervescent heating mechanism .
we can suppose that the agn is intermittent ( rb02 ; @xcite ) and heats the icm through a succession of outbursts . during each outburst ,
the agn injects a population of bubbles which subsequently rise buoyantly . during the `` off '' periods ,
the bubbles continue their rise heating the cluster atmosphere .
if the outbursts follow each other on a timescale which is short with respect to the rising timescale of the bubbles , then the flux of the bubbles reaches a quasi steady state .
in fact , the ratio @xmath177 between the rise timescale @xmath178 and the intermittence timescale @xmath179 gives the number of populations injected within the icm within a time @xmath178 .
the larger this ratio is , the larger the number of bubble populations rising into the cluster atmosphere and the mechanism approaches the steady state .
the radio galaxies are likely to be intermittent on a timescale as short as @xmath180 yr ( rb02 ; @xcite ) . in the next section
we will see that the risetime is @xmath181 yr or even larger .
the value of the ratio @xmath182 is @xmath183 or more ; therefore the assumption of steady state is tenable and the released heating may be treated in a time - averaged sense .
rb02 also include a convection term in eq .
( [ eq : heating ] ) which we have neglected .
the reason for this choice is twofold .
first of all , the convection must be limited to the innermost regions of the cluster , in order to allow the presence of metallicity gradients in cluster cores @xcite .
most importantly , a negative gradient for the entropy is a necessary condition for the onset of the convection .
in fact , the condition of instability : @xmath184 ( where @xmath185 is the ratio of the specific heats @xmath186 and has the value @xmath187 for a highly ionized gas ) must be fulfilled for the convection to operate and it is equivalent to requiring that @xmath188 .
this condition is not satisfied in m87 where we verified that the entropy is an increasing function of the radius , as we show in fig .
[ fig : entropy ] .
we compare the values inferred for the extra - heating @xmath139 term in m87 reported in fig .
[ fig : heating ] ( filled circles ) , with the predictions from the rb02 model derived according to eq .
( [ eq : hbr ] ) , in order to assess whether , for a reasonable choice of the parameters , the heat flux required to balance the cooling is compatible with the heat injected by the central agn . in order to determine the pressure and the pressure gradient in eq .
( [ eq : hbr ] ) , we use the analytical expressions ( [ eq:2beta ] ) and ( [ eq : tfit ] ) for @xmath12 and @xmath10 with the best - fit parameters obtained by fitting the deprojected electron density and temperature profiles .
we use eqs .
( [ eq : hbr ] ) - ( [ eq : qint ] ) as fitting functions for the extra - heating , where @xmath174 , @xmath175 and the total normalization @xmath189 : @xmath190 are the free parameters . the solid line in fig .
[ fig : heating ] is the derived best fit .
the model proposed by rb02 seems to provide a fair description of the heating needed to balance the cooling flow in the inner @xmath164 kpc of the cluster , where conduction is not sufficient .
the line seems to follow adequately the behavior of the data points
. nevertheless , the derived fit values have large errors . by fixing @xmath191 ( which is the adiabatic index for relativistic bubbles ) ,
we find @xmath192 kpc and @xmath193 ( the dashed curve in fig .
[ fig : heating ] ) . by using eq .
( [ eq : normfit ] ) , we can also derive the central agn luminosity @xmath194 @xmath195 required to stop the cooling .
the three external points in fig .
[ fig : heating ] seem to require an additional extra heating , showing an excess with respect to the general behavior of the data at large radii and with respect to the best fit function shape . this excess is related to the flattening of the temperature profile in the external regions which dampens conduction .
however , it must be noted that the values of the heating @xmath139 required in these three bins are respectively : @xmath196 , @xmath197 and @xmath198 and are all in agreement with zero within @xmath199 .
hence , the evidence for the excess is not particularly strong .
our estimated radius @xmath200 kpc , is comparable to the extension of the agn jet as seen in the _ chandra _ image of m87 ( e.g. * ? ? ?
* ; * ? ? ?
the fact that @xmath175 is of the same order of magnitude of the jet is consistent with a scenario where the effervescent bubbles are generated through the interaction of the radio jet with the cluster atmosphere .
understanding the precise nature of such interaction will require considerable efforts both on the observational and theoretical side .
our aim here is simply to note that our fit does not rule out the possibility that the bubbles are generated through the interaction of the radio jet with the cluster atmosphere , as would have been the case if , for example , the fitting of the effervescent model to m87 had returned an @xmath175 value 10 times larger than the one actually measured .
the inferred luminosity value is similar to the luminosity evaluated for the m87 agn ( * ? ? ?
* oek hereafter ) , @xmath201 ( see also * ? ? ?
slightly different values for the luminosity could be inferred with different choices of @xmath157 and @xmath143 which of course provide different best fit parameters values and different related luminosities .
when a larger @xmath157 is considered , the higher contribution of the thermal conduction reduces the amount of heating needed to balance the radiative cooling . setting the conductivity to the spitzer value ( @xmath202 ) we find best fit values which are similar to those inferred for @xmath158 and
provide a slightly lower luminosity ( @xmath203 ) . on the contrary ,
when smaller values for @xmath157 are considered , some additional heating is necessary .
for @xmath204 we derive an agn luminosity @xmath205 , which is somewhat larger than the oek estimations .
however , for such small efficiencies , the shape of the rb02 model no longer provides a good description of the data , especially in the central regions .
thus , if the contribution of the thermal conduction is too small , the `` effervescent heating '' model is not suitable to describe the heating necessary to balance the cooling flow .
some variations are found also for different initial values of @xmath206 .
we tried 16 and 0.16 m@xmath207/yr corresponding to 10 and 0.1 times the original @xmath143 value we considered . as expected , for large values of @xmath143 the corresponding agn luminosity is significantly enhanced ( @xmath208 ) since the central agn must provide a larger quantity of energy to the outflowing bubbles .
the variations when smaller @xmath143 are considered , are modest , slightly reducing the luminosity to @xmath209 .
note that it is not necessary for the agn luminosity , obtained by requiring that the `` effervescent heating '' model balances the cooling flow to be exactly equal to the agn luminosity derived from radio observations .
in fact , the required luminosity from the model should be regarded as a time - averaged power of the agn , as the agn dynamical times are smaller than the radiative cooling flow scale times .
one should also keep in mind that only a fraction of the total power of the agn is used to quench the cooling flow and that the luminosity required from the model can be significantly smaller than the real agn luminosity . from our analysis , we can infer that the values of @xmath173 derived with different choices of @xmath157 and @xmath143 are of the same order of estimates by oek and @xcite .
while the luminosity is slightly affected for different choices of @xmath143 and @xmath157 , @xmath175 variations are quite modest and the inferred values of the scale radius are always of the order of @xmath210 kpc , which approximatively corresponds to the agn jet extension .
starting from the results inferred in the previous section , we can try to draw a more general picture , using also informations coming from radio observations of m87 .
we can suppose that the buoyant bubbles are radio bubbles filled with magnetic field and relativistic particles @xcite responsible for the synchrotron emission in m87 .
as already outlined by oek , the radio structures are highly filamented .
this suggests that the dimensions of the bubbles are small .
@xcite discussed the dynamics of the rise of buoyant light bubbles within the cluster atmosphere ( see also * ? ? ?
* ; * ? ? ?
the buoyant bubble rapidly reaches a terminal velocity @xmath211 . in the limit of small bubbles
, @xmath211 can be estimated by balancing the buoyancy force with the ram pressure ( drag force ) of the cluster gas .
the buoyancy force is @xmath212 where @xmath213 is the volume of the bubble , @xmath214 is the bubble radius , @xmath215 is the local gravity acceleration at the radius @xmath115 ( @xmath114 is the gravitational mass within the radius @xmath115 ) ; @xmath133 and @xmath216 are respectively the density of the icm and of the bubble . @xmath217
is the density contrast .
the drag force for subsonic motion can be approximated by @xmath218 with the drag coefficient @xmath219 . by equating eqs .
( [ eq : fbuoy ] ) and ( [ eq : drag ] ) the velocity of the bubble can be determined .
@xcite derived @xmath211 under the assumption that the density of the icm is well described by an isothermal @xmath83-model and the density contrast @xmath220 . in this case
@xmath211 is a fraction ( @xmath221 ; @xmath222 is the core radius of the cluster ) of the sound velocity .
correspondingly , for small bubbles , @xmath211 is subsonic which is consistent with the fact that no shocks are detected in m87 .
we derived the rise velocity of the bubbles in m87 at a radius @xmath223 kpc , using the results from the deprojection for @xmath224 and @xmath133 .
the density contrast @xmath225 can be inferred considering that the bubbles filled with relativistic plasma are in pressure equilibrium with the icm .
indeed , the pressure equilibrium between radio bubbles and thermal plasma implies that @xmath226 ( where @xmath83 is a factor of the order of the unity which accounts for the contribution to the pressure of the magnetic field which is almost in equipartition with the particle energy ) .
@xmath227 is the mean energy of the relativistic particles and @xmath228 is the numerical density of the relativistic particles .
the relativistic particles mean energy is much larger than that of the thermal electrons in the icm , leading to a density contrast @xmath220 .
the values of @xmath211 for the bubble radius @xmath214 varying in @xmath229 $ ] kpc range are plotted in fig .
[ fig : velb ] and the motion is indeed subsonic .
it is worth considering that , in a recent analysis of these data ( * ? ? ?
* m02 hereafter ) , we have found that , cospatially with the radio lobes , there exists a thermal component with @xmath230 .
this component is likely structured in blobs with typical scales smaller than @xmath53 pc .
the filling factor of these blobs has been estimated to be of the order of the percent .
some informations about the filling factor of the radio bubbles can be obtained using recent results from radio observation of m87 . using the standard minimum pressure analysis ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , oek evaluate the magnetic field in the lobes and in the filaments visible in the radio map and evaluate the minimum pressure of the magnetic field and relativistic particles .
they assume that the proton - to - electron energy is @xmath231 and that @xmath232 where @xmath132 is the filling factor of the relativistic particles and of the magnetic field and @xmath233 is the ratio of the true magnetic pressure ( comprehensive of the tension ) to the magnetic pressure @xmath234 . for a tangled magnetic field @xmath235 .
the estimations for the minimum pressure derived by oek can be compared with the thermal pressure that we can infer from our deprojected @xmath10 and @xmath12 profiles . in agreement with oek
, we find that the pressure of the relativistic plasma is smaller than the thermal pressure by about an order of magnitude . by keeping the condition @xmath231 and assuming @xmath236
, we can derive the filling factor which reconciles the minimum pressure @xmath237 with the thermal pressure of the plasma , considering that the minimum pressure scales with @xmath132 according to : @xmath238 the discrepancy of a factor @xmath239 between the two pressures can be eliminated by assuming a filling factor @xmath240 . by using the estimations of @xmath237 from oek in different plasma of the radio lobes and filaments ,
we infer filling factors of few @xmath241 . assuming @xmath235 ( tangled magnetic field ) reduces @xmath132 by a factor @xmath242 .
the filling factors of the radio bubbles and of the cold thermal blobs are of the same order .
the survival of the cold thermal blobs in the hotter icm requires that thermal conduction be substantially suppressed ; this may happen if these blobs are tied to the radio bubbles and magnetic fields shield them from collisions with icm particles .
since the filling factors of blobs and bubbles are similar and the density of the blobs is about twice that of the icm ( m02 ) , the mean density of each bubble+blob is about that of the surrounding icm .
thus , assuming that the blobs are tied to the bubbles and that they occupy similar volumes , their density contrast with the surrounding icm will be small . in fig .
[ fig : velb ] the dotted line plots the bubble rise velocity @xmath211 for a @xmath243 . starting from the rise velocity , it is straightforward to derive the rise timescale of the bubbles . for a density @xmath220 and a bubble radius of @xmath53 pc ,
the risetime is @xmath181 yr at @xmath244 kpc and it becomes even larger for smaller bubbles and smaller density contrasts , holding @xmath245 . as outlined previously , this value is significantly larger than the duty cycle timescale @xmath179 of intermittency of the agn , so that the ratio @xmath246 and the mechanism approaches the steady state .
it is worth noting that the above picture refers to radio bubbles and cool x - ray blobs located in the lobes .
however , the heating process related to the rise of the bubbles must be isotropic throughout the cluster in order to balance the cooling flow . under the assumption that the radio galaxy is intermittent on a short timescale
, the mechanism is expected to heat isotropically the icm since it is likely that no direction is preferred for the agn ejection ; we also recall that when the agn is turned off the bubbles continue their rise within the icm heating it up ; within this picture , by averaging on a cooling time , the radio bubble populations are likely to be isotropically distributed in the cluster . nevertheless , the cool x
ray blobs are detected only in the lobes regions .
so the picture emerging here is that of an agn which injects radio bubbles , in all the directions , through a succession of outbursts . in the lobes ,
the outburst is occurring at the present time and , here , also the cool x - ray blobs are present .
they are tied to the radio bubbles and the magnetic fields shield them from collisions with icm particles .
the thermal conduction is inhibited allowing the cool blobs to survive in a hotter ambient medium . in the halo , where the radio bubbles have been injected during a past activity of the agn
, only `` old '' populations of bubbles which are buoyantly rising are present .
it is likely that the cool blobs which were tied to the radio bubbles when the agn was active there , have thermalized , the magnetic fields slow down but do not stop entirely thermalization so that only a single phase gas is detected in the halo regions .
the crisis of the standard cooling flow model brought about by _
chandra _ and observations of galaxy clusters , has led to the development of several heating models with the aim of identifying a mechanism able to quench the cooling flow .
we have used observations of virgo / m87 to inspect the dynamics of the gas in the center of the cluster , and to study the heating processes able to balance radiative losses .
we combined the observations of three satellites , namely , _
chandra _ and _ beppo - sax_. by means of the spectral deprojection technique , we inferred the profiles for the temperature @xmath10 and the electron density @xmath12 for m87 .
the temperature profile drops by a factor of @xmath127 in the inner ( @xmath247 kpc ) .
the electron density @xmath12 is well described by a 2-@xmath83 profile .
starting from these profiles , we derived some related physical quantities such as the gravitational mass and the entropy profile .
the gravitational mass profile shows a plateaux at @xmath248 kpc , which could correspond to the edge of the central cd . in agreement with @xcite and @xcite
results on a set of cluster , we found that the thermal conduction in m87 can balance the cooling flow only in the outer part of the core ( say , @xmath249 kpc ) , while in the inner @xmath126 kpc , where the temperature profile becomes flatter and the conduction is no longer efficient , some extra - heating is required .
we have determined the extra - heating needed to balance the cooling flow in m87 , assuming a thermal conduction efficiency @xmath158 with respect to the spitzer value , and an outflow mass rate @xmath250 .
the high quality of our combined dataset allows us to inspect properly the innermost regions , deriving an accurate profile for the extra - heating term @xmath251 in the regions where thermal conduction is not sufficient to quench the cooling flow .
several models and simulations concerning heating mechanisms through buoyant gas in the cluster icm have been recently proposed in the literature ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we have assumed that the heating is provided by the central agn by means of deposition of buoyant bubbles in the icm according to the model proposed by @xcite .
the bubbles rise through the icm and expand doing work on the surrounding plasma and heating it up .
we fitted the extra - heating required in m87 , with the heating functions proposed by rb02 ( see eqs .
[ eq : hbr ] - [ eq : qint ] ) .
the rb02 model seems suitable to describe the behavior of the data . by fixing the adiabatic index @xmath252 ( relativistic bubbles ) , we find a scale radius @xmath192 kpc and @xmath253 , which corresponds to a central agn luminosity @xmath254 .
the scale radius is of the order of the extension of the agn jet and the inferred agn luminosity is similar to the one estimated by @xcite and @xcite @xmath255 .
smaller conductivity efficiencies ( @xmath204 ) or larger outflow mass rates ( @xmath256 ) provide agn luminosities ( @xmath257 ) which are somewhat higher than the estimations of @xcite and @xcite . however , if the efficiency of the thermal conduction is reduced ( @xmath204 ) the model functions of rb02 seem no longer suitable to describe the heating needed to balance the cooling flow . for higher conduction efficiencies ( @xmath258 ) or smaller outflow rates values ( @xmath259 ) the inferred luminosity
is slightly reduced ( @xmath260 ) .
the different values derived for @xmath173 are of the same order of the luminosity measures for the agn luminosity in m87 obtained by @xcite and @xcite . in all the cases considered ,
the inferred scale radius @xmath175 , which fixes the radius at which bubbles are deposited and start rising , is @xmath210 kpc which approximatively corresponds to the agn jet extension . finally , we discussed a scenario where the bubbles are filled with relativistic particles and magnetic field , responsible for the radio emission in m87 . in this scenario
the density contrast @xmath225 is expected to be as large as 1 .
the buoyant velocity @xmath211 of the bubbles can be derived by balancing the buoyant force with the drag force .
for small bubbles , the rise velocity is subsonic . under the hypothesis of equipartition between relativistic particles and magnetic field in the bubbles and of equilibrium pressure with the thermal icm
, we evaluated the filling factor @xmath132 of the radio bubbles .
we find @xmath261 which is of the same order of the filling factor of the cool thermal component observed in the regions of the radio lobes ( m02 ) .
hence , we suggest that this cool thermal component is structured in blobs tied to the radio bubbles . the thermal conduction , which should rapidly thermalized the cool blobs , is suppressed by the magnetic fields of the radio bubbles .
the density contrast of the buoyant bubble+blob system is @xmath262 further reducing the rising velocity .
the radio galaxies are likely to be intermittent on a timescale @xmath263 yr and they are supposed to heat the icm through a succession of outbursts . during each outburst a population of radio bubbles is injected into the icm .
the bubbles rise buoyantly in the intracluster gas , heating it up .
the outbursts follow each other on small timescale @xmath179 which is much shorter than the rise timescale @xmath264 yr of the bubbles so that the mechanism is isotropic throughout the cluster and approaches the steady state .
the x ray cool blobs are detected in the radio lobes where the injection of radio bubbles is occurring at the present time .
the blobs are tied to the radio bubbles so that the thermal conduction is highly suppressed by the magnetic fields . in the radio halo ,
the radio bubbles injection occurred in the past .
the radio bubbles are buoyantly rising in the cluster atmosphere and it is likely that the x ray cool blobs which were tied to the bubbles when the agn was active in those directions , have in the meanwhile thermalized so that only a single phase thermal component is present here . the authors wish to thank luigina feretti and mariachiara rossetti for useful discussions .
the authors are pleased to acknowledge the referee c. kaiser whose useful comments and suggestions have significantly improved the paper .
by using eq . ( [ eq : depro ] ) the deprojected variable @xmath62 is obtained from @xmath63 by subtracting off the contribution of the outer shells , starting from the outermost annulus and moving inwards .
this basic prescription assumes that there is no emission beyond the maximum radius @xmath77 .
thus , the basic recipe ( [ eq : depro ] ) must be corrected to account for a contribution to the x ray spectra from the gas beyond @xmath77 . where @xmath266 is the area of the ring , @xmath63 the variable to be deprojected and @xmath267 the correction factor of the ring .
@xmath268 and @xmath269 are the area and the variable to be deprojected of the outer ring . in order to determine @xmath267 , some assumptions on the shape at large radii for the different
_ deprojected _ quantities must be made .
the classic solution consists in assuming that all @xmath270 , @xmath75 and @xmath12 have an @xmath271 dependence at large radii , with @xmath272 . for this standard assumption
, the correction factor @xmath273 can be determined analytically @xcite . however , the standard assumption @xmath272 is quite simplistic and holds only for an isothermal cluster with @xmath278 at large radii ( e.g. in the usual @xmath83-model with @xmath279 ) . on the contrary ,
we prefer considering the dependence @xmath280 with a generic @xmath281 , different for each deprojected quantity .
while for @xmath272 the correction factor could be calculated analytically , for a generic @xmath281 it must be determined numerically .
we evaluated the @xmath273 factor of eq .
( [ eq : gj ] ) by truncating the integrals at a @xmath282 external radius , with steps @xmath283 wide .
we verified that the numerical method for the simple case @xmath284 provides results which differ by no more than a few percent from the analytical values , and in any case the effect is always limited to the outermost bins . in order to find out the correct @xmath281 for each quantity to be deprojected , we applied iteratively the deprojection .
we started from reasonable values of @xmath281 .
then we applied the deprojection and derive the new @xmath281 values from the slopes of the deprojected profiles at large radii .
we used these new @xmath281 values to determine the correction factor of eq .
( [ eq : gj ] ) and applied again the deprojection working out new @xmath281 values .
we stopped when all the new @xmath281 values differed from the starting values by less than 4@xmath241 .
bhringer , h. , belsole , e. , kennea , j. , matsushita , k. , molendi , s. , worrall , d. , mushotzky , r. f. , ehle , m. , guainazzi , m. , sakelliou , i. , stewart , g. , vestrand , w. t. , dos santos , s. , 2001 , , 365 , l181 .emission - weighted temperatures in kev and normalized emission integral ( nei ) per unit area ;
@xmath285 is given in units , i.e. @xmath286 , where @xmath48 is the angular distance of m87 in cm , @xmath49 the redshift and @xmath42 in @xmath50 .
the @xmath51 is in arcmin@xmath52.[tab : t2d - ei ] [ cols="<,^,^,^",options="header " , ] | the crisis of the standard cooling flow model brought about by _
chandra _ and _ xmm - newton _ observations of galaxy clusters , has led to the development of several models which explore different heating processes in order to assess if they can quench the cooling flow . among the most appealing mechanisms are thermal conduction and heating through buoyant gas deposited in the icm by agns .
we combine virgo / m87 observations of three satellites ( _ chandra _ , _ xmm - newton _ and _ beppo - sax _ ) to inspect the dynamics of the icm in the center of the cluster . using the spectral deprojection technique
, we derive the physical quantities describing the icm and determine the extra - heating needed to balance the cooling flow assuming that thermal conduction operates at a fixed fraction of the spitzer value .
we assume that the extra - heating is due to buoyant gas and we fit the data using the model developed by ruszkowski and begelman ( 2002 ) .
we derive a scale radius for the model of @xmath0 kpc , which is comparable with the m87 agn jet extension , and a required luminosity of the agn of a @xmath1 , which is comparable to the observed agn luminosity .
we discuss a scenario where the buoyant bubbles are filled of relativistic particles and magnetic field responsible for the radio emission in m87 .
the agn is supposed to be intermittent and to inject populations of buoyant bubbles through a succession of outbursts .
we also study the x ray cool component detected in the radio lobes and suggest that it is structured in blobs which are tied to the radio buoyant bubbles . |
in this paper we develop further the study of immersions of two - dimensional surfaces in multidimensional euclidean or minkowski spaces by means of @xmath0 models , which was carried out in a series of papers @xcite .
the key point is the formulation of the equations defining the immersion directly in the matrix form ( cf .
equation [ 3_2 ] ) where the immersion takes values in the lie algebra @xmath2 , identified by means of the negative of the killing form with the euclidean space @xmath1 .
this allows us to formulate explicitly the structural equations for the immersion ( the gauss
weingarten and the gauss codazzi
ricci equations ) directly in matrix terms .
in particular in proposition [ prop3 ] we establish an explicit form of the gauss - weingarten equations satisfied by the moving frame on a surface corresponding to the @xmath0 model .
this is done in a fashion independent of any specific parametrization .
then we use this result to establish various geometric characteristics of the studied immersions like curvatures and curvature vectors .
all these quantities are directly derived from the map describing the relevant @xmath0 model . for the simplest case @xmath5 the equation ( [ 3_2 ] )
defining the immersion takes the form @xmath6 where @xmath7 are the usual pauli matrices , and the differentials of coordinate functions of the immersion @xmath8 are given in terms of the affine coordinate @xmath9 of @xmath10 model by equation ( [ 4_3 ] ) . as follows from the equation ( [ 1_22 ] ) describing the @xmath10 model , a particular class of solutions of this model
is given by an arbitrary holomorphic function @xmath9 in this case the immersion is minimal ( i.e. represents a surface with zero mean curvature ) and @xmath9 expresses the gauss map of the surface by means of the stereographic projection .
this is directly related to the classical weierstrass enneper formulae for an immersion of a minimal surface in @xmath11 .
in fact , almost one and half century ago weierstrass and enneper showed @xcite that every minimal surface in @xmath11 can be represented locally in terms of two holomorphic functions @xmath12 and @xmath13 defined on a domain @xmath14 by the following expressions @xmath15 this implies that the complex tangent vector of the immersion is given by @xmath16 where @xmath17 denotes the ( complex ) derivative with respect to @xmath18 .
moreover , the metric of the minimal surface is conformal and is expressed in terms of local parameters @xmath18 and @xmath19 by the formula @xmath20 this implies , in particular , that the coordinate lines @xmath21 and @xmath22 describe geodesics on this surface .
the ideas of enneper and weierstrass have since been developed by many authors with a purpose of extending them to construct immersions of more general types of surfaces . for a classical presentation
we refer to a treatise by eisenhardt@xcite and for a modern approach to the subject see e.g. @xcite and references therein , in particular the recent books by f. helein @xcite and k. kenmotsu @xcite .
this topic has been further explored by , among others , b. konopelchenko @xcite .
he established a direct connection between generic surfaces and trajectories of an infinite - dimensional hamiltonian system .
namely , he considered a nonlinear dirac type system of equations for two complex - valued functions @xmath23 and @xmath24 of @xmath25 @xmath26 he showed @xcite that for any solutions @xmath27 of the system ( [ 0_4 ] ) and an arbitrary curve @xmath28 in the complex plane , the following integrals over the bilinear combinations of @xmath29 , @xmath30 @xmath31 determine the coordinates of the radius vector @xmath32 describing a constant mean curvature ( cmc ) surface immersed in @xmath33 ( the integrals are independent of the integration path , since the integrands are exact differentials ) .
to see how to reduce more general cases down to this case see e.g. ref @xcite . in accordance with @xcite
we will refer to equations ( [ 0_4 ] ) and ( [ 0_5 ] ) as the generalized weierstrass formulae .
it was shown in our previous work @xcite that the generalized weierstrass formulae for two - dimensional surfaces with non - vanishing mean curvature in multi - dimensional spaces are equivalent to @xmath0 sigma models .
this determination has opened a new way for constructing and studying two - dimensional surfaces .
the further advantage of use of the @xmath0 models in this context lies in the fact that they allow us to replace the methods based on dirac - type equations by the formalism connected with completely integrable systems , for example lax pairs , hamiltonian structures , or systems defining infinite number of conserved quantities .
an original procedure for constructing the general classical solutions admitting finite action of the euclidean two - dimensional @xmath0 model was devised by a.m. din et al @xcite .
these solutions are obtained by repeated applications of a certain transformation to the basic solution expressed in terms of holomorphic functions . as a result ,
one gets three classes of solutions : holomorphic , anti - holomorphic and the `` mixed '' ones . in this paper
we show that to each of these solutions we can associate a surface in @xmath34 . in the holomorphic ( or antiholomorphic ) case we are able to integrate completely the equations of the immersion .
it turns out that in the @xmath35 case the surface is a part of an euclidean sphere , cf .
section [ sec : cp1 ] .
however for arbitrary @xmath36 other situations are possible . in example
@xmath37 in section [ sec : cp2 ] we present a one - parameter family of surfaces for which the curvature is not constant but for some specific values of this parameter it reduces to a constant .
the second and third examples in section [ sec : cp2 ] are concerned with mixed solutions of the @xmath38 model . in one case
we obtain a surface in @xmath39 , which happens to be immersible in @xmath11 , but the immersion does not come from a @xmath35 model , since the curvature is not constant .
the other mixed solution leads to a generic surface in @xmath39 with nonconstant curvatures .
all these results raise interesting questions which require further investigations concerning general properties of immersions given by the @xmath0 models - either holomorphic or mixed . finally , let us note that the outlined approach to the study of surfaces in @xmath40 lends itself to numerous potential applications .
it is useful for description of monodromy of solutions of higher order painlev equations and their connection with theory of surfaces .
it can also lead to the development of numerical computing tools in the study of surfaces through the techniques of completely integrable systems .
surfaces immersed in lie groups , lie algebras and homogeneous spaces appear in many areas of physics , chemistry and biology@xcite .
the algebraic approach to structural equations of these surfaces has often proved to be very difficult from the computational point of view .
a natural geometric approach to derivation and classification of such equations which we propose here seems therefore to be of importance for applications in physics and other sciences .
this paper is organized as follows . in section [ sec : cpn ] we introduce basic material on @xmath0 models in presenting it we focus on the use of a compatibility condition , rather then on the usual approach via the euler
lagrange equations .
the next section is devoted to the presentation of required notions and facts on the structure of complex projective spaces . here
we prove in detail a certain decomposition of the group @xmath41 , which was previously noted in the paper @xcite of rowe et al . in section [ sec : sun ] we show how to use the equations of the @xmath0 model to construct an immersion of a surface in the lie algebra @xmath2 .
the obtained formula extends the classical weierstrass formula for the conformal immersion of a @xmath42-dimensional surface into the @xmath43-dimensional euclidean space .
we derive the equations of the moving frame for the above immersion in @xmath2 and derive some geometrical characteristics of these surfaces .
this analysis is developed further in sections [ sec : cp1 ] and [ sec : su3 ] , where , using the conformality of the surfaces obtained from the @xmath0 models for @xmath44 , we establish an explict formula for the moving frame in terms of the data of the model .
let us note that the case @xmath45 produces surfaces immersed in @xmath46 . in section [ sec
: cp2 ] we illustrate our theoretical considerations with some examples based on explicit solutions of the @xmath38 sigma model .
the last section contains remarks and sugestions regarding possible further developments .
from a large supply of geometrical models of immersions @xcite we concentrate in this paper on a particular class of models , the so called @xmath0 sigma models .
the @xmath0 sigma model can be defined in terms of functions @xmath47 defined on an simply connected domain ( i.e. an open connected subset ) @xmath48 of the complex plane and satisfying the constraint @xmath49 .
here and below we employ the standard notation where points of the complex coordinate space @xmath50 are denoted by @xmath51 and the standard hermitian inner product in @xmath50 , respectively the norm , by @xmath52 the unit sphere in @xmath50 corresponding to this norm is given by @xmath53 we shall use @xmath54 to denote ordinary derivatives and @xmath55 for the covariant derivatives defined according to the formula @xmath56 with this notation the lagrangian density for such a model is given by ( cf .
e.g. @xcite ) @xmath57 and the solutions of the @xmath0 model are stationary points of the action functional @xmath58 the physically relevant case concerns fields which can be extended to the whole riemann sphere @xmath59 , but the case of an arbitrary @xmath48 is also of some interest , especially for questions of ( local ) differential geometry .
next we note that since @xmath60 is not changed by the transformations @xmath61 with @xmath62 , it is actually defined by the map @xmath63 : \omega \rightarrow { { { { { \mathbb c}}}}p^{n}}$ ] , where @xmath63= \{e^{i\varphi}z\mid \varphi \in { { { \mathbb r}}}\}$ ] is the element of the projective space @xmath64 corresponding to @xmath65 .
we find it often more convenient to use this latter point of view and describe the model in terms of unnormalized " fields @xmath66 related to the `` @xmath67 s '' above by @xmath68 we shall refer to the `` @xmath67 s '' above as inhomogeneous and to the `` @xmath69 s '' as homogeneous coordinates of the model .
now , using the customary notation for holomorphic and antiholomorphic derivatives @xmath70 and introducing the orthogonal projector on the orthogonal complement to the complex line in @xmath71 determined by @xmath69 given by @xmath72 we may express the action functional ( [ 1_4 ] ) in terms of @xmath69 s by @xmath73 since @xmath74 is an orthogonal projector , it satisfies @xmath75 the map @xmath63 $ ] is determined by a solution of the euler
lagrange equations which is associated with the action ( [ 1_7 ] ) . in terms of homogeneous coordinates @xmath69 s the equations take the form @xmath76 = 0.\ ] ] using the projector p we can rewrite ( [ 1_11 ] ) as @xmath77=0,\ ] ] or equivalently as the conservation law @xmath78+\bar{\partial}\left[\partial p,\,p\right ] = 0.\ ] ] further , introducing the @xmath79 by @xmath79 matrix @xmath80 @xmath81 = \frac{\bar{\partial } f \otimes \fd - f\otimes \bar{\partial}\fd}{\fd\cdot f } + \frac{f\otimes \fd}{(\fd \cdot f)^2 } \left[(\bar{\partial}\fd \cdot f ) - ( \fd \cdot \bar{\partial } f)\right]\ ] ] and noting that its hermitian conjugate is @xmath82 = \frac{-\partial f \otimes
\fd + f\otimes \partial \fd}{\fd\cdot f } + \frac{f\otimes \fd}{(\fd\cdot f)^2 } \left[(\partial \fd \cdot f ) - ( \fd \cdot \partial f)\right],\ ] ] we can reformulate ( [ 1_13 ] ) succinctly as @xmath83 it follows from the above ( [ 1_16 ] ) that @xmath84 is an hermitian matrix , i.e. @xmath85 .
one can check by a straightforward computation that the complex - valued functions @xmath86 satisfy @xmath87 for any solution @xmath69 of the euler
lagrange equations ( [ 1_11 ] ) .
note that @xmath88 and @xmath89 are invariant under global @xmath90 transformation , i.e. @xmath91 .
in this section we collect several facts concerning realization of projective space @xmath0 as a homogeneous space of the special unitary group @xmath41 and discuss related decompositions of the group and its lie algebra @xmath2 .
the standard reference , where all the details missing here can be found , is the book of helgason@xcite . as is well known , the space @xmath0 consists of complex lines through the origin @xmath92 in @xmath93 ( the one dimensional complex subspaces of @xmath93 ) .
we denote by @xmath94 the map which associates to any nonzero vector @xmath95 the line passing through the origin and @xmath96 , so that @xmath97,\ ] ] the numbers @xmath98 , determined up to a nonzero complex number , are called the homogeneous coordinates of the line @xmath99 .
the restriction of @xmath94 to the unit sphere @xmath100 remains surjective the resulting map @xmath101 is known as the hopf fibration .
observe that if the line @xmath102 passes through the point @xmath103 , then the fiber over @xmath102 , @xmath104 , is just the great circle in @xmath105 passing through @xmath106 .
for any given @xmath107 one introduces the so called affine or inhomogeneous coordinates defined in the complement of the set @xmath108 by the prescription @xmath109\mapsto \biggl(\frac{z_{0}}{z_{j}},\,\ldots,\,\frac{z_{j-1}}{z_{j}},\,\frac{z_{j+1}}{z_{j}},\,\ldots,\,\frac{z_{n}}{z_{j}}\biggr)\ ] ] which sets up a natural isomorphism of @xmath110 with @xmath111 . in the particular case of the affine coordinates defined in the set @xmath112 we shall write @xmath113 . by transitivity of the action of @xmath41 on the set of lines in @xmath93
one has a natural identification @xmath114 , with @xmath115 denoting the isotropy group of the standard reference point @xmath116 .
now @xmath117 and the identification above is written as @xmath118 . passing to the lie algebra level and denoting the respective lie algebras by @xmath119 and @xmath120
one has the direct sum decomposition of the isotropy lie algebra @xmath121 with @xmath122 being its center . to study other decompositions
we first recall that the killing form of @xmath119 is given by the formula @xmath123 and is negative definite .
the space @xmath124 of skew - hermitian matrices can thus be given the structure of a real euclidean space of dimension @xmath125 by taking the negative of the killing form as the inner product .
the orthogonal complement to @xmath126 with respect to this inner product consists of matrices of the form @xmath127 where @xmath128 and @xmath129 is the @xmath130 zero matrix , and this yields the orthogonal decomposition @xmath131 .
this later fact is a starting point for introducing a useful parametrization of the projective space , analogous to the spherical coordinates on the euclidean sphere .
observe that the adjoint action of the isotropy group @xmath115 on @xmath132 reduces to the action of @xmath133 on @xmath110 given by the following formula @xmath134 the action is clearly transitive on the unit sphere in @xmath132 and essentially this fact implies validity of the next result ( for a general form of such decompositions cf .
@xcite ) . to formulate it we first introduce more notations . set @xmath135 and let @xmath136 .
the lie subgroup @xmath137 of @xmath41 consists of matrices @xmath138 where we have set @xmath139 and which is isomorphic to @xmath140 .
the corresponding maximal torus in @xmath0 is @xmath141 \mid \za\in { { { \mathbb r}}}\right \}\ ] ] denoting further by @xmath142 the centralizer and by @xmath143 the normalizer of @xmath144 in @xmath115 , i.e. @xmath145 and @xmath146 , we see that @xmath147 the factor group @xmath148 is the weyl group associated with @xmath0 .
we can now formulate the main result of this section , which will be used extensively later on .
it describes a certain decomposition of the group @xmath41 related to the spherical parametrization of the projective space and is stated as the point * c ) * of the proposition below .
the points * a ) * and * b ) * are included for readers convenience and comprise the classical decompositions found e.g. in @xcite ) .
it should be pointed out that * c ) * is an elaboration of the result stated in @xcite , but proved there only for @xmath149 .
[ theo : polar decomp . ]
+ let @xmath150 , @xmath151 , and let @xmath152 denote the image of @xmath153 in @xmath115 by means of the injection @xmath154 .
the remaining notations are as explained above . * a ) * every element of @xmath155 can be written as a product @xmath156 with @xmath157 .
more precisely , the map @xmath158 arising from the group multiplication , is a smooth surjection .
* b ) * the map @xmath159 is a smooth surjection and is a double covering on the complement of the point @xmath160 . *
c ) * let @xmath161 be the diagonal embedding of @xmath162 into @xmath163 .
the map @xmath164 is a smooth surjection .
we are going to prove only part * c ) * of the statement and this follows by simple matrix calculations from the polar decomposition given in equation ( [ eqn : cartan_decomp ] ) .
assume that we have a product of the form @xmath165 where @xmath166 and @xmath167 for @xmath168 . by splitting a factor of the form @xmath169 from the matrix on the left hand side in this product and commuting it with the middle term
, we can bring the entries in the top left corners of the matrices on both sides of equation ( [ eqn : prelim_prod ] ) to coincide with each other , thus obtaining the product @xmath170 with the relations @xmath171 for @xmath168 still satisfied .
now it remains only to split off the factors of the form @xmath172 from the both extreme terms to get the sought for form ( [ eqn : rowe_decomp ] ) of the product .
[ rem : product]the statement in * c ) * above is simply that every element of @xmath41 can be written as a product of four matrices of the above given form , and writing down the product of the middle terms in equation ( [ eqn : rowe_decomp ] ) explicitely we obtain @xmath173 so that the right hand side of the decomposition ( [ eqn : rowe_decomp ] ) reduces to @xmath174 where @xmath175 and @xmath176 , which up to unimportant changes in parametrization , is the expression given in ( * ? ? ?
* equation ( 2 ) on p. 3605 ) . the polar decomposition ( [ eqn : cartan_decomp ] ) reduces to the following decomposition of @xmath149 which can be found in the paper of d. j. rowe and all @xcite .
[ coro : cartan_for_su3 ] each element of the @xmath149 group can be decomposed into the following product : @xmath177 where @xmath178 for @xmath168 and @xmath179 with @xmath180 with @xmath181 . writing this more explicitely we obtain @xmath182 where @xmath183 for @xmath168 satisfy @xmath184 and @xmath180 is of modulus @xmath37 ;
clearly @xmath185 we finish this section by presenting the explicit orthogonal bases for the lie algebras @xmath3 and @xmath4 , which will be used in our future discussion of the @xmath186 and @xmath187 models . for uniformity
we use the inner product @xmath188 where the killing form is given by the formula @xmath189 the orthogonal basis with respect to the killing form is given for the case @xmath5 by the matrices @xmath190 , for @xmath191 , where @xmath192 denote the pauli matrices @xmath193 now , @xmath194 is eight - dimensional ( over @xmath195 ) and consists of matrices of the form @xmath196 for this case we can choose a basis adapted to the decomposition @xmath197 , where the isotropy lie subalgebra @xmath126 is given by @xmath198 and may be further decomposed as @xmath199 accordingly , as its orthogonal basis , we take @xmath200 , where @xmath201 its orthogonal complement @xmath132 consists then of matrices defined in ( [ eqn : z(x ) ] ) which take the form @xmath202 we suplement the above defined basis of @xmath126 by the following basis for @xmath132 @xmath203 and observe that these bases are orthogonal with respect to the killing form . thus the matrices @xmath204 form an orthogonal basis for @xmath194 .
in order to investigate immersions defined by means of solutions of the @xmath0 models and , in particular , to envisage the moving frames and the corresponding gauss - weingarten and the gauss - codazzi - ricci equations , it is convenient to exploit the euclidean structure of the @xmath2 lie algebras leading to an identification @xmath205 described in the previous section
. let us assume that the matrix @xmath80 given by ( [ 1_14 ] ) is constructed from a solution @xmath69 of the euler lagrange equation ( [ 1_11 ] ) defined on an open connected and simply connected set @xmath206 .
the conservation law ( [ 1_16 ] ) then holds and implies that the matrix - valued 1-form @xmath207 is closed @xmath208 and takes values in the lie algebra @xmath2 of antihermitian matrices . by decomposing @xmath209 into the real and imaginary parts we write @xmath210 where the @xmath37-forms @xmath211 and @xmath212 with values in @xmath213 are anti - symmetric and symmetric , respectively , i.e. @xmath214 with the superscript @xmath215 denoting the transposition . from the closedness of the 1-form
@xmath209 it follows that the integral @xmath216 locally depends only on the end points of the curve @xmath28 ( i.e. it is locally independent of the trajectory in the complex plane @xmath217 ) .
the integral defines a mapping @xmath218 which will be called the generalized weierstrass formula for the immersion .
the tangent vectors of this immersion , by virtue of ( [ 3_2 ] ) , are @xmath219 and the complex tangent vectors are @xmath220 hence a surface @xmath221 associated with the @xmath0 model ( [ 1_11 ] ) by means of the immersion ( [ 3_8 ] ) satisfies the following [ theo : metric_form ] + components of the metric form induced on @xmath221 by the euclidean structure in @xmath2 defined by the negative of the killing form ( [ eqn : killing ] ) are given by @xmath222 .
\nonumber \end{aligned}\ ] ] the components of the metric form with respect to the complex tangent vectors are given by the following expressions @xmath223 where @xmath88 and @xmath89 are functions defined by ( [ 1_18 ] ) and @xmath224 is a non - negative ( real valued ) function given by @xmath225 consequently the first fundamental form @xmath226 of the surface @xmath221 is given with respect to the complex coordinates @xmath227 by @xmath228 in section [ sec : cp1 ] we compute explicitely coefficients of the metric form in the case of @xmath35 model . as usual
, we denote @xmath229 the determinant of the metric form .
it is known that the gaussian curvature of the surface @xmath221 with respect to the induced metric is given by @xmath230 .
\label{k4_9}\ ] ] the quantity @xmath231 defined on @xmath221 , called hopf differential , is invariant with respect to conformal changes of coordinates .
we use this freedom to simplify the corresponding equations .
our next task is to determine a moving frame on the surface @xmath221 and to write the corresponding gauss - weingarten equations expressed in terms of a solution @xmath69 satisfying the @xmath0 sigma model equations ( [ 1_11 ] ) .
using the gramm - schmidt orthogonalization procedure to construct and write explicitly expressions for the normals @xmath232 to a given surface in @xmath233 can , in general , be a complicated task .
an alternative way we propose here involves the use of the isomorphism of @xmath233 with the lie algebra @xmath2 . in this representation
, the equations for a moving frame on the surface can be written in terms of @xmath79 by @xmath79 skew - hermitian matrices .
to simplify , in the following calculations we suppress the normalization factor @xmath234 in the definition of the inner product in @xmath2 cf .
( [ eqn : killing ] ) .
we introduce real normal vectors @xmath235 to the surface @xmath221 and consider the frame @xmath236 with components satisfying the following conditions @xmath237 next we define @xmath238 now we can formulate the following [ prop3 ] + for any solution @xmath69 of the @xmath0 sigma model equations ( [ 1_11 ] ) , such that the determinant of the induced metric @xmath239 is nonzero in some neighborhood of a regular point @xmath240 in @xmath217 , there exists in this neighborhood a moving frame @xmath241 on this surface which satisfies the following gauss
weingarten equations @xmath242 where the @xmath125 by @xmath125 matrices @xmath243 and @xmath244 have the form @xmath245 and @xmath246 the elements of @xmath243 and @xmath244 take the form @xmath247 where @xmath248 the gauss
ricci equations are given by @xmath249=0\ ] ] and coincide with the equations of the @xmath0 sigma model ( [ 1_11 ] ) .
we note that the elements @xmath250 are the usual christoffel symbols .
* note that for any solution of the @xmath0 equations ( [ 1_11 ] ) the matrices @xmath251 and @xmath252 are defined by ( [ 3_10 ] ) .
as can be checked by a straightforward computation using ( [ 1_11 ] ) , the mixed derivatives @xmath253 and @xmath254 coincide and are normal to the surface @xmath255\nonumber\\ & = & \frac{1}{\fd f } ( p \p f \otimes \bp \fd p - p \bp f \otimes
\p \fd p ) \nonumber\\ & & \qquad + \frac{1}{(\fd f)^2 } ( \p \fd p \bp f - \bp \fd p \p f ) f \otimes \fd\nonumber\\ & = & -[\bp p , \p p ] = \bp \p x .\end{aligned}\ ] ] combining this equation with the @xmath0 equations ( [ 1_11 ] ) , expressed in terms of the projector p , we obtain @xmath256 \cdot [ \p p , p ] ) } = 0,\\ & \tr{(\p \bp x \cdot \bp x ) } = \tr{([\p p , \bp p ] \cdot [ \bp p , p ] ) } = 0 .
\end{split}\ ] ] as a direct consequence of differentiation of the normals ( [ 3_17 ] ) we get @xmath257 and @xmath258 using the expressions ( [ 3_18])-([3_20 ] ) and ( [ 3_11 ] ) ) we come to the following set of linear equations @xmath259 which allow us to determine elements @xmath260 and @xmath261 in terms of the coefficients of the metric form , @xmath262 and @xmath263 . as can be easily calculated , they take the form ( [ 3_21 ] ) postulated in proposition [ prop3 ] .
the second derivatives @xmath264 and @xmath265 are @xmath266,\\ \bp^2 x & = \frac{1}{\fd f } ( f \otimes \bp^2 \fd p - p \bp^2 f \otimes \fd)\\ & + \frac{2}{(\fd f)^2 } \left[(\fd \cdot \bp f ) p \bp f \otimes \fd - ( \bp \fd \cdot f ) f \otimes \bp \fd p\right ] . \end{split}\ ] ] let us observe that the following traces ( and their complex conjugates ) vanish : @xmath267 this means that the vectors corresponding to the matrices @xmath268 and @xmath269 are normal to the surface determined by ( [ 3_18 ] ) . substituting ( [ 3_30 ] ) into equations ( [ 3_27 ] ) and solving the obtained linear systems we can determine @xmath270 , @xmath271 which prove to be of the form ( [ 3_22 ] ) . + finally , the gauss codazzi ricci ( gcr ) equations are the necessary and sufficient conditions for a local existence of a surface and are the compatibility conditions of the gauss - weingarten equations . in our case
the gcr equations coincide with the @xmath0 sigma model equations ( [ 1_11 ] ) and are given in a matrix form by ( [ 3_23 ] ) .
so , with any solution @xmath69 of the @xmath0 model we can associate a surface defined by ( [ 3_7 ] ) .
* + making use of the expressions for the second derivatives of @xmath272 , the induced metric and for the elements @xmath270 appearing in matrices @xmath243 and @xmath244 , we can write explicitly the second fundamental form of the surface in terms of the model @xmath273 where the symbol @xmath274 denotes the normal part of matrices @xmath275 and the indices @xmath276 stand for @xmath18 or @xmath19 .
the quantities @xmath270 are given by ( [ 3_22 ] ) .
+ the mean curvature vector can also be expressed in terms of the model as follows : @xmath277 -2g_{\xi,\bar\xi } \p \bp x \right.\\ & \left .
\qquad + g_{\xi,\xi}[\bp^2 x -a_{2,1 } \p x -a_{2,2 } \bp x]\right\}. \end{split}\ ] ] the willmore functional of a surface has the form @xmath278 when a solution @xmath69 satisfying the @xmath279 model ( [ 1_11 ] ) is defined on the whole riemannian sphere @xmath280 then the integral @xmath281 ) } \d \xi \d \bar{\xi } \nonumber\\ & = & \frac{1}{8\pi } \int_{s^2 } \frac{1}{\fd f } ( \p \fd p \bp f - \bp \fd p \p f ) \d \xi \d \bar{\xi}\end{aligned}\ ] ] is an invariant of the surface and is known as the topological charge of the model . if the integral ( [ 3_35 ] ) exists then it is an integer which characterizes globally the surface under consideration . + summarizing
, we can now state the following analog of the bonnet theorem , cf .
@xcite . for the complex - valued function @xmath69 satisfying the @xmath0 sigma model equations ( [ 1_11 ] ) , the generalized weierstrass formula for immersion ( [ 3_8 ] ) , i.e. @xmath282d\xi + [ \bar{\partial}p , p]d\bar{\xi}\mbox{,}\ ] ] describes a surface in @xmath2 .
this surface is determined by its first and second fundamental forms ( [ 3_14 ] ) and ( [ 3_31 ] ) uniquely up to euclidean motions . finally , it is worth noting that the method described above may be of use in the study of the elliptic periodic two - dimensional toda lattice ( @xmath283 ) which is related to surfaces immersed in @xmath2 lie algebra @xcite .
the equations of @xmath283 can be written in a matrix form as the zero curvature equations @xmath284 $ ] , formally identical with the gauss codazzi
ricci equation ( [ 3_23 ] ) , where the two @xmath79 by @xmath79 matrices @xmath243 and @xmath244 are defined as follows @xmath285 where for @xmath286 , we set @xmath287 , @xmath288 and @xmath289 . it is known @xcite that the zero - curvature equation ( [ 3_23 ] ) for matrices ( [ 4.30 ] ) implies the existence of a complex valued function @xmath290 such that @xmath291 so , according to the proposition [ prop3 ] , we can identify ( [ 4.31 ] ) with the complex tangent vectors ( [ 3_10 ] ) of the immersion ( [ 3_8 ] ) .
hence the @xmath283 equations can be viewed as being associated with the specific form ( [ 1_13 ] ) of the @xmath0 model .
establishing this link could be useful for determining certain geometric characteristics of surfaces corresponding to the elliptic periodic two - dimensional toda lattice , but this point will not be considered here .
in this section we sketch an application of the techniques developed in the previous sections to the case of @xmath35 sigma model .
this allows us to put the results obtained in the earlier publications@xcite in a broader perspective , as well as to point out some further geometrical properties of surfaces obtained from this model .
the fields of the @xmath35 model in the notation of the section [ sec : sun ] are given by @xmath63=[f_0,\,f_1]$ ] , where @xmath292 , but it is customary to replace here the homogeneous coordinates @xmath293 by the affine coordinate @xmath294 .
the euler lagrange equation ( [ 1_11 ] ) reduces to @xmath295 and the matrix @xmath80 of the eqn .
( [ 1_14 ] ) is then given by @xmath296 the @xmath37-form @xmath297 of the eqn .
( [ 3_2 ] ) defines an immersion into @xmath3 to pass to the euclidean space @xmath11 we first compute its real and imaginary parts and obtain @xmath298,\end{aligned}\ ] ] and @xmath299.\end{aligned}\ ] ] these in turn can be expressed in terms of the pauli matrices @xmath300 ( cf .
( [ pauli ] ) ) as follows @xmath301 where the real - valued @xmath37-forms @xmath302 , are given by @xmath303 \d \xi\\ & + \left[\bp \bw + \bw^2\bp w + \bp w + w^2 \bp \bw\right ] \d \bar{\xi } \big{\}},\\ \d x_2 & = \frac{1}{2(1+|w|^2)^2 } \big{\{}\left[-\p \bw - \bw^2\p w + \p w + w^2\p \bw\right ] \d \xi \\ & + \left[\bp \bw + \bw^2\bp w - \bp w -w^2 \bp \bw\right ] \d \bar{\xi } \big{\}},\\ \d x_3 & = \frac{1}{(1+|w|^2)^2 } \left\{\left[w \p\bw - \bw \p w\right ] \d \xi + \left[\bar{w}\bp w - w \bp \bw\right]\d \bar{\xi}\right\ } , \end{split}\ ] ] which is the generalized weierstrass formula for an immersion into @xmath304 .
an interested reader may check that the equations ( [ 4_3 ] ) yield the classical weierstrass formula ( [ minimal ] ) under the substitution @xmath294 . starting from a particular solution @xmath9 of the @xmath305 sigma model equation ( [ 1_22 ] ) one constructs an immersion in @xmath33 by the use of the formulae ( [ 4_3 ] ) .
the following is now readily obtained form the proposition [ theo : metric_form ] . for the immersion given by equations ( [ 4_3 ] ) the coefficients of the induced metric
are given by the following expressions @xmath306 the complex form of the induced metric in this case is given by @xmath307 for solutions of ( [ 1_22 ] ) which are defined over @xmath308 , the function @xmath9 can be only holomorphic or antiholomorphic ( cf .
@xcite ) . for holomorphic @xmath9 equations ( [ metric ] )
reduce to @xmath309 implying that the immersion is conformal .
these relations , as shown earlier @xcite , imply also @xmath310 .
the gaussian curvature is @xmath311 , and the first and the second fundamental forms for the immersion are equal , @xmath312 moreover , as shown by k. kenmotsu in @xcite , the function @xmath9 represents the complex gauss map of the surface .
geometrically , all that means that solutions of the @xmath313 model ( [ 1_22 ] ) defined over @xmath308 parametrize the standardly immersed sphere in @xmath33 . this had already been shown in the case of instanton solutions of the @xmath314 sigma model in @xcite .
here we apply our considerations to the @xmath315 model for which we construct the associated immersion of a surface @xmath221 in @xmath316 and compute some of its geometric characteristics . for the case of
@xmath45 we can pass from the representation @xmath63=[f_0,f_1,f_2]$ ] to the inhomogeneous ( affine ) coordinates @xmath317 and @xmath318 .
now the euler lagrange equations ( [ 1_11 ] ) take the form @xmath319 where @xmath320 as was noted in @xcite , the metric induced by the immersion @xmath321 in @xmath322 is conformal for holomorphic solutions of the @xmath38 model defined over @xmath308 and is then given by @xmath323 we shall write @xmath324 where @xmath325 is a complex - valued function of @xmath326 given by @xmath327 , @xmath30 of the @xmath38 sigma model equations ( [ 1_23 ] ) defined over @xmath308 such that the induced metric is nonzero in some neighbourhood of a regular point @xmath328 , determines a conformal parametrization of a surface @xmath221 immersed in the @xmath4 lie algebra . its moving frame on @xmath221 can be written in terms of @xmath43 by @xmath43 skew - hermitian matrices and is of the form @xmath329 where the complex tangent vectors to the surface @xmath221 are given by the following traceless matrices @xmath330 and where we have defined @xmath331 * remark 2 . * the explicit expressions for the complex normals to this surface immersed in @xmath4 , in terms of @xmath332 and @xmath333 ,
can be found in the appendix to @xcite .
* proof . * due to the normalization of the function @xmath272 ( given by equations ( [ 4_4 ] ) and ( [ 3_10 ] ) ) we can express the moving frame @xmath334 on a surface @xmath221 in terms of the adjoint @xmath4 representation @xmath335 where @xmath336 note that @xmath337 span over @xmath338 , the same space as @xmath339 . using the polar decomposition of the @xmath340 group given in section [ sec : sun ] , cf .
( [ eqn : rowe_decomp ] ) , a general @xmath340 matrix @xmath341 can be decomposed into a product of three @xmath342 factors .
performing the multiplication in the expression ( [ factor ] ) and setting @xmath343 and @xmath344 , we obtain @xmath345 where the complex - valued functions @xmath346 of @xmath347 satisfy and @xmath348 are real - valued functions of @xmath349 .
the requirement that the parametrization of a surface @xmath221 is conformal implies the following relations @xmath350 and @xmath351 now we have to determine the form of a 8-parameter representation of the matrix @xmath341 in terms of @xmath352 , compatible with the @xmath38 sigma model ( [ 1_23 ] ) . using the @xmath43 by @xmath43 projector matrix
@xmath353 we can write the euler lagrange equations ( [ 1_23 ] ) in the form of a conservation law ( [ 1_16 ] ) for the matrix @xmath354 where we have defined the following expression @xmath355 and the quantity @xmath243 is given by equation ( [ 1_25 ] ) . according to ( [ 1_14 ] ) and ( [ 3_10 ] ) , the matrices @xmath356 and @xmath357 take the required form ( [ 4_6 ] ) .
let us note that to satisfy the compatibility condition for ( [ 4_7 ] ) ( i.e. @xmath358 it is sufficient , in view of the conservation laws ( [ 1_13 ] ) , to postulate that the condition ( [ 3_10 ] ) holds for the matrix @xmath80 given by ( [ kk ] ) .
so , we can determine functions @xmath359 and @xmath360 , appearing in the matrix @xmath361 , in terms of @xmath362 and @xmath363 , @xmath30 . by a straightforward algebraic computation
we get @xmath364 where we have used the notation introduced in ( [ dd ] ) and have set @xmath365 .
given the above form of the matrix @xmath341 , the matrices @xmath366 , @xmath367 and the @xmath368 , @xmath369 the moving frame ( [ 4_7 ] ) adopts the required forms ( [ 4_5])-([4_6 ] ) .
one can check directly that it satisfies the gauss - weingarten equations ( [ 3_18 ] ) . in our case the corresponding gcr equations , which are the compatibility conditions for ( [ 3_18 ] ) , coincide with the @xmath38 sigma model equations ( [ 1_23 ] ) .
thus we have proved that any holomorphic solution of the @xmath38 model defined over @xmath308 gives a surface conformally immersed in @xmath316 with the moving frame given by ( [ 4_5 ] ) and ( [ 4_6 ] ) . * q.e.d * +
based on the results of the previous sections we can now construct certain classes of two - dimensional surfaces immersed in @xmath316 . for this purpose
we use the @xmath38 sigma model defined over @xmath308 . for this model
all solutions of the euler
lagrange equations ( [ 1_23 ] ) are known @xcite . if we require the finiteness of the action([1_7 ] )
they split into three classes : holomorphic ( i.e. @xmath370 ) , antiholomorphic ( i.e. @xmath371 ) and the mixed ones .
the latter ones can be determined from either the holomorphic or antiholmorphic functions by the following procedure @xcite : + consider three arbitrary holomorphic functions @xmath372 and define for any pair of them the wronskian functions @xmath373 then one can check that the map @xmath374 , where @xmath375 is a solution of the @xmath38 sigma model , so called the mixed one , and hence the ratios @xmath376 satisfy equations ( [ 1_23 ] ) .
an alternative approach starts with any antiholomorphic functions @xmath377 and constructs functions @xmath378 and consequently @xmath362 as above but using @xmath379 instead of @xmath17 in the definition of @xmath380 .
it yields results which are complementary to the ones obtained by the first approach .
let us note here that the requirement of finite action ( [ 1_7 ] ) excludes solutions which admit painlev transcendent ( i.e. all critical points are fixed independent of initial data ) , branch points or essential singularities .
+ now , let us discuss some classes of surfaces immersed in @xmath4 algebra which can be obtained directly by applying the weierstrass representation ( [ 3_8 ] ) .
for the @xmath381 model ( [ 1_23 ] ) , the matrix @xmath80 in terms of @xmath362 and @xmath382 , @xmath30 is given by ( [ kk ] ) . from the equation ( [ 3_2 ] ) we obtain for the real and imaginary parts of the @xmath37-form @xmath383 the expressions @xmath384,\\ \mathrm{d}x^2 & = \frac{1}{2 } \left[(\mathbb{k}^{\dagger } + \bar{\mathbb{k}})\mathrm{d}\xi + ( \mathbb{k}+\mathbb{k}^t ) \mathrm{d } \bar{\xi}\,\right ] .
\end{split}\ ] ] clearly , the matrices @xmath211 and @xmath212 are antisymmetric and symmetric , respectively , and hence can be decomposed in terms of the chosen basis in @xmath4 given by ( [ s_4])([s_8 ] ) as follows @xmath385 . \end{split}\ ] ] as a result of the decomposition ( [ 5_7 ] ) , there exists eight real - valued functions @xmath386 @xmath387 which determine the generalized weierstrass representation of surfaces associated with the @xmath38 model ( [ 1_23 ] ) . considering the off - diagonal entries of the matrices @xmath211 and
@xmath212 we get @xmath388\d \xi + [ \frac{1}{a}(\bp w_1 - \bp \bw_1 ) + \frac{\bar{\rho}}{a^2 } ( \bw_1+w_1)]\d \bar{\xi},\\ \d x_2 & = -i\big\{[\frac{1}{a } ( \p w_1 + \p \bw_1 ) + \frac{\rho}{a^2}(\bw_1 - w_1)]\d \xi \\ & \quad + [ -\frac{1}{a } ( \bp \bw_1 + \bp w_1 ) + \frac{\bar{\rho}}{a^2 } ( \bw_1 - w_1 ) ] \d \bar{\xi}\big\},\\ \d x_5 & = -i\big\{[\frac{1}{a } ( \p w_2 + \p \bw_2 ) + \frac{\rho}{a^2}(\bw_2-w_2)]\d \xi \\ & \quad + [ -\frac{1}{a}(\bp \bw_2 + \bp w_2 ) + \frac{\bar{\rho}}{a^2 } ( \bw_2 - w_2)]\d \bar{\xi}\big\ } , \end{split}\ ] ] @xmath389\d \xi\\ & \quad + [ \frac{1}{a}(\bw_2 \bp w_1 - w_1 \bp \bw_2 - \bw_1 \bp w_2 + w_2 \bp \bw_1 ) + \frac{\bar{\rho}}{a^2 } ( w_1\bw_2 - \bw_1 w_2 ) ] \d \bar{\xi } \big\},\\ \d x_7 & = [ \frac{1}{a } ( w_1 \p \bw_2 - \bw_2 \p w_1 + w_2 \p \bw_1 - \bw_1 \p w_2 ) + \frac{\rho}{a^2 } ( w_1 \bw_2 + \bw_1 w_2 ) ] \d \xi \\ & \quad + [ \frac{1}{a}(\bw_2 \bp w_1 - w_1 \bp \bw_2 + \bw_1 \bp w_2 - w_2 \bp \bw_1 ) + \frac{\bar{\rho}}{a^2 } ( w_1 \bw_2 + \bw_1 w_2)]\d \bar{\xi},\\ \d x_8 & = [ \frac{1}{a}(\p \bw_2 - \p w_2 ) + \frac{\rho}{a^2 } ( \bw_2+w_2)]\d \xi + [ \frac{1}{a } ( \bp w_2 - \bp \bw_2 ) + \frac{\bar{\rho}}{a^2 } ( \bw_2 + w_2 ) ]
\d \bar{\xi}. \end{split}\ ] ] from the diagonal entries of the matrix @xmath212 we obtain @xmath390\d \xi \\ & \quad + [ \frac{1}{a } ( \bw_1 \bp w_1 - w_1 \bp \bw_1 ) + \frac{\bar{\rho}}{a^2 } |w_1|^2 ] \d \bar{\xi}\big\},\\ \d x_4 & = 2 \big\{[\frac{1}{a } ( w_2 \p \bw_2 - \bw_2 \p w_2 ) + \frac{\rho}{a^2 } |w_2|^2]\d \xi \\ & \quad + [ \frac{1}{a } ( \bw_2 \bp w_2 - w_2 \bp \bw_2 ) + \frac{\bar{\rho}}{a^2 } |w_2|^2 ] \d \bar{\xi } \big\}. \end{split}\ ] ] note that by virtue of the conservation law ( [ 1_16 ] ) , the @xmath37-forms ( [ 5_8 ] ) and ( [ 5_9 ] ) are the exact differentials of real - valued functions .
the functions @xmath391 @xmath392 constitute the coordinates of the radius vector @xmath393 of a two - dimensional surface in @xmath316 .
thus , if the complex - valued functions @xmath362 , @xmath30 correspond to any solution of the @xmath38 sigma model ( [ 1_13 ] ) , then we can use the generalized weierstrass formulae ( [ 5_8 ] ) and ( [ 5_9 ] ) to construct a two - dimensional surface in @xmath316 uniquely defined by this solution .
let us note that in the limiting case when @xmath394 or when we put @xmath395 or @xmath396 , the weierstrass formulae ( [ 5_8 ] ) and ( [ 5_9 ] ) reduce to formulae ( [ 4_3 ] ) describing the immersion in the the @xmath35 case .
these limits characterize some properties of solutions of both systems ( [ 1_22 ] ) and ( [ 1_23 ] ) .
now , let us discuss some classes of surfaces immersed in @xmath316 which can be determined directly by applying the weierstrass representation ( [ 5_8 ] ) and ( [ 5_9 ] ) .
+ * example 1 .
* as well known , the simplest case of solutions of the @xmath38 model is obtained by taking @xmath362 to be analytic . in this case
@xmath397 and so many expressions in ( [ 5_8 ] ) and ( [ 5_9 ] ) vanish .
in fact we get , ( with @xmath398 denoting the complex conjugate ) @xmath399 these expressions can be easily integrated giving us , up to overall constants that can be added to any @xmath400 : @xmath401 note that in general we have a surface in @xmath316 .
using ( [ 4_2 ] ) it is very easy to calculate the curvature as we know that @xmath402 where the dot denotes the differentiation with respect to @xmath18 .
then the gaussian curvature is given by @xmath403 now , by setting @xmath396 the above model is reduced to the @xmath35 case , with @xmath404 and the remaining components @xmath405 ( for @xmath406 ) . note
that in this case our surface is the surface of an appropriately located sphere . to see this note that @xmath407 then we have @xmath408 and so we see that all the points lie on the surface of a sphere of unit radius , centred at @xmath409
of course , the number of times this surface is covered depends on the degree of @xmath332 , i.e. the topological charge of the map .
this is , of course , consistent with ( [ holo1 ] ) , which gives a constant . in the @xmath381 the situation is more complicated but also more can be said about the surface ; i.e. , for example , all points lie on the hyperellipsoid surface @xmath410 however , the gaussian curvature is not necessarily constant . to see this we use ( [ holo1 ] ) and
consider very specific fields , namely : @xmath411 then @xmath412 and so @xmath413 then @xmath414 which shows that the curvature is constant only when @xmath415 or @xmath416 , but not in other cases .
hence , in general , the surfaces do not have a constant gaussian curvature .
* example 2 . *
the simple mixed solution obtained by choosing @xmath417 , @xmath418 , @xmath419 in the formulae ( [ 6_1])([5_3 ] ) gives us the following @xmath420 the weierstrass representation ( [ 5_8 ] ) , ( [ 5_9 ] ) can be integrated and it leads to the following expression for the immersion of our surface in @xmath316 in polar coordinates @xmath421 by @xmath422 the curvatures can be calculated , but the expressions are rather involved , so we omit them here .
+ * example 3 .
* another interesting class of mixed solutions of the @xmath381 model is given by @xmath423 in this case the system of euler lagrange equations ( [ 1_23 ] ) simplifies considerably and therefore the weierstrass formulae ( [ 5_8 ] ) , ( [ 5_9 ] ) can be easily integrated . in polar coordinates , setting @xmath424 , we obtain @xmath425 this describes a surface of revolution which is contained in a subspace of dimension @xmath43 ( see fig .
1 ) . the first and second fundamental forms are @xmath426\mbox{,}\\ ii & = { 4\over ( 1+r^2)^2(r^4 + 6r^2 + 1)^{1/2}}\left[(r^2 + 3)\d r^2 + r^2(r^2 - 1)\d \varphi^2\right]\mbox{. } \end{split } \label{5_22}\ ] ] the gaussian and mean curvature are @xmath427 since the curvatures are not constant , the surface can not be obtained from the @xmath35 model .
there are reasons to expect that the association between the weierstrass representation of surfaces immersed in @xmath2 lie algebras and the solutions of the euclidean two - dimensional @xmath279 sigma models , described in this paper , can be found also for more general sigma models . a good object of the investigation in this direction are the complex grassmannian sigma models which take values on symmetric spaces @xmath428 .
these models share many important common properties with the @xmath279 models considered here .
they possess an infinite number of conserved quantities , as well as infinite - dimensional dynamical symmetries which generate the kac - moody algebra .
the grassmannian sigma model equations , just like those of the @xmath0 models , have a hamiltonian structure and complete integrability with a well formulated linear spectral problem .
many classes of solutions of these equations are known , see eg.@xcite ; they can be expressed in terms of holomorphic functions and functions obtained from them by a procedure which is a generalization of the transformation which generates all solutions of the @xmath0 models . the complex grassmannian sigma model in two euclidean dimensions
are defined in terms of fields @xmath429 where @xmath430 , taking values in the complex grassmann manifold @xmath431 , where @xmath432 . by decomposing @xmath239 into two blocks @xmath433 where @xmath434 are @xmath435component orthonormal column vectors , @xmath436
we define the projector matrix @xmath74 ( @xmath437 , @xmath438 ) as @xmath439 in general , it has higher rank than the corresponding matrix for the @xmath0 model . however , the equation of motion in terms of @xmath74 in this case has the same form as ( [ 1_77 ] ) and is obtained by minimizing the action of the lagrangian @xmath440,\ ] ] where @xmath441 is the covariant derivative for @xmath272 , @xmath442 the above fact implies that our method can be successfully used for constructing surfaces associated with the complex grassmannian sigma models .
the question of the diversity and complexity of these surfaces , however , remains open and has to be answered in further work . in this paper
we have shown how to generalize the old idea of enneper @xcite and weierstrass @xcite in connection with the @xmath0 sigma models and their group properties .
we have found the structural equations of surfaces immersed in @xmath2 lie algebras and expressed them in terms of any solution of the @xmath0 model .
the most important advantage of the presented method is that it is quite general . in constructing surfaces we proceeded directly from the given @xmath0 model , without refering to any additional considerations .
another important advantage of our method is that , due to the conservation laws of the @xmath0 model , the obtained expressions for surfaces are given at least in the form of quadratures .
we have discussed in detail the geometrical aspects of the constructed surfaces .
namely , we have demonstrated through the use of moving frame that one can derive , via the @xmath0 models , the first and the second fundamental forms of a given surface as well as relations between them as expressed in the gauss
weingarten and the gauss codazzi
ricci equations .
we have illustrated the proposed method of constructing surfaces in the case of low dimensional @xmath2 lie algebras .
a systematic application of the group theory makes its possible to obtain large number of particular solutions of the @xmath0 equations and associated surfaces in @xmath1 .
a question arises whether these solutions , corresponding to specific boundary conditions , are actually observable in nature .
the answer depends to a large degree on their stability .
stable solutions should be observable and should also provide the starting point for perturbative calculations
. these should in turn provide good approximative solutions relevant for situations in which the group - theoretical solutions no longer apply . in this context
another question arises , namely what physical insight one gains from exact analytic expressions for surfaces .
a particular answer is that they show up qualitative features that might be difficult to detect numerically .
we hope that our approach and our results may be useful in applications to the study of surfaces which arise in physics , chemistry and biology , by providing explicit models in situations which have been well investigated experimentally but for which the theory is not yet well developed .
further exploration of relations between various properties of harmonic maps @xmath443 and properties of surfaces in planed for future work . * acknowledgments * + partial support for a.m.g .
work was provided by a research grant from nserc of canada and fonds fqrnt du gouvernement du qubec .
wishes to acknowledge and thank the centre de recherches mathmatiques ( universit de montral ) and nserc of canada for the financial support provided for his visit to montreal .
99 a. bobenko , _ surfaces in terms of 2 by 2 matrices _ in : * harmonic maps and integrable systems * , ed .
a. fordy and j.c .
wood ; vieweg , braunschweig , 1994 .
a. bobenko and u. eitner , * painlev ' e equations in the differential geometry of surfaces * , lect .
notes math .
1753 , springer , berlin , 2000 .
m. p. do carmo , * riemannian geometry , _ mathematics : theory and applications _ * , birkh " auser , boston , 1992 r. carroll and b. konopelchenko , _ generalised weierstrass - enneper inducing conformal immersions and gravity _ , int . j. mod . phys . * a * 11 , ( 7 ) , 1183 - 1216 ( 1996 ) . f. chen , * introduction to plasma physics and controlled fusion * , vol .
1 , plasma physics , plenum press , new york , 1983 .
a. davidov , * solitons in molecular systems * , kluwer , new - york , 1991 .
f. david , p. ginsparg and y. zinn - justin , eds .
, * fluctuating geometries in statistical mechanics and field theory * , elsevier , amsterdam , 1996 .
a. m. din and w.j .
zakrzewski , _ general classical solutions of the @xmath444 model _ , nucl .
* b 174 * , 397 - 403 ( 1980 ) .
l. eisenhardt , * a treatise on the differential geometry of curves and surfaces * , ginn and company , 1909 .
a. enneper , nachr .
augustus univ .
gottingen * 12 * ( 1868 ) , 258 - 277 . e.v .
ferapontov and a.m. grundland , _ links between different analytic descriptions of constant mean curvature surfaces _ , j. nonlin .
phys . 7 , 14 - 21 , ( 2000 ) a. s. fokas and i. m. gelfand , _ surfaces on lie groups , on lie algebras and their integrability _ , comm .
phys . * 177 * , 203220 ( 1996 ) .
a. s. fokas , i. m. gelfand , f. finkel and q. m. liu , _ a formula for constructing infinitely many surfaces on lie algebras and integrable equations _ , selecta math .
new series * 6 * , 347375 ( 2000 ) .
d. j. gross , t. piran and s. weinberg , * two - dimensional quantum gravity and random surfaces * , world scientific , singapore , 1992 .
a. m. grundland and w. j. zakrzewski , _ on certain geometric aspects of @xmath445 harmonic maps _ , j. math
. phys . * 44 * , 1 , 328 - 337 ( 2003 ) .
a. m. grundland , a. strasburger and w. j. zakrzewski , _ surfaces immersed in @xmath2 lie algebras obtained from the @xmath446 sigma models _
, preprint crm-3180 , febr .
2005 , and dg/0502482 ( 2005 ) .
m. a. guest , _ harmonic two - spheres in complex projective space and some open problems , _ expo . math . * 10 * ( 1992 ) , 6187 .
m. a. guest , * harmonic maps , loop groups and integrable systems , * cambridge university press , cambridge , 1997 . a. m. grundland and l. snobl , _ description of surfaces associated with @xmath445 sigma models on minkowski space _ , j. geom
( accepted for publication , 2005 ) .
a. m. grundland and l. snobl , _ description of surfaces associated with grassmannian sigma models on minkowski space _
, j. math .
phys . * 46 * , 8 , 3508 - 3520 ( 2005 ) .
a. m. grundland and w. j. zakrzewski , _
@xmath445 harmonic maps and the weierstrass problem _
, j. math .
phys . * 44 * , 8 , 3370 - 3382 ( 2003 ) .
a. m. grundland , w. j. zakrzewski , _ the weierstrass represention for surfaces immersed in @xmath316 and the @xmath381 maps _ , j. math .
43,6 , 3352 - 3362 ( 2002 ) .
f. helein , * harmonic maps , conservation laws and moving frames * , cambridge univ . press , cambridge , 2002 .
f. helein , * constant mean curvature surfaces , harmonic maps and integrable systems * , lect . math . , birkhuser , boston , 2001 . s. helgason , * differential geometry , lie groups and symmetric spaces * , academic press , new york 1978 .
k. kenmotsu , _ weierstrass formula for surfaces of prescribed mean curvature _ , math .
245 , 8999 ( 1979 ) k. kenmotsu , * surfaces with constant mean curvatures * , ams , providence , ri , 2003 .
b. konopelchenko and g. landolfi , _ induced surfaces and their integrable dynamics ii .
generalized weierstrass representations in 4-d spaces and deformations via ds hierarchy _
, stud . in appl .
104 , 129 - 169 ( 1999 ) s. kobayashi and k. nomizu , * foundation of differential geometry * , john wiley , new - york , vol.1 and 2 , 1963 .
b. konopelchenko , _ induced surfaces and their integrable dynamics _
96 , 9 - 51 ( 1996 ) b. konopelchenko and i. taimanov , _ constant mean curvature surfaces via an integrable dynamical system _ , j. phys . a : math .
29 , 1261 - 1265 ( 1996 ) g. landolfi , _ on the canham - helfrich membrane model _ , j. phys . a : math .
gen . , * 36 * , 46994715 ( 2003 ) .
d. nelson , t. piran and s. weinberg , * statistical mechanics of membranes and surfaces * world scientific , singapore , 1992 .
r. osserman , a survey of minimal surfaces , dover , new - york 1996 .
z. ou - yang , j. lui and y. xie , * geometric methods in elastic theory of membranes in liquid crystal phases * , world scientific , singapore , 1999 .
f. pedit and u. pinkall , _ quaternionic analysis on riemann surfaces and differential geometry _ , doc
. math . * extra volume icm * , 389400 ( 1998 ) .
u. pinkall and i. sterling , _ on the classification of constant mean curvature tori _ ,
ann . math . *
130 * , 407451 ( 1989 ) .
u. pinkall , _ regular homotopy classes of immersed surfaces _ , topology * 24 * , 421434 ( 1985 )
. j. polchinski and a. strominger , _ effective string theory _ ,
* 67 * , 16811684 ( 1991 ) .
rowe , b.c .
sanders and h. de guise , _ representations of the weyl group and wigner functions for su(3 ) _ , j. math . phys . *
40 * , 36043615 ( 1999 ) . s. a. safram , * statistical thermodynamics of surfaces , interfaces and membranes * , addison - wesley , new - york , 1994 . a. sommerfeld , * lectures on theoretical physics * , vol . 13 , acad .
press , new york , 1952 .
r. s. ward , _ sigma models in @xmath447 dimensions _ , in : * harmonic maps and integrable systems * , ed .
a. fordy and j.c .
wood ; vieweg , braunschweig , 1994 . k. weierstrass , _ fortsetzung der untersuchung uber die minimalflachen _ , mathematische werke , vol 3 , verlagsbuchhandlung , hillesheim , ( 1866 ) , 219 - 248 .
t. j. willmore , * riemannian geometry * , clarendon press , oxford , 1993 . v. e. zakharov and a. v. mikhailov , _ relativistically invariant two - dimensional models of field theory which are integrable by means of the inverse scattering problem method _ , zh . eksp .
fiz . * 74 * , 1953 ( 1978 ) ; ( sov .
jetp * 47*,1017 ( 1979 ) ) .
w. j. zakrzewski , * low dimensional sigma models * , adam hilger , bristol , 1989 . | we study some geometrical aspects of two dimensional orientable surfaces arrising from the study of @xmath0 sigma models . to this aim
we employ an identification of @xmath1 with the lie algebra @xmath2 by means of which we construct a generalized weierstrass formula for immersion of such surfaces . the structural elements of the surface like its moving frame , the gauss - weingarten and the gauss codazzi
ricci equations are expressed in terms of the solution of the @xmath0 model defining it .
further , the first and second fundamental forms , the gaussian curvature , the mean curvature vector , the willmore functional and the topological charge of surfaces are expressed in terms of this solution .
we present detailed implementation of these results for surfaces immersed in @xmath3 and @xmath4 lie algebras .
[ firstpage ] |
a long - standing dichotomy in nebular astrophysics is that the heavy element abundances derived from optical recombination lines ( orls ) are systematically higher than those derived from collisionally excited lines ( cels ; see peimbert & peimbert 2006 and liu 2006 for recent reviews ) .
temperature fluctuations ( peimbert 1967 ) , density inhomogeneities ( rubin 1989 ; viegas & clegg 1994 ) and x - ray irradiation of quasineutral chemically homogeneous clumps ( ercolano 2009 ) have been proposed to explain this discrepancy . however , extensive observations of planetary nebulae ( pne ) have shown that they are incapable of accounting for all observational features ( liu 2006 ) . in order to explain the cel / orl abundance dichotomy , liu et al .
( 2000 , l2000 hereafter ) presented a bi - abundance model .
the model assumed that the nebula contains two gas components of different abundances the diffuse component of `` normal '' composition that dominates the emission of cels and another consisting of cold h - deficient inclusions posited in the diffuse gas that dominates the emission of orls .
the model explains well spectroscopic observations from the ultraviolet ( uv ) to the infrared ( ir ) .
however , detailed photoionization models are still required to test this scenario and address issues such as : ` how can the nebular physical conditions and its elemental abundances be reliably determined in the exiseence of chemical inhomogeneities ? ' , and ` what is the nature and origin of the h - deficient clumps ? ' to address these issues , this paper presents the results of three - dimensional photoionization modelling for the pn ngc6153 .
ngc6153 is an archetypal nebula showing a particularly large orl versus cel abundance discrepancy factor ( adf ) of about ten ( l2000 ) .
a large number of orls from a variety of heavy - element ions have been detected in this pn .
ngc6153 is thus an ideal object to model and to investigate the possible causes of the cel / orl abundance discrepancy problem .
early optical studies by pottasch et al .
( 1984 , 1986 ) suggested that ngc6153 has a very high metallicity .
the result was further corroborated by the observations of the short wavelength spectrometer on board the infrared space observatory ( _ iso_-sws ) in the mid - ir ( pottasch et al .
l2000 presented a comprehensive spectroscopic analysis of ngc6153 from the uv to the far - ir and found that the c , n , o , and ne abundances derived from orls are about ten times higher than those derived from cels . using simple empirical models
, l2000 showed that the discrepancy can be explained by assuming the existence of a h - deficient component embedded in the diffuse gas . using optical integral field spectroscopic observations of three pne including ngc6153 with the very large telescope fibre large array multi element spectrograph argus integral field unit , tsamis et al .
( 2008 ) provided further evidence for the existence of cold h - deficient inclusions and constrained their physical sizes to be smaller than @xmath21,000 astronomical units .
pquignot et al .
( 2002 , 2003 ) constructed one - dimensional photoionization models that contain two gas components of different abundances .
the models reproduced satisfactorily most of the observed spectral features of ngc6153 .
it is however difficult to self - consistently treat the diffuse radiation fields in one - dimensional models and investigate the size and spatial distribution of the postulated h - deficient clumps .
this motivated us to carry out further modelling using the monte carlo photoionization code mocassin ( ercolano et al .
2003a , 2005 , 2008 ) capable of dealing with an arbitrary nebular geometry and density and abundance distributions .
the paper is organized as follows . in section 2 ,
previous modelling work and observational data used to constrain our models are summarized . in section 3
, we introduce the code used for the modelling and present the modelling results . in section 4 , we discuss the implications and limitation of our results , the properties of the h - deficient inclusions and their impacts on nebular studies .
a summary is given in section 5 .
in this section , we summarize observations available to constrain the photoionization models of ngc6153 and introduce the earlier one - dimensional bi - abundance work of pquignot et al .
( 2002 , 2003 ) , which serves as the starting point of our new three - dimensional models .
spectra of ngc6153 from the uv to the far - ir are available from the literature ( e.g. , l2000 , pottasch et al .
1984 , 1986 , 2003 ) . for a given wavelength range
, there may exist more than one set of observations . due to varying aperture sizes , orientations and measurement uncertainties ,
different studies often obtain different fluxes for a given emission line . for the current work ,
we have adopted line fluxes of l2000 , who present measurements in the optical obtained with ground - based facilities , in the uv obtained with the international ultraviolet explorer ( _ iue _ ; 11503300 ) , and in the mid- to far ir obtained with the infrared astronomical satellite ( _ iras_;7.622.7@xmath3 m ) and the long wavelength spectrometer on board the _ iso _ ( _ iso_-lws ; 43197@xmath3 m ) .
l2000 obtained two sets of optical spectra .
the set of spectra with the slit oriented along the nebular minor axis allows one to investigate the spatial brightness distribution of a given line along the slit .
the other set , obtained by scanning the slit across the whole nebular surface , when combined with the total h@xmath4 flux [ @xmath5@xmath6(h@xmath4 ) = @xmath710.86ergs @xmath8s@xmath9 ; cahn et al .
1992 ] , yields integrated fluxes of the whole nebula for all the detected lines .
the integrated line fluxes can be directly compared with the model predictions .
note that l2000 used a logarithmic extinction , _
c_(h@xmath4 ) = 1.30 , deduced from the balmer decrement , * and the galactic reddening law of howarth ( 1983 ) * to deredden the uv , optical and near - ir spectra .
they adopted a different value _
c_(h@xmath4 ) = 1.19 , deduced from the observed radio continuum flux density and the h@xmath4 flux , to deredden the mid- and far - ir spectra .
the same dereddening algorithm has been used in the current study .
here we have added a few mid - ir cels falling within the spectral window of the _ iso_-sws in our modelling .
given the fact that the size of the _
iso_-sws diaphragm was not large enough to cover the entire nebula and the telescope pointing was offset by @xmath10 east and @xmath11 south of the nebular centre , aperture corrections have been applied by comparing the line fluxes obtained with _ iso_-sws and those with _
iras_. all strong emission lines considered in the previous modelling ( l2000 , pquignot et al .
2003 ) were included in the current work , c.f . the next section for details .
in addition to line fluxes from the literature , we have supplemented the data with new observations using the b&c spectrograph mounted on the european southern observatory ( eso ) 1.52 m telescope .
the spectra , obtained on 2001 may 12 , were taken in order to measure several near - ir emission lines ( he i @xmath1210830 , [ c i ] @xmath129850 , and [ s iii ] @xmath12@xmath129069,9531 ) , important for model constraints .
the he i @xmath1210830 line is particularly important to constrain the helium abundance in a bi - abundance model ( c.f . section 4.1 ) .
the spectra covered the wavelength range 770013254 , with the slit placed along the nebular minor axis .
two long exposures , each of 1800s , and one short exposure of 300s were made .
the spectra were reduced using the long92 package in midas following the standard procedures .
the spectra were bias - subtracted , flat - fielded , cosmic rays cleaned and wavelength - calibrated using exposures of a he - ar calibration lamp .
they were then flux - calibrated with the iraf package using wide - slit observations of _ hst _ standard stars .
after dereddening , the spectra were normalized such that hi _
i_(p12)/_i_(h@xmath4 ) = 0.01106 , as predicted by the recombination theory , caseb ( storey & hummer 1995 ) for @xmath13 = 6,080 k and @xmath14 = 3,500 @xmath15 .
the @xmath13 and @xmath14 values here are taken from l2000 .
fluxes deduced from these observations are tabulated in table 1 ( see section 3 ) . in an attempt to detect directly the postulated h - deficient knots and to constrain their physical sizes ,
we have also carried out long - slit spectroscopy of ngc6153 using the stis instrument on board the hubble space telescope ( _ hst _ ) . on 2002
june 34 , two gratings , g430 m and g750 m , were used to cover the spectral range @xmath12@xmath123050 - 5600 and @xmath12@xmath125450 - 10100 , respectively .
a long slit of @xmath16 was placed along the nebular minor axis through the central star to cover the brightest parts of the nebula .
the spatial resolution along the slit was @xmath17/pixel .
the standard pipeline procedures in iraf / stsdas were used to reduce the data .
unfortunately , these data suffered from poor signal - to - noise ratio as well as severe contamination by cosmic rays and turned out to be of very limited value for the current investigation ( see section 4.4 ) . [ modelresults ]
lrrccccccccc line & @xmath12 ( )@xmath18 & @xmath19 & s & e1 & e2 & b@xmath20 & b@xmath21 & b & b@xmath22 & b@xmath23 & b@xmath24 + + h@xmath4 & 4861.3 & 100 & 1.01 & 1.01 & 1.00 & 0.95 & 0.06 & 1.01 & 0.94&0.06&1.00 + h@xmath25 & 6562.8 & 298.08 & 0.97 & 0.97 & 0.97 & 0.90 & 0.08 & 0.98 & 0.90 & 0.08 & 0.98 + h@xmath26 & 4340.5 & 48.92 & 0.95 & 0.95 & 0.95 & 0.89 & 0.06 & 0.95 & 0.89 & 0.06 & 0.95 + he i & 3888.6 & 11.88 & 1.10 & 1.07 & 1.06 & 0.78 & 0.17 & 0.95 & 0.78 & 0.17 & 0.95 + he i & 4471.5 & 6.47 & 1.02 & 1.02 & 1.02 & 0.68 & 0.27 & 0.95 & 0.68 & 0.27 & 0.95 + he i & 5875.7 & 18.85 & 1.01 & 1.02 & 1.02 & 0.67 & 0.35 & 1.02 & 0.68 & 0.34 & 1.02 + he i & 6678.2 & 4.83 & 1.08 & 1.09 & 1.09 & 0.72 & 0.40 & 1.12 & 0.72 & 0.39 & 1.11 + he i & 7065.2 & 4.35 & 1.23 & 1.37 & 1.34 & 0.86 & 0.11 & 0.97 & 0.86 & 0.11 & 0.97 + he i & 7281.6 & 0.56 & 1.89 & 1.91 & 1.91 & 1.32 & 0.23 & 1.55 & 1.32 & 0.23 & 1.55 + he i&10830.3 & 59.96 & 1.59 & 1.79 & 1.74 & 1.21 & 0.10 & 1.31 & 1.21 & 0.10 & 1.31 + he ii & 1640.0 & 82.19 & 1.03 & 0.98 & 0.99 & 0.86 & 0.13 & 0.99 & 0.81 & 0.21 & 1.02 + he ii & 4686.0 & 12.85 & 1.02 & 0.97 & 0.98 & 0.84 & 0.13 & 0.97 & 0.80 & 0.21 & 1.01 + bj / h@xmath4 & 3646 & 0.60 & 0.8 & 0.8 & 0.8 & 0.75 & 0.26 & 1.01 & 0.75&0.26&1.01 + + c ii & 4267.2 & 2.42 & 0.16 & 0.16 & 0.16 & 0.10 & 0.85 & 0.95 & 0.10 & 0.82 & 0.92 + c iii&4650+(m1 ) & 0.50 & 1.11 & 0.76 & 0.76 & 0.41 & 0.27 & 0.68 & 0.40 & 0.39 & 0.79 + c
iii&4187(m18 ) & 0.08 & 1.45 & 1.00 & 1.00 & 0.54 & 1.1 & 1.64 & 0.52 & 1.60 & 2.12 + n ii&4041.3 & 0.24 & 0.19 & 0.21 & 0.21 & 0.15 & 0.89 & 1.04 & 0.16 & 0.84 & 1.00 +
n ii&4241.8 & 0.22 & 0.14 & 0.15 & 0.15 & 0.11 & 0.66 & 0.77 & 0.11 & 0.62 & 0.73 + n
ii&5679.0+(v3 ) & 0.94 & 0.26 & 0.29 & 0.28 & 0.21 & 0.72 & 0.93 & 0.21 & 0.69 & 0.90 + n
iii&4379.0+(m17 ) & 0.66 & 0.60 & 0.45 & 0.46 & 0.30 & 0.94 & 1.24 & 0.30 & 1.35 & 1.65 + o ii&4075.0+(v10 ) & 4.12 & 0.15 & 0.15 & 0.16 & 0.10 & 0.64 & 0.74 & 0.10 & 0.67 & 0.77 + o ii&4089.3 & 0.55 & 0.13 & 0.13 & 0.14 & 0.09 & 0.83 & 0.92 & 0.09 & 0.86 & 0.95 + o ii&4649.1 & 1.39 & 0.23 & 0.23 & 0.24 & 0.16 & 0.89 & 1.05 & 0.16 & 0.92 & 1.08 + o ii&4651.0+(v1 ) & 3.68 & 0.21 & 0.21 & 0.23 & 0.15 & 0.84 & 0.99 & 0.15 & 0.87 & 1.02 + ne ii&3340.0+(v2 ) & 1.96 & 0.09 & 0.09 & 0.09 & 0.08 & 0.56 & 0.64 & 0.08 & 0.61 & 0.69 + ne ii&3700.0+(v1 ) & 1.07 & 0.10 & 0.09 & 0.10 & 0.08 & 0.54 & 0.62 & 0.08 & 0.60 & 0.68 + ne ii&4392.0 & 0.15 & 0.11 & 0.11 & 0.11 & 0.06 & 0.56 & 0.65 & 0.09 & 0.61 & 0.70 + mg ii&4481.0 & 0.05 & & & & 0.66 & 0.26 & 0.92 & 0.67 & 0.27 & 0.93 + + [ c i]&9850.3 & 0.17 & 0.06 & 0.17 & 0.34 & 0.25 & 0.07 & 0.32 & 0.26 & 0.05 & 0.31 + c iii]&1907.0 + & 46.48 & 1.22 & 1.28 & 1.06 & 0.90 & 0.07 & 0.97 & 0.87 & 0.11 & 0.98 + c iv&1549.0 + & @xmath2730.0 & 1.08 & 0.61 & 0.61 & 0.53 & 0.00 & 0.53 & 0.49 & 0.00 & 0.49 + [ n i]&5199.0 + & 0.24 & 0.02 & 0.05 & 1.06 & 1.16 & 0.06 & 1.22 & 1.12 & 0.04 & 1.16 + [ n ii]&5754.6 & 0.83 & 0.30 & 0.61 & 0.75 & 1.03 & 0.21 & 1.24 & 1.08 & 0.20 & 1.28 + [ n ii]&6583.5 + & 64.20 & 0.31 & 0.72 & 0.99 & 1.08 & 0.05 & 1.13 & 1.11 & 0.05 & 1.16 + n iii ] & 1744.0 + & 16.30 & 0.32 & 0.44 & 0.33 & 0.42 & 0.00 & 0.42 & 0.40 & 0.00 & 0.40 + [ o i]&6300.3 & 0.78 & 0.00 & 0.02 & 0.74 & 0.64 & 0.00 & 0.64 & 0.69 & 0.00 & 0.69 + [ o ii]&3726.1 & 18.9 & 0.40 & 0.75 & 0.89 & 0.86 & 0.22 & 1.08 & 0.90 & 0.25 & 1.15 + [ o ii]&3728.8 & 9.69 & 0.47 & 0.80 & 0.91 & 0.94 & 0.20 & 1.14 & 0.97 & 0.22 & 1.18 + [ o iii ] & 4363.2 & 4.19 & 0.90 & 1.05 & 0.93 & 0.92 & 0.00 & 0.92 & 0.88 & 0.01 & 0.89 + [ o iii ] & 5006.8 + & 1189.90 & 1.10 & 1.19 & 1.16 & 0.96 & 0.00 & 0.96 & 0.94 & 0.00 & 0.94 + [ ne iii]&3869.0 & 93.91 & 0.96 & 1.00 & 0.95 & 1.01 & 0.00 & 1.01 & 0.99 & 0.00 & 0.99 + [ s ii]&4068.6 & 1.03 & 0.13 & 0.31 & 0.52 & 0.56 & 0.00 & 0.56 & 0.59 & 0.00 & 0.59 + [ s ii]&4076.4 & 0.35 & 0.13 & 0.31 & 0.52 & 0.55 & 0.00 & 0.55 & 0.58 & 0.00 & 0.58 + [ s ii]&6716.4 & 3.16 & 0.22 & 0.38 & 0.57 & 0.65 & 0.00 & 0.65 & 0.65 & 0.00 & 0.65 + [ s ii]&6730.8 & 5.26 & 0.19 & 0.36 & 0.56 & 0.61 & 0.00 & 0.61 & 0.62 & 0.62 & 0.62 + [ s iii]&6312.1 & 1.29 & 0.97 & 1.33 & 1.16 & 1.39 & 0.00 & 1.39 & 1.38 & 0.00 & 1.38 + [ s iii]&9068.9 & 39.74 & 0.86 & 1.10 & 1.03 & 1.09 & 0.00 & 1.09 & 1.09 & 1.09 & 1.09 + [ s iii]&9531.0 & 113.68 & 0.74 & 0.95 & 0.89 & 0.95 & 0.00 & 0.95 & 0.95 & 0.00 & 0.95 + [ cl iii]&5517.7 & 0.55 & 1.07 & 1.13 & 1.01 & 1.08 & 0.00 & 1.08 & 1.07 & 0.00 & 1.07
+ [ cl iii]&5537.7 & 0.70 & 0.88 & 1.01 & 0.90 & 0.95 & 0.00 & 0.95 & 0.94 & 0.00 & 0.94 + [ ar iii]&5191.8 & 0.10 & 0.83 & 0.84 & 0.77 & 0.94 & 0.00 & 0.94 & 0.91 & 0.00 & 0.91 + [ ar iii]&7135.8 & 19.00 & 1.00 & 0.96 & 0.97 & 1.01 & 0.00 & 1.01 & 1.00 & 0.00 & 1.00 + [ ar iv]&4711.4 & 2.54 & 1.00 & 0.97 & 0.98 & 0.95 & 0.00 & 0.95 & 0.93 & 0.00 & 0.93 + [ ar iv]&4740.2 & 2.35 & 1.00 & 0.99 & 0.99 & 0.96 & 0.00 & 0.96 & 0.93 & 0.00 & 0.93 + + [ n iii]&57.3@xmath3 m & 72.43 & 1.14 & 1.03 & 1.02 & 0.85 & 0.16 & 1.01 & 0.86 & 0.16 & 1.02 + [ o iii]&51.8@xmath3 m & 267.60 & 0.91 & 0.79 & 0.85 & 0.62 & 0.23 & 0.85 & 0.63 & 0.25 & 0.88 + [ o iii]&88.4@xmath3 m & 74.44 & 0.84 & 0.60 & 0.65 & 0.49 & 0.13 & 0.62 & 0.49 & 0.14 & 0.63 + lrrccccccccc line & @xmath12 ( )@xmath18 & @xmath19 & s & e1 & e2 & b@xmath20 & b@xmath21 & b & b@xmath22 & b@xmath23 & b@xmath24 + [ o iv]&25.9@xmath3 m & 118.10 & 1.18 & 1.18 & 1.29 & 1.10 & 0.16 & 1.26 & 1.04 & 0.29 & 1.33 + [ ne ii]&12.8@xmath3 m & 23.14 & 0.07 & 0.11 & 0.15 & 0.16 & 1.00 & 1.16 & 0.15 & 0.68 & 0.83 + [ ne iii]&15.6@xmath3 m & 253.50 & 1.00 & 0.94 & 1.00 & 0.88 & 0.50 & 1.38 & 0.88 & 0.59 & 1.47 + [ ne iii]&36.0@xmath3 m & 32.61 & 0.65 & 0.60 & 0.64 & 0.57 & 0.18 & 0.75 & 0.57 & 0.21 & 0.78 + [ s iii]&18.7@xmath3 m & 47.28 & 1.03 & 1.20 & 1.18 & 1.18 & 0.03 & 1.21 & 1.18 & 0.03 & 1.21 + [ s iii]&33.5@xmath3 m & 32.61 & 0.81 & 0.78 & 0.77 & 0.80 & 0.02 & 0.82 & 0.80 & 0.01 & 0.81 + [ s iv]&10.5@xmath3 m & 362.16 & 1.09 & 0.94 & 0.93 & 0.87 & 0.01 & 0.88 & 0.87 & 0.01 & 0.88 + [ ar iii ] & 9.0@xmath3 m & 56.34 & 0.40 & 0.37 & 0.39 & 0.37 & 0.01 & 0.38 & 0.37 & 0.01 & 0.38 + [ ar iii]&21.8@xmath3 m & 2.69 & 0.53 & 0.48 & 0.51 & 0.49 & 0.00 & 0.49 & 0.49 & 0.00 & 0.49 + @xmath18 : : ` + ' refers to the total intensity of the multiplet .
imaging observations of ngc6153 have been carried out with ground - based telescopes ( pottasch et al .
1986 ) and the _ hst _ ( l2000 ) . fig .
1 shows the _ hst _ images taken under programme go-8594 ( pi liu ) with the wfpc2 camera in the f502n and f656n narrow - band filters on 2000 august 13 .
two 600s and two 500s exposures were made in the f502n and f656n filters , respectively .
the pixel size was @xmath28 .
the images were co - added , cosmic - rays removed and calibrated following the standard recipes ( e.g. holtzmann et al . 1995 ) .
they were then dereddened using an extinction coefficient @xmath29 .
the two filters trace respectively emission of the [ o iii ] @xmath125007 and h@xmath25 lines , although there could be some contamination from the nebular continuum emission and the [ n ii ] @xmath12@xmath126584,6548 lines in the case of f656n .
we performed aperture photometry and found that h@xmath25 and [ o iii ] @xmath125007 line fluxes deduced from _
images agreed within 10 per cent with the corresponding values yielded by the optical scanning spectrum of l2000 , suggesting insignificant contamination to the _ hst _ f656n image from the [ n ii ] lines and the nebular continuum emission .
the azimuthally averaged radial surface brightness distribution of h@xmath25 was deduced from the _ hst _ f656n image and plotted in fig . 2 .
the profile is used in our models to constrain the nebular density distribution .
[ sphfluxdist ] previous empirical and one - dimensional photoionization models of ngc6153 provide valuable insight and a useful guide for setting the initial parameters of the new 3d photoionization models .
l2000 constructed several parametrized empirical models to match the observed fluxes of selected orls and cels .
recombination contributions to cels were considered .
they obtained the following results : 1 ) models of a uniform chemical composition and density distribution can not reproduce the large strengths of orls ; 2 ) chemically homogeneous models with density variations predict too weak orls and the hydrogen balmer discontinuity and a too strong [ ne iii ] fine - structure line compared to observations ; and 3 ) bi - abundance models consisting of two components , each with a different temperature , density , and chemical composition , can account for most of the observed patterns of ngc6153 .
two different bi - abundance models were presented by l2000 .
one assumed some low - density ( @xmath14@xmath30= 700 @xmath15 ) and low - temperature ( @xmath13@xmath30= 500 k ) h - deficient material embedded in the `` normal '' nebular gas of @xmath14@xmath30= 5,500 @xmath15 and @xmath13@xmath30= 9,500 k. in the other model , the hydrogen - depleted material was assumed to be fully ionized and have a high density ( @xmath31 @xmath15 ) and a moderate temperature of 4,700k . in both models ,
the helium and heavy element abundances with respect to hydrogen in the h - deficient component were respectively @xmath32 and@xmath33 times higher than the corresponding values in the diffuse gas .
pquignot et al .
( 2002 , 2003 ) constructed detailed bi - abundance photoionization models of ngc6153 using the one - dimensional code nebu .
their models were composed of sectors extracted from a number of spherically symmetric models , each made of different input parameters except the central ionizing source . in the models , a h - deficient component with a small filling factor was mixed with a normal - composition shell .
they found that models containing a small amount of dense , cold h - deficient material [ @xmath34 = 0.0031@xmath35 , @xmath36 = 4410 @xmath15 , @xmath13@xmath30= 1390 k ] in pressure equilibrium with the surrounding gas of `` normal '' composition [ @xmath34 = 0.38@xmath35 , @xmath37 = 1170 @xmath15 , @xmath13@xmath30= 9040 k ] can well account for the observed fluxes of most orls and cels .
compared to the normal gas , the h - deficient component is enhanced in helium and cnone elements by factors of 6 and 100 , respectively .
nevertheless , the average elemental abundances of the entire nebula are only slightly affected by the presence of the h - deficient component , given its small mass .
pquignot et al .
( 2002 ) pointed out that helium abundance relative to hydrogen will be overestimated if composition fluctuations are not taken into account and it is also essential to consider collisional contribution to excitation of he i lines to correctly determine helium abundance .
the modelling is carried out using mocassin ( version 2.02.12 ; ercolano et al . 2003a , 2005 ) ,
a fully three - dimensional monte carlo photoionization code that solves the radiative transfer self - consistently in an iterative way and is designed to tackle problems involving density variations and chemical inhomogeneities .
the code is therefore suitable for the current investigation aiming to study the effects of the complicated geometry and possible chemical inhomogeneities of ngc6153 . to reduce degree of freedom , the ionizing star is assumed to have a blackbody spectral energy distribution in our models .
a more realistic representation of the ionizing star is left to a later work .
the nebula is approximated by a cubical cartesian grid of cells of individually pre - defined density and chemical composition .
for each cell , the radiation fields ( including the stellar and the diffuse component ) and physical conditions ( electron temperature , density and ionic fractions of all elements considered ) are calculated iteratively . based on an initial guess of the temperature and ionization structures , the code calculates radiation fields , solves the ionization and thermal equilibrium equations and then updates the temperature and ionization structures .
this process is iterated until physical conditions are converged in most cells .
finally , the emission spectrum of the model nebula is calculated by integrating over the volume of the nebula .
the code has been applied to construct photoionization models of the pn ngc3918 , the h - deficient knots in the `` born - again '' pn abell30 , the wolf - rayet pn ngc1501 ( ercolano et al . 2003b , c , 2004 ) and pn ngc6781 ( schwarz & monteiro , 2006 ) . in our models , we have assumed that ngc6153 has an azimuthally symmetrical structure in the x and y plane and a reflection symmetry in z and placed the central star at the centre of a corner cell of a @xmath38 cubic grid .
modelling in this way can effectively save the computational time since only one eighth of the nebula is simulated .
the _ hst _ images of ngc6153 show that the nebula can be approximated by an ellipsoidal shell with enhanced emission in the equator .
the brightest part of the nebula exhibits point symmetry with respect to the central star , probably implying the existence of an ellipsoidal ring lying along the major axis in the south - east to the north - west direction .
a faint halo can also be seen surrounding the bright nebula . in order to better constrain the input parameters ,
we have constructed a series of models with progressively increasing complexity .
these include a spherical model ( model s ) , two ellipsoidal models without and with a torus ( models e1 and e2 , respectively ) , and a bi - abundance model ( model b ) .
the spherical and ellipsoidal models are chemically homogeneous .
table 2 gives the adopted model parameters that yield the best match to the observations . in table 1 , we compare the predicted line fluxes with the observations .
further details of our models are presented below .
.model parameters .
[ cols="<,<,<,<,<,<,<",options="header " , ] @xmath18 : : from l2000 for optical / uv cels .
@xmath39 : : from pottasch et al .
@xmath40 : : from asplund et al .
( 2009 ) .
given that adf values measured for pne are always larger than unity , h - deficient inclusions are presumably a genuine feature of pne .
their presence in pne is however not predicted by current theories of stellar evolution . while it has been proposed ( iben et al .
1983 ) that an evolved star undergoing a very late helium flash ( the so called `` born - again '' pne ) may harbor h - deficient material , the central stars of most pne exhibiting large adf s are found not to be h - deficient .
hydrogen - deficient knots have been clearly detected in the two `` born - again '' pne , abell30 and abell58 .
on the other hand , wesson et al .
( 2003 , 2008 ) found that the h - deficient knots in abell30 and abell 58 are oxygen - rich , in contrast to the expectation of the `` born - again '' scenario .
their chemical composition has more in common with neon novae than with sakurai s object .
the latter is believed to have recently experienced a final he - shell flash .
lau et al .
( in preparation ) explore the possibility of binary evolution to account for the observed high oxygen and neon abundances of the h - deficient knots in abell58 , assuming that the knots are from the ejecta of a neon nova explosion .
they consider a number of scenarios but none fit perfectly with the observations .
their best scenario consists of binary stars where the primary one white dwarf provided the neon nova explosion and the secondary asymptotic giant branch star evolved into the pn .
this scenario requires very fine - tuned initial conditions , particularly the initial separation , to make the neon nova occur just after the final flash of the secondary asymptotic giant branch companion .
therefore , this scenario does not predict a large number of pne with h - deficient knots .
another conjecture is that the h - deficient inclusions are evaporating solid bodies , such as metal - rich planetesimals formed in planetary disks ( liu et al .
this scenario has been subsequently investigated by henney & stasiska ( 2010 ) , who show that the destruction of solid bodies during the short pn phase itself is not able to provide enough metal - rich material to explain the observed cel / orl abundance discrepancies , due to a low sputtering rate . on the other hand
, the sublimation of solid bodies during the final stages of the asymptotic giant branch phase , which is much longer than the pn phase , might provide enough metal - rich material and serve as a possible mechanism causing the abundance discrepancies .
although other explanations to the ( moderate ) adf of hiirs are still possible , the h - deficient inclusions may also be present in h ii regions ( tsamis et al .
2003 ; tsamis & pquignot , 2005 ) .
they could be supernovae ejecta in the form of metal - rich droplets that fall back into the interstellar medium after a long journey lasting about 10@xmath41 yr ( stasiska et al .
they might also originate from evaporating proto - planetary disks around newly formed stars , such as the proplyds discovered in the orion nebula .
we have constructed detailed three - dimensional photoionization models of ngc6153 of spherical or ellipsoidal geometry , with and without h - deficient inclusions .
the model results are compared with spectroscopic observations from the uv , optical to the far ir as well as narrow - band imaging observations . in those models ,
the main nebula is assumed to be chemically homogeneous and of a `` normal '' composition .
the spherical model can reproduce the strengths of hydrogen and helium recombination lines and of high excitation cels . in order to explain the strengths of low - excitation cels , we have considered an ellipsoidal shell where the density decreases from the equator to the poles , supplemented with an equatorial torus having the same chemical composition but of a higher density .
the chemically homogeneous ellipsoidal model with a torus reproduces the strengths of all cels within the measurement uncertainties , but fails by an order of magnitude for heavy element orls . in all cases ,
chemically homogeneous models yield small temperature fluctuations , and are therefore incapable to explain the large discrepancy between @xmath42})$ ] and @xmath43 .
in contrast , a bi - abundance model , incorporating small - scale chemical inhomogeneities in the form of a small amount of metal - rich inclusions , not only reproduces the strengths of heavy element orls and the hydrogen balmer discontinuity , but also improves the overall fitting of cels . in our best - fit bi - abundance model , the cnone abundances in the metal - rich knots are enhanced by 40100 times compared to those of the main nebula .
the presence of those metal - rich knots increases the average metal content of the whole nebula by 30 per cent .
optical / uv cels arise mainly from the main diffuse nebula of `` normal '' composition and thus probe the physical conditions and chemical abundances of this component .
ir cels originate from both components , and thus yield elemental abundances higher than optical / uv cels but lower than orls .
in addition , temperatures deduced from ir cels tend to be lower than values deduced from optical / uv cels .
for he i lines , although a large fraction of their fluxes arises from the h - deficient component , given the relatively low abundance enhancement factor of about 5 , the helium abundance of the normal component remains a fair representation of the average value of the whole nebula .
nevertheless , in the presence of h - deficient inclusions , the traditional analysis assuming a single uniform chemical composition for the whole nebula , will greatly overestimate the helium abundances of the normal component .
one should bear in mind that in the current work we were unable to separate helium emission originating from the two components to obtain reliable estimates of the helium abundance in each of the two components .
indeed , we have assumed he / h @xmath44 for the normal component .
the collisional excitation dominant he i @xmath12 10830 line can , in principle , be used to estimate the helium abundance in the normal component .
its usage is however hampered by observational and theoretical uncertainties .
the electron temperature inside the metal - rich knots is as low as 800k .
the main coolants are ir fine - structure lines from heavy element ions , including the [ o iii ] 52 , 88@xmath3 m , the [ ne iii ] 15.6 , 36@xmath3 m and [ ne ii ] 12.8@xmath3 m lines .
the metal - rich knots have a high density , and are roughly in pressure equilibrium with the surrounding `` normal '' gas .
the metal - rich knots weaken and soften the uv radiation field .
the cometary tails of `` normal '' gas behind the knots have relatively low temperatures and ionization degrees . we show that the presence of metal - rich knots may lead to significant errors in nebular properties derived from the traditional methods assuming a single nebular component .
we discuss physical and chemical properties of the metal - rich knots in ngc6153 as well as their possible origins . to better constrain the nature of metal - rich knots in pne , new observations and techniques are useful .
for example , both imaging and spectroscopic observations of ir fine - structure lines at high spatial resolution and sensitivity will certainly shed new insight into the properties of these metal - rich knots . in future work
, we plan to develop more comprehensive methods to probe these cold knots by involving more observational features ( e.g. he i discontinuities ; zhang et al . 2005a , 2009 ) and up - to - date recombination coefficients of orls for low - temperature plasma ( fang et al .
in preparation ) . we will also construct detailed three - dimensional models for more pne exhibiting large adf s with resolved metal - rich knots , such as m 1 - 42 .
these are the subjects of further papers .
* acknowledgments * * we would like to thank the referee for the valuable comments , which helped improve the quality of the paper . *
the modelling was carried out on the sgi altix330 system at the department of astronomy at peking university , hp supercomputer operated by the center for computational science and engineering at peking university and the columbia supercomputer operated by the nasa advanced supercomputing ( nas ) division at ames research center .
we acknowledge our awards , smd-08 - 0633 and smd-09 - 1154 , of high - end computing ( hec ) time on nas and nccs resources by the nasa science mission directorate ( smd ) .
we thank johnny chang for his able and generous support of our research using columbia .
this research has made use of nasa s astrophysics data system bibliographic services .
be is supported by an stfc advanced fellowship .
yz acknowledges financial support from the seed funding programme for basic research in hku ( 200909159007 ) .
ali b. , blum r. d. , bumgardener t. e. , cranmer s. r. , ferland g. j. , haefner r. l. , tiede g. p. , 1991
, pasp , 103 , 1182 asplund m. , grevesse n. , sauval a.j .
, scott p. , 2009
, ara&a , 47 , 481 * aver e. , olive k. a. , skillman e. d. , 2010 , jcap , 5 , 3 * benjamin r. a. , skillman e. d. , smits d. p. , 1999 ,
apj , 514 , 307 bernard - salas j. , houck j. r. , morris p. w. , sloan g. c. , pottasch s. r. , barry d. j. , 2004 , apj , 154 , 271 blagrave k. p. m. , martin p. g. , rubin r. h. , dufour r. j. , baldwin j. a. , hester j. j. , walter d. k. , 2007 , apj , 655 , 299 cahn j. h. , kaler j. b. , stanghellini l. , 1992 , a&as , 94 , 399 clegg r. e. s. , harrington j. p. , 1989
, mnras , 239 , 869 dudziak g. , pquignot d. , zijlstra a. a. , walsh j. r. , 2000 , a&a , 363 , 717 ercolano b. 2009 , mnras,397 , l69 ercolano b. , barlow m. j. , storey p. j. , liu x .- w . , 2003a , mnras , 340 , 1136 ercolano b. , barlow m. j. , storey p. j. , liu x .- w . , rauch t. , werner k. , 2003c , mnras , 344 , 1145 ercolano b. , barlow m. j. , storey p. j. 2005 , mnras , 362 , 1038 ercolano b. , barlow m. j. , de marco o. , rauch t. , liu x .- w . , 2004 , mnras , 354 , 558 ercolano b. , morisset c. , barlow m. j. , storey p. j. , liu x .- w .
, 2003b , mnras , 340 , 1153 ercolano b. , young p. r. , drake j. j. , raymond j. c. 2008 , apjs , 175 , 534 fang x. , storey p. j. , liu x .- w . , 2010 , in prep ferland g. j. , korista k. t. , verner d. a. , ferguson j. w. , kingdon j. b. , verner e. m. , 1998 , pasp , 110 , 761 garnett d. r. , dinerstein h. l. , 2001 , apj , 558 , 145 henney w. j. , stasiska g. , 2010 , apj , 711 , 881 holtzmann j.a . , burrows c. j. , casertano s. , et al . , 1995 , pasp , 107 , 1065 * howarth i. d. , 1983 , mnras , 203 , 301 * iben i. , jr kaler j. b. , truran j. w. , renzini a. , 1983 , apj , 264 , 605 kingdon j. , ferland g. j. , 1995 , apj , 403 , 211 lau h. b , de marco o. , liu x .- w . , 2010 , in prep liu x .- w .
, 2006 , in iau symp .
234 planetary nebulae , eds .
barlow m. j. , mndez r. h. ( cambridge : cambridge university press ) , p.219 liu x .- w .
, luo s .- g . , barlow m. j. , danziger i. j. , storey p. j. , 2001 , mnras , 327 , 141 liu x .- w . , storey p. j. , barlow m. j. , danziger i. j. , cohen m. , bryce , m. , 2000 , mnras , 312 , 585 luo s .- g . , liu x .- w . , barlow m. j. , 2001 , mnras , 326,1049 nussbaumer h. , storey p. j. , 1984 , a&as , 56 , 293 peimbert m. , 1967 , apj , 150 , 825 peimbert m. , peimbert a. , 2006 , in iau symp .
234 planetary nebulae , eds .
barlow m. j. , mndez r. h. ( cambridge : cambridge university press ) , p.227 peimbert m. , torres - peimbert s. , 1987a , rev .
astr . astrofis .
, 14 , 540 peimbert m. , torres - peimbert s. , 1987b , rev .
astrofis . , 15 , 117 pquignot d. , amara m. , liu x .- w .
, barlow m. j. , storey p. j. , morisset c. , torres - peimbert s. , peimbert m. , 2002 , rmxac , 12 , 142 `` photoionization models for planetary nebulae with inhomogeneous chemical composition '' p ' equignot d. , liu x .- w . , barlow m. j. , storey p. j. , morisset c. , 2003 , in planetary nebulae : their evolution and role in the universe , iau symp .
209 , eds kwok s. , dopita m. , sutherland r. ( asp ) , p. 347
pquignot d. , petitjean p. , boisson c. , 1991 , a&a , 251 , 680 porter r. l. , ferland g. j. , macadam k. b. , 2007 , apj , 657 , 327 * porter r. l. , bauman r. p. , ferland g. j. , macadam k. b. 2005 , apjl , 622 , l73 * pottasch s. r. , beintema d. a. , raimond e. , et al . , 1984 ,
apj , 278 , 33 pottasch s. r. , bernard - salas j. , beintema d. a. , wesselius p. r. , 2003 , apj , 406 , 175 pottasch s. r. , mo j. -e .
, dennefeld m. , 1986 , a&a , 155 , 397 robbins r. r. , 1968 , apj , 151 , 511 rubin r. h. , 1989 , apjs , 69 , 897 schwarz h. e. , monteiro h. , 2006 , apj , 648 , 430 smits d. p. , 1996 ,
mnras , 278 , 683 stasiska g. , szczerba r. , 2001 , a&a , 379 , 1024 stasiska g. , tenorio - tagle g. , rodrguez m. , henney w. j. , 2007 , a&a , 471 , 193 storey p. j. , hummer d. g. , 1995 , mnras , 272 , 41 stoy r. h. , 1933 , mnras , 93 , 588 tsamis y. g. , barlow m. j. , liu x .- w .
, danziger i. j. , storey p. j. , 2003 , mnras , 338 , 687 tsamis y. g. , pquignot d. , 2005 , mnras , 364 , 687 tsamis y. g. , walsh j. r. , pquignot d. , barlow m. j. , danziger i. j. , liu x .- w . , 2008 , mnras , 386 , 22 viegas s. , clegg r. e. s. , 1994 , mnras , 271 , 993 wesson r. , liu x .- w . , barlow m. j. , 2003 , mnras , 340 , 253 wesson r. , barlow m. j. , liu x .- w . , storey p. j. , ercolano b. , de marco o. , 2008 , mnras , 383 , 1639 zanstra h. , 1931 , pub . dominion astrophys .
obs . , 4 , 209 zhang y. , ercolano b. , liu x .- w . , 2007 , a&a , 464 , 631 zhang y. , liu x .- w .
, liu y. , rubin r. h. , 2005a , mnras , 358 , 457 zhang y. , liu x .- w . ,
luo s .- g .
, pquignot d. , barlow m. j. , 2005b , a&a , 442 , 249 zhang y. , liu x .- w . , wesson r. , storey p. j. , liu y. , danziger i. j. , 2004 , mnras,351 , 935 zhang y. , yuan h .- b . , hua c .- t . , liu x .- w . , nakashima j. , kwok s. , 2009 , apj , 695 , 488 | deep spectroscopy of the planetary nebula ( pn ) ngc6153 shows that its heavy element abundances derived from optical recombination lines ( orls ) are ten times higher than those derived from collisionally excited lines ( cels ) , and points to the existence of h - deficient inclusions embedded in the diffuse nebula . in this study , we have constructed chemically homogeneous and bi - abundance three - dimensional photoionization models , using the monte carlo photoionization code mocassin .
we attempt to reproduce the multi - waveband spectroscopic and imaging observations of ngc6153 , and investigate the nature and origin of the postulated h - deficient inclusions , as well as their impacts on the empirical nebular analyses assuming a uniform chemical composition .
our results show that chemically homogeneous models yield small electron temperature fluctuations and fail to reproduce the strengths of orls from c , n , o and ne ions .
in contrast , bi - abundance models incorporating a small amount of metal - rich inclusions ( @xmath0 per cent of the total nebular mass ) are able to match all the observations within the measurement uncertainties .
the metal - rich clumps , cooled down to a very low temperature ( @xmath1 k ) by ionic infrared fine - structure lines , dominate the emission of heavy element orls , but contribute almost nil to the emission of most cels .
we find that the abundances of c , n , o and ne derived empirically from cels , assuming a uniform chemical composition , are about 30 per cent lower than the corresponding average values of the whole nebula , including the contribution from the h - deficient inclusions .
ironically , in the presence of h - deficient inclusions , the traditional standard analysis of the optical helium recombination lines , assuming a chemically homogeneous nebula , overestimates the helium abundance by 40 per cent .
[ firstpage ] ism : abundances planetary nebulae : individual : ngc6153 |
first revealed by the _ einstein _
observatory , the spectroscopic temperature of the intracluster medium ( icm ) in many cd clusters of galaxies decreases toward the center by a factor of @xmath4 ( e.g. , canizares et al . 1988 ; mushotzky & szymkowiak 1988 ; see also fabian 1994 for a review ) .
generally , the appearance of such a central cool component ( hereafter ccc ) has been described in two possible but competing ways .
one is a single - phase ( hereafter 1p ) scenario , in which a mono - nature plasma showing an inward temperature decrease is assumed to permeate the whole cluster ( e.g. , allen et al .
2001 ; vikhlinin et al .
the other is a two - phase ( hereafter 2p ) scenario , in which the cluster s central region is assumed to be occupied by a mixture of two gas components characterized by discrete temperatures ( i.e. , hot and cool ) , of which the relative volume filling factors slowly vary with radius ( e.g. , fukazawa et al .
1994 , 1998 ; takahashi et al . 2009
; see makishima et al .
2001 for a review ) .
although in many cases it is difficult to distinguish between the 1p and 2p models due to insufficient data quality , the implications of these models on the formation and evolution of the ccc are intrinsically different . in the 1p scenario the ccc
is simply interpreted as the dense cluster core , where radiative losses have significantly lowered the gas temperature ( see , e.g. , peterson & fabian 2006 for a review ) , while the 2p scenario favors an interpretation that the ccc is caused by the cd galaxy , rather than being a cooling portion of the icm ( makishima et al .
2001 and references therein ) .
furthermore , the choice between the two scenarios has considerable impact on the accuracies of the measurements of both gas metallicities ( e.g. , the `` fe - bias '' discussed in buote 2000 ) and , in some extreme cases , gravitating mass distributions in clusters .
for the latter one , the systematic mass bias induced by different icm models can reach @xmath5% for the cluster s central region , which is comparable to the typical deviation between the results obtained in x - ray and lensing studies ( e.g. , nagai et al .
2007 ; mahdavi et al . 2008 ) .
hence , it is of fundamental and urgent importance to determine which of the two scenarios is valid , based on the analysis of existing high quality x - ray data of nearby , bright clusters .
so far , only a few works have been done to examine the preference between the 1p and 2p icm models of cd groups and clusters . of these ,
fukazawa et al . ( 1994 ) and ikebe et al .
( 1999 ) used the _ asca _ data to show that a 2p model with temperatures of 1.4 kev and 4.4 kev for the two components is required to describe the ccc of the centaurus cluster .
their results were recently confirmed by the analysis of the high quality data acquired with the european photon imaging camera ( epic ) and reflection grating spectrometer ( rgs ) onboard ( takahashi et al . 2009 ) .
the preference for the 2p scenario was also reported in the works on the fornax cluster ( ikebe et al .
1996 ; matsushita et al . 2007 ) and the ngc 5044 group ( buote & fabian 1998 ; buote et al . 2003 ; tamura et al .
2003 ) with _ asca _ , _ chandra _ , _ xmm - newton _ , and _ suzaku _ observations .
however , due to the lack of such studies on more representative relaxed clusters or groups that exhibit strong cccs , it is not yet clear whether or not the 2p scenario do surpass the 1p scenario in general , and if yes , to what extent the 2p model can be applied to constrain the origin of the ccc . to address these issues , we study in this work the rich , bright cd cluster of galaxies abell 1795 ( a1795 hereafter ) , by jointly analyzing the archival _ chandra _ , _ xmm - newton _ , and data .
a1795 is suited to our purpose as it is a relaxed , nearby ( at a redshift of z @xmath6 ) cluster of galaxies with a luminous x - ray emission ( with a @xmath7 kev luminosity of @xmath8 ergs @xmath9 ; xu et al .
1998 ) , which is found peaked at the giant cd galaxy pgc 049005 ( jones & forman 1984 ; briel & henry 1996 ) . under the 2p assumption , xu et al .
( 1998 ) resolved a cool ( @xmath10 kev ) and a hot ( @xmath11 kev ) component in the central 220@xmath12 kpc region of this cluster with _
asca _ , although ettori et al .
( 2002 ) employed the 1p model instead to fit _
spectra extracted from the same region , and found an inward temperature decrease to @xmath13 kev . in the past 11 years , x - ray observations of this cluster have been accumulated to @xmath14 ks , 140 ks , and 270 ks with _
chandra _ , _ xmm - newton _ , and _ suzaku _ , respectively .
clearly this is an ideal target for utilizing as much of existing high - quality x - ray data as possible to judge which spectral model is more correct .
the layout of this paper is as follows .
section 2 gives a brief description of our data reduction procedure .
the data analysis and results are described in 3 .
we discuss the physical implication of our analysis in 4 , and summarize our work in 5 . throughout the paper
we assume a hubble constant of @xmath15@xmath16 km s@xmath17 mpc@xmath17 , a flat universe with the cosmological parameters of @xmath18 and @xmath19 , and quote errors by the 90% confidence level unless stated otherwise . at the redshift of this cluster
, 1 corresponds to about 72@xmath12 kpc . to compare with previous results ,
we adopt the solar abundance standards of anders & grevesse ( 1989 ) .
in table 1 , we list 11 datasets of a1795 obtained with the advanced ccd imaging spectrometer ( acis ) , which are all used in the analysis . among the 11 observations ,
five acis - s3 exposures focused on the cluster center , and six acis - i exposures covered off - center regions up to @xmath20 mpc . all data were telemetered in vfaint mode , except for the one with obsid 494 in faint mode . using ciao v4.4 and caldb v4.4.7
, we removed bad pixels and columns , as well as events with grades 1 , 5 , and 7 .
we then executed the gain , cti , and astrometry corrections .
the cumulative exposure time was reduced from @xmath21 ks to @xmath22 ks after removing intervals contaminated by occasional background flares with count rate @xmath23@xmath24% of mean value , which were detected by examining @xmath25 kev lightcurves extracted from source free regions near the ccd edges ( e.g. , gu et al .
2009 ) . when available , acis - s1 data were also used to crosscheck the determination of the contaminated intervals .
the obtained clean exposure time for each observation is listed in table 1 . in figure 1_a
_ we show the combined acis image , which has been corrected for exposure but not for background .
was used to observe the central and the south periphery regions of a1795 on 2000 june 26 and 2003 january 13 , respectively , both with the epic operated in the full window mode with thin filter , and with the rgs in the spectroscopy mode ( table 1 ) .
data reduction and calibration were carried out with sas v11.0.1 . in the screening process we set @xmath26 , and kept events with @xmath27 012 for mos cameras and events with @xmath27 04 for pn camera . by examining lightcurves extracted in @xmath28 kev and @xmath29 kev from source free regions , we rejected time intervals affected by hard- and soft - band flares , respectively , in which the count rate exceeds a 2@xmath30 limit above the quiescent mean value ( e.g. , katayama et al . 2004 ; nevalainen et al .
the cleaned mos1 , mos2 , and pn datasets of the central pointing have exposure times of 43.6 ks , 40.3 ks , and 30.2 ks , respectively ( table 1 ) .
the rgs data were screened using the method described in tamura et al .
( 2001 ) .
as listed in table 1 , a1795 was also observed by on 2005 december 10 , with five pointings aimed at its central region , near south and north regions ( offset @xmath31 ) , and far south and north regions ( offset @xmath32 ) . the onboard x - ray imaging spectrometer ( xis ; koyama et al .
2007 ) and the hard x - ray detector ( hxd ; takahashi et al .
2007 ) were both operated in normal modes .
the same xis datasets were already utilized by bautz et al .
( 2009 ) . in analyzing the obtained data , we used the software heasoft 6.11.1 and latest caldb ( 20111109 for the xis and 20110913 for the hxd ) .
we started with version 2.0 processing data , and removed the data obtained either near south atlantic anomaly or at low elevation angles from the earth rim ( @xmath33 and @xmath34 for night and day , respectively ) .
cut - off rigidity criteria of @xmath35 gv for hxd - pin data were also applied . for the xis instrument
, we further examined @xmath25 kev lightcurves of a source free region in each ccd , and filtered off anomalous time bins with count rate above the 2@xmath30 limit of the quiescent mean value .
the combined xis image is shown in figure 1_b_. the data obtained in the far north field are contaminated seriously by solar wind charge exchange emission ( fujimoto et al .
2007 ) , and thus are not included in the following data analysis .
the obtained clean exposure times are listed in table 1 .
we do not consider hxd - gso data in this paper .
we excluded point sources detected beyond a @xmath36 threshold on the acis images with the ciao tool celldetect , and masked the corresponding regions on the epic and xis images after considering the differences in the point spread functions ( psfs ) of the instruments .
the mask regions have radii of @xmath37 and @xmath38 for the epic and xis , respectively .
the point sources outside the acis sky coverage were detected and masked using the epic images .
the ancillary response files ( arfs ) and the redistribution matrix files ( rmfs ) were calculated with the ciao tools mkwarf and mkacisrmf for the acis , and with the sas tools arfgen and rmfgen for the epic .
we calculated the rmfs for the rgs using the sas tool rgsrmfgen , after blurring the line spread function by convolving the raw rmfs with the surface brightness profile ( 3.3.1 ) extracted from the @xmath39 kev acis image along the dispersion direction .
as for the xis , we generated the response files using xissimarfgen ( version 2008 - 04 - 05 ) and xisrmfgen ( version 2007 - 05 - 14 ) .
to enhance the precision of the monte - carlo simulation carried out by xissimarfgen , we created the initial feed photon list by combining all the exposure - corrected and background - subtracted acis - i images .
the method to build the response of hxd - pin is described in 3.4 .
we estimated the background as a combination of three independent components , i.e. , non x - ray background ( nxb ) , cosmic x - ray background ( cxb ) , and galactic emission .
for all the xis data , we created the nxb spectra using a dark earth observation database .
the cxb and galactic emission components in the xis data were estimated by analyzing a spectrum extracted from a region about @xmath40 ( @xmath41@xmath12 mpc at the distance of a1795 ) south of the cluster center , which is covered by the dataset with obsid = 800012050 ( table 1 ) ; a similar background region was used in bautz et al .
( 2009 ) . since this region is outside the virial radius @xmath42@xmath12 mpc , where the brightness of cluster emission is expected to be @xmath43 @xmath44 @xmath45 @xmath46 @xmath47 in @xmath48 kev ( roncarelli et al .
2006 ) , lower than the detection limit of the xis , i.e. , @xmath49 @xmath44 @xmath45 @xmath46 @xmath47 ( bautz et al .
2009 ) , the icm component can be ignored in the background fitting . after subtracting the nxb from the extracted spectrum , we fitted the resulting spectrum with an absorbed power law model ( photon index @xmath50 ) describing the cxb , and two unabsorbed optically thin thermal models ( abundance = 1 @xmath51 , temperatures = 0.08 kev and 0.2 kev ; snowden et al .
1998 ) describing the galactic emission .
the @xmath7 kev cxb flux was estimated as @xmath52 @xmath44 @xmath45 @xmath9 @xmath53 , which is consistent with the results of kushino et al .
the obtained cxb + galactic background template was applied to all the subsequent xis spectral analysis assuming the same uniform cxb + galactic emission distribution over the field of view ( e.g. , sato et al .
2008 ) . in order to achieve an accordant background model for the _ chandra _ acis
, we utilized the above cxb + galactic background template and calculated the acis nxb spectra in an adaptive way as follows .
for all acis - i observations , background spectra were extracted from regions @xmath54 ( @xmath55@xmath12 mpc ) away from the cluster center , where the icm brightness was measured to be weak but can not be neglected , on a level of 10% to 40% of the cxb + galactic components ( bautz et al . 2009 ) .
we subtracted the cxb + galactic components from the extracted spectra , whose brightness was fixed at the xis result obtained above , and fitted the resulting spectra with an empirical nxb model that consists of a broken power law and five narrow gaussian lines ( e.g. , humphrey & buote 2006 ) , plus a thermal apec model for the cluster emission . since the temperature and metal abundance of this icm component can not be well constrained , due to the relatively high nxb level of the acis , we fixed them to previous _
results ( temperature = 2.5 kev , abundance = 0.1 @xmath51 ) reported in bautz et al . ( 2009 ) . in the @xmath56 region ,
average surface brightness of the icm component was obtained as @xmath57 @xmath44 @xmath45 @xmath46 @xmath47 in @xmath48 kev , which is about @xmath58% of the cxb + galactic components , in rough consistent with bautz et al .
the acis - i nxb model was thus determined .
the same procedure was applied to all acis - s data , but source free regions on the s1 chip were used as the background regions instead . to crosscheck above background model with the blank sky templates , we applied both backgrounds to the spectral analysis in 3.1.1 . by fitting the background - subtracted acis spectra extracted from @xmath59 with a single - phase thermal model
, we found that the icm temperatures with the two background models differ by 0.35 kev , which is less than the 1@xmath30 uncertainties of 0.5 kev .
the error caused by background model is much smaller for the inner regions , where the icm emission contributes to @xmath60% of the total counts in @xmath61 kev ( table 2 ) . the background model for the _ xmm - newton _ epic data
was determined with a method similar to that used for the _ chandra _ acis data .
the epic spectra extracted from @xmath62 ( @xmath63@xmath12 mpc ) away from the cluster center , covered by the data with obsid = 0109070201 , were fit with a model consisting of the cxb + galactic components , the epic nxb model including a broken power law and six gaussian lines ( e.g. , gastaldello et al .
2007 ; zhang et al . 2009 ) , and a cluster component approximated by an apec model .
the temperature and abundance of the cluster component were fixed at 2.4 kev and 0.1 @xmath51 ( bautz et al .
2009 ) , respectively , and the @xmath48 kev surface brightness was obtained as @xmath64 @xmath44 @xmath45 @xmath46 @xmath47 ( @xmath65% of the cxb + galactic components ) , in agreement with bautz et al .
( 2009 ) and our _ chandra _ estimate .
the best - fit cxb + nxb + galactic models were used to create background spectra . in figure 2_a
_ , we plot the nxb - subtracted background spectra against the best - fit cxb and galactic components for the acis , epic , and xis , as well as the cluster emission component for the former two instruments . for the rgs data , we employed the blank sky template ( e.g. , takahashi et al .
2009 ) in the following spectral analysis .
errors quoted below in the spectral fittings were estimated by taking into account both statistical and systematic uncertainties .
the former was calculated by scanning over the parameter space with the xspec command _ steppar _ , as the fitting was repeated for a few iterations at each step to ensure that the actual minimum @xmath66 is found . for the latter , we altered the normalization of the cxb spectra by 10% ( kushino et al .
2002 ) to approximate its field - to - field variation , and similarly , re - normalized the nxb components by 2% , 2% , and 1% for the acis , epic , and xis data ( e.g. , hickox & markevitch 2006 ; de luca & molendi 2004 ; tawa et al .
2008 ) , respectively , to assess the error ranges .
we have also tried assigning a systematic error of 2% to approximate the unassessed calibration uncertainties of the epic data ( e.g. , takahashi et al . 2009 ) , which turned out to have negligible impacts on the results and thus was not included in the actual fittings .
we extracted the acis and epic spectra from eight thin annuli , i.e. , @xmath67@xmath12 kpc ( @xmath68 ) , @xmath69@xmath12 kpc ( @xmath70 ) , @xmath71@xmath12 kpc ( @xmath72 ) , @xmath73@xmath12 kpc ( @xmath74 ) , @xmath75@xmath12 kpc ( @xmath76 ) , @xmath77@xmath12 kpc ( @xmath78 ) , @xmath79@xmath12 kpc ( @xmath80 ) and @xmath81@xmath12 kpc ( @xmath82 ) , and the xis spectra from four thick annuli , i.e. , @xmath83@xmath12 kpc ( @xmath84 ) , @xmath85@xmath12 kpc ( @xmath86 ) , @xmath87@xmath12 kpc ( @xmath88 ) and @xmath89@xmath12 kpc ( @xmath90 ) . when fitting these spectra , the lower energy cut was set at 0.7 kev , 0.7 kev , and 0.6 kev for the acis , epic , and xis , respectively , while the upper cut was fixed at 8.0 kev for all the three detectors .
also , the si k edge ( @xmath91 kev ) was excluded from all the xis spectra .
independent fittings were carried out for the _ chandra _ ( acis - s and acis - i ) , _ xmm - newton _ ( mos and pn ) , and ( bi and fi ) spectrum sets , by applying a common absorbed apec model and linking all annuli with xspec model projct , which performs projection of 3-d shells onto 2-d annuli to evaluate the projected emission of outer shells on inner ones . for each shell , the gas temperature , metal abundance , and column density of the neutral absorber were set free .
the best - fit model yielded @xmath92 , 1040/881 , and 845/775 for the acis , epic , and xis spectra , respectively .
figure 2_b _ and 2_c _ show the best - fit deprojected 1p fittings to the spectra extracted from @xmath87@xmath12 kpc and core regions ( @xmath93@xmath12 kpc for the acis and epic , and @xmath83@xmath12 kpc for the xis ) , respectively . as shown in figure 3_a _ , the best - fit deprojected 1p temperature drops inwards from @xmath94 kev at @xmath95@xmath12 kpc to @xmath96 kev in the central 30@xmath12 kpc , which shows apparent diagnostic of a cool core ( e.g. , sanderson et al .
on the other hand , in the cluster s outskirt ( @xmath97@xmath12 kpc ) , the temperature declines outwards steeply down to @xmath96 kev .
generally , our 1p temperature profiles are consistent with previous reports ( e.g. , ettori et al .
2002 ; vikhlinin et al .
2006 ; snowden et al . 2008 ; bautz et al .
incidentally , in @xmath98@xmath12 kpc , the temperature obtained with the acis spectra is by about @xmath99 kev higher than those measured with the epic and xis spectra .
this discrepancy can be ascribed to a local temperature structure ( see 3.2.2 ) , which is smoothed to some extent in the and data .
although the best - fit 1p model can reproduce the three data groups with reasonable goodness , i.e. , @xmath100 , the fitting residuals become drastically significant in the central 80@xmath12 kpc regions , where the averaged gas temperature is measured to be @xmath101 kev ( fig .
3_a _ ) . as shown in figure 2_c
_ , the 1p model significantly underestimates the fe - l blend and the continuum in @xmath23 5.0 kev .
this indicates that these annular spectra exhibit stronger multi - temperature nature than is predicted by foreground and background contributions from different radii that are accounted for by the projct model . to further examine the issues with the 1p icm model for these regions , following , e.g. , cavagnolo et al .
( 2008 ) , we fitted the acis spectra in @xmath102 kev and @xmath103 kev with the 1p model , both extracted from the central 80@xmath12 kpc and corrected for projection effects .
the 1p gas temperature obtained in @xmath102 kev ( @xmath104 kev ) became lower than its counterpart in @xmath103 kev ( @xmath105 kev ) at the 99% confidence level .
the bandpass dependence can not be ascribed to the calibration problem of _ chandra _ at soft band , because a similar dependence can be obtained using the _ xmm - newton _ data , with @xmath106 kev and @xmath107 kev .
thus , the 1p modeling of the icm is considered inadequate , and a 2p icm spectral model is hence invoked .
next we investigate whether or not the 2p icm model , as inferred in the above 1p analysis , is applicable to the spectra extracted in the inner 80@xmath12 kpc region , after the projection effects are taken into account . here , the 2p icm components were both represented by thermal apec models , which were constrained to have the same metallicity and absorption .
both the cool- and hot - phase icm temperatures ( @xmath108 and @xmath109 , respectively ) were tied among all the thin shells in central 80@xmath12 kpc , while the 1p icm model ( 3.1.1 ) was retained for the outer parts .
this fitting yielded overall @xmath110 = 1608/1494 and 929/879 for the acis and epic spectrum sets , respectively .
the fitting goodness was compared to the 1p one ( 3.1.1 ) using an f - test , which yielded @xmath111-statistic of 66.0 and 52.5 for the acis and epic data , respectively , indicating that the 2p model gives a better fit than the 1p model at @xmath23 99% confidence level . as shown in figure 2 _ c _ , the 2p icm model apparently better reproduces the fe - l blend and the continuum in @xmath23 5.0 kev than its 1p counterpart .
this shows that the 2p icm model is strongly and consistently required for the inner 80@xmath12 kpc region even after removing ( via projct ) foreground and background contributions from the outer shells .
the 2p temperatures ( @xmath108 , @xmath109 ) were determined as ( @xmath112 , @xmath113 ) kev and ( @xmath114 , @xmath115 ) kev by using the acis and epic data , respectively . for the regions with @xmath116@xmath12 kpc , the 2p model does not significantly improve the fitting over the 1p model .
the same 2p icm model was also applied to the thick shell spectra obtained with the xis .
compared to the 1p case ( 3.1.1 ; @xmath117 ) , the 2p icm model again gave significantly better fit to the xis spectra within central @xmath118@xmath12 kpc region , with reduced chi - squared improved to @xmath110 = 802/773 . since the psf of the xis is much wider than those of the acis and epic , with the xis alone we can not constrain the boundary between the 1p and 2p regions to a small radius , as we did with the acis and epic data . to examine the psf effect on the spectral fitting of the innermost thick shell ( @xmath83@xmath12 kpc )
, we performed a ray tracing simulation ( e.g. , ishisaki et al .
2007 ; reiprich et al . 2009 ) based on the acis image , and found that the @xmath39 kev counts in this shell are mostly ( @xmath119% ) from the emission originated in @xmath93@xmath12 kpc , only a small fraction ( @xmath120% ) come from @xmath75@xmath12 kpc region where a hotter component ( @xmath121 kev ; figure 3_a _ ) is detected with the acis data .
hence this hot component can not bias the 2p result significantly .
in fact , the best - fit 2p temperatures for this region , @xmath122 kev and @xmath123 kev , are consistent with those obtained in the thin shell analysis .
thus , the psf of the xis little affects the 2p results obtained for the innermost thick shell . as the most consistent form of our 2p analysis
, we simultaneously fitted the thin shell spectra ( the acis and epic ) and those from the thick shell ( the xis ) , by allowing to vary independently the pair of 2p temperatures of each inner shell , i.e. , @xmath124@xmath12 kpc and @xmath125@xmath12 kpc for the acis / epic and xis cases , respectively .
nearly the same fit goodness was achieved ( table 2 ) , and the best - fit @xmath108 and @xmath109 , as shown in figure 4_a _ and table 2 , exhibit nearly insignificant spatial variations across the inner shells .
the values of @xmath108 and @xmath109 are consistent among the three instruments , and broadly consistent with the result , ( @xmath108 , @xmath109 ) = ( @xmath126 , @xmath127 ) kev , reported in xu et al .
( 1998 ) . in short ,
the icm in the cool core of this cluster can be described by two discrete temperatures , @xmath128 kev and @xmath129 kev , which are both consistent with being spatially constant .
this makes the simple 2p icm picture ( makishima et al .
2001 ) a natural and reasonable description . employing the 1p formalism , fabian et al .
( 2001 ) reported a filamentary structure in the inner 50@xmath12 kpc region of a1795 , with a possible temperature of @xmath130 kev , apparently lower than the value of @xmath108 obtained in our 2p analysis .
to look for such a component , we added a third apec component to the 2p icm model describing the xis data from the central 144@xmath12 kpc region .
indeed , we obtained a significantly better fit ( @xmath131 ) by adding a third apec component , whose temperature is @xmath132@xmath133 kev , while its metal abundance , absorption , and redshift were tied to those of the 2p icm components .
an @xmath111-test indicates that the probability of this improvement being caused by chance is @xmath134 .
the @xmath25 kev luminosity of this 0.8 kev component is @xmath135 ergs @xmath9 , which is consistent with the luminosity of the filamentary structure measured with the acis in fabian et al .
( 2001 ; @xmath136 ergs @xmath9 ) .
as shown in table 2 , the values of @xmath108 and @xmath109 , as well as the icm abundances , remain nearly the same by adding this 0.8 kev component , because it contributes only @xmath137% of the total @xmath25 kev luminosity of the central 144@xmath12 kpc region ( @xmath138 ergs @xmath9 ) . in order to crosscheck the xis result
, we analyzed the rgs and deprojected epic spectra extracted from the central 80@xmath12 kpc region by fitting them with three component spectral model , i.e. , 2p icm plus a third apec component .
since the rgs is not so sensitive to hot gas components , we fixed @xmath108 and @xmath109 at the best - fit epic results , i.e. , 2.0 kev and 5.0 kev , respectively . by fitting the rgs1 and rgs2 spectra simultaneously in @xmath139 ,
a weak 0.8 kev component was again detected , with a probability of @xmath140 for the detection to be caused by chance .
the fitting of the epic spectra was improved to @xmath141 by including the 0.8 kev component , significantly better than previous 2p fitting in terms of @xmath111-test ( @xmath142% confidence level ) .
the @xmath25 kev luminosities of this component determined with the rgs and the epic are @xmath143 ergs @xmath9 and @xmath144 ergs @xmath9 , respectively , nicely consistent with the xis value and the acis result in fabian et al .
( 2001 ) . using the high resolution _
chandra _ , _ xmm - newton _ , and _ suzaku _ data
, we have demonstrated that the 2p icm model gives a significantly better description of the icm thermal condition in the central 80@xmath12 kpc of a1795 than the 1p counterpart . the cool and hot phase icm components have @xmath25 kev luminosities of about @xmath145 ergs @xmath46 and @xmath146 ergs @xmath9 , respectively .
also a weak 0.8 kev component has been detected in the same region , which has a @xmath25 kev luminosity of about @xmath147 ergs @xmath9 .
the 2p icm picture implicitly assumes that these phases with different temperatures coexist , each occupying a certain fraction of the total volume under study in the cluster center , which can be described as volume filling factor ( e.g. , ikebe et al .
1999 ; makishima et al . 2001 ; takahashi et al .
2009 ) . to quantify this quantity
, we followed the method described in ikebe et al .
( 1999 ) and calculated the filling factor of the cool phase gas , @xmath148 , where @xmath149 denotes 3-d radius . with @xmath150 ,
the specific emission measures of the two phases , @xmath151 and @xmath152 , are described as @xmath153 where @xmath154 and @xmath155 are the density distributions of the cool and hot phases , respectively . assuming a pressure balance between the two phases @xmath156 we have @xmath157 ^{-1}.\ ] ] figure 4_b _ shows radial profiles of @xmath158 , calculated with the best - fit deprojected acis , epic , and xis model parameters as listed in table 2 .
thus , the acis and epic datasets consistently indicate that the hot component dominates in volume over its cool counterpart not only in outer regions , but also in the core region ( @xmath159@xmath12 kpc ) .
the cool phase gas occupies up to @xmath160 of the volume in the central @xmath161@xmath12 kpc , whereas @xmath150 declines steeply to below 0.05 outside the central @xmath162@xmath12 kpc region . the xis result for
the inner @xmath163@xmath12 kpc is consistent with those of the acis and epic , whereas the relatively large value of @xmath150 indicated by the outer @xmath85@xmath12 kpc xis bin can be attributed to the broad psf of the xis .
in fact , by performing a ray tracing simulation ( see 3.1.2 for more details ) , the outer xis bin at @xmath85@xmath12 kpc was found to be contaminated by photons scattered from the inner regions , in such a way that @xmath164% of the emission in this region actually comes from the central 144@xmath12 kpc . in this case , a weighted mean of @xmath165 measured with the xis over inner @xmath163@xmath12 kpc and @xmath166 measured with the other two missions at the @xmath85@xmath12 kpc region becomes @xmath167 , in agreement with the outer xis point .
hence , after correcting for the psf effect , all data indicate the same 2p configuration in the central region .
the icm in the central @xmath163@xmath12 kpc may alternatively in a multi - phase condition ( e.g. , kaastra et al .
2004 ) , to be described by a more complicated emission measure distribution in a wide temperature range than that used in 3.1.2 . to examine this possibility , we analyzed the deprojected xis spectra from the inner 144@xmath12 kpc region with the same multi - temperature fitting approach as used in tamura et al .
( 2001 ) and takahashi et al .
that is , the spectra were modeled as a cumulative contribution of seven apec components , whose temperatures are given as @xmath168 , 1.5@xmath168 , ( 1.5@xmath169@xmath168 , ... , and ( 1.5@xmath170@xmath168 , where @xmath168 is a base temperature and left free in the fitting .
the seven components were constrained to have the same metal abundance , and suffer the same absorption .
the fit has been acceptable , and the goodness of fitting ( @xmath171 ) is as good as that of the 2p icm plus a 0.8 kev component model ( table 2 ) .
the base temperature was constrained as @xmath172 kev . as shown in figure 5 , in the multi - phase model , only one weak component at @xmath137 kev and two strong components at 2.4 kev and 5.3 kev
remain significant ( 68% confidence level ) , while the rest components can not be constrained .
this is essentially identical to the 2p icm plus 0.8 kev modeling .
our result is consistent with the multi - phase fitting with the epic data in kaastra et al .
( 2004 ) , which also shows a two - temperature structure ( 2.5 kev and 4.9 kev , see their table 6 ) in the cluster center .
hence , the xis data prefers the discrete 2p model , to a continuous temperature distribution , although the latter can not be ruled out based on available data . as shown in figure 3_b _ , the deprojected metal abundance profiles appear roughly consistent between the 1p and 2p icm modelings ( 3.1.1 and 3.1.2 , respectively ) .
both profiles show a peak in the shell of @xmath69@xmath12 kpc , and a mild decline outwards .
this abundance profile , however , should be regarded as an average among those of different elements , because we have so far assumed the solar abundance ratios . to examine the icm for possible deviation from the solar ratios , we employed two vapec models and reran the deprojected 2p fittings of the xis spectra extracted from the central 320@xmath12 kpc region .
specifically , we left the abundances of o , mg , si and fe free , fixed the abundances of he , c and n at the solar value ( anders & grevesse 1989 ) , and tied the abundance of ni to that of fe and those of other elements ( ne , al , s , ar and ca ) to that of si .
this model gave nearly the same set of temperatures for the 2p icm , and a fit goodness ( @xmath173 ) slightly better than that of the best - fit apec model ( @xmath174 ) .
as shown in table 3 , the best - fit fe and si abundances increase significantly towards the center , while the o and mg abundances are nearly constant throughout the cluster .
this confirms the previous acis result reported in ettori et al .
( 2002 ; see also matsushita et al .
we also added another apec model ( @xmath175 = 0.5 @xmath51 ; see table 2 ) in the central 144@xmath12 kpc to account for the 0.8 kev component , while the model gave nearly the same best - fit abundance profiles for the 2p icm components . to examine the possible existence of any intrinsic absorption , we compared the absorption obtained in the spectral analysis with the galactic value .
regardless of the 1p or 2p modeling , the acis and xis fitting results revealed a significant excess absorption by @xmath176 @xmath45 ( table 3 ) beyond the galactic value of @xmath177 @xmath45 , at least in the central 30@xmath12 kpc region .
the excess absorption could not be mitigated by varying the abundances of specific elements ( e.g. , oxygen ) whose emission lines couple with the absorption feature .
this result is in good agreement with the acis result of ettori et al .
( 2002 ; @xmath178 @xmath45 ) . on the contrary ,
our best - fit epic result ( @xmath179 @xmath45 ) , as shown in table 3 , implies lack of excess absorption , in agreement with nevalainen et al .
( 2007 ) who used the same mos data .
the discordance among the three detectors maybe caused by a temporal solar wind charge exchange emission , which biased the epic absorption low .
as shown in wargelin et al .
( 2004 ) , the brightness of charge exchange can reach @xmath180 photon @xmath9 @xmath181 @xmath45 in @xmath182 kev , which is sufficient to undermine the absorption by @xmath183 @xmath45 .
the origin of the central excess absorption detected with the acis and xis , on the other hand , still remains unclear .
as shown above , the azimuthally - averaged spectral analysis prefers a view of the 2p icm in the central 80@xmath12 kpc of a1795 than the 1p counterpart .
however , we can not exclude at present the possibility that the 2p preference is artificially caused by an anisotropic spatial distribution of 1p gas temperature that fluctuates between the range from @xmath108 to @xmath109 . to address this issue
, we examine two - dimensional ( 2-d ) temperature distribution in the central region of a1795 .
below , 2-d icm abundance and @xmath184/@xmath185 distributions are also presented .
combining five acis - s pointings onto the central region together , we have collected sufficient photons for a high resolution 2-d spectral analysis . following the procedure described in detail in gu et al .
( 2009 ) , a set of @xmath186 discrete points ( @xmath187 , @xmath188 ) , or `` knots '' , were chosen in the central @xmath189@xmath12 kpc , which are randomly distributed with a separation of @xmath190@xmath12 kpc between any two adjacent knots . to each knot we assigned a circular region , which is centered on the knot and has an adaptive radius of @xmath191@xmath12 kpc , ensuring that it encloses @xmath192 photons in @xmath39 kev after all detected point sources ( 2.2 )
were excluded .
the spectrum extracted from each circular region was fitted with the 1p and 2p apec models , both subjected to an absorption that was set free . in the 2p spectral fitting , the temperatures of the cool and hot components were fixed at 2.2 kev and 5.7 kev ( i.e. , best - fit acis results ; 3.1.2 ) , respectively , and the two components were assumed to have the same abundance and absorption . for each knot
@xmath187 , we obtained the best - fit 1p gas temperature , 1p/2p abundance , and specific emission measure ratio between the cool and hot components ( @xmath184/@xmath185 ) , as well as their @xmath193 errors .
then , following gu et al .
( 2009 ) , we calculated continuous maps based on the obtained knots . for any position @xmath194 within the map region ,
we defined a scale @xmath195 , so that there are @xmath192 net photons in a circular region centered at @xmath194 , whose radius is @xmath195 .
the 1p temperature at @xmath194 was then calculated by a weighted mean of all the knots @xmath187 in the circular region , @xmath196 where @xmath197 is the distance from @xmath194 to @xmath198 , and @xmath199 is the gaussian kernel whose scale parameter @xmath30 is fixed at @xmath200 .
the use of compact gaussian kernel guarantees an angular resolution of @xmath201@xmath12 kpc within central 100@xmath12 kpc region .
the abundance and @xmath184/@xmath185 maps , along with their @xmath193 error maps , are calculated in the same way .
the resulting 1p temperature and abundance maps are shown in figure 6 , and the 2p abundance ( hereafter @xmath202 ) and @xmath203/@xmath185 maps are presented in figure 7 . the 0.8 kev component ( 3.1.3 ) was not considered in the above 2-d analysis , because it will introduce negligible effects , i.e. , uncertainties of about 0.05 kev and 0.02 @xmath51 , to the temperature and abundance measurements , respectively .
the primary purpose of 2-d spectral analysis is to assess the validity of spherically symmetric temperature distribution that we assumed in the radial 1p/2p analysis . to do this , we examined the obtained 1p temperature map for any significant anisotropic distribution . as shown in figure 6_a _
, the 1p temperature distribution in the cluster central region exhibits an approximate elliptical symmetry ; the temperature variation in the azimuthal direction is about 0.5 kev , 1.0 kev , and 1.0 kev at @xmath204@xmath12 kpc , @xmath205@xmath12 kpc , and @xmath162@xmath12 kpc , respectively , which is insufficient , at least by a factor of three , to imitate the obtained 2p icm condition .
similar morphology is seen in the @xmath184/@xmath185 map of the core region as shown in figure 7_b_. this confirms that the 2p view is not an artifact caused by a 1p icm with large azimuthal asymmetry .
it also indicates that the two phases , in terms of the 2p view , must be separated on scales smaller than the spatial resolution allowed by the present analysis ( @xmath65@xmath12 kpc ) .
the most prominent feature on the 1p temperature map ( fig . 6_a _ ) is a high - temperature arc ( @xmath206 kev according to the 1p model ) located at @xmath207@xmath12 kpc southeast ( se ) of the cluster s center , with an open angle of @xmath23@xmath208 .
given the temperature difference of about 2 kev , the hot structure is significant over the ambient on 90% confidence level .
this feature agrees with a high temperature bin ( @xmath75@xmath12 kpc ) on the deprojected acis temperature profile shown in figure 3_a_. a same high temperature structure was found by markevitch et al .
( 2001 ) , who carried out a 1p azimuthal spectral analysis for the southern part of the cluster .
markevitch et al .
( 2001 ) also reported a density jump by a factor of @xmath209 near the inner edge ( @xmath210@xmath12 kpc from the cluster center ) of the high - temperature arc , and ascribed the density jump to a cold front .
since both the high - temperature component and the cold front locate outsides of the 2p region ( i.e. , @xmath211@xmath12 kpc ) , they are unable to affect the obtained 2p result significantly , even for the xis thick shell as indicated by the ray tracing simulation shown in 3.1.2 . a comparison of the @xmath202 map ( fig .
7_a _ ) with the @xmath184/@xmath185 map ( fig .
7_b _ ) suggests a spatial correlation on a scale of @xmath3@xmath12 kpc between the two quantities .
both maps reveal strongly inhomogeneous distributions outside the relaxed core region , with substructures protruding towards southwest , northeast , and northwest of the cluster center by @xmath212@xmath12 kpc .
most of these substructures show both higher metal abundances , typically by a factor of @xmath4 , and larger cool phase fractions , than their neighborhoods .
the best - fit @xmath202 and @xmath203/@xmath185 values for all the knot - centered circular regions in @xmath3@xmath12 kpc , obtained in our 2-d spectral analysis ( 3.2.1 ) , were compared directly in figure 7_c_. indeed , it reveals a positive linear correlation , which is represented by an analytic form as @xmath213/@xmath214 @xmath51 .
the linear correlation coefficient was obtained as 0.97 . to quantify the suggested correlation , following , e.g. , shibata et al .
( 2001 ) , we calculated a 2-d cross correlation function as , @xmath215 where @xmath216 and @xmath217 are averages of @xmath202 and @xmath218 maps , respectively , @xmath219 is the distance between @xmath220 and @xmath221 , and the bracket is ensemble average . the obtained correlation function is shown in figure 7_d_. a reference profile ( @xmath222 ) was also calculated , based on a series of random maps , @xmath223 and @xmath224 , which were obtained by randomizing the maps within the same data range and smoothing to the same spatial resolution as the original @xmath202 and @xmath225 maps .
the error bars shown in figure 7_d _ were calculated from the variances of @xmath226 when scattering the @xmath202 and @xmath218 maps by their @xmath193 error maps .
comparing to the @xmath222 , the @xmath227 profile shows a significant excess within 100@xmath12 kpc , reconfirming the result that cool - phase icm is indeed more metal - rich .
given the detected correlation , the assumption made in 3.1.2 and 3.1.6 that the two phases icm have approximately the same abundances may not stand now .
therefore , we refitted the deprojected xis spectra of the central @xmath163@xmath12 kpc region with the 2p icm ( vapec + vapec ) + 0.8 kev component ( apec ; 0.5 @xmath51 ) model , same as the one in 3.1.6 , except that the iron abundances of the cool and hot phases were let float separately .
the absorption column densities of the three components were again tied together in the fitting as a free parameter .
we found that the abundance of cool phase icm ( @xmath228 ) does appear higher than that of the hot one ( @xmath229 ) ; i.e. , ( @xmath230 ) = ( @xmath231 , @xmath232 ) @xmath51 .
compared to the model presented in 3.1.6 ( table 3 ) , the current model improves fit goodness to @xmath233 from @xmath234 , with a probability of @xmath235 for the improvement to be caused by chance . following , e.g. , simionescu et al . ( 2008 ) , the iron abundance to be obtained with a 2p model assuming a single common metallicity can be estimated by @xmath236 eliminating @xmath184 and @xmath185 with eq.(1 ) , we have @xmath237 adopting @xmath238 ( fig . 4_b _ ) and ( @xmath230 ) obtained above , @xmath239 is then calculated as 0.51@xmath51 , which agrees with that derived with the xis spectra ( @xmath240 @xmath51 , table 3 ) .
thus , all the acis and xis results obtained so far can be interpreted as evidence for the relatively high metal abundance of the cool phase icm . in many cool - core clusters
, the x - ray surface brightness exhibits an central excess over a @xmath241 model that well fits the outer region .
makishima et al . (
2001 ) argued that the central excess is a combined result of two major effects ; the presence of a cool phase component , and the existence of hierarchical potential structure . according to their definition ,
the hierarchical potential is specified as a halo - in - halo structure , i.e. , a smaller potential component is nested on a larger one , so that the overall potential shows a central deepening relative to a simple king profile .
here we examined the surface brightness profile of a1795 for such a central excess . in order to measure the surface brightness profile precisely by taking advantages of both high spatial resolution of the acis and stable low background of the xis
, we calculated exposure - corrected surface brightness profiles from the inner 1000@xmath12 kpc acis and @xmath242@xmath12 kpc xis data , i.e. , @xmath243 and @xmath244 , respectively , where @xmath245 is projected radius .
the method described in , e.g. , markevitch et al .
( 1998 ) , was employed to compensate discrepancies on calibration and instrumental background between the two instruments .
first , we performed simulations using the xissim tool ( ishisaki et al .
2007 ) to smooth the _
images with the _ suzaku _ psf , and extracted the simulated surface brightness profile in the central 1000@xmath12 kpc , i.e. , @xmath246 . then , @xmath247 was scaled to @xmath244 by solving @xmath248 , where @xmath249 and @xmath250 represent the differences in normalization and nxb , respectively .
a modified @xmath243 with the original acis resolution was obtained by applying @xmath249 and @xmath250 to @xmath243 in the same way as we normalized @xmath246 .
then , the modified @xmath243 and @xmath251 , shown in figure 8 , were fitted jointly with the @xmath241 model , @xmath252 , where @xmath253 is the central brightness , @xmath254 is the core radius , and @xmath255 is the background . to represent the possible halo - in - halo structure ,
a double-@xmath241 model was also tested as @xmath256 , where subscripts 1 and 2 denote compact and extended components , respectively .
as shown in table 4 , the fit with the @xmath241 model can be rejected on the 95% confidence level , because it significantly underestimates the surface brightness in central @xmath257@xmath12 kpc , while the double-@xmath241 model gives an acceptable fit to the data .
the successful fits in figure 8 are based on a still more sophisticated model , a @xmath241+double-@xmath241 gas density model , to be explained later .
the properties of the central excess brightness , represented by the compact @xmath241 component ( @xmath258 , @xmath259@xmath12 kpc ) , are consistent within errors between the @xmath260 kev and @xmath261 kev bands .
such an energy - independent central excess is likely to be caused by an intrinsic hierarchical potential shape , rather than the presence of cool phase component .
this result agrees with the earlier result reported in xu et al .
( 1998 ) . to further quantify the central emission excess
, we corrected the surface brightness profile for the projection effect using the best - fit deprojected 2p gas temperature and abundance profiles ( table 2 ; 3.1.2 ) .
the observed surface brightness profile was modeled as @xmath262 where @xmath263 is the cooling function , @xmath184 and @xmath185 are the same as in eq.(1 ) , and @xmath255 is the averaged background value . following ikebe et al .
( 1999 ) , we represented the specific emission measure profiles by a 2p @xmath241+double-@xmath241 model , which consists of a @xmath241 component for the cool phase icm , @xmath264 ^{-3\beta_{\rm c } } \mbox { } \mbox { } \mbox { } \mbox { } r \leq 80h_{71}^{-1}\mbox { } { \rm kpc } , \\ 0 \mbox { } \mbox { } \mbox { } \mbox { } \mbox { } \mbox { } \mbox { } \mbox { } \mbox { } \mbox { } \mbox { } \mbox { } \mbox { } \mbox { } \mbox { } \mbox { } \mbox { } \mbox { } \mbox { } \mbox { } \mbox { } \mbox { } \mbox { } \mbox { } \mbox { } \mbox { } \mbox { } \mbox { } \mbox { } \mbox { } r > 80h_{71}^{-1}\mbox { } { \rm kpc } , \end{array } \right .\ ] ] and a double-@xmath241 component for the hot phase icm , @xmath265 ^{-3\beta_{\rm h1 } } + q_{\rm 0,h2 } \left [ 1 + \left(\frac{r}{r_{\rm c , h2}}\right)^{2 } \right ] ^{-3\beta_{\rm h2}},\ ] ] where @xmath266 , @xmath267 , and @xmath268 are the model normalizations of the cool phase , the compact hot phase , and the extended hot phase components , respectively .
the filling factor @xmath150 is renormalized into these parameters . with these preparations
, we fitted the @xmath39 kev acis + xis surface brightness profiles using eq.(8 ) , where @xmath263 was calculated from the best - fit 2p spectral parameters .
the fit has been successful with @xmath269 .
the best - fit values of @xmath270 , @xmath271 , @xmath272 , @xmath273 , @xmath274 , and @xmath275 are given in table 4 .
the fittings with surface brightness profiles extracted in different energies ( i.e. , @xmath260 kev and @xmath261 kev bands ) are shown in figure 8 . as presented in table 4 , all the fittings yielded consistent parameters for the 2p @xmath241+double-@xmath241 model . in some clusters ,
the central excess can be alternatively explained by a cuspy dark matter distribution proposed in navarro , frenk , & white ( 1996 ; hereafter nfw model ) .
the nfw model predicts potential structure with a single spatial scale , instead of the dual structure assumed by the @xmath241+double-@xmath241 model .
given the nfw dark matter density profile , @xmath276 where @xmath277 is the critical density , @xmath278 is characteristic density , and @xmath279 is scale radius , the icm density profile is expressed as @xmath280 where @xmath281 is model normalization and @xmath282 is a parameter related to the icm temperature . then the projected icm surface brightness profile was modeled as a single component , @xmath283 we found this nfw model can not give acceptable fits to the @xmath39 kev surface brightness profile for the entire cluster , with a minimum @xmath284 . as shown in figure 8 , the central excess emission is too strong to be explained by the nfw model that best - fits the @xmath285@xmath12 kpc region .
hence , we reconfirmed the _
result reported in xu et al .
( 1998 ) that the halo - in - halo hierarchical model is preferred to the single - halo nfw model in a1795 .
next we calculated the total gravitating mass profile @xmath286 based on the best - fit 2p temperatures ( 3.1.2 ) and the emission measure profiles obtained with the @xmath241+double-@xmath241 model ( 3.3.1 ) . with the help of eqs.(1)-(3 ) , the best - fit emission measure profiles , @xmath287 and @xmath152 , were converted to gas density profiles of the cool and hot phases , @xmath154 and @xmath155 , respectively .
then , under the assumptions of spherical symmetry and hydrostatic equilibrium , the gravitating mass profile @xmath286 was calculated as @xmath288 where @xmath289 = @xmath290 = @xmath291 is the gas pressure , @xmath292 n_{\rm h}(r ) \}$ ] is the averaged gas mass density , @xmath293 is the approximated mean molecular weight , and @xmath294 is the proton mass . as shown in figure 9 , the obtained mass profile exhibits a significant shoulder - like feature at @xmath295@xmath12 kpc , enclosing an excess mass ( @xmath296 ) of about @xmath297 above a flat core mass distribution calculated with the extended hot phase component alone .
our result agrees with the result , @xmath298@xmath12 kpc and @xmath299 , as reported in xu et al .
( 1998 ) . to examine systematic errors of the mass profile due to different modelings of the icm temperature distribution
, we also calculated the gravitating mass profile using the best - fit deprojected 1p spectral parameters ( 3.1.1 ) .
the projected x - ray surface brightness profile was modeled with a single icm component as @xmath300 where the specific emission measure is given by a double-@xmath241 model , @xmath301 ^{-3\beta_{\rm 1 } } + q_{\rm 0,2 } \left [ 1 + \left(\frac{r}{r_{\rm c,2}}\right)^{2 } \right ] ^{-3\beta_{\rm 2}}.\ ] ] the fit is as good as the 2p case , with @xmath302 . the resulting mass profile , derived by substituting @xmath289 = @xmath303 and @xmath304 to eq.(14 ) , is shown in figure 9 by a red curve . thus , the shoulder - like potential structure is again found , although the 1p model systematically underestimates the gravitating mass within 80@xmath12 kpc by @xmath305% .
this bias is negligible in @xmath60 kpc .
another concern is that , the assumption of hydrostatic equilibrium may not hold for the icm near the cold front located southeast of the cluster center ( 3.2.2 ) , casting doubt on the validity of mass profile obtained in the related region ( markevitch et al .
2001 ) . to investigate this
, we excluded the southern half of the cluster and recalculated mass profile for the northern half using the best - fit 2p spectral results and surface brightness profile with the acis data .
the 2p temperatures , @xmath306 kev and @xmath307 kev , are consistent with previous acis results .
the same is the obtained @xmath308 profile , except for the annulus of the cold front ( i.e. , @xmath309@xmath12 kpc ) , where @xmath310 is lower than @xmath155 by @xmath311% in @xmath312@xmath12 kpc . after comparing @xmath313 to @xmath286 in figure 9
, we found that the shoulder - like structure , though diminished by @xmath314% , still remains significant after the high - temperature arc is excluded .
hence we conclude that the existence of central hierarchical structure is unambiguously confirmed in a1795 .
here we examined the @xmath315 kev hxd - pin data for possible existence of any extra emission component .
first , following , e.g. , nakazawa et al .
( 2009 ) , we calculated the effective area of the hxd - pin considering the location and extension of the source .
based on the observed best - fit gas temperature and gas density profiles ( fig .
4_a _ and table 4 ) , we calculated the icm surface brightness distribution map @xmath316 in @xmath315 kev .
then we divided the whole hxd field of view into @xmath317 blocks , calculated the point source arf for each block at @xmath318 using the ftool hxdarfgen , and convolved all the monochromatic arfs with the normalized @xmath316 .
the pin / xis cross normalization factor of 1.132 ( memo 2007 - 11 ) was also adopted in the resulting arf . to assess the nxb level
, we utilized the `` tuned '' nxb model , which provides an optimized reproductivity by making use of hxd - gso information ( memo 2008 - 03 ) .
the cxb model was assumed as @xmath3191.0 kev@xmath320 @[email protected] kev ) photons @xmath45 @xmath9 @xmath323 , where the normalization was set to match the hxd - pin opening angle of 0.32 @xmath324 ( e.g. , nakazawa et al . 2009 ) . as shown in figure 10 ,
the nxb - subtracted xis and hxd - pin spectra in @xmath325 kev were tentatively fitted with a cxb + icm model .
the cxb component was modeled by a cutoff powerlaw with index of 1.29 ( boldt 1987 ) , and the icm component was represented by the 2p model derived in 3.1.2 .
the 2p temperatures were set to ( @xmath108 , @xmath109 ) = ( @xmath326 , @xmath327 ) kev , and their relative normalizations were fixed at the best - fit value in our 2p analysis with the xis . the fitting is acceptable with @xmath328 , and no additional hard x - ray component is required in @xmath329 kev band . to obtain an upper limit on any hard x - ray component , we fitted the combined xis - hxd spectra with a 2p icm plus @xmath330 power - law model ( nakazawa et al .
taking account of the nxb systematic uncertainty of 2.0% at 90% confidence level ( fukazawa et al .
2009 ) , the upper limit flux of power - law emission in @xmath315 kev band was estimated to be @xmath331 ergs @xmath45 @xmath9 . a similar upper limit on any thermal hard x - ray excess , @xmath332 ergs @xmath45 @xmath9 , was obtained by alternatively fitting the combined spectra with 2p icm plus a 10 kev apec model .
we have analyzed _ chandra _ , _ xmm - newton _ , and _ suzaku _ data of the relaxed galaxy cluster a1795 .
the deprojected spectra extracted from the central 80@xmath12 kpc region have been successfully reproduced by invoking two major icm components with discrete temperatures , indicating a 2p property .
a third weak 0.8 kev component is marginally detected in the core region . in
3.2 , we analyzed the _
chandra _ data ,
under 2-d ( spatially ) and 2p ( spectroscopically ) formalism .
the icm metallicity was found to show a similar spatial distribution to that of the cool phase component , indicating that the cool phase icm is more metal - enriched than the hot phase one .
a joint fitting of the thin shell spectra from the acis and epic , and thick shell spectra from the xis , consistently indicates a clear preference of the two - phase icm nature , over the single - phase one , in the central 80@xmath12 kpc region of a1795 .
the icm therein can be characterized by two representative temperatures , @xmath333 kev and @xmath334 kev for the hot and cool phases , respectively , whose @xmath25 kev luminosities were estimated to be @xmath335 ergs @xmath9 and @xmath336 ergs @xmath46 .
both components show insignificant temperature gradients in the 80@xmath12 kpc region .
the hot phase icm dominates in volume over the cool phase one by a factor of @xmath337 , even in the innermost region .
the 2p view is consistent with previous results reported in xu et al .
( 1998 ) . employing the 2p icm model and fitting the cool phase and hot phase components with @xmath241 and double-@xmath241 gas density models , respectively
, we quantified the shape of cluster potential .
as shown in figure 9 , the total gravitating mass profile exhibits a central mass excess ( @xmath338 ) , which has a similar spatial scale to the region of clear 2p property .
this indicates a potential hierarchy , with a 100@xmath12 kpc level halo nested in the center of a cluster - scale one . as shown in figure 3_b
_ , the icm metallicity is also enhanced in the 2p region .
furthermore , by comparing the 2-d metallicity and the emission measure ratio maps created with the 2p icm model , we found a possible spatial correlation on @xmath339 level between the cool - phase fraction ( or @xmath150 ) and the icm metallicity in the @xmath3@xmath12 kpc region .
the 2-d maps also revealed a filamentary distribution of the cool , metal - rich gas in this region .
the subsequent 2p fitting to the xis spectra confirmed the metallicity enhancement in the cool phase relative to the hot phase , giving the best - fit iron abundances of the cool phase and hot phase icm to be 0.80 @xmath51 and 0.36 @xmath51 , respectively .
such a picture resembles the previous findings of the metal - rich , multi - phase arms in m87 ( simionescu et al .
2008 ) . in the innermost core of a1795 ,
i.e. , @xmath340@xmath12 kpc , however , a dip is seen in the abundance distribution , which does not correlate with the cool phase fraction .
such a feature is also seen in the coolest spots of some other clusters , while its origin still remains unclear ( e.g. , sanders et al . 2006 ) . in 3.1.3 and
3.1.5 , we showed that the xis and rgs spectra of the central 144@xmath12 kpc region can be better reproduced by including an additional weak 0.8 kev component , without significantly revising the properties of the cool and hot icm components .
this 0.8 kev component has a @xmath25 kev luminosity of @xmath135 ergs @xmath9 , and is dimmer than the hot icm component by two orders of magnitude . since the temperature and luminosity of this component are consistent with those of an x - ray / h@xmath282 filament in the core of a1795 as reported in fabian et al .
( 2001 ) , the 0.8 kev emission might be associated with a portion of ism of the cd galaxy , which is presumably confined within some filamentary magnetic structures .
the 2p formalism has also been successful for the central region of the centaurus cluster , as firstly reported by fukazawa et al .
( 1994 ) and ikebe et al .
( 1999 ) with _ asca _ data , and recently reinforced by takahashi et al .
( 2009 ) using _ xmm - newton_. according to takahashi et al .
( 2009 ) , the cool phase and hot phase coexist and dominate in the central 70@xmath12 kpc of the centaurus cluster , exhibiting a temperature ratio @xmath341 .
these properties are analogous to those of a1795 . in addition , a similar central excess in the total gravitating mass profile , with a radius of 50@xmath12 kpc , has also been found in the centaurus cluster ( ikebe et al .
thus , the two clusters are very similar in the icm properties , as well as in the total mass distribution . as shown in 3.1.4 and figure 4_b _ , the cool ( @xmath342 kev ) phase component of a1795 occupies up to only 20% volume even in the innermost region , and in figure 6 and 7 , no apparent separation on the spatial distribution is seen between the cool phase and hot phase icm in the cluster core .
this indicates that the cool phase is substantially mixed into the surrounding hot phase , instead of forming any `` cool - phase - only '' core on scale of @xmath343@xmath12 kpc .
since such a 2p structure is seen predominately around a cd galaxy , with an enhanced metallicity ( particularly in the cool phase ) , the phenomenon is very likely related to the cd galaxy .
one possible formation mechanism of the 2p structure could be an agn - driven gas transport . as indicated by numerical simulations , e.g. , churazov et al .
( 2001 ) and guo & mathews ( 2010 ) , the buoyant bubbles created by agn outbursts can drag a certain amount of surrounding cool , metal - rich gas to larger radii .
for example , the icm in central 30@xmath12 kpc regions of the hydra a cluster and m87 are known to be in a 2p or multi - phase form ( ikebe et al .
1997 ; molendi 2002 ) .
more recently , the distributions of cool phase icm were found to coincide with the powerful radio lobes and x - ray cavities , hence , the bubble uplifting model has successfully been invoked for both hydra a cluster and m87 ( nulsen et al .
2002 ; simionescu et al .
these properties may be useful to explain the origin of the co - existing 2p icm .
as for a1795 , by adopting the obtained distributions of total gravitating mass and cool phase gas mass ( @xmath286 and @xmath344 , respectively ; 3.3.2 ) , the energy required to uplift the cool component from cd galaxy center to current position ( average @xmath345@xmath12 kpc ) is estimated as @xmath346 ergs , which is consistent with the amount of mechanical energy injected from its central agn outburst ( rafferty et al .
2006 ) . nevertheless , the faint x - ray cavity detected in a1795 is located @xmath347@xmath12 kpc from the center ( birzan et al . 2004 ) , apparently much smaller than the radius of the cool phase icm as shown in the obtained 2-d temperature and the @xmath184/@xmath185 maps ( fig . 6 & 7 ) .
hence the appearance of the 2p icm can not be solely ascribed to bubble uplifting via recent agn activity .
neither can the cool phase icm be bubble remnant , since the uplifted gas would conductively mix and reach thermal equilibrium with the surrounding hot phase icm quickly , i.e. , within @xmath348 yrs ( e.g. , simionescu et al . 2008 ) , after the bubble is torn apart .
the most natural and effective way to sustain a stable 2p structure is to separate the two icm phases by magnetic fields .
in such a case , the two phases become thermally insulated with each other , since the gyroradius of a thermal electron in a @xmath349 @xmath350 g field is smaller than its mean free path by about @xmath351 orders of magnitude ( sarazin 1988 ) . actually , cluster central regions are often threaded by rather strong ( @xmath5 @xmath350 g ; ge & owen 1993 ) magnetic fields , which are possibly related to the cd galaxy . employing a general topological classification of such cd - galaxy - related magnetic field lines into closed and open ones , makishima et al .
( 2001 ) proposed that the closed loops ( with their both ends anchored to the cd galaxy ) are filled with the cool phase icm , while the open - line regions , connecting to outer regions , are permeated by the hot phase .
thus , the cool phase component is expected to emerge as numerous filamentary substructures in the core region , while the hot phase distributes throughout the vast cluster volume .
in addition , the cool phase is naturally more metal - rich than the hot phase in this scenario , as the supernova yield of the cd galaxy may be largely confined within the loops ( takahashi et al .
2009 ) , which are surrounded by intruding , less contaminated hot phase icm .
this `` cd corona '' model is expected to provide a natural account for the appearance of metal - rich cool phase icm in the center of a1795 . as the cool phase is thermally insulated from the hot phase , and the radiative transport between the two phases
can be ignored ( sarazin 1988 ) , a continuous , efficient heating source located in the central region is required to prevent it from collapsing due to radiative cooling ( see aschwanden et al .
2001 for a review ) . as first pointed out by rosner , tucker & vaiana ( 1978 ; hereafter rtv ) for solar corona , and applied successfully to the centaurus cluster by takahashi et al .
( 2009 ) , a loop structure as illustrated in figure 11 has a built - in feedback mechanism to maintain thermal stability for the cool phase plasma . in order to characterize the loop heating mechanism in a quantitative way
, we employed an updated version of rtv model , an analytical solution to the hydrostatic model introduced in aschwanden & schrijver ( 2002 ; as02 hereafter ) for solar corona and applied it to the 2p icm . this model considers a thin , arch - like loop delineating magnetic field lines , for which the pressure between the loop - interior and surrounding plasmas are in equilibrium . to constrain the temperature and density structures of the loop , as02 assumed mass and momentum conservations along the loop , as well as a total energy balance involving radiative loss rate , heating rate , and thermal conductive flux .
the heating rate function has been defined in form of @xmath352 @xmath353 , where @xmath354 is heating rate at the bottom of loop , @xmath355 is the height in the loop plane , and @xmath356 is the scale height of the heating source ( fig .
the analytic solution by as02 gave a loop temperature distribution increasing with @xmath355 in general , and the exact temperature gradient was predicted to be sensitive to the heating scale @xmath356 . as shown in figure 9 of as02 , for a uniform heating , i.e. , @xmath357 , where @xmath358 is the loop height , the loop temperature exhibits a mild decrease towards the loop bottom by @xmath359% , while for a small scale heating at the bottom , i.e. , @xmath360 , a large part of the loop becomes near isothermal , with only a sharp , small - scale temperature drop seen at the footpoints .
such a loop also obey a scaling law , as defined in eq.(29 ) of as02 ( also see eq.19 below ) , which describes the loop maximum temperature @xmath361 as a function of @xmath356 , @xmath358 , and external pressure at footpoint @xmath362 .
another important property of the loop - confined plasma is that , as shown in eq.(30 ) of as02 , @xmath361 shows rather weak dependence on the heating rate @xmath363 , i.e. , @xmath364 , because a decrease ( or increase ) in @xmath363 will make the loop thinner ( or thicker ) , so as to keep the plasma luminosity nearly equal to @xmath363 .
specifically , following as02 , the temperature distribution of the loop - interior plasma is expected as @xmath365^{a(h / h_{\rm h})}\right\}^{b(h / h_{\rm h})},\ ] ] and the density distribution can be calculated as @xmath366 , % p(h)=p_{0}{\rm exp}\left[-\frac{h}{\lambda_{\rm p}(h , h , h_{\rm h})}\right],\ ] ] where @xmath367 is directly derived from the scaling law ( eq.29 of as02 ) , @xmath368 is the gravitational acceleration calculated from the obtained total mass profile ( 3.3.2 ) , @xmath369 and @xmath370 are the best - fit analytic functions , as defined in eqs.(20 ) and ( 21 ) of as02 , respectively , to reproduce the dependence of @xmath371 on @xmath372 derived in a numerical way , and similarly , @xmath373 is the best - fit empirical function of scaling law factor given in eq.(35 ) of as02 .
we adopted the best - fit coefficients of the hydrostatic solutions presented in table 1 of as02 to characterize @xmath369 , @xmath370 , and @xmath373 , and then , to calculate @xmath361 , @xmath374 , and @xmath375 based on the loop properties . in short ,
given a set of ( @xmath376 ) , we may determine the temperature and density distributions of the loop - interior plasma .
next we applied the as02 model to the x - ray observations of a1795 and , for a comparison , the centaurus cluster , both harboring a prominent cool phase in the central region . in the calculation we determined the external pressures @xmath362 by the temperatures and densities of the hot phases measured at the innermost regions , and took the loop heights @xmath377@xmath12 kpc and 70@xmath12 kpc for a1795 and the centaurus cluster , respectively , as indicated by the radii of cool phases ( 3.1.2 ; takahashi et al .
figure 12_a _ shows the predictions of eq.(19 ) , i.e. , @xmath361 as a function of @xmath356 , in comparison with the measured maximum cool phase temperature , i.e. , @xmath378 kev and @xmath379 kev for a1795 and the centaurus cluster ( fig .
4_a _ ; takahashi et al .
2009 ) , respectively .
this comparison yields @xmath380@xmath12 kpc for a1795 and 18@xmath12 kpc for the centaurus cluster , which indicates that the heating is concentrated at the bottoms of the loops . since the sizes of radio lobes of the central agns in a1795 and the centaurus cluster ( @xmath314@xmath12 kpc and @xmath381@xmath12 kpc , respectively ; ge & owen 1993 ; taylor et al .
2002 ) agree well with @xmath356 estimated above , we speculate that the agn feedback is a candidate heating source for the coronal loops .
the life time of the radio lobes can be estimated by @xmath382 yrs , roughly consistent with the cooling time of the cool phase icm .
furthermore , as shown in rafferty et al .
( 2006 ) , the heating rates provided by the central agn outbursts in a1795 and the centaurus cluster were estimated to be @xmath383 ergs @xmath9 and @xmath384 ergs @xmath385 , respectively , which agree well with their cool phase luminosities ( @xmath386 ergs @xmath9 for a1795 and @xmath387 ergs @xmath9 for the centaurus cluster ) .
having obtained the estimates of @xmath358 and @xmath356 for the two clusters , we then calculated , via eq.(17 ) , the @xmath374 profiles , and show the results in figure 12_b_. thus , a mild inward temperature decrease is expected , especially for the centaurus cluster . actually , takahashi et al .
( 2009 ) showed ( in their fig .
4_a _ ) that the cool phase temperature of the centaurus cluster decreases mildly , from @xmath388 kev ( as already used in fig .
12_a _ ) at @xmath389@xmath12 kpc , to @xmath390 kev = 0.72 @xmath361 at @xmath391@xmath12 kpc .
this latter value , when plotted on figure 12 _ b _ ( with a dashed cross ) , agrees with the prediction .
the same is true in the a1795 case , which shows a more flat @xmath374 profile that is consistent with the measured value of @xmath392 at 15@xmath12 kpc ( solid cross on fig .
12_b _ ) . as shown in figure 12_c _ , we applied the obtained @xmath371 to eq.(18 ) and calculated the normalized loop gas density profile @xmath393 for a1795 , which again agrees well with the observed profile . in summary ,
the cd corona view , combined with the as02 modeling of the loop - confined plasmas , can account for several important properties of inner regions of a1795 and the centaurus cluster .
these properties include the stable 2p structure , the higher metallicity of the cool phase , the absolute values of @xmath108 , and the spatial @xmath108 and density distributions .
by analyzing the _ chandra _ , _ xmm - newton _ , and _ suzaku _ data of the x - ray bright galaxy cluster a1795 , we report clear preference for the 2p icm model in the central 80@xmath12 kpc , which consists of a cool phase ( @xmath394 kev ) and a hot phase ( @xmath395 kev ) component .
this 2p model provides significantly better fit to the deprojected spectra than the 1p model with continuous temperature profile , while the latter can not be fully ruled out based on current data . combining the _ suzaku _ xis and the _ xmm - newton _ epic & rgs , we have marginally detected a third weak 0.8 kev component in the inner 144@xmath12 kpc region that can be ascribed to a portion of ism component of the cd galaxy .
based on a 2-d spectral analysis with the acis data , we have revealed a possible spatial correlation between the cool phase and metal - rich gas in the @xmath3@xmath12 kpc region .
a follow - up xis analysis shows consistent result , that the cool phase icm does exhibit a higher metallicity ( @xmath396 @xmath51 ) than the hot phase one ( @xmath397 @xmath51 ) .
hence , we have successfully resolved a 2 kev , metal - rich component associated with the cd galaxy , which is spatially mixed but thermally separated with the surrounding 5 kev cluster component .
all these properties can be explained by the cd corona model incorporating with the as02 solution for quiescent coronal loops .
this work was supported by the grant - in - aid for scientific research ( s ) , no
. 18104004 , titled `` study of interactions between galaxies and intra - cluster plasmas '' , and by the national science foundation of china ( grant no . 10878001 and 10973010 , and national science fund for distinguished young scholars ) and the ministry of science and technology of china ( grant no .
2009cb824900 and 2009cb824904 ) .
l. g. was supported by the grand - in - aid for jsps fellows , through the jsps postdoctoral fellowship program for foreign researchers .
allen , s. w. , schmidt , r. w. , & fabian , a. c. 2001 , mnras , 328 , l37 anders , e. , & grevesse , n. 1989 , geochim .
acta , 53 , 197 aschwanden , m. j. , poland , a. i. , & rabin , d. m. 2001 , ara&a , 39 , 175 aschwanden , m. j. , & schrijver , c. j. 2002 , apjs , 142 , 269 ( as02 ) bautz , m. w. , miller , e. d. , sanders , j. s. , arnaud , k. a. , mushotzky , r. f. , porter , f. s. , hayashida , k. , henry , j. p. , hughes , j. p. , kawaharada , m. , makishima , k. , sato , m. , & tamura , t. 2009 , pasj , in press [ preprint arxiv:0906.3515 ] birzan , l. , rafferty , d. a. , mcnamara , b. r. , wise , m. w. , & nulsen , p. e. j. 2004 , apj , 607 , 800 boldt , e. 1987 , phys .
, 146 , 215 canizares , c. r. , markert , t. h. , & donahue , m. e. 1988 , in @xmath398 @xmath399 @xmath400 @xmath401 @xmath402 @xmath403 , ed .
a. c. fabian ( dordrecht : kluwer ) , p. 63
cavagnolo , k. w. , donahue , m. , voit , g. m. , & sun , m. 2008 , apj , 682 , 821 churazov , e. , br@xmath404ggen , m. , kaiser , c. r. , b@xmath405hringer , h. , & forman , w. 2001 , apj , 554 , 261 de luca , a. , & molendi , s. 2004 , a&a , 419 , 837 ettori , s. , fabian , a. c. , allen , s. w. , & johnstone , r. m. 2002 , mnras , 331 , 635 fabian , a. c. 1994 , ara&a , 32 , 277 fabian , a. c. , sanders , j. s. , ettori , s. , taylor , g. b. , allen , s. w. , crawford , c. s. , iwasawa , k. , & johnstone , r. m. 2001 , mnras , 321 , l33 fujimoto , r. , mitsuda , k. , mccammon , d. , takei , y. , bauer , m. , ishisaki , y. , porter , s. f. , yamaguchi , h. , hayashida , k. , & yamasaki , n. y. 2007 , pasj , 59 , 133 fukazawa , y. , ohashi , t. , fabian , a. c. , canizares , c. r. , ikebe , y. , makishima , k. , mushotzky , r. f. , & yamashita , k. 1994 , pasj , 46 , l55 fukazawa , y. , makishima , k. , tamura , t. , ezawa , h. , xu , h. , ikebe , y. , kikuchi , k. , & ohashi , t. 1998 , pasj , 50 , 187 fukazawa , y. , et al .
2009 , pasj , 61 , 17 gastaldello , f. , buote , d. a. , humphrey , p. j. , zappacosta , l. , bullock , j. s. , brighenti , f. , & mathews , w. g. 2007 , apj , 669 , 158 ge , j. p. , & owen , f. n. 1993 , aj , 105 , 778 gu , l. , xu , h. , gu , j. , wang , y. , zhang , z. , wang , j. , qin , z. , cui , h. , & wu , x .-
2009 , apj , 700 , 1161 guo , f. , & mathews , w. g. 2010 , apj , 717 , 937 hickox , r. c. , & markevitch , m. 2006 , apj , 645 , 95 humphrey , p. j. , & buote , d. a. 2006 , apj , 639 , 136 koyama , k. , et al .
2007 , pasj , 59 , 23 kushino , a. , ishisaki , y. , morita , u. , yamasaki , n. y. , ishida , m. , ohashi , t. , & ueda , y. 2002 , pasj , 54 , 327 mahdavi , a. , hoekstra , h. , babul , a. , & henry , j. p. 2008
, mnras , 384 , 1567 makishima , k. , ezawa , h. , fukuzawa , y. , honda , h. , ikebe , y. , kamae , t. , kikuchi , k. , matsushita , k. , nakazawa , k. , ohashi , t. , takahashi , t. , tamura , t. , & xu , h. 2001 , pasj , 53 , 401 matsushita , k. , b@xmath405hringer , h. , takahashi , i. , & ikebe , y. 2007 , a&a , 462 , 953 molendi , s. 2002 , apj , 580 , 815 mushotzky , r. f. , & szymkowiak , a. e. 1988 , in @xmath398 @xmath399 @xmath400 @xmath401 @xmath402 @xmath403 , ed .
a. c. fabian ( dordrecht : kluwer ) , p. 53
nagai , d. , vikhlinin , a. , & kravtsov , a. v. 2007 , apj , 655 , 98 nakazawa , k. , sarazin , c. l. , kawaharada , m. , kitaguchi , t. , okuyama , s. , makishima , k. , kawano , n. , fukazawa , y. , inoue , s. , takizawak , m. , wik , d. r. , finoguenov , a. , & clarke , t. e. 2009 , pasj , 61 , 339 navarro , j. f. , frenk , c. s. , & white , s. d. m. 1996 , apj , 462 , 563 nevalainen , j. , markevitch , m. , & lumb , d. 2005 , apj , 629 , 172 nevalainen , j. , bonamente , m. , & kaastra , j. 2007 , apj , 656 , 733 rafferty , d. a. , mcnamara , b. r. , nulsen , p. e. j. , & wise , m. w. 2006 , apj , 652 , 216 reiprich , t. h. , hudson , d. s. , zhang , y .- y . , sato , k. , ishisaki , y. , hoshino , a. , ohashi , t. , ota , n. , & fujita , y. 2009 , a&a , 501 , 899 roncarelli , m. , ettori , s. , dolag , k. , moscardini , l. , borgani , s. , & murante , g. 2006 , mnras , 373 , 1339 rosner , r. , tucker , w. h. , & vaiana , g. s. 1978 , apj , 220 , 643 sanders , j. s. , & fabian , a. c. 2006 , mnras , 370 , 63 sanderson , a. j. r. , ponman , t. j. , & osullivan , e. 2006 , mnras , 372 , 1496 sarazin , c. j. 1988 , @xmath406-@xmath407 @xmath408 @xmath409 @xmath401 @xmath410 @xmath403 ( cambridge university press ) simionescu , a. , werner , n. , finoguenov , a. , b@xmath405hringer , h. , & br@xmath404ggen , m. 2008 , a&a , 482 , 97 shibata , r. , matsushita , k. , yamasaki , n. y. , ohashi , t. , ishida , m. , kikuchi , k. , b@xmath405hringer , h. , & matsumoto , h. 2001 , apj , 549 , 228 snowden , s. l. , egger , r. , finkbeiner , d. , freyberg , m. j. , & plucinsky , p. p. 1998 , apj , 493 , 715 snowden , s. l. , mushotzky , r. f. , kuntz , k. d. , & davis , d. s. 2008 , a&a , 478 , 615 takahashi , i. , kawaharada , m. , makishima , k. , matsushita , k. , fukazawa , y. , ikebe , y. , kitaguchi , t. , kokubun , m. , nakazawa , k. , okuyama , s. , ota , n. , & tamura , t. 2009 , apj , 701 , 377 takahashi , t. , et al .
2007 , pasj , 59 , 35 tamura , t. , kaastra , j. s. , makishima , k. , & takahashi , i. 2003 , a&a , 399 , 497 tawa , n. , hayashida , k. , nagai , m. , nakamoto , h. , tsunemi , h. , yamaguchi , h. , ishisaki , y. , miller , e. d. , mizuno , t. , dotani , t. , ozaki , m. , & katayama , h. 2008 , pasj , 60 , 11 taylor , g. b. , fabian , a. c. , & allen , s. w. 2002 , mnras , 334 , 769 wargelin , b. j. , markevitch , m. , juda , m. , kharchenko , v. , edgar , r. , & dalgarno , a. 2004 , apj , 607 , 596 vikhlinin , a. , kravtsov , a. , forman , w. , jones , c. , markevitch , m. , murray , s. s. , & van speybroeck , l. 2006 , apj , 640 , 691 xu , h. , makishima , k. , fukazawa , y. , ikebe , y. , kikuchi , k. , ohashi , t. , & tamura , t. 1998 , apj , 500 , 738 zhang , y .- y . ,
reiprich , t. h. , finoguenov , a. , hudson , d. s. , & sarazin , c. l. 2009 , apj , 699 , 1178 20/12/1999 & _ chandra _ acis - s & 494 & 13 48 56.5 & + 26 36 26.1 & 19.8/17.5 + 21/03/2000 & _ chandra _ acis - s & 493 & 13 48 49.2 & + 26 36 27.3 & 19.9/18.9 + 10/06/2002 & _ chandra _ acis - s & 3666 & 13 48 48.9 & + 26 34 32.3 & 14.6/13.7 + 14/01/2004 & _ chandra _ acis - s & 5287 & 13 48 55.0 & + 26 36 35.0 & 14.5/13.5 + 18/01/2004 & _ chandra _ acis - i & 5289 & 13 48 55.1 & + 26 36 45.0 & 15.2/14.7 + 23/01/2004 & _ chandra _ acis - i & 5290 & 13 49 00.8 & + 26 42 07.5 & 15.1/14.7 + 20/03/2005 & _ chandra _ acis - i & 6159 & 13 48 32.9 & + 26 40 45.4 & 15.1/14.6 + 20/03/2005 & _ chandra _ acis - s & 6160 & 13 48 49.4 & + 26 36 27.1 & 15.0/14.1 + 28/03/2005 & _ chandra _ acis - i & 6161 & 13 49 19.5 & + 26 31 05.4 & 13.8/13.3 + 28/03/2005 & _ chandra _ acis - i & 6162 & 13 48 48.0 & + 26 36 21.7 & 13.8/13.3 + 31/03/2005 & _ chandra _ acis - i & 6163 & 13 48 47.6 & + 26 36 16.4 & 15.1/14.8 + 26/06/2000 & _ xmm - newton _ epic - pn & 0097820101 & 13 48 52.3 & + 26 35 21.6 & 47.7/30.2 + 26/06/2000 & _ xmm - newton _ epic - mos1 & 0097820101 & 13 48 52.3 & + 26 35 21.6 & 50.1/43.6 + 26/06/2000 & _ xmm - newton _ epic - mos2 & 0097820101 & 13 48 52.3 & + 26 35 21.6 & 50.1/40.3 + 26/06/2000 & _ xmm - newton _ rgs & 0097820101 & 13 48 52.3 & + 26 35 21.6 & 66.6/40.1 + 13/01/2003 & _ xmm - newton _ epic - pn & 0109070201 & 13 48 40.0 & + 26 22 16.4 & 70.3/48.1 + 13/01/2003 & _ xmm - newton _ epic - mos1 & 0109070201 & 13 48 40.0 & + 26 22 16.4 & 64.4/54.1 + 13/01/2003 & _ xmm - newton _ epic - mos2 & 0109070201 & 13 48 40.0 & + 26 22 16.4 & 64.4/54.2 + 10/12/2005 & _ suzaku _ xis & 800012010 & 13 48 53.8 & + 26 36 03.6 & 13.1/13.0 + 10/12/2005 & _ suzaku _ xis & 800012020 & 13 48 53.3 & + 26 47 57.5 & 24.4/24.0 + 11/12/2005 & _ suzaku _ xis & 800012030 & 13 48 53.5 & + 26 59 58.2 & 30.6/@xmath411 + 11/12/2005 & _ suzaku _ xis & 800012040 & 13 48 53.5 & + 26 24 02.5 & 26.1/25.5 + 12/12/2005 & _ suzaku _ xis & 800012050 & 13 48 53.5 & + 26 12 00.4 & 40.1/37.2 + 10/12/2005 & _ suzaku _ hxd & 800012010 & 13 48 53.8 & + 26 36 03.6 & 10.4/10.4 + & & + @xmath67 & 99.8% & @xmath412 & @xmath413 & 1750/1496 & @xmath414 & @xmath415 & @xmath416 & 1608/1494 + @xmath69 & 99.6% & @xmath417 & @xmath418 & & @xmath419 & @xmath420 & @xmath421 & + @xmath71 & 99.0% & @xmath422 & @xmath423 & & @xmath424 & @xmath425 & @xmath426 & + @xmath73 & 97.8% & @xmath427 & @xmath428 & & & & & + @xmath75 & 93.4% & @xmath429 & @xmath430 & & & & & + @xmath77 & 86.7% & @xmath431 & @xmath432 & & & & & + @xmath79 & 82.4% & @xmath433 & @xmath434 & & & & & + @xmath81 & 10.2% & @xmath435 & @xmath436& & & & & + & & + @xmath67 & 99.8% & @xmath437 & @xmath438 & 1040/881 & @xmath439 & @xmath440 & @xmath441 & 929/879 + @xmath69 & 99.7% & @xmath442 & @xmath443 & & @xmath444 & @xmath445 & @xmath446 & + @xmath71 & 99.5% & @xmath447 & @xmath448 & & @xmath449 & @xmath450 & @xmath451 & + @xmath73 & 99.1% & @xmath452 & @xmath453 & & & & & + @xmath75 & 97.3% & @xmath454 & @xmath455 & & & & & + @xmath77 & 91.7% & @xmath456 & @xmath457 & & & & & + @xmath79 & 85.5% & @xmath458 & @xmath459 & & & & & + @xmath81 & 13.8% & @xmath460 & @xmath461 & & & & & + & & + @xmath83 & 99.4% & @xmath462 & @xmath463 & 845/775 & @xmath464 & @xmath465 & @xmath466 & 802/773 + @xmath83 & 99.4% & & & & @xmath467 & @xmath468 & @xmath463 & 791/771 + @xmath85 & 96.4% & @xmath469 & @xmath470 & & [ @xmath464 ] & [ @xmath465 ] & @xmath473 & + @xmath87 & 86.9% & @xmath474 & @xmath475 & & & & & + @xmath89 & 14.3% & @xmath476 & @xmath477 & & & & & + @xmath83 & 0.44@xmath478 & 0.53@xmath479 & 0.54@xmath480 & 0.48@xmath481 & 3.26@xmath482 & 343/328 & 2.17@xmath483 & 1.08@xmath484 + @xmath85 & 0.46@xmath485 & 0.50@xmath486 & 0.37@xmath487 & 0.34@xmath481 & 2.07@xmath488 & 299/284 & 1.20@xmath489 & 1.34@xmath490 + @xmath87 & 0.54@xmath491 & 0.39@xmath492 & 0.10@xmath493 & 0.31@xmath494 & 1.41@xmath495 & 145/149 & 1.21@xmath496 & 1.07@xmath496 + @xmath260 & @xmath497 & @xmath498 & 810/143 & & @xmath499 & @xmath500 & @xmath501 & @xmath502 & 143/140 + @xmath261 & @xmath503 & @xmath504 & 484/121 & & @xmath505 & @xmath506 & @xmath507 & @xmath508 & 135/118 + & & & & + & & & & + energy ( kev ) & & @xmath509&@xmath241&@xmath510 & & @xmath279 & @xmath511 & @xmath510 + @xmath260 & & @xmath512 & @xmath513 & 734/143 & & @xmath514&@xmath515 & 351/143 & + @xmath261 & & @xmath516 & @xmath517 & 364/121 & & @xmath518&@xmath519 & 286/121 & + & & + energy ( kev ) & & @xmath270 & @xmath520&@xmath271 & @xmath521&@xmath272 & @xmath275&@xmath510 & + @xmath260 & & @xmath522 & @xmath523 & @xmath524 & @xmath525 & @xmath526 & @xmath525 & 146/137 & + @xmath261 & & @xmath527 $ ] & @xmath528 $ ] & @xmath529&@xmath530 & @xmath531 & @xmath530 & 143/115 & + @xmath39 & & @xmath532 & @xmath533 & @xmath534&@xmath525 & @xmath535 & @xmath536 & 161/151 & + @xmath12 kpc regions that were used to create the background template ( 2.2 ) , plotted against the best - fit models composed of cxb ( blue ) , galactic ( green ) , and icm ( red ) components .
( b ) nxb - subtracted deprojected acis - i ( upper ) , epic - pn ( middle ) , and xis - fi ( lower ) spectra extracted from @xmath87@xmath12 kpc , along with the corresponding best - fit cxb ( blue ) , galactic ( green ) , and icm ( red ) models .
( c ) nxb - subtracted deprojected acis - s ( upper ) , epic - pn ( middle ) , and xis - fi ( lower ) spectra extracted from the central @xmath162@xmath12 kpc , @xmath162@xmath12 kpc , and @xmath163@xmath12 kpc regions , respectively , fitted with 1p ( red solid ; 3.1.1 ) and 2p ( blue solid ; 3.1.2 ) models .
the prediction of each best - fit 2p component is shown with a blue dash line .
for clarity the cxb and galactic emission models are omitted from the plots . ] , derived by the deprojected analysis of the acis ( red ) , epic ( green ) , and xis ( black ) spectra ( 3.1.4 ) . the result obtained by fitting the acis surface brightness in @xmath39 kev ( 3.3.1 )
is plotted with a solid line .
the acis and epic points on both panels are slightly displaced for clarity . ]
/@xmath185 map of the central region , obtained with the 2p spectral analysis of the acis data without incorporating deprojection ( 3.2.1 ) . in panels ( a ) and ( b ) , the x - ray intensity contours and the 2p abundance contours are superposed , respectively .
the map center and the meaning of the cross is the same as in figure 6 .
( c ) a scatter plot between the values in panel ( b ) and those in panel ( a ) in @xmath3@xmath12 kpc .
error bars are given at the 68% confidence level .
for clarity , the abundances were averaged over 10% intervals in @xmath184/@xmath185 .
solid line shows an analytic fit to the abundance-@xmath184/@xmath185 relation .
( d ) two point correlation function between the obtained 2p abundance map and the @xmath184/@xmath185 map ( black ) . a reference profile ( red )
is given after randomizing the two maps ( see 3.2.3 for more details ) . ]
kev and ( b ) @xmath261 kev , obtained with the acis ( @xmath537@xmath12 kpc ) and xis ( @xmath242@xmath12 kpc ) after the instrumental normalizations are readjusted ( 3.3.1 ) .
the best - fit @xmath241+double@xmath538 models , based on the best - fit deprojected 2p results , are plotted with solid lines . for comparison , the best - fit @xmath241 model and nfw model , based on 1p results that can describe the surface brightness profiles extracted from @xmath23 100@xmath12 kpc regions , are also plotted with dash and dotted lines , respectively . ]
+ double-@xmath241 fitting of the gas density profile ( 3.1.2 and 3.3.1 , respectively ) . for comparison ,
the corresponding mass profiles obtained with the best - fit 1p spectral analysis and double-@xmath241 gas density models ( 3.1.1 and 3.3.2 , respectively ) are also plotted with red solid lines .
error bars are at the 68% confidence level . ]
kpc ( black ) and hxd - pin ( red ) spectrum , compared with the cxb ( green ) and icm ( blue ) components .
the 2p icm components were determined in the @xmath539 kev range , and extrapolated into the hxd - pin range ( see 3.4 ) . the systematic nxb uncertainty of 2.0%
is plotted in cyan . ] as a function of heating scale heights @xmath356 , calculated for a1795 ( solid ) and the centaurus cluster ( dashed ) .
the horizontal lines mark the observed @xmath361 for these clusters .
( b ) predicted @xmath374 profiles of the loops for a1795 ( solid ) and the centaurus cluster ( dashed ) , where @xmath355 is the height in the loop plane .
the crosses show the observed @xmath374 for the innermost shells presented in 3.1.2 and takahashi et al .
( c ) predicted @xmath540 profile of the loops for a1795 ( solid ) . for comparison ,
data points with vertical error bars show the observed @xmath540 profile obtained with the 2p @xmath241+double-@xmath241 fitting of the acis surface brightness distribution ( 3.3.1 ) . ] | based on a detailed analysis of the high - quality _ chandra _ , _ xmm - newton _ , and _ suzaku _ data of the x - ray bright cluster of galaxies abell 1795 , we report clear evidence for a two - phase intracluster medium ( icm ) structure , which consists of a cool ( with a temperature @xmath0 kev ) and a hot ( @xmath1 kev ) component that coexist and dominate the x - ray emission at least in the central 80 kpc . a third weak emission component ( @xmath2 kev ) is also detected within the innermost 144 kpc and is ascribed to a portion of inter - stellar medium ( ism ) of the cd galaxy .
deprojected spectral analysis reveals flat radial temperature distributions for both the hot phase and cool phase components .
these results are consistent with the measurements reported in xu et al .
( 1998 ) , and resemble the previous findings for the centaurus cluster ( e.g. , takahashi et al . 2009 ) . by analyzing the emission measure ratio and gas metal abundance maps created from the _ chandra _ data
, we find that the cool phase component is more metal - enriched than the hot phase one in @xmath3 kpc region , which agrees with that found in m87 ( simionescu et al . 2008 ) .
the coexistence of the cool phase and hot phase icm can not be realized by bubble uplifting from active galactic nuclei ( agn ) alone . instead , the two - phase icm properties are better reconciled with a cd corona model ( makishima et al .
2001 ) . in this model
, the cool phase may be ascribed to the plasmas confined in magnetic loops , which are surrounded by the intruding hot phase icm and have been polluted by metals synthesized in the cd galaxy .
agn feedback energy released in the innermost 10 kpc can serve as the heating source to prevent the loop - interior gas from cooling down to temperatures much lower than the observed value .
the total gravitating mass profile exhibits a hierarchical structure , regardless of the icm temperature modeling . |
n - body codes are one of the most important tools of theoretical cosmology @xcite because they offer the possibility of simulating most of the gravitational processes driving the formation of the large scale structure of the universe ( hereafter lss ) @xcite@xcite@xcite .
these simulations are often used to check cosmological models and to constrain the free parameters of these models which can not be fixed either theoretically or observationally .
+ the typical mass scale for gravitational instability , the jeans mass , has a value of @xmath12 solar masses ( 1 solar mass @xmath13 ) at the recombination epoch , and it gives the approximate size of the first objects forming by gravitational collapse at that epoch . on the other hand , the largest structure we see in our universe today , the `` supercluster '' of galaxies , has a mass in excess of @xmath14 solar masses .
moreover , the gravitational force has a truly long - range character , which makes it impossible to introduce reasonable upper cutoffs in the mass range .
for all these reasons , one would like to be able to perform simulations spanning more than 12 orders of magnitude in mass , but present - day state - of - the - art software and hardware technology does not allow simulations with more than @xmath15 bodies . for these reasons ,
the quest for increasingly efficient algorithms is still in progress .
however , the importance of making n - body simulations is clear to several authors @xcite@xcite@xcite . during the past years n - body codes
have been much improved and applied successfully to various problems in galaxy dynamics , galaxy formation , and cosmological large structure formation .
nevertheless , the computational expense has remained prohibitive for @xmath16 , even using tree - based algorithms on the most powerful computers .
+ the situation is even worse for other n - body algorithms .
the n - body direct evolution method scales as @xmath17 , which makes it impossible to run simulations with more than @xmath18 particles . to overcome this difficulty , and when high accuracy is required , alternative numerical methods based on hierarchical force - computation algorithms are widely used .
the recent effort has addressed the production of new software and algorithms for the new generation of high - performance computer systems .
the ultimate target is an implementation of the tree n - body algorithm to run simulations with higher accuracy and particle number , decreasing the cost of the simulation in terms of cpu time and increasing performance in terms of number of particles / second elaborated when running on mpp systems .
+ among the tree algorithms designed to compute the gravitational force in n - body systems , one of the most used and powerful in modern cosmology is that by barnes and hut ( bh ) @xcite .
the bh octal - tree recursive method is inherently adaptive and allows one to achieve a higher mass resolution even if parallel implementation of this algorithm @xcite@xcite suffers from a serious drawback : it can easily run into imbalance as soon as the configuration evolves , causing performance degradation . in this paper
we present a modified version of the bh algorithm in which we have introduced an enhanced grouping strategy .
we will show how this feature allows an increase in performance when we consider n - body simulation with a large number of particles ( @xmath19 ) .
the code we present incorporates fully periodic boundary conditions using the ewald method , without the use of fast fourier transform techniques @xcite .
+ in section 2 we give a brief description of our n - body parallel code , based on the bh tree algorithm , and the dynamic load balance ( dlb ) policy adopted . in section 3 we describe our enhanced grouping strategy . in section 4
we show the results of our tests and in section 5 we report our conclusions .
since the publication of the monograph by hockney and eastwood in 1981 @xcite , a new class of particle simulation methods @xcite@xcite@xcite@xcite@xcite has emerged as an alternative to particle - particle ( pp ) @xcite@xcite@xcite particle - mesh ( pm ) ( for a review of this method see @xcite ) and particle - particle - particle - mesh ( @xmath20 ) @xcite methods .
these new methods are characterized by the particles being arranged into a hierarchy of clusters , which span the full range of length scales from the minimum interparticle spacing up to the diameter of the entire system .
these methods are usually known as _ tree methods _ or _ tree codes _ because of the data structures used . with these new methods ,
the short - range force on a particle @xmath5 is calculated as a direct sum over nearby particles .
remote bodies are organized into groups which become progressively larger with the distance from the particle ; then a multipole expansion of the potential of each cluster about its center of mass is performed .
the long - range contribution to the acceleration is given by the sum of the particle - cluster interactions .
the bh algorithm works using a hierarchy of cubes arranged in an octal - tree structure ; that is , each node in the tree has eight siblings and each node represents a physical volume of the space .
the total mass of all particles within a given volume and their centers of mass are stored at the corresponding node .
thus , the system is first surrounded by a single cell ( cube ) encompassing all the particles . this main cell ( called _ root
is subdivided into eight subcells of equal volume , each containing its own subset of particles . each subcell in turn is subdivided into eight new subcells and so on .
this procedure is repeated until each cell at the lowest level contains only one particle .
the force on any given @xmath5 is then the sum of the forces by the nearby particles plus the force by the distant cells whose mass distributions are approximated by multipole series truncated typically at the quadrupole order @xcite .
the criterion for determining whether a cell is sufficiently distant for a multipole force evaluation ( that is , for approximating the cell as a multipole ) is based on an opening angle parameter @xmath21 given by @xmath22 where @xmath23 is the size of the cell and @xmath24 is the distance of @xmath5 from the center of mass of the cell .
smaller values of @xmath21 lead to more cell opening and more accurate forces ( for @xmath25 we have an error lower than 1% on the accelerations @xcite ) .
the equations of the dynamics are solved using the leapfrog integrator .
+ in our parallel implementation of the bh tree algorithm , using the pghpf / craft ( an implementation of high performance fortran by the portland group ) programming environment for the cray t3e system , we have exploited both the _ data sharing _ and the _ work sharing _ programming models .
the flexibility of the pghpf / craft environment allows one to mix these two modes in order to gain the maximum efficiency and speed - up .
we can distinguish two main phases in our code structure : the tree_formation ( tf ) and the force_compute ( fc ) .
a data distribution in contiguous blocks !
distribute particle_attribute(block , * ) and alternatively , a fine grain distribution ! hpf$ distribute tree_attribute(cyclic , * ) were adopted to distribute the particle data properties and the tree data properties .
hpf$ distribute directive of the pghpf / craft compiler allows us to consider an array like particle_attribute ( or tree_attribute ) as a unique large array , accessible from all the processors , the array being physically distributed in the local memory of all the processors .
we used two different sets of initial conditions , namely uniform and clustered distributions having 2 million particles each , and they were carried out using from 16 to 128 pes .
our results show that the higher code performances are obtained using a fine grain tree data distribution and a coarse grain bodies data distribution .
a detailed description can be found in @xcite .
the static array distribution , fixed as described above , allows each pe to cooperate during the tf phase by using principally the do independent structure of pghpf that is a synchronous mechanism , and then to execute the fc phase in asynchronous mode . to minimize the communication overhead
, each pe executes the fc phase mainly on the local residing bodies .
the block distribution arranges bodies with the nearest logical number ( near in the space ) in the same pe local memory , or in the nearest pes . using the above mentioned data distribution
, each pe has a block of closed bodies in the local memory ( @xmath26 , where n$pes is the number of processors used for the simulation);in an initial condition with a uniform distribution , the pes having extreme numeration in the pool of available pes have a lower load at each time - step .
the load imbalance is enhanced when a clustered situation occurs during the system evolution .
the pes having bodies in clustered regions have a greater workload since the load of the fc phase increases as the mass density grows .
the technique we follow to perform a load redistribution among the pes is to assign each pe to execute this phase _ only _ for a fixed portion of the bodies residing in the local memory @xmath27 given by + @xmath28 where @xmath29 const .
( @xmath30 ) . the fc phase for all the remaining bodies + @xmath31 is executed by all the pes that have concluded the fc phase for the assigned @xmath27 bodies .
no correlation is considered between the pe memory location of the body belonging to the @xmath32 set and the pe that computes the fc phase on it .
the results imply that it is possible to fix a @xmath33 value that allows the best code performances .
data already presented in @xcite show that it is convenient to fix the @xmath33 value near @xmath34 , which is the value that maximizes the load balance for n - body simulations of the lss both in uniform and clustered situation .
our work and data sharing - parallel tree ( wdsh - pt ) code is principally aimed at running lss cosmological simulations with a number of particles as high as possible using supercomputers such as cray t3e systems . in order to increase the code efficiency ,
we adopt initially the grouping method proposed by barnes @xcite and introduce a modified implementation of his grouping policy yielding very high gains in the code performances with the same accuracy .
to compute the force on a body , the bh algorithm needs to build an interaction list ( @xmath35 ) for each particle @xmath10 .
starting from the root cell , a tree inspection is done and the opening angle parameter @xmath36 is used to evaluate whether a cell must be opened or closed as mentioned above . if a cell has dimension @xmath23 and distance @xmath24 from the particle @xmath10 so that eq .
( [ eq : un ] ) is verified , the cell is closed , it is added to the @xmath35 , and its subcells are not investigated further . otherwise the cell is opened and its subcells are investigated in the same way .
bodies belonging to an opened cell are added to the @xmath35 .
+ next , the force on the body is computed using the monopole and quadrupole momenta for all the cells in the list . + [ [ bh - timing ] ]
bh timing + + + + + + + + + the tree inspection phase represents a sizeable task to compute the force because the cell opening criterion is applied many times for each particle .
the cpu time @xmath37 to compute the force in a time - step for all the @xmath38 particles is @xmath39 where @xmath40 is the average time to build an @xmath35 and @xmath41 is the average time to compute the force on each particle using the interaction list .
+ [ [ b90-timing ] ] b90 timing + + + + + + + + + + the basic idea of b90 was to build a unique interaction list that allows the force for a _ group _ of particles inside a region ; i.e. , a cell @xmath2 of the tree _
( grouping cell ) _ , to be computed reducing the number of tree inspections to build the @xmath42 .
b90 builds an il that applies everywhere within @xmath2 and reuses this il for each particle @xmath43 in turn . in this way it is possible to reduce the tree inspection phase .
the cpu time @xmath44 for b90 may be written as @xmath45 where * @xmath46 is the number of grouping cells ( assuming that each body is inside a group region ) ; * @xmath47 is the average time to build an interaction list for a group ; * @xmath48 is the average time to compute the force on a particle using the list formed for the group . in the following paragraphs
we will compare the @xmath44 time with the @xmath37 time considering the generic case @xmath49 .
we notice that different values of @xmath36 give similar results , as shown by the accompanying figures .
[ [ b90-opening - criterion ] ] b90 opening criterion + + + + + + + + + + + + + + + + + + + + + the original bh algorithm adopts an opening criterion @xmath50 , based on the distance between the position of the @xmath10 particle and the center of mass of the remote cells , the @xmath35 length ( @xmath51 ) being proportional to @xmath52 . in order to have the same accuracy as the original algorithm , the interaction list of the grouping cell is formed using eq .
( [ eq : un ] ) but now the @xmath24 term is computed in terms of the distance from the center of mass of an inspected cell and the edge of the grouping cell , as shown in fig . 1 ( @xmath53 is used instead of @xmath54 ) .
this implies that the @xmath35 formed using the grouping cell contains more elements than the @xmath35 formed by applying the original bh algorithm .
+ [ [ b90-interaction - list - increment ] ] b90 interaction list increment + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the b90 adopts an opening criterion ( @xmath55 ) based on the distance between the edge of @xmath2 and the center of mass of the remote cells . in this case the @xmath56 length will be proportional to @xmath57
. moreover , the b90 criterion uses @xmath55 numerically equal to @xmath50 when using the original bh algorithm .
we found experimentally the relation between @xmath55 and @xmath50 , using 2 million particles in a uniform distribution ( see fig .
2 ) : this relation agrees with data in salmon @xcite .
a typical value used as opening criterion to run simulations for the lss is @xmath58 .
consequently we consider @xmath59 , which , in terms of @xmath35 length increment , corresponds to running a simulation with @xmath60
. + [ [ b90-timing - vs - bh - timing ] ] b90 timing vs bh timing + + + + + + + + + + + + + + + + + + + + + + + figs .
3 and 4 show the relationship @xmath61 and @xmath62 with @xmath36 ranging between @xmath63 and @xmath64 . considering @xmath65 and @xmath66 with @xmath67 , eq .
( [ eq : du ] ) may be rewritten using the above relations as follows : @xmath68 where @xmath69 , and @xmath70 is the average number of particles in a grouping cell .
+ for a large number of systems and in particular for our wdsh - pt code running on the t3e system , @xmath41 ranges from @xmath71 to @xmath72 at @xmath49 . considering these figures ,
we obtain from eqs .
( [ eq : tre ] ) and ( [ eq : tgl1 ] ) , respectively , @xmath73 and @xmath74 and then @xmath75 . +
a real gain could be obtained using b90 if the cpu time spent to form the interaction list were longer than the phase to compute the force on the particle .
the results reported in @xcite demonstrate that if @xmath76 the code performance is between two and three times faster of the bh algorithm , and a good choice for the @xmath77 value is about 32 @xcite .
we will now describe the modification we introduce in the 1999 version of our wdshpt code ( wd99 ) to increase performance even if @xmath78 .
the basic idea is to assign the same @xmath56 to each particle within a cell @xmath2 , containing a maximum of @xmath3 particles .
+ we will not use the b90 criterion to build the interaction list .
instead , we will use the same @xmath50 criterion used in the original bh algorithm .
this criterion is applied to a hypothetical particle placed in the center of mass of the @xmath2 , hereafter vb ( virtual body ) ( fig .
moreover , we consider the @xmath56 as formed by two parts given by @xmath79 @xmath80 and @xmath81 being two subsets of the interaction list . an element is included in one of the two subsets , using the following sphere criterion for all the elements that satisfy eq .
( [ eq : un ] ) .
@xmath82 + + + @xmath83 + @xmath84 + + @xmath81 + + moreover all @xmath43 are included in @xmath81 . using the two subsets it is possible to compute the force @xmath6 on a particle @xmath4 as the sum of two components , @xmath85 where @xmath8 is a force component due to the elements listed in @xmath84 and @xmath9 is the force component due to the elements in @xmath81 .
we assume the component @xmath8 to be the same for each particle @xmath43 and compute it considering the gravitational interaction between the vb and only the elements listed in @xmath84 , while the @xmath9 component is computed separately for each @xmath10 particle by the direct interaction with the elements listed in @xmath81 .
moreover , @xmath9 contains a restricted number of elements in comparison with the @xmath8 list , so we expect a net gain in performance even if @xmath86 .
the gain that is possible depends on several parameters ( @xmath77 , the size of the @xmath2 and the _ sphere radius _ ) , whose ranges of variation are constrained by the maximum allowed value of the overall error of the method , as we will describe in the following sections . before showing the performance of our wd99 procedure in an n - body simulation of the large scale structure of the universe ,
it is important that we perform an error analysis of the procedure itself .
considering that the cumulative error , when simulations for the lss studies are run using the original bh algorithm , is lower than @xmath87 @xcite , fixing the opening criterion @xmath88 , we will give some constraint concerning the size of @xmath2 , the @xmath77 value , and the _ sphere radius _ needed to have negligible cumulative error .
the following sub - sections discuss the two main sources of error .
the first error source is that wd99 uses the sphere criterion and the vb to create an interaction list @xmath89 and wd99 applies the @xmath90 to all bodies @xmath43 .
this approximation could introduce an error in the force value on the @xmath10 particle if the @xmath91 , created using the original bh algorithm , and the @xmath90 have a difference in the elements greater then @xmath87 . as we found with our tests , in order to decrease this difference it is necessary to limit the size of @xmath2 that is equivalent to fixing a _ critical level _ of the tree structure : cells above the critical level can not form a grouping cell .
the user has to fix the critical level considering the density of local bodies in the box where bodies are arranged : the critical level must be chosen in such a way as to make the difference between @xmath91 and @xmath90 negligible ( no more than @xmath87 of the elements ) .
it seems reasonable for a lss simulation in a 50 mpc box with more than 2 million particles to fix the critical level as the sixth level of the tree .
the next section shows in detail the obtained results . in any case
, the cumulative error is very small considering the increase in accuracy compared with the original bh algorithm , due to the inclusion of all @xmath43 in the interaction list .
the second error source is due to the assignment of @xmath8 , computed for the vb , to each @xmath43 .
we found that the sphere criterion allows us to reduce this error to values much lower than 0.01% for @xmath92 if the dimension of the @xmath2 cell is fixed with the critical level as mentioned above , and the _ sphere radius _ is three times the radius of the sphere enclosing the @xmath2 cell ( fig .
+ another important constraint to be fixed is the value of @xmath77 .
all the elements @xmath43 are listed in the @xmath81 list and there is a direct body - body interaction among the @xmath93 ( @xmath94 ) elements forming the group .
this introduces a term @xmath95 in the algorithm complexity . in order to avoid a decrease of the code efficiency and to maintain a good code accuracy , as with the original bh algorithm
, it seems reasonable , running lss simulation with more than 1 million particles , to maintain @xmath96 .
we adopted , in our runs , a safe value @xmath97 even if we obtained good results with @xmath98 .
+ in the next section we show the errors obtained using the above - mentioned constraints when applying wd99 to lss simulation with both uniform and clustered distributions .
we carried out many tests to estimate the error introduced in the wd99 and obtained increased performances using several values of @xmath77 .
therefore , this section is subdivided as follows : first we test whether our algorithm increases the average length of the interaction list , then we measure the resulting percentage error , and we conclude with an overall performance analysis . as a test case we ran a simulation using 2 million particles for lss in a cubic region of 50 mpc , starting from a homogeneous initial condition ( redshift @xmath99 ) and reaching a clustered configuration ( redshift @xmath100 ) .
we used an opening angle parameter @xmath36 ranging from @xmath101 to @xmath64 .
our tests were executed on a cray t3e system and the results will be shown in the following sections . the aim of this first test is to verify that the wd99 algorithm does not introduce a significant computational cost when the force for a generic particle is computed .
this measurement is substantially performed on the average length of the @xmath35 we form adopting our code .
6 reports the result we obtain when the simulation evolves at redshift @xmath99 .
tests were executed for several values of redshift , but the differences between bh and our algorithm was computed only at the end of the run .
the curves were obtained by fixing @xmath98 and varying the critical level from 5 to 8 . in all cases
the differences we obtained are negligible , which means that the computed @xmath35 for the vb ( with the adopted sphere criterion ) is about equal to the @xmath35 we obtain for a generic particle with the original bh algorithm .
the first important result is that wd99 does not produce any increment in the il length and consequently @xmath102 .
we carry out this measurement in two phases .
first we run a single time - step of the 2-million - particle simulation at redshift @xmath99 .
we compare the values we obtain running the bh original algorithm and the wd99 . as a reference case
, we adopt the critical level equal to 6 .
a similar comparison is made at @xmath103 and the bh and wd99 histograms of the forces of each component are compared .
the comparison shows a negligible difference in the force distribution in a single time - step , at least an order of magnitude less than the error of the original bh algorithm .
+ the second measurement is made analysing an entire system evolution .
we start with the initial condition of 2 million particles with redshift @xmath99 , @xmath104 , @xmath67 , and particle mass about @xmath105 solar masses .
the system evolution is carried out up to redshift @xmath100 .
the evolution is executed with the original bh algorithm and with the wd99 code . as reference case
, we adopt the critical level equal to 6 .
we measure the absolute error @xmath106 in the position of particles and in the velocities of particles in the mean square sense , @xmath107 and @xmath108 the study of the final evolution is described in the next sub - section . here
we give the measured value at the end of the simulation : @xmath109 and @xmath110 .
similar values are measured when running simulations with more than 2 million particles and with a critical level equal to 6 .
the obtained @xmath106 values lead us to conclude that the wd99 procedure does not introduce significant errors in comparison with the bh algorithm .
the final stages obtained running simulations with the bh algorithm and the wd99 code are very similar .
the two - point correlation function is defined as @xmath111 @xmath112 being the number of pairs of particles with separations between @xmath113 and @xmath114 , @xmath115 the volume considered , @xmath116 the particle number taken as centres , and @xmath117 the mean particle density .
we calculate this function at redshift @xmath103 for wd99 and bh algorithms .
the values we obtain are perfectly equal , and the substructures we form ( number and size ) are identical . to conclude our wd99 description , we report the performances measured for the wd99 code ( fig .
7 ) ( including the boundary periodic conditions using the ewald method @xcite ) and the performances of the original bh algorithm . the measured performances lead us to the conclusion that when the system evolution is clustered ( @xmath103 ) the wd99 does not decrease the performance as the bh algorithm .
this effect is due to the nature of the wd99 algorithm , which has a structure that increases the efficiency when clusters of particles are well closed .
this important effect allows us to run simulations with very clustered systems , obtaining very good performance and negligible errors . moreover , the efficiency of the wd99 increases by a factor of up to five at the redshift @xmath103 .
the gain is enhanced when bigger simulations are run : a recent simulation with 16 million particles performed on the cray t3e system using wd99 showed an increase in performance by a factor of 7 at the redshift @xmath103 .
we note that the gain obtained , in comparison with that obtained by the original bh algorithm , is greater using a lower critical level ( 5 or 6 ) .
the gain is incremented using a sphere criterion with _ sphere radius _ lower than the value we consider , having only a small increment in the global error .
the code wd99 is mainly used for lss studies , but it could be tested and used for other applications where accuracy not higher than 1% is necessary . considering the high performances we obtained , the wd99 method may be very successfully applied when clustered configurations such as galaxies or clusters of galaxies have to be studied .
the new approach could be applied also to other fields of physics where collisionless systems are to be simulated , as in plasma and hydrodynamic studies .
+ the code is written in fortran 90 with pghpf / craft , but the latest version ( written in f90 and c languages ) uses the one - side communication library shmem , allowing it to run on the origin 2000 systems .
a new version will be implemented using dynamical array allocation , and we are studying the implementation of the parallel out - of - core @xcite , moving data in the disk . this version will be developed for a cc - numa machine with mpi-2 . we plan to have a freely available version of wd99 in october 2000 .
all the tests were carried out using the cray t3e 1200/256 machine at the cineca ( casalecchio di reno ( bo ) , italy ) , a 256 pe system , using the financial support of the italian consortium cnaa ( consorzio nazionale per lastronomia e lastrofisica ) .
we thank dr .
g. erbacci of cineca and dr .
a. f. lanza of catania astrophysical observatory for their useful help .
aarseth , _ astrophys .
space sci . _ * 14 * , 118 ( 1971 ) .
appel , _ siam j. sci .
comput . _ * 6 * , 85 ( 1985 ) .
j. barnes , in _ use of supercomputers in stellar dynamics _ , edited by p. hut and s. mcmillan ( springer - verlag , berlin , 1986 ) , p. 175 .
j. barnes , _ j. comput .
_ * 87 * , 161 ( 1990 ) .
barnes and p. hut , _ nature _ * 324 * , 446 ( 1986 ) .
barnes and p. hut , _ astrophys .
_ * 70 * , 389 ( 1989 ) .
u. becciani , v. antonuccio - delogu and a. pagliaro , _ comput .
commun . _ * 99 * , 1 ( 1996 ) .
u. becciani , r. ansaloni , v. antonuccio - delogu , g. erbacci , m. gambera and a pagliaro , _ comput .
commun . _ * 106 * , 1 ( 1997 ) .
e. bertschinger and j.m .
gelb , _ comput .
_ * 5 * , 164 ( 1991 ) .
birdsall and a.b .
langdon , _ plasma physics via computer simulation _ ( mcgraw - hill international , new york , 1985 ) .
j. dubinski , m.sc .
thesis ( university of toronto , 1988 ) j. dubinski , _ new astronomy _ * 133 * , 1 ( 1996 ) . j.w . eastwood and r.w .
hockney , _
j. comput .
_ * 16 * , 342 ( 1974 ) .
m. hnon , _ ann .
dastrophys . _
* 27 * , 83 ( 1964 ) .
l. hernquist , _ astrophys .
j. suppl . _ * 64 * , 715 ( 1987 ) .
l. hernquist , f.r .
bouchet and y. suto _ astrophys . j. suppl . _ * 75 * , 231 ( 1991 )
. r.w . hockney and j.w.eastwood ,
_ computer simulation using particles _ ( mcgraw - hill international , new york , 1981 ) .
jernigan , in _ iau symposium 113 , dynamics of star cluster _ , edited by j. goodman and p. hut ( reidel , dordrecht , 1985 ) , p. 275 .
a. kravtsov , a.a .
klypin and a.m. khokhlov , _ astrophys .
j. suppl . _
* 111 * , 73 ( 1997 ) . b. kuhlman , a.l .
melott and s.f .
shandarin , _ astrophys .
j. lett . _
* 470 * , l41 ( 1996 ) .
d. porter , ph.d .
thesis , ( university of california , berkeley , 1985 ) .
press , _ in use of supercomputers in stellar dynamics _ , edited by p. hut and s. mcmillan ( springer - verlag , berlin 1986 ) , p. 184 .
j. salmon , ph.d .
thesis ( california institute of technology , 1990 ) . j. salmon and m.s .
warren , in _ proc . of the eight conf . on parallel processing for scientific computing _
( soc . for industr .
& appl . math .
philadelphia ) , 1997 ) .
s. von hoerner , _ z. astrophys . _
* 50 * , 184 ( 1960 ) . | n - body codes for performing simulations of the origin and evolution of the large scale structure of the universe have improved significantly over the past decade in terms of both the resolution achieved and the reduction of the cpu time .
however , state - of - the - art n - body codes hardly allow one to deal with particle numbers larger than a few @xmath0 , even on the largest parallel systems . in order to allow simulations with larger resolution
, we have first reconsidered the grouping strategy as described in j. barnes ( 1990 , _ j. comput .
phys . _ * 87 * , 161 ) ( hereafter b90 ) and applied it with some modifications to our wdsh - pt ( work and data sharing - parallel tree ) code ( u. becciani et al .
, 1996 , _ comput .
phys .
comm . _ * 99*,1 ) . in the first part of this paper we will give a short description of the code adopting the algorithm of j. e. barnes and p. hut ( 1986 , _ nature _ , * 324 * , 446 ) and in particular the memory and work distribution strategy applied to describe the _ data distribution _ on a cc - numa machine like the cray - t3e system . in very large simulations
( typically @xmath1 ) , due to network contention and the formation of clusters of galaxies , an uneven load easily verifies . to remedy this
, we have devised an automatic work redistribution mechanism which provided a good dynamic load balance without adding significant overhead . in the second part of the paper
we describe the modification to the barnes grouping strategy we have devised to improve the performance of the wdsh - pt code .
we will use the property that nearby particles have similar interaction lists .
this idea has been checked in b90 , where an interaction list is built which applies everywhere within a cell @xmath2 containing a small number of particles @xmath3 .
b90 reuses this interaction list for each particle @xmath4 in the cell in turn .
we will assume each particle @xmath5 to have the same interaction list .
we consider that the agent force @xmath6 on a particle @xmath5 can be decomposed into two terms @xmath7 . the first term @xmath8 is the same for each particle in the cell and is generated by the interaction between a hypothetical particle placed in the center of mass of the @xmath2 and the farther cells contained in the interaction list .
@xmath9 is different for each particle @xmath10 and is generated by the interaction between @xmath5 and the elements near @xmath2 .
thus it has been possible to reduce the cpu time and increase the code performance .
this enable us to run simulations with a large number of particles ( @xmath11 ) in nonprohibitive cpu times .
0.5 cm -2 cm |
a recent new scientist article @xcite deals with errors in courts due to `` bad mathematics '' , advocating the use of the so - called bayesian methods to avoid them .
although most examples of resulting `` rough justice '' come from real life cases , the first `` probabilistic pitfall '' is taken from crime fiction , namely from a `` 1974 episode of the cult us television series '' _ columbo _ , in which a `` society photographer has killed his wife and disguised it as a bungled kidnapping . ''
the pretended mistake happens in the concluding scene , when `` the hangdog detective [ ] induces the murderer to grab from a shelf of 12 cameras the exact one used to snap the victim before she was killed . ''
according to the article author ( or to experts on which scientific journalists often rely on ) the question is that `` killer or not , anyone would have a 1 in 12 chance of picking the same camera at random .
that kind of evidence would never stand up in court . ''
then a sad doubt is raised , `` or would it ?
in fact , such probabilistic pitfalls are not limited to crime fiction . '' being myself not particularly fond of this kind of entertainment ( perhaps with a little exception of the columbo series , that i watch casually ) , i can not tell how much crime fiction literature and movies are affected by `` probabilistic pitfalls '' .
instead , i can give firm witness that scientific practice is plenty of mistakes of the kind reported in ref.@xcite , that happen even in fields the general public would hardly suspect , like frontier physics , whose protagonists are supposed to have a skill in mathematics superior to police officers and lawyers .
but it is not just a question of math skill ( complex calculations are usually done without mistakes ) , but of _ probabilistic reasoning _
( what to calculate ! ) .
this is a quite old story .
in fact , as david hume complained 260 years ago @xcite , _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ `` the celebrated monsieur leibniz has observed it to be a defect in the common systems of logic , that they are very copious when they explain the operations of the understanding in the forming of demonstrations , but are too concise when they treat of probabilities , and those other measures of evidence on which life and action entirely depend , and which are our guides even in most of our philosophical speculations . ''
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ it seems to me that the general situation has not improved much .
yes , ` statistics ' ( a name that , meaning too much , risks to mean little ) is taught in colleges and universities to students of several fields , but distorted by the ` frequentistic approach ' , according to which one is not allowed to speak of probabilities of causes .
this is , in my opinion , the original sin that gives grounds for a large number of probabilistic mistakes even by otherwise very valuable scientists and practitioners ( see e.g. chapter 1 of ref .
@xcite ) .
going back to the `` shambling sleuth columbo '' , being my wife and my daughter his fans , it happens we own the dvd collections of the first seven seasons .
it occurred then i watched with them , not much time ago ( perhaps last winter ) , the ` incriminated ' , superb episode _ negative reaction _
@xcite , one of the best performances of peter falk playing the role of the famous lieutenant .
however , reading the mentioned new scientist article , i did not remember i had a ` negative reaction ' from the final scene , although i use and teach bayesian methods for a large variety of applications .
did i overlook something ?
i watched again the episode and i was again convinced columbo s last move was a conclusive checkmate . ) .
therefore , rather than chess , the name of the game is _ poker _ , and columbo s bluff is able to induce the murderer to provide a crucial piece of evidence to finally incriminate him .
] then i have invited some friends , all with physics or mathematics degree and somewhat knowledgeable of the bayesian approach , to enjoy an evening together during the recent end of year holidays in order to let them make up their minds whether columbo had good reasons to take paul galesco , magnificently impersonated by dick van dyke , in front of the court ( bayes or not , we had some fun ) .
the verdict was unanimous : columbo was fully absolved or , more precisely , there was nothing to reproach the story writer , peter s. fischer .
the convivial after dinner jury also requested me to write a note on the question , possibly with a short , self - contained introduction to the ` required math ' .
not only to ` defend columbo ' or , more properly , his writer , but , and more seriously , to defend the bayesian approach , and in particular its applications in forensic science .
in fact , we all deemed the beginning paragraphs of the new scientist article could throw a bad light on the rest of the contents .
imagine a casual reader of the article , possibly a lawyer , a judge or a student in forensic science , to which the article was virtually addressed , and who might have seen _ negative reaction_. most likely he / she considered legitimate the charges of the policemen against the photographer .
the ` negative reaction ' would be that the reader would consider the rest of the article a support of dubious validity to some ` strange math ' that can never substitute the human intuition in a trial . ]
not a good service to the ` bayesian cause ' .
( imagine somebody trying to convince you with arguments you hardly understand and who begins asserting something you consider manifestly false . ) in the following section i introduce the basic elements of bayesian reasoning ( subsection [ ss : jl ] can be skipped on first reading ) , using a toy model as guiding example in which the analysis of ref .
@xcite ( `` 1 in 12 '' , or , more precisely `` 1 in 13 '' ) holds . section [ sec : columbo_priors ]
shows how such a kind of evidence would change columbo s and jury s opinion .
then i discuss in section [ sec : finale ] why a similar argument does not apply to the clip in which columbo finally frames galesco , and why all witnesses of the crucial actions ( including tv watchers , with the exception of the author of ref .
@xcite and perhaps a few others ) and an hypothetical court jury ( provided the scene had been properly reported ) had to be absolutely positive the photographer killed his wife ( or at least he knew who did it in his place ) . the rest of the paper might be marginal , if you are just curious to know why i have a different opinion than ref .
@xcite , although i agree on the validity of bayesian reasoning .
in fact , at the end of the work , this paper is not the ` short note ' initially planned .
the reason is that the past months i had many discussions on some of the questions treated here with people from several fields .
i have realized once more that it is not easy to put the basic principles at work if some important issues are not well understood .
people are used to solving their statistical problems with ` ad hoc ' formulae ( see appendix h ) and therefore tend to add some ` bayesian recipes ' in their formularium .
it is then too high the risk that one looks at simplified methods
bayesian methods require a bit more thinking and computation that others ! that are even advertised as ` objective ' . or one just refuses to use any math , on the defense of pure intuition .
( by the way , this is an important point and i will take the opportunity to comment on the apparent contradictions between intuition and formal evaluation of beliefs , defending both , but encouraging the use of the latter , superior to the former in complex situations see in particular appendix c ) .
so , to conclude the introduction , this document offers several levels of reading : * if you are only interested to columbo s story , you can just jump straight to section [ sec : finale ] . *
if you also ( or regardless of columbo ) want to have an opportunity to learn the basic rules of bayesian inference , subsections [ ss : bayes_theorem ] , [ ss : bayes_theorem_2 ] and [ ss : many_pieces ] , based on a simple master example , have been written on the purpose
. then you might appreciate the advantage of logarithmic updating ( section [ ss : jl ] ) and perhaps see how it applies to the aids example of appendix f. + * if you already know the basics of the probabilistic reasoning , but you wonder how it can be applied into real cases , then section [ sec : real_life ] should help , together with some of the appendices . *
if none of the previous cases is yours ( you might even be an expert of the field ) , you can simply browse the document .
perhaps some appendices or subsections might still be of your interest .
* finally , there is the question of the many footnotes , which can break the pace of the reading .
they are not meant to be necessarily read sequentially along with the main text and could be skipped on a first fast reading ( in fact , this document is closer to an hypertext than to a standard article . ) enjoy !
let us leave aside columbo s cameras for a while and begin with a different , simpler , stereotyped situation easier to analyze .
imagine there are two types of boxes , @xmath0 , that only contain white balls ( @xmath1 ) , and @xmath2 , that contain one white balls and twelve black ( incidentally , just to be precise , although the detail is absolutely irrelevant , we have to infer from columbo s words , _ `` you did nt touch any of these twelve cameras .
you picked up that one '' _ , the cameras were thirteen ) .
you take at random a box and extract a ball .
the resulting color is _
white_. you might be interested to evaluate the probability that the box is of type @xmath0 , in the sense of stating in a quantitative way how much you believe this hypothesis .
in formal terms we are interested in @xmath3 , knowing that @xmath4 and @xmath5 , a problem that can be sketched as @xmath6 [ here ` @xmath7 ' stands for ` given ' , or ` conditioned by ' ; @xmath8 is the general ( ` background ' ) status of information under which this probability is assessed ; ` @xmath9 ' or ` @xmath10 ' after ` @xmath7 ' indicates that both conditions are relevant for the evaluation of the probability . ]
a typical mistake at this point is to confuse @xmath3 with @xmath11 , or , more often , @xmath12 with @xmath13 , as largely discussed in ref .
hence we need to learn how to turn properly @xmath11 into @xmath3 using the rules of probability theory .
the ` probabilistic inversion ' @xmath14 can _ only _ and appendix h. ] be performed using the so - called _
bayes theorem _ , a simple consequence of the fact that , given the _ effect _
@xmath15 and some _ hypotheses _ @xmath16 concerning its possible cause , the joint probability of @xmath15 and @xmath16 , conditioned by the _ background information _ represents all we know about the hypotheses and the effect considered .
writing @xmath8 in all expressions could seem a pedantry , but it is nt .
for example , if we would just write @xmath17 in these formulae , instead of @xmath18 , one might be tempted to take this probability equal to one , because the observed event is a well established fact , that has happened and is then certain .
but it is not this certainty that enters these formulae , but rather the probability ` that fact could happen ' in the light of ` everything we knew ' about it ( ` @xmath8 ' ) . ]
@xmath8 , can be written as @xmath19 where ` @xmath20 ' stands for a logical ` and ' . from the second equality of the last equation we get@xmath21 that is one of the ways to express bayes theorem . ]
valid if we deal with a class of incompatible hypotheses [ i.e. @xmath22 and @xmath23 .
in fact , in this case a general rule of probability theory [ eq . ( [ eq : rul6 ] ) in appendix a ] allows us to rewrite the denominator of eq .
( [ eq : phi|e ] ) as @xmath24 . in this note , dealing only with two hypotheses , we prefer to reason in terms of probability ratios , as shown in eq .
( [ eq : bayes_factor ] ) . ] since a similar expression holds for any other hypothesis @xmath25 , dividing member by member the two expressions we can restate the theorem in terms of the relative beliefs , that is @xmath26 the initial ratio of beliefs ( ` odds ' ) is updated by the so - called _ bayes factor _ , that depends on how likely _ each _ hypothesis can produce that effect . , the probabilities of the effects @xmath27 and @xmath28 have usually nothing to do with each other . ] introducing @xmath29 and @xmath30 , with obvious meanings , we can rewrite eq .
( [ eq : bayes_factor ] ) as @xmath31 note that , if the initial odds are unitary , than the final odds are equal to the updating factor .
_ bayes factors can be interpreted as odds due only to an individual piece of evidence , the two hypotheses were considered initially equally likely_. and [ fn : jaynes ] , as well as appendix h. ) ] this allows us to rewrite @xmath32 as @xmath33 , where the tilde is to remind that _ they are not properly odds _ , but rather ` _ pseudo - odds _ ' .
we get then an expression in which all terms have _ virtually _ uniform meaning : @xmath34 if we have only two hypotheses , we get simply @xmath35 .
if the updating factor is unitary , then the piece of evidence does not modify our opinion on the two hypotheses ( no matter how small can numerator and denominator be , as long as their ratio remains finite and unitary !
see appendix g for an example worked out in details ) ; when @xmath36 vanishes , then hypothesis @xmath0 becomes impossible ( `` _ it is falsified _ '' ) ; if instead it is infinite ( i.e. the denominator vanishes ) , then it is the other hypothesis to be impossible .
( the undefined case @xmath37 means that we have to look for other hypotheses to explain the effect . , but an absolute truth , i.e. @xmath38 , depends on which class of hypotheses is considered . stated in other words , in the realm of probabilistic inference
_ falsities can be absolute , but truths are always relative_. ] ) applying the updating reasoning to our box game , the bayes factor of interest is @xmath39 as it was remarked , this number would give the required odds _ if _ the hypotheses were initially equally likely . but how strong are the initial relative beliefs on the two hypotheses ? ` unfortunately ' , we can not perform a probabilistic inversion if we are unable to assign somehow _ prior probabilities _ to the hypotheses we are interested in . and appendix h. [ fn : nopriors ] ] indeed , in the formulation of the problem i on purpose passed over the relevant pieces of information to evaluate the prior probabilities ( it was said that `` there are two types of boxes '' , not `` there are two boxes '' ! ) .
if we specify that we had @xmath40 boxes of type @xmath0 and @xmath41 of the other kind , then the initial odds are @xmath42 and the final ones will be @xmath43 from which we get ( just requiring that the probability of the two hypotheses have to sum up to one and @xmath44 are generic , complementary hypotheses we get , calling @xmath45 the bayes factor of @xmath46 versus @xmath44 and @xmath47 the initial odds to simplify the notation , the following convenient expressions to evaluate the probability of @xmath46 : @xmath48 ] ) @xmath49 if the two hypotheses were _ initially _ considered equally likely , then the evidence @xmath1 makes @xmath0 13 times more believable than @xmath2 , i.e. @xmath50 , or approximately 93% . on the other hand , if @xmath0 was _ a priori _ much less credible than @xmath2 , for example by a factor 13 , just to play with round numbers , the same evidence made @xmath0 and @xmath2 equally likely . instead
, if we were initially in strong favor of @xmath0 , considering it for instance 13 times more plausible than @xmath2 , that evidence turned this factor into 169 , making us 99.4% confident _ highly confident _ , some would even say ` practically sure ' ! that the box is of type @xmath0 .
imagine now the following variant of the previous toy experiment .
after the white ball is observed , you put it again in the box , shake well and make a second extraction .
you get white the second time too . calling @xmath51 and
@xmath52 the two observations , we have now : ) , although we are dealing now with more complex events and complex hypotheses , logical and of simpler ones . moreover , eq . ( [ eq : seq:2 ] )
is obtained from eq .
( [ eq : seq:1 ] ) making use of the formula ( [ eq : joint_prob ] ) of joint probability , that gives @xmath53 and an analogous formula for @xmath2 . note also that , going from eq .
( [ eq : seq:2 ] ) to eq .
( [ eq : seq:3 ] ) , @xmath54 has been rewritten as @xmath55 to emphasize that the probability of a second white ball , conditioned by the box composition and the result of the first extraction , depends indeed only on the box content and not on the previous outcome ( ` extraction after re - introduction ' ) . ]
@xmath56 that , using the compact notation introduced above , we can rewrite in the following enlighting forms .
the first is [ eq . ( [ eq : seq:4 ] ) ] @xmath57 that is , _ the final odds after the first inference become the initial odds of the second inference _ ( and so on , if there are several pieces of evidence ) .
therefore , beginning from a situation in which @xmath0 was thirteen times more credible than @xmath2 is exactly equivalent to having started from unitary odds updated by a factor 13 due to a piece of evidence .
the second form comes from eq .
( [ eq : seq:3 ] ) : @xmath58 i.e. ) follows from eq .
( [ eq : seq:3a ] ) because a bayes factor can be defined as the ratio of final odds over the initial odds , depending on the evidence .
therefore @xmath59 ] @xmath60 _ bayes factors _ due to _ independent _ , that we have used above to turn eq .
( [ eq : seq:2 ] ) into eq .
( [ eq : seq:3 ] ) and that can be expressed , in general terms as @xmath61 i.e. , _ under the condition of a well precise hypothesis _ ( @xmath16 ) , the probability of the effect @xmath62 does not depend on the knowledge of whether @xmath63 has occurred or not . note that , in general , although @xmath63 and @xmath62 are independent given @xmath16 ( they are said to be _ conditionally independent _ ) , they might be otherwise _ dependent _ , i.e. @xmath64 .
( going to the example of the boxes , it is rather easy to grasp , although i can not enter in details here , that , if we do not know the kind of box , the observation of @xmath51 changes our opinion about the box composition and , as a consequence , the probability of @xmath52 see the examples in appendix j ) ] pieces of evidence _ multiply_.
that is , two independent pieces of evidence ( @xmath51 and @xmath52 ) are equivalent to a single piece of evidence ( ` @xmath65 ' ) , whose bayes factor is the product of the individual ones . in our case @xmath66 . in general , if we have several hypotheses @xmath16 and several _ independent _ pieces of evidence , @xmath63 , @xmath62 , ,
@xmath67 , indicated all together as @xmath68 , then eq .
( [ eq : bayes_factor ] ) becomes @xmath69 \times o_{i , j}(i)\ , , \label{eq : product_odds}\end{aligned}\ ] ] i.e. @xmath70 where @xmath71 stand for ` product ' ( analogous to @xmath72 for sums ) . the remark that bayes factors due to independent pieces of evidence multiply together and the overall factor finally multiplies the initial odds suggests a change of variables in order to play with additive quantities .
this can be done taking the logarithm of both sides of eq .
( [ eq : product_odds ] ) , that then become @xmath73 & = & \sum_{k=1}^n \log_{10}[\tilde o_{i , j}(e_k , i ) ] + \log_{10}[o_{i , j}(i)]\ , , \label{eq : sum_bf}\end{aligned}\ ] ] respectively , where the base 10 is chosen for practical convenience because , as we shall discuss later , what substantially matters are powers of ten of the odds . introducing the new symbol jl
, we can rewrite eq .
( [ eq : sum_bf ] ) as @xmath74 or @xmath75 where @xmath76 \label{eq : jl_alle}\\ \mbox{jl}_{i , j}(i ) & = & \log_{10}\left[o_{i , j}(i ) \right]\label{eq : jl_0}\\ \delta\mbox{jl}_{i , j}(e_k , i ) & = & \log_{10}\left[\tilde o_{i , j}(e_k , i)\right ] \label{eq : deltajl_k } \\ \delta\mbox{jl}_{i ,
j}(\mvec{e},i ) & = & \sum_{k=1}^n \delta\mbox{jl}_{i , j}(e_k , i)\ , .
\label{eq : sum_deltajl_k}\end{aligned}\ ] ] the letter ` l ' in the symbol is to remind _
logarithm_. but it has also the mnemonic meaning of _ leaning _ , in the sense of ` inclination ' or ` propension ' . the ` j ' is for _ judgment_. therefore ` jl ' stands for _ judgement leaning _ , that is an inclination of the judgement , an expression i have taken the liberty to introduce , using words not already engaged in probability and statistics , because in these fields many controversies are due to different meanings attributed to the same word , or expression , by different people ( see appendices b and g for further comments ) .
jl can then be visualized as the indicator of the ` justice balance ' to which @xmath77 refers stands for ` guilty ' . ]
( figure [ fig : jl ] ) , that displays zero if there is no unbalance , but it could move to the positive or the negative side depending on the weight of the several arguments pro and con .
the role of the evidence is to vary the jl indicator by quantities @xmath78 s equal to base 10 logarithms of the bayes factors , that have then a meaning of _ weight of evidence _ , an expression due to charles sanders peirce @xcite ( see appendix e ) .
but the judgement is rarely initially unbalanced .
this the role of @xmath79 , that can be considered as a a kind of _ initial weight of evidence _ due to our prior knowledge about the hypotheses @xmath16 and @xmath25 [ and that could even be written as @xmath80 , to stress that it is related to a 0-th piece of evidence ] to understand the rationale behind a possible uniform treatment of the prior as it would be a piece of evidence , let us start from a case in which you now _ absolutely nothing_. for example have to state your beliefs on which of my friends , dino or paolo , will first run next rome marathon .
it is absolutely reasonable you assign to the two hypotheses equal probabilities , i.e. @xmath81 , or @xmath82 ( your judgement is perfectly balanced ) .
this is because in brain these names are only possibly related to italian males .
nothing more .
( but nowadays search engines over the web allow to modify your opinion in minutes . ) as soon as you deal with _ real _ hypotheses of your interest , things get quite different .
it is in fact very rare the case in which the hypotheses tell you not more than their names .
it is enough you think at the hypotheses ` rain ' or ` not rain ' , the day after you read these lines in the place where you live . in general
the information you have in your brain related to the hypotheses of your interest can be considered the initial piece of evidence _ you _ have , usually different from that somebody else might have ( this the role of @xmath8 in all our expressions ) .
it follows that prior odds of 10 ( @xmath83 ) will influence your leaning towards one hypothesis , exactly like unitary odds ( @xmath84 ) followed by a bayes factor of 10 ( @xmath85 ) .
this the reason they enter on equal foot when `` balancing arguments '' ( to use an expression la peirce see the appendix e ) pro and against hypotheses .
.a comparison between probability , odds and judgement leanings [ cols="^,^,^,^,^,^,^ " , ] as we see from this table , and as we better understand from figure [ fig : two_normal ] , numbers large in module are in favor of @xmath44 , and very large ones are in its strong favor . instead
, the numbers laying in the interval defined by the two points marked in the figure by a cross provide evidence in favor of @xmath46 .
however , while individual pieces of evidence in favor of @xmath46 can only be weak ( the maximum of @xmath86jl is about 0.3 , reached around @xmath87 , namely @xmath88 , to be precise , for which @xmath86jl reaches 0.313 ) , those in favor of the alternative hypothesis can be sometimes very large .
it follows then that one gets easier convinced of @xmath44 rather than of @xmath46 .
we can check this by a little simulation .
we choose a model , extract 50 random variables and analyze the data as if we did nt know which generator produced them , although considering @xmath46 and @xmath44 equally likely .
we expect that , as we go on with the extractions , the pieces of evidence accumulate until we _ possibly _ reach a level of practical certainty .
obviously , the individual pieces of evidence do not provide the same @xmath86jl , and also the sign can fluctuate , although we expect more positive contributions if the points are generated by @xmath46 and the other way around if they came from @xmath44 .
therefore , as a function of the number of extractions the accumulated weight of evidence follows a kind of _ asymmetric random walk _
( imagine the jl indicator fluctuating as the simulated experiment goes on , but drifting ` in average ' in one direction ) .
figure [ fig : simulazioni ] shows 200 inferential stories , half per generator .
we see that , in general , we get practically sure of the model after a couple of dozens of extractions .
but there are also cases in which we need to wait longer before we can feel enough sure on one hypothesis .
it is interesting to remark that the leaning in favor of each hypothesis grows , _ in average _ , linearly with the number of extractions .
that is , a little piece of evidence , which is in average positive for @xmath46 and negative for @xmath44 , is added after each extraction . however , around the average trend , there is a large varieties of individual inferential histories .
they all start at @xmath89 for @xmath90 , but in practice there are no two identical ` trajectories ' .
all together they form a kind of ` fuzzy band ' , whose ` effective width ' grows also with the number of extractions , _ but not linearly_. the widths grows as the square root of @xmath91 .. we can also evaluate the _ uncertainty of prevision _
, quantified by the standard deviation .
we get for the two hypotheses @xmath92 & = & 0.15 \\
\sigma[\delta\mbox{jl}_{1,2}(h_1 ) ] & = & 0.24 \\ u_r[\delta\mbox{jl}_{1,2}(h_1 ) ] & = & 1.6 \end{array}\right . \hspace{0.5
cm } & & \hspace{0.5 cm } \left\{\begin{array}{rcl } \mbox{e}[\delta\mbox{jl}_{1,2}(h_2)]&=&-0.38 \\ \sigma[\delta\mbox{jl}_{1,2}(h_2)]&=&0.97 \\
u_r[\delta\mbox{jl}_{1,2}(h_2 ) ] & = & 2.6 \end{array}\right.\end{aligned}\ ] ] where also the _ relative uncertainty _
@xmath93 has been reported , defined as the uncertainty divided by the absolute value of the prevision .
the fact that the uncertainties are relatively large tells clearly that we _ do not expect _ that a single extraction will be sufficient to convince us of either model .
but this does not mean we can not take the decision because the number of extraction _ has been _ too small .
if a very large fluctuation provides a @xmath86jl of @xmath94 ( the table in this section shows that this is not very rare ) , we have already got a very strong evidence in favor of @xmath44 . repeating what has been told several time , what matters is the cumulated judgement leaning .
it is irrelevant if a jl of @xmath94 comes from ten individual pieces of evidence , only from a single one , or partially from evidence and partially from prior judgement .
+ when we plan to make @xmath91 extractions from a generator , probability theory allows us to calculate expected value and uncertainty of @xmath95 : @xmath96 & = & n \times \mbox{e}[\delta\mbox{jl}_{1,2}(h_i ) ] \\
\sigma[\delta\mbox{jl}_{1,2}(n , h_i ) ] & = & \sqrt{n } \times \sigma[\delta\mbox{jl}_{1,2}(h_i)]\\ u_r[\delta\mbox{jl}_{1,2}(n , h_i ) ] & = & \frac{1}{\sqrt{n}}\times u_r[\delta\mbox{jl}_{1,2}(h_i)]\,.\end{aligned}\ ] ] in particular , for @xmath97 we get @xmath98 ( @xmath99 ) and @xmath100 ( @xmath101 ) , that explain the gross feature of the bands in figure [ fig : simulazioni ] .
] this is the reason why , as @xmath91 increases , the bands tend to move away from the line @xmath84 .
nevertheless , individual trajectories can exhibit very ` irregular ' that do not follow the general trend _ are not exceptions _ , being generated by the same rules that produces all of them . ]
behaviors as we can also see in figure [ fig : simulazioni ] .
some comments on _ likelihood _ are also in order , because the reader might have heard this term and might wonder if and how it fits in the scheme of reasoning expounded here .
one of the problems with this term is that it tends to have several meanings , and then to create misunderstandings . in plane english
` likelihood ' is `` 1 . the condition of being likely or probable ; probability '' , or `` 2 .
something that is probable '' ; but also `` 3 .
( mathematics & measurements / statistics ) the probability of a given sample being randomly drawn regarded as a function of the parameters of the population '' .
technically , with reference to the example of the previous appendix , the likelihood is simply @xmath102 , where @xmath103 is fixed ( the observation ) and @xmath16 is the ` parameter ' .
then it can take two values , @xmath104 and @xmath105 .
if , instead of only two models we had a continuity of models , for example the family of all gaussian distributions characterized by central value @xmath106 and ` effective width ' ( standard deviation ) @xmath107 , our likelihood would be @xmath108 , i.e. @xmath109 written in this way to remember that : 1 ) a likelihood is a function of the model parameters and not of the data ; 2 ) @xmath110 _ is not _ a probability ( or a probability density function ) of @xmath106 and @xmath107 .
anyway , for the rest of the discussion we stick to the very simple likelihood based on the two gaussians . that is , instead of a double infinity of possibilities , our space of parameters is made only of two points , @xmath111 and @xmath112 .
thus the situation gets simpler , although the main conceptual issues remain substantially the same . in principle
there is nothing bad to give a special name to this function of the parameters .
but , frankly , i had preferred statistics gurus named it after their dog or their lover , rather than call it ` likelihood . '
was introduced by r. a. fisher with the object of _ avoiding _ the use of bayes theorem '' @xcite . ]
the problem is that it is very frequent to hear students , teachers and researcher explaining that the ` likelihood ' tells `` how likely the parameters are '' ( this is _ the probability of the parameters ! not the ` likelihood ' _ ) . or
they would say , with reference to our example , `` it is the probability that @xmath103 comes from @xmath16 '' ( again , this expression would be the probability of @xmath16 given @xmath103 , and not the probability of @xmath103 given the models ! ) imagine if we have only @xmath46 in the game : @xmath103 comes with certainty from @xmath46 , although @xmath46 does not yield with certainty @xmath103 .
is '' ( note the quote mark of ` likely ' , as in the example of footnote [ fn : james ] ) .
but , fortunately we find in http://en.wikipedia.org/wiki/likelihood_function that `` this is not the same as the probability that those parameters are the right ones , given the observed sample .
attempting to interpret the likelihood of a hypothesis given observed evidence as the probability of the hypothesis is a common error , with potentially disastrous real - world consequences in medicine , engineering or jurisprudence .
see prosecutor s fallacy [ * ] for an example of this . ''
( [ * ] see http://en.wikipedia.org/wiki/prosecutor%27s_fallacy . )
+ now you might understand why i am particular upset with the name likelihood .
[ fn_wiki_likelihood ] ] several methods in ` conventional statistics ' use somehow the likelihood to decide which model or which set of parameters describes at best the data .
some even use the likelihood ratio ( our bayes factor ) , or even the logarithm of it ( something equal or proportional , depending on the base , to the weight of evidence we have indicated here by jl )
. the most famous method of the series is the _ maximum likelihood principle_. as it is easy to guess from its name , it states that the _ best estimates _ of the parameters are those which maximize the likelihood .
all that _ seems _ reasonable and in agreement with what it has been expounded here , but it is not quite so .
first , for those who support this approach , likelihoods are not just a part of the inferential tool , they are everything .
priors are completely neglected , more or less because of the objections in footnote [ fn : nopriors ] .
this can be acceptable , if the evidence is overwhelming , but this is not always the case .
unfortunately , as it is now easy to understand , neglecting priors is mathematically equivalent to consider the alternative hypotheses equally likely ! as a consequence of this statistics miseducation ( most statistics courses in the universities all around the world only teach
` conventional statistics ' and never , little , or badly probabilistic inference ) is that too many unsuspectable people fail in solving the aids problem of appendix b , or confuse the likelihood with the probability of the hypothesis , resulting in misleading scientific claims ( see also footnote [ fn_wiki_likelihood ] and ref .
@xcite ) .
the second difference is that , since `` there are no priors '' , the result can not have a probabilistic meaning , as it is openly recognized by the promoters of this method , who , in fact , do not admit we can talk about probabilities of causes ( but most practitioners seem not to be aware of this ` little philosophical detail ' , also because frequentistic gurus , having difficulties to explain what is the meaning of their methods , they say they are ` probabilities ' , but in quote marks![multiblock footnote omitted ] ) . as a consequence , the resulting ` error analysis ' , that in human terms means to assign different beliefs to different values of the parameters , is cumbersome . in practice
the results are reasonable only if the possible values of the parameters are initially equally likely and the ` likelihood function ' has a ` kind shape ' ( for more details see chapters 1 and 12 of ref .
in most cases ( and practically always in courts ) pieces of evidence are not acquired directly by the person who has to form his mind about the plausibility of a hypothesis .
they are usually accounted by an intermediate person , or by a chain of individuals .
let us call @xmath113 the report of the evidence @xmath15 provided in a _
testimony_. the inference becomes now @xmath114 , generally different from @xmath115 . in order to apply bayes theorem in one of its form
we need first to evaluate @xmath116 .
probability theory teaches us how to get it [ see eq .
( [ eq : rul4 ] ) in appendix a ] : @xmath117 ( @xmath113 could be due to a true evidence or to a fake one ) .
three new ingredients enter the game : * @xmath118 , that is the probability of the evidence to be correctly reported as such . * but the testimony could also be incorrect the other way around ( it could be incorrectly reported , simply by mistake , but also it could be a ` fabricated evidence ' ) , and therefore also @xmath119 is needed . note that the probabilities to lie could be in general asymmetric , i.e. @xmath120 , as we have seen in the aids problem of appendix f , in which the response of the analysis models false witness well . * finally , since @xmath121 enters now directly , the ` nave ' bayes factor , only depending on @xmath122 , is not longer enough . taking our usual two hypotheses , @xmath123 and @xmath124 , we get the following bayes factor based on the _ testified evidence _
@xmath113 ( hereafter , in order to simplify the notation , we use the subscript ` @xmath125 ' in odds and bayes factors , instead of ` @xmath126 ' , to indicate that they are in favor of @xmath125 and against @xmath127 , as we already did in the aids example of appendix f ) : @xmath128 as expected , this formula is a bit more complicate that the bayes factor calculated taking @xmath15 for granted , which is recovered if the lie probabilities vanish @xmath129 { } { \ \ \
\tilde o_{h}(e , i)}\,,\end{aligned}\ ] ] i.e. only when we are absolutely sure the witness does not err or lie reporting @xmath15 ( but peirce reminds us that `` absolute certainty , or an infinite chance , can never be attained by mortals '' @xcite ) .
in order to single out the effects of the new ingredients , eq .
( [ eq : bf_lie ] ) can be rewritten as and @xmath130 respectively in the numerator and in the denominator , eq .
( [ eq : bf_lie ] ) becomes @xmath131 then @xmath132 can be indicated as @xmath133 , @xmath121 is equal to @xmath134 and , finally , @xmath130 can be written as @xmath135 . ]
@xmath136 } { 1 + \lambda(i ) \cdot \left [ \frac { \tilde o_{h}(e , i)}{p(e\,|\,h , i ) } -1 \right ] } \ , , \label{eq : weight_et}\end{aligned}\ ] ] where @xmath137 under the _ condition _ can not be factorized .
the effective odds @xmath138 can however be written in the following convenient forms @xmath139 although less interesting than eq .
( [ eq : weight_et ] ) .
[ fb : peh=0 ] ] @xmath140 and @xmath141 , i.e. _ @xmath142 positive and finite_. the parameter @xmath133 , ratio of the _ probability of fake evidence _ and the _ probability that the evidence is correctly accounted _ , can be interpreted as a kind of _ lie factor_. given the human roughly logarithmic sensibility to probability ratios , it might be useful to define , in analogy to the jl , @xmath143\,.\end{aligned}\ ] ] let us make some instructive limits of eq .
( [ eq : weight_et ] ) .
@xmath144{}{}&\tilde o_{h}(e , i)\\ \tilde o_{h}(e_t , i ) & \xrightarrow [ \mbox{{\footnotesize $ \lambda(i ) \rightarrow 1 $ } } ] { } { } & 1 \\ \tilde o_{h}(e_t , i ) & \xrightarrow [ \mbox{{\footnotesize $ p(e\,|\,h , i)\rightarrow 0 $ } } ] { } { } & 1\\ \tilde o_{h}(e_t , i ) & \xrightarrow [ \mbox{{\footnotesize $ \tilde o_{h}(e , i)\rightarrow \infty $ } } ] { } { } & \frac{p(e\,|\,h , i)}{\lambda(i ) } + 1 - p(e\,|\,h , i)\end{aligned}\ ] ] as we have seen , the ideal case is recovered if the lie factor vanishes . instead ,
if it is equal to 1 , i.e. @xmath145 , the reported evidence becomes useless .
the same happens if @xmath146 vanishes [ this implies that @xmath130 vanishes too , being @xmath147 . however , the most remarkable limit is the last one .
it states that , even if @xmath142 is very high , the effective bayes factor can not exceed the inverse of the lie factor : @xmath148}\,,\end{aligned}\ ] ] or , using logarithmic quantities @xmath149 } \,.\end{aligned}\ ] ] at this point some numerical examples are in order ( and those who claim they can form their mind on pure intuition get all my admiration
_if _ they really can ) .
let us imagine that @xmath15 would ideally provide a weight of evidence of 6 [ i.e. @xmath150 .
we can study , with the help of table [ tab : jl_et ] , + @xmath151 & & + & @xmath152 : & 10 & @xmath153 & @xmath154 & @xmath155 & @xmath156 & @xmath157 & @xmath158 & @xmath159 + @xmath160 & & 6.00 & 6.00 & 6.00 & 6.00 & 6.00 & 6.00 & 6.00 & 6.00 + @xmath161 & & 6.00 & 6.00 & 6.00 & 6.00 & 5.99 & 5.95 & 4.96 & @xmath162 + @xmath163 & & 5.96 & 5.96 & 5.96 & 5.95 & 5.92 & 5.68 & 4.00 & @xmath164 + @xmath165 & & 5.70 & 5.70 & 5.70 & 5.68 & 5.52 & 4.92 & 3.00 & @xmath166 + @xmath94 & & 4.96 & 4.96 & 4.95 & 4.92 & 4.68 & 3.95 & 2.00 & @xmath167 + @xmath168 & & 4.00 & 4.00 & 3.99 & 3.95 & 3.70 & 2.96 & 1.04 & @xmath169 + @xmath158 & & 3.00 & 3.00 & 3.00 & 2.96 & 2.70 & 1.96 & 0.30 & @xmath170 + @xmath171 & & 2.00 & 2.00 & 2.00 & 1.95 & 1.70 & 1.00 & 0.04 & @xmath172 + @xmath157 & & 1.00 & 1.00 & 1.00 & 0.96 & 0.74 & 0.26 & 0.004 & @xmath173 + 0 & & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 + + + @xmath151 & & + & @xmath152 : & 10 & @xmath153 & @xmath154 & @xmath155 & @xmath156 & @xmath157 & @xmath158 & @xmath159 + @xmath160 & & 3.00 & 3.00 & 3.00 & 3.00 & 3.00 & 3.00 & 3.00 & 3.00 + @xmath94 & & 3.00 & 3.00 & 3.00 & 3.00 & 2.99 & 2.95 & 1.96 & @xmath167 + @xmath168 & & 2.96 & 2.96 & 2.96 & 2.95 & 2.92 & 2.68 & 1.04 & @xmath169 + @xmath158 & & 2.70 & 2.70 & 2.70 & 2.68 & 2.52 & 1.93 & 0.30 & @xmath170 + @xmath171 & & 1.96 & 1.96 & 1.96 & 1.92 & 1.68 & 1.00 & 0.04 & @xmath172 + @xmath157 & & 1.00 & 1.00 & 0.99 & 0.96 & 0.74 & 0.26 & 0.004 & @xmath173 + 0 & & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 + + + @xmath151 & & + & @xmath152 : & 10 & @xmath153 & @xmath154 & @xmath155 & @xmath156 & @xmath157 & @xmath158 & @xmath159 + @xmath160 & & 1.00 & 1.00 & 1.00 & 1.00 & 1.00 & 1.00 & 1.00 & 1.00 + @xmath158 & & 1.00 & 1.00 & 1.00 & 1.00 & 0.99 & 0.96 & 0.26 & @xmath170 + @xmath171 & & 0.96 & 0.96 & 0.96 & 0.96 & 0.93 & 0.72 & 0.04 & @xmath172 + @xmath157 & & 0.72 & 0.72 & 0.72 & 0.70 & 0.58 & 0.23 & 0.003 & @xmath173 + @xmath174 & & 0.41 & 0.41 & 0.41 & 0.39 & 0.27 & 0.07 & @xmath175 & @xmath173 + 0 & & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 + how the _ weight of the reported evidence _
@xmath176 depends on the other beliefs [ in this table logarithmic quantities have been used throughout , therefore @xmath177 is the base ten logarithm of the odds in favor of @xmath15 given the hypothesis @xmath125 ; the table provides , for comparisons , also @xmath176 from @xmath178 equal to 3 and 1 ] .
the table exhibits the limit behaviors we have seen analytically . in particular , if we fully trust the report , i.e. @xmath179 , then @xmath176 is exactly equal to @xmath180 , as we already know .
but as soon as the absolute value of the lie factor is close to @xmath181 , there is a sizeable effect .
the upper bound can be the be rewritten as @xmath182\,,\\ \mbox{or}\hspace{3cm}\mbox { } & & \mbox { } \nonumber\\ \delta\mbox{jl}_h(e_t , i ) & \le & \mbox{min}\,[\delta\mbox{jl}_h(e , i),\,-\mbox{j}\lambda(i ) ] \,,\end{aligned}\ ] ] a relation valid in the region of interest when thinking about an evidence in favor of @xmath125 , i.e. @xmath183 and @xmath184 .
this upper bound is very interesting .
since minimum conceivable values of @xmath151 for human beings can be of the order of @xmath165 ( to perhaps @xmath185 or @xmath186 , but in many practical applications @xmath171 or @xmath158 can already be very generous ! ) , in practice the effective weights of evidence can not exceed values of about @xmath187 ( i have no strong opinion on the exact value of this limit , my main point is that _ you consider there might be such a practical human limit_. ) this observation has an important consequence in the combination of evidences , as anticipated at the end of section [ ss : agatha ]
. should we give more consideration to a single strong piece of evidence , virtually weighing @xmath188 , or 10 independent weaker evidences , each having a @xmath86jl of 1 ?
as it was said , in the ideal case they yield the same global leaning factor .
but as soon as human fallacy ( or conspiracy ) is taken into account , and we remember that our belief is based on @xmath113 and not on @xmath15 , then we realize that @xmath189 is well above the range of jl that we can reasonably conceive .
instead the weaker pieces of evidence are little affected by this doubt and when they sum up together , they really can provide a @xmath86jl of about 10 .
let us go back to our toy model of section [ sec:1in13 ] and let us complicate it just a little bit , adding the possibility of incorrect testimony ( but we also simplify it using uniform priors , so that we can focus on the effect of the _ uncertain evidence _ ) .
for example , imagine you do not see directly the color of the ball , but this is reported to you by a collaborator , who , however , might not tell you always the truth .
we can model the possibility of a lie in following way : after each extraction he tosses a die and reports the true color only if the die gives a number smaller than 6 . using the formalism of appendix i
, we have @xmath190 the resulting _ belief network _ , .
since the connections between the _ nodes _ of the resulting _ network _ have usually the meaning of probabilistic links ( but also deterministic relations can be included ) , this graph is called a _
belief network_. moreover , since bayes theorem is used to update the probabilities of the possible _ states _ of the nodes ( the node ` box ' , with reference to our toy model , has states @xmath0 and @xmath2 ; the node ` ball ' has states @xmath1 and @xmath191 ) , they are also called _ bayesian networks_. for more info , as well as tutorials and demos of powerful packages having also a friendly graphical user interface , i recommend visiting hugin @xcite and netica @xcite web sites .
( my preference for hugin is mainly due to the fact that it is multi - platform and runs nicely under linux . ) for a book introducing bayesian networks in forensics , ref .
@xcite is recommended . for a monumental probabilistic network on the ` * case that will never end * ' , see ref .
@xcite ( if you like classic thrillers , the recent paper of the same author might be of your interest @xcite ) . ] relative to five extractions and to the corresponding five reports is shown in figure [ fig : bn ] , redrawn in a different way in figure [ fig : bn_mon_0 ] . in this diagram
the _ nodes _ are represented by ` monitors ' that provide the probability of each _ state _ of the _ variable_. the green bars mean that we are in condition of uncertainty with respect to all states of all variable .
let us describe the several nodes : * initial box compositions have probability 50% each , that was our assumption . *
the probability of white and black are the same for all extractions , with white a bit more probable than black ( 14/26 versus 12/26 , that is 53.85% versus 46.15% ) .
* there is also higher probability that the ` witness ' reports white , rather than black , but the difference is attenuated by the ` lie factors . ' and one for @xmath192 . for simplicity
we assume here they have the same value .
] in fact , calling @xmath193 and @xmath194 the reported colors we have @xmath195 all probabilities of the network have been updated ( hugin @xcite has nicely done the job for us ) .
we recognize the 93% of box @xmath0 , that we already know .
we also see that the increased belief on this box makes us more confident to observe white balls in the following extractions ( after re - introduction ) .
the fact that the witness could lie reduces , with respect to the previous case , our confidence on @xmath0 and on white balls in future extractions . as an exercise on what we have learned in appendix h
, we can evaluate the ` effective ' bayes factor @xmath196 that takes into account the testimony .
applying eq .
( [ eq : weight_et ] ) we get @xmath197 } { 1 + \lambda(i ) \cdot \left [ \frac { \tilde o_{h}(w , i)}{p(w\,|\,h , i ) } -1 \right ] } \\ & = & 13\times \frac{5}{17 } = 3.82\,,\end{aligned}\ ] ] or @xmath198 , about a factor of two smaller than @xmath199 , that was 1.1 ( this mean we need two pieces of evidence of this kind to recover the loss of information due to the testimony ) .
the network gives us also the probability that the witness has really told us the truth , i.e. @xmath200 , that is _ different _ from @xmath201 , the reason being that white was initially a bit more probable than black .
the most interesting thing that comes from the result of the network is how the probabilities that the two witness lie change .
first we see that they are the same , about 95% , as expected for symmetry . but
the surprise is that the probability the the first witness said the truth has increased , passing from 85% to 95% .
we can justify the variation because , in qualitative agreement with intuition , if we have concordant witnesses , _ we tend to believe to each of them more than what we believed individually_. once again , the result is , perhaps after an initial surprise , in qualitative agreement with intuition .
the important point is that intuition is unable to get quantitative estimates .
again , the message is that , once we agree on the basic assumption and we check , whenever it is possible , that the results are reasonable , it is better to rely on automatic computation of beliefs .
this last information reduces the probability of @xmath0 , but does not falsify this hypothesis , as if , instead , we had _ observed _ black .
obviously , it does also reduce the probability of white balls in the following extractions .
the other interesting feature concerns the probability that each witness has reported the truth .
our belief that the previous two witnesses really saw what they said is reduced to 83% .
but , nevertheless we are more confident on the first two witnesses than on the third one , that we trust only at 76% , although the lie factor is the same for all of them .
the result is again in agreement with intuition : if many witnesses state something and fewer say the opposite , _ we tend to believe the majority _
, if we initially consider all witnesses equally reliable .
but a bayesian network tells us also how much we have to believe the many more then the fewer .
let us do , also in this case the exercise of calculating the effective bayes factor , using however the first formula in footnote [ fb : peh=0 ] : the effective odds @xmath202 can be written as @xmath203 i.e. @xmath204 } = { 13}/{61 } = 0.213 $ ] , smaller then 1 because they provide an evidence against box @xmath0 ( @xmath205 ) .
it is also easy to check that the resulting probability of 75.7% of @xmath0 can be obtained summing up the three weights of evidence , two in favor of @xmath0 and two against it : @xmath206 , i.e. @xmath207 , that gives a probability of @xmath0 of 3.1/(1 + 3.1)=76% . only in this case
we become certain that the box is of the kind @xmath2 , and the game is , to say , finished .
but , nevertheless , we still remain in a state on uncertainty with respect to several things .
the first one is the probability of a white ball in future extractions , that , from now becomes 1/13 , i.e. 7.7% , and does not change any longer .
but we also remain uncertain on whether the witnesses told us the truth , because what they said is not incompatible with the box composition .
but , and again in qualitative agreement with the intuition , we trust much more whom told black ( 1.6% he lied ) than the two who told white ( 70.6% they lied ) . another interesting way of analyzing the final network is to consider the probability of a black ball in the five extractions considered .
the fourth is one , because we have seen it .
the fifth is 92.3% ( @xmath209 ) because we know the box composition . but in the first two extractions the probability is smaller than it ( 70.6% ) , while in the third is higher ( 98.4% ) .
that is because in the two different cases we had an evidence respectively against and in favor of them . | triggered by a recent interesting new scientist article on the too frequent incorrect use of probabilistic evidence in courts , i introduce the basic concepts of probabilistic inference with a toy model , and discuss several important issues that need to be understood in order to extend the basic reasoning to real life cases . in particular , i emphasize the often neglected point that degrees of beliefs are updated not by ` bare facts ' alone , but by all available information pertaining to them , including how they have been acquired . in this light
i show that , contrary to what claimed in that article , there was no `` probabilistic pitfall '' in the columbo s episode pointed as example of `` bad mathematics '' yielding `` rough justice '' .
instead , such a criticism could have a ` negative reaction ' to the article itself and to the use of bayesian reasoning in courts , as well as in all other places in which probabilities need to be assessed and decisions need to be made . anyway , besides introductory / recreational aspects , the paper touches important questions , like : role and evaluation of priors ; subjective evaluation of bayes factors ; role and limits of intuition ; ` weights of evidence ' and ` intensities of beliefs ' ( following peirce ) and ` judgments leaning ' ( here introduced ) , including their uncertainties and combinations ; role of relative frequencies to assess and express beliefs ; pitfalls due to ` standard ' statistical education ; weight of evidences mediated by testimonies . a small introduction to bayesian networks , based on the same toy model ( complicated by the possibility of incorrect testimonies ) and implemented using hugin software ,
is also provided , to stress the importance of formal , computer aided probabilistic reasoning .
# 1 |
detecting galaxies at high - redshift provides a powerful observational way to probe structure formation in the early universe .
there have now been several hundreds of galaxies found using the lyman - break technique ( e.g. @xcite .
although very successful , this method is limited to luminous objects .
a complementary technique to select high - redshift galaxies is the detection of systems in absorption in the spectra of background quasars .
since their selection is independent of luminosity and morphology , the study of galaxy evolution through these systems circumvents the biases inherent to traditional emission studies .
+ damped lyman-@xmath0 systems ( hereafter dlas ) have hydrogen column densities , log n(hi ) @xmath6 20.3 and are thought to be associated with parts of high - redshift galaxies , just as they are at lowest redshift @xcite . from our knowledge of the local universe
we expect that high - redshift galaxies , especially those forming stars , should contain some amount of dust .
therefore , the possibility that dlas are dusty objects is currently hotly debated . indeed , smoothed - particle hydrodynamic simulations @xcite predict that dlas should be 1/3@xmath8 solar at @xmath9 and even more metal - rich toward lower redshifts whilst current observations measure typically a metallicity around 1/10@xmath10 solar .
dust is therefore invoked to reconcile results from the simulations with current observations .
in addition , one can estimate the dust content of individual dla from the ratio of depleted onto undepleted elements it contains .
it is assumed that a fraction of the elements is not detected in the gas phase because it is locked into dust grains ( e.g. @xcite .
such studies have shown that the dust depletion levels observed in dlas are small , lower than those measured in the galactic interstellar - medium clouds with similar n(hi ) column densities .
in fact , the observations resemble the dust depletion of the galactic halo warm gas @xcite . another more direct way to probe the dust content of dlas is to look for a 2175 feature ( the so - called `` dust bump '' ) in the quasar spectrum .
this feature is known to be more predominant in the milky way than in the small and large magellanic clouds .
many searches lead to no - detections @xcite . using sloan digital sky survey ( sdss ) spectra
, @xcite claimed the detections of bumps in three individual mgii absorbers at z=1.5 , while @xcite have detected a clear bump in a dla at z=0.524 toward a bl lac at z=0.94 .
other searches remained unsuccessful , indicating that most dlas do not show a predominant 2175 feature . a promising tool for finding the signatures of dust are searches for the strong silicate absorption features in the infrared .
@xcite , using spitzer data , report on the first indication of 9.7 @xmath11 m silicate absorption in a dla at z@xmath12 , and find it to be stronger than expected from extrapolation of the relation from galactic diffuse ism .
+ a consequence of dlas being dusty is that it might prevent the inclusion of their background quasar in colour - colour selected samples . in order to overcome such biases ,
radio - selected quasar samples have been built recently : while @xcite did not discover a missing population , a more recent survey finds 4 fields ( out of @xmath260 ) where the radio - flux selected quasars have no optical counterparts down to very faint limits @xcite .
these objects are the potential candidates of dusty dlas obscuring background quasars .
+ in this paper , we investigate yet another possible signature of the dust content of dlas . indeed , dust in a quasar absorber will redden the colour of its background quasar .
@xcite were the first to look for reddening of the quasar continua in a sample containing dlas and detect reddening at 99.999% level .
more recently , @xcite have used 79 dlas from sdss data release 2 to compare the slope of power laws fitting quasar continua with and without dlas .
these authors found a colour excess due to small magellanic cloud type of dust of reddening @xmath5e(b - v)@xmath6 @xmath50.01 mag . using colour excess measurements for sdss sightlines containing dlas and matching sightlines without them , @xcite report a detection of reddening in a sample of dr5 sdss qsos towards sightlines containing dlas at the mean level of @xmath5e(b - v)@xmath6 @xmath26.3x10@xmath13 .
they derive their estimate by comparing colours of qsos with intervening dlas to the mean colour of a set of qso of similar emission redshift and brightness , and bootstrapping the resulting broad distribution of colour excesses to ascertain the slightly redder colours of the dla qsos .
a byproduct of the @xcite work , who performed a bayesian analysis on a variety of optical and radio datasets to determine the loss fraction of qsos due to dust obscuration , is an estimate of the probability density distribution for dust reddening in optical spectra , that also peaks around the same value for @xmath5e(b - v)@xmath6 . here , we build composite spectra of quasar spectra with dlas from sdss data releases 5 and 7 , and compare them with reference composite spectra built to match the original spectra in magnitudes and redshifts .
the organisation of the paper is as follows : after the introduction , we describe the selection of the dla samples in section 2 , followed by the construction of the matching non - absorber samples , and the procedure for obtaining composite spectra in section 3 , where we also presents the method for deriving extinction as well as the results for the full samples . after discussing subsample analyses ( section 4 ) , we summarise and conclude in section 5 .
we start the selection of our samples from two lists of sdss surveys for dlas .
@xcite searched in data release ( dr ) 5 , following a partially automated procedure tested on earlier data releases @xcite .
their cleanest ( statistical ) sample contains 738 dlas .
we have chosen not to include objects that are listed in their non - statistical sample , as these can be towards qsos with strong intrinsic absorption , or have z@xmath14 z@xmath15 .
broad absorption line qso ( bals ) are known to be more strongly reddened in the rest - frame uv than non - bal qsos , and hence this list of dlas is then cross - correlated with the bal catalogue of @xcite , in order to avoid such bals entering our analysis . furthermore , we have restricted ourselves to those sightlines towards qsos in the @xcite sample that only exhibit one single dla .
thus , we exclude the 59 respectively 5 cases , where two or three dlas are found along the same line of sight .
this leaves us with a sample of 526 sdss dr5 qso sightlines with securely identified dla well suited for the analysis of their dust content .
+ @xcite have surveyed dr7 , and present a list of 1426 strong hi absorbers at redshifts 2.15@xmath16 z @xmath16 5.2 , of which 937 systems have log n(hi @xmath17 20.3 . applying the same criteria for filtering out potentially problematic sightlines as above
, we arrive at 731 sdss dr7 qso sightlines for this dla sample .
while both lists have a substantial overlap ( 371 of prochaska s sources appear also in the noterdaeme list , albeit 31 of these are estimated there to fall below the log n(hi @xmath17 20.3 criterion ) , the two methods for finding dlas differ significantly enough to warrant keeping both datasets separate . it should , however , be kept in mind that the two samples are not independent from each other .
note also that the sdss treatment of spectroscopic calibration has improved from dr5 to dr7 , especially the spectrophotometric algorithms .
therefore , we have decided to strictly use only the versions of the spetra that have been processed by the dr7 pipeline .
figure [ imag_and_zem ] shows the distribution of both samples in qso emission redshift and the @xmath3 band magnitudes .
in addition , we show the underlying sdss qso lists ( cleaned of bals and the dlas from both lists , as detailed below ) from which we can draw our matching spectra . the insert shows the histogram of the absorber redshifts for the two samples .
the vast majority ( @xmath18 ) of the dlas are in the redshift range @xmath22.2 @xmath16 z@xmath19 3.3 .
there are a variety of different methods to derive estimates for the reddening of each qso spectrum , or the average reddening of the whole sample ( e.g. @xcite .
we follow closely the suggestion of @xcite and @xcite , who have analysed the dust reddening in mgii absorbers towards sdss qsos .
the basic principle is to define a sample of qsos , where for each dla quasar in question a close match in both emission redshift and @xmath3-band magnitude ( and hence absolute i - band magnitude ) is found , that does not contain a dla . as @xcite point out , this matched non - absorber qso spectrum will be of very similar data quality and hence sensitivity to similar levels of absorption at each wavelength , because of the regulation of exposure levels directly at the telescope and the exposure level check of all sdss spectra before they are admitted to the database .
+ given that there is a dispersion in the spectral slopes of qso of similar magnitude and redshift , and hence a difficulty in matching spectra one by one , the strength of the method proposed here relies upon averaging over an as large as possible set of qsos with individually different slopes to obtain a robust measure of the ( geometric ) _ mean _ spectra for both the samples of qsos containing dlas and those free of dlas .
sdss provides for both the largest homogeneous samples of dlas as well as the largest reservoir of qsos from which to draw comparison spectra . to find an appropriate match for a dla - containing qso spectrum , we search the sdss dr5 database after cleaning it from bals and
also the dlas found in both lists .
this resulted in a pool of 8371 qsos with redshifts 2.24@[email protected] , from which to draw the comparison spectra .
we minimise the difference of the match to the qso in the two - dimensional parameter space of the @xmath3 magnitude and emission redshift , defined as @xmath21 .
note that we have corrected the @xmath3 magnitudes in each case for galactic reddening .
the normalisation factors in the distance metric were chosen such that the number of outliers in both quantities are minimal , as detailed in the following paragraph .
+ figure [ delta_z_delta_i ] shows the distribution of deviations between the dla containing spectra and their best matches in redshift and brightness for the two samples .
the shape of the brightness distribution of the quasars in sdss within the redshift range probed here , introduces a systematic malmquist bias : in general there is a slightly higher chance to select a _ fainter _
match at _ higher _ redshift .
this effect is particularly pronounced in the brightest dla qsos , because the number of potential counterparts below @xmath3 mag @xmath1618.0 becomes too low to select an adequate match .
hence , we decided to exclude all dla sightlines from the following analyses that deviate more than 0.08 in @xmath3 mag and/or 0.007 in relative redshift difference ( dotted lines in the two panels of [ delta_z_delta_i ] ) .
this reduces the dr5 sample to 475 , and the dr7 sample to 676 dlas .
the mean @xmath22 band magnitude difference @xmath23(dla - match ) for these is -0.0029 , and -0.0022 mag , whereas the mean relative redshift differences are @xmath24(qso - match)/z = @xmath25 , and @xmath26 . + in principle , we could increase the number of matches for each dla spectrum to more than one counterpart .
however , the gain in s / n ( or reduced scatter ) for the ratio of dla / match , dominated then by the dla composite s noise , which can not be reduced , does not compensate for the loss in matching precision : the more matches are being culled from the pool , the more severe the bias becomes . of special importance here
are they increasing redshift differences that may distort the precision at which the composite match spectrum reproduces the shape of the hence , we have decided to utilise only the best match in each case .
+ the redshift distribution for the dla systems in conjunction with the sdss spectral coverage ( 3800 @xmath27 ) allows in principle the construction of a composite spectrum in the dlas restframe from @xmath2 600 to @xmath2 3000 . obviously , however , the flux distribution of each qso blueward of 1215 in the quasar restframe is strongly suppressed due to the neutral h column densities sampled inside the forest , in a manner specific to each sightline differently . hence
, in order to reliably determine the effects of dust reddening by comparison with an unreddened continuum , we have to restrict ourselves in each spectrum to the part redward of the beginning of the lyman forest .
we have therefore chosen to exclude all pixels in a spectrum that lie below 1240 in the qso restframe .
this introduces a cutoff in the restframe of the absorbers that varies from spectrum to spectrum individually depending on the combination of z@xmath4 and z@xmath28 . + geometric mean spectra
were then generated from the absorber and non - absorber sample in the following manner : we extend the sdss wavelength grid with a stepsize of 0.001 in log @xmath29 to the lowest wavelengths covered by the spectra in the dla restframe , and , as specified above , redward of the lyman forest of the qso itself .
for each wavelength data point @xmath30 in this grid we compute for each sightline separately which pixel in the original spectrum comes to lie closest when transforming by division through the absorber redshift , i.e. we take the nearest integer value of @xmath31 where @xmath32 denotes the wavelength of the first pixel in each spectrum , @xmath33 is the absorption redshift of the dla , and @xmath34 is the pixel number starting from the blue end of the spectrum . before computing the geometric mean of the flux distribution in the sample of the thus determined pixels , we clear it of those individual pixels flagged by the spectroscopic pipeline as possibly corrupted @xcite . we have made no attempt in masking additional pixels , e.g. at the location of known absorption at different redshifts than the dla .
there is also no weighing of pixels according to e.g. s / n , as this possibly distorts the power - law nature of the combined spectrum .
we have corrected the individual spectra for galactic reddening @xcite .
the geometric mean composite spectra for the non - absorber sets were generated in exactly the same way , using the redshift of the corresponding dla in each case for equation [ pixel_number ] .
note that we use in each case the spectra delivered by the dr7 pipeline , which includes the latest version of the spectrophotometric calibration .
the advantage of the _ geometric _ mean is that it yields a meaningful average of the extinction at each wavelength by essentially averaging optical depths .
+ note that our method does not rely upon any interpolation of the original spectra - at the price of sacrificing spectral resolution , because we bin together flux values that may reside in the dla restframe separated by 1/2 of a pixel at the extreme .
we have tested the effects of this by comparison of the resulting composite spectra to two cases of a more refined treatment : i. rebinning each spectrum after deredshifting it , and ii . computing the weighted average of the flux values of the two pixels in the original spectra that always straddle one wavelength element in the new dla restframe grid .
as expected , the differences over large wavelength ranges are negligible , and manifest themselves to a noticable degree only on a two to three pixel scale .
hence , we believe our method , being as close as possible to the data , is robust enough not to have to rely on the intricacies of interpolation as in case i. because we have to throw out pixels with fluxes f@xmath35 0.0 , method ii from above results in having to flag out twice as many datapoints , which is therefore deemed less acceptable despite the higher spectral resolution .
+ figure [ number_contributing ] shows the number of spectra contributing to each wavelength bin for the construction of the composite spectra in the dla restframe .
it also highlights to which location in these spectra the observed central wavelength of the @xmath3 band ( 7481 ) is shifted in the dla frames .
furthermore we have indicated the average brightness of the qsos contributing at each wavelength , and the average dla absorber redshift .
note that the region covered by the bulk of the spectra extends from @xmath21400 to @xmath22300 , and exhibits fairly constant values of the mean brightness and absorber redshifts .
the need for staying redward of the lyman @xmath0 emission , and the decreasing high ( restframe ) wavelength with increasing redshift of each qso , result of course in fewer spectra contributing to the composite towards the edges of the wavelength regime covered by our samples .
hence , these spectral areas carry less weight due to their increased noise / scatter in the reddening analyses ( as detailed in 3.3 . ) . + the geometric mean composite spectra of the 475 ( dr5 ) and 676 ( dr7 ) absorber spectra , and the corresponding non - absorbers are shown in the top panel of figures [ prochaska_spectrum ] and [ noterdaeme_spectrum ] .
note how the qsos emission lines for both samples are equally smeared out due to the different shifts applied for each specific combination of @xmath36 and @xmath37 , whereas only for the dla sample the narrow absorption lines appear in the expected places . because of the lower numbers of spectra contributing to each wavelength bin towards the blue and red end of the composite , the noise ( or scatter ) starts to increase .
the rise of the mean flux towards the red side is an effect of the qsos contributing to this part of the spectra being on average of lower redshift and brighter ( cf .
figure [ imag_and_zem ] ) .
additionally , the abrupt change in the flux beyond 2800 can be attributed to the stack becoming essentially the average of a very few spectra from this point on .
+ the lower panels of figures [ prochaska_spectrum ] and [ noterdaeme_spectrum ] show the ratios of the two composite spectra .
the cyan dotted line indicates the ratio expected assuming an extinction of @xmath5e(b - v)@xmath6 = 0.01 for an smc curve @xcite , and represents the upper limit on @xmath5e(b - v)@xmath6 derived by @xcite .
it is immediately obvious that very little room is left for the signature of _ any _ dust extinction in the dla sample .
we can quantitatively assess how much extinction can be tolerated by our data with the following procedure .
+ suppose an unknown dla stacked spectrum , @xmath38 is reddened by dust within the absorbers , @xmath39 . then
, the ratio of these two spectra would simply yield the extinction : @xmath40 .
the underlying assumption for our method is that the stacked spectrum of the matches is a good representation of the unreddened dla stack .
our matching scheme , however , leads us to an unreddened spectrum @xmath41 , that , by definition , has the same flux as the ( supposedly ) dust reddened dla spectrum at some fiducial wavelength @xmath42 .
hence , the ratio of the two composites in this case is modified by the extinction @xmath43 at this matching wavelength @xmath44 given an extinction law , the ratio becomes a function of e(b - v ) only : @xmath45 . here
, we have chosen the smc extinction curve of @xcite , and assume an @xmath46 as typical for the total - to - selective ratio .
note that the choice of the smc reddening curve is the most constraining regarding the absence of dust signatures - lmc or even mw dust type induces stronger reddening .
because we are fitting for the slope of this function , the potential offsets in @xmath43 are of no relevance in this circumstance , and hence the normalisation of the ratio curve is not important .
+ figure [ chi2 ] shows the results of these one - parameter fits for dlas ( this work ) and two sdss samples of mgii absorbers @xcite . to obtain the latter two , we have applied exactly the same averaging scheme as outlined above . before the calculation of the goodness of fit statistic , @xmath47 ( where @xmath48 denotes the degrees of freedom for the fit ) ,
we have taken out areas of @xmath75 from the known , narrow metal lines in the spectra .
we find that we can constrain the amounts of extinction for those two samples of mgii absorbers very tightly : @xmath5e(b - v)@xmath6 = [email protected] for the intervening , and @xmath5e(b - v)@xmath6 = [email protected] for the associated cases .
we therefore reproduce the results of @xcite and @xcite well , despite the fact that we do not explicitly also remove regions of the spectra that are affected by strong absorption at redshifts different from the mgii absorbers .
we take this test as a sign that both our stacking algorithms and the method to estimate the reddening are robust and accurate for these mgii samples .
we note , however , that the lower redshift of these absorbers ( compared to our samples of dlas ) may reduce the problem of the foreground reddening due to the lower redshift path sampled , and the intrinsically lower density of absorbers in the low - z universe .
+ we have also applied this spectral method to the samples of dlas and pool qsos used by vladilo et al .
( 2008 ) for their photometric analysis .
these authors have taken care to take out qsos that may be afflicted by absorption unrelated to the dlas .
again , we are able to reproduce their results very well , and derive an @xmath49 , indicating that both methods yield similar results when applied to the same samples . here , the strength of our method is specifically that all the unrelated effects will affect both the dla and the pool spectra in the same way , and hence we choose to increase the sample size to ensure that these effects are averaged out .
we are making use of both the largest databases for dlas _ and _ comparison spectra ( increasing the number of dlas by almost a factor of 3 for the dr7 sample , and the number of the objects in the comparison pool by a factor of 5 compared to the vladilo samples ) . + a word of caution is necessary when trying to quantify the confidence intervals for our results .
nominally , the number of degrees of freedom ( d.o.f . ) for the two samples are 2543 ( dr7 ) and 2583 ( dr7 ) , after taking out the pixels affected by metal - line absorption , and restricting ourselves to the wavelength region below 2500 to avoid the large drop in both numbers of spectra contributing and average qso brightness .
however , as we have detailed in the section on the construction of the composite spectra , adjacent pixels are not truly independent of each other .
we have tested the robustness of the fits by binning the composites ( with bin sizes up to 5 pixels , hence reducing the d.o.f . by the same factor ) .
hence , we arrive at measurements of the mean @xmath5e(b - v)@xmath6 = -0.0013@xmath7 0.0025 ( dr5 ) and @xmath5e(b - v)@xmath6 = [email protected] ( dr7 ) for the full samples , where the error budget adopts a generous 3@xmath50 level for the approach with the 5-pixel bin . + in order to test whether we could have detected an @xmath5e(b - v)@xmath6 distribution as predicted by @xcite and measured by @xcite , we have artificially reddened each dla quasar spectrum by a value drawn from a lognormal distribution of width 0.4 dex and mean of log @xmath51 , broadly consisted with those two studies .
figure [ artificial_reddening ] shows that we not only could clearly detect such a reddening , but also could retrieve the input value very accurately ( red curve in figure [ chi2 ] ) . in fact , the two @xmath52 curves for our spectral analyses and the fake reddening are excluded by several @xmath50 .
+ we have also performed a null test by comparing the composite spectra of the matched samples with composites created from the samples of the 2nd best , 3rd best , etc .... matching samples . as expected , the analysis of the ratios of these stacked spectra yields no detectable reddening , and the range of @xmath5e(b - v)@xmath6 values obtained by this procedure ranges from -0.002 @xmath16@xmath5e(b - v)@xmath6 @xmath160.002 for dr5 , and -0.0018 @xmath16@xmath5e(b - v)@xmath6 @xmath160.001 for dr7 .
hence we encompass the two measurements with the dla samples .
we note in passing that the main difficulty in matching the shape of the composites and their comparison counterparts lies in the strong emission lines of the qsos , especially blueward of @xmath21700 , where the fe - complexes have a profound influence on the flux above the continuum . indeed , when inspecting the ratio curves for the two samples , around that wavelength there seems to appear a kink in them , which may be caused by a mismatch in the emission lines .
this is the main driver towards these ( very low ) negative @xmath5e(b - v)@xmath6 values we obtain formally from the fitting . if we restrict ourselves to either the ratios blueward or redward of this feature , the fitting results in exact null values ( at the expense of leverage in wavelength ) .
+ one possible complication in matching the quasars luminosities based upon their emission redshifts and the observed i band magnitudes is that the latter could systematically underestimate the true continuum flux for the dla sample compared to their non - dla counterparts because of a higher fraction of strong metal absorption lines due to the presence of the dla .
a rough estimate of the size of this effect indicates that in sightlines where a very strong absorber ( of the order of 5 restframe equivalent width ) is present in the i band , it will result in a photometric measurement that is decreased from the normal qso brightness by 2% .
thus we may in such cases be selecting qsos as matches that are systematically too faint .
we have tested whether such a possible bias may influence our analysis of the dust reddening signal by creating samples of matches following exactly the procedures described above , but introducing an artificial offset @xmath53 in the i band magnitude by minimising @xmath54 .
even shifts as large as @xmath53=-0.2 magnitudes do not change the overall shape of the composite spectra , yielding almost exactly the same results for the reddening estimates when fitting for the slope of the ratio curves .
while the full samples of dla sightlines do not show any appreciable sign of dust reddening to a high degree of accuracy , it is still possible that there are certain subsets of dlas and/or sightlines that do exhibit an amount of wavelength dependent extinction , which is simply too small to be detected when averaging over such large datasets . in order to check for such potential trends
, we have divided the dla lists into subsamples according to various criteria concerning the dlas themselves ( namely absorption redshift , hi column density , and the ab- or presence of metal lines ) or the underlying qso sightlines ( apparent brightness or emission redshift ) . to keep a high enough number of dlas we have in almost all cases decided to simply split the full samples into two halves .
+ figure [ subsample_nh ] shows the results of a split ( for the dr5 sample ) into the highest ( left panel ) and lowest ( right panel ) hi column densities for the dla .
the mean of these subsamples are at log n@xmath55=20.91 , and log n@xmath55=20.42 . clearly , there is no difference in the fact that there is a null detection for both subsamples .
+ figure [ subsample_zabs ] shows how differentiating the samples according to the absorber redshifts affects the dust reddening analysis .
again , both for the low and the high redshift subsamples ( mean z@xmath28=2.63 , and 3.39 , respectively ) there is no reddening detectable .
note how the higher redshift sample leads to a lower wavelength cut - off , and due to the generally fainter spectra a slightly higher noise in the composite .
+ figure [ subsample_metals ] demonstrates that the dlas classified by @xcite as containing no visible narrow metal absorption lines in individual sdss spectra , do also not exhibit any such signs in their composite .
only 88 of the 475 dlas fall into this class ( 18.5% ) , and hence the noise in the composite is much higher than in the other subsamples . yet , both classes do not show reddening .
+ the separation into qsos of low and high redshifts ( shown in figure [ subsample_zem ] , mean z@xmath28=2.31 and 3.78 ) is similar to the cut in qso brightness ( mean @xmath3 band psf magnitude of @xmath3=18.62 , and 19.48 mag , corrected for galactic reddening ) , because of the trend for fainter sources with higher redshift .
the subsample with the brightest qsos ( for dr5 ) is the only one that does show a positive value of the colour excess , but not signifcantly different from a null detection : @xmath5e(b - v)@xmath6 = [email protected] . +
to summarise , none of these simple subselections do give rise to the assumption that there are significant trends in reddening regarding either the observational constraints ( brightness and emission redshift of sources ) or dla properties ( absorption redshift , additional narrow metal line absorption , and hi column density ) .
while our analysis technique reproduces well past results for other classes of absorbers at different redshifts @xcite , and dlas using other methods @xcite and @xcite where some extinction could be detected , we could no detect any significant extinction in samples of dlas drawn from the data releases 5 and 7 of the sdss .
these findings are based on the largest dla sample ever studied , but also on the largest sample of quasars from which to draw a reference spectrum .
the implication is that most of the dlas observed in sdss have a very small dust content .
+ @xcite have analysed sdss spectra with intervening strong mgii absorbers , and detect clear dust reddening . more specifically , their subsamples with the highest mgii equivalent width ( subsamples 8 and 25 in their notation , table a4 ) show the highest degrees of reddening ( @xmath5e(b - v)@xmath6 = 0.032 and 0.085 ) .
these strong mgii absorbers are then classified , based on the column densities of the accompanying metal lines , as akin to low redshift dlas .
we note however , that a direct measurement of the hi absorption in these absorbers was not possible due to the low wavelength cutoff of sdss , and the mean hi column densities for the subsamples are derived by using an smc relationship between @xmath5e(b - v)@xmath6 and hi .
@xcite also caution that not all high ew mgii absorbers are indeed dlas , although @xcite point out that the specific requirement of the qsos probed by @xcite with hst being uv - bright introduces a bias against red qsos , that may have lead them to an underestimate of the dla fraction .
hence , it is a priori not clear whether our results contradict these findings .
also note that there may be a redshift evolution effect .
in fact , @xcite have noted that the rest - frame colour excess e(b - v ) for their mgii samples shows a strong decline with increasing redshift ( their figure 10 , lower right panel ) .
thus , it is possible that the difference in the mean redshifts of our samples ( [email protected] ) compared to the lower [email protected] of @xcite may partly explain the seemingly discrepant results .
+ yet another strategy involving spectral analysis is fitting the qso continuum spectra by a power - law , and checking for deviations in the slope of these for spectra with and without dlas . employing this technique
, @xcite can place an upper limit for a sample of 70 dlas from sdss dr2 of @xmath5e(b - v)@xmath6 = 0.01 .
note that we can rule out such a high value for our samples with a very high degree of confidence .
+ in addition , there have been claims of firm detections of dust reddening towards sightlines afflicted by intervening dlas .
specifically , @xcite report a mean e(r - z)=27(@xmath79)x10@xmath13 for a subsample of the same dr5 dla sample that we use .
when applying our method to these subsamples of dlas _ and _ the subsample of dr5 pool qsos for the constructing of the comparison spectrum , we recover their results .
note that @xcite have removed all quasars from their samples that may have additional absorption unrelated to the dla , reducing the number of objects in each sample by a large fraction .
we conclude that the differences do not reside in the methods , but the sample selection .
we note that the reddening distribution of undesired interlopers probably has a tail with a small number of objects exhibiting much larger than the typical reddening expected from dla systems .
it may hence be that for such low frequency events , that cases of e.g. bal qsos , despite efforts to filter them out , enter the dla samples more often than the comparison pool , and hence could wash out the signal .
increasing the sample size reduces the possible impact of such special , very red qsos .
our result is also at odds with expectations from the evolution of high - redshift galaxies ( but see @xcite .
it should be emphasised however that there are two separate issues : i ) what is the dust content of the quasar absorbers we know of today and ii ) which fraction of the quasar absorbers are missed because of the extinction they induce on their background quasar such that it is not selected in the first place .
indeed , models of galactic chemical evolution ( e.g. @xcite predict the existence of regions of high metallicity and column density which are mostly not detected in current dla surveys @xcite .
one way to reconcile these two findings is to invoke a dust - bias in the observations .
these apparent discrepancies could indeed be explained by the dust content of the absorbers themselves : they could absorb enough light from the quasar to make the latter too faint to be included in current quasar samples , which barely go beyond magnitude [email protected] .
in fact , an extension of the @xcite calculations taking into account recent observations addresses both these issues : @xcite propose that whilst the dust content of the dlas in current samples is not high , the missing fraction is possibly quite important ranging from 30% to 50% at @xmath56 .
one way to to test such a hypothesis is to look at fainter objects than what can be probed with sdss .
using a bayesian analysis of these issues in light of different techniques and datasets @xcite posit that probably less than 10% of dlas have been missed in current optical samples , and also derive an @xmath5e(b - v)@xmath6 posterior probability distribution peaking strongly around log @xmath5e(b - v)@xmath6 @xmath57 .
+ on the other hand , if the absence of the detection of extinction in dlas is not due to an observational bias , then selecting in absorption implies that we are picking a population of high - redshift galaxies which have probably undergone little star formation , e.g. @xcite and @xcite .
this is difficult to reconcile with the fact that these system are gas - rich by construction .
this work has benefitted from support of the `` agence nationale de la recherche '' with reference anr-08-blan-0316 - 01 .
we would like to thank bruno milliard , jean - michel - deharveng , jochen liske , and michael murphy for helpful comments refining the analysis substantially .
we thank the referee for a thorough review of the results and valuable suggestions , that improved the mansucript .
+ funding for the sdss and sdss - ii has been provided by the alfred p. sloan foundation , the participating institutions , the national science foundation , the u.s .
department of energy , the national aeronautics and space administration , the japanese monbukagakusho , the max planck society , and the higher education funding council for england .
the sdss web site is http://www.sdss.org/. + the sdss is managed by the astrophysical research consortium for the participating institutions .
the participating institutions are the american museum of natural history , astrophysical institute potsdam , university of basel , university of cambridge , case western reserve university , university of chicago , drexel university , fermilab , the institute for advanced study , the japan participation group , johns hopkins university , the joint institute for nuclear astrophysics , the kavli institute for particle astrophysics and cosmology , the korean scientist group , the chinese academy of sciences ( lamost ) , los alamos national laboratory , the max - planck - institute for astronomy ( mpia ) , the max - planck - institute for astrophysics ( mpa ) , new mexico state university , ohio state university , university of pittsburgh , university of portsmouth , princeton university , the united states naval observatory , and the university of washington . , s. , panagia , n. , & stiavelli , m. 2000 , in astronomical society of the pacific conference series , vol . 215 , cosmic evolution and galaxy formation : structure , interactions , and feedback , ed
. j. franco , l. terlevich , o. lpez - cruz , & i. aretxaga , 65+ | we searched for evidence of the reddening of background qso spectra from the sloan digital sky survey ( sdss ) due to dust in intervening damped lyman @xmath0 ( dla ) systems .
we utilise the two data releases dr5 and dr7 to arrive at sample sizes of 475 ( dr5 ) and 676 ( dr7 ) absorbers , based on two different published lists of sdss dlas .
both samples span roughly the redshift range of @xmath1 , with a mean of [email protected] , and the majority of the dlas ( 75% ) below z=3.3 .
we construct geometric mean spectra in the absorber restframes ranging from 1240 to @xmath2 2800 , and composite spectra of samples matching the dla qsos in @xmath3 band magnitude and emission redshift z@xmath4 , but without absorption lines . by comparing the slopes of these composite spectra with their matched counterparts , we find no sign of reddening in the ensemble of the absorbers from these samples . owing to both the unprecedently large sizes of the dla samples themselves and the non - dla sdss qso sample ,
from which we can draw our matching spectra , we can place very tight limits for this non - detection ( @xmath5e(b - v)@xmath6 = -0.0013@xmath7 0.0025 ( dr5 ) and @xmath5e(b - v)@xmath6 = [email protected] ( dr7 ) .
interestingly , when applying our technique to the samples of york et .
al . ( 2006 ) , vandenberk et al . (
2008 ) ( intervening and intrinsic mgii absorbers ) and the smaller dla - subsample and pool of comparison qsos of vladilo et al . ( 2008 ) , we do recover their results , i.e. detect the same amount of reddening as these authors do .
furthermore , we have tested whether subsamples of our large sample in categories involving the absorbers ( hi column densities , presence or absence of accompanying metal absorption , absorber redshift ) or the background quasars ( emission redshift , brightness ) do reveal dust extinction , but found no trends .
these results are at odds with both detections of dust reddening from previous studies , and also with expectations from observations of high - redshift galaxies .
[ firstpage ] quasar absorption lines - damped lyman alpha systems - dust |
[ [ summary - of - the - contents . ] ] summary of the contents .
+ + + + + + + + + + + + + + + + + + + + + + + + in @xcite mostovoy showed ( modulo certain errors that were corrected in @xcite ) that if @xmath6 then the space @xmath7 of holomorphic maps from @xmath8 to @xmath9 has the same homotopy type of the space @xmath10 of coresponding continuous maps up to a certain dimension , which generalizes the classical result of segal @xcite for @xmath5 .
in @xcite we used a variant of mostovoy s method to the analogous problem for the space @xmath11 of algebraic maps from @xmath12 to @xmath13 for @xmath14 .
these our results can also be seen as generalizations of the results of @xcite and @xcite . in @xcite
we used analogous methods to prove a homotopy ( or homology ) approximation theorem for the space @xmath15 of real algebraic maps from @xmath12 to @xmath9 when @xmath2 . combing this result with the main theorem of @xcite
, we obtained a @xmath16-equivariant homotopy approximation result in @xcite ( where the @xmath16-action on @xmath9 is induced by complex conjugation ) , which is itself a generalization of ( * ? ? ? * theorem 3.7 ) . in his recent paper @xcite , mostovoy ,
in addition to correcting the mistakes in @xcite , introduced a new idea that allowed him to improve the bounds on the degree of homotopy groups in his homotopy approximation theorem . with the help of analogous methods and some additional techniques of ours , in the recent work @xcite
, we improved the bounds of the homotopy approximation theorem given in @xcite .
the first purpose of this article is to obtain a improved version of the equivariant homotopy approximation theorem of @xcite . since the arguments are analogous to those in @xcite , we only state our new results and refer to the references for detailed proofs ( cf .
remark [ remark : proof ] below ) .
the second purpose is to consider the possibility of approximating the space of continuous maps by its subspaces of algebraic maps of a fixed degree , analogously to the results of segal @xcite and mostovoy @xcite . because the situation is analogous , in this paper
we only consider the space of algebraic maps from @xmath12 to @xmath13 .
for this purpose we only need to prove that the projection maps @xmath17 are homotopy equivalences .
it is easy to see that both of these maps have contractible fibres , however that alone is not sufficient to prove that the maps are homotopy equivalencies .
in fact , in ( * ? ? ? * conjecture 3.8 ) , we conjectured that this is true in general . here
we provide an informal argument that shows that this is true for @xmath5 .
a proof of this fact already appears in ( * ? ? ?
* proposition 2.1 ) .
however we find that the argument given there is unconvincing since it does not seem to make any use of any properties of the map @xmath18 ( e.g. the fact that is a quasi - fibration ) . in the remainder of this section
we briefly describe the notation and the definitions .
[ [ notation . ] ] notation .
+ + + + + + + + + the notation of this paper is essentially analogous to the one used in @xcite and @xcite .
note that the presence of @xmath19 indicates that a complex case is being considered ( i.e. maps take values in @xmath9 rather than @xmath13 or polynomials have coefficients in @xmath19 rather than @xmath20 , etc . ) .
let @xmath0 and @xmath1 be positive integers such that @xmath21 .
we choose @xmath22\in { \bbb r\mbox{{\rm p}}}^m$ ] and @xmath23\in\bbb { \bbb c { \rm p}}^n$ ] as the base points of @xmath12 and @xmath24 , respectively .
let @xmath25 denote the space consisting of all based maps @xmath26 .
when @xmath27 , we denote by @xmath28 the corresponding path component of @xmath29 for each @xmath30 ( @xcite ) .
similarly , let @xmath31 denote the space of all free maps @xmath32 and @xmath33 the corresponding path component of @xmath31 .
a map @xmath34 is called a _ algebraic map of degree _ @xmath35 if it can be represented as a rational map of the form @xmath36 $ ] such that @xmath37 $ ] are homogeneous polynomials of the same degree @xmath35 with no common _ real _ roots except @xmath38 .
we denote by @xmath39 ( resp .
@xmath40 ) the space consisting of all ( resp . based )
algebraic maps @xmath34 of degree @xmath35 .
it is easy to see that there are inclusions @xmath41_2}({\bbb
r\mbox{{\rm p}}}^m,{\bbb c { \rm p}}^n ) , \quad { \mbox{{\rm alg}}}^*_d({\bbb r\mbox{{\rm p}}}^m,{\bbb c { \rm p}}^n)\subset { \mbox{{\rm map}}}^*_{[d]_2}({\bbb r\mbox{{\rm p}}}^m,{\bbb c { \rm p}}^n),\ ] ] where @xmath42_2\in{\bbb z}/2=\{0,1\}$ ] denotes the integer @xmath35 mod @xmath43 .
let @xmath44 denote the space consisting of all @xmath45-tuples @xmath46^{n+1}$ ] of homogeneous polynomials of degree @xmath35 with coefficients in @xmath19 and without non - trivial common real roots ( but possibly with non - trivial common _ non - real _ ones ) .
since @xmath47 acts on @xmath48 freely , one can define the projectivisation @xmath49 by the orbit space @xmath50 .
let @xmath51 be the subspace consisting of all @xmath45-tuples @xmath52 such that the coefficient of @xmath53 in @xmath54 is 1 and @xmath55 in the other @xmath56 s @xmath57 ) .
then there are natural projection maps @xmath58 for @xmath27 and @xmath59 a fixed algebraic map , we denote by @xmath60 and @xmath61 the spaces defined by @xmath62_2}^*({\bbb r\mbox{{\rm p}}}^m,{\bbb c { \rm p}}^n):f\vert { \bbb r\mbox{{\rm p}}}^{m-1}=g\}. \end{cases}\ ] ] it is well - known that there is a homotopy equivalence @xmath63 ( @xcite ) .
let @xmath64 denote the space of all homogenous polynomials @xmath65 $ ] of degree @xmath35 .
we choose a fixed tuple @xmath66 such that @xmath67 .
in this situation , we denote by @xmath68 the subspace given by @xmath69 , and define the subspace @xmath70 by @xmath71 because @xmath72 for any @xmath73 , one can define the projection @xmath74 by the restriction @xmath75 let @xmath76_2}^*({\bbb r\mbox{{\rm p}}}^m,{\bbb c { \rm p}}^n ) , \\
j_{d,\bbb c}:{\mbox{{\rm alg}}}_d({\bbb r\mbox{{\rm p}}}^m,{\bbb c { \rm p}}^n)\stackrel{\subset}{\rightarrow } { \mbox{{\rm map}}}_{[d]_2}({\bbb r\mbox{{\rm p}}}^m,{\bbb c { \rm p}}^n ) , \\
i_{d,\bbb c}^{{\prime}}:{\mbox{{\rm alg}}}_d^{\bbb c}(m , n;g)\stackrel{\subset}{\rightarrow } f(m , n;g)\simeq \omega^m{\bbb c { \rm p}}^n \end{cases}\ ] ] denote the inclusions and @xmath77_2}^*({\bbb r\mbox{{\rm p}}}^m,{\bbb c { \rm p}}^n ) , \\ j_d^{\bbb c}=j_{d,\bbb c}\circ \gamma_d^{\bbb c}:\tilde{a}_d^{\bbb c}(m , n)\to { \mbox{{\rm map}}}_{[d]_2}^*({\bbb r\mbox{{\rm p}}}^m,{\bbb c { \rm p}}^n ) , \\
i_d^{{\prime}{\prime}}=i^{{\prime}}_{d,{\bbb c}}\circ { \psi_d^{{\bbb c}}}^{{\prime } } : a_d^{{\bbb c}}(m , n;g ) \to f(m , n;g ) \end{cases}\ ] ] the natural maps .
the notations used in this paper can be summarized in the following two diagrams , where @xmath78 denotes a fixed based algebraic map of degree @xmath35 and we omit their details of the notations in ( [ diagramr ] ) and refer the reader to @xcite .
@xmath79_2}({\bbb r\mbox{{\rm p}}}^m,{\bbb c { \rm p}}^n ) & { \mbox{{\rm map}}}^*_{[d]_2}({\bbb r\mbox{{\rm p}}}^m,{\bbb c { \rm p}}^n ) \ar@{_{(}->}[l ] & f^{{\bbb c}}(m , n;g ) \ar@{_{(}->}[l ] \\
{ \mbox{{\rm alg}}}_d({\bbb r\mbox{{\rm p}}}^m,{\bbb c { \rm p}}^n ) \ar@{^{(}->}[u]^{j_{d,{\bbb c } } } & { \mbox{{\rm alg}}}_d^*({\bbb r\mbox{{\rm p}}}^m,{\bbb c { \rm p}}^n ) \ar@{^{(}->}[u]^{i_{d,{\bbb c } } } \ar@{_{(}->}[l ] & { \mbox{{\rm alg}}}^{{\bbb c}}_d(m , n;g ) \ar@{^{(}->}[u]^{i_{d,{\bbb c}}^{{\prime } } } \ar@{_{(}->}[l]\\ \tilde{a}_d^{{\bbb c}}(m , n ) \ar@{->>}[u]^{\gamma^{{\bbb c}}_{d } } \ar@/^4pc/[uu]^>>>>>>{j_d^{{\bbb c } } } & a_d^{{\bbb c}}(m , n ) \ar@{->>}[u]^{\psi^{{\bbb c}}_d } \ar@/^4pc/[uu]^>>>>>>{i_d^{{\bbb c } } } & a_d^{{\bbb c}}(m , n;g ) \ar@{- > > } [ u]^{{\psi_d^{{\bbb c}}}^{{\prime } } } \ar@/^4pc/[uu]^>>>>>>{i_d^{{\prime}{\prime } } } } \ ] ] @xmath80_2}({\bbb r\mbox{{\rm p}}}^m,{\bbb r\mbox{{\rm p}}}^n ) & { \mbox{{\rm map}}}^*_{[d]_2}({\bbb r\mbox{{\rm p}}}^m,{\bbb r\mbox{{\rm p}}}^n ) \ar@{_{(}->}[l ] & f(m , n;g ) \ar@{_{(}->}[l ] \\ { \mbox{{\rm alg}}}_d({\bbb r\mbox{{\rm p}}}^m,{\bbb r\mbox{{\rm p}}}^n ) \ar@{^{(}->}[u]^{j_{d,{\bbb r } } } & { \mbox{{\rm alg}}}_d^*({\bbb r\mbox{{\rm p}}}^m,{\bbb r\mbox{{\rm p}}}^n ) \ar@{^{(}->}[u]^{i_{d,{\bbb r } } } \ar@{_{(}->}[l ] & { \mbox{{\rm alg}}}_d(m , n;g ) \ar@{^{(}->}[u]^{i_{d,{\bbb r}}^{{\prime } } } \ar@{_{(}->}[l]\\ \tilde{a}_d(m , n ) \ar@{->>}[u]^{\gamma_d } \ar@/^4pc/[uu]^>>>>>>{j_d } & a_d(m , n ) \ar@{->>}[u]^{\psi_d } \ar@/^4pc/[uu]^>>>>>>{i_d } & a_d(m , n;g ) \ar@{->>}[u]^{\psi_d^{{\prime } } } \ar@/^4pc/[uu]^>>>>>>{i_d^{{\prime } } } } \ ] ]
in this section we state the main results of this paper .
first define the positive integers @xmath81 and @xmath82 by @xmath83 where @xmath84 is the integer part of a real number @xmath85 and we remark that the equality @xmath86 holds .
we first recall the following two results .
[ thm : ky1 ] if @xmath87 , the natural map @xmath88_2}({\bbb r\mbox{{\rm p}}}^1,{\bbb r\mbox{{\rm p}}}^n)\simeq \omega s^n$ ] is a homotopy equivalence up to dimension @xmath89 [ thm : ky4 ] let @xmath0 and @xmath1 be positive integers such that @xmath14 . 1 .
let @xmath90 be an algebraic map of degree @xmath35 .
then the natural map @xmath91 is a homotopy equivalence up to dimension @xmath92 if @xmath93 and a homology equivalence up to dimension @xmath92 if @xmath94 .
2 . the natural maps @xmath95_2}^*({\bbb r\mbox{{\rm p}}}^m,{\bbb r\mbox{{\rm p}}}^n ) \\
j_d:\tilde{a}_d(m , n)\to { \mbox{{\rm map}}}_{[d]_2}({\bbb r\mbox{{\rm p}}}^m,{\bbb r\mbox{{\rm p}}}^n ) \end{cases}\ ] ] are homotopy equivalences up to dimension @xmath92 if @xmath93 , and homology equivalences up to dimension @xmath92 if @xmath94 . [ remark : aky1 ] ( i ) the above theorem was recently proved in @xcite and is an improvement of the main result of @xcite .
\(ii ) a map @xmath96 is called _ a homotopy _
( resp . _ a homology _ ) _ equivalence up to dimension _ @xmath97
if @xmath98 ( resp .
@xmath99 ) is an isomorphism for any @xmath100 and an epimorphism for @xmath101 .
\(iii ) let @xmath102 be a finite group and let @xmath96 be a @xmath102-equivariant map .
then a map @xmath96 is called _ a @xmath102-equivariant homotopy _ ( resp . _ homology _ ) _ equivalence up to dimension _
@xmath97 if the restriction map @xmath103 is a homotopy equivalence ( resp .
homology ) equivalence up to dimension @xmath97 for any subgroup @xmath104 .
now we state the main results of this paper as follows .
[ thm : ky5-i ] let @xmath0 and @xmath1 be positive integers such that @xmath2 . 1 .
let @xmath105 be a fixed algebraic map of degree @xmath35 .
the natural map @xmath106 is a homotopy equivalence up to dimension @xmath107 if @xmath108 and a homology equivalence up to dimension @xmath107 if @xmath109 .
2 . the natural maps @xmath110_2}({\bbb r\mbox{{\rm p}}}^m,{\bbb
c { \rm p}}^n ) \\
j_d^{{\bbb c}}:\tilde{a}_d^{{\bbb c}}(m , n)\to { \mbox{{\rm map}}}_{[d]_2}({\bbb r\mbox{{\rm p}}}^m,{\bbb c { \rm p}}^n ) \end{cases}\ ] ] are homotopy equivalences up to dimension @xmath107 if @xmath108 and homology equivalences up to dimension @xmath107 if @xmath109 . note that the complex conjugation on @xmath4 naturally induces @xmath3-actions on the spaces @xmath111 and @xmath112 . in the same way
, it also induces a @xmath3-action on @xmath9 and it extends to @xmath3-actions on the spaces @xmath113 and @xmath114 , where we identify @xmath115 and regard @xmath12 as a @xmath3-space with the trivial @xmath3-action . since @xmath116 , @xmath117 and @xmath118 , by theorem [ thm : ky4 ] and theorem [ thm : ky5-i ]
we obtain the following result .
[ cor : ky5-ii ] let @xmath0 and @xmath1 be positive integers such that @xmath14 . 1 .
let @xmath59 be an algebraic map of degree @xmath35 .
the natural map @xmath106 is a @xmath3-equivariant homotopy equivalence up to dimension @xmath92 if @xmath93 and a @xmath3-equivariant homology equivalence up to dimension @xmath92 if @xmath94 .
the natural maps @xmath110_2}({\bbb
r\mbox{{\rm p}}}^m,{\bbb c { \rm p}}^n ) \\
j_d^{{\bbb c}}:\tilde{a}_d^{{\bbb c}}(m , n)\to { \mbox{{\rm map}}}_{[d]_2}({\bbb r\mbox{{\rm p}}}^m,{\bbb c { \rm p}}^n ) \end{cases}\ ] ] are @xmath3-equivariant homotopy equivalences up to dimension @xmath92 if @xmath93 and are @xmath3-equivariant homology equivalences up to dimension @xmath92 if @xmath94 . [ remark : proof ] remark that @xmath119 for @xmath120 .
thus , we may regard theorem [ thm : ky5-i ] and corollary [ cor : ky5-ii ] as the improvements of ( * ? ? ?
* theorems 1.4 , 1.5 and corollary 1.7 ) .
the method of the proof of theorem [ thm : ky5-i ] consists of applying the ideas used in the proof of theorem [ thm : ky4 ] to the argument of the proof of ( * ? ? ?
* theorems 1.4 and 1.5 ) .
we shall therefore omit the details and refer the reader to @xcite and @xcite .
there is a mistake in the statement of ( * ? ? ?
* corollary 1.7 ) ; the condition @xmath2 should be replaced by @xmath14 as in corollary [ cor : ky5-ii ] .
in this section , we consider the following conjecture stated in @xcite . it is easy to see that @xmath121 is a homotopy equivalence if @xmath18 is so .
thus , we only consider the projection @xmath18 .
there is much evidence that suggests that @xmath18 is a homotopy equivalence .
indeed , the fibre of it over @xmath125 is homeomorphic to the space of everywhere positive @xmath126-coefficient polynomials in @xmath127-variables of degree @xmath128 with leading coefficient @xmath129 , which is convex and contractible .
if it is a quasi - fibration , it is a homotopy equivalence .
although we can not prove this in general , we can do so it for @xmath5 .
_ we will exploit the convenient fact that spaces of tuples of polynomials in one variable can be identified with certain configuration spaces of points or particles in the complex plane .
more exactly , we identify the space @xmath130 with the space of @xmath35 particles of each of @xmath131 different colours , in the case @xmath132 , say , red , blue and yellow , located in @xmath133 symmetrically with respect to the real axis and such that no three particles of different colour lie at the same point on the real axis .
note that off the real axis the particles are completely unrestricted .
the space of algebraic maps @xmath130 can also be thought of as a configuration space of @xmath134 particles of each of @xmath131 different colours as above , where @xmath135 , but with the additional property that when @xmath131 different particles meet ( off the real line ) they ( and their conjugate particles ) disappear ( this space is topologised as the obvious quotient of the preceding one ) . finally , we need one more configuration space , introduced in @xcite .
let @xmath136 denote the space of no more than @xmath137 particles of colour @xmath138 , where @xmath139 mod @xmath43 , on the real axis , with the property that any even number of particles of the same colour at the same point on the real axis vanish , and of course , as before no @xmath131 particles of different colours lie at the same point .
in other words , @xmath136 is a configuration space modulo 2 .
there is a natural map @xmath140 which factors through @xmath18 , @xmath141 in the fibre over this configuration , all points in the upper half plane are sent linearly to the fixed point @xmath145 and those in the lower half plane to @xmath146 . for a @xmath128 or @xmath147 fold particle lying on the real line ,
@xmath134 of the particles are moved to @xmath148 and @xmath134 are moved to to @xmath146 leaving @xmath55 or @xmath129 particles in place .
this argument shows that the fibres of both @xmath149 and @xmath142 are contractible .
next , we will show that the maps @xmath144 and @xmath142 are both quasi - fibrations .
consider a point ( configuration of particles ) in @xmath136 . by a singular sub - configuration for colour @xmath138 we mean a collection @xmath1 particles of distinct colours other than the @xmath138-th colour which are all located at the same point .
intuitively , we think of a singular configuration for the @xmath138-th colour as an obstacle through which a particle of the @xmath138-th colour can not pass . by a singular sub - configuration ( without specifying the colour ) we mean a singular sub - configuration for any colour . clearly , it is easy to see that the topology of the fibre over a particular configuration in @xmath136 is determined by two quantities : the number of particles of each colour in @xmath136 and the number of singular sub - configurations ( for each colour ) .
hence , it the proof we will consider filtrations with respect to the number of particles of each colour and the number of singular sub - configurations . for non - negative integers @xmath150 with @xmath151 , let @xmath152 denote the subspace of @xmath136 consisting of configurations with precisely @xmath153 particles of the @xmath138-th colour .
we start by proving that the restrictions of the maps @xmath149 and @xmath142 to the pre - images of @xmath152 are quasi - fibrations . for any integer @xmath154 ,
let @xmath155 denote the subspace of @xmath152 consisting of configurations containing exactly @xmath134 singular sub - configurations .
it is easy to show that the restriction of @xmath144 and @xmath142 to the pre - image of @xmath155 is a locally trivial fibre bundle .
now we filter the space @xmath152 by closed subspaces @xmath156 of configurations containing @xmath157 singular sub - configurations
. set theoretic differences between these spaces are the spaces @xmath155 over which the maps are locally trivial fibre bundles and hence quasifibrations .
now we apply the dold - thom criterion ( * ? ? ?
* lemma 4.3 ) . to do so we have to construct an open neighbourhood of @xmath158 in @xmath159 , a deformation of this neighbourhood onto @xmath158 , together with a corresponding covering neighbourhood and a covering deformation required by the dold - thom criterion ( since all the fibres are contractible the condition that the induced maps on the fibres are homotopy equivalences is automatically satisfied ) .
such neighbourhoods and deformations are easy to describe intuitively .
the set of points of the required neighbourhood of @xmath158 consists of the points of @xmath158 together with those configurations in @xmath159 with at least one non - singular sub - configuration of @xmath1 particles of different colours contained in a sufficiently small interval . here
sufficiently small refers to the requirement that the particles in this sub - configuration be much nearer to each other than they are to any other particle and that the length of the minimal interval containing the sub - configuration be much less than the length of any interval containing a collection of @xmath131 particles of different colour .
the deformation can now be defined by introducing a force of attraction between particles of different colour ( e.g. a force field satisfying an inverse - square law ) . by induction on @xmath134
we show that the maps @xmath144 and @xmath142 are quasifibrations over @xmath160 .
we now fix @xmath161 and filter the space @xmath136 according to the number of points of the @xmath131-th colour .
the set theoretic differences between the terms of the filtration are precisely the spaces @xmath152 and we have already proved that the restriction of the maps @xmath149 and @xmath142 to the inverse images of these spaces are quasi - fibrations . we apply again the dold - thom criterion . for this purpose
we need to construct open neighbourhoods of spaces of configurations with no more than @xmath162 particles of the last colour in the space of configurations of no more than @xmath134-particles of the last colour .
the method is again analogous to the one we used earlier .
our deformation will pull together pairs of particles of the last colour which are very close by means of a gravitational force field between particles of the last colour .
for this purpose we must avoid hitting a singular sub - configuration .
again , it is easy to see that we can choose open neighbourhoods and deformations with the right properties .
this proves that the maps restricted to the pre - images of spaces with the number of particles of the first @xmath163 colours fixed are quasifibrations .
now we filter these spaces according to the number of particles of the last but one colour .
proceeding by induction we see that @xmath144 and @xmath142 are quasi - fibrations . 99 adamaszek , a. , kozlowski , a. , and yamaguchi , k. , spaces of algebraic and continuous maps between real algebraic varieties , _
@xmath164to appear in@xmath165 quart .
_ ( published online in q.j.m .
doi:10.1093/qmath / haq029 ) ( cf .
math.at ; arxiv : 0809.4893 ) .
boyer , c.p . ,
hurtubise , j.c . , and milgram , r.j . , stability theorems for spaces of rational curves , _ int . j. math . _
* 12 * ( 2001 ) , 223 - 262 .
bochnak , j. , coste , m. , and roy , m - f . , real algebraic geometry , a series of modern surveys in math . * 36 * , springer - verlag , 1991 .
cohen , r.l . ,
jones , j.d.s . , and segal , g.b . , stability for holomorphic spheres and morse theory , _ contemporary math . _ * 258 * ( 2000 ) , 87106 . crabb , m.c . , and sutherland , w.a . , function spaces and hurewicz - radon numbers , _ math .
_ * 55 * ( 1984 ) , 6790 .
guest , m.a . ,
kozlowski , a. , and yamaguchi , k. , spaces of polynomials with roots of bounded multiplicity , _ fund . math . _
* 116 * ( 1999 ) , 93117 .
hatcher , a. , algebraic topology , cambridge university press , 2001 .
kozlowski , a. , and yamaguchi , k. , topology of complements of discriminants and resultants , _
* 52 * ( 2000 ) , 949959 .
kozlowski , a. , and yamaguchi , k. , spaces of algebraic maps from real projective spaces into complex projective spaces , _ contemporary math .
_ * 519 * ( 2010 ) , 145164 .
kozlowski , a. , and yamaguchi , k. , simplicial resolutions and spaces of algebraic maps between real projective spaces , preprint .
mostovoy , j. , spaces of rational loops on a real projective space , _ trans .
* 353 * ( 2001 ) , 19591970 .
mostovoy , j. , spaces of rational maps and the stone - weierstrass theorem , _ topology _ * 45 * ( 2006 ) , 281293 .
mostovoy , j. , truncated simplicial resolutions and spaces of rational maps , _
@xmath164to appear in@xmath165 quart .
_ ( published online in q.j.m .
doi:10.1093/qmath / haq031 ) .
sasao , s. , the homotopy of @xmath166 , _ j. london math . soc . _
* 8 * ( 1974 ) , 193197 .
segal , g.b .
, the topology of spaces of rational functions , _ acta math . _ * 143 * ( 1979 ) , 3972 .
vassiliev , v.a .
, complements of discriminants of smooth maps , topology and applications , amer . math . soc . , translations of math .
monographs * 98 * , 1992 ( revised edition 1994 ) .
yamaguchi , k. , complements of resultants and homotopy types , _ j. math .
kyoto univ .
_ * 39 * ( 1999 ) , 675684 . | we study the homotopy types of certain spaces closely related to the spaces of algebraic ( rational ) maps from the @xmath0 dimensional real projective space into the @xmath1 dimensional complex projective space for @xmath2 ( we conjecture this relation to be a homotopy equivalence ) . in @xcite
we proved that natural maps of these spaces to the spaces of all continuous maps are @xmath3-equivariant homotopy equivalences , where the @xmath3-equivariant action is induced from the conjugation on @xmath4 . in the same article we also proved that the homotopy types of the terms of the natural degree filtration approximate closer and closer the homotopy type of the space of continuous maps and obtained bounds that describe the closeness of the approximation in terms of the degrees of the maps . in this paper , we improve the estimates of the bounds by using new methods invented in @xcite and used in @xcite . in addition ,
in the the last section , we prove a special case ( @xmath5 ) of the conjecture stated in @xcite that our spaces are homotopy equivalent to the spaces of algebraic maps . |
the analyticity of local @xmath0-smooth cr diffeomorphisms between two essentially finite generic real analytic submanifolds of @xmath1 is established in @xcite , as a kind of reflection principle , provided that all the components of the cr diffeomorphism extend holomorphically to a fixed wedge ( the so - called notion of _ essential finiteness _ appeared in the work @xcite by k. diederich and s.m .
webster ; a further geometric reflection principle for locally finite @xmath0-smooth cr mappings appeared in the work @xcite by k. diederich and j.e .
fornss ) . in @xcite , @xcite and
more recently in @xcite , @xcite , @xcite , @xcite ( _ cf .
_ also the applications @xcite , @xcite ) , separate assumptions on the map and on the target have been unified : instead , it is assumed that the so - called _ characteristic variety _ is zero - dimensional at the central point .
however , in these references , it is always supposed that the source generic submanifold is minimal at the central point , whereas , in @xcite , no minimality assumption was needed . this article is devoted to fill the gap between these two trends of thought , applying the technique of normal deformations of analytic discs borrowed from @xcite , @xcite , @xcite .
let @xmath2 or @xmath3 . let @xmath4 .
let @xmath5 .
set @xmath6 .
for @xmath7 , denote @xmath8 .
consider a local @xmath0-smooth cr mapping between two local generic submanifolds @xmath9 in @xmath1 and @xmath10 in @xmath11 , defined _ precisely _ as follows ( background material may be found in @xcite , @xcite ) .
the purpose is to avoid the ( ambiguous ) language of germs .
a _ local @xmath0-smooth cr mapping _ consists of the following data .
* a _ local generic submanifold @xmath9 in @xmath1 _ of positive codimension @xmath12 and of positive cr dimension @xmath13 passing through a point @xmath14 and defined in coordinates @xmath15 vanishing at @xmath16 as a graph : @xmath17 where the functions @xmath18 are real analytic on @xmath19 , for some @xmath20 .
it is also required for all @xmath21 with @xmath22 , that @xmath23 if @xmath24 , namely @xmath9 is a uniformly approximatively horizontal graph .
of course , after perhaps shrinking @xmath20 , this condition is automatically satisfied if the coordinates are adjusted at the beginning in order that @xmath25 .
in fact , it is more convenient to work with a local representation of @xmath9 by complex defining equations @xmath26 obtained by applying the implicit function theorem to , where of course , one may assume that the @xmath27 converge normally for @xmath28 and that for all @xmath21 with @xmath22 , one has @xmath29 if @xmath30 . * a local generic submanifold @xmath10 in @xmath11 with central point @xmath31 of positive codimension @xmath32 and of positive cr dimension @xmath33 passing through a point @xmath34 and defined in coordinates @xmath35 similarly as in . *
a @xmath36-smooth mapping @xmath37 with @xmath38 which is defined in @xmath39 and which satisfies for some two radii @xmath40 , @xmath41 with @xmath42 , @xmath43 , the condition @xmath44 by @xcite ( and also @xcite ) , after shrinking ( if necessary ) @xmath20 and @xmath45 with @xmath46 : * every @xmath0-smooth cr function defined on @xmath47 ( and in particular the @xmath48 components @xmath49 of @xmath50 ) is a uniform limit of polynomials on @xmath51 .
a _ complete wedge in @xmath52 with edge @xmath53 _ is a subset of @xmath1 of the form @xmath54 , where @xmath55 is some open strictly convex infinite ( _ i.e. _ not truncated ) cone in @xmath56 . as in @xcite
, it will be assumed that : * there exists a complete wedge @xmath57 in @xmath58 with edge @xmath59 such that the @xmath60 components of @xmath50 extend holomorphically to @xmath57 .
put @xmath61 for @xmath62 and and @xmath63 for @xmath64 .
let @xmath65 be an arbitrary basis of @xmath66 vector fields tangent to @xmath9 having real analytic coefficients ( the most convenient is written in below ) .
consider the _
first characteristic variety of @xmath50 at @xmath16 _ to be the complex analytic subset @xmath67 of @xmath68 consisting of elements @xmath69 satisfying the equations @xmath70
\right\vert_{\bar t=0}=0 , \ , \
\text{\rm for all } \
j'=1,\dots , d ' \ \text{\rm and all } \
\beta\in\n^m.\ ] ] it is indeed a complex analytic subset defined as the zero set of an infinite collection of functions which are holomorphic in @xmath71 . by ,
@xmath72 , for @xmath73 and for all @xmath74 .
it follows that the origin @xmath31 belongs to the complex analytic subset @xmath67 .
the focus is on the dimension at @xmath75 of @xmath76 .
the map @xmath50 will be called _ essentially finite at _
@xmath16 if @xmath77 .
denote by @xmath78 the ( not local ) cr orbit of @xmath16 in @xmath9 .
the main result is as follows .
let @xmath79 be a local @xmath36-smooth cr mapping between two real analytic local generic submanifolds of @xmath1 and of @xmath80 .
assume that there exists a complete wedge @xmath57 in @xmath58 with edge @xmath81 such that the @xmath48 components of @xmath50 extend holomorphically to @xmath57 , assume that there exist points @xmath82 arbitrarily close to @xmath16 at which @xmath50 is real analytic and assume that @xmath50 is essentially finite at @xmath16 .
then there exists a radius @xmath83 with @xmath84 such that @xmath50 extends holomorphically to @xmath85 . in the version of theorem 2.9 published in @xcite ,
@xcite , it is assumed that @xmath9 is minimal at @xmath16 , which entails the holomorphic extendability assumption * ( v ) * , thanks to @xcite . in @xcite ,
@xcite , there is a crucial proposition about envelopes of meromorphy ( same statement and same proof in the two references albeit notations differ ) , relying on subtle geometric arguments which stem from the theory of deformations of analytic discs developed in @xcite , @xcite , @xcite .
in this article , the purpose is to clean up and to simplify these geometric arguments , by means of the propagation of wedge extendability theorem established in @xcite , @xcite ( _ cf .
_ @xcite for a preliminary version ) .
the stronger theorem 2.9 will be established thanks to this change of geometric point of view .
an elementary lemma applies to recover from theorem 2.9 the main result of @xcite , @xcite .
let @xmath79 be a local @xmath0-smooth cr mapping between two real analytic local generic submanifolds of @xmath1 and of @xmath80 .
assume that @xmath9 is minimal at @xmath16 , so that @xmath86 contains @xmath87 for some @xmath88 and so that ( thanks to @xcite ) after perhaps shrinking @xmath45 , the assumption * ( v ) * holds .
if @xmath50 is essentially finite at @xmath16 , then there exist points @xmath89 arbitrarily close to @xmath16 at which @xmath50 is real analytic .
finally , in order to recover the main result of @xcite , remind that the essential finiteness of @xmath10 at @xmath31 together with the cr diffeomorphism assumption entails the essential finiteness of @xmath50 at @xmath16 ( @xcite , lemma 4.1 ; a more general version is corollary 1.3 in @xcite ; the most general version appears as theorem 4.3.1 * ( 3 ) * in @xcite , in which it is shown that cr - transversality of @xmath50 at @xmath16 together with essential finiteness of @xmath10 at @xmath31 implies that @xmath50 is essentially finite at @xmath16 ) . in @xcite
, it is observed that essential finiteness of a hypersurface @xmath9 at one of its points @xmath16 implies its minimality ( finite type in the sense of lie - chow - kohn - bloom - graham ) at @xmath16 . here
are further observations .
if @xmath10 is a local generic submanifold of @xmath80 passing through a point @xmath31 which is essentially finite at @xmath31 , then @xmath90 and the cr orbit @xmath91 itself is essentially finite at @xmath31 .
furthermore , in the case where @xmath9 is a real analytic hypersurface , essential finiteness of @xmath50 at @xmath16 implies that @xmath9 is minimal at @xmath16 .
assume that @xmath92 and that @xmath50 is a cr diffeomorphism .
then there exists a zariski - dense open subset of points @xmath93 at which @xmath10 is finitely nondegenerate .
it follows that @xmath50 itself is finitely nondegenerate at @xmath94 , and by a known result ( @xcite , @xcite , @xcite , @xcite ) , @xmath50 is real analytic at @xmath95 . applying then theorem 2.9 : ( @xcite ) let @xmath79 be a local @xmath96-smooth cr mapping which is a cr diffeomorphism . if @xmath9 is essentially finite at @xmath16 and if the components of @xmath50 extend holomorphically to a wedge at @xmath16 , then @xmath50 is real analytic at @xmath16 . further applications ( in the spirit of @xcite , @xcite ) that may be stated are left to the interested reader .
the remainder of this article is devoted to the proofs of theorem 2.9 , of lemma 2.11 and of lemma 2.12 .
denote by @xmath97 the basis of @xmath66 vector fields tangent to @xmath9 defined explicitely by @xmath98 here , the coefficients of the vector fields @xmath99 are holomorphic with respect to @xmath100 .
since it will be more convenient for later use to have antiholomorphic dependence with respect to @xmath100 , replace @xmath100 by @xmath101 ( which is possible when @xmath102 belongs to @xmath9 ) , and write the vector fields under the form @xmath103 apply the derivations @xmath104 , @xmath105 , to @xmath106 , which yields @xmath107 for @xmath108 and for @xmath109 .
as the coefficients of the vector fields @xmath99 are holomorphic with respect to @xmath110 , the differentiated equations may be rewritten under the developed form : @xmath111 here , @xmath112 denotes the @xmath113-th jet of @xmath114 . by construction , the functions @xmath115 are holomorphic with respect to @xmath116 , @xmath117 with @xmath118 , @xmath119 , they are holomorphic with respect to the zero - th order jet @xmath120 with @xmath121 , they are relatively polynomial with respect to the nonzero derivatives @xmath122 , and they are relatively holomorphic with respect to @xmath69 with @xmath123 . by the main assumption of essential finiteness , there exists an integer @xmath124 such that the complex analytic subset defined by the equations @xmath125 which passes through the origin in @xmath80 , is of dimension zero at the origin . by @xcite , chapter 5 , it follows that there exists @xmath126 with @xmath127 , there exists @xmath128 with @xmath129 , there exists @xmath130 with @xmath131 and for all @xmath132 , there exist monic weierstrass polynomials @xmath133 of the form @xmath134 with coefficients @xmath135 being holomorphic with respect to @xmath116 , @xmath117 with @xmath136 , @xmath137 and with respect to @xmath138 with @xmath139 , with moreover @xmath140 for all @xmath141 , such that the complex analytic set @xmath142 is contained in the zero - set of all weierstrass polynomials @xmath143 , namely the set : @xmath144 thanks to hilbert s nullstellensatz , there exists an integer @xmath145 such that for @xmath146 , the powers @xmath147 belong to the ideal generated by the @xmath148 .
consequently , each component @xmath149 satisfies the monic polynomial equation @xmath150 for all @xmath151 . here ,
notably , each component @xmath152 is separated from the others .
to establish lemma 2.11 , let @xmath40 be as in its statement and let @xmath126 be as in 3 just above .
shrinking @xmath126 if necessary , assume @xmath153 to fix ideas .
fix @xmath154 and consider the following trivial dichotomy : either @xmath155 for all @xmath156 or @xmath157 over @xmath158 . in the first case , replace the equation @xmath159 by the equation , which is a monic polynomial of degree @xmath160 in @xmath149 .
after a finite number of steps , the second case holds , with a monic polynomial of lower degree , still denoted by @xmath161 . pick a point @xmath162 together with an open neighborhood @xmath163 of @xmath164 so that @xmath165 for all @xmath166 , whereas @xmath167 over @xmath168 .
fix now @xmath169 .
apply the same dichotomy as , to @xmath170 , but for @xmath171 running only over @xmath168 .
process as in the case @xmath172 to replace @xmath173 by a monic polynomial of minimal degree , chose @xmath174 and @xmath175 , _ etc . _ after @xmath48 steps , there exists a point @xmath176 , there exists an open neighborhood @xmath177 of @xmath95 , there exists monic polynomials , still denoted by @xmath178 which are of the form , such that @xmath179 for all @xmath180 but @xmath181 for all @xmath182 .
apply then the implicit function theorem to the identities , to solve @xmath183 for all @xmath171 in a ( possibly smaller ) neighborhood of @xmath95 , where @xmath184 is a certain complex analytic @xmath11-valued mapping defined in a neighborhood of @xmath185 in @xmath186 .
it then follows from standard arguments which are easy modifications of the original phenomenon discovered in @xcite that @xmath50 extends holomorphically to a neighborhood of @xmath95 in @xmath1 , or equivalently that @xmath50 is real analytic in a neighborhood of @xmath95 in @xmath9 . the positive radius @xmath126 could have been shrunk from the beginning to be arbitrarily small , and the same reasoning provides a point @xmath95 , arbitrarily close to @xmath16 , at which @xmath50 is real analytic .
the proof of lemma 2.11 is complete .
let @xmath187 be as in definition 2.2 , let @xmath40 be as in the assumptions of theorem 2.9 and let @xmath126 be as in 3 above . by hypothesis , there exists at least one point @xmath188 at which @xmath50 is real analytic ( it will not be necessary to deal with such points which are closer to @xmath16 ) . by assumption
, the components @xmath152 extend holomorphically to the complete wedge @xmath57 in @xmath52 with edge @xmath189 . define a complete wedge @xmath190 in @xmath191 with edge @xmath158 as the intersection of @xmath57 with @xmath191 .
here is an illustration , where @xmath190 is on the top and its symmetric @xmath192 ( to be introduced later ) is on the bottom : polydiscs-wedges.pstex_t in these conditions , the components @xmath152 of @xmath50 extend holomorphically to @xmath193 .
also , the jet @xmath194 in the arguments of the functions @xmath195 in extends antiholomorphically to @xmath196 .
it is now time to state an independent general proposition , where @xmath10 disapears , where each component @xmath197 is replaced by a @xmath0-smooth cr function @xmath198 , not necessarily coming from a cr mapping , where the jet @xmath199 is replaced by an independent vector valued mapping @xmath200 which extends antiholomorphically to @xmath190 , for similar statements , _ see _ @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite . as above , let @xmath9 be a real analytic local generic submanifold defined as a graph in @xmath201 by , let @xmath126 with @xmath202 and let @xmath190 be a complete wedge in @xmath85 with edge @xmath158 .
let @xmath203 be a @xmath0-smooth cr function defined over @xmath204 which extends holomorphically to the complete wedge @xmath193 and which is real analytic in a neighborhood of at least one point @xmath205 .
let @xmath206 , let @xmath131 and let @xmath207 be a @xmath208-valued @xmath0-smooth cr mapping defined on @xmath209 which satisfies @xmath210 for all @xmath156 , which extends holomorphically to @xmath193 and which is real analytic in a neighborhood of the same point @xmath211 .
let @xmath212 with @xmath213 , and for @xmath214 , let @xmath215 be some functions which are holomorphic for @xmath136 , for @xmath137 and for @xmath216 .
assume that @xmath203 satisfies the ( not necessarily monic ) polynomial equation @xmath217 for all @xmath151 and that the @xmath96-smooth functions @xmath218 are not all identically zero .
then there exists an open neighborhood @xmath219 of @xmath220 in @xmath1 to which @xmath203 extends holomorphically .
this proposition , applied to each @xmath197 , completes the proof of theorem 2.9 .
the remainder of this section is devoted to describe its proof , intuitively speaking .
although the equations ( of a form similar to ) were obtained by applying the @xmath66 vector fields tangent to @xmath9 , it will be crucial to reapply the vector fields @xmath99 to .
but before applying the @xmath221 , it is also crucial to assume that @xmath222 is the smallest possible integer with the property that there exists a relation of the form on @xmath223 .
this assumption is of course free .
since @xmath203 is cr , it may be considered as a constant by the derivations @xmath99 .
suppose for a while that dividing by @xmath224 is allowed in some sense .
then rewrite as follows : @xmath225 applying now the derivations @xmath99 , the term @xmath226 $ ] vanishes ( crucial fact ) , which yields the identities @xmath227-h_\ell\left(z,\bar t , \overline{b(t)}\right)\ , \overline{l}_k\left [ h_0\left(z,\bar t , \overline{b(t)}\right)\right ] \over \left[h_0\left(z,\bar t , \overline{b(t)}\right)\right]^2}\ , a(t)^{n-\ell}=0,\ ] ] for @xmath228 .
now , after chasing the unnecessary denominator , observe that since the coefficients of the @xmath99 are holomorphic in @xmath229 , there exist new holomorphic functions @xmath230 , @xmath231 , such that the identity may be rewritten as @xmath232 where of course @xmath233 setting @xmath234 , the relation is totally similar to , if some freedom is allowed about the number of functions @xmath235 .
however , the degree @xmath236 of the relation is strictly less than the degree @xmath222 of the relation .
_ because the degree @xmath222 was chosen to be the smallest possible one , this relation has to be trivial_. equivalently : @xmath237 over @xmath158 , for @xmath238 and @xmath239 . in other words ,
the quotient @xmath240 ( which has to be defined carefully in some sense ) is cr on @xmath241 .
informally speaking , such an identity should be exceptional , because the term @xmath242 is anti - cr .
in fact , one may expect intuitively that if the relation is satisfied , then there are no terms @xmath242 at all , and hence the quotient @xmath243 extends meromorphically ( but not holomorphically , because of the presence of a quotient ) to a neighborhood @xmath244 of @xmath245 in @xmath1 .
this is true and will be proved below in main lemma 7.1 formulated below , where the assumption that @xmath203 and @xmath207 are already real analytic in a neighborhood of the point @xmath246 is used . according to the works of k. oka and e.e .
levi , a meromorphic function defined in a domain of @xmath1 is always the global quotient of two holomorphic functions ( _ see _
@xcite for instance ) .
it follows that there exist functions @xmath247 holomorphic in @xmath248 such that @xmath249 for all @xmath250 . replacing then in and chasing the denominator : @xmath251 for all @xmath252 . here
, the function @xmath253 is not identically zero . finally ,
by reproducing section 3.3 of @xcite or proposition 6.4 of @xcite ( _ see _ also @xcite ) , it follows that @xmath203 is real analytic at every point of @xmath254 , hence ( thanks to the severi - tomassini theorem ) it extends holomorphically to an open neighborhood @xmath255 of @xmath256 in @xmath1 .
sections 6 , 7,8,9 and 10 are devoted to complete rigorously all the details of this strategy of proof .
inspired by the preceding discussion , introduce the set @xmath257 of functions of the form @xmath258 , where @xmath259 , where @xmath207 is a @xmath208-valued @xmath0-smooth cr mapping defined on @xmath260 which extends holomorphically to @xmath196 and which is real analytic at the point @xmath261 and where @xmath262 is holomorphic for @xmath136 , @xmath263 and for @xmath264 and where @xmath265 for all @xmath266 .
the main feature of @xmath267 is the following .
every function @xmath268 admits a real analytic extension to @xmath193 which is antiholomorphic with respect to the complex transversal coordinates @xmath269 .
indeed , @xmath270 admits an antiholomorphic extension to @xmath193 by assumption , the function @xmath271 is holomorphic with respect to its variables , but it also depends on the holomorphic variables @xmath272 in general .
consequently , the antiholomorphicity with respect to @xmath116 ( only ) is lost .
the set @xmath273 is an entire ring which is stable under differentiation by the @xmath66 vector fields tangent to @xmath9 : @xmath274 the set @xmath273 is obviously stable under addition and multiplication .
suppose that there exist two elements @xmath275 , @xmath276 such that @xmath277 for all @xmath278 .
then there exists a nonempty open subset @xmath279 of @xmath280 on which @xmath281 or @xmath282say @xmath283 to fix ideas
vanishes identically . one must show that that @xmath284 vanishes identically on @xmath241 . according to lemma 6.1
, the function @xmath285 extends real analytically and antiholomorphically with respect to @xmath100 into the wedge @xmath193 . by the principle of analytic continuation ( for real analytic functions ) , it suffices to show that @xmath286 vanishes identically on a nonempty open subset of @xmath193 .
let @xmath287 and let @xmath288 be an open polydisc centered at @xmath289 with @xmath290 . for @xmath291 , the intersection @xmath292 is maximally real in the slice @xmath293 .
also , the function @xmath294 extends antiholomorphically with respect to @xmath100 into the sliced wedge @xmath295 . by the generic uniqueness principle ( for antiholomorphic functions ) , it follows that @xmath296 vanishes identically in the sliced wedge @xmath297 .
since @xmath298 was arbitrary , it follows that @xmath299 vanishes identically in the nonempty open subset @xmath300 of @xmath193 , as desired .
finally , since the coefficients of the vector fields @xmath99 are holomorphic for @xmath301 and @xmath302 , the second property follows from the chain rule .
the proof of lemma 6.2 is complete .
a slightly more precise result has in fact been established .
the zero - set of a nonzero function @xmath303 in @xmath304 is a closed subset of @xmath158 with nonempty interior . in the sequel ,
denote by @xmath305 the zero set of @xmath306 on @xmath241 .
since the ring @xmath273 is entire , it is allowed to consider its _ field of fractions _
@xmath307 which consists formally of quotients of the form @xmath308 and which may be viewed as a standard complex - valued function on the dense open subset @xmath309 .
since @xmath310 is a field , the notion of algebraic dependence makes sense .
precisely , a @xmath0-smooth cr function @xmath203 which is defined on @xmath158 is called _ algebraic over _
@xmath311 if there exists a nonzero polynomial @xmath312 with coefficients in @xmath313 which annihilates @xmath203 , _ i.e. _ such that @xmath314 for all @xmath315 .
equivalently , after chasing the denominators , the @xmath316-smooth cr function @xmath203 is algebraically dependent over the ring @xmath273 .
another notion of algebraic dependence is the following .
let @xmath317 denote the field of functions which are meromorphic in some connected open neighborhood of @xmath318 in @xmath1 . in this field ,
the classical algebraic operations are defined up to some shrinking of the domains of definition .
then a @xmath316-smooth cr function @xmath203 defined on @xmath223 is called _ algebraic over _
@xmath317 if there exists a nonzero polynomial @xmath319 with coefficients @xmath320 which annihilates @xmath203 , _
i.e. _ such that @xmath321 for all @xmath171 in some neighborhood @xmath244 of @xmath322 in @xmath1 , outside the union of the polar sets of the @xmath323 .
it is now time to reformulate and to slightly generalize the heuristic discussion of 5.4 . in the following lemma ,
a main lemma 7.1 ( to be formulated and to be proved later ) , is hidden .
let @xmath203 be a @xmath316-smooth cr function on @xmath223 which extends holomorphically to the complete wedge @xmath193 in @xmath85 with edge @xmath241 .
assume that @xmath203 is real analytic at one point @xmath324 .
if the function @xmath203 is algebraic over the field @xmath307 ( or equivalently over the ring @xmath273 ) , then it is algebraic over the field @xmath325 . without loss of generality , assume that @xmath203 satisfies a polynomial relation @xmath326 for all @xmath156 , _ whose degree @xmath213 is the smallest possible_. hence @xmath327 does not vanish identically . proceeding exactly as in the heuristic discussion of 5.4 ,
divide by @xmath327 , which yields @xmath328 apply now the derivations @xmath99 , which yields @xmath329-h_\ell \left(z,\bar t , \overline{b(t ) } \right)\ , \overline{l}_k\left [ h_0 \left(z,\bar t , \overline{b(t ) } \right)\right ] \over \left[h_0 \left(z,\bar t , \overline{b(t ) } \right ) \right]^2}\ , a(t)^{n-\ell}=0.\ ] ] now , after chasing the unnecessary denominator , and using , rewrite this identity under the form @xmath330 where @xmath331 belongs to @xmath273 . thus a relation which is totally similar to , but which is of degree strictly less than @xmath222 , has been constructed . because the degree @xmath222 was chosen to be the smallest possible one , this relation has to be trivial .
consequently , the @xmath66 vector fields @xmath221 annihilate the quotients @xmath240 outside the zero - set @xmath332 . by main lemma 7.1
( to be discussed later ) , this implies that for each @xmath333 , the quotient @xmath334 coincides in some neighborhood @xmath244 of @xmath322 in @xmath1 with the restriction to @xmath335 \backslash \mathcal{z}_{h_0}$ ] of a meromorphic function @xmath336 . finally , replacing @xmath334 by @xmath337 in the original relation
, it follows that @xmath203 is algebraic over @xmath325 .
the proof of lemma 6.12 is complete . thanks to this lemma 6.12 ,
the proof of main proposition 5.2 is complete , as explained after .
in sum , everything relies upon a statement , interesting in itself , which is reformulated carefully , including all the assumptions . as in main proposition 5.2 ,
let @xmath9 be a real analytic local generic submanifold defined in @xmath338 , as in definition 2.2 * ( i)*. let @xmath126 with @xmath339 and let @xmath193 be a complete wedge in @xmath340 with edge @xmath223 .
let @xmath341 , let @xmath342 and let @xmath343 be two @xmath344- and @xmath345-valued @xmath316-smooth cr mappings defined on @xmath223 which satisfy @xmath346 , @xmath347 , which extend both holomorphically to @xmath193 and which are real analytic in a neighborhood of the same point @xmath348 .
let @xmath349 and @xmath350 be two holomorphic functions defined for @xmath351 , for @xmath352 and for @xmath353 , @xmath354 .
assume that @xmath355 does not vanish identically on @xmath241 and consider the quotient @xmath356 which belongs to @xmath307 and remind that by corollary 6.5 , the zero - set @xmath357 of @xmath358 is a closed subset of @xmath241 with nonempty interior .
then the following three conditions are equivalent : * there exists a meromorphic function in @xmath325 of the form @xmath359 , where @xmath360 and @xmath361 are holomorphic functions defined in some open connected neighborhood @xmath244 of @xmath318 in @xmath1 with @xmath362 such that @xmath363 for all @xmath364 outside the zero set of @xmath276 ; * the quotient @xmath365 is cr over the dense open subset @xmath366 ; * there exists a nonempty open subset @xmath279 of the dense open subset @xmath366 on which the quotient @xmath365 is cr . obviously , * ( 2 ) * @xmath367 * ( 3)*. treat first the two implications * ( 1 ) * @xmath367 * ( 2 ) * and * ( 3 ) * @xmath367 * ( 2 ) * , which are easy . indeed , by lemma 6.1 , for @xmath368 , the cr derivatived functions @xmath369 also belong to @xmath370 .
if * ( 1 ) * or * ( 3 ) * holds , then @xmath371 is cr on a nonempty open subset @xmath279 of @xmath366 , so corollary 6.5 implies that @xmath372 vanishes identically on @xmath373 , which yields * ( 2)*. next , begin the proof of the implication * ( 2 ) * @xmath367 * ( 1 ) * , which is by far the main task .
even if it is already known that * ( 2 ) * and * ( 3 ) * are equivalent , property * ( 1 ) * will be established assuming only that @xmath365 is cr over a nonempty open subset @xmath279 , as in * ( 3)*. define the schwarz reflection across @xmath9 which stabilizes the slices @xmath374 explicitely by @xmath375 shrinking @xmath126 a bit if necessary , one can construct a slightly smaller complete wedge @xmath376 which is contained in @xmath377 , as depicted in figure 1 above . what is important is that @xmath193 and @xmath378 are directed by two opposite vectors at every point of @xmath241 , because a version of the edge of the wedge theorem will be applied in the end of the proof .
since @xmath379 is antiholomorphic with respect to @xmath100 and since @xmath380 and @xmath381 extend as real analytic functions in @xmath193 which are antiholomorphic with respect to @xmath100 , it follows by composition with the reflection @xmath382 that @xmath275 and @xmath276 extend to be real analytic in @xmath383 and holomorphic with respect to @xmath100 .
denote by @xmath384 and @xmath385 these two real analytic extensions .
since the antiholomorphic reflection coincides with the identity mapping on @xmath386 , the @xmath36 boundary values of the real analytic extensions @xmath387 , @xmath388 are just @xmath275 and @xmath276 .
since @xmath389 is cr on the nonempty open subset @xmath279 of @xmath390 , and since it is holomorphic with respect to the variable @xmath100 in the vertical conelike slices @xmath391 , it follows from an elementary ( and known ) separate cauchy - riemann principle that there exists a neighborhood @xmath392 of @xmath279 in @xmath1 such that @xmath393 extends holomorphically to @xmath394 .
apply now the following general lemma to deduce that @xmath395 can be represented as a quotient @xmath396 , where @xmath161 and @xmath397 are holomorphic in @xmath378 .
let @xmath398 be a domain , let @xmath399 and @xmath400 be two real analytic functions in @xmath401 .
suppose that there exist a point @xmath402 and a nonempty open neighborhood of @xmath289 such that @xmath403 is contained in @xmath404 and such that the restriction to @xmath403 of @xmath405 is holomorphic in @xmath403 .
then there exist two holomorphic functions @xmath161 and @xmath397 in @xmath401 such that @xmath406 in this proof ( only ) , denote some complex coordinates on @xmath1 by @xmath407 and @xmath408 . by assumption , for @xmath409 and for @xmath410
, it holds that @xmath411 which yields by analytic continuation : @xmath412 for all @xmath413 .
it follows that the quotient @xmath405 is holomorphic on @xmath414 , even if @xmath414 is not ( locally ) connected .
in fact , to achieve the proof , the local connectedness of @xmath415 is needed . to fix this point , it is sufficient to show that @xmath416 is in fact of real dimension @xmath417 , hence in particular @xmath415 is locally connected .
indeed , proceeding by contradiction , in a neighborhood of a point @xmath418 at which @xmath416 is geometrically smooth and of dimension @xmath419 , make a translation of coordinates so that @xmath420 is the origin and @xmath416 is locally represented by @xmath421 .
then write @xmath422,\ ] ] for some integer @xmath423 and some real analytic function @xmath424 independent of @xmath425 which is not identically zero . without loss of generality , assume that @xmath426 , hence @xmath427,\ ] ] for some integer @xmath428 , where the function @xmath429 is also not identically zero .
since by assumption @xmath430 , it holds that @xmath431 .
compute then for @xmath432 , which yields @xmath433,\ ] ] a contradiction .
thus @xmath405 extends holomorphically to the _ locally connected _ open set @xmath434 . to pursue the proof ,
choose an arbitrary point @xmath435 , center the coordinates @xmath116 at @xmath420 and consider the polydisc @xmath436 , where @xmath437 .
then @xmath438 and @xmath439 may be developed in power series with respect to @xmath440 and @xmath441 with radius of convergence at least equal to @xmath442 .
after perhaps making a unitary transformation , assume that the maximally real submanifold @xmath443\cap \delta_n(\rho_q)$ ] is not contained in @xmath416 .
then complexify the variable @xmath444 to be the complex variable @xmath445 and introduce the element of meromorphic function @xmath446 which is defined in @xmath447 . by construction ,
the restrictions to the maximally real subspace @xmath448 \backslash \mathcal{z}_{a_2}\ ] ] of @xmath449 and of @xmath450 coincide .
because @xmath451 is locally connected , it follows from the principle of analytic continuation that @xmath449 and @xmath452 coincide all over @xmath453 . in summary , at every point @xmath420 of @xmath401 ,
a meromorphic extension of the holomorphic function @xmath454 defined in @xmath455 has been constructed .
it follows from @xcite that this meromorphic function can be represented as a quotient of holomorphic functions in @xmath401 , which completes the proof of the lemma .
thus @xmath456 extends meromorphically to @xmath457 . of course
, neither @xmath392 nor a neighborhood ( in @xmath1 ) of the dense subset @xmath458 need contain a point of the ( thin in the nonminimal case ) subset @xmath459 .
here comes the assumption that the cr mappings @xmath460 and @xmath461 extend holomorphically to a neighborhood @xmath462 ( in @xmath1 ) of some point @xmath463 .
it follows that @xmath464 and @xmath465 extend real analytically to @xmath466 .
thanks to the ( already established ) equivalence between * ( 2 ) * and * ( 3 ) * , the quotient @xmath467 is cr over the dense open subset @xmath468 .
in particular , there exists a point @xmath469 in a neighborhood of which @xmath276 is nonzero .
then , thanks to the severi - tomassini extension theorem , @xmath467 extends holomorphically to a neighborhood @xmath470 ( in @xmath1 ) of @xmath471 .
applying lemma 7.5 just above , it follows that @xmath365 extends meromorphically to @xmath462 .
forget the open subset @xmath392 and summarize the obtained extension result .
the cr quotient @xmath365 extends meromorphically to @xmath472 .
the goal , to which the remainder of the paper is devoted , is to prove that @xmath365 extends meromorphically to some open neighborhood @xmath244 of @xmath322 in @xmath1 . to this aim , define the set @xmath473 of points @xmath474 such that there exists a small nonempty open polydisc @xmath475 centered at @xmath420 and a meromorphic extension @xmath476 of @xmath477 to @xmath475 .
this set is nonempty , since @xmath95 belongs to @xmath473 by assumption .
it follows from the uniqueness principle on a generic edge that the various meromorphic functions @xmath478 glue together to provide a well defined meromorphic function @xmath479 defined in the open neighborhood @xmath480 of @xmath473 in @xmath1 .
state this property as a step lemma .
the cr quotient extends meromorphically to @xmath481 .
if @xmath482 , main lemma 7.1 would be proved , almost gratuitously . suppose therefore that the complement @xmath483 of @xmath473 in @xmath484 is nonempty .
to conclude the proof of main lemma 7.1 , it will suffice to derive a contradiction in the following form : establish that there exists in fact _ at least one point _
@xmath485 at which @xmath371 extends meromorphically .
so , assume that @xmath483 ( the bad set ) and @xmath473 ( the good set ) are nonempty , with @xmath486 and with @xmath487 . for technical convenience ,
it is better to pick a special point @xmath488 so that @xmath483 lies behind a real analytic generic `` wall '' @xmath489 passing through @xmath490 , as depicted in figure 2 below ( _ cf . _
@xcite , @xcite ) .
there exists a point @xmath485 and a real analytic hypersurface @xmath491 passing through @xmath490 which is generic in @xmath1 such that @xmath492 lies , near @xmath490 , in one side of @xmath489 .
let @xmath493 be an arbitrary point and let @xmath494 be a piecewise real analytic curve running in complex tangential directions to @xmath9 ( cr - curve ) which links the point @xmath420 with the point @xmath495 .
such a curve exists because the cr orbit @xmath484 is locally minimal at every point . after shortening @xmath494 and changing the points @xmath420 and @xmath95 if necessary
, one can assume that @xmath494 is a smoothly embedded segment , that @xmath420 and @xmath95 belong to @xmath494 and are close to each other .
therefore @xmath494 can be described as a part of an integral curve of some nonvanishing real analytic section @xmath496 of @xmath497 defined in a neighborhood of @xmath420 .
blowing-ellipsoids.pstex_t let @xmath498 be a small @xmath499-dimensional real analytic hypersurface passing through @xmath420 such that @xmath500 is not tangent to @xmath501 at @xmath420 .
integrating @xmath496 with initial conditions in @xmath501 , one obtains real analytic coordinates @xmath502 so that for fixed @xmath503 , the segments @xmath504 are contained in the trajectories of @xmath496 . after a translation
, one may assume that the origin @xmath505 corresponds to a point of @xmath494 close to @xmath420 which is not contained in @xmath483 , again denoted by @xmath420 .
fix a small @xmath342 and for real @xmath506 , define the ellipsoids ( _ see _ figure 2 above ) @xmath507 there exists the smallest @xmath508 with @xmath509 . then @xmath510 and @xmath511 .
observe that every boundary @xmath512 is transverse to the trajectories of @xmath496 out off the equatorial set @xmath513 which is contained in @xmath473 .
hence @xmath514 is transverse to @xmath496 at every point of @xmath515 .
so @xmath516 is generic in @xmath1 , since @xmath496 is a section of @xmath497 .
to conclude , it suffices to choose a point @xmath517 and to take for @xmath489 a small real analytic hypersurface passing through @xmath490 which is tangent to @xmath518 and satisfies @xmath519 .
the proof of lemma 8.1 .
is complete .
choose now a point @xmath485 as in lemma 10.1 and choose @xmath520 such that the polydisc @xmath521 ( in the new coordinates centered at @xmath490 ) with center @xmath490 is contained in the polydisc @xmath85 ( in the old coordinates centered at @xmath16 ) with center @xmath16 and localize everything in @xmath521 ( _ cf .
_ figure 3 below , where @xmath483 has been redrawn on the left ) .
denote by @xmath522 the negative , left open one - sided neighborhood of @xmath489 in @xmath9 such that @xmath492 lies in @xmath522 in a neighborhood of @xmath490 , and denote by @xmath523 the other side , which is contained in @xmath473 by assumption .
choose affine coordinates vanishing at @xmath490 , still denoted by @xmath524 in order that @xmath25 and @xmath525 .
denote by @xmath526 the coordinates @xmath527 , which are decomposed in real and imaginary part as @xmath528 .
then @xmath9 is defined by @xmath529 , @xmath530 with @xmath531 , @xmath532 and @xmath489 is defined by a supplementary equation @xmath533 , with @xmath534 , @xmath535 .
after possibly replacing @xmath489 by a new hypersurface which is contained in @xmath536 ( _ cf .
_ the end of the proof of lemma 8.1 ) and after possibly making a dilatation of coordinates , one may assume that the supplementary equation of @xmath489 is simply given by @xmath537 , and that @xmath523 is given by @xmath538 ( for the readability of figure 3 below , the curvature of @xmath489 has been reversed , hence the picture of @xmath489 is slightly incorrect ) .
intersect everything with the polydisc @xmath521 centered at @xmath490 in particular , define @xmath539 to reach the desired contradiction ( _ cf .
_ the last sentence of 7 ) which achieves the proof of main lemma 7.1 , the principal goal is now to show that the quotient @xmath308 which already extends holomorphically to @xmath540 extends holomorphically to a neighborhood of @xmath490 in @xmath1 . for this purpose
, the technique of normal deformations of analytic discs enters the scene ( @xcite , @xcite , @xcite ) .
let @xmath541 with @xmath542 and for every @xmath21 with @xmath543 , consider an analytic disc @xmath544 where @xmath545 is given by @xmath546 the disc @xmath547 should be attached to @xmath548 and should satisfy @xmath549 ( remind that @xmath490 is the origin in the chosen coordinates ) .
a necessary and sufficient condition is that @xmath550 satisfies the so - called bishop functional equation @xmath551(\zeta),\ ] ] for all @xmath552 . here
, @xmath553 denotes the harmonic conjugate operator ( hilbert transform on @xmath554 ) normalized at @xmath555 , namely satisfying @xmath556 for every @xmath557 . by @xcite
, the solution exists , is unique and yields a family of analytic discs @xmath558 which is smooth ( and in fact real analytic ) with respect to @xmath21 and @xmath559 .
of course , for @xmath560 , the disc @xmath561 is the constant disc which maps @xmath562 to @xmath490 .
disc-wall.pstex_t the following elementary lemma shows that the boundary @xmath563 of the disc @xmath564 meets the wall @xmath489 only at @xmath490 , as shown in figure 3 just above .
the notion of analytic isotopy is appropriate to apply the continuity principle ( _ cf . _
@xcite , @xcite ) .
there exists @xmath541 with @xmath565 such that the following two properties are satisfied : * for every @xmath21 with @xmath566 , the mapping @xmath567 is an embedding ; moreover , each @xmath547 is analytically isotopic to the point @xmath490 ; * for every @xmath21 with @xmath568 : @xmath569 the fact that @xmath547 is an embedding for @xmath570 is obvious , since this is the case for @xmath571 .
then the analytic isotopy is obtained by letting @xmath21 decrease to @xmath572 .
this proves part * ( 1)*. for part * ( 2 ) * , put @xmath573 , where @xmath574 and compute @xmath575 since the real analytic solution @xmath576 vanishes identically for @xmath560 and vanishes at @xmath577 , there exists a constant @xmath578 such that @xmath579 for all @xmath580 .
one then deduces from the elementary inequalities @xmath581 and @xmath582 that @xmath583 recall that @xmath584 .
hence it suffices to choose @xmath585 in order to insure that @xmath586 for all @xmath587 , which completes the proof .
the goal is to construct a meromorphic extension of @xmath371 to a neighborhood of @xmath490 .
let @xmath588 be an open neighborhood of the point @xmath589 .
thanks to lemma 2.7 in @xcite ( a slight modification of the geometric constructions in @xcite ) , it is possible to include the disc @xmath590 in a _ regular _ ( in the sense of definition 1.8 in @xcite ; _ see _ also p. 493 of @xcite ) family of analytic discs @xmath591 , where @xmath592 satisfies @xmath593 for some @xmath594 , where @xmath595 satisfies @xmath596 for some @xmath597 and where @xmath598 is contained in @xmath599 essentially , the disc is deformed near @xmath600 in order that its direction at exit at @xmath601 covers an open cone at @xmath490 , by means of the parameter @xmath602 , as in @xcite , @xcite , @xcite .
then the parameter @xmath379 achieves translation along @xmath9 . by an application of the continuity principle ( where property * ( 1 ) * of lemma 9.4 is needed )
, there exist @xmath603 with @xmath604 and @xmath471 with @xmath605 , @xmath606 such that the following set covered by pieces of analytic discs @xmath607 is a ( curved ) local wedge of edge @xmath9 at @xmath490 ( _ see _ figure 4 just below ) to which meromorphic functions in @xmath608 extend meromorphically .
edge-of-the-wedge.pstex_t in the preceding situation , three extension results hold : * for @xmath609 , the @xmath610-valued @xmath316-smooth cr function @xmath611 extends holomorphically to @xmath612 ; * the cr quotient @xmath365 extends meromorphically to a symmetric wedge @xmath613 contained in @xmath614 which is only slightly smaller ; * the cr quotient @xmath365 extends meromorphically to @xmath612 . as a preliminary , define translations of geometric objects in the normal directions @xmath615 as follows .
if a unitary vector @xmath616 with zero @xmath116-component and zero @xmath617-component is given , namely @xmath618 is of the form @xmath619 with @xmath620 , then for every @xmath621 very small in comparison with @xmath622 , define the translation @xmath623 as @xmath624 and @xmath625 are ( approximatively ) symmetric to each other , there exists a unitary vector @xmath626 such that @xmath627 in other words , @xmath628 and @xmath629 .
translate also the analytic discs , which yields @xmath630 .
prove now * ( 1 ) * of lemma 10.3 . by construction , the two @xmath631-smooth cr functions @xmath632 and @xmath633 extend holomorphically to @xmath624 .
since for every @xmath634 , the discs @xmath635 have their boundaries contained in @xmath636 and are analytically isotopic to a point in @xmath624 , it follows from the continuity principle that @xmath632 and @xmath633 extend holomorphically to the wedge @xmath637 . by letting @xmath621 tend to zero , it follows that @xmath632 and @xmath633 extend holomorphically to @xmath612 .
next , prove * ( 2 ) * of lemma 10.3 . since @xmath632 and @xmath633 extend holomorphically to @xmath612 , by reasoning as in the beginning of the proof of main lemma 5.2
, it follows that @xmath365 extends meromorphically to the symmetric wedge @xmath638 , which is contained in @xmath614 , but only slightly smaller .
importantly , there exists a unitary vector @xmath639 with coordinates of the form @xmath640 such that @xmath641 and @xmath642 ( _ see _ again figure 4 ) .
finally , prove * ( 3 ) * of lemma 10.3 . in the domain @xmath643 , the meromorphic extension of the cr quotient @xmath644 ,
can be represented as a quotient @xmath396 . since for every @xmath645
, the discs @xmath646 have their boundaries contained in @xmath647 and are analytically isotopic to a point in @xmath648 , it follows from the continuity principle that @xmath161 and @xmath173 extend holomorphically to @xmath649 . by letting @xmath621 tend to zero ,
one deduces that @xmath161 and @xmath173 extend holomorphically to @xmath612 .
hence the cr quotient @xmath467 extends meromorphically to @xmath612 .
the proof of lemma 10.3 is complete .
the proof of main lemma 7.1 is almost achieved . by lemma 10.3
, the cr quotient @xmath365 extends meromorphically to the union @xmath650 this union is connected ( _ cf .
_ figure 5 just below .
denote again by @xmath396 this meromorphic extension , where @xmath161 and @xmath173 are holomorphic in the domain @xmath651 .
deformation.pstex_t introduce a one parameter family of smooth deformations @xmath652 , @xmath653 , of @xmath9 localized in a neighborhood of @xmath490 , with @xmath654 , by pushing @xmath9 near @xmath490 inside @xmath655 .
for this , it suffices to use a cut - off function @xmath656 with support in a neighborhood of the origin , and for @xmath657 , @xmath658 , to define @xmath652 by the vectorial equation @xmath659 where @xmath641 and @xmath660 .
since the resolution of bishop s equation is stable under perturbation , there exists a smoothly deformed family @xmath661 of analytic discs with @xmath662 and a deformed wedge @xmath663 since @xmath612 and @xmath655 have opposite directions , it is clear that one can insure that @xmath664 contains the point @xmath490 . for @xmath665 ,
let the parameter @xmath21 vary in the interval @xmath666 $ ] to deduce that the discs @xmath667 are analytically isotopic to a point in @xmath668 . by a further application of the continuity principle
, it follows that @xmath161 and @xmath173 extend holomorphically to @xmath664 , hence to a neighborhood of @xmath490 . in conclusion
, the cr quotient @xmath365 extends meromorphically to a neighborhood of @xmath490 .
the proof of main lemma 7.1 is complete . in conclusion ,
the proof of theorem 2.9 is complete .
establish first the second sentence .
assume that the source @xmath9 is a hypersurface and that @xmath50 is essentially finite at @xmath16 .
proceeding by contradiction , suppose that @xmath9 is not minimal at @xmath16 .
equivalently , the cr orbit of @xmath16 is an @xmath669-dimensional complex hypersurface passing through @xmath16 , which coincides in fact with the segre variety @xmath670 .
choose holomorphic coordinates @xmath671 vanishing at @xmath16 in which @xmath672 and represent @xmath9 by a single complex equation @xmath673 , where @xmath674 , whence @xmath675 .
it follows that the usual basis of @xmath66 vector fields tangent to @xmath9 satisfies @xmath676 . in the fundamental equations
@xmath677 for @xmath73 , specify @xmath678 , which yields : @xmath679 since @xmath50 is cr and @xmath680 is contained in @xmath9 , the mapping @xmath681 is holomorphic , hence it is justified to write @xmath682 instead of @xmath683 . by assumption , @xmath50 is essentialy finite at @xmath16 .
hence apply the derivations @xmath684 , for @xmath685 , to , which amounts to applying the derivations @xmath686 to , which yields expressions of the form @xmath687 exactly as in 3 , it follows from the essential finiteness assumption that there exist weierstrass polynomials such that @xmath688 putting @xmath689 , it follows that @xmath690 is a constant for @xmath691 , hence vanishes identically .
however , if @xmath692 vanishes identically , it is clearly impossible for @xmath50 to be essentially finite at @xmath16 , since differentiation with respect to @xmath693 of @xmath694 gives nothing else than the constant zero , so @xmath67 is defined by @xmath695 , hence is positive - dimensional .
this establishes the second sentence of lemma 2.12 . to establish the first sentence ,
remind again that if @xmath696 , then the cr orbit of @xmath31 is an @xmath697-dimensional complex manifold passing through @xmath31 , which coincides in fact with the segre variety @xmath698 .
assume that the coordinates @xmath699 are such that @xmath700 .
then the complex defining equations of @xmath10 are of the form @xmath701 , where @xmath702 for @xmath73 .
however , this clearly contradicts the two classical characterizations of essential finiteness ( @xcite )
. in general , there exist coordinates @xmath703 centered at @xmath31 , where @xmath704 is the holomorphic codimension of the intrinsic complexification @xmath705 and where @xmath706 , such that @xmath10 is represented by @xmath707 where @xmath708 for @xmath709 . without loss of generality , assume that the coordinates are normal . by the characterization of essential finiteness in normal coordinates , the ideal in @xmath710 generated by the partial derivatives @xmath711 , @xmath712 , and the partial derivatives @xmath713 , @xmath714 , should be of finite codimension , where @xmath715 runs in @xmath716 . however , the second collection vanishes identically .
thus , the ideal in @xmath710 generated by the partial derivatives @xmath717 is of finite codimension .
this shows that the cr orbit @xmath718 is essentially finite at the origin .
the proof of lemma 2.12 is complete . | an analytico - geometric reflection principle is established by means of normal deformations of analytic discs .
= 0.35 cm * table of contents * * 1 .
introduction 1 .
* * 2 .
preliminaries and statement of the main theorem 1 . * * 3 .
polynomial identities 4 .
* * 4 .
proof of lemma 2.11 5 .
* * 5 .
description of the proof of theorem 2.9 6 .
* * 6 .
rings of cr functions and their fields of fractions 8 .
* * 7 .
statement of main lemma 7.1 11 .
* * 8 .
localization at a nice boundary point 14 .
* * 9 .
construction of analytic discs 16 . * * 10 .
meromorphic extension 17 .
* * 11 .
proof of lemma 2.12 20 .
* * 12 .
references 21 . * |
= 15000 = 15000 this paper updates our derivation ( kormendy 1988 , 1990 ; kormendy & freeman 1996 ) of scaling laws for dm halos of sc i m and dwarf spheroidal ( dsph ) galaxies .
we show that dm halos in less luminous galaxies have smaller core radii @xmath6 , higher central densities @xmath7 , and smaller central velocity dispersions @xmath8 .
these scaling laws are analogous to the fundamental plane relations for elliptical galaxies ( djorgovski & davis 1986 , 1987 ; faber et al .
1987 ; dressler et al . 1987 ; djorgovski , de carvalho , & han 1988 ; see kormendy & djorgovski 1989 for a review ) , and they are interesting for the same reason : they provide new constraints on galaxy formation and evolution .
simple conclusions are discussed in 4 .
a detailed discussion will be published in kormendy & freeman ( 2003 ) .
halo parameters for giant galaxies are derived by decomposing rotation curves @xmath1 into visible matter and dm contributions ( van albada et al .
1985 ) . at galaxy absolute magnitudes @xmath17
, rotation curve decomposition becomes impossible as @xmath5 decreases ( tully & fisher 1977 ) and becomes similar to the velocity dispersion of the gas .
pressure - supported galaxies are not flat .
dm central densities can still be derived , e.g. , by fitting king ( 1966 ) models to the density and velocity dispersion profiles of dsph galaxies .
but dm @xmath6 and @xmath8 can no longer be measured . in this paper , we combine these data to investigate the systematic properties of dm halos over a large range of galaxy luminosities . only sc i m and dwarf spheroidal ( dsph ) galaxies are included .
galaxies of type e
sbc are omitted for two reasons that result from their bulge components .
( 1 ) rotation curve decomposition must deal with two visible - matter components that have different unknown mass - to - light ratios .
therefore it is less reliable .
( 2 ) gravitational compression of the dm by the baryons has substantially modified the halo when the visible mass density is high .
sbc galaxies satisfy the dm correlations , but others deviate in the direction of small @xmath6 and large @xmath7 ( kormendy 1988 , 1990 ) .
this is consistent with baryonic compression .
further evidence for baryonic compression is presented in athanassoula , bosma , & papaioannou ( 1987 , hereafter abp ) .
baryonic compression corrections are omitted here but will be included in kormendy & freeman ( 2003 ) .
consider first the rotation curve of an isothermal sphere in the ideal case where we can measure a massless disk embedded in it . then at @xmath18 , -1pt @xmath19 and at @xmath20 , @xmath21 where @xmath22 is the gravitational constant . if we observe only the @xmath23 part of the rotation curve , we can measure @xmath7 but not @xmath6 or @xmath8 . because of this , @xmath7 is often the only halo parameter that we can measure in low - luminosity galaxies .
in contrast , if the measurements reach far enough into the @xmath24 constant part of the rotation curve , then all three parameters can be measured .
dwarf galaxies come closest to the above ideal , because visible matter contributes only a small fraction of the total mass .
more generally , visible matter dominates the central part of the rotation curve , and a multicomponent mass model is required .
the rotation curve of the visible matter is calculated from the brightness distribution assuming that the mass - to - light ratio @xmath3 of each component is constant with radius .
values of @xmath3 are adjusted to fit as much of the inner rotation curve as desired .
hi gas is taken into account separately .
molecular gas is assumed to follow the light distribution , so it is included in @xmath3 .
then , given the total rotation curve @xmath2 of the visible matter , the halo rotation curve is @xmath25 . a model such as an isothermal
is then fitted to @xmath26 to derive the halo asymptotic velocity @xmath27 , @xmath6 , and @xmath7 .
rotation curve decompositions have now been published for @xmath28 galaxies . of these , 55 survive our selection cuts ( 2.3 ) .
many authors emphasize that mass modeling is uncertain ( van albada et al .
1985 ; skillman et al .
1987 ; lake & feinswog 1989 ) .
lake and feinswog point out that if measurement errors are interpreted strictly , few observations of rotation curves reach large enough radii to determine halo parameters uniquely
. therefore the present results depend on the following assumptions : 1 rotation curves that flatten out to @xmath29 constant are assumed to stay flat outside the radius range measured . 2 we use maximum disk decompositions .
this requires discussion .
the unknown mass - to - light ratio of the disk is a problem the ratio of visible to dark mass can be varied greatly while preserving a good fit to @xmath1 . as the amount of visible mass is reduced , the central dm density must be increased and its core radius must be decreased .
the extreme models ( van albada et al . 1985 , fig
. 4 and 8) are usually the maximum disk mass that does not require the halo to have a hollow core and a solution with @xmath30 . in giant galaxies ,
these solutions are very different .
if we had no additional constraints , we could say little about halo properties .
fortunately , we have other constraints . we can not let @xmath3 get arbitrarily small .
we observe structures such as bars and spiral density waves that require disks to be self - gravitating .
abp turned this qualitative remark into a practical constraint on @xmath3 by applying toomre s ( 1981 ) swing amplifier instability criterion and requiring that the disk have the proper density to give the observed spiral structure ( i.e. , two arms but not one ) .
the results converged on essentially the maximum disk solutions for 18 of 21 sc
im galaxies studied .
the resulting mass - to - light ratios imply plausible young stellar populations . in general
, some evidence favors maximum disks ( e.g. , taga & iye 1994 ; sackett 1997 ; debattista & sellwood 1998 ; bosma 1999 ; sellwood & moore 1999 ; weiner , sellwood , & williams 2001 ; gerhard 2004 ; athanassoula 2004 ; weiner 2004 ) , and other evidence suggests that some disks are submaximal ( e.g. , bottema 1993 , 1997 ; courteau & rix 1999 ) .
our choice of maximum disk solutions affects only giant galaxies ; dwarfs are so dm dominated that @xmath3 uncertainties have little effect . if we used `` bottema disks '' instead of maximum disks , parameter correlations with galaxy luminosity would be shallower . 3 halo are assumed to have non - singular isothermal mass distributions . iau symposium 220 focuses in part on the well known collision ( moore 1994 ) between the prediction that cold dm ( cdm ) has cuspy central density profiles @xmath31 ( e.g. , navarro , frenk , & white 1996 , 1997 [ nfw ] , who find that @xmath32 ) and observational evidence that at least dwarf galaxies have flat cores .
for the purposes of this paper , the difference between isothermals and nfw profiles is nontrivial but relatively benign .
an analogous problem arose with elliptical galaxies : early studies of cores , including the discovery of fundamental plane correlations ( kormendy 1984 ; lauer 1985 ; kormendy 1987b , c ) , were published before _ hubble space telescope _
( _ hst _ ) showed that high - luminosity ellipticals have cuspy cores with projected densities @xmath33 , @xmath34 to 0.25 at small radii ( e.g. , lauer et al .
but pre-_hst _ observations of core radii and central densities probe the same physics as ( and , in fact , are roughly proportional to ) _ hst _ measurements of profile break radii and densities ( kormendy et al .
most results deduced from ground - based photometry remain valid .
we expect that the present dm parameters will prove to measure the relevant physics when the form of the halo density profile is better known .
that is , we consider @xmath6 as an approximate profile break radius and @xmath7 as a measure of the density at @xmath35 or averaged inside @xmath35 ; this should be valid whether halos are isothermal or not .
we take dm parameters from published decompositions with as few changes as possible consistent with the assumption that halos are isothermal , with a uniform distance scale , and with the following selection criteria : 1 morphological types are restricted to sc
i m and dsph , as noted above .
late - type and spheroidal galaxies are physically related ( kormendy 1985 , 1987c ; binggeli & cameron 1991 ; ferguson & binggeli 1994 ) .
most dsph companions of our galaxy have had episodes of star formation in the past 1 8 gyr ( da costa 1994 and mateo 1998 provide reviews ) ; they presumably turned from irregulars into spheroidals since that time ( kormendy & bender 1994 ) .
in fact , the distinction between galaxy types has blurred as hi gas has been found in or near a few spheroidals ( e.g. , sculptor : knapp et al .
1978 ; carignan et al . 1998
; see mateo 1998 for a review ) .
it seems physically reasonable to include sc
i m and dsph galaxies in the same parameter correlation diagrams . 2
we discard most galaxies with inclinations @xmath36 .
broeils ( 1992 ) remarks that hi rotation curve derivations are less accurate when the galaxy is too face - on .
also , there is some danger that oval distortions ( bosma 1978 ; kormendy 1982 ) result in incorrect estimates of inclinations .
because they provide much - needed leverage at high luminosities , we kept four @xmath36 galaxies from abp : m101 , ngc 5236 , ngc 6946 , and ic 342 .
inclination is not a critical selection criterion ; most nearly face - on galaxies satisfy the dm correlations . 3
the most important selection cut is to ensure that rotation curves reach large enough radii to constrain the dm parameters .
after some experimentation , we decided to require that the rotation curve measurements reach out to at least 4.5 exponential scale lengths of the disk .
the peak in @xmath1 for an exponential occurs at 2.2 scale lengths ( freeman 1970 ) , so the above choice ensures that the outer disk rotation curve drops significantly over the radius range in which we have velocity data . observing a flat rotation curve then provides good constraints on dm parameters .
the fussy choice of the ratio 4.5 resulted from a desire to keep a few galaxies that provide leverage at the high - luminosity end of the correlations .
the radius cut is not applied slavishly ; we keep a few galaxies that slightly violate the above criterion ( ratio 3.5 4.4 ) but that are sufficiently halo dominated that the dm parameters are well determined .
these galaxies are ddo 127 , ddo 154 , ddo 168 , ngc 247 , and ic 2574 . in general , the radius cut is important ; if we do not use it , we get a substantially larger galaxy sample , and it mostly is consistent with the dm correlations , but it shows considerably larger scatter than figures 2 4 . besides the selection cuts , we adopt the following procedures to make parameters from different sources be as consistent with each other as possible .
following broeils ( 1992 ) , we base distances on the virgocentric flow model of kraan - korteweg ( 1986 ) . however , the zeropoint is based on distances from cepheids and from surface brightness fluctuations ( ferrarese et al . 2000 ; tonry et al .
the distance to the virgo cluster is taken to be @xmath37 = 16.5 mpc , corresponding to a hubble constant of @xmath38 km s@xmath39 mpc@xmath39 .
the center of the virgo cluster is assumed to be at ( @xmath40 , @xmath41 ) = ( 281@xmath42 , 75@xmath42 ) ( binggeli , tammann , & sandage 1987 ) .
the virgocentric infall velocity of the local group is assumed to be 220 km s@xmath39 .
when accurate distances of our galaxies are known , e.g. , from cepheids , they are adopted .
most sources used essentially the above distance scale ; when an author did not , we assumed that @xmath43 and that @xmath44 .
this is not strictly correct , because gas and dynamical masses scale differently with distance .
but the errors are small on the scales of figures 2 4 .
galactic absorption corrections are from burstein & heiles ( 1984 ) or equivalently from the rc3 ( de vaucouleurs et al .
absolute magnitudes are corrected for internal absorption as in tully & fouqu ( 1985 ) . in carrying out rotation curve decompositions ,
some authors model the dm with a nonsingular isothermal sphere ; then we adopt the parameters @xmath6 , @xmath7 , and @xmath8 without modification .
other authors use the `` pseudo - isothermal sphere '' ( hereafter the pits ) , i.e. , the approximation that the volume density is @xmath45 they then derive parameters @xmath7 , @xmath46 , and the asymptotic circular velocity @xmath47 as @xmath48 .
no compelling physical argument favors one model over the other .
however , if we want to combine data from different sources , we need to correct parameters determined using the pits to the ones that would have have been measured using the isothermal .
an exact correction is not possible , because equation ( 3 ) is a poor approximation to the isothermal sphere except as @xmath48 .
figure 1 shows three possible scalings of the rotation and velocity dispersion curves of the pits to the isothermal .
all three scalings are shown to emphasize the physical difference between the two models .
the scalings are not equally plausible , but each has a simple underlying motivation .
we will find that the best scaling is intermediate between the middle and right panels of figure 1 . in all three scalings ,
the central densities of the two models are the same .
the panels differ in how radii and velocities are scaled . at left
, the pits is scaled in @xmath49 and @xmath5 so that the central densities and asymptotic circular velocities are the same
. then the core radius @xmath6 of the isothermal is @xmath50 .
the dispersion scale in the top panel is the velocity dispersion of the isothermal sphere , but this is equal to the local dispersion of the pits only as @xmath48 . for most galaxies , the hi distribution does not extend beyond @xmath51 . here
the _ local _ dispersion of the pits is less than that of the isothermal ( fig . 1
, _ top left _ ) .
the reason is that an isotropic system with the density distribution of equation ( 3 ) is not isothermal ; its velocity dispersion is @xmath52 where @xmath53 .
since the analytic model matches the isothermal badly in both @xmath1 and @xmath54 , we discard this scaling .
given the limited extent of hi data , it seems more realistic to scale the pits so that its rotation curve is similar to that of the isothermal at @xmath55 .
we get such a scaling if we solve our problem with the halo dispersion by making the velocity dispersion of the corresponding isothermal be equal to the _ central _ dispersion of the pits . then @xmath56 and @xmath57 .
the resulting parameters are related in the normal way for an isothermal sphere , @xmath58 .
the virtue of this scaling is that the two halo rotation curves are similar over the radius range in which they are actually fitted to the data ( fig . 1 , _ bottom right _ ) .
however , there is a problem ( fig . 1 , _ top right _ ) .
this scaling forces @xmath47 to be much larger than the velocity of the flat part of the rotation curve .
an examination of published decompositions shows that authors almost never extrapolate implicitly to such a large maximum rotation velocity
one that is never observed .
the exceptions are some rotation curve decompositions in which @xmath59 over the whole radius range of the data , i.e. , cases in which @xmath6 is completely unconstrained .
examples are the decompositions of ngc 7331 and ngc 6674 in broeils ( 1992 ) .
we discard these decompositions anyway .
this suggests a compromise scaling like the one in the middle panel . here
@xmath47 is scaled to the _ maximum _ , not the asymptotic , rotation velocity of the isothermal .
then , when authors use equation ( 3 ) to derive @xmath7 , @xmath46 , and @xmath47 via rotation curve decomposition , we would adopt @xmath60 , the authors quoted value of @xmath7 , and a halo velocity dispersion @xmath61 .
again , these parameters are related as normal for an isothermal sphere . in practice , there is no guarantee that any of the above scalings represents what happens when one author uses the pits and another uses the isothermal in rotation curve decomposition .
we note again that figure 1 is included to emphasize that the pits is not very isothermal .
however , we will find in 3 that enough decompositions have been published using each of the above models so that we can derive the dm parameter correlations separately for each model .
we will then derive the best scaling of one model to the other by matching their respective dm correlations .
this scaling is used to construct figure 4 , which combines data from all sources .
it is intermediate between the middle and right - hand panels of figure 1 , i.e. , @xmath62 ; @xmath63 , and @xmath64 .
the smallest dsph galaxies allow us to greatly increase our leverage on correlations of dm parameters with galaxy luminosity .
note that only the halo central density can be measured .
pioneering observations by aaronson ( 1983 ) and by aaronson & olszewski ( 1987 ) showed that the stellar velocity dispersions in draco and umi are @xmath65 km s@xmath39 , much larger than expected if the galaxies are in equilibrium and if they consist only of old , metal - poor stars ( @xmath66 @xmath67 2.5 ) .
high dispersions imply mass - to - light ratios @xmath68
. moreover , the central dm densities @xmath69 to 1 @xmath70 pc@xmath71 are `` shockingly high . indeed , these are the highest central dm densities seen in any galaxy so far '' ( kormendy 1987a ) .
the latter result was an early sign of the correlations in figures 2 4 .
dm in dsph galaxies has important implications , so considerable effort has gone into trying to find some escape .
early worries included small - number statistics , measurement errors , atmospheric velocity jitter in the most luminous agb stars , and especially unrecognized binary stars .
these have largely been laid to rest as more and fainter stars have been measured , as the time baseline on measurements of individual stars has increased beyond a decade , and as different authors have proved to agree ( e.g. , armandroff , olswezski , & pryor 1995 ; olszewski , pryor , & armandroff 1996 , see tremaine 1987 ; pryor 1992 ; mateo 1994 , 1997 , 1998 for reviews ) . as the number of dsph galaxies with dispersion measurements has increased ( above reviews and mateo et al .
1998 ; cook et al . 1999 ; ct et al . 1999 ; and gallart et al . 2001 ) , escape routes that depend on rare events have become implausible .
these include the suggestion ( kuhn & miller 1989 ; kuhn 1993 ) that the stars formerly in dsph galaxies are unbound because of galactic tides , so we overestimate the masses of systems that are far from equilibrium .
the required orbital resonance works best if the dispersion is only marginally larger than the escape velocity and if not too many systems need special engineering .
but @xmath3 ratios of 10 100 ( not 2.5 ! ) imply velocities that are inflated by factors of @xmath72 .
piatek & pryor ( 1995 ) , oh , lin , & aarseth ( 1995 ) , sellwood & pryor ( 1998 ) , klessen , grebel , & harbeck ( 2003 ) , and wilkinson ( 2004 ) argue convincingly that tides do not inflate velocity dispersions this much , especially not without producing velocity gradients across the galaxies that would have been seen .
this remains true even though apparently extratidal stars have been seen in some dsphs ( e.g. , irwin & hatzidimitriou 1995 ; piatek et al .
2001 , 2002 ; palma et al .
in addition , some dsph galaxies are too far from our galaxy to be affected by tides ( mateo 1998 ) .
all nine dsph galaxies with dynamical analyses appear dm dominated .
it has become difficult to argue that we get fooled by special circumstances .
one dm alternative is not addressed by the above arguments : modified newtonian dynamics ( milgrom 1983a , b , c ; milgrom & bekenstein 1987 ) .
mond has been much debated in recent years and will be revisited during this conference .
thirty years of failure to identify all constituents of dm persuade us to treat even exotic alternatives with respect .
however , while the jury is out , we assume conventional gravity and treat measurements of high velocity dispersions in dsph galaxies as detections of dm .
accordingly , figures 2 and 4 include dm central densities for all galaxies with @xmath7 tabulated in mateo ( 1998 ) plus and ii ( ct et al .
this includes four di galaxies with virial measurements of @xmath7 ; three of these are based on hi and one ( lgs3 ) is based on stellar dynamics ( cook et al .
the central mass density has been corrected for visible matter by subtracting 2.5 times the central , @xmath5-band volume brightness .
for all galaxies except fornax , the correction is small . in fornax ( @xmath73 ) ,
visible matter accounts for approximately half of the central density . for this galaxy
, @xmath7 is very uncertain .
the `` error bars '' on @xmath7 require discussion .
the plotted densities are derived by assuming that @xmath6 is the same for the visible and dark matter . if @xmath74 as in the other galaxies in the figures , then @xmath7 is smaller by a factor of 0.46 ( pryor & kormendy 1990 ; lake 1990 ) .
anisotropic velocity dispersions reduce @xmath7 estimates by as much as an order of magnitude over the isotropic case ( pryor & kormendy 1990 ; mateo et al .
1993 ) , although the extreme models are not good fits to the data .
finally , model - independent lower limits to the dm density ( merritt 1987 ) are typically @xmath75 @xmath70 pc@xmath71 .
these limits are mostly too weak to be interesting . however , for leo ii , umi , and draco , they are @xmath76 , @xmath77 , and @xmath78 , respectively .
even these low values are reasonably consistent with the extrapolation of the dm correlations to low luminosities .
the correlations between halo @xmath6 , @xmath7 , and @xmath8 and galaxy absolute magnitude @xmath79 are illustrated in figures 2 4 .
figure 2 shows results derived using isothermal halos .
figure 3 shows results derived using pseudo - isothermal halos .
figure 4 combines all of the data , as discussed below .
the main result of this paper is that dm satisfies well defined scaling laws .
halos in less luminous galaxies have smaller core radii , higher central densities , and smaller central velocity dispersions .
the scatter about the correlations is slightly smaller in figure 2 than in figure 3 .
it is not surprising that decompositions based on the isothermal sphere are better behaved , given the slow rise of the pits rotation curve to its asymptotic value ( fig . 1 ) .
however , rotation curve decompositions based on isothermal and pits models separately give essentially the same correlations .
there are 31 galaxies with isothermal dm decompositions and 37 with pits decompositions .
only 13 galaxies are common to both samples .
so the results in figures 2 and 3 are largely independent .
clearly , we want to combine the two samples .
we do this by scaling the correlations in figure 3 to those in figure 2 .
least - squares fits to the correlations in figure 2 ( omitting ngc 4605 ) give : @xmath80 least - squares fits to the correlations in figure 3 give : @xmath81 the two samples have essentially the same average absolute magnitude : @xmath82 for objects analyzed with isothermals and @xmath83 for those analyzed with pseudo - isothermals .
the average of the above values is @xmath84 . requiring that the correlations agree at @xmath85 provides the scaling of the pits results to those measured with the isothermal sphere ( its ) : @xmath86 here we explicitly identify the parameters derived with the different dm models .
this scaling is intermediate between the middle and right panels of figure 1 .
figure 4 shows the combined sample after scaling of the pseudo - isothermal results to those determined with isothermal halos .
based on 55 galaxies spanning 9 magnitudes in @xmath79 , the correlations are quite robust .
uncertainties in rotation curve decomposition are unlikely to threaten them , although the slopes are still uncertain .
if we keep only the galaxies with the most leverage on the dm , say the 11 galaxies whose rotation curves reach out to at least 10 times the scale length of the disk ( one of these is ngc 3198 , in which the ratio is 11.4 ) , then we get essentially the same correlations with essentially the same scatter .
least - squares fits to the points in figure 4 give : @xmath87 in physically more transparent terms ( @xmath88 is @xmath89 ) , @xmath90 equations 26 and 27 imply that halo surface density , which is proportional to @xmath91 , is nearly independent of galaxy luminosity .
this near - independence is illustrated explicitly in figure 5 .
it implies a tully - fisher ( 1977 ) relation similar to equation 29 ( cf .
fall 2002 ) . since mass @xmath92 and since @xmath93 constant , @xmath94 .
then @xmath95 implies that @xmath96 .
we find that @xmath97 .
this suggests that @xmath98 .
if we correct for the slight dependence of halo surface density on @xmath99 ( fig . 5 ) , then this becomes @xmath100 .
slope uncertainties are significant , so we should be cautious in how much we interpret this result . however , it is a plausible estimate of the degree to which small galaxies are more dm dominated than large ones .
ugc 2885 ( @xmath101 ) is brighter than umi ( @xmath102 ) by 14.7 mag . over 15 mag
, @xmath100 implies that @xmath103 changes by a factor of 26 .
this is entirely reasonable . -35pt
scaling laws are new constraints on the nature of dm and on galaxy formation and evolution . most of these remain to be explored .
simple implications include : 1 the surprisingly high dm densities in dwarf spheroidals are normal for galaxies of such low luminosity .
this implies that dsphs are real galaxies and not tidal fragments .
tides almost certainly pull bound fragments out of more luminous galaxy progenitors , but they can not retain even the relatively low dm densities in those progenitors ( barnes & hernquist 1992 ) , much less increase the dm density to the high values characteristic of dwarf spheroidal galaxies . 2 dwarf spheroidal galaxies are not included in the least - squares fits in figures 2 and 4 because only @xmath7 can be derived for their halos . however , these @xmath7 values lie on the extrapolation to low luminosity of the correlations for spiral and irregular galaxies . that is , the dm halos of dsph and sc
i m galaxies appear to form a single physical sequence as a function of dm core mass .
3 since virialized density depends on collapse redshift @xmath11 , @xmath12 , the smallest dwarfs formed at least @xmath104 earlier than the biggest spirals .
correction for baryonic dm compression will make @xmath7 smaller for giant galaxies
. this will slightly increase @xmath105 .
4 the visible matter parameters @xmath6 and @xmath8 of dsphs are a factor of about 2 smaller than their extrapolated dm parameters .
this is reasonably consistent with the hypothesis that extreme dsphs have low visible matter densities ( @xmath106 ) because they lost most of their baryons early .
possible reasons include galactic winds ( e.g. , dekel & silk 1986 ) or the difficulty of holding onto baryons in shallow dm potential wells when the universe was ionized ( e.g. , klypin et al .
1999 ) . in the absence of a dark halo
, the loss of most baryons would unbind the few stars that had already formed .
but since these galaxies contain dm halos , we expect instead that the distribution of stars has expanded to fill the halo s core . unlike the situation in giant galaxies , visible matter and dm
would then have similar scale parameters . 5
the fact that , as luminosity decreases , dwarf galaxies become much more numerous and also more nearly dominated by dm raises the possibility that there exists a large population of objects that are completely dark ( freeman 1987 ; kormendy 1990 ; see also tully 2004 ) .
undiscovered dm dwarfs would help to solve the well known problem that the spectrum of initial density fluctuations predicted by cdm theory predicts far too many dwarf satellites of giant galaxies ( moore et al .
1999 ; klypin et al .
the favored explanation for why these dwarfs are not seen is that they virialized early , before or during the reionization of the universe , and therefore lost or never captured the canonical fraction of baryons because those baryons were too hot to be confined in the puny potential wells of the dark dwarfs .
our observations suggest that empty halos if they exist are likely to be small and dense and to have small total masses . they would be darker versions of draco and umi .
6 djorgovski ( 1992 ) has compared an earlier version of the dm parameter correlations to the scaling laws predicted by hierarchical clustering ( peebles 1974 ; gott & rees 1975 ) . for a power spectrum of initial density fluctuations that is a power law in wavenumber @xmath107 , @xmath15
, the size @xmath108 , density @xmath109 , and velocity dispersion @xmath8 of a bound object are related approximately by @xmath110 here we have used the relation @xmath111 for an isothermal sphere .
djorgovski pointed out that the dm parameter correlations in kormendy ( 1990 ) imply that @xmath112 , close to the value @xmath113 expected for giant galaxies in cdm theory . with the more accurate fits in equations 20 22 ,
@xmath114 we get @xmath115 , @xmath116 , and @xmath117 , respectively .
note that these values are not independent .
their average is @xmath118 .
if we use the fits ( equations 5 7 ) determined from decompositions using isothermal dm , then the average is @xmath119 .
both values are remarkably close to the value @xmath120 expected in @xmath121cdm theory at a halo mass of @xmath122 ( shapiro & iliev 2002 ) .
we need to correct the slopes for baryonic dm compression ; this will be done in kormendy & freeman ( 2003 ) .
the above comparison provides a measure of the slope of the fluctuation power spectrum on mass scales that are smaller than those accessible to most other methods .
shapiro & iliev ( 2002 ) have made a more detailed comparison of the dm parameter correlations published by kormendy & freeman ( 1996 ) with their predictions based on _
cobe_-normalized cdm fluctuation spectra .
they found that the agreement between predictions and observations was best for @xmath121cdm .
it is interesting to note a consequence of the theoretical prediction that the slope @xmath14 gets steeper at smaller mass scales .
if @xmath123 for the smallest dwarfs ( shapiro & iliev 2002 ; ricotti 2002 ) , then the straight lines in the left panels of figure 4 should curve downward toward the visible matter parameters of dsph galaxies .
this would strengthen the inference that the visible and dark matter in these galaxies is distributed similarly
. it will be important to look for curvature in the correlations as more data become available for dwarf galaxies .
finally , we note that the scatter in figures 2 4 has surely been increased by problems with the data .
( 1 ) distance errors are not negligible . for our calibrating galaxies , we can compare accurate distances to those given by our virgocentric flow field machinery .
this implies errors in @xmath124 of @xmath125 . since @xmath44 , distance errors are a significant although not the dominant source of scatter in equations 5 25 .
( 2 ) if some disks are submaximal , then this affects the scatter in the correlations . if the degree to which they are submaximal depends on @xmath79 ( kranz , slyz , & rix 2003 ) ,
this affects the correlation slopes , too .
( 3 ) the assumption that dm halos have isothermal cores is challenged by cdm theory , although it is suppported by many observations .
it will be important to see how the correlations are affected if nfw halos are used .
( 4 ) the correlations in figures 2 4 require correction for dm compression by the baryons before a definitive comparison with theory can be made
. we will address these issues in future papers .
jk is grateful to the staff of mt .
stromlo observatory for their hospitality during three visits when part of this work was done .
we thank s. djorgovski , s. m. fall , and p. shapiro for helpful discussions on the comparison of predicted and observed dm scaling laws .
this work used the nasa / ipac extragalactic database , which is operated by the jet propulsion laboratory , california institute of technology , under contract with nasa .
kormendy , j. , & freeman , k. c. 1996 , in ringberg proceedings 1996 of sonder - forschungsbereich 375 , research in particle astrophysics , ed . r. bender et al .
( munich : technische universitt mnchen , ludwig - maximilians - universitt , max - planck - institut fr physik , & max - planck - institut fr astrophysik ) , 13 | published mass models fitted to galaxy rotation curves are used to study the systematic properties of dark matter ( dm ) halos in late - type and dwarf spheroidal ( dsph ) galaxies .
halo parameters are derived by fitting non - singular isothermals to @xmath0 , where @xmath1 is the observed rotation curve and @xmath2 is the rotation curve of the visible matter .
the latter is calculated from the surface brightness assuming that the mass - to - light ratio @xmath3 is constant with radius .
`` maximum disk '' values of @xmath3 are adjusted to fit as much of the inner rotation curve as possible without making the halo have a hollow core .
rotation curve decomposition becomes impossible fainter than absolute magnitude @xmath4 , where @xmath5 becomes comparable to the velocity dispersion of the gas . to increase the luminosity range
further , we include dsph galaxies , which are physically related to spiral and irregular galaxies . combining the data
, we find that dm halos satisfy well defined scaling laws analogous to the `` fundamental plane '' relations for elliptical galaxies .
halos in less luminous galaxies have smaller core radii @xmath6 , higher central densities @xmath7 , and smaller central velocity dispersions @xmath8 . scaling laws
provide new and detailed constraints on the nature of dm and on galaxy formation and evolution .
some simple implications include : 1 a single , continuous physical sequence of increasing mass extends from dsph galaxies with @xmath9 to sci galaxies with @xmath10 . 2 the high dm densities in dsph galaxies are normal for such tiny galaxies . since virialized density depends on collapse redshift @xmath11 , @xmath12 , the smallest dwarfs formed at least @xmath13 earlier than the biggest spirals .
3 the high dm densities of dsphs implies that they are real galaxies formed from primordial density fluctuations .
they are not tidal fragments .
tidal dwarfs can not retain even the low dm densities of their giant - galaxy progenitors .
in contrast , dsphs have higher dm densities than do giant - galaxy progenitors . 4
the fact that , as luminosity decreases , dwarf galaxies become much more numerous and also more nearly dominated by dm raises the possibility that there exists a large population of objects that are completely dark .
such objects are a canonical prediction of cold dm theory . if they exist , `` empty halos '' are likely to be small and dense that is , darker versions of draco and umi
. 5 the slopes of the dm parameter correlations provide a measure on galactic mass scales of the slope @xmath14 of the power spectrum @xmath15 of primordial density fluctuations .
our preliminary results not yet corrected for baryonic compression of dm give @xmath16 .
this is consistent with cold dm theory .
# 1 # 1#1 # 1#1 # 1 # 1_#1 _ # 1_#1 _ = # 1 1.25 in .125 in .25 in |
the present work is a sequel to a work by bettinelli @xcite , whose aim is to investigate the topology of scaling limits for random maps of arbitrary genus .
a map is a cellular embedding of a finite graph ( possibly with multiple edges and loops ) into a compact connected orientable surface without boundary , considered up to orientation - preserving homeomorphisms . by _
, we mean that the faces of the map the connected components of the complement of edges are all homeomorphic to disks .
the genus of the map is defined as the genus of the surface into which it is embedded . for technical reasons
, it will be convenient to deal with rooted maps , meaning that one of the half - edges or oriented edges is distinguished .
we will particularly focus on bipartite quadrangulations : a map is a quadrangulation if all its faces have degree @xmath5 ; it is bipartite if each vertex can be colored in black or white , in such a way that no edge links two vertices that have the same color .
although in genus @xmath6 , all quadrangulations are bipartite , this is no longer true in positive genus @xmath7 . a natural way to generate a large random bipartite quadrangulation of genus
@xmath0 is to choose it uniformly at random from the set @xmath8 of all rooted bipartite quadrangulations of genus @xmath0 with @xmath3 faces , and then consider the limit as @xmath3 goes to infinity .
a natural setting for this problem is to consider quadrangulations as metric spaces endowed with their graph metric , properly rescaled by the factor @xmath4 @xcite and to study their limit in the gromov
hausdorff topology @xcite . from this point of view , the planar case @xmath6 has largely been studied during the last decade .
le gall @xcite showed the convergence of these metric spaces along some subsequence .
it is believed that the convergence holds without the `` along some subsequence '' part in the last sentence , and le gall gave a conjecture for a limiting space to this sequence @xcite .
although the whole convergence is yet to be proved , some information is available on the accumulation points of this sequence .
le gall and paulin @xcite proved that every possible limiting metric space is almost surely homeomorphic to the two - dimensional sphere .
miermont @xcite later gave a variant proof of this fact .
we showed in @xcite that the convergence along some subsequence still holds in any fixed positive genus @xmath0 . in this work , we show that the topology of every possible limiting space is that of the genus @xmath0-torus @xmath9 .
we will work in fixed genus @xmath0 . on the whole
, we will not let it figure in the notation , in order to lighten them .
as the case @xmath6 has already been studied , we suppose @xmath7 .
recall that the gromov
hausdorff distance between two compact metric spaces @xmath10 and @xmath11 is defined by @xmath12 where the infimum is taken over all isometric embeddings @xmath13 and @xmath14 of @xmath15 and @xmath16 into the same metric space @xmath17 , and @xmath18 stands for the usual hausdorff distance between compact subsets of @xmath19 .
this defines a metric on the set @xmath20 of isometry classes of compact metric spaces @xcite , theorem 7.3.30 , making it a polish space . for any map @xmath21 , we call @xmath22 its set of vertices
. there exists on @xmath22 a natural graph metric @xmath23 : for any vertices @xmath24 and @xmath25 , the distance @xmath26 is defined as the number of edges of any shortest path linking @xmath24 to @xmath27 .
the main result of @xcite is the following .
[ cvq ] let @xmath2 be uniformly distributed over the set @xmath8 of all bipartite quadrangulations of genus @xmath0 with @xmath3 faces .
then , from any increasing sequence of integers , we may extract a subsequence @xmath28 such that there exists a metric space @xmath29 satisfying @xmath30 in the sense of the gromov
hausdorff topology , where @xmath31 moreover , the hausdorff dimension of the limit space @xmath29 is almost surely equal to @xmath5 , regardless of the choice of the sequence of integers .
remark that the constant @xmath32 is not necessary in this statement ( simply change @xmath33 into @xmath34 ) .
we kept it for the sake of consistency with @xcite , and because of our definition of @xmath33 later in the paper , although it is irrelevant for the moment .
note also that , a priori , the metric space @xmath35 depends on the subsequence @xmath28 .
similarly to the planar case , we believe that the extraction in proposition [ cvq ] is not necessary , and we conjecture the space @xmath36 for the limit , where @xmath37 was defined at the end of section 6.3 in @xcite .
we also believe that the space @xmath38 is somewhat universal , in the sense that we conjecture it as the scaling limit of more general classes of random maps .
more precisely , we think that proposition [ cvq ] still holds while replacing the class of quadrangulations with some other `` reasonable '' class of maps , as well as the constant @xmath32 , which is inherent to the class of quadrangulations , with the appropriate constant . in particular , our approach can be generalized to the case of @xmath39-angulations , @xmath40 , by following the same lines as le gall in @xcite .
we may now state our main result , which identifies the topology of @xmath41 , regardless of the subsequence @xmath28 .
[ cvq2 ] the metric space @xmath42 is a.s .
homeomorphic to the @xmath0-torus @xmath9 .
in the general picture , we rely on the same techniques as in the planar case .
the starting point is to use a bijection due to chapuy , marcus and schaeffer @xcite between bipartite quadrangulations of genus @xmath0 with @xmath3 faces and so - called well - labeled @xmath0-trees with @xmath3 edges .
the study of the scaling limit as @xmath43 of uniform random well - labeled @xmath0-trees with @xmath3 edges was the major purpose of @xcite .
this study leads to the construction of a continuum random @xmath0-tree , which generalizes aldous s crt @xcite .
the first step of our proof is to carry out the analysis of le gall @xcite in the nonplanar case and see the space @xmath44 as a quotient of this continuum random @xmath0-tree via an equivalence relation defined in terms of brownian labels on it .
we then adapt miermont s approach @xcite , and use the notion of 1-regularity introduced by whyburn @xcite and studied by whyburn and begle @xcite in order to see that the genus remains the same in the limit . [ cols="^,^ " , ]
finally , we deduce the technical estimates we need from the planar case thanks to a bijection due to chapuy @xcite between well - labeled @xmath0-trees and well - labeled plane trees with @xmath0 distinguished triples of vertices .
we will use the background provided in @xcite .
we briefly recall it in section [ secprel ] . in section [ secrgt ] , we define real @xmath0-trees and explain how we may see @xmath35 as a quotient of such objects .
theorem [ ip ] in section [ secip ] gives a criteria telling which points are identified in this quotient , and section [ secsurf ] is dedicated to the proof of theorem [ cvq2 ] .
finally , we expose in section [ secprle ] chapuy s bijection , and use it to prove four technical lemmas stated during section [ secip ] .
in this section , we recall the notation , settings and results from @xcite that we will need for this work .
we refer the reader to @xcite for more details .
we use the following formalism for maps . for any map
@xmath21 , we denote by @xmath22 and @xmath45 , respectively , its sets of vertices and edges .
we also call @xmath46 its set of half - edges , and @xmath47 its root . for any half - edge @xmath48
, we write @xmath49 its reverse so that @xmath50as well as @xmath51 and @xmath52 its origin and end .
finally , we say that @xmath53 is an orientation of the half - edges if for every edge @xmath54 exactly one of @xmath48 or @xmath55 belongs to @xmath56.=-1 the first main tool we will need consists of the chapuy
schaeffer bijection @xcite , corollary 2 to theorem 1 , which allows us to code ( rooted ) quadrangulations by so - called well - labeled ( rooted ) @xmath0-trees .
a @xmath0-tree is a map of genus @xmath0 with only one face .
this notion naturally generalizes the notion of plane tree : in particular , @xmath57-trees are plane trees
. it may be convenient to represent a @xmath0-tree @xmath58 with @xmath3 edges by a @xmath59-gon whose edges are pairwise identified ( see figure [ gon ] ) .
we note @xmath60 , @xmath61 the half - edges of @xmath58 arranged according to the clockwise order around this @xmath59-gon .
the half - edges are said to be arranged according to the _ facial order _ of @xmath58 .
informally , for @xmath62 , @xmath63 is the `` first half - edge to the left after @xmath64 . ''
we call _ facial sequence _ of @xmath58 the sequence @xmath65 , @xmath66 defined by @xmath67 and for @xmath68 , @xmath69 .
imagine a fly flying along the boundary of the unique face of @xmath58 .
let it start at time @xmath57 by following the root @xmath70 , and let it take one unit of time to follow each half - edge , then @xmath71 is the vertex where the fly is at time @xmath72 .
let @xmath58 be a @xmath0-tree .
two vertices @xmath73 are said to be _ neighbors _ , and we write @xmath74 , if there is an edge linking them .
a _ well - labeled @xmath0-tree _ is a pair @xmath75 where @xmath58 is a @xmath0-tree and @xmath76 is a function ( thereafter called _ labeling function _ ) satisfying : @xmath77 , where @xmath70 is the root of @xmath58 ; if @xmath78 , then @xmath79 .
we call @xmath80 the set of all well - labeled @xmath0-trees with @xmath3 edges .
a _ pointed quadrangulation _ is a pair @xmath81 consisting in a quadrangulation @xmath82 together with a vertex @xmath83 .
we call @xmath84 the set of all pointed bipartite quadrangulations of genus @xmath0 with @xmath3 faces .
the chapuy marcus
schaeffer bijection is a bijection between the sets @xmath85 and @xmath86 .
we briefly describe here the mapping from @xmath85 onto @xmath86 , and we refer the reader to @xcite for a more precise description .
let @xmath87 be a well - labeled @xmath0-tree with @xmath3 edges and @xmath88 .
as above , we write @xmath65 , @xmath89 its facial sequence . the pointed quadrangulation @xmath90 corresponding to @xmath91 is then constructed as follows .
first , shift all the labels in such a way that the minimal label is equal to @xmath92 .
let us call @xmath93 this shifted labeling function . then , add an extra vertex @xmath94 carrying the label @xmath95 inside the only face of @xmath58 . finally , following the facial sequence , for every @xmath96 , draw an arc without crossing any edge of @xmath58 or arc already drawn between @xmath71 and @xmath97 , where @xmath98 is the _ successor _ of @xmath72 , defined by @xmath99 where @xmath100 , and with the conventions @xmath101 , and @xmath102 .
the quadrangulation @xmath82 is then defined as the map whose set of vertices is @xmath103 , whose edges are the arcs we drew and whose root is the first arc drawn , oriented _ from _
@xmath65 if @xmath104 or _ toward _
@xmath65 if @xmath105 ; see figure [ cms ] . . on the bottom left picture
, the vertex @xmath94 has a thicker ( red ) borderline . ] because of the way we drew the arcs of @xmath82 , we see that for any vertex @xmath106 , @xmath107 . when seen as a vertex in @xmath108 , we write @xmath109 instead of @xmath71
. in particular , @xmath110 .
we end this section by giving an upper bound for the distance between two vertices @xmath109 and @xmath111 , in terms of the labeling function @xmath112 : @xmath113\\[-8pt ] & & \qquad\le\mathfrak { l}({{\mathfrak{t}}}(i ) ) + \mathfrak{l}({{\mathfrak{t}}}(j ) ) - 2 \max\bigl(\min_{\mbox{\fontsize{8.36pt}{8.36pt}\selectfont{$k \mbox{$\in$ } \overrightarrow{{[\![}i , j { ] \!]}}$ } } } \mathfrak{l}({{\mathfrak{t}}}(k)),\min _ { \mbox{\fontsize{8.36pt}{8.36pt}\selectfont{$k \mbox{$\in$ } \overrightarrow { { [ \![}j , i { ] \!]}}$ } } } \mathfrak{l}({{\mathfrak{t}}}(k ) ) \bigr)+2,\nonumber\end{aligned}\ ] ] where we note , for @xmath114 , @xmath115\!]}{:=}[i , j ] \cap { \mathbb{z}}= \ { i , i+1 , \ldots , j \}$ ] , and @xmath116\ ! ] } } { : = } \cases { { [ \![}i , j { ] \ ! ] } , & \quad if $ i \le j$,\cr { [ \![}i,2n { ] \!]}\cup{[\![}0,j { ] \ ! ] } , & \quad if $ j < i$. } \ ] ] we refer the reader to @xcite , lemma 4 , for a detailed proof of this bound .
we explained in @xcite how to decompose a @xmath0-tree into simpler objects .
roughly speaking , a @xmath0-tree is a scheme ( a @xmath0-tree whose all vertices have degree at least @xmath117 ) in which every half - edge is replaced by a forest .
we adapt the standard formalism for plane trees as found in @xcite for instance to forests .
let us call @xmath118 , where @xmath119 .
if @xmath120 , we write @xmath121 . for @xmath122 , , we let @xmath123 be the concatenation of @xmath124 and @xmath125 .
if @xmath126 for some @xmath127 , we say that @xmath124 is a _ ancestor _ of @xmath128 and that @xmath128 is a _ descendant _ of @xmath124 . in the case
where @xmath129 , we may also use the terms _ parent _ and _ child _ instead . a _
forest _ is a finite subset @xmath130 satisfying the following : there is an integer @xmath131 such that @xmath132\!]}$ ] ; if @xmath133 , @xmath134 , then its parent belongs to @xmath135 ; for every @xmath133 , there is an integer @xmath136 such that @xmath137 if and only if @xmath138 ; @xmath139 . the integer @xmath140 is called the _ number of trees _ of @xmath135 . for @xmath141 , we call @xmath142 its oldest ancestor
. a _ tree _ of @xmath135 is a level set for @xmath143 : for @xmath144 , the @xmath145th tree of @xmath135 is the set @xmath146 .
the integer @xmath147 hence records which tree @xmath124 belongs to .
we call @xmath148 the _ floor _ of the forest @xmath135 . for @xmath149
, we write @xmath78 if either @xmath124 is a parent or child of @xmath125 , or @xmath150 and @xmath151 .
it is convenient , when representing a forest , to draw edges between @xmath124 s and @xmath125 s such that @xmath152 ; see figure [ dfs ] . we say that an edge drawn between a . ]
parent and its child is a _ tree edge _ whereas an edge drawn between two consecutive tree roots , that is , between some @xmath72 and @xmath153 , will be called a _
floor edge_. we call @xmath154 the set of all forests with @xmath155 trees and @xmath156 tree edges .
a _ well - labeled forest _ is a pair @xmath157 where @xmath135 is a forest , and @xmath158 is a function satisfying : for all @xmath159 , @xmath160 ; if @xmath78 , @xmath79 .
let @xmath161 be the set of well - labeled forests with @xmath155 trees and @xmath156 tree edges .
[ [ encoding - by - contour - and - spatial - contour - functions ] ] encoding by contour and spatial contour functions + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + there is a very convenient way to code forests and well - labeled forests .
let @xmath162 be a forest .
let us begin by defining its _ facial sequence _ @xmath163 as follows ( see figure [ dfs ] ) : @xmath164 , and for @xmath165 : * if @xmath166 has children that do not appear in the sequence @xmath167 , then @xmath168 is the first of these children , that is , @xmath169 where @xmath170 * otherwise , if @xmath166 has a parent [ i.e. , @xmath171 , then @xmath172 is this parent ; * if neither of these cases occur , which implies that @xmath173 , then @xmath174 . .
the paths are dashed on the intervals corresponding to floor edges . ]
each tree edge is visited exactly twice once going from the parent to the child , once going the other way around whereas each floor edge is visited only once from some @xmath72 to @xmath153 . as a result , @xmath175 . the _ contour pair _ @xmath176 of @xmath177 consists in the _ contour function _
@xmath178\to{\mathbb{r}}_+$ ] of @xmath135 and the _ spatial contour function _ @xmath179\to
{ \mathbb{r}}$ ] defined by @xmath180 and linearly interpolated between integer values ( see figure [ wlf ] ) .
it entirely determines @xmath177 .
we explain here how to decompose a well - labeled @xmath0-tree .
see @xcite for a more precise description .
we call _ scheme _ of genus @xmath0 a @xmath0-tree with no vertices of degree one or two .
a scheme is said to be _ dominant _ when it only has vertices of degree exactly three .
we call @xmath181 the finite set of all schemes of genus @xmath0 and @xmath182 the set of all dominant schemes of genus @xmath0 .
let us first explain how to decompose a @xmath0-tree ( without labels ) into a scheme , a family of forests and an integer .
let @xmath183 be a scheme .
we suppose that we have forests @xmath184 , @xmath185 , where for all @xmath48 , @xmath186 , as well as an integer @xmath187 , @xmath188\!]}$ ] , where @xmath189 denotes the root of @xmath183 .
we construct a @xmath0-tree as follows .
first , we replace every edge @xmath190 in @xmath183 with a chain of @xmath191 edges .
then , for every half - edge @xmath185 , we replace the chain of half - edges corresponding to it with the forest @xmath192 , in such a way that its floor matches with the chain . in other words , we `` graft '' the forest @xmath193 to the left of @xmath48 . finally , the root of the @xmath0-tree is the half - edge linking @xmath194 to @xmath195 in the forest grafted to the left of @xmath70 .
[ decompnolab ] the above description provides a bijection between the set of all @xmath0-trees and the set of all triples @xmath196 where @xmath197 is a scheme ( of genus @xmath0 ) , the forests @xmath198 are such that @xmath199 for all @xmath48 and @xmath200\!]}$ ] .
moreover , @xmath0-trees with @xmath3 edges correspond to triples satisfying the condition @xmath201 .
let @xmath58 be a @xmath0-tree and @xmath196 be the corresponding triple .
we say that @xmath183 is the scheme of @xmath58 and that the forests @xmath192 , @xmath202 , are its forests .
the set @xmath203 may be seen as a subset of @xmath58 ; we call _ nodes _ its elements . finally , we call _ floor _ of @xmath58 the set @xmath204 of vertices we obtain after replacing the edges of @xmath183 by chains of edges ( see figure [ structure ] )
. -tree @xmath58 into its scheme @xmath183 , the collection of its motzkin paths @xmath205 and the collection of its well - labeled forests @xmath206 . in this example ,
the integer @xmath207 .
the floor of @xmath58 is more thickly outlined , and its two nodes are even more thickly outlined . ]
we now deal with well - labeled @xmath0-trees .
we will need the following definition : we call _ motzkin path _ a sequence @xmath208 for some @xmath209 such that @xmath210 and for @xmath211 , @xmath212 .
we write @xmath213 its lifetime .
let @xmath183 be a scheme .
we suppose that we have well - labeled forests @xmath214 , @xmath215 , where for all @xmath48 , @xmath216 , as well as an integer @xmath217\!]}$ ] .
suppose moreover that we have a family of motzkin paths @xmath218 such that @xmath219 is defined on @xmath220\!]}$ ] and @xmath221 for some family of integers @xmath222 with @xmath223 .
we suppose that the motzkin paths satisfy the following relation : @xmath224 we will say that a quadruple @xmath225 satisfying these constraints is _
compatible_. we construct a well - labeled @xmath0-tree as follows .
we begin by suitably relabeling the forests . for every half - edge @xmath48 ,
first , we shift the labels of @xmath226 by @xmath227 so that it goes from @xmath227 to @xmath228
. then we shift all the labels of @xmath229 tree by tree according to the motzkin path : precisely , we change @xmath230 into @xmath231 .
then we replace the half - edge @xmath48 with this forest , as in the previous section . as before , we find the position of the root thanks to @xmath124 . finally , we shift all the labels for the root label to be equal to @xmath57.=-1 [ decwl ] the above description provides a bijection between the set of all well - labeled @xmath0-trees and the set of all compatible quadruples . moreover ,
@xmath0-trees with @xmath3 edges correspond to quadruples satisfying the condition @xmath232 .
if we call @xmath233 the contour pair of @xmath234 , then we may retrieve the oldest ancestor of @xmath235 thanks to @xmath236 by the relation @xmath237 where we use the notation @xmath238 } x\ ] ] for any process @xmath239 .
the function @xmath240 then records the labels of the forest @xmath192 , once shifted tree by tree according to the motzkin path @xmath226 .
this function will be used in section [ secset ] . through the chapuy marcus
schaeffer bijection , a uniform random quadrangulation corresponds to a uniform random well - labeled @xmath0-tree .
it can then be decomposed into a scheme , a collection of well - labeled forests , a collection of motzkin paths and an integer , as explained above .
the following section exposes the scaling limits of these objects .
let us define the space @xmath241 of continuous real - valued functions on @xmath242 killed at some time @xmath243,{\mathbb{r}}).\ ] ] for an element @xmath244 , we will define its lifetime @xmath245 as the only @xmath246 such that @xmath247,{\mathbb{r}})$ ] .
we endow this space with the following metric : @xmath248 throughout this section , @xmath156 and @xmath155 will denote positive real numbers and @xmath249 will be any real number .
we define here the brownian bridge @xmath250}^{0\to l}$ ] on @xmath251 $ ] from @xmath57 to @xmath249 and the first - passage brownian bridge @xmath252}^{0\to -\sigma}$ ] on @xmath251 $ ] from @xmath57 to @xmath253 . informally ,
@xmath250}^{0 \to l}$ ] and @xmath252}^{0 \to-\sigma}$ ] are a standard brownian motion @xmath254 on @xmath251 $ ] conditioned , respectively , on the events @xmath255 and @xmath256 .
because both theses events occur with probability @xmath57 , we need to define these objects properly .
there are several equivalent ways to do so ; see for example @xcite .
we call @xmath257 the density of a centered gaussian variable with variance @xmath24 , as well as @xmath258 its derivative @xmath259 let @xmath260 be a standard brownian motion . as explained in @xcite , proposition 1 , the law of the brownian bridge is characterized by the equation @xmath250}^{0\to l}(m ) = l$ ] and the formula @xmath261}^{0\to
l}(t)\bigr)_{0 \le t \le m ' } \bigr ) \bigr ] = \mathbb { e}\biggl [ f ( ( \beta_t)_{0 \le t \le m ' } ) \frac{{p}_{m - m'}(l-\beta _ { m'})}{{p}_m(l ) } \biggr]\ ] ] for all bounded measurable functions @xmath262 on @xmath241 , for all @xmath263 .
we define the law of the first - passage brownian bridge in a similar way , by letting @xmath264}^{0\to-\sigma}(t)\bigr)_{0 \le t \le m ' } \bigr ) \bigr]\nonumber\\[-8pt]\\[-8pt ] & & \qquad= \mathbb{e}\biggl [ f ( ( \beta_t)_{0 \le t \le m ' } ) \frac{{p}_{m - m'}'(-\sigma - \beta_{m'})}{{p}_m'(-\sigma ) } \mathbh{1}_{\ { { \underline{\beta}}_{m ' } > - \sigma\ } } \biggr ] \nonumber\end{aligned}\ ] ] for all bounded measurable functions @xmath262 on @xmath241 , for all @xmath265 and@xmath252}^{0\to-\sigma}(m ) = -\sigma$ ] . for any real numbers @xmath266 , @xmath267 , @xmath268
, we define the bridge on @xmath251 $ ] from @xmath266 to @xmath267 and the first - passage bridge on @xmath251 $ ] from @xmath269 to @xmath270 by @xmath271}^{l_1 \to l_2}(s ) { : = } l_1 + b_{[0,m]}^{0 \to l_2-l_1 } \quad\mbox{and}\quad f_{[0,m]}^{\sigma_1 \to\sigma_2 } { : = } \sigma_1 + f_{[0,m]}^{0 \to \sigma_2-\sigma_1}.\ ] ] see @xcite , section 5.1 , for a more precise description of these objects .
in particular , @xcite , lemmas 10 and 14 , show that these objects appear as the limits of their discrete analogs .
conditionally given a first - passage brownian bridge @xmath272}^{\sigma\to0}$ ] , we define a gaussian process @xmath273}(s))_{0\le s \le m}$ ] with covariance function @xmath274}(s ) , z_{[0,m]}(s')\bigr ) = \inf_{[s\wedge s',s\vee s ' ] } ( f - \underline{f}).\ ] ] the process @xmath275}^{\sigma\to0 } , z_{[0,m]})$ ] has the law of the so - called brownian snake s head ; see @xcite for more details .
recall that @xmath182 is the set of all dominant schemes of genus @xmath0 , that is , schemes with only vertices of degree @xmath117 . for any @xmath197
, we identify an element @xmath276 with an element of @xmath277 by setting : * @xmath278 , * @xmath279 for every @xmath280 , * @xmath281 .
we write @xmath282^{\vec e({{\mathfrak{s } } } ) } , \sum_{\mathfrak{e}\in\vec e({{\mathfrak{s } } } ) } x_\mathfrak{e}=1 \biggr\},\ ] ] the simplex of dimension @xmath283 .
note that @xmath156 lies in @xmath284 as long as @xmath285 .
we define the probability @xmath286 by , for all measurable function @xmath287 on @xmath288 $ ] , @xmath289 where @xmath290 , the measure @xmath291 is the lebesgue measure on the set @xmath292^{\vec e({{\mathfrak{s}}}){\setminus}\{\mathfrak{e}_*\}}\times ( { \mathbb{r}}_+^*)^{{\check}e({{\mathfrak{s}}})}\times{\mathbb{r}}^{v({{\mathfrak{s}}}){\setminus}\{\mathfrak{e}^-_*\}}\times[0,1]\ ] ] and @xmath293 is a normalization constant .
we gave a nonintegral expression for this constant in @xcite .
let @xmath294 be uniformly distributed over the set @xmath80 of well - labeled @xmath0-trees with @xmath3 vertices .
we call @xmath295 its scheme and we define , as in section [ decomp ] , @xmath296 its well - labeled forests , @xmath297 and @xmath298 , respectively , their sizes and lengths , @xmath299 the shifted labels of its nodes , @xmath300 its motzkin paths and @xmath301 the integer recording the position of the root in the first forest @xmath302 .
we call @xmath303 the contour pair of the well - labeled forest @xmath304 , and we extend the definition of @xmath305 to @xmath306 $ ] by linear interpolation .
we then define the rescaled versions of these objects [ recall that @xmath307 @xmath308 throughout this paper , the notation with a parenthesized @xmath3 will always refer to suitably rescaled objects , as in the definitions above .
we described in @xcite the limiting law of these objects : [ cvint ] the random vector @xmath309 converges in law toward the random vector @xmath310 whose law is defined as follows : * the law of the vector @xmath311 is the probability @xmath286 , * conditionally given @xmath312 : * * the processes @xmath313 , @xmath314 , and @xmath315 , @xmath316 , are independent ; * * the process @xmath313 has the law of a brownian snake s head on @xmath317 $ ] going from @xmath318 to @xmath57 @xmath319}^{\sigma _ \infty^\mathfrak{e}\to0 } , z_{[0,m_\infty^\mathfrak{e}]}\bigr);\ ] ] * * the process @xmath315 has the law of a brownian bridge on @xmath320 $ ] from @xmath57 to @xmath321 @xmath322}^{0 \to l_\infty^\mathfrak{e}};\ ] ] * * the motzkin paths are linked through the relation @xmath323 applying skorokhod s representation theorem , we may and will assume that this convergence holds almost surely . as a result ,
note that for @xmath3 large enough , @xmath324 .
let @xmath2 be uniformly distributed over the set @xmath8 of bipartite quadrangulations of genus @xmath0 with @xmath3 faces .
conditionally given @xmath326 , we take @xmath327 uniformly over @xmath328 so that @xmath329 is uniform over the set @xmath86 of pointed bipartite quadrangulations of genus @xmath0 with @xmath3 faces . recall that every element of @xmath8 has the same number of vertices , @xmath330 .
through the chapuy marcus
schaeffer bijection , @xmath331 corresponds to a uniform well - labeled @xmath0-tree with @xmath3 edges @xmath332 .
the parameter @xmath333 appearing in the bijection will be irrelevant to what follows .
recall the notation @xmath334 , @xmath335 and @xmath336 , @xmath337 from section [ seccms ] . for technical reasons , it will be more convenient , when traveling along the @xmath0-tree , not to begin by its root but rather by the first edge of the first forest . precisely , we define @xmath338 where @xmath301 is the integer recording the position of the root in the first forest of @xmath339 .
we define @xmath340 in a similar way , and endow @xmath341\!]}$ ] with the pseudo - metric @xmath342 defined by @xmath343 we define the equivalence relation @xmath344 on @xmath345\!]}$ ] by declaring that @xmath346 if @xmath347 , that is , if @xmath348 .
we call @xmath349 the canonical projection from @xmath341\!]}$ ] to @xmath341\!]}_{/\sim_n}$ ] , and we slightly abuse notation by seeing @xmath342 as a metric on @xmath341\!]}_{/\sim_n}$ ] defined by @xmath350 . in
what follows , we will always make the same abuse with every pseudo - metric .
the metric space @xmath351\!]}_{/\sim_n},d_n ) $ ] is then isometric to @xmath352 , which is at @xmath353-distance at most @xmath92 from the space @xmath354 .
we extend the definition of @xmath342 to noninteger values by linear interpolation and define its rescaled version : for @xmath355 $ ] , we let @xmath356 the spatial contour function of the pair @xmath294 is the function @xmath357 \to{\mathbb{r}}$ ] , defined by @xmath358 and linearly interpolated between integer values .
its rescaled version is @xmath359 recall definition ( [ lele ] ) of the process @xmath360 .
we define its rescaled version by @xmath361 proposition [ cvint ] shows that @xmath362 converges in the space @xmath363 toward @xmath364 we can express @xmath365 in terms of the processes @xmath362 s by concatenating them . for @xmath366 two functions started at @xmath57 , we call @xmath367 their concatenation defined by @xmath368 and , for @xmath369 , @xmath370 we arrange the half - edges of @xmath295 according to its facial order , beginning with the root @xmath371 , so that @xmath372 . by continuity of the concatenation , @xmath365 converges in @xmath373 toward @xmath374 , where the half - edges of @xmath375 are arranged in the same way . bound ( [ dlemme ] ) provides us with an upper bound on @xmath376 .
we define @xmath377\!]}}$ } } } { { \mathfrak{l}}}_n(k ) , \min _ { \mbox{\fontsize{8.36pt}{8.36pt}\selectfont{$k \mbox{$\in$ } \overrightarrow { { [ \![}j , i { ] \!]}}$ } } } { { \mathfrak{l}}}_n(k ) \bigr)+2,\ ] ] we extend it to @xmath378 $ ] by linear interpolation and define its rescaled version @xmath379 as we did for @xmath342 by ( [ rescdn ] ) .
we readily obtain that @xmath380 moreover , the process @xmath381 converges in @xmath382 ^ 2,{\mathbb{r}}),\|\cdot\|_\infty)$ ] toward the process @xmath383 defined by @xmath384}$ } } } { { \mathfrak{l}}}_\infty(x),\min _ { \mbox{\fontsize{8.36pt}{8.36pt}\selectfont{$x \mbox{$\in$ } \overrightarrow{[t , s]}$ } } } { { \mathfrak{l}}}_\infty(x ) \bigr),\ ] ] where @xmath385 } { : = } \cases { [ s , t ] , & \quad if $ s \le t$ , \cr [ s,1 ] \cup[0,t ] , & \quad if $ t < s$. } \ ] ] in @xcite , lemma 19 , we showed the tightness of the processes @xmath376 s laws thanks to the inequality ( [ bound2 ] ) . as a result , from any increasing sequence of integers
, we may extract a ( deterministic ) subsequence @xmath28 such that there exists a function @xmath386 ^ 2,{\mathbb{r}})$ ] satisfying @xmath387 by skorokhod s representation theorem , we will assume that this convergence holds almost surely .
we can check that the function @xmath388 is actually a pseudo - metric .
we define the equivalence relation associated with it by saying that @xmath389 if @xmath390 , and we call @xmath391_{/\sim _ \infty}$ ] .
we proved in @xcite that @xmath392 in the sense of the gromov
hausdorff topology .
from now on , we fix such a subsequence @xmath28 .
we will always focus on this particular subsequence in the following , and we will consider convergences when @xmath393 to hold along this particular subsequence .
in the discrete setting , it is sometimes convenient to work directly with the space @xmath339 instead of @xmath341\!]}$ ] . in the continuous setting
, we will see @xmath394 as a quotient of a continuous version of a @xmath0-tree , which we will call real @xmath0-tree .
in other words , we will see the identifications @xmath395 as of two different kinds : some are inherited `` from the @xmath0-tree structure , '' whereas the others come `` from the map structure . '' as @xmath0-trees generalize plane trees in genus @xmath0 , real @xmath0-trees are the objects that naturally generalize real trees .
we will only use basic facts on real trees in this work .
see , for example , @xcite for more detail .
we consider a fixed dominant scheme @xmath396 .
let @xmath397 and @xmath398 be two families of positive numbers satisfying @xmath399 and @xmath400 for all @xmath48 . as usual , we arrange the half - edges of @xmath183 according to its facial order , @xmath401 . for every @xmath402 , there exists a unique @xmath403 such that @xmath404 we let @xmath405 and @xmath406 . by convention ,
we set @xmath407 and @xmath408 . beware that these notions depend on the family @xmath409 .
there should be no ambiguity in what follows .
let us suppose we have a family @xmath410 of continuous functions @xmath411 \to{\mathbb{r}}_+$ ] such that @xmath412 and @xmath413 .
it will be useful to consider their concatenation : we define the continuous function @xmath414 \to{\mathbb{r}}_+$ ] going from @xmath415 to @xmath57 by @xmath416 we define the relation @xmath417 on @xmath325 $ ] as the coarsest equivalence relation for which @xmath418 if one of the following occurs : @xmath419 } h ; \\ \label{eqit22 } & & h(s ) = \underline h ( s),\qquad h(t ) = \underline h ( t),\nonumber\\[-8pt]\\[-8pt ] & & \qquad\mathfrak{e}(s)=\overline{\mathfrak{e}(t ) } \quad\mbox{and}\quad h^{\mathfrak{e}(s)}(\langle s \rangle ) = \sigma^{\mathfrak{e}(t ) } - h^{\mathfrak{e}(t)}(\langle t\rangle);\nonumber \\ \label{eqit23 } & & \langle s\rangle = \langle t\rangle= 0\quad\mbox{and}\quad \mathfrak{e}(s)^- = \mathfrak{e}(t)^-.\end{aligned}\ ] ] if we see the @xmath420 s as contour functions ( in a continuous setting ) , the first item identifies numbers coding the same point in one of the forests .
the second item identifies the floors of forests `` facing each other '' : the numbers @xmath421 and @xmath422 should code floor points ( two first equalities ) of forests facing each other ( third equality ) and correspond to the same point ( fourth equality ) .
finally , the third item identifies the nodes .
we call _ real @xmath0-tree _ any space @xmath423_{/\simeq}$ ] obtained by such a construction . , the two nodes @xmath3 and @xmath156 , an example of tree @xmath424 and an example of tree @xmath425 to the left of @xmath426}$ ] rooted at @xmath427 .
the set @xmath428 $ ] . on this picture ,
@xmath24 is an ancestor of @xmath27 and @xmath429 , and we can see the sets @xmath430}$ ] , @xmath431}$ ] and @xmath192 . ]
we now define the notions we will use throughout this work ( see figure [ not ] ) . for @xmath432 $ ] , we write @xmath433 its equivalence class in the quotient @xmath434_{/\simeq}$ ] . similarly to the discrete case , the floor of @xmath435 is defined as follows .
we call _ floor _ of @xmath435 the set @xmath436 . for @xmath437 ,
let @xmath438 and @xmath439 . the set @xmath440)$ ] is a real tree rooted at @xmath441 .
we call _ tree _ of @xmath435 a set of the form @xmath424 for any @xmath442 . if @xmath443 , we simply set @xmath444 .
let @xmath425 be a tree of @xmath435 rooted at @xmath427 , and @xmath445 .
we call @xmath446}$ ] the range of the unique injective path linking @xmath24 to @xmath27
. in particular , the set @xmath447}$ ] will be of interest .
it represents the ancestral lineage of @xmath24 in the tree @xmath425 .
we say that @xmath24 is an _ ancestor _ of @xmath27 , and we write @xmath448 , if @xmath449}$ ] .
we write @xmath450 if @xmath448 and @xmath451 .
let @xmath452 and @xmath453}{\setminus}\ { \rho_b , b\}$ ] .
let @xmath454 and @xmath455 .
then , provided @xmath456 , we call the _ tree to the left _ of @xmath457}$ ] rooted at @xmath427 the set @xmath458)$ ] .
we define the _ tree to the right _ of @xmath457}$ ] rooted at @xmath427 in a similar way , by replacing `` @xmath459 '' with `` @xmath460 '' in the definitions of @xmath461 and @xmath462 .
we call _ subtree _ of @xmath435 any tree of @xmath435 , or any tree to the left or right of @xmath457}$ ] for some @xmath463
. note that subtrees of @xmath435 are real trees , and that trees of @xmath435 are also subtrees of @xmath435 . for a subtree @xmath425 , the maximal interval @xmath464 $ ] such that @xmath465)$ ]
is called the _ interval coding _ the subtree @xmath425 . for @xmath215 , we call the _ forest to the left of @xmath48 _ the set @xmath466
. the _ nodes _ of @xmath435 are the elements of @xmath467 . in what follows
, we will identify the nodes of @xmath435 with the vertices of @xmath183 .
in particular , the two nodes @xmath468 and @xmath469 lie in @xmath192 .
we extend the definition of @xmath446}$ ] to the floor of @xmath192 : for @xmath470 , let @xmath471 be such that @xmath472 and @xmath473 .
we define @xmath474}{:=}{\mathscr{t}}([s\wedge t , s\vee t ] ) \cap{\mathit{fl}}\ ] ] the range of the unique because @xmath183 is a dominant scheme . ]
injective path from @xmath24 to @xmath27 that stays inside @xmath192 . for clarity
, we write the set @xmath475}$ ] simply as @xmath476}$ ] .
note that , in particular , @xmath476}= \mathfrak{f } ^\mathfrak{e}\cap\mathfrak{f}^{\bar\mathfrak{e}}=\mathfrak { f}^\mathfrak{e}\cap{\mathit{fl}}$ ] .
let @xmath477 .
there is a natural way , there are other possible ways to explore the @xmath0-tree between them . indeed , a point of @xmath204 is visited twice or three times if it is a node when we travel around @xmath204 .
in particular , this definition depends on the position of the root in @xmath183 for such points . in what follows
, we never use this definition for such points , so there will be no confusion . ] to explore @xmath435 from @xmath24 to @xmath27 . if @xmath478 , then let @xmath479 and @xmath480 .
if @xmath481 , then let @xmath482 and @xmath483 .
we define @xmath484 } { : = } { \mathscr{t}}(\overrightarrow{[s , t ] } ) , \ ] ] where @xmath485}$ ] is defined by ( [ orast ] ) .
we call @xmath486 ( resp .
, @xmath487 ) the real @xmath0-tree obtained from the scheme @xmath295 ( resp .
, @xmath375 ) and the family @xmath488 [ resp . , @xmath489 .
for the sake of consistency with @xcite , we call @xmath490 and @xmath491 the functions obtained by ( [ defconcat ] ) in this construction .
we also call @xmath492 and @xmath493 the corresponding equivalence relations .
when dealing with @xmath487 , we add an @xmath494 symbol to the notation defined above : for example , the floor of @xmath487 will be noted @xmath495 , and its forest to the left of @xmath48 will be noted @xmath496 . it is more natural to use @xmath339 rather than @xmath486 in the discrete setting .
as @xmath339 may be viewed as a subset of @xmath486 , we will use for @xmath339 the formalism we defined above simply by restriction . note that the notions of floor , forests , trees and nodes are consistent with the definitions we gave in section [ decomp ] in that case .
note that , because the functions @xmath497 s are first - passage brownian bridges , the probability that there exists @xmath498 such that @xmath499 for all @xmath500 is equal to @xmath57 . as a result , there are almost surely no trees rooted at the nodes of @xmath487 . moreover , the fact that the forests @xmath501 and @xmath502 are independent yields that , almost surely , we can not have a tree in @xmath192 and a tree in @xmath503 rooted at the same point . as a consequence ,
we see that , almost surely , all the points of @xmath487 are of order less than @xmath117 . consistently with the notation @xmath504 and @xmath505 in the discrete setting , we call @xmath506 [ resp . , @xmath507 the equivalence class of @xmath508 $ ] in @xmath509_{/\simeq_\infty}$ ] ( resp .
, in @xmath510_{/\sim_\infty}$ ] ) .
[ coarse ] the equivalence relation @xmath493 is coarser than @xmath511 , so that we can see @xmath394 as the quotient of @xmath487 by the equivalence relation on @xmath487 induced from @xmath512 . by definition of @xmath493 , it suffices to show that if @xmath513 satisfy ( [ eqit21 ] ) , ( [ eqit22 ] ) or ( [ eqit23 ] ) , then @xmath514 .
let us first suppose that @xmath421 and @xmath422 satisfy ( [ eqit21 ] ) , that is , @xmath515 } { { \mathfrak{c}}}_\infty.\ ] ] in a first time , we moreover suppose that @xmath516 for all @xmath517 .
using proposition [ cvint ] , we can find integers @xmath518 such that @xmath519 and @xmath520 } { { \mathfrak{c}}}_{(n)}$ ] .
the latter condition imposes that
@xmath521 so that @xmath522 and @xmath389 by ( [ dinfty ] ) .
equation ( [ f ] ) shows that , for every @xmath48 , the law of @xmath497 is absolutely continuous with respect to the wiener measure on any interval @xmath523 $ ] , for @xmath498 . because local minimums of brownian motion are pairwise distinct , this is also true for any @xmath497 , and thus for the whole process @xmath491 by construction .
if there exists @xmath517 for which @xmath524 , it is thus unique .
we may then apply the previous reasoning to @xmath525 and @xmath526 and find that @xmath527 and @xmath528 , so that @xmath395 .
let us now suppose that @xmath421 and @xmath422 satisfy ( [ eqit22 ] ) . if there is @xmath529 such that @xmath530 , then @xmath531 by ( [ eqit21 ] ) .
the same holds with @xmath422 instead of @xmath421 .
we may thus restrict our attention to @xmath421 and @xmath422 for which @xmath532 for all @xmath533 and @xmath534 for all @xmath535 .
let us call @xmath536 . in order to avoid confusion
, we use the notation @xmath537 and @xmath538 when dealing with the functions @xmath539 s . we know that for @xmath3 large enough , we have @xmath540 .
we only consider such @xmath3 s in the following .
we first find @xmath541 such that @xmath542 , @xmath543 , and @xmath544 .
we define @xmath545\!]}\dvtx \mathfrak{e } _ n(r)=\bar\mathfrak{e } , { { \mathfrak{c}}}_{(n)}^{\bar\mathfrak{e}}(\langle r \rangle_n ) = \sigma _ { ( n)}^{\mathfrak{e } } - { { \mathfrak{c}}}_{(n)}^{\mathfrak{e}}\bigl(\bigl\langle s_{(n)}\bigr\rangle_n\bigr ) \biggr\},\ ] ] so that @xmath546 , and then @xmath547 . taking an extraction
if needed , we may suppose that @xmath548 . by construction , @xmath549 and @xmath550 .
so @xmath551 and @xmath422 fulfill requirement ( [ eqit21 ] ) and @xmath552 by the above argument . the case of ( [ eqit23 ] ) is easier and may be treated in a similar way .
this lemma allows us to define a pseudo - metric and an equivalence relation on @xmath487 , still denoted by @xmath388 and @xmath553 , by setting @xmath554 and declaring @xmath555 if @xmath389 . the metric space @xmath35 is then isometric to @xmath556 .
we define @xmath557 on @xmath487 by letting @xmath558 we will see in lemma [ min ] that there is a.s . only one point where the function @xmath559 reaches its minimum . on this event ,
the following lemma holds .
[ dssb ] @xmath560let @xmath561 be the unique point where @xmath559 reaches its minimum
. then @xmath562 moreover , @xmath395 implies @xmath563 .
this readily comes from the discrete setting .
let @xmath564 be an integer where @xmath565 reaches its minimum . by extracting if necessary
, we may suppose that @xmath566 converges and its limit is necessarily @xmath561 .
let @xmath567 be such that . from the chapuy
schaeffer bijection , @xmath568 .
letting @xmath569 after renormalizing yields the first assertion .
the second one follows from the first one and the triangle inequality . as a result of lemmas [ coarse ] and [ dssb ]
, we can define @xmath559 on @xmath487 by @xmath570 .
when @xmath571 , we have @xmath572 } { { \mathfrak{l}}}_\infty(x),\min _ { x \in[b , a ] } { { \mathfrak{l}}}_\infty(x ) \bigr),\ ] ] where @xmath428 $ ] was defined by ( [ ab ] ) .
this section is dedicated to the proof of the following theorem : [ ip ] almost surely , for every @xmath573 , @xmath574 is equivalent to @xmath575 .
we already know that @xmath575 implies @xmath574 from the bound @xmath576 .
we will show the converse through a series of lemmas .
we adapt the approach of le gall @xcite to our setting .
let us begin by giving some information on the process @xmath577 .
[ min ] the set of points where @xmath559 reaches its minimum is a.s . a singleton .
let @xmath578 \to{\mathbb{r}}$ ] be a continuous function .
we say that @xmath402 is a _ right - increase point _ of @xmath262 if there exists @xmath579 $ ] such that @xmath580 for all @xmath581 .
a _ left - increase point _ is defined in a symmetric way .
we call @xmath582 the set of all ( left or right ) increase points of @xmath262 .
[ pc ] a.s .
, @xmath583 and @xmath584 are disjoint sets .
as the proofs of these lemmas are rather technical and unrelated to what follows , we postpone them to section [ secprle ] . in what follows , every discrete path denoted by the letter `` @xmath585 '' will always be a path in the _ map _ , never in the tree , that is , a path using the edges of the map .
let @xmath425 be a subtree of @xmath339 and @xmath586 be a path in @xmath2 that avoids the base point @xmath327 .
we say that the arc @xmath587 enters the subtree @xmath425 from the left ( resp .
, from the right ) if @xmath588 , @xmath589 and @xmath590 [ resp . , @xmath591 .
we say that the path @xmath585 _ passes through _ the subtree @xmath425 between times @xmath72 and @xmath145 , where @xmath592 , if : * @xmath593 ; @xmath594\ ! ] } ) \subseteq\tau$ ] ; @xmath595 , * @xmath596 .
the first condition states that @xmath585 `` visits '' @xmath425 , whereas the second one ensures that it really goes `` through . ''
it enters and exits @xmath425 going `` in the same direction . ''
we say that a vertex @xmath597 converges toward a point @xmath598 if there exists a sequence of integers @xmath599\!]}$ ] coding @xmath600 [ i.e. , @xmath601 such that @xmath602 admits a limit @xmath421 satisfying @xmath603 .
let @xmath604\!]}$ ] be the intervals coding subtrees @xmath605 .
we say that the subtree @xmath606 converges toward a subtree @xmath607 if the sequences @xmath608 and @xmath609 admit limits @xmath249 and @xmath610 such that the interval coding @xmath425 is @xmath611 $ ] .
the following lemma is adapted from le gall @xcite , end of proposition 4.2 .
[ legalllem ] with full probability , the following occurs .
let @xmath612 be such that @xmath613 .
we suppose that there exists a subtree @xmath425 rooted at @xmath427 such that @xmath614 .
we further suppose that we can find vertices @xmath600 , @xmath615 and subtrees @xmath606 in @xmath339 converging , respectively , toward @xmath24 , @xmath27 , @xmath425 and satisfying the following property : for infinitely many @xmath3 s , there exists a geodesic path @xmath616 in @xmath2 from @xmath600 to @xmath617 that avoids the base point @xmath327 and passes through the subtree @xmath606 .
then , @xmath618 .
the idea is that if @xmath24 and @xmath27 were identified , then all the points in the discrete subtrees close ( in a certain sense ) to the geodesic path would be close to @xmath24 in the limit .
fine estimates on the sizes of balls yield the result .
we proceed to the rigorous proof .
we reason by contradiction and suppose that @xmath619 .
we only consider integers @xmath3 for which the hypothesis holds .
we call @xmath620 the root of @xmath606 , and we set , for @xmath498 , @xmath621 } , { { \mathfrak{l}}}_\infty ( x ) > { { \mathfrak{l}}}_\infty(a)+\frac{\varepsilon}8 \biggr\}.\ ] ] we first show that @xmath622 , where @xmath623 denotes the closed ball of radius @xmath624 centered at @xmath24 in the metric space @xmath625
. let @xmath626 .
we can find @xmath627 converging toward @xmath628 .
for @xmath3 large enough , we have @xmath629}\qquad \mathfrak{l}_n(x ) \ge\mathfrak { l}_n(a_n ) + \frac{\varepsilon}{16 } n^{1/4}.\end{aligned}\ ] ] the first inequality comes from the fact that @xmath574 .
the second inequality is a consequence of the first one .
the third inequality holds because @xmath630 and @xmath631 .
finally , the fourth inequality follows by compactness of @xmath632}$ ] . from now on ,
we only consider such @xmath3 s .
we call @xmath633 the last integer coding @xmath634 , and @xmath635\!]}$ ] the interval coding @xmath606 .
we also call @xmath636 two integers such that @xmath616 passes through @xmath606 between times @xmath72 and @xmath145 . for the sake of simplicity
, we suppose that @xmath616 enters @xmath606 from the left
. enters @xmath606 from the right may be treated by considering the path @xmath637 instead of @xmath616 . ]
notice that the path @xmath616 does not intersect @xmath638}$ ] , because the labels on @xmath638}$ ] are strictly greater than the labels on @xmath616 .
let @xmath639 be the largest integer in @xmath640\!]}$ ] such that @xmath641 belongs to the set @xmath642\!]})$ ] .
then @xmath643\!]})$ ]
. moreover , @xmath644 : otherwise , all the vertices in @xmath645 $ ] would have labels greater than @xmath646 , and it is easy to see that this would prohibit @xmath616 from exiting @xmath606 by going `` to the right , '' in the sense that we would not have @xmath647 . as a result , when performing the chapuy marcus schaeffer bijection for the arc linking @xmath641 to @xmath648 , we have to visit @xmath634 .
then , going through consecutive successors of @xmath649 , we are bound to hit @xmath648 , so that @xmath650 .
this yields that @xmath651 , and , by taking the limit , @xmath652 .
we conclude thanks to two lemmas , whose proofs are postponed to section [ secprle ] .
they are derived from similar results in the planar case : @xcite , lemma 2.4 , and @xcite , corollary 6.2 .
we call @xmath653 the volume measure on @xmath654 , that is , the image of the lebesgue measure on @xmath325 $ ] by the canonical projection from @xmath325 $ ] to @xmath394 .
[ leme2 ] almost surely , for every @xmath655 and every subtree @xmath425 rooted at @xmath427 , the condition @xmath656 implies that @xmath657 } , { { \mathfrak{l}}}_\infty(x ) > { { \mathfrak{l}}}_\infty ( \rho)-\eta + \frac{\varepsilon}8 \biggr\}\biggr ) > 0.\end{aligned}\ ] ] [ leme4 ] let @xmath658 $ ] . for every @xmath659 , @xmath660 < \infty.\ ] ] we apply lemma [ leme2 ] to @xmath425 and @xmath661 , and we find that , for @xmath662 small enough , @xmath663 the inclusion @xmath664 yields that @xmath665 lemma [ leme4 ] applied to @xmath666 and @xmath667 yields that @xmath668 is integrable , so that @xmath669 a.s .
this is a contradiction .
we call @xmath670 the floor of @xmath339 . let @xmath671\!]}$ ] , and let @xmath98 be its successor in @xmath294 , defined by ( [ suc ] ) .
is the set of ( blue ) large dots . ]
we moreover suppose that @xmath672 .
we say that the arc linking @xmath504 to @xmath673 _ overflies _ the set @xmath674\ ! ] } } ) \cap{\mathit{fl}}_n,\ ] ] where @xmath675\!]}}$ ] was defined by ( [ oraij ] ) .
we define the set overflown by a path @xmath585 in @xmath2 that avoids the base point @xmath327 as the union of the sets its arcs overfly ; see figure [ of ] .
we denote it by @xmath676 .
[ lemof ] let @xmath677 and @xmath678 , @xmath679 .
we suppose that , for @xmath3 sufficiently large , there exist vertices @xmath680 , @xmath681 and @xmath600 , @xmath615 converging , respectively , toward @xmath678 , @xmath254 , @xmath24 and @xmath27 . if , for infinitely many @xmath3 s , there exists a geodesic path @xmath616 from @xmath600 to @xmath617 that overflies @xmath682}$ ] , then for all @xmath683}$ ] , @xmath684 moreover , if there exists @xmath683}$ ] for which @xmath685 , then .
let @xmath686}$ ] .
we can find vertices @xmath687}$ ] converging to @xmath429 . by definition
, there is an arc of @xmath616 that overflies @xmath688 .
say it links a vertex labeled @xmath249 to a vertex @xmath125 labeled @xmath689 . from the chapuy
schaeffer construction , we readily obtain that @xmath690 .
using the fact that @xmath691 , we find @xmath692 moreover , we can construct a path from @xmath688 to @xmath125 going through consecutive successors of @xmath688 . as a result , @xmath693 , so that @xmath694 both claims follow by taking limits in these inequalities after renormalization , and by using the fact that @xmath695 .
we proceed in three steps .
we first show that points of @xmath495 are not identified with any other points , then that points can not be identified with their strict ancestors , and finally theorem [ ip ] .
[ lemfl ] a.s . , if @xmath696 and @xmath697 are such that @xmath698 , then @xmath699 .
let @xmath700 and @xmath701 be such that @xmath574 .
we first suppose that @xmath24 is not a node .
there exists @xmath702 such that @xmath703 , and we can find @xmath421 , @xmath422 satisfying @xmath704 , @xmath705 and @xmath706 . without loss of generality
, we may suppose that @xmath707 . until further notice
, we will moreover suppose that @xmath708}$ ] .
we restrict ourselves to the case @xmath324 , which happens for @xmath3 sufficiently large .
we can find @xmath709 and @xmath710 converging toward @xmath24 and @xmath27 and satisfying .
let @xmath616 be a geodesic path ( in @xmath2 , for @xmath711 ) from @xmath600 to @xmath617 .
it has to overfly at least @xmath712}$ ] or @xmath713}$ ] .
indeed , every pair @xmath714}\times{[[}a_n , \mathfrak{e}^+ { ] ] } $ ] breaks @xmath339 into connected components , and the points @xmath600 and @xmath617 do not belong to the same of these components .
there has to be an arc of @xmath616 that links a point belonging to the component containing @xmath600 to one of the other components .
such an arc overflies @xmath246 or @xmath628 .
let us suppose that , for infinitely many @xmath3 s , @xmath616 overflies @xmath715}$ ] .
then , lemma [ lemof ] ensures that @xmath716 for all @xmath717}$ ] .
properties of brownian snakes show that the labels on @xmath718}$ ] are brownian . precisely , we may code @xmath476}$ ] by the interval @xmath719 $ ] as follows . for @xmath720 $ ] ,
we define @xmath721 . then @xmath476}= { \mathscr{t}}_\infty ( \{t_x , 0 \le x \le\sigma_\mathfrak{e}\})$ ] , and @xmath722 where , conditionally given @xmath312 , the process @xmath723 ( defined during proposition [ cvint ] ) has the law of a certain brownian bridge .
using the fact that local minimums of brownian motion are distinct , we can find @xmath724}{\setminus}\{a\}$ ] such that @xmath725 for all @xmath726}{\setminus}\{a\}$ ] . because @xmath696 , @xmath421 and @xmath422 are both increase points of @xmath491 and thus are not increase points of @xmath559 , by lemma [ pc ] . as a result ,
there exist two trees @xmath727 and @xmath728 rooted at @xmath729 , @xmath730}{\setminus}\{a\}$ ] satisfying @xmath731 ( see figure [ tau1tau2 ] ) . and @xmath732 . ]
passing through the tree @xmath733 . ]
similarly , if for infinitely many @xmath3 s , @xmath616 overflies @xmath734}$ ] , then we can find two trees @xmath735 and @xmath736 rooted at @xmath737 , @xmath738}{\setminus}\{a\}$ ] satisfying @xmath739 , and @xmath740 for all @xmath741}$ ] .
three cases may occur : for @xmath3 large enough , @xmath616 does not overfly @xmath742}$ ] ( and therefore overflies @xmath715}$ ] ) ; for @xmath3 large enough , @xmath616 does not overfly @xmath712}$ ] ( and therefore overflies @xmath713}$ ] ) ; for infinitely many @xmath3 s , @xmath616 overflies @xmath743}$ ] , and for infinitely many @xmath3 s , @xmath616 overflies @xmath744}$ ] . in case ( i ) , the trees @xmath745 and @xmath732 are well defined .
let @xmath746 , @xmath747 be trees rooted at @xmath748 , @xmath749}$ ] converging to @xmath745 and @xmath732 .
we claim that , for @xmath3 sufficiently large , @xmath616 passes through @xmath733 or @xmath750 .
first , notice that , for @xmath3 large enough , @xmath751}= \varnothing$ ] .
otherwise , for infinitely many @xmath3 s , we could find @xmath752}$ ] , and , up to extraction , we would have @xmath753}\subseteq{[[}a , d { ] ] } { \setminus}\{a\}$ ] .
furthermore , @xmath754 so that @xmath755 , and @xmath756 by lemma [ dssb ] , which is impossible .
for @xmath3 even larger , it holds that @xmath757 . roughly speaking
, @xmath616 can not go from a tree located at the right of @xmath733 ( resp .
, at the left of @xmath758 ) to a tree located at its left in @xmath759 ( resp . , to a tree located at its right in @xmath760 ) without entering it .
then @xmath761 has to enter @xmath733 from the right or @xmath750 from the left and pass through one of these trees ( see figure [ taun12 ] ) .
more precisely , we call @xmath762\!]}$ ] and @xmath763\!]}$ ] the sets coding the subtrees @xmath733 and @xmath750 . let @xmath764}$ ] be a point that is not overflown by @xmath616 , @xmath765 and @xmath766 . then , we let @xmath767\ ! ] } } \cup \overrightarrow{{[\![}q_n , s_n^2 { ] \ ! ] } } ) .\ ] ] we call @xmath768 the last point of @xmath616 belonging to @xmath769 .
such a point exists because @xmath770 and @xmath771 .
the remarks in the preceding paragraphs yield that neither @xmath768 nor @xmath772 belong to @xmath773}$ ] , and , because of the way arcs are constructed in the chapuy
schaeffer bijection , we see that . without loss of generality
, we may assume that @xmath774 .
because @xmath616 does not overfly @xmath775 , it enters @xmath733 from the right at time @xmath72 , that is , @xmath776 .
let @xmath777 be the first point after @xmath778 not belonging to @xmath733 .
it exists because @xmath779 .
then , because @xmath780 and @xmath616 does not overfly @xmath775 , we see that @xmath781 , so that @xmath616 passes through @xmath733 between times @xmath72 and @xmath145 . in case ( ii )
, we apply the same reasoning with @xmath782 and @xmath783 instead of @xmath745 and @xmath732 . in case ( iii ) , the four trees @xmath784 , @xmath732 , @xmath782 and @xmath783 are well defined , and we obtain that @xmath616 has to pass through one of their discrete approximations .
we then conclude by lemma [ legalllem ] that @xmath618 , which contradicts our hypothesis .
we treat the case where @xmath785}{\setminus}\{a\}$ ] in a similar way , simply by replacing @xmath52 ( resp .
, @xmath51 ) by @xmath786 if @xmath787}$ ] ( resp . , @xmath788}$ ] ) .
when @xmath24 is a node , we apply the same arguments , finding up to six trees ( one for each forest containing @xmath24 ) .
finally , if @xmath789 , then @xmath24 is a strict ancestor of @xmath27 .
this will be a particular case of lemma [ lemad ] .
[ lemad ] a.s .
, for every @xmath573 such that @xmath790 and @xmath791 , we have @xmath792 .
the proof of this lemma uses the same kind of arguments we used in section [ secfp ] , is slightly easier than the proof of lemma [ lemfl ] and is very similar to le gall s proof for proposition 4.2 in @xcite , so that we leave the details to the reader .
[ lemst ] a.s .
, for every tree @xmath607 rooted at @xmath794 and all @xmath795 satisfying @xmath698 , we have @xmath793 .
let @xmath607 be a tree rooted at @xmath796 and @xmath795 satisfying @xmath451 and @xmath797 . by lemma [ lemad ] , we know that @xmath798 and @xmath799 . as a consequence
, we have either @xmath513 for all @xmath800 or @xmath801 for all @xmath802 . without loss of generality
, we will assume that the first case occurs .
let us suppose that @xmath803 .
by lemma [ dssb ] , we know that @xmath613 , and by ( [ dzero ] ) , we have both @xmath804 } { { \mathfrak{l}}}_\infty < { { \mathfrak{l}}}_\infty(a)$ ] and @xmath805 } { { \mathfrak{l}}}_\infty < { { \mathfrak{l}}}_\infty(a)$ ] . as a result , there are two subtrees @xmath806 $ ] and @xmath807 $ ] rooted at @xmath808}{\setminus}\{a , b\}$ ] and @xmath809}\cup{[[}\rho , b { ] ] } \cup{\mathit{fl}}_\infty ) { \setminus}\{a , b\}$ ] satisfying @xmath810 .
let @xmath605 be a tree rooted at @xmath620 and @xmath600 , @xmath615 be points converging to @xmath425 , @xmath24 , and @xmath27 .
let @xmath811 $ ] and @xmath812 $ ] be subtrees rooted at @xmath813}{\setminus}\{a_n , b_n\}$ ] and @xmath814}\cup{[[}\rho_n , b_n { ] ] } \cup { \mathit{fl}}_n ) { \setminus}\{a_n , b_n\}$ ] converging toward @xmath745 and @xmath732 .
we consider a geodesic path @xmath616 from @xmath600 to @xmath617 .
recall that @xmath698 implies that @xmath815 . because every point in @xmath816}$ ] is a strict ancestor to @xmath24 or @xmath27 , for @xmath3 large enough , @xmath616 does not intersect @xmath817}$ ] .
otherwise , we could find an accumulation point @xmath678 identified with @xmath24 and @xmath27 , such that @xmath818 or @xmath819 ( possibly both ) , and this would contradict lemma [ lemad ] . if @xmath820 , for @xmath3 large , @xmath616 does not intersect @xmath821}$ ] either .
the same reasoning yields that @xmath616 does not intersect @xmath670 for @xmath3 sufficiently large , because of lemma [ lemfl ] .
let @xmath762\!]}$ ] and @xmath822\!]}$ ] be the sets coding the subtrees @xmath733 and @xmath750 .
we let @xmath823\ ! ] } } ) \quad\mbox{and}\quad b_n { : = } { \dot}{{\mathfrak{t}}}_n(\overrightarrow{{[\![}t_n^1,s_n^2 { ] \ ! ] } } ) .\vadjust{\goodbreak}\ ] ] by convention , if @xmath824 , we set @xmath825}{:=}\varnothing$ ] .
it is easy to see that @xmath770 , @xmath826 , @xmath827}\cup { [ [ } \rho_n , \rho_n^2 { ] ] } \cup{\mathit{fl}}_n$ ] and @xmath828 .
we conclude as in the proof of lemma [ lemfl ] .
we call @xmath768 the last point of @xmath616 belonging to @xmath769 .
such a point exists because @xmath770 and @xmath771 .
the remarks in the preceding paragraphs yield that , for @xmath3 large enough , neither @xmath829 nor @xmath778 belong to @xmath830 . for @xmath3
even larger , @xmath831 , and because of the way arcs are constructed in the chapuy
schaeffer bijection , we see that @xmath832 .
the path @xmath616 either enters @xmath733 from the left or enters @xmath750 from the right . without loss of generality , we may suppose that @xmath833 .
let @xmath834 be the first point after @xmath778 not belonging to @xmath733 . then @xmath835 . if @xmath616 passes through @xmath733 between times @xmath72 and @xmath836 , we are done .
otherwise , @xmath837 because of the condition @xmath838 ( informally , @xmath616 can not pass over @xmath750 without entering it ) .
we consider the first point @xmath839 after @xmath840 not belonging to @xmath750 , and reiterate the argument . because @xmath616 is a finite path
, we see that @xmath616 will eventually pass through @xmath733 or @xmath750 ; see figure [ fig10 ]
. passing through the subtree @xmath733 .
] if @xmath616 passes through @xmath733 ( resp . ,
@xmath750 ) for infinitely many @xmath3 s , a reasoning similar to the one we used in the proof of lemma [ lemof ] yields that @xmath841 [ resp . , @xmath842 .
we conclude by lemma [ legalllem ] that @xmath619 .
this is a contradiction .
[ lemdt ] a.s .
, for all @xmath843 such that @xmath844 and @xmath574 , we have @xmath793 .
the proof of this lemma is very similar to that of lemma [ lemst ] .
let @xmath843 be such that @xmath845 and @xmath574 . here again , we may suppose that @xmath513 for all @xmath800 , and we can find two subtrees @xmath846 $ ] and @xmath847 $ ] rooted at @xmath848}\cup{[[}\rho _ b , b { ] ] } \cup{\mathit{fl}}_\infty){\setminus}\{a , b\}$ ] satisfying @xmath849 . as before , we consider the discrete approximations @xmath600 , @xmath617 , @xmath850\!]})$ ] and @xmath851\!]})$ ] of @xmath24 , @xmath27 , @xmath745 and @xmath732 .
let @xmath616 be a geodesic path from @xmath600 to @xmath617 . we still define @xmath823\ ! ] } } ) \quad\mbox{and}\quad b_n { : = } { \dot}{{\mathfrak{t}}}_n(\overrightarrow{{[\![}t_n^1,s_n^2 { ] \ ! ] } } ) , \ ] ] and we see by the same arguments as in lemma [ lemst ] that , for @xmath3 sufficiently large , @xmath616 does not intersect @xmath830 .
we then conclude exactly as before .
theorem [ ip ] follows from lemmas [ lemfl ] , [ lemad ] , [ lemst ] and [ lemdt ] .
a straightforward consequence of theorem [ ip ] is that , if the equivalence class of @xmath603 for @xmath511 is not trivial , then @xmath421 is an increase point of @xmath852 . by lemma [ pc ] ,
the equivalence class of @xmath24 for @xmath493 is then trivial .
such points may be called _ leaves _ by analogy with tree terminology .
the goal of this section is to prove theorem [ cvq2 ] . to that end , we use the notion of regular convergence , introduced by whyburn @xcite .
recall that @xmath853 is the set of isometry classes of compact metric spaces , endowed with the gromov
hausdorff metric .
we say that a metric space @xmath10 is a _ path metric space _ if any two points @xmath854 may be joined by a path isometric to a real segment necessarily of length @xmath855 .
we call @xmath856 the set of isometry classes of path metric spaces . by @xcite ,
theorem 7.5.1 , @xmath856 is a closed subset of @xmath20 .
we say that a sequence @xmath857 of path metric spaces is _ @xmath92-regular _ if for every @xmath858 , there exists @xmath859 such that for @xmath3 large enough , every loop of diameter less than @xmath860 in @xmath861 is homotopic to @xmath57 in its @xmath662-neighborhood .
this definition is actually slightly stronger than whyburn s original definition @xcite .
see the discussion in the second section of @xcite for more details .
we also chose here not to restrict the notion of 1-regularity only to converging sequences of path metric spaces , as it was done in @xcite , because the notion of 1-regularity ( as stated here ) is not directly related to the convergence of the sequence of path metric spaces .
our main tool is the following theorem , which is a simple consequence of begle @xcite , theorem 7 .
[ begle ] let @xmath857 be a sequence of path metric spaces all homeomorphic to the @xmath0-torus @xmath9 .
suppose that @xmath861 converges toward @xmath15 for the gromov
hausdorff topology , and that the sequence @xmath857 is 1-regular
. then @xmath15 is homeomorphic to @xmath9 as well . in order to apply proposition [ begle ] ,
we construct a path metric space @xmath862 homeomorphic to @xmath9 , and an embedded graph that is a representative of the map @xmath2 , such that the restriction of @xmath862 to the embedded graph is isometric to @xmath863 .
we use the method provided by miermont in @xcite , section 3.1 .
we write @xmath864 the set of faces of @xmath2 .
let @xmath865 , @xmath866 be @xmath3 copies of the hollow bottomless unit cube @xmath867 ^ 3{\setminus}\bigl((0,1)^2 \times[0,1 ) \bigr)\ ] ] endowed with the intrinsic metric @xmath868 inherited from the euclidean metric .
( the distance between two points @xmath246 and @xmath628 is the euclidean length of a minimal path in @xmath869 linking @xmath246 to @xmath628 . )
let @xmath870 , and let @xmath871 , @xmath872 , @xmath873 and @xmath874 be the four half - edges incident to @xmath262 , ordered according to the counterclockwise order .
for @xmath875 , we define : @xmath876 in this way , we associate with every half - edge @xmath877 a path along one of the four edges of the square @xmath878 , where @xmath262 is the face located to the left of @xmath879.=-1 we then define the relation @xmath880 as the coarsest equivalence relation for which @xmath881 for all @xmath882 and @xmath883 $ ] .
this corresponds to gluing the spaces @xmath869 s along their boundaries according to the map structure of @xmath2 .
the topological quotient @xmath884 is a two - dimensional cw - complex satisfying the following .
its @xmath92-skeleton @xmath885 is an embedding of @xmath2 with faces @xmath886 . to the edge
@xmath887 corresponds the edge of @xmath888 made of the equivalence class of the points in @xmath889)$ ] .
its @xmath57-skeleton @xmath890 is in one - to - one correspondence with @xmath328 .
its vertices are the equivalence classes of the corners of the squares @xmath878 .
we endow the space @xmath891 with the largest pseudo - metric @xmath892 compatible with @xmath868 , @xmath893 and @xmath880 , in the sense that @xmath894 for @xmath895 and @xmath896 whenever @xmath897 .
its quotient still noted @xmath892then defines a pseudo - metric on @xmath888 ( which actually is a true metric , as we will see in proposition [ surf ] ) . as usual , we define @xmath898 its rescaled version .
we rely on the following proposition .
it was actually stated in @xcite for the two - dimensional sphere but readily extends to the @xmath0-torus .
[ surf ] the space @xmath862 is a path metric space homeomorphic to @xmath9 .
moreover , the restriction of @xmath888 to @xmath890 is isometric to @xmath899 , and any geodesic path in @xmath888 between points in @xmath890 is a concatenation of edges of @xmath888 .
finally , @xmath900 , so that , by proposition [ cvq ] , @xmath901 in the sense of the gromov
hausdorff topology .
we prove here that @xmath42 is a.s .
homeomorphic to @xmath9 by means of propositions [ begle ] and [ surf ] . to this end
, we only need to show that the sequence @xmath902 is 1-regular . at first
, we only consider simple loops made of edges .
we proceed in two steps : lemma [ eps0 ] shows that there are no noncontractible `` small '' loops ; then lemma [ 1reg ] states that the small loops are homotopic to @xmath57 in their @xmath662-neighborhood .
[ eps0 ] a.s .
, there exists @xmath903 such that for all @xmath639 large enough , any noncontractible simple loop made of edges in @xmath904 has diameter greater than @xmath905 .
the basic idea is that a noncontractible loop in @xmath888 has to intersect @xmath670 and to `` jump '' from a forest to another one . at the limit , the loop transits from a forest to another by visiting two points that @xmath512 identifies .
if the loops vanish at the limit , then these two identified points become identified with a point in @xmath495 , creating an increase point for both @xmath559 et @xmath491 .
we proceed to the rigorous proof .
we argue by contradiction and assume that , with positive probability , along some ( random ) subsequence of the sequence @xmath28 , there exist noncontractible simple loops @xmath616 made of edges in @xmath888 with diameter tending to @xmath57 ( with respect to the rescaled metric @xmath906 ) .
we reason on this event . because @xmath616 is noncontractible
, it has to intersect @xmath670 : if not , @xmath616 would entirely be drawn in the unique face of @xmath295 , which is homeomorphic to a disk , by definition of a map .
it would thus be contractible , by the jordan curve theorem .
let @xmath907 .
up to further extraction , we may suppose that @xmath908 .
notice that every time @xmath616 intersects @xmath670 , it has to be `` close '' to @xmath600 .
precisely , if @xmath909 tends to @xmath27 , then @xmath910 , which yields @xmath619 , and @xmath699 by lemma [ lemfl ] . moreover , for @xmath3 sufficiently large , the base point @xmath911 : otherwise , for infinitely many @xmath3 s , @xmath912 , so that @xmath559 would reach its minimum at @xmath24 , and we know by lemma [ min ] that this is not the case .
let us first suppose that @xmath24 is not a node of @xmath487 .
there exists @xmath702 such that @xmath913 and for @xmath3 large enough , @xmath914 . for @xmath3
even larger , the whole loop @xmath616 `` stays in @xmath915 . ''
precisely , for all @xmath916 , we have @xmath917 .
otherwise , since @xmath918 is finite , there would exist @xmath919 such that for infinitely many @xmath3 s , we can find @xmath920 .
up to extraction , @xmath921 , so that @xmath922 ( @xmath24 is not a node ) and @xmath923 , which is impossible , by lemma [ lemfl ] . at @xmath600 and
`` jumping '' from @xmath924 to @xmath925 at @xmath617 . ]
we claim that there exists an arc of @xmath616 linking a point @xmath926 to some point in @xmath927 that overflies either @xmath928}$ ] or @xmath929}$ ] ( see figure [ archmag ] ) .
let us suppose for a moment that this does not hold .
in particular , there is no arc linking a point in @xmath930 to a point in @xmath931
. it will be more convenient here to write @xmath616 as @xmath932 where the @xmath933 s are vertices , and the @xmath934 s are arcs . let @xmath935\!]}\dvtx v_j \in{\mathit{fl}}_n \}$ ] be the index of the first time @xmath616 returns to @xmath670 .
then @xmath936 belong to the same set @xmath937 or @xmath931 , and @xmath938 is thus drawn inside the face of @xmath295 . as a result , the path @xmath939 is homotopic to the segment @xmath940}$ ] . repeating the argument for every `` excursion '' away from @xmath670
, we see that @xmath616 is homotopic to a finite concatenation of segments all included in the topological segment @xmath941}$ ] , where we used the notation of section [ secfor ] for the forest @xmath924 ; see figure [ archmag ] .
it follows that @xmath616 is contractible , which is a contradiction .
we consider the case where the arc from the previous paragraph overflies @xmath942}$ ] .
the other case is treated in a similar way . from the construction of the chapuy marcus
schaeffer bijection , we can find integers @xmath943 such that @xmath944 , @xmath945 and for all @xmath946 , @xmath947 . up to further extraction , we may suppose that @xmath948 and @xmath949 .
therefore , for all @xmath950 , @xmath951 . moreover , the fact that @xmath952 yields @xmath953 , so that @xmath421 is an increase point for @xmath559 .
but @xmath954 and @xmath421 has to be an increase point for @xmath491 . by lemma [ pc ] , this can not happen .
if @xmath24 is a node , there are three half - edges @xmath955 , @xmath956 and @xmath957 such that @xmath958 . a reasoning similar to what precedes yields the existence of an arc of @xmath616 linking a point @xmath617 in one of the three sets @xmath959 , @xmath960 ( where we use the convention @xmath961 ) to a point lying in another one of these three sets that overflies either , if @xmath962 , @xmath963}\cup{[[}\mathfrak{e } _ { i+1 } { ] ] } $ ] or @xmath964}$ ] , or , if @xmath965 , @xmath966}$ ] or @xmath967}\cup { [ [ } a,\rho_{b_n } { ] ] } $ ] .
we conclude by similar arguments .
we now turn our attention to contractible loops .
let @xmath585 be a contractible simple loop in @xmath888 made of edges .
then @xmath585 splits @xmath888 into two domains .
only one of these is homeomorphic to a disk . , which is either the plane when @xmath968 , or the unit disk when @xmath969 ; see , for example , @xcite , theorem 1.7 .
] we call it the _ inner domain _ of @xmath585 , and we call the other one the _ outer domain _ of @xmath585 . in particular , these domains are well defined for loops whose diameter is smaller than @xmath905 , when @xmath3 is large enough .
[ 1reg ] a.s .
, for all @xmath498 , there exists @xmath970 such that for all @xmath639 sufficiently large , the inner domain of any simple loop made of edges in @xmath904 with diameter less than @xmath860 has diameter less than @xmath662 .
we adapt the method used by miermont in @xcite .
the idea is that a contractible loop separates a whole part of the map from the base point .
then the labels in one of the two domains it separates are larger than the labels on the loop . in the @xmath0-tree
, this corresponds to having a part with labels larger than the labels on the `` border . '' in the continuous limit , this creates an increase point for both @xmath491 and @xmath559 .
suppose that , with positive probability , there exists @xmath971 for which , along some ( random ) subsequence of the sequence @xmath28 , there exist contractible simple loops @xmath616 made of edges in @xmath888 with diameter tending to @xmath57 ( with respect to the rescaled metric @xmath906 ) and whose inner domains are of diameter larger than @xmath662 .
let us reason on this event .
first , notice that , because @xmath7 , the outer domain of @xmath616 contains at least one noncontractible loop , so that its diameter is larger than @xmath972 by lemma [ eps0 ]
. let @xmath561 be the unique point where @xmath559 reaches its minimum , and @xmath973 be an integer where @xmath565 reaches its minimum .
we call @xmath974 the corresponding point in the @xmath0-tree .
this is a vertex at @xmath975-distance @xmath92 from @xmath976 .
let us take @xmath977 in the domain that does not contain @xmath978 , such that the distance between @xmath977 and @xmath616 is maximal .
( if @xmath979 , we take @xmath977 in either of the two domains according to some convention . ) let @xmath980}\cup { \mathit{fl}}_n \cup{[[}\rho_{x_n},x_n { ] ] } ) $ ] be such that there exists an injective path[multiblock footnote omitted ] @xmath981 in @xmath339 from @xmath977 to @xmath634 that intersects @xmath761 only at @xmath634 .
in other words , when going from @xmath977 to @xmath982 along some injective path , @xmath634 is the first vertex belonging to @xmath616 we meet ; see figure [ xyzn ] .
such a point exists because @xmath977 and @xmath978 do not belong to the same of the two components delimited by @xmath616 .
up to further extraction , we suppose that @xmath983 , @xmath984 and @xmath985 .
we call @xmath986}\cup{\mathit{fl}}_\infty\cup{[[}\rho_{x},x { ] ] } $ ] the injective path corresponding to @xmath981 in the limit , that is , the path defined as @xmath981 `` without the subscripts @xmath3 . '' because the distance between two points in the same domain as @xmath977 is smaller than @xmath987 , we obtain that @xmath988 , as soon as @xmath989 .
in particular , we see that @xmath990 , and that the path @xmath991 is not reduced to a single point .
intersects @xmath606 .
this figure represents the case where @xmath992}$ ] . ]
let us first suppose that @xmath993 .
( in particular , @xmath994 for @xmath3 large , so that there is no ambiguity on which domain to chose @xmath977 . ) in that case , @xmath995}\cup{\mathit{fl}}_\infty\cup{[[}\rho _ { x},x { ] ] } ) { \setminus}\{x , w^\bullet\}$ ] , so that the points in @xmath996 are increase points of @xmath491 . by lemma [ pc ] , we can find a subtree[multiblock footnote omitted ] @xmath425 , not containing @xmath628 , satisfying @xmath997 and rooted on the path @xmath991 .
we consider a discrete approximation @xmath606 rooted on @xmath998 . because the loop @xmath616 is contractible
, all the labels of the points in the same domain as @xmath977 are larger than @xmath999 .
indeed , the labels represent the distances ( up to an additive constant ) in @xmath2 to the base point , and every geodesic path from such a point to the base point has to intersect @xmath616 .
for @xmath3 large enough , it holds that @xmath1000 . as a consequence
, @xmath606 can not entirely be included in the domain containing @xmath977 . therefore , the set @xmath1001 is not empty , so that we can find @xmath1002 .
up to extraction , we may suppose that @xmath1003 . on one hand , @xmath1004 ,
so that @xmath1005 . on the other hand , @xmath1006 and @xmath1007 ,
so that @xmath1008 .
because @xmath628 is not a leaf , this contradicts theorem [ ip ] .
when @xmath1009 , we use a different argument .
let @xmath1010 and @xmath1011 be , respectively , in the inner and outer domains of @xmath616 , such that their distance to @xmath616 is maximal .
because @xmath600 and @xmath617 do not belong to the same domain , we can find @xmath1012\ ! ] } } \quad\mbox{and}\quad t_n^2 \in \overrightarrow{{[\![}\beta_n,\alpha_n { ] \!]}}\ ] ] such that @xmath1013 , @xmath1014 . up to extraction , we suppose that @xmath1015 } \quad\mbox{and}\quad \frac{t_n^2}{2n } \to t^2 \in\overrightarrow{[\beta,\alpha]}.\ ] ] because @xmath1016 , we have @xmath1017 .
moreover , the argument we used to prove that @xmath1018 yields that @xmath1019 and @xmath1020 . as a result , we obtain that @xmath1021 .
this contradicts lemma [ min ] .
it remains to deal with general loops that are not necessarily made of edges .
we reason on the set of full probability where lemmas [ eps0 ] and [ 1reg ] hold .
we fix @xmath1022 .
let @xmath905 be as in lemma [ eps0 ] and @xmath1023 as in lemma [ 1reg ] . for @xmath639
sufficiently large , the conclusions of both lemmas hold , together with the inequality .
now , take any loop @xmath1024 drawn in @xmath904 with diameter less than @xmath1025 .
consider the union of the closed faces visited by @xmath1024 .
the boundary of this union consists in simple loops made of edges in @xmath904 .
let us call @xmath1026 the set of these simple loops . because every face of @xmath904 has a diameter smaller than @xmath1027
, we see that for all @xmath1028 , @xmath1029 .
then , by lemma [ eps0 ] , @xmath653 is contractible and , by lemma [ 1reg ] , its inner domain is of diameter less than @xmath662 . by definition , for all @xmath1028 , @xmath1024
entirely lies either inside the inner domain of @xmath653 , or inside its outer domain .
we claim that there exists one loop in @xmath1026 such that @xmath1024 lies in its inner domain .
then , it will be obvious that @xmath1024 is homotopic to @xmath57 in its @xmath662-neighborhood .
let us suppose that @xmath1024 lies in the outer domain of every loop @xmath1028 .
then , every face of @xmath904 is either visited by @xmath1024 , or included in the inner domain of some loop @xmath1030 . as a result
, we obtain that @xmath1031 .
this is in contradiction with our choice of @xmath1023 .
in order to prove lemmas [ min ] , [ pc ] , [ leme2 ] and [ leme4 ] , we rely on similar results for the brownian snake driven by a normalized excursion @xmath1032 .
this means that @xmath1033 has the law of a normalized brownian excursion , and , conditionally given @xmath1033 , the process @xmath1034 is a gaussian process with covariance @xmath1035}\mathbh{e}.\ ] ] we first focus on the proofs of lemmas [ min ] and [ pc ] .
lemmas 3.1 and 3.2 in @xcite state that , a.s .
, @xmath1034 reaches its minimum at a unique point , and that , a.s . , @xmath1036 and @xmath1037 are disjoint sets
. we will use a bijection due to chapuy @xcite to transfer these results to our case .
chapuy s bijection consists in `` opening '' @xmath0-trees into plane trees .
we briefly describe it here .
see @xcite for more details .
let @xmath58 be a @xmath0-tree whose scheme @xmath183 is dominant .
such a @xmath0-tree will be called _
dominant _ in the following . as usual
, we arrange the half - edges of @xmath183 according to its facial order : @xmath1038 .
let @xmath125 be one of the nodes of @xmath58 .
we can see it as a vertex of @xmath183 .
let us call @xmath1039 , @xmath1040 and @xmath1041 the three half - edges starting from @xmath125 ( i.e. , @xmath1042 ) , where @xmath1043 .
we say that @xmath125 is _ intertwined _ if the half - edges @xmath1039 , @xmath1040 , @xmath1041 are arranged according to the counterclockwise order around @xmath125 ( see figure [ slice ] )
. when @xmath125 is intertwined , we may _ slice _ it : we define a new map , denoted by @xmath1044 , by slicing the node @xmath125 into three new vertices @xmath1045 , @xmath1046 and @xmath1047 ( see figure [ slice ] ) . .
] the map obtained by such an operation turns out to be a dominant @xmath1048-tree . after repeating @xmath0 times this operation ,
we are left with a plane tree . in that regard ,
we call _ opening sequence _ of @xmath58 a @xmath0-uple @xmath1049 such that @xmath1050 is an intertwined node of @xmath58 , and for all @xmath1051 , the vertex @xmath933 is an intertwined node of @xmath1052
. we can show that every @xmath0-tree has exactly @xmath1053 intertwined nodes , and thus @xmath1054 opening sequences . to reverse the slicing operation
, we have to intertwine and glue back the three vertices together .
we then need to record which vertices are to be glued together .
this motivates the following definition : we call _ tree with @xmath0 triples _ a pair @xmath1055 , where : * @xmath1056 is a ( rooted ) plane tree ; * for @xmath1057 , @xmath1058 is a set of three vertices of @xmath1056 ; * the vertices @xmath1059 , @xmath1057 , @xmath1060 , are pairwise distinct ; * the vertices of the tree @xmath1061}\ ] ] have degree at most @xmath117 , and the @xmath1059 s have degree exactly @xmath92 in that tree .
( as in the case of @xmath0-trees , the set @xmath1062}$ ] represents the range of the unique path linking @xmath24 and @xmath27 in the tree . )
let @xmath58 be a @xmath0-tree together with an opening sequence @xmath1063 . for all @xmath1057 ,
let us call @xmath1064 the triple of vertices obtained from the slicing of @xmath933 , as well as @xmath1065 the plane tree .
we define @xmath1066 .
then @xmath1067 is a bijection from the set of all dominant @xmath0-tree equipped with an opening sequence into the set of all trees with @xmath0 triples .
now , when the @xmath0-tree is well - labeled , we can do the same slicing operation , and the three vertices we obtain all have the same label .
we call _ well - labeled tree with @xmath0 triples _ a tree with @xmath0 triples @xmath1068 carrying a labeling function @xmath1069 such that : * @xmath1070 , where @xmath879 is the root of @xmath1056 ; * for every pair of neighboring vertices @xmath1071 , we have @xmath1072 ; * for all @xmath1073 , we have @xmath1074 .
we call @xmath1075 the set of all well - labeled trees with @xmath0 triples having @xmath3 edges .
the bijection @xmath1067 then extends to a bijection between dominant well - labeled @xmath0-trees equipped with an opening sequence and well - labeled trees with @xmath0 triples .
the contour pair of an opened @xmath0-tree can be obtained from the contour pair of the @xmath0-tree itself ( and vice versa ) .
the labeling function is basically the same , but read in a different order .
the contour function is slightly harder to recover , because half of the forests are to be read with the floor directed `` upward '' instead of `` downward . '' because we will deal at the same time with @xmath0-trees and plane trees in this section , we will use a gothic font for objects related to @xmath0-trees , and a boldface font for objects related to plane trees . in the following ,
we use the notation of section [ decomp ] .
let @xmath75 be a well - labeled dominant @xmath0-tree with scheme @xmath183 and @xmath1076 be one of the @xmath1054 corresponding opened well - labeled trees .
the intertwined nodes of the @xmath0-tree correspond to intertwined nodes of its scheme , so that the opening sequence used to open @xmath75 into @xmath1076 naturally corresponds to an opening sequence of @xmath183 .
let @xmath1077 be the tree obtained by opening @xmath183 along this opening sequence .
we identify the half - edges of @xmath183 with the half - edges of @xmath1077 , and arrange them according to the facial order of @xmath1077 : @xmath1078 .
( beware that this is not the usual arrangement according to the facial order of @xmath183 . ) now , the plane tree @xmath1056 is obtained by replacing every half - edge @xmath1079 of @xmath1077 with the corresponding forest @xmath1080 of proposition [ decwl ] , as in section [ secdec ] .
we call @xmath1081 the contour pair of @xmath234 , we let @xmath1082 and we define @xmath1083 by ( [ lele ] ) . for any edge @xmath1084 with @xmath1085 , we will visit the forest @xmath1086 while `` going up '' and the forest @xmath1087 while `` coming down '' when we follow the contour of @xmath1088 .
precisely , we define @xmath1089 the first function is the concatenation of the contour functions of the trees in @xmath1086 with an extra `` up step '' between every consecutive trees .
the second one is the concatenation of the contour functions of the trees in @xmath1087 with an extra `` down step '' between every consecutive trees .
it is merely the contour function of @xmath1087 shifted in order to start at @xmath57 . what -tree .
the squares form one triple and the triangles the other one .
the ( blue ) short dashes correspond to the upward - directed floors and the ( green ) long dashes to the downward - directed floors .
the ( red ) solid line on the right of the root corresponds to the part of the tree containing the root that has to be visited at the end .
the forest @xmath1090 is also represented on this figure . ]
happens to the forests @xmath1091 and @xmath1092 is a little more intricate .
let us first call ( see figure [ opened ] ) @xmath1093\\[-8pt ] y & { : = } & \inf\{s \dvtx { { \mathfrak{c}}}^{\bar\mathfrak{e}_*}(s)= - \sigma ^{\mathfrak{e } _ * } - \underline { { \mathfrak{c}}}^{\mathfrak{e}_*}(u ) \}. \nonumber \ ] ] when visiting the forest @xmath1094 , the floor is directed downward up to time @xmath628 and then upward : @xmath1095 } { { \mathfrak{c}}}^{\bar\mathfrak{e } _ * } + { { \mathfrak{c}}}^{\bar\mathfrak{e}_*}(y ) \bigr)_{0 \le s \le m^{\bar\mathfrak{e } _ * } - y}.\ ] ] finally , the forest @xmath1091 is visited twice .
the first time ( when beginning the contour ) , it is visited between times @xmath124 and @xmath1096 , and the floor is directed upward : @xmath1097 } { { \mathfrak{c}}}^{\mathfrak{e } _ * } + { { \mathfrak{c}}}^{\mathfrak{e}_*}(u ) \bigr)_{0 \le s \le m^{\mathfrak{e } _ * } - u}.\ ] ] the second time ( when finishing the contour ) , we visit it between times @xmath57 and @xmath246 with the floor directed downward , then we visit a part of the tree containing the root between times @xmath246 and @xmath124 : @xmath1098 } { { \mathfrak{c}}}^{\mathfrak{e } _ * } + \underline{{{\mathfrak{c}}}}^{\mathfrak{e}_*}(u ) \bigr)_{0 \le s \le u - x}.\ ] ] the contour pair of @xmath1076 is then given by @xmath1099 where @xmath1100 as in section [ secsl ] , we let @xmath294 be uniformly distributed over the set @xmath80 of well - labeled @xmath0-trees with @xmath3 edges , and , applying skorokhod s representation theorem , we assume that the convergence of proposition [ cvint ] holds almost surely .
let @xmath1101 be a sequence of i.i.d .
random variables uniformly distributed over @xmath1102\!]}$ ] and independent of @xmath1103 . with any dominant scheme @xmath1104 and integer @xmath1105\!]}$ ] , we associate a deterministic opening sequence . when @xmath294 is dominant , we may then define @xmath1106 as the opened tree of @xmath294 according to the opening sequence determined by the integer @xmath1107 . in this case , we call @xmath1108 the contour pair of @xmath1106 .
when @xmath1109 is not dominant , we simply set @xmath1110 , where we write @xmath1111 \mapsto 0 $ ] .
we also let @xmath1112 be the rescaled versions of @xmath1113 and @xmath1114 .
we now work at fixed @xmath1115 for which proposition [ cvint ] holds , @xmath1116 , and such that for all @xmath1117\!]}$ ] , @xmath1118 .
note that the set of such @xmath1115 s is of full probability .
for @xmath3 large enough , @xmath1119 , so that @xmath1106 is well defined . for all @xmath3
such that @xmath324 and @xmath1120 , we always open the @xmath0-tree @xmath294 according to the same opening sequence , so that the ordering @xmath1121 of the half - edges of @xmath295 is always the same . as a result , we obtain that @xmath1122 where @xmath1123 is defined by ( [ op1])([op5 ] ) and ( [ contopen ] ) when replacing every occurrence of @xmath1124 by @xmath1125 and every occurrence of @xmath1083 by @xmath1126 . note that @xmath1127 has exactly @xmath1128 a priori distinct accumulation points , each corresponding to one of the ways of opening the real @xmath0-tree @xmath487 .
now , because every @xmath1126 goes from @xmath57 to @xmath57 , it is easy to see that for all @xmath1129 , the points where @xmath559 reaches its minimum are in one - to - one correspondence with the points where @xmath1130 reaches its minimum .
moreover , we can see that if @xmath491 and @xmath559 have a common increase point , then at least one of the pairs @xmath1123 will also have a common increase point .
indeed , let us suppose that @xmath491 and @xmath559 have a common increase point .
then , there exists @xmath1131 such that @xmath1132 and @xmath1126 have a common increase point @xmath1133 $ ] .
we use the following lemma : [ iptransfert ] let @xmath1134 \to{\mathbb{r}}$ ] be a function . *
if @xmath1135 is an increase point of @xmath262 , then @xmath421 is an increase point of @xmath1136 as well .
* if @xmath1137 $ ] is an increase point of @xmath262 , then @xmath421 is an increase point of @xmath1138}{f}$ ] .
we postpone the proof of this lemma and finish our argument .
if @xmath1139 , then @xmath421 is a common increase point of @xmath1140 and @xmath1141 thanks to lemma [ iptransfert ] . when @xmath1142 , this fact remains true if we define @xmath1143 .
note that @xmath246 is an increase point of @xmath1140 , even if @xmath57 is not an increase point of the second function defining @xmath1144 in ( [ op5 ] ) . in this case ,
for all @xmath1129 , @xmath1145 and @xmath1130 have a common increase point .
let us now suppose that @xmath1146 , and let us fix @xmath1105\!]}$ ] .
we consider the opening corresponding to @xmath1129 .
if @xmath1147 is visited while coming down in the contour of the opened tree , then we conclude as above .
if both @xmath1148 and @xmath1149 are visited while going up , then @xmath57 will be an increase point of @xmath1150 , so that @xmath1145 and @xmath1130 will still have a common increase point . in the remaining case
where @xmath1148 is visited while going up and @xmath1149 is visited while coming down ( i.e. , @xmath1151 ) , we can not conclude that @xmath1145 and @xmath1130 have a common increase point .
this , however , only happens when the node @xmath52 belongs to the opening sequence .
but when we pick an opening sequence , we can always choose not to pick a given node , because at each stage of the process , we have at least @xmath1152 intertwined nodes .
this implies that at least one of the opening sequences will not contain @xmath52 , and the corresponding pair @xmath1153 will have a common increase point .
proof of lemma [ iptransfert ] let @xmath1135 be an increase point of @xmath262 .
if @xmath421 is a right - increase point of @xmath262 , then @xmath580 when @xmath950 for some @xmath1154 . for such @xmath610 s , @xmath1155 , so that @xmath1156 , and @xmath421 is a right - increase point of @xmath1157 .
if @xmath421 is a left - increase point of @xmath262 , then @xmath580 when @xmath1158 for some .
if @xmath1159 , then , using the fact that @xmath1160 } f$ ] , we obtain that @xmath1155 when @xmath1161 and conclude as above that @xmath421 is a left - increase point of @xmath1162 . finally ,
if @xmath1163 , then for all @xmath1164 , we have @xmath1165 , and because @xmath1166 , we conclude that @xmath421 is a right - increase point of @xmath1157 .
we obtain the second assertion of the lemma by applying the first one to @xmath1167 and the function @xmath1168 .
conditionally on the event @xmath1169 , the distribution of @xmath1108 is that of the contour pair of a uniform well - labeled tree with @xmath0 triples .
we use this fact to see that the law of @xmath1127 converges weakly toward a law absolutely continuous with respect to the law of @xmath1170 .
let @xmath1171 be uniformly distributed over the set @xmath1172 of all well - labeled plane trees with @xmath3 edges .
we call @xmath1173 the contour pair of @xmath1174 and define as usual the rescaled versions of both functions , @xmath1175 for all @xmath1 , @xmath1176 and @xmath1177 , we define @xmath1178 respectively , the profile and rescaled profile of @xmath1174 .
we let @xmath1179 be the one - dimensional ise ( random ) measure defined by @xmath1180 for every nonnegative measurable function @xmath1181 .
by @xcite , theorem 2.1 , it is known that @xmath1179 a.s . has a continuous density @xmath1182 with compact support . in other words ,
@xmath1183 for every nonnegative measurable function @xmath1181 .
[ jd ] the triple @xmath1184 converges weakly toward the triple @xmath1185 in the space @xmath1186,{\mathbb{r}})^2\times\mathcal{c}_c({\mathbb{r}})$ ] endowed with the product topology .
it is known that the pair @xmath1187 converges weakly to @xmath1188 : in @xcite , theorem 5 , chassaing and schaeffer proved this fact with @xmath1189 instead of @xmath1190 in the definition ( [ scuncp ] ) .
the claim as stated here easily follows by using the uniform continuity of @xmath1188 . using @xcite ,
theorem 3.6 , and the fact that @xmath1182 is a.s .
uniformly continuous @xcite , theorem 2.1 , we also obtain that the sequence @xmath1191 converges weakly to @xmath1182 . as a result ,
the sequences of the laws of the processes @xmath1192 , @xmath1193 and @xmath1191 are tight .
the sequence @xmath1194 of the laws of @xmath1184 is then tight as well , and , by prokhorov s lemma , the set @xmath1195 is relatively compact .
let @xmath1196 be an accumulation point of the sequence @xmath1194 .
there exists a subsequence along which @xmath1197 converges weakly toward a random variable @xmath1198 with law @xmath1196 .
thanks to skorokhod s theorem , we may and will assume that this convergence holds almost surely along this subsequence .
we know that @xmath1199 it remains to see that @xmath1200 is the density of the occupation measure of @xmath1201 , that is , @xmath1202 for all @xmath1181 continuous with compact support .
first , notice that @xmath1203 & = & \int_{{\mathbb{r } } } dx \,x_{(n ) } ( x ) h ( { \gamma^{-1 } n^{-1/4}}{\lfloor}{\gamma n^{1/4}}x { \rfloor})\\[-2pt ] & \to&\int_{\mathbb{r}}dx \,f'(x ) h(x)\end{aligned}\ ] ] by dominated convergence , a.s . as @xmath43 along the subsequence we consider .
it is convenient to introduce now the notation @xmath1204 defined as follows : for @xmath1205 , we set @xmath1206 then , if we denote by @xmath1207 the @xmath72th vertex of the facial sequence of @xmath606 , and by @xmath620 the root of @xmath606 , we obtain that the time the process @xmath1208 spends at each vertex @xmath1209 is exactly @xmath1152 .
so we have @xmath1210 & & \qquad= \frac1 n \sum_{v \in\tau_n{\setminus}\{\rho_n \ } } h ( { \gamma^{-1 } n^{-1/4}}\lambda_n(v ) ) + \frac1 n h(0)\\[-2pt ] & & \qquad= \frac1 { 2n } \int_0^{2n } ds \ , h ( { \gamma^{-1 } n^{-1/4}}\lambda_n ( { \langle\!\langle}s { \rangle\!\rangle}_n ) ) + \frac1 n h(0)\\[-2pt ] & & \qquad= \int_0 ^ 1 ds \ , h ( { \gamma^{-1 } n^{-1/4}}\lambda_n({\langle\!\langle}2ns { \rangle\!\rangle}_n ) ) + \frac 1 n h(0)\\[-2pt ] & & \qquad\to\int_0 ^ 1 dt \ , h(z_t')\end{aligned}\ ] ] a.s . along the subsequence considered .
we used the fact that @xmath1211 which is obtained by using the uniform continuity of @xmath1201 .
this proves that @xmath1198 has the same law as @xmath1212 .
thus the only accumulation point @xmath1196 of the sequence @xmath1194 is the the law of the process @xmath1185 . by relative compactness of the set @xmath1195
, we obtain the weak convergence of the sequence @xmath1194 toward @xmath1196 .
we define @xmath1213}.\ ] ] this quantity is well defined @xcite , lemma 10 .
we also define the law of the pair @xmath1214 by the following formula : for every bounded borel function @xmath287 on @xmath1186,{\mathbb{r}})^2 $ ] , @xmath1215 = \mathbb{e } [ w \varphi ( \mathbh{e},z ) ] .\ ] ] [ cs2 ] the pair @xmath1216 converges weakly toward the pair @xmath1214 in the space @xmath1217,{\mathbb{r}})^2 , \| \cdot\|_\infty)$ ] of pair of continuous real - valued functions on @xmath1218 $ ] endowed with the uniform topology .
let @xmath262 be a bounded continuous function on @xmath1186,{\mathbb{r}})^2 $ ] .
we have @xmath1219 & = & { \mathbb{p}}(d_n ) \mathop{\sum_{(\tau,\lambda ) \in\mathcal{t}^0_n}}_{(\tau , \lambda ) \leftrightarrow({\mathbf{c}},{\mathbf{l } } ) } f({\mathbf{c}},{\mathbf{l } } ) { \mathbb{p}}\bigl((\tau _ n,\lambda_n ) = ( \tau , \lambda ) | d_n \bigr)\\ & & { } + { \mathbb{p}}(\overline{d}_n ) f ( { \mathbf{0}}_{2n},{\mathbf{0}}_{2n } ) , \end{aligned}\ ] ] where we used the notation @xmath1220 to mean that the well - labeled tree @xmath1221 is coded by the contour pair @xmath1222 .
it was shown in @xcite , lemma 8 , that the number of well - labeled trees with @xmath0 triples having @xmath3 edges is equivalent to the number of well - labeled plane trees having @xmath3 edges , together with @xmath0 triples of vertices ( not necessarily distinct and not arranged ) such that all the vertices of the same triple have the same label .
more precisely , we have @xmath1223 and , because @xmath262 is bounded and @xmath1224 , we obtain that @xmath1225 \sim\frac{|\mathcal { t}^0_n| } { |{\mathcal{w}}_n| } \mathbb{e}\biggl [ \biggl(\sum_{k \in{\mathbb{z } } } x_n(k)^3 \biggr)^g f \bigl(\gamma_{(n ) } , \lambda _ { ( n ) } \bigr ) \biggr].\ ] ] using the asymptotic formulas @xmath1226 , as well as @xmath1227 for some positive constant @xmath1228 only depending on @xmath0 ( @xcite , lemma 8) , as well as the computation @xmath1229 we see that there exists a positive constant @xmath429 such that @xmath1225 \sim c \mathbb{e}\biggl [ \biggl(\int _ { \mathbb{r}}dx \ , x_{(n)}(x)^3\biggr)^g f \bigl(\gamma_{(n ) } , \lambda_{(n ) } \bigr ) \biggr].\vadjust{\goodbreak}\ ] ] now , let @xmath858 .
thanks to @xcite , lemma 10 , we see that both quantities @xmath1230 $ ] and @xmath1231 $ ] are finite .
then , using the fact that @xmath1232 \le\frac1 l \mathbb{e}\biggl [ \biggl(\int x_{(n)}^3 \biggr)^{g+1 } \biggr],\ ] ] we obtain that , for @xmath1233 sufficiently large , @xmath1234 < { \varepsilon}\ ] ] and @xmath1235 < { \varepsilon}.\ ] ] thanks to the proposition [ jd ] , for @xmath3 sufficiently large , @xmath1236\\ & & \hspace*{48.7pt } { } - \mathbb { e}\biggl [ \biggl(\int { f_{\mathrm{ise}}}^3 \biggr)^g f ( \mathbh{e},z ) \mathbh{1}_{\{\int{f_{\mathrm{ise}}}^3
\le l\ } } \biggr ] \biggr| < { \varepsilon}.\end{aligned}\ ] ] this yields the existence of a constant @xmath1237 such that @xmath1238 { { \mathop{\longrightarrow}}_{n\to\infty}}c \mathbb{e}\biggl [ \biggl(\int { f_{\mathrm{ise}}}^3 \biggr)^g f ( \mathbh{e},z ) \biggr],\ ] ] and we compute the value of @xmath1237 by taking @xmath1239 .
thanks to ( [ defw ] ) , we see that the properties that hold almost surely for the pair @xmath1032 also hold almost surely for @xmath1240 .
we may now conclude thanks to @xcite , lemma 3.1 , that @xmath1241 and , by @xcite , lemma 3.2 , @xmath1242 this concludes the proof of lemmas [ min ] and [ pc ] .
chapuy s bijection may naturally be transposed in the continuous setting .
let @xmath1105\!]}$ ] be an integer corresponding to an opening sequence , and @xmath1243 the real tree coded by @xmath1145 .
the interval @xmath325 $ ] may be split into @xmath1244 intervals coding the two halves of @xmath1245 and the other forests of @xmath487 . through the continuous analog of chapuy s bijection
, these intervals are reordered into an order corresponding to the opening sequence .
we call @xmath1246 \to[0,1]$ ] the bijection accounting for this reordering .
it is a cadlag function with derivative @xmath92 satisfying @xmath1247 for all @xmath508 $ ] . in order to see that lemma [ leme2 ] is a consequence of @xcite , lemma 2.4 ,
let us first see what happens to subtrees of @xmath487 through the continuous analog of chapuy s bijection .
it is natural to call root of @xmath487 the point @xmath1248 , where the real number @xmath1249 was defined in proposition [ cvint ] as the limit of the integer coding the root in @xmath339 , properly rescaled .
using classical properties of the brownian motion together with proposition [ cvint ] , it is easy to see that , almost surely , @xmath1250 is a leaf of @xmath487 , so that @xmath1251 is well defined .
any subtree of @xmath487 not included in @xmath1251 ( these subtrees require extra care , we will treat them separately ) is transformed through chapuy s bijection into some subtree of the opened tree @xmath1252
( i.e. , into some tree to the left or right of some branch of @xmath1252 ) .
this is easy to see when the subtree is not rooted at a node of @xmath487 , and we saw at the end of section [ secdefgtree ] that , almost surely , all the subtrees are rooted outside the set of nodes of @xmath487 .
we reason by contradiction to rule out these subtrees .
we call @xmath1024 the lebesgue measure on @xmath325 $ ] .
let us suppose that there exist @xmath1253 , and some subtree @xmath425 , coded by @xmath611 $ ] , not included in @xmath1254 , such that @xmath1255 } { { \mathfrak{l}}}_\infty<{{\mathfrak{l}}}_\infty ( l)-\eta $ ] , and @xmath1256 \dvtx { { \mathfrak{l}}}_\infty(s ) < { { \mathfrak{l}}}_\infty(l ) - \eta+ { \varepsilon } ; \nonumber\\ & & \hspace*{72pt } \forall x \in [ { { \mathfrak{c}}}_\infty(l ) , { { \mathfrak{c}}}_\infty(s)],\\ & & \hspace*{72pt}{{\mathfrak{l}}}_\infty\bigl(\sup\{t\le s \dvtx { { \mathfrak{c}}}_\infty(t)=x\}\bigr ) > { { \mathfrak{l}}}_\infty ( l)-\eta + \frac{\varepsilon}8 \biggr\ } \biggr ) = 0 .
\nonumber\end{aligned}\ ] ] note that , by definition of @xmath1145 , the function @xmath1257 is constant on @xmath611 $ ] . let us call @xmath1258 and @xmath1259 .
it is easy to see that ( [ liminf ] ) remains true when replacing , respectively , @xmath249 , @xmath610 , @xmath491 and @xmath559 with @xmath461 , @xmath462 , @xmath1260 and @xmath1261 .
thanks to proposition [ cs2 ] , the conclusion of @xcite , lemma 2.4 , is also true for the opened tree @xmath1252 , and the fact that @xmath1262 $ ] codes a subtree of the opened tree yields a contradiction .
we then use a re - rooting argument to conclude . with positive probability ,
@xmath1251 is no longer the tree containing the root in the uniformly re - rooted @xmath0-tree .
let us suppose that , with positive probability , there exists a subtree of @xmath487 included in @xmath1251 , satisfying the hypotheses but not the conclusion of lemma [ leme2 ] . then , with positive probability
, there will exist a subtree not included in the tree containing the root of the uniformly re - rooted @xmath0-tree , satisfying the hypotheses but not the conclusion of lemma [ leme2 ] .
the fact that the uniformly re - rooted @xmath0-tree has the same law as @xmath487 yields a contradiction . using the same arguments as in @xcite
, we can see that lemma [ leme4 ] is a consequence of the following lemma ( see @xcite , corollary 6.2 ) : for every @xmath659 and every @xmath658 $ ] , there exists a constant @xmath1263 such that , for every @xmath858 , @xmath1264 \le c_{p,\delta } { \varepsilon}^{4p-\delta}.\ ] ] this readily comes from @xcite , lemma 6.1 , stating that for every @xmath659 and every @xmath658 $ ] , there exists a constant @xmath1265 such that , for every @xmath858 , @xmath1266 \le c'_{p,\delta } { \varepsilon}^{4p-\delta}.\ ] ] obviously , this still holds for @xmath1267 $ ] . using the link between @xmath559 and @xmath1268 , as well as proposition [ cs2 ] , we see that , for @xmath659 and @xmath658 $ ] , @xmath1269 & = & \mathbb{e}\biggl [ \biggl(\int_0 ^ 1 \mathbh{1}_{\ { { \mathbf{l}}_\infty(s ) \le\min{\mathbf{l}}_\infty+ { \varepsilon}\}}\ , ds\biggr)^p \biggr]\\ & = & \mathbb{e}\biggl [ w \biggl(\int_0 ^ 1 \mathbh{1}_{\{z_s \le\min z + { \varepsilon}\}}\ , ds\biggr)^p \biggr]\\ & \le&(\mathbb{e } [ w^2 ] c'_{2p,2\delta})^{1/2 } { \varepsilon}^{4p-\delta } \\ & = & c_{p,\delta } { \varepsilon}^{4p-\delta},\end{aligned}\ ] ] where @xmath1270 c'_{2p,2\delta})^ { 1/2 } < \infty$ ] , by @xcite , lemma 10 .
the author is sincerely grateful to grgory miermont for the precious advice and support he provided during the accomplishment of this work . | we discuss scaling limits of large bipartite quadrangulations of positive genus . for a given @xmath0 , we consider , for every @xmath1 , a random quadrangulation @xmath2 uniformly distributed over the set of all rooted bipartite quadrangulations of genus @xmath0 with @xmath3 faces . we view it as a metric space by endowing its set of vertices with the graph metric .
as @xmath3 tends to infinity , this metric space , with distances rescaled by the factor @xmath4 , converges in distribution , at least along some subsequence , toward a limiting random metric space .
this convergence holds in the sense of the gromov
hausdorff topology on compact metric spaces .
we show that , regardless of the choice of the subsequence , the limiting space is almost surely homeomorphic to the genus @xmath0-torus . . |
this computation in this work has been done using the facilities of the supercomputer center , institute for solid state physics , university of tokyo .
this work has also been supported partially by a grant - in - aid for scientific researach ( b ) from the ministry of education , science , sports and culture of japan .
g. binning , h. rohrer , ch .
gerber , and e. weibel , phys .
lett . * 49 * , 57 ( 1982 ) .
g. binning , c. f. quate , and ch .
gerber , phys .
lett . * 56 * , 930 ( 1986 ) .
r. wiesendanger , h. -j .
gntherodt , g. gntherodt , r. j. gambino , and r. ruf , phys .
* 65 * , 247 ( 1990 ) .
s. f. alvarado and p. renaud , phys .
68 , 1387 ( 1992 ) .
k. sueoka , k. mukasa , and k. hayakawa , jpn .
j. appl . phys . *
32 * , 2989 ( 1993 ) .
y. martin and h. k. wickramasinghe , appl .
50 * , 1455 ( 1987 ) ; j. j. saenz , n. garcia , p. grtter , e. meyer , h. heinzelmann , r. wiesendanger , l. rosenthaler , h. r. hidber , and h. -j .
gntherodt , j. appl . phys .
* 62 * , 4293 ( 1987 ) ; p. grtter , e. meyer h. heinzelmann , l. rosenthaler , h. -r .
hidber , and h. -j .
gntherodt , j. vac .
& technol . a * 6 * , 279 ( 1988 ) . k. mukasa , h. hasegawa , y. tazuke , k. sueoka , m. sasaki , and k. hayakawa , jpn . j. appl
. phys . * 33 * , 2692 ( 1994 ) .
k. mukasa , k. sueoka , h. hasegawa , y. tazuke , and k. hayakawa , material science and engineering b * 31 * , 69 ( 1995 ) . h. ness and f. gautier , phys . rev .
b * 52 * , 7352 ( 1995 ) . o. k. andersen , phys . rev .
b * 12 * , 3060 ( 1975 ) .
l. hedin and b. i. lundquist , j. phys .
c * 4 * , 2064 ( 1971 ) ; u. von barth and l. hedin , _
ibid_. * 5 * , 1629 ( 1972 ) . j. m. soler and a. r. williams ,
b * 40 * , 1560 ( 1989 ) . c. s. wang and a. j. freeman , phys .
b * 24 * , 4364 ( 1981 ) .
s. ohnishi and a. j. freeman , and m. weinert , phys .
b * 28 * , 6741(1983 ) . m. a. ruderman and c. kittel , phys . rev . *
96 * , 99 ( 1954 ) ; t. kasuya , prog .
phys . , * 16 * , 45 ( 1956 ) ; k. yosida , phys
. rev . * 106 * , 893 ( 1957 ) .
y. martin , c. c. williams , and h. k. wickramasinghe , j. appl . phys . * 61 * , 4723 ( 1987 ) . e. meyer and h. heinzelmann , in _ scanning tunneling microscopy _ ii , springer series in surface sciences 28 , ed . by r. wiesendanger and h. -j
gntherodt ( springer - verlag , berlin , 1992 ) p.99 .
jarvis , h. yamada , s. -i .
yamamoto , h. tokumoto , and j. b. pethica , nature * 384 * , 247 ( 1996 ) . | first - principles calculations have been performed for the exchange interaction and exchange force between two magnetic fe(100 ) thin films .
it is shown that the magnitude of the calculated exchange force is of the order of @xmath0n at @xmath1 and of the order of @xmath2n at @xmath3 , where @xmath4 denotes the distance between surfaces of the two thin films and @xmath5 the lattice constant of bulk fe .
the obtained forces are sufficiently larger than the sensitivity of the current atomic force microscopy of @xmath6n , which suggests the feasibility of the _ exchange force microscopy_. * i. introduction * surface magnetism has attracted much attention both in fundamental and applied researches .
experimental techniques to investigate the magnetic structure of the surface having been developed , mainly detect the spin polarization of electrons emitted from surface and the average magnetization on sample surface within a sub - micron range is usually observed .
none of the techniques has so far probed the local magnetic structure in an atomic scale . the scanning tunneling microscopy ( stm)@xcite and the atomic force microscopy ( afm)@xcite provide us with a way to investigate the surface structure with atomic resolution . the stm probes the surface of conducting materials by measuring tunneling current between the tip and sample , whereas the afm senses the atomic force between them .
the afm has a great advantage that it can be applied not only to conducting materials but also to insulators .
the extension of the stm to measurements of the magnetic structure , the so - called spin - polarized stm ( sp - stm ) , is currently being developed by various groups with different approaches.@xcite the magnetic force microscopy ( mfm)@xcite detects the force arising from the long - range interaction between magnetic dipoles of the tip and sample .
the typical tip - sample separations in the mfm are of the order of more than 10 nm and the spatial resolution is of the order of 10 nm to 100 nm .
it is fairly certain that improvement of the resolution might be made by probing the short - range exchange force rather than the long - range magnetic dipole force.@xcite in previous papers,@xcite we pointed out the possibility of the _ exchange force microscopy _ ( efm ) which probes the short - range exchange force between the tip and sample . since microscopic understanding of the exchange interaction between the tip and sample is of great importance and crucial in designing the efm , we previously studied the exchange interaction between ferromagnetic tip and ferromagnetic sample based on a one - dimensional electron gas model.@xcite by using a tight - binding model , ness and gautier@xcite calculated the exchange interaction between a ferromagnetic fe tip and magnetic cr(001 ) and ni(001 ) surfaces . since these calculations inevitably employ semi - phenomenological parameters , it is highly desirable to make @xmath7 evaluation of the exchange interaction and the exchange force for more realistic systems on the basis of first - principles electronic structure calculations , which is the purpose of the present paper .
the paper is organized as follows : in sec.ii we discuss the adopted model and calculation method .
the calculated results are presented in sec.iii .
the final sec.iv is devoted to conclusion and discussion . *
ii . method of calculation
* by modeling the tip and samples to be used in actual experiments , we adopt the two three - atomic - layer fe(100 ) films which are separated by the distance @xmath4 , as shown in fig .
[ fig1 ] .
surface atoms of the one film are assumed to be facing the hollow sites on the surface of the other film .
the translational symmetry is preserved in the surface - normal ( @xmath8 ) direction , in which a set of the two three - layer films are periodically located with the five - atomic - layer vacuum gap ( repeated slab model ) . for a calculation of the electronic structure of the adopted films ,
we employ the local - spin - density approximation to the density - functional theory , and the linear augmented - plane - wave ( lapw ) method@xcite with hedin - lundquist exchange correlation.@xcite the potential and charge density are expanded with plane waves in the whole space and with spherical waves inside muffin - tin spheres , and the force is calculated by following the prescription given by soler and williams.@xcite because bulk fe is in the ferromagnetic ground state , magnetic moments in each three - layer film is taken to be in the ferromagnetic alignment . as far as the relative orientation of moments in the two films
are concerned , however , we consider the parallel ( p ) and anti - parallel ( ap ) configurations in order to calculate the exchange interaction and the exchange force between the two magnetic films .
our lapw calculations are carried out by changing the film - film separation @xmath4 from 1.4to 5.0with an assumption that the internal atomic coordinates of each film are rigidly fixed .
* iii
. calculated results * _ a. magnetic moments _
we discuss first the magnetic moments in the adopted fe thin films .
figure [ fig2 ] shows the spin magnetic moments calculated within the muffin - tin spheres of fe atoms in the three layers of the upper film , which are referred to as the @xmath9 , @xmath10 and @xmath11 layers ( fig .
[ fig1 ] ) . the film - film separation @xmath4 in the abscissa is denoted relative to the lattice constant of bulk fe ( @xmath5 = 2.83 ) .
we should note that the separation of @xmath12 corresponds just to the interlayer distance of bulk fe .
due to the symmetry in our model , the magnetic moments in the @xmath13 layer of the lower film are the same as those in the @xmath14 layer of the upper film in the p configuration , while in the ap configuration those in the @xmath13 layer has the same magnitude but with the opposite sign as those in the @xmath14 layer . for both p and ap configurations , the magnetic moments at the outer @xmath11 layer
is enhanced to be about @xmath15 , which is almost identical to the surface magnetic moment obtained by slab calculations.@xcite the magnetic moments at the central @xmath10 layer is about @xmath16 , which is close to the experimental bulk value and to calculated one at the central layer of the slab.@xcite the magnetic moments of the @xmath10 and @xmath11 layers for both p and ap configurations are almost independent of the film - film separation . on the contrary , magnetic moments on the @xmath9 layer change considerably at @xmath17 , where the magnetic moments reduce significantly from the surface value to the bulk one as decreasing @xmath4 .
nevertheless , the change in the magnitude of moments of the @xmath9 layer becomes insignificant for @xmath18 .
we notice that the magnetic moments for the ap configuration are always smaller than that for the p configuration . _
b. exchange interaction _ the total energies , @xmath19 and @xmath20 , for the p and ap configurations
are shown in fig .
[ fig3 ] as a function of @xmath21 . when the films are moved from the bulk position at @xmath22 , the total energies for both @xmath19 and @xmath20 increase . the exchange interaction energy , @xmath23 , defined by @xmath24 is also plotted in fig .
[ fig3 ] . in the all region ( @xmath25 ) investigated , we get the positive @xmath23 , which means that the p configuration is more favorable than the ap one .
this is expected from the fact that the resulting magnetic configuration corresponds to a natural stacking of ferromagnetic fe in bulk .
the exchange interaction energy has an rkky - type oscillation@xcite with a period of about @xmath26 ( @xmath27 2 ) .
the peak values of @xmath28 are @xmath29ev at @xmath30 and @xmath31ev at @xmath32 .
note that the exchange interaction can be realized even at relatively far separation at @xmath33 ( [email protected] ) . _ c. exchange force
_ figure [ fig4 ] shows the atomic forces acting on atoms at the @xmath9 , @xmath10 and @xmath11 layers of the upper film .
the force direction is perpendicular to the film surface due to symmetry . as in the case of magnetic moments ( fig .
[ fig2 ] ) , the atomic force for the @xmath9 layer has a significant @xmath4 dependence whereas forces for the @xmath10 and @xmath11 layers are almost independent of @xmath4 .
the total forces summed up these forces : @xmath35 , for the p and af configurations , are shown in fig .
[ fig5 ] .
the exchange force defined by @xmath36 has an oscillation against @xmath21 with the period of about @xmath37 , which is identical with the period of the exchange interaction ( fig .
[ fig3 ] ) .
the magnitudes of the first and second peaks are @xmath38 n and @xmath39 n , respectively . in the case of @xmath1 , where the relevant @xmath4 orbitals must have a large overlap , the exchange force may arise from the direct exchange couplings between @xmath4 states of the two films . in the case of @xmath18 ,
on the other hand , the exchange coupling is expected to be mediated through delocalized @xmath40 and @xmath41 electrons . *
iv . conclusion and discussion * in summary , we have performed the first - principles calculation of the exchange interaction and the exchange force between two magnetic fe(100 ) thin films , to show that the magnitude of the calculated exchange force is of the order of @xmath42 at @xmath1 while it is of the order of @xmath43n at @xmath3 .
it has been shown that at @xmath44 , the perturbation of the approaching ferromagnetic film to magnetic moments in the other film , is negligibly small .
the calculated magnitude of the exchange force is in fairly good agreement with our previous result of @xmath45 in spite of a crudeness of the one - dimensional electron - gas model.@xcite@xcite our result is also comparable with the exchange force of @xmath46 between a fe tip and cr(100 ) surface calculated by ness and gautier@xcite with the use of the tight - binding model .
because the sensitivity of the current afm is of the order of @xmath47 to @xmath48n@xcite our calculation suggests the possibility of the efm , although it would be necessary to take into account an actual shape of the apex tip before we deduce definite conclusions .
ness and gautier have been pointed out@xcite that magnitudes of the exchange energy and force do not strongly depend on the tip morphology , although tip s shape varies the tip - sample distance where exchange interaction and force have the maximum values .
it has been also shown@xcite from calculations adopting sharp and blunt tips that an atomically sharp tip is not necessary for an actual observation with a constant - force mode , although a sharp tip is indispensable for a high , lateral spatial resolution .
these results@xcite justify , to some extent , our calculation employing a thin film as a tip ; a first - principles calculation adopting more realistic tips is under consideration . in actual measurements of the surface magnetic structure by using the efm
, we would have the two possibilities : a constant - force mode and a constant - distance mode , just as in the stm.@xcite in the constant - force mode , we detect the variation of the tip - sample distance due to the force difference between the parallel and anti - parallel configurations . such a experiment@xcite has been made to detect the antiferromagnetic structure from one terrace to the next on cr(100 ) surface , by using the constant - current mode in the sp - stm . on the contrary , in the constant - distance mode , we directly detect a change of the exchange forces between the parallel and anti - parallel configurations at a given tip - sample separation .
if such a measurement is performed by changing the tip - sample distance , we would face technical difficulties such as the snap - in effect : when the tip - sample distance comes in the point of force instability where the gradient of the force exceeds the magnitude of the lever stiffness , the tip and surface would snap into contact.@xcite in order to measure the rkky - type exchange force , we require an exploration of the way to avoid the snap - in effect . recently
, a continuous measurement of the force curve between tip and sample , as approaching towards contact without the snap - in effects , can be provided by means of force - controlled feedback system.@xcite this technique may possibly provide the basis of the efm measurement . |
in the standard framework of three massive neutrinos , the parameters which describe neutrino oscillation phenomena are three mixing angles @xmath0 , @xmath1 , @xmath2 , a cp phase @xmath3 , and two mass squared differences @xmath4 , @xmath5 .
the parameters ( @xmath6 , @xmath1 ) and ( @xmath7 , @xmath0 ) were determined by the experiments of atmospheric and accelerator neutrinos , and by those of solar and long baseline reactor neutrinos , respectively@xcite . on the other hand ,
the value of @xmath2 was determined recently by the two accelerator neutrino experiments , t2k@xcite and minos@xcite and by the three reactor neutrino experiments , double - chooz@xcite , daya bay@xcite and reno@xcite .
the results of the global analysis are given in refs.@xcite .
the quantities which are not yet determined are the pattern of mass hierarchy ( in other words the sign of @xmath8 ) , the cp phase @xmath3 and the octant of @xmath1 ( in other words the sign of @xmath9 ) . from @xmath10 .
however , we need more statistics before we draw any conclusion . ]
the next things to do is to determine the pattern of mass hierarchy and the octant of @xmath1 , before we achieve the final goal of the neutrino oscillation study , i.e. , measurement of the cp phase @xmath3 .
these tasks are expected to be done in the future experiments .
these experiments include so - called super beam experiments ( such as no@xmath11a@xcite , t2k phase 2@xcite , lbne@xcite , lbno@xcite ) , in which neutrinos are produced in pion decays , a neutrino factory@xcite , in which neutrinos are produced in muon decays , or a beta beam@xcite , in which neutrinos are produced in beta decays of radioactive isotopes .
in the accelerator neutrino experiments , the appearance oscillation probabilities @xmath12 and @xmath13 ( @xmath14 and @xmath15 in the case of a neutrino factory and a beta beam ) are measured , and one would naively expect that the oscillation probabilities of these two channels are sufficient to determine the cp phase @xmath3 .
it is known , unfortunately , that even if the values of the oscillation probabilities @xmath12 and @xmath13 are exactly given , we can not determine uniquely the values of the oscillation parameters due to eightfold parameter degeneracy .
this eightfold parameter degeneracy@xcite consists of three kinds of degeneracies , the intrinsic degeneracy@xcite , the sign degeneracy@xcite ( @xmath16 ) and the octant degeneracy@xcite ( @xmath17 ) . to see the eightfold degeneracy , it is convenient for the plot to give eight different points for different eight solutions . in ref.@xcite
it was shown that the solution specified by @xmath18 const . and @xmath19 const .
gives a quadratic curve ( a hyperbola in most cases ) in the @xmath20 plane . in fig.[fig:1
] , @xmath21 const . and @xmath22 const .
gives two different quadratic curves ( due to the sign degeneracy ) depending on the pattern of mass hierarchy , and each curve in general has two intersections ( due to the intrinsic degeneracy ) with one of two horizontal lines @xmath23 ( two lines appear due to the octant degeneracy ) . _ the eight - fold degeneracy in the @xmath20 plane .
when the two probabilities @xmath24 const . and @xmath25 const .
are given , it gives us a quadratic curve for each mass hierarchy .
nh ( solid ) and ih ( dashed ) stand for normal ( @xmath26 ) and inverted ( @xmath27 ) hierarchy , respectively . _ ] _ the allowed regions at best - fit ( the shaded area bounded by the thin quadratic curve ) and at 90%cl ( the wider area bounded by the thick quadratic curves and the line @xmath28 ) from the t2k appearance result@xcite .
the regions bounded by solid ( dashed ) curves are for normal ( inverted ) hierarchy .
the horizontal ( vertical ) band @xmath29 ( @xmath30 ) bounded by thick straight lines stands for the allowed region at 90%cl of the atmospheric ( reactor ) neutrino experiments , respectively .
the horizontal ( vertical ) thin straight line(s ) @xmath31 , @xmath32 ( @xmath33 ) stand for the best - fit value(s ) for the atmospheric ( reactor ) neutrino data . _ ] _ the dependence of the value of the cp phase @xmath34 which is obtained with a wrong ansatz with respect to the intrinsic , sign and octant degeneracies as a function of the true value of @xmath3 from the t2k appearance results on @xmath35 and @xmath36 .
@xmath37 , @xmath38 is assumed .
_ ] each quadratic curve in fig.[fig:1 ] shrinks to a straight line when the experiment is performed at the oscillation maximum @xmath39 , and this is the case at the t2k experiment .
however , the present result@xcite is for the neutrino mode @xmath35 only , and @xmath21 const . gives the inside of the region bounded by some quadratic curve ( the shaded area bounded by the thin quadratic curve in fig.[fig : t2k ] ) , instead of a straight line .
since the t2k result comes with the experimental error , the allowed region at 90 % cl becomes ( bounded by the thick quadratic curve and the line @xmath28 ) much wider than that by the best - fit value in fig.[fig : t2k ] . in the future the t2k experiment will measure both @xmath40 and @xmath41 with intense neutrino beams , and the error is expected to be reduced . even with the reduced error , however , it is not obvious whether parameter degeneracy can be resolved .
it turns out that the sign degeneracy is most serious to determine @xmath3 .
this can be seen from fig.[fig:2 ] , where the value of @xmath42 , which is obtained with a wrong ansatz with respect to the intrinsic , sign , and octant degeneracies , respectively , is plotted as a function of the true value of @xmath43 in the case of t2k setup for @xmath37 and @xmath44 .
since the sign degeneracy can be lifted with a long baseline experiment with a baseline length @xmath45 1000 km because of a large matter effect , it is expected to be resolved by future experiments with very long baselines .
it is expected that the accelerator long baseline experiments with intense neutrino beams will enable us not only to measure precisely the oscillation parameters of the three flavor framework but also to probe new physics by looking for deviation from the standard scenario with three massive neutrinos .
in particular , at accelerator neutrino experiments with high energy neutrinos we can study the channels @xmath47 and @xmath48 with opera - like detectors .
these experiments include the minsis proposal @xcite , in which the numi beam is used to look for @xmath46 at short baseline , and a neutrino factory@xcite .
is expected to be difficult in this scenario . however , @xmath46 detection may be possible in a future extension to the high energy neutrino factory . ]
these channels are expected to give stronger constraints on new physics than those which are obtained by the present experiments . here
i will describe a few examples of new physics , where @xmath46 detection can improve on the limits .
let us suppose that a non - standard interaction @xmath49 exists due to new physics , where @xmath50 is the coefficient of the non - standard interaction , normalized in terms of the weak scale , @xmath51 and @xmath52 are neutrinos and charged leptons of flavor @xmath53 , and @xmath54 and @xmath55 stand for fermions ( the only relevant ones are electrons , u and d quarks ) .
then this interaction predicts the exotic reactions such as @xmath56 or @xmath57 in the process of production and detection of neutrinos@xcite .
sensitivity to @xmath58 has been investigated by several groups . among others ,
@xcite showed that the sensitivity of a neutrino factory to @xmath59 ( @xmath60 ) is excellent . by various experiments . ]
it is known @xcite that in generic see - saw models the kinetic term gets modified after integrating out the right handed neutrino and unitarity is expected to be violated . in the case of the so - called minimal unitarity violation , in which only three light neutrinos are involved and sources of unitarity violation are assumed to appear only in the neutrino sector , unitarity violation is strongly constrained .
its constraint mostly comes from the bounds of rare decays of charged leptons , and deviation from unitarity is smaller than @xmath61(1% ) @xcite .
when the mixing matrix @xmath62 is nonunitary , deviation from unitarity is expressed as @xmath63 .
@xcite studied sensitivity to @xmath64 at a neutrino factory , and the bound is stronger than that of the present one from @xmath65 .
the anomaly which was announced by the lsnd group@xcite would imply mass squared difference of @xmath61(1 ) ev@xmath66 if it is interpreted as a phenomenon due to neutrino oscillation @xmath67 .
the standard three flavor scheme has only two independent mass squared differences , i.e. , @xmath68ev@xmath66 ( @xmath69ev@xmath66 ) for the solar ( atmospheric ) neutrino oscillation .
to accommodate a neutrino oscillation scheme to the lsnd anomaly , therefore , the extra state should be introduced .
this extra state should be sterile neutrino , which is singlet with respect to the gauge group of the standard model , because the number of weakly interacting light neutrinos should be three from the lep data@xcite . to test the lsnd anomaly ,
the miniboone experiment has been performed , but their results@xcite seem to be inconclusive . on the other hand , the flux of the reactor neutrino was recalculated in ref.@xcite and it was claimed that the normalization is shifted by about + 3% on average .
this claim is qualitatively consistent with an independent calculation in ref.@xcite .
if their claim on the reactor neutrino flux is correct , then neutrino oscillation with @xmath701ev@xmath66 may be concluded from a re - analysis of 19 reactor neutrino results at short baselines@xcite . this is called reactor anomaly .
furthermore , it was pointed out in ref.@xcite that the measured and predicted @xmath71ge production rates differ in the gallium radioactive source experiments gallex and sage , and this is called gallium anomaly
. the anomalies of lsnd , reactor and gallium constitute the main motivation for study of sterile neutrino oscillations .
sensitivity to sterile neutrino mixing at the minsis proposal was studied in ref.@xcite , and it was shown that minsis can improve on the current bound by a factor 100 for the @xmath47 channel . in a neutrino scheme with one sterile neutrino state , there are three cp phases , and it was shown in ref.@xcite that the largest cp violation can potentially occur in the @xmath47 channel with the present constraints from other experiments . recently a project called nustorm@xcite was proposed as a low energy prototype for a neutrino factory . although the low energy neutrino beam of nustorm does not allow us to create @xmath72 s , it has a good sensitivity to the sterile neutrino mixing angles .
in the standard framework of three massive neutrinos , all the three mixing angles have been determined , and the remaining quantities to be measured are the sign of @xmath8 , the sign of @xmath9 , and the cp phase @xmath3 .
these quantities are expected to be determined in the future experiments , by resolving parameter degeneracies .
if the neutrino energy is sufficiently high as in the minsis proposal or the high energy option of a neutrino factory , then one can detect @xmath46 with an opera - like detector . in this case
, we can probe new physics , such as non - standard interactions , unitarity violation , or light sterile neutrino scenarios , so @xmath46 detection in neutrino oscillations deserves further study .
i would like to thank the organizers for invitation and hospitality during the workshop .
this research was partly supported by a grant - in - aid for scientific research of the ministry of education , science and culture , under grant no .
24540281 .
j. beringer _
[ particle data group collaboration ] , phys .
d * 86 * ( 2012 ) 010001 .
k. abe _ et al . _
[ t2k collaboration ] , phys .
lett . * 107 * ( 2011 ) 041801 [ arxiv:1106.2822 [ hep - ex ] ] .
p. adamson _ et al .
_ [ minos collaboration ] , phys . rev .
* 107 * ( 2011 ) 181802 [ arxiv:1108.0015 [ hep - ex ] ] . y. abe _ et al . _
[ double - chooz collaboration ] , phys .
* 108 * ( 2012 ) 131801 [ arxiv:1112.6353 [ hep - ex ] ] .
an _ et al . _
[ daya - bay collaboration ] , phys .
* 108 * ( 2012 ) 171803 [ arxiv:1203.1669 [ hep - ex ] ] .
j. k. ahn _
[ reno collaboration ] , phys .
* 108 * ( 2012 ) 191802 [ arxiv:1204.0626 [ hep - ex ] ] .
d. v. forero , m. tortola and j. w. f. valle , phys .
d * 86 * ( 2012 ) 073012 [ arxiv:1205.4018 [ hep - ph ] ] .
g. l. fogli , e. lisi , a. marrone , d. montanino , a. palazzo and a. m. rotunno , phys .
d * 86 * ( 2012 ) 013012 [ arxiv:1205.5254 [ hep - ph ] ] .
m. c. gonzalez - garcia , m. maltoni , j. salvado and t. schwetz , arxiv:1209.3023 [ hep - ph ] .
y. itow , talk at _
25th international conference on neutrino physics and astrophysics _
( neutrino 2012 ) , kyoto , japan , june 39 , 2012 . d. s. ayres _ et al . _ [ nova collaboration ] , hep - ex/0503053 .
y. itow _ et al . _
[ t2k collaboration ] , hep - ex/0106019 .
k. abe , t. abe , h. aihara , y. fukuda , y. hayato , k. huang , a. k. ichikawa and m. ikeda _ et al .
_ , arxiv:1109.3262 [ hep - ex ] .
a. badertscher , t. hasegawa , t. kobayashi , a. marchionni , a. meregaglia , t. maruyama , k. nishikawa and a. rubbia , arxiv:0804.2111 [ hep - ph ] .
j. goon _
[ lbne collaboration ] , arxiv:1204.2295 [ physics.ins-det ] .
a. stahl , c. wiebusch , a. m. guler , m. kamiscioglu , r. sever , a. u. yilmazer , c. gunes and d. yilmaz _ et al .
_ , cern - spsc-2012 - 021
. s. geer , phys . rev .
d * 57 * ( 1998 ) 6989 [ erratum - ibid .
d * 59 * ( 1999 ) 039903 ] [ hep - ph/9712290 ] .
a. bandyopadhyay _ et al .
_ [ iss physics working group collaboration ] , rept .
* 72 * ( 2009 ) 106201 [ arxiv:0710.4947 [ hep - ph ] ] .
p. zucchelli , phys .
b * 532 * ( 2002 ) 166 .
v. barger , d. marfatia and k. whisnant , phys .
d * 65 * ( 2002 ) 073023 [ hep - ph/0112119 ] .
j. burguet - castell , m. b. gavela , j. j. gomez - cadenas , p. hernandez and o. mena , nucl .
b * 608 * ( 2001 ) 301 [ hep - ph/0103258 ] .
h. minakata and h. nunokawa , jhep * 0110 * ( 2001 ) 001 [ hep - ph/0108085 ] .
g. l. fogli and e. lisi , phys .
d * 54 * ( 1996 ) 3667 [ hep - ph/9604415 ] .
o. yasuda , new j. phys .
* 6 * ( 2004 ) 83 [ hep - ph/0405005 ] .
s. bertolucci , a. blondel , a. cervera , a. donini , m. dracos , d. duchesneau , f. dufour and r. edgecock _ et al .
_ , arxiv:1208.0512 [ hep - ex ] . the minsis proposal , + http://www-off-axis.fnal.gov/minsis/index.html .
r. alonso , s. antusch , m. blennow , p. coloma , a. de gouvea , e. fernandez - martinez , b. gavela and c. gonzalez - garcia _ et al .
_ , arxiv:1009.0476 [ hep - ph ]
. s. choubey _ et al . _
[ ids - nf collaboration ] , arxiv:1112.2853 [ hep - ex ] .
y. grossman , phys .
b * 359 * ( 1995 ) 141 [ arxiv : hep - ph/9507344 ] .
m. c. gonzalez - garcia , y. grossman , a. gusso and y. nir , phys .
d * 64 * ( 2001 ) 096006 [ arxiv : hep - ph/0105159 ] .
t. ota , j. sato and n. a. yamashita , phys .
d * 65 * ( 2002 ) 093015 [ arxiv : hep - ph/0112329 ] .
s. davidson , c. pena - garay , n. rius and a. santamaria , jhep * 0303 * ( 2003 ) 011 [ hep - ph/0302093 ] .
s. antusch , c. biggio , e. fernandez - martinez , m. b. gavela and j. lopez - pavon , jhep * 0610 * ( 2006 ) 084 [ arxiv : hep - ph/0607020 ] .
e. fernandez - martinez , m. b. gavela , j. lopez - pavon and o. yasuda , phys .
b * 649 * ( 2007 ) 427 [ arxiv : hep - ph/0703098 ] .
s. antusch , m. blennow , e. fernandez - martinez and j. lopez - pavon , phys .
d * 80 * , 033002 ( 2009 ) [ arxiv:0903.3986 [ hep - ph ] ] .
a. aguilar _ et al . _
[ lsnd collaboration ] , phys .
d * 64 * ( 2001 ) 112007 [ arxiv : hep - ex/0104049 ] .
a. a. aguilar - arevalo _ et al .
_ [ the miniboone collaboration ] , phys . rev .
* 98 * ( 2007 ) 231801 [ arxiv:0704.1500 [ hep - ex ] ] .
a. a. aguilar - arevalo _ et al . _
[ the miniboone collaboration ] , phys .
* 105 * ( 2010 ) 181801 [ arxiv:1007.1150 [ hep - ex ] ] .
t. a. mueller _ et al .
_ , phys .
c * 83 * ( 2011 ) 054615 [ arxiv:1101.2663 [ hep - ex ] ] .
huber , arxiv:1106.0687 [ hep - ph ] .
g. mention , m. fechner , t. lasserre , t. a. mueller , d. lhuillier , m. cribier and a. letourneau , phys .
d * 83 * ( 2011 ) 073006 [ arxiv:1101.2755 [ hep - ex ] ] . c. giunti and m. laveder , phys .
d * 77 * ( 2008 ) 093002 [ arxiv:0707.4593 [ hep - ph ] ] .
a. donini , k. -i .
fuki , j. lopez - pavon , d. meloni and o. yasuda , jhep * 0908 * ( 2009 ) 041 [ arxiv:0812.3703 [ hep - ph ] ] .
p. kyberd _ et al .
_ [ nustorm collaboration ] , arxiv:1206.0294 [ hep - ex ] . | status of neutrino oscillation study is briefly reviewed .
some aspects relevant to tau physics are also described .
neutrino oscillation , parameter degeneracy , new physics |
cobham s theorem is related to numeration systems and can be considered as a classical result in formal languages theory .
it is formulated as follows .
let @xmath0 be two multiplicatively independent integers ( i.e. , the only integers satisfying @xmath3 are @xmath4 ) .
if a subset @xmath5 of integers is both @xmath1- and @xmath2-recognizable then it is a finite union of arithmetic progressions ( i.e. , @xmath6 is an _ ultimately periodic _ set ) . recall that @xmath7 is said to be _
@xmath1-recognizable _ if the language @xmath8 of the @xmath1-ary representations ( without leading zeroes ) of the elements in @xmath6 is a regular language accepted by a finite automaton ( see for instance ( * ? ? ?
this famous result has been widely studied from various points of view ( we give here just a few references ) : extension to non - standard numeration systems @xcite or to the framework of @xmath9-regular sequences @xcite , study of the multidimensional case ( known as cobham - semenov s theorem ) @xcite , alternative proofs using the formalism of the first order logic @xcite , . the original proof due to cobham
is widely considered as rather difficult @xcite . in his book , s. eilenberg proposed as a challenge to find an easier proof @xcite .
the major improvements in the simplification of the proof of cobham s theorem were made by g. hansel in @xcite where he makes use of the notion of syndeticity and sketches the key - points leading to the result . recall that an infinite set of integers @xmath10 is said to be _
if there exists @xmath11 such that for all @xmath12 , @xmath13 .
( notice that hansel s ideas about syndeticity also hold in a wider framework than @xmath1-ary numeration systems @xcite . ) afterwards , a great work of presentation relying on the main ideas found in @xcite was made by several authors @xcite .
unfortunately , in these last two documents a same mistake can be found ( statement [ lem : faux ] below is not correct and example [ exa : c ] is a counter - example ) . in this note ,
our modest contribution is to correct this error using as simple arguments as possible . in the spirit , we are naturally close to @xcite and @xcite but new ideas appear in our reasoning .
finally , we hope that this erratum added to @xcite or @xcite will now give a complete presentation of the proof of cobham s theorem .
let us set @xmath14 as the alphabet of the @xmath1-ary digits . in @xcite , the following result is presented .
[ lem : faux ] if an infinite @xmath1-recognizable set @xmath5 is such that @xmath15 is right dense , i.e. , for all @xmath16 there exists @xmath17 such that @xmath18 , then @xmath6 is syndetic .
[ exa : c ] as stated above , statement [ lem : faux ] is not correct .
an easy counter - example is given by the following set @xmath6 of integers @xmath19 indeed , this set is @xmath20-recognizable : @xmath21 , and trivially right dense but not syndetic . in the literature ,
statement [ lem : faux ] is generally presented to obtain the following proposition .
[ pro:1 ] ( * ? ? ?
5 ) let @xmath0 be two multiplicatively independent integers . if an infinite set of integers if both @xmath1- and @xmath2-recognizable , then it is syndetic . in substance ,
this latter result can naturally be found in cobham s work ( see ( * ? ? ? * lemma 3 ) ) . in this note ,
our aim is to give an alternative proof of proposition [ pro:1 ] not using statement [ lem : faux ] .
our approach relies on five easy lemmas .
we assume that the reader has some basic knowledge in automata theory ( see for instance @xcite ) .
if @xmath5 is a set of integers , we define a mapping ( or a right - infinite word ) @xmath22 such that @xmath23 if and only if @xmath24 .
if @xmath25 is a finite word , @xmath26 denotes its length .
[ lem:1 ] let @xmath27 be a dfa ( deterministic finite automaton ) with @xmath28 as transition function .
for any state @xmath29 , the set @xmath30 is such that @xmath31 is ultimately periodic , i.e. , there exist @xmath32 and @xmath33 such that for all @xmath34 , @xmath35 . for any state @xmath29
, we define a mapping @xmath36 since @xmath37 is finite , there exist @xmath38 and @xmath39 such that @xmath40 and @xmath41 .
obviously , for any @xmath42 , @xmath43 .
consequently for all @xmath44 , @xmath45 in other words , @xmath46 is ultimately periodic : @xmath47 if @xmath48 . to conclude the proof ,
observe that @xmath49 where @xmath50 .
it is enough to find integers @xmath56 satisfying @xmath57 this is a direct consequence of kronecker s theorem ( because @xmath58 and @xmath59 are still multiplicatively independent hence @xmath60 is irrational ) @xcite .
[ lem:3 ] let @xmath61 and @xmath5 be an infinite @xmath1-recognizable set .
then there exist integers @xmath62 such that for all @xmath63 , the set @xmath64 is nonempty .
moreover , the integer @xmath65 can be chosen arbitrarily large .
let @xmath27 be a dfa recognizing @xmath8 .
since @xmath6 is infinite , there exists @xmath66 arbitrarily large such that @xmath67 is prefix of an infinite number of elements in @xmath8 .
let @xmath68 .
by lemma [ lem:1 ] , there exist @xmath69 and @xmath70 such that @xmath71 for all @xmath72 . for any @xmath73
, the interval @xmath74 contains all the integers having a @xmath1-ary representation of the form @xmath75 with @xmath76 .
since the set @xmath77 is infinite , there exists a word @xmath78 such that @xmath79 is the @xmath1-ary representation of an element in @xmath6 with @xmath80 .
take @xmath81 .
consequently , the interval @xmath82 contains an element belonging to @xmath6 .
the conclusion follows from the periodicity of @xmath31 : @xmath83 , for all @xmath84 .
recall that a state @xmath85 is said to be _ accessible _ ( resp .
_ coaccessible _ ) if there exists a word @xmath25 such that @xmath86 ( resp .
@xmath87 ) .
the _ trimmed _ minimal automaton of a language @xmath88 is obtained by taking only states which are accessible and coaccessible .
[ lem:4 ] let @xmath61 and @xmath5 be an infinite @xmath1-recognizable set such that @xmath89 is the trimmed minimal automaton of @xmath8 . if there exists a state @xmath85 such that @xmath90 is infinite , then there exist integers @xmath62 such that for all @xmath63 , the set @xmath64 is empty .
let @xmath85 be a state such that @xmath90 is infinite .
without loss of generality , we may assume that @xmath91 and there exists @xmath66 such that @xmath92 .
( indeed , if @xmath93 is infinite then the same property holds for some other state @xmath85 . )
we use the same reasoning as in the previous proof .
thanks to lemma [ lem:1 ] , there exist @xmath94 and @xmath70 such that @xmath71 for all @xmath72 .
since @xmath90 is infinite , there exists @xmath95 such that no word @xmath78 of length @xmath96 is such that @xmath97 .
in other words , if @xmath98 then @xmath99 and the interval @xmath82 does not contain any element of @xmath6 . once again , the conclusion follows from the periodicity of @xmath31 .
[ lem:5 ] let @xmath100 be two multiplicatively independent integers and @xmath5 be an infinite @xmath1- and @xmath2-recognizable set of integers .
if @xmath89 is trimmed minimal automaton of @xmath101 , then for any state @xmath102 , the set @xmath103 is cofinite .
assume that @xmath111 .
let @xmath27 be the trimmed minimal automaton of @xmath101 . for all @xmath112 ,
we write @xmath113 .
thanks to lemma [ lem:5 ] , @xmath114 is cofinite .
this means that for all @xmath44 , there exists @xmath115 such that for all @xmath116 , @xmath9 belongs to @xmath114 .
clearly , @xmath115 depends only on the state @xmath117 and there are a finite number of such states .
let @xmath118 .
consequently , for any @xmath119 , there exists a word @xmath120 of length @xmath121 such that @xmath122 .
in other words , for any @xmath119 , there exist @xmath123 such that @xmath124 .
we conclude that any interval of length @xmath125 contains at least an element belonging to @xmath6 . c. michaux , r. villemaire , presburger arithmetic and recognizability of sets of natural numbers by automata : new proofs of cobham s and semenov s theorems , _ ann .
. logic _ * 77 * ( 1996 ) , 251277 . | in this note , we give an alternative proof of the following result .
let @xmath0 be two multiplicatively independent integers .
if an infinite set of integers is both @xmath1- and @xmath2-recognizable , then it is syndetic .
notice that this result is needed in the classical proof of the celebrated cobham s theorem .
therefore the aim of this paper is to complete @xcite and @xcite to obtain an accessible proof of cobham s theorem . |
quasiparticle self - consistency is a construction that determines the noninteracting green s function @xmath1 that is minimally distant from the true green s function @xmath2 .
a measure of distance , or metric is necessary ; a good choice @xcite results in an effective static potential : @xmath56+{\rm re}[\sigma({\varepsilon_j})]}_{ij } \right\ } \langle\psi_j| .
\label{eq : veff}\end{aligned}\ ] ] @xmath57 is the self - energy in the basis of single - particle eigenstates @xmath58 , which becomes @xmath59 in the _ gw _ approximation .
starting from a trial @xmath60 , e.g. the lda , @xmath61 is determined through _ gw _ , which determines a new @xmath60 .
the cycle is repeated until self - consistency .
another key property of eq .
[ eq : veff ] is that , at self - consistency , the poles of @xmath3 coincide with the peaks in @xmath4
. therefore the band structure generated by @xmath64 coincides with the peaks of the specral function @xmath19 .
this is significant , because it means there is no many - body `` mass renormalization '' of the noninteracting hamiltonian . in other words ,
the attribution of mass renormalization to correlation effects , a concept widely used in the literature @xcite , is ill - defined : it depends on an arbitrary reference , e.g. the lda .
the absence of mass renormalization is a very useful property : we can directly associate qs__gw _ _ energy bands @xmath5 with peaks in the spectral function @xmath6 . in comparing the areas of electron pockets in the fermi surface of fe , we pointed out the small discrepancy between measured and simulated values .
as a measure of the discrepancy we provide the rigid shift of the fermi energy that would lead the computed area to equate the measured one ( see table ii in main text ) .
this measure is justified under the assumption that qs__gw _ _ shift rigidly , which is actually not the case for pocket vi , for which the agreement is poorer . according to a thomas - fermi model of screening @xcite , the elecron - phonon interaction renormalizes @xmath65 by a factor @xmath66 .
band vi is roughly spherical , enabling us to evaluate @xmath67 analytically : @xmath68\ ] ] estimating @xmath69 @xmath22 from the fe electron density , this leads to a renormalization factor of @xmath70 . in both qs__gw
_ _ calculations of fe and ni , a @xmath52 mesh of 12@xmath2812@xmath2812 divisions was found to be sufficient for calculating @xmath10 .
the one - body part was evaluated on a 24@xmath2824@xmath2824 mesh .
fe 3__p _ _ and 4__d _ _ states were included through local orbitals : omitting these and treating 3@xmath73 as core levels @xcite can shift qp levels by as much as 0.1 ev in the fl regime .
other parameters @xcite , such as broadening the pole in @xmath2 in constructing @xmath74 , the basis of eigenfunction products , and the energy cutoff for the off - diagonal parts of @xmath10 , were also carefully monitored . when set to tight tolerances qp levels near @xmath39 were stable to a resolution of 0.05 ev .
qp levels are calculated including spin - orbit coupling ( soc ) , though it is omitted in the calculation of @xmath10 .
the effect of soc on @xmath10 was found to make small changes to @xmath61 . concerning the qs__gw__+dmft calculation on ni , we projected the lattice problem on the ni @xmath0 orbitals following the prescription of haule @xcite .
we compute higher level diagrams locally using the hybridization expansion version of the numerically exact continuous time qmc method @xcite . in order to single out the correlated subspace , a procedure of projection / embedding which
was originally introduced in @xcite in the lapw basis of the wien2k package , is developed in the full - potential linear muffin - tin orbitals ( fp - lmto ) basis @xcite .
this projector maps the full space green s function @xmath75 ( with band and @xmath52-point label @xmath76 ) to the local green s function @xmath77 defined only on the correlated subspace .
the compact index @xmath78 , collects information on the atom type @xmath79 , site @xmath80 , spin coordinate @xmath81 , and angular momentum components @xmath82 and @xmath83 .
the projection operation can be cast in the following form : @xmath84 where the coefficients @xmath85 account for localization inside the sphere , while @xmath86 gives an estimate of correlations relative to the specific orbital component @xmath82 .
more specifically @xmath85 are linear combinations of spherical harmonics @xmath87 and the qs__gw _ _ quasiparticle eigenfunctions in the fp - lmto basis .
the terms @xmath88 are radial integral of the kind @xmath89 where the index @xmath90 in @xmath91 indicates the possibility of selecting the radial solution @xmath92 of the schrdinger equation inside the mto , its energy derivative @xmath93 , and its local orbitals contributions @xmath94 @xcite . by means of these definitions
we ensure that the localized orbitals are centred on the correlated atom corresponding to the muffin - tin site @xmath80 .
the local green s function is defined on a grid of matsubara frequencies @xmath97 and it is employed to calculate the hybridization function of the system , which feeds the ctqmc impurity solver .
the result of the impurity solver is the local impurity self - energy @xmath98 also defined on the matsubara axis . in order to update the full green s function @xmath75 with this local self - energy , so to iterate the dmft loop to self - consistency , an embedding procedure is needed .
because of the specific properties of the transformation @xmath95 , the embedding procedure @xmath99 can be operated by means of the same matrices , even though this is not a general requirement of the theory @xcite .
the output of the dmft loop is the impurity self - energy @xmath103 and the corresponding impurity green s function @xmath104 . at self - consistency
they correspond to the local self - energy and local green s function of the correlated subset ( @xmath0-electrons of the metal in this case ) .
though , since the ctqmc solver works in the matsubara s frequency space extrapolations to zero - energy or some kind of analytical continuation has to be employed to extract physically meaningful quantities .
we evaluated the scattering rate from the intercept at @xmath105 of a forth - order polynomial obtained from a fit of @xmath106 in the vicinity of @xmath105 .
the coefficient @xmath107 of the linear term is then related to the quasiparticle weight @xmath108 .
the intercept being @xmath109 where @xmath7 is the scattering rate , we have been able to evaluate @xmath7 between @xmath110 and @xmath111 for both fe and ni , corresponding to a coherent charge - scattering regime .
the spectral function of the correlated subset can be obtained by continuing on the real - frequency axis the quantity @xmath112 . in order to do that , we utilised a statistical technique based on the maximum entropy method @xcite . as model function
we used the magnetic qs__gw _ _ spectral function of the @xmath0 orbitals .
the resulting spectral function @xmath113 contains all the effects of the dynamical local response including features beyond the quasiparticle peaks , as satellites at higher energy .
this is indeed the case for fe , reported in figures [ fig : fe_spf ] . in our application to ni
, the projectors used for the ni @xmath102 are constructed from 5 bands below @xmath39 and 3 bands above @xmath39 , which correspond to a window of @xmath114 ev . by choosing a wide energy window , @xmath95 becomes nearly static @xcite .
the corresponding on - site coulomb parameters were chosen to be @xmath95=10 ev and @xmath115=0.9 ev , as calculated by constrained rpa @xcite .
the matsubara frequency grid is defined over 2000 points with an inverse temperature @xmath116 ev@xmath22 .
standard procedure : : we first perform a full qs__gw _ _ loop
. then we extract the spin - averaged green s function from which the local hybridization function @xmath117 is obtained .
the dmft solver uses @xmath117 , @xmath95 and @xmath115 to produce the local self - energy @xmath98 . by keeping the qs__gw _ _ part of the green s function fixed , the calculation of the self - energy is repeated until convergence of the magnetic moment ( dmft loop )
. then a new density @xmath118 is recomputed summing over all matsubara s frenquencies according to @xmath119 where @xmath120 are the right ( @xmath121 ) and left ( @xmath122 ) eigenfunctions of the dmft hamiltonian @xmath123 with corresponding eigenvalues @xmath124 . a new qs__gw _ _ loop
is then converged by keeping the density fixed , that produces a new qs__gw _ _ green s function .
the output is used to initialise a new dmft loop and so on until qs__gw _ _ self - energy and dmft updated density are self - consistent .
+ the number of iterations in each dmft loop varies between 10 and 20 , and the number of random moves per iteration are roughly @xmath125 .
this method is basically the equivalent of state - of - the - art dmft implementations in the framework of dft+dmft .
we believe it is superior to other @xmath45+dmft implementations because of the possibility to close the full self - consistent loop and because of the higher quality of the low - level theory chosen .
qs__gw__+sldmft : : from the converged dmft self - energy , we first extrapolate the static limit of @xmath98 , then we embed it into the lattice problem and we keep only its symmetrized real part @xmath126/2 .\ ] ] this is done for the spin - up and spin - down channels separately .
we finally retain only the spin - flip component @xmath127 and we add it to the charge component @xmath128 computed at qs__gw _ _ level .
this procedure allowed us to prevent counting twice the magnetic contributions to the self - energy . to further confirm that the local momentum is not relevant in the description of fe
, we applied the fully dynamical qs__gw__+dmft approach also to fe .
the parameters @xmath129 ev and @xmath130 ev used are the same as in@xcite , where they have been derived from gw within the same projection scheme .
four valence and four conduction bands have been used to project onto the correlated system .
the resulting magnetic moment @xmath49=2.20 bohr , as the qs__gw _ _ prediction , confirming that local corrections are absolutely negligible in fe .
we extract the scattering rate @xmath7 from the intercept and the linear term of a forth - order polynomial fitting im@xmath10(@xmath8 ) in the vicinity of @xmath105 as explained above .
we find that fe has extremely coherent charge scattering .
we also present the analytically continued img(i@xmath8)/@xmath131 for fe reported in figure [ fig : fe_spf ] .
99 j. schfer , m. hoinkis , e. rotenberg , p. blaha , and r. claessen , phys .
b * 72 * , 155115 ( 2005 ) .
j. callaway and c. s. wang , phys .
b*16 * , 2095 ( 1977 ) j. snchez - barriga , et al , phys .
* 103 * , 267203 ( 2009 ) a. l. walter , j. d. riley , and o. rader , new j. phys . * 12 * , 013007 ( 2010 ) l. hedin , j. phys . : condens
. matter * 11 * , r489 ( 1999 ) t. kotani , m. van schilfgaarde , and s. v. faleev , phys . rev .
b * 76 * , 165106 ( 2007 ) .
our code was adapted from the original ecalj code ( github.com/tkotani/ecalj ) .
s. ismail - beigi , arxiv:1406.0772 ( 2014 ) d. r. baraff , phys .
b * 8 * , 3439 ( 1973 ) a. gold , j. low temp .
16 * , 3 ( 1974 ) a. santoni and f. j. himpsel , phys .
b * 43 * , 1305 ( 1991 ) j. pendry , surface science * 57 * , 679 ( 1976 ) r. matzdorf , surface science reports * 30 * , 153 ( 1998 ) v. n. strocov , j. elec . spect . and rel .
phenomena * 130 * , 65 ( 2003 ) v. n. strocov , r. claessen , g. nicolay , s. hfner , a. kimura , a. harasawa , s. shin , a. kakizaki , p. nilsson , h. starnberg , _ et al .
lett * 81 * , 4943 ( 1998 ) p. j. feibelman and d. eastman , phys .
b * 10 * , 4932 ( 1974 ) s. tanuma , c. j. powell , and d. r. penn , surface and interface analysis * 21 * , 165 ( 1994 ) g. g. lonzarich , in _ electrons at the fermi surface _ , edited by m. springford ( cambridge university press , 1980 ) a. v. gold , l. hodges , p. t. panousis , and d. r. stone , int . j. magn * 2 * , 375 ( 1971 ) for further details on the electron - phonon coupling model and on the qs__gw__+dmft implementation and application to fe and ni , see supplemental material . k. d. belashchenko , v. p. antropov , and n. e. zein , phys . rev .
b * 73 * , 073105 ( 2006 ) f. j. himpsel , j. a. knapp , and d. e. eastman , phys . rev .
b * 19 * , 2919 ( 1979 ) k. karlsson and f. aryasetiawan , phys .
b * 62 * , 3006 ( 2000 ) i. mazin , d. singh , and a. aguyao , arxiv : cond - mat/0401563 m. shimizu , rep . prog
. phys . * 44 * , 329 ( 1981 ) w. eberhardt and e. w. plummer , phys .
b * 21 * , 3245 ( 1980 ) s. biermann , f. aryasetiawan and a. georges , phys .
* 90 * , 086402 ( 2003 ) .
the main accomplishment of this work was to reproduce within gw+dmft the satellite at @xmath46 - 6 ev .
this is also captured by lda+dmft @xcite .
k. haule , phys .
b * 75 * , 155113 ( 2007 ) n. w. ashcroft and d. mermin , _ solid state physics _
( brooks / cole , 1976 ) k. haule , c. h. yee , and k. kim , phys .
b * 81 * , 195107 ( 2010 ) p. werner , a. comanac , l. de medici , m. troyer , and a. j. millis , phys
lett . * 97 * , 076405 ( 2006 ) m. methfessel , m. van schilfgaarde and r. a. casali , _ electronic structure and physical properties of solids : the uses of the lmto method , lecture notes in physics _ , vol .
535 ( springer - verlag , berlin , 2000 ) t. kotani , m. van schilfgaarde , and s. faleev , phys .
b * 76 * , 165106 ( 2007 ) s. ismail - beigi , arxiv:1406.0772 ( 2014 ) a. l. walter , j. d. riley , and o. rader , new j. phys . *
12 * , 013007 ( 2010 ) n. w. ashcroft and d. mermin , _ solid state physics _ ( brooks / cole , 1976 ) j. schfer , m. hoinkis , e. rotenberg , p. blaha , and r. claessen , phys .
b * 72 * , 155115 ( 2005 ) .
k. haule , c. yee , and k. kim , phys .
b * 81 * , 195107 ( 2010 ) p. werner , a. comanac , l. de medici , m. troyer , and a. j. millis , phys . rev . lett .
* 97 * , 076405 ( 2006 ) k. haule , phys .
b * 75 * , 155113 ( 2007 ) m. methfessel , m. van schilfgaarde and r. a. casali , _ electronic structure and physical properties of solids : the uses of the lmto method , lecture notes in physics _ , vol
. 535 ( springer - verlag , berlin , 2000 ) j. e. gubernatis , m. jarrel , r. n. silver and d. s. sivia , phys .
b * 44 * , 6011 ( 1991 ) s. choi , a. kutepov , k. haule , m. van schilfgaarde , and g. kotliar , arxiv:1504.07569 ( 2015 ) meng wang , chenglin zhang , xingye lu , guotai tan , huiqian luo , yu song , miaoyin wang , xiaotian zhang , e.a .
goremychkin , t.g .
perring , t.a .
maier , zhiping yin , kristjan haule , gabriel kotliar , pengcheng dai , nature communications * 4 * , 2874 ( 2013 ) | we present a detailed study of local and non - local correlations in the electronic structure of elemental transition metals carried out by means of the quasiparticle self - consistent gw ( qs__gw _ _ ) and dynamical mean field theory ( dmft ) .
recent high resolution arpes and haas - van alphen data of two typical transition metal systems ( fe and ni ) are used as case study .
( i ) we find that the properties of fe are very well described by qs__gw__. agreement with cyclotron and very clean arpes measurements is excellent , provided that final - state scattering is taken into account .
this establishes the exceptional reliability of qs__gw _ _ also in metallic systems .
( ii ) nonetheless qs__gw _ _ alone is not able to provide an adequate description of the ni arpes data due to strong local spin fluctuations .
we surmount this deficiency by combining nonlocal charge fluctuations in qs__gw _ _ with local spin fluctuations in dmft ( qs__gw__+``magnetic dmft '' ) .
( iii ) finally we show that the dynamics of the local fluctuations are actually not crucial .
the addition of an external static field can lead to similarly good results if non - local correlations are included through qs__gw__. high - resolution spectroscopy is limited in transition metals , in part because it is difficult to make sufficiently high quality samples .
fe and ni are elements of which high quality films have been grown , and high - resolution angle - resolved photoemission spectroscopy ( arpes ) performed @xcite .
these experiments provides a good reference to test the validity of different approximations of the electronic structure .
there are also not many calculations of spectral functions in these materials .
fe has been studied in the local - density approximation ( lda ) @xcite and with corrections through dynamical - mean field theory ( dmft ) @xcite .
it is not surprising that the lda does not track the arpes experiment well @xcite , but it has been found that lda+dmft also fails to properly account for arpes data @xcite .
the _ gw _ approximation @xcite is widely applied to many kinds of insulators , but how well it describes 3@xmath0 transition metals is much less established . through quasiparticle self - consistency ( qs__gw _ _ )
one determines the noninteracting green s function @xmath1 which is minimally distant from the true green s function @xmath2 @xcite . within qs__gw
_ _ many electronic properties are in excellent agreement with experiment @xcite , most notably the quasiparticle band structures . moreover , at self - consistency the poles of qs__gw _ _ @xmath3 coincide with the peaks in @xmath4 .
this means that there is no many - body `` mass renormalization '' of the noninteracting hamiltonian , which allows for a direct association of qs__gw _ _ energy bands @xmath5 with peaks in the spectral function @xmath6 .
thus , qs__gw
_ _ provides an optimum framework to test the range of validity , and the limitations to the _ gw _ approximation . in this work
, we compare qs__gw _ _ results to various experimental data in elemental 3@xmath0 materials in the fermi liquid ( fl ) regime , with a heavy focus on fe because of the high quality of arpes @xcite and de haas - van alphen ( dhva ) @xcite data available .
we will show that qs__gw _ _ and arpes spectral functions agree to within experimental resolution , with the proviso that the final state scattering is properly accounted in interpreting the experimental data .
by contrast , discrepancies appear in ni a classical itinerant ferromagnet .
this can be attributed to the lack of spin fluctuations in _ gw _ diagrams .
however we find out that there is no need to include finite - energy spin fluctuations , instead a static correction to the qsgw self - energy is sufficient to correct for the size of the local moment .
while this finding is completely new within such an extensive formalism , it opens up an avenue to test the validity of a similar argument for other transition metals .
the lda or lda+dmft should be problematic , as nonlocality in the self - energy can be important ( see supplemental material ) .
fig .
[ fig : qsgwbands ] compares the calculated qs__gw _ _ band structure of fe to peaks in arpes spectra of ref .
@xcite , along with some inverse photoemission data @xcite . while agreement appears to be very good , there are some discrepancies , particularly along the @xmath7-h line ( see also fig . [
fig : gamh](a ) ) .
as noted earlier , the qs__gw _ _ band structure reflects the peaks of @xmath6 with no renormalizations from the @xmath8- or @xmath9- dependence of @xmath10 . in the fl regime , arpes spectra @xmath11
are generally thought to be a fairly direct measure of @xmath6 .
but the two are not identical even in the fl regime , independently of the precision of the experimental setup . assuming a one - step model @xcite for the photoemission process ( initial and final state coupled through fermi s golden rule @xcite ) @xmath11 can be written as @xmath12 is the spectral function of the final state , broadened by scattering of the photoelectron as it approaches the surface @xcite .
@xmath13 is the final - state surface transmission amplitude and @xmath14 the photoexcitation matrix element ( taken to be constant and @xmath15-independent @xcite ) .
thus the final state is considered to be a damped bloch wave , taking the form of a lorentzian distribution centred in @xmath16 and broadened by @xmath17 @xcite , while the initial state is an undamped bloch function with an energy broadening @xmath18 , obtained through the qs__gw _ _ spectral function .
this approximation is reasonable since in the fl regime @xmath19 is sharply peaked around the qp level .
@xmath20 is directly related to the inverse of the electron mean free path . for photon energy in the range 100 - 130ev , @xmath21@xmath22 @xcite .
+ + the final - state scattering broadens @xmath23 ; but it also can shift the peak @xmath24 in @xmath23 . the most significant discrepancy between arpes and qs__gw
_ _ is found in the v@xmath25 band , fig .
[ fig : gamh](@xmath26 ) between @xmath27 and 0.4@xmath28h . fig .
[ fig : gamh](@xmath29 ) shows @xmath30 calculated by qs__gw _ _ , and the corresponding @xmath31 calculated from eq .
( [ eq : pes ] ) . estimating the peak shift change from @xmath32 , we find @xmath33 ev at @xmath7 , increasing to @xmath34ev for _ k _ between 0.1h and 0.3h .
@xmath35=0.06ev tallies closely with the discrepancy between the v@xmath25 band and the measured arpes peak for [email protected] .
there is also a significant discrepancy in the ii@xmath37 band near @xmath38h . where it crosses @xmath39 , the qs__gw _ _ bands deviate from the arpes peak by nearly 0.15ev . but
arpes simulated by eq .
( [ eq : pes ] ) is much closer to experiment ( fig .
[ fig : gamh](@xmath40 ) ) .
this is understood from fig .
[ fig : gamh](@xmath0 ) , which plots the qs__gw _ _ dispersion along a line @xmath41 normal to the film surface , passing through [ 0,0,0.77h ] . a measurement that includes contributions from this line biases the arpes peak in the direction of @xmath39 since e@xmath42 is minimum at @xmath43 .
thus we attribute most of the discrepancy in the fermi surface crossing ( red star in [ fig : qsgwbands](@xmath29 ) ) to an artifact of final - state scattering .
to better pin down the errors in qs__gw _ _ , we turn to de haas - van alphen ( dhva ) measurements .
extremal areas of the fs cross sections can be extracted to high precision from dhva and magnetoresistance experiments .
areas normal to [ 110 ] and [ 111 ] are given in table [ tab : dhva ] , along with areas calculated by qs__gw__. fig.[fig : qsgwbands ] shows the qs__gw _ _ fermi surface , which closely resembles the one inferred by lonzarich ( version b ) @xcite .
there is some ambiguity in resolving the small viii@xmath25 pocket at n because its tiny area is sensitive to computational details .
discrepancies in the extremal areas are not very meaningful : it is more sensible to determine the change @xmath44 in fermi level needed to make the qs__gw _ _ area agree with dhva measurements .
this amounts to the average error in the qs__gw _ _ qp levels , assuming that the bands shift rigidly .
this assumption is well verified for all pockets , except for the small vi one owing to strong electron - phonon coupling @xcite . .
[ cols= " < , < , < , > , < , < , > " , ] local spin fluctuations are well captured by localized non perturbative approaches such as dmft .
we can reasonably expect that the addition of spin - flip diagrams to qs__gw _ _ would be sufficient to incorporate these effects .
a @xmath45+dmft study of ferromagnetic ni can be found in @xcite , but the dependency of @xmath45 on the starting point , together with all the advantages mentioned at the beginning , motivated us to devise a qs__gw__-based approach .
here we adopt a novel implementation merging qs__gw _ _ with dmft .
we adapted haule s continuous time quantum monte carlo solver with the projection and embedding schemes described in ref .
@xcite , and which are outlined in the supplemental material @xcite . the fully consistent qs__gw__+dmft calculation is composed by alternately repeated dmft and qs__gw _ _ loops .
first the qs__gw _ _ green s function is converged at fixed density , then it is projected on the ni @xmath0-orbitals and finally , within the dmft loop , the local self - energy is obtained .
updating the total density with the locally corrected green s function and repeating the procedure leads to complete self - consistency .
this method fully takes into account the dynamics of the local spin - fluctuations included in the dmft diagrams .
results are reported in table [ tab : dmft_results ] and they confirm that dmft adds the correct local diagrams missing in the qs__gw _ _ theory .
moreover by carefully continuing the resulting green s function on the real - frequency axis , we find an exchange splitting of @xmath460.3 ev and a satellite at @xmath465 ev @xcite . in order to investigate the importance of the dynamics in the local spin - spin channels , we carry out a qs__gw__+dmft calculation by retaining only the static limit of the dmft loop ( we call it qs__gw__+``sldmft '' ) .
once the dmft loop converges , we take the zero frequency limit of the magnetic part of the dmft self - energy and add it to the spin - averaged qs__gw _ _ hamiltonian @xcite .
as it is clearly shown in table [ tab : dmft_results ] and figure [ fig : mvse ] , this static hamiltonian reproduces very accurately magnetic moment and details of the band structure .
this is a strong indication that for ni the dynamics of local spin fluctuations is not crucial
. this will be the case if the quasiparticle picture is a reasonable description of ni , even if qs__gw _ _ alone does not contain enough physics to yield an optimum quasiparticle approximation . to verify this further , we model spin fluctuations by carrying out the qs__gw _ _ self - consistent cycle in the presence of a magnetic field @xmath47 , and tuning @xmath48 to reduce @xmath49 ( see inset , figure [ fig : mvse ] )
our key finding is that when @xmath47 is tuned to make @xmath49 agree with experiment , @xmath50 does also , reproducing arpes spectra to high precision in the fl regime , as clearly reported in figure [ fig : mvse ] and table [ tab : dmft_results ] . both the qs__gw _ _ and lsda overestimate @xmath49 for itinerant systems , but the latter also _ underestimates _ it in local - moment systems ( fig .
[ fig : mmom](@xmath29 ) ) . in the lsda treatment of ni ,
these effects cancel and render the moment fortuitously good .
when spin fluctuations are folded in through @xmath47 , the lsda moment becomes too small .
this finding must be interpreted as a sign of the superior level of internal consistency in the qs__gw _ _ theory with respect to lsda .
without such a degree of consistency spin fluctuations could not be approximated by a static field .
we have performed detailed qs__gw _ _ calculations of the electronic band structure of several 3@xmath0 metallic compounds to assess the reliability of this theory in the fermi liquid regime and the importance of the non - local terms in the self - energy . _ * fe : * _ through de haas - van alphen and cyclotron measurements we established that qs__gw _ _ qp levels at @xmath39 have an error of @xmath460.05 ev , and effective masses are well described .
comparable precision is found below @xmath39 by comparing to arpes data , provided final state scattering is taken into account .
the qs__gw _ _ @xmath0 bandwidth falls in close agreement with arpes , and is approximately @xmath51 times that of the lda ( fig .
[ fig : qsgwbands ] ) .
if @xmath10 is @xmath52-averaged to simulate a local self - energy , the qs__gw _ _ band structure changes significantly and resembles the lda .
thus non - locality in the self - energy is important in transition metals , and its absence explains why lda+dmft does not yield good agreement with arpes @xcite .
_ * ni : * _ qs__gw _ _ @xmath0 bandwidths , the @xmath53 splitting , the @xmath54 alignment , are all in excellent agreement with experiment , while @xmath55 and @xmath50 are too large .
however through the addition of a uniform static external field qs__gw _ _ can give both in good agreement , indicating a high level of consistency in the theory , contrary to lsda in which is not possible to have both quantities correct at the same time . to account for spin fluctuations in an _ ab initio _ framework
, we constructed a novel qs__gw__+dmft implementation and we utilised it at different degrees of approximations demonstrating that in itinerant magnets as ni ( i ) the dynamics of fluctuations is irrelevant ( ii ) their effect can be very well approximated by a static field as long as the non - local correlation part is treated accurately .
this work was supported by the simons many - electron collaboration , and epsrc grant ep / m011631/1 .
the authors gratefully acknowledge computer resources from the gauss centre for supercomputing e.v .
( www.gauss-centre.eu ) .
we also acknowledge the partnership for advanced computing in europe ( prace ) for awarding us access to the following resources : curie fn and tn based in france at the trs grand centre de calcul ( tgcc ) , and supermuc , based in germany at leibniz supercomputing centre . |
energy markets have been liberalised worldwide in the last two decades .
since then we have witnessed the increasing importance of such commodity markets which organise the trade and supply of energy such as electricity , oil , gas and coal .
closely related markets include also temperature and carbon markets .
there is no doubt that such markets will play a vital role in the future given that the global demand for energy is constantly increasing .
the main products traded on energy markets are spot prices , futures and forward contracts and options written on them .
recently , there has been an increasing research interest in the question of how such energy prices can be modelled mathematically . in this paper , we will focus on modelling energy _
prices , which include day - ahead as well as real - time prices .
traditional spot price models typically allow for _ mean - reversion _ to reflect the fact that spot prices are determined as equilibrium prices between supply and demand .
in particular , they are commonly based on a gaussian ornstein uhlenbeck ( ou ) process , see schwartz @xcite , or more generally , on weighted sums of ou processes with different levels of mean - reversion , see , for example , benth , kallsen and meyer - brandis @xcite and klppelberg , meyer - brandis and schmidt @xcite . in such a modelling framework , the mean - reversion is modelled directly or physically , by claiming that the price change is ( negatively ) proportional to the current price . in this paper
, we interpret the mean - reversion often found in commodity markets in a _
weak sense _ meaning that prices typically concentrate around a mean - level for demand and supply reasons . in order to account for such a weak form mean - reversion , we suggest to use a modelling framework which allows to model spot prices ( after seasonal adjustment ) directly in _
stationarity_. this paper proposes to use the class of volatility modulated lvy - driven volterra ( @xmath0 ) processes as the building block for energy spot price models . in particular , the subclass of so - called lvy semistationary ( @xmath1 ) processes turns out to be of high practical relevance .
our main innovation lies in the fact that we propose a modelling framework for energy spot prices which ( 1 ) allows to model deseasonalised energy spot prices directly in _ stationarity _ , ( 2 ) comprises _ stochastic volatility _ , ( 3 ) accounts for the possibility of _ jumps _ and _ spikes _ , ( 4 ) features great flexibility in terms of modelling the _ autocorrelation structure _ of spot prices and of describing the so - called _ samuelson effect _ , which refers to the finding that the volatility of a forward contract typically increases towards maturity .
we show that the new class of @xmath0 processes is analytically tractable , and we will give a detailed account of the theoretical properties of such processes .
furthermore , we derive explicit expressions for the forward prices implied by our new spot price model .
in addition , we will see that our new modelling framework encompasses many classical models such as those based on the schwartz one - factor mean - reversion model , see schwartz @xcite , and the wider class of continuous - time autoregressive moving - average ( carma ) processes . in that sense , it can also be regarded as a unifying modelling approach for the most commonly used models for energy spot prices .
however , the class of @xmath0 processes is much wider and directly allows to model the key special features of energy spot prices and , in particular , the stochastic volatility component .
the remaining part of the paper is structured as follows .
we start by introducing the class of @xmath0 processes in section [ sectionlss ] .
next , we formulate both a geometric and an arithmetic spot price model class in section [ sectionmodel ] and describe how our new models embed many of the traditional models used in the recent literature . in section [ sectionforward ] , we derive the forward price dynamics of the models and consider questions like affinity of the forward price with respect to the underlying spot .
section [ emp ] contains an empirical example , where we study electricity spot prices from the european energy exchange ( eex ) .
finally , section [ sectionconclusion ] concludes , and the contains the proofs of the main results .
throughout this paper , we suppose that we have given a probability space @xmath2 with a filtration @xmath3 satisfying the ` usual conditions , ' see karatzas and shreve @xcite , definition i.2.25 .
let @xmath4 denote a cdlg lvy process with lvy
khinchine representation @xmath5 for @xmath6 , @xmath7 and @xmath8 for @xmath9 , @xmath10 and the lvy measure @xmath11 satisfying @xmath12 and @xmath13 . we denote the corresponding characteristic triplet by @xmath14 . in a next step ,
we extend the definition of the lvy process to a process defined on the entire real line , by taking an independent copy of @xmath15 , which we denote by @xmath16 and we define @xmath17 for @xmath18 . throughout the paper
@xmath19 denotes such a two - sided lvy process .
the class of volatility modulated lvy - driven volterra ( @xmath0 ) processes , introduced by barndorff - nielsen and schmiegel @xcite , has the form @xmath20 where @xmath21 is a constant , @xmath22 is the two - sided lvy process defined above , @xmath23 are measurable deterministic functions with @xmath24 for @xmath25 , and @xmath26 and @xmath27 are cdlg stochastic processes which are ( throughout the paper ) assumed to be _ independent _ of @xmath22 . in addition , we assume that @xmath28 is positive .
note that such a process generalises the class of convoluted subordinators defined in bender and marquardt @xcite to allow for stochastic volatility .
a very important subclass of @xmath0 processes is the new class of lvy semistationary ( @xmath1 ) processes : we choose two functions @xmath29 such that @xmath30 and @xmath31 with @xmath32 whenever @xmath33 , then an @xmath1 process @xmath34 is given by @xmath35 note that the name lvy semistationary processes has been derived from the fact that the process @xmath36 is stationary as soon as @xmath28 and @xmath37 are stationary . in the case that @xmath38 is a two - sided brownian motion , we call such processes brownian semistationary ( @xmath39 ) processes , which have recently been introduced by barndorff - nielsen and schmiegel @xcite in the context of modelling turbulence in physics .
the class of @xmath1 processes can be considered as the natural analogue for ( semi- ) stationary processes of lvy semimartingales ( @xmath40 ) , given by @xmath41 the class of @xmath0 processes can be embedded into the class of ambit fields , see barndorf - nilsen and schmiegel @xcite , barndorff - nielsen , benth and veraart @xcite . also , it is possible to define @xmath0 and @xmath1 processes for _ singular _ kernel functions @xmath42 and @xmath43 , respectively ; a function @xmath42 ( or @xmath43 ) defined as above is said to be singular if @xmath44 ( or @xmath45 ) does not exist or is not finite . in order to simplify the exposition ,
we will focus on the stochastic integral in the definition of an @xmath0 ( and of an @xmath46 ) process only .
that is , throughout the rest of the paper , let @xmath47 in this paper , we use the stochastic integration concept described in basse - oconnor , graversen and pedersen @xcite where a stochastic integration theory on @xmath48 , rather than on compact intervals as in the classical framework , is presented . throughout the paper , we assume that the filtration @xmath49 is such that @xmath22 is a lvy process with respect to @xmath49 , see basse - oconnor , graversen and pedersen @xcite , section 4 , for details .
let @xmath14 denote the lvy triplet of @xmath22 associated with a truncation function @xmath50 .
according to basse - oconnor , graversen and pedersen @xcite , corollary 4.1 , for @xmath51 the process @xmath52 with @xmath53 is integrable with respect to @xmath22 if and only if @xmath52 is @xmath49-predictable and the following conditions hold almost surely : @xmath54 when we plug in @xmath55 , we immediately obtain the corresponding integrability conditions for the @xmath1 process .
[ ex1 ] in the case of a gaussian ornstein uhlenbeck process , that is , when@xmath56 for @xmath57 and @xmath58 , then the integrability conditions above are clearly satisfied , since we have @xmath59 for many financial applications , it is natural to restrict the attention to models where the variance is finite , and we focus therefore on lvy processes @xmath22 with finite second moment . note that the integrability conditions above do not ensure square - integrability of @xmath60 even if @xmath22 has finite second moment .
but substitute the first condition in ( [ mart - int - cond ] ) with the stronger condition @xmath61 \,\mathrm{d}s<\infty,\ ] ] then @xmath62 is square integrable . clearly , @xmath63 $ ] is constant in case of stationarity . for the lebesgue integral
part , we need @xmath64<\infty.\ ] ] according to the cauchy
schwarz inequality , we find @xmath65\leq\int _ { -\infty}^t\bigl|g(t , s)\bigr|^{2a } \,\mathrm{d}s\int _ { -\infty } ^t\bigl|g(t , s)\bigr|^{2(1-a ) } \mathbb{e}\bigl [ \omega_s^2\bigr ] \,\mathrm{d}s\ ] ] for any constant @xmath66 .
thus , a sufficient condition for ( [ leb - int - l2-cond ] ) to hold is that there exists an @xmath66 such that @xmath67 \,\mathrm{d}s < \infty,\end{aligned}\ ] ] which simplifies to @xmath68 \,\mathrm{d}s<\infty,\end{aligned}\ ] ] in the @xmath1 case . given a model for @xmath28 and @xmath43 , these conditions are simple to verify .
let us consider an example .
in example [ ex1 ] , we showed that for the kernel function @xmath69 and in the case of constant volatility , the conditions ( [ mart - int - cond ] ) are satisfied . next , suppose that there is stochastic volatility , which is defined by the barndorff - nielsen and shephard @xcite stochastic volatility model , that is @xmath70 , for @xmath71 , @xmath72 and a subordinator @xmath73 .
suppose now that @xmath73 has cumulant function @xmath74 for a lvy measure @xmath75 supported on the positive real axis , and that @xmath76 has finite expectation . in this case
, we have that @xmath77 = \int_0^{\infty}z \ell_u(\mathrm{d}z)<\infty$ ] for all @xmath78 .
thus , both ( [ mart - int - l2-cond ] ) and ( [ leb - int - l2-cond ] ) are satisfied ( the latter can be seen after using the sufficient conditions ) , and we find that @xmath79 is a square - integrable stochastic process .
this section presents the new modelling framework for energy spot prices , which is based on @xmath0 processes . as before , for ease of exposition , we will disregard the drift part in the general @xmath0 process for most of our analysis and rather use @xmath80 with @xmath81 as the building block for energy spot price , see ( [ eqsimpleambit ] ) for the precise definition of all components . throughout the paper ,
we assume that the corresponding integrability conditions hold
. we can use the @xmath0 process defined in ( [ defyspot ] ) as the building block to define both a geometric and an arithmetic model for the energy spot price .
also , we need to account for trends and seasonal effects .
let @xmath82 denote a bounded and measurable deterministic seasonality and trend function . in a _
geometric _ set up
, we define the spot price @xmath83 by @xmath84 in such a modelling framework , the deseasonalised , logarithmic spot price is given by a @xmath0 process .
alternatively , one can construct a spot price model which is of _ arithmetic _ type .
in particular , we define the electricity spot price @xmath85 by @xmath86 ( note that the seasonal function @xmath87 in the geometric and the arithmetic model is typically not the same . ) for general asset price models , one usually formulates conditions which ensure that prices can only take positive values .
we can easily ensure positivity of our arithmetic model by imposing that @xmath22 is a lvy subordinator and that the kernel function @xmath42 takes only positive values .
we have formulated the new spot price model in the general form based on a @xmath0 process to be able to account for non - stationary effects , see , for example , burger _
et al . _
@xcite , burger , graeber and schindlmayr @xcite .
if the empirical data analysis , however , supports the assumption of working under stationarity , then we will restrict ourselves to the analysis of @xmath1 processes with stationary stochastic volatility . as mentioned in the , traditional models for energy spot prices
are typically based on mean - reverting stochastic processes , see , for example , schwartz @xcite , since such a modelling framework reflects the fact that commodity spot prices are equilibrium prices determined by supply and demand .
stationarity can be regarded as a weak form of mean - reversion and is often found in empirical studies on energy spot prices ; one such example will be presented in this paper . in order to be able to have a stationary model , the lower integration bound in the definition of the @xmath0 process , and in particular for the @xmath1 process , is chosen to be @xmath88 rather than 0 .
clearly , in any real application , we observe data from a starting value onwards , which is traditionally chosen as the observation at time @xmath89 .
hence , while @xmath0 processes are defined on the entire real line , we only define the spot price for @xmath90 .
the observed initial value of the spot price at time @xmath89 is assumed to be a _ realisation _ of the random variable @xmath91 and @xmath92 , respectively .
such a choice guarantees that the deseasonalised spot price is a stationary process , provided we are in the stationary @xmath1 framework . since @xmath0 and @xmath1 processes are driven by a general lvy process
@xmath22 , it is possible to account for price jumps and spikes , which are often observed in electricity markets . at the same time , one can also allow for brownian motion - driven models , which are very common in , for example , temperature markets , see , for example , benth , hrdle and cabrera @xcite .
a key ingredient of our new modelling framework which sets the model apart from many traditional models is the fact that it allows for stochastic volatility .
volatility clusters are often found in energy prices , see , for example , hikspoors and jaimungal @xcite , trolle and schwartz @xcite , benth @xcite , benth and vos @xcite , koopman , ooms and carnero @xcite , veraart and veraart @xcite .
therefore , it is important to have a stochastic volatility component , given by @xmath28 , in the model .
note that a very general model for the volatility process would be to choose an @xmath0 process , that is , @xmath93 and @xmath94 where @xmath95 denotes a deterministic , positive function and @xmath73 is a lvy subordinator . in fact , if we want to ensure that the volatility @xmath96 is stationary , we can work with a function of the form @xmath97 , for a deterministic , positive function @xmath98 . the kernel function @xmath42 ( or @xmath43 ) plays a vital role in our model and introduces a flexibility which many traditional models lack : we will see in section [ secevarcov ] that the kernel function together with the autocorrelation function of the stochastic volatility process determines the autocorrelation function of the process @xmath99 .
hence our @xmath0 based models are able to produce various types of autocorrelation functions depending on the choice of the kernel function @xmath42 .
it is important to stress here that this can be achieved by using _
one _ @xmath0 process only , whereas some traditional models need to introduce a multi - factor structure to obtain a comparable modelling flexibility . also due to the flexibility in the choice of the kernel function
, we can achieve greater flexibility in modelling the shape of the samuelson effect often observed in forward prices , including the hyperbolic one suggested by bjerksund , rasmussen and stensland @xcite as a reasonable volatility feature in power markets .
note that we obtain the modelling flexibility in terms of the general kernel function @xmath42 here since we specify our model directly through a stochastic integral whereas most of the traditional models are specified through evolutionary equations , which limit the choices of kernel functions associated with solutions to such equations . in that context , we note that a @xmath0 or an @xmath1 process can not in general be written in form of a stochastic differential equation ( due to the non - semimartingale character of the process ) . in section [ sectsmcond ] , we will discuss sufficient conditions which ensure that an @xmath1 process is a semimartingale .
as already mentioned above , energy spot prices are typically modelled in stationarity , hence the class of @xmath1 processes is particularly relevant for applications . in the following
, we will show that many of the traditional spot price models can be embedded into our @xmath1 process - based framework .
our new framework nests the stationary version of the classical one - factor schwartz @xcite model studied for oil prices . by letting @xmath22
be a lvy process with the pure - jump part given as a compound poisson process , cartea and figueroa @xcite successfully fitted the schwartz model to electricity spot prices in the uk market .
benth and altyt .
e benth @xcite used a normal inverse gaussian lvy process @xmath22 to model uk spot gas and brent crude oil spot prices .
another example which is nested by the class of @xmath1 processes is a model studied in benth @xcite in the context of gas markets , where the deseasonalised logarithmic spot price dynamics is assumed to follow a one - factor schwartz process with stochastic volatility .
a more general class of models which is nested is the class of so - called carma - processes , which has been successfully used in temperature modelling and weather derivatives pricing , see benth , altyt .
e benth and koekebakker @xcite , benth , hrdle and lpez cabrera @xcite and hrdle and lpez cabrera @xcite , and more recently for electricity prices by garca , klppelberg and mller @xcite , benth _ et al .
_ @xcite .
a carma process is the continuous - time analogue of an arma time series , see brockwell @xcite , brockwell @xcite for definition and details .
more precisely , suppose that for nonnegative integers @xmath100 @xmath101 where @xmath102 and @xmath103 is a @xmath104-dimensional ou process of the form @xmath105 with @xmath106 .\ ] ] here we use the notation @xmath107 for the @xmath108-identity matrix , @xmath109 the @xmath104th coordinate vector ( where the first @xmath110 entries are zero and the @xmath104th entry is 1 ) and @xmath111 $ ] is the transpose of @xmath112 , with @xmath113 and @xmath114 for @xmath115 . in brockwell
@xcite , it is shown that if all the eigenvalues of @xmath116 have negative real parts , then @xmath117 defined as @xmath118 is the ( strictly ) stationary solution of ( [ multi - ou - carma ] ) .
moreover , @xmath119 is a @xmath120 process .
hence , specifying @xmath121 in ( [ carma ] ) , the log - spot price dynamics will be an @xmath1 process , but without stochastic volatility .
garca , klppelberg and mller @xcite argue for @xmath122 dynamics as an appropriate class of models for the deseasonalised log - spot price at the singapore new electricity market .
the innovation process @xmath22 is chosen to be in the class of stable processes . from benth ,
e benth and koekebakker @xcite , brownian motion - driven @xmath123 models seem appropriate for modelling daily average temperatures , and are applied for temperature derivatives pricing , including forward price dynamics of various contracts .
more recently , the dynamics of wind speeds have been modelled by a brownian motion - driven @xmath124 model , and applied to wind derivatives pricing , see benth and altyt .
e benth @xcite for more details . finally note that the arithmetic model based on a superposition of @xmath1 processes nests the non - gaussian ornstein uhlenbeck model which has recently been proposed for modelling electricity spot prices , see benth , kallsen and meyer - brandis @xcite .
we emphasis again that , beyond the fact that @xmath1 processes can be regarded as a unifying modelling approach which nest many of the existing spot price models , they also open up for entirely _ new _ model specifications , including more general choices of the kernel function ( resulting in non - linear models ) and the presence of stochastic volatility .
next , we study the second order structure of volatility modulated volterra processes @xmath125 , where @xmath126 , assuming the integrability conditions ( [ mart - int - cond ] ) hold and that in addition @xmath99 is square integrable .
let @xmath127 and @xmath128 .
recall that throughout the paper we assume that the stochastic volatility @xmath28 is independent of the driving lvy process . note that proofs of the following results are easy and hence omitted .
the conditional second order structure of @xmath99 is given by @xmath129 \operatorname{cov}\bigl((\overline y_{t+h } , \overline y_t)|\omega\bigr)&= & \kappa_2 \int_{-\infty}^{t } g(t+h , s ) g(t , s ) \omega_s^2\,\mathrm{d}s \qquad\mbox{for } t \in\mathbb{r } , h \geq0.\end{aligned}\ ] ] the conditional second order structure of @xmath36 is given by @xmath130 \operatorname{cov}\bigl((y_{t+h } , y_t)|\omega\bigr)&= & \kappa_2 \int_0^{\infty } g(x+h)g(x)\omega_{t - x}^2 \,\mathrm{d}x \qquad\mbox{for } t \in\mathbb{r } , h \geq0.\end{aligned}\ ] ] the unconditional second order structure of @xmath99 is then given as follows .
the second order structure of @xmath99 for stationary @xmath28 is given by @xmath131 \operatorname{var}(\overline y_t ) & = & \kappa_2 \mathbb{e } \bigl ( \omega_0 ^ 2 \bigr ) \int_{-\infty}^t g(t , s)^2 \,\mathrm{d}s + \kappa_1 ^ 2 \int _ { -\infty}^t \int_{-\infty}^t g(t , s)g(t , u ) \gamma \bigl(|s - u|\bigr ) \,\mathrm{d}s \,\mathrm{d}u , \\[-2pt ] \operatorname{cov}(\overline y_{t+h } , \overline y_t ) & = & \kappa_2 \mathbb{e } \bigl(\omega_0 ^ 2 \bigr ) \int_{-\infty}^{t } g(t+h , s)g(t , s ) \,\mathrm{d}s \\[-2pt ] & & { } + \kappa_1 ^ 2 \int_{-\infty}^{t+h } \int_{-\infty}^{t } g(t+h , s)g(t , u ) \gamma\bigl(|s - u|\bigr ) \
, \mathrm{d}s \,\mathrm{d}u,\end{aligned}\ ] ] where @xmath132 denotes the autocovariance function of @xmath28 , for @xmath133 . the unconditional second order structure of @xmath36
is then given as follows .
the second order structure of @xmath36 for stationary @xmath28 is given by @xmath134 \operatorname{var}(y_t ) & = & \kappa_2 \mathbb{e } \bigl ( \omega_0 ^ 2 \bigr ) \int_0^{\infty } g(x)^2 \,\mathrm{d}x + \kappa_1 ^ 2 \int _ 0^{\infty } \int_0^{\infty } g(x)g(y ) \gamma \bigl(|x - y|\bigr ) \,\mathrm{d}x \,\mathrm{d}y , \\[-2pt ] \operatorname{cov}(y_{t+h } , y_t ) & = & \kappa_2 \mathbb{e } \bigl ( \omega_0 ^ 2 \bigr ) \int_{0}^{\infty } g(x+h)g(x ) \,\mathrm{d}x + \kappa_1 ^ 2 \int _ { 0}^{\infty } \int_0^{\infty } g(x+h)g(y ) \gamma\bigl(|x - y|\bigr ) \,\mathrm{d}x \,\mathrm{d}y,\end{aligned}\ ] ] where @xmath135 denotes the autocovariance function of @xmath28 , for @xmath133 .
hence , we have @xmath136 \\[-8pt ] \nonumber & & \quad=\frac{\kappa_2 \mathbb{e } ( \omega_0 ^ 2 ) \int_0^{\infty } g(x+h)g(x ) \,\mathrm{d}x + \kappa_1 ^ 2 \int_0^{\infty } \int_0^{\infty } g(x+h)g(y ) \gamma(|x - y|)\,\mathrm{d}x \,\mathrm{d}y}{\kappa_2 \mathbb{e } ( \omega_0 ^ 2 ) \int_0^{\infty } g(x)^2 \,\mathrm{d}x + \kappa_1 ^ 2 \int_0^{\infty } \int_0^{\infty}g(x)g(y ) \gamma(|x
- y|)\,\mathrm{d}x \,\mathrm{d}y}.\end{aligned}\ ] ] if @xmath137 or if @xmath28 has zero autocorrelation , then @xmath138 the last corollary shows that we get the same autocorrelation function as in the @xmath139 model . from the results above , we clearly see the influence of the general damping function @xmath43 on the correlation structure . a particular choice of @xmath43 , which is interesting in the energy context is studied in the next example .
[ ex - bjerksund - g ] consider the case @xmath140 for @xmath141 and @xmath58 , which is motivated from the forward model of bjerksund , rasmussen and stensland @xcite , which we shall return to in section [ sectionforward ] .
we have that @xmath142 this ensures integrability of @xmath143 over @xmath144 with respect to any square integrable martingale lvy process @xmath22 .
furthermore , @xmath145 .
thus , @xmath146 observe that since @xmath43 can be written as @xmath147 it follows that the process @xmath148 is a semimartingale according to the knight condition , see knight @xcite and also basse @xcite , basse and pedersen @xcite , basse - oconnor , graversen and pedersen @xcite .
we pointed out that the subclass of @xmath1 processes are particularly relevant for modelling energy spot prices since they allow one to model directly in stationarity .
let us focus on this class in more detail .
clearly , an @xmath1 process is in general not a semimartingale .
however , we can formulate sufficient conditions on the kernel function and on the stochastic volatility component which ensure the semimartingale property .
the sufficient conditions are in line with the conditions formulated for @xmath39 processes in barndorff - nielsen and schmiegel @xcite , see also barndorff - nielsen and basse - oconnor @xcite .
note that the proofs of the following results are provided in the .
[ thmsm ] let @xmath36 be an @xmath1 process as defined in ( [ lss ] ) .
suppose the following conditions hold : @xmath149 .
the function values @xmath45 and @xmath150 exist and are finite .
the kernel function @xmath43 is absolutely continuous with square integrable derivative @xmath151 .
the process @xmath152 is square integrable for each @xmath153 .
the process @xmath154 is integrable for each @xmath153 .
then @xmath155 is a semimartingale with representation @xmath156 where @xmath157 for @xmath158 and @xmath159 an example of a kernel function which satisfies the above conditions is given by @xmath160 for @xmath161 , @xmath36 is given by a volatility modulated ornstein uhlenbeck process . in a next step ,
we are now able to find a representation for the quadratic variation of an @xmath1 process provided the conditions of proposition [ thmsm ] are satisfied .
[ thmqv ] let @xmath36 be an @xmath1 process and suppose that the sufficient conditions for @xmath36 to be a semimartingale ( as formulated in proposition [ thmsm ] ) hold .
then , the quadratic variation of @xmath36 is given by @xmath162_t = g(0+)^2 \int_0^t \omega_{s-}^2\,\mathrm{d}[l]_s\qquad \mbox{for } t \geq0.\ ] ] note that the quadratic variation is a prominent measure of accumulated stochastic volatility or intermittency over a certain period of time and , hence , is a key object of interest in many areas of application and , in particular , in finance .
the question of deriving semimartingale conditions for @xmath1 processes is closely linked to the question whether a spot price model based on an @xmath1 process is prone to arbitrage opportunities . in classical financial theory , we usually stick to the semimartingale framework to ensure the absence of arbitrage
. nevertheless one might ask the question whether one could still work with the wider class of @xmath1 processes which are not semimartingales . here
we note that the standard semimartingale assumption in mathematical finance is only valid for _ tradeable _ assets in the sense of assets which can be held in a portfolio .
hence , when dealing with , for example , electricity spot prices , this assumption is not valid since electricity is essentially non - storable .
hence , such a spot price can not be part of any financial portfolio and , therefore , the requirement of being a martingale under some equivalent measure @xmath163 is not necessary .
guasoni , rsonyi and schachermayer @xcite have pointed out that , while in frictionless markets martingale measures play a key role , this is not the case any more in the presence of market imperfections .
in fact , in markets with transaction costs , _ consistent price systems _ as introduced in schachermayer @xcite are essential . in such a set - up ,
even processes which are not semimartingales can ensure that we have _ no free lunch with vanishing risk _ in the sense of delbaen and schachermayer @xcite .
it turns out that if a continuous price process has _
conditional full support _
, then it admits consistent price systems for arbitrarily small transaction costs , see guasoni , r ' asonyi and schachermayer @xcite .
it has recently been shown by pakkanen @xcite , that under certain conditions , a @xmath39 process has conditional full support .
this means that such processes can be used in financial applications without necessarily giving rise to arbitrage opportunities .
let us briefly point out some model extensions concerning a multi - factor structure , non - stationary effects , multivariate models and alternative methods for incorporating stochastic volatility .
a straightforward extension of our model is to study a superposition of @xmath1 processes for the spot price dynamics .
that is , we could replace the process @xmath36 by a superposition of @xmath164 factors : @xmath165 and where all @xmath166 are defined as in ( [ defyspot ] ) for independent lvy processes @xmath167 and independent stochastic volatility processes @xmath168 , in both the geometric and the arithmetic model .
such models include the benth , kallsen and meyer - brandis @xcite model as a special case .
a superposition of factors @xmath169 opens up for separate modelling of spikes and other effects .
for instance , one could let the first factor account for the spikes , using a lvy process with big jumps at low frequency , while the function @xmath43 forces the jumps back at a high speed .
the next factor(s ) could model the `` normal '' variations of the market , where one observes a slower force of mean - reversion , and high frequent brownian - like noise , see veraart and veraart @xcite for extensions along these lines .
note that all the results we derive in this paper based on the one factor model can be easily generalised to accommodate for the multi - factor framework .
it should be noted that this type of `` superposition '' is quite different from the concept behind supou processes as studied in , for example , barndorff - nielsen and stelzer @xcite . in order to study various energy spot prices simultaneously
, one can consider extensions to a multivariate framework along the lines of barndorff - nielsen and stelzer @xcite , veraart and veraart @xcite .
in addition , another interesting aspect which we leave for future research is the question of alternative ways of introducing stochastic volatility in @xmath0 processes .
so far , we have introduced stochastic volatility by considering a stochastic proportional of the driving lvy process , that is , we work with a stochastic integral of @xmath28 with respect to @xmath22 .
an alternative model specification could be based on a stochastic time change @xmath170 , where @xmath171 .
such models can be constructed in a fashion similar to that of volatility modulated non - gaussian ornstein uhlenbeck processes introduced in barndorff - nielsen and veraart @xcite .
we know that outside the brownian or stable lvy framework , stochastic proportional and stochastic time change are not equivalent .
whereas in the first case the jump size is modulated by a volatility term , in the latter case the speed of the process is changed randomly .
these two concepts are in fact fundamentally different ( except for the special cases pointed out above ) and , hence , it will be worth investigating whether a combination of stochastic proportional and stochastic time change might be useful in certain applications .
in this subsection , we are concerned with the calculation of the forward price @xmath172 at time @xmath6 for contracts maturing at time @xmath173 .
we denote by @xmath174 a finite time horizon for the forward market , meaning that all contracts of interest mature _ before _ this date .
note that in energy markets , the corresponding commodity typically gets delivered over a delivery period rather than at a fixed point in time .
extensions to such a framework can be dealt with using standard methods , see , for example , benth , altyt benth and koekebakker @xcite for more details .
let @xmath175 denote the spot price , being either of geometric or arithmetic kind as defined in ( [ eqgambit - spot ] ) and ( [ eqaambit - spot ] ) , respectively , with @xmath176 where the stochastic volatility @xmath28 is chosen as previously defined in ( [ eqambitsv ] ) .
clearly , the corresponding results for @xmath1 processes can be obtained by choosing .
we use the conventional definition of a forward price in incomplete markets , see duffie @xcite , ensuring the martingale property of @xmath177 , @xmath178,\qquad 0 \leq t \leq t \leq t^*,\ ] ] with @xmath163 being an equivalent probability measure to @xmath179 . here , we suppose that @xmath180 , the space of integrable random variables . in a moment
, we shall introduce sufficient conditions for this . in finance
, one usually uses equivalent martingale measures @xmath163 , meaning that the equivalent probability measure @xmath163 should turn the discounted price dynamics of the underlying asset into a ( local ) @xmath163-martingale .
however , as we have already discussed , this restriction is not relevant in , for example , electricity markets since the spot is not tradeable .
thus , we may choose any equivalent probability @xmath163 as pricing measure . in practice
, however , one restricts to a parametric class of equivalent probability measures , and the standard choice seems to be given by the esscher transform , see benth , altyt benth and koekebakker @xcite , shiryaev @xcite .
the esscher transform naturally extends the girsanov transform to l ' evy processes . to this end , consider @xmath181 defined as the ( generalised ) esscher transform of @xmath22 for a parameter @xmath182 being a borel measurable function .
following shiryaev @xcite ( or benth , altyt benth and koekebakker @xcite , barndorff - nielsen and shiryaev @xcite ) , @xmath181 is defined via the radon nikodym density process @xmath183 for @xmath184 being a real - valued function which is integrable with respect to the lvy process on @xmath185 $ ] , and @xmath186 ( for @xmath187 ) being the log - moment generating function of @xmath188 , assuming that the moment generating function of @xmath188 exists .
a special choice is the ` constant ' measure change , that is , letting @xmath189 in this case , if under the measure @xmath179 , @xmath22 has characteristic triplet @xmath190 , where @xmath191 is the drift , @xmath192 is the squared volatility of the continuous martingale part and @xmath11 is the lvy measure in the l ' evy khinchine representation , see shiryaev @xcite , a fairly straightforward calculation shows that , see shiryaev @xcite again , the esscher transform preserves the lvy property of @xmath22 , and the characteristic triplet under the measure @xmath193 on the interval @xmath185 $ ] becomes @xmath194 , where @xmath195 this comes from the simple fact that the logarithmic moment generating function of @xmath22 under @xmath181 is @xmath196 it is important to note here that the choice of @xmath182 ( as , e.g. , in ( [ constantmc ] ) ) forces us to choose a _ starting time _ since the function @xmath197 will _ not _ be integrable with respect to @xmath22 on the unbounded interval @xmath144 .
recall that the only reason why we model from @xmath88 rather than from @xmath198 is the fact that we want to be able to obtain a stationary process under the probability measure @xmath179 . throughout this section
, we choose the starting time to be zero , which is a convenient choice since @xmath199 , and it is also practically reasonable since this can be considered as the time from which we start to observe the process . with such a choice , we do not introduce any risk premium for @xmath200 . in the general case , with a time - dependent parameter function @xmath201 , the characteristic triplet of @xmath22 under @xmath181 will become time - dependent , and hence the lvy process property is lost . instead , @xmath22 will be an independent increment process ( sometimes called an additive process ) .
note that if @xmath38 , a brownian motion , the esscher transform is simply a girsanov change of measure where @xmath202 for @xmath203 and a @xmath204-brownian motion @xmath205 .
similarly , we do a ( generalised ) esscher transform of @xmath73 , the subordinator driving the stochastic volatility model , see ( [ eqambitsv ] ) . we define @xmath206 to have the radon nikodym density process @xmath207 for @xmath208 being a real - valued function which is integrable with respect to @xmath73 on @xmath209 $ ] , and @xmath210 being the log - moment generating function of @xmath76 . since @xmath73 is a subordinator , we obtain @xmath211 where @xmath212 and @xmath75 denotes the lvy measure associated with @xmath73 .
our discussion above on choosing a starting value applies to the measure transform for the volatility process as well , and hence throughout the paper we will work under the assumption that @xmath213 for @xmath214 . note in particular , that this assumption implies that under the risk - neutral probability measure , the characteristic triplets of @xmath22 and @xmath73 only change on the time interval @xmath185 $ ] . on the interval @xmath215
, we have the same characteristic triplet for @xmath22 and @xmath73 as under @xmath179 .
choosing @xmath216 , with a constant @xmath217 , an esscher transform will give a characteristic triplet @xmath218 , which thus preserves the subordinator property of @xmath219 under @xmath206 .
for the general case , the process @xmath73 will be a time - inhomogeneous subordinator ( independent increment process with positive jumps ) .
the log - moment generating function of @xmath76 under the measure @xmath206 is denoted by @xmath220 . in order to ensure the existence of the ( generalised ) esscher transforms , we need some conditions .
we need that there exists a constant @xmath221 such that @xmath222 , and where @xmath223 .
( similarly , we must have such a condition for the lvy measure of the subordinator driving the stochastic volatility , that is , @xmath75 ) .
also , we must require that exponential moments of @xmath188 and @xmath76 exist .
more precisely , we suppose that parameter functions @xmath184 and @xmath224 of the ( generalised ) esscher transform are such that @xmath225 the exponential integrability conditions of the lvy measures of @xmath22 and @xmath73 imply the existence of exponential moments , and thus that the esscher transforms @xmath181 and @xmath206 are well defined .
we define the probability @xmath226 as the class of pricing measures for deriving forward prices . in this respect , @xmath182
may be referred to as the market price of risk , whereas @xmath227 is the market price of volatility risk .
we note that a choice @xmath228 will put more weight to the positive jumps in the price dynamics , and less on the negative , increasing the `` risk '' for big upward movements in the prices under @xmath229 .
let us denote by @xmath230 the expectation operator with respect to @xmath229 , and by @xmath231 the expectation with respect to @xmath206 .
suppose that the spot price is defined by the geometric model @xmath232 where @xmath99 is defined as in ( [ shortnot ] ) . in order to have the forward price @xmath172 well defined ,
we need to ensure that the spot price is integrable with respect to the chosen pricing measure @xmath229 .
we discuss this issue in more detail in the following .
we know that @xmath28 is positive and in general not bounded since it is defined via a subordinator .
thus , @xmath233 ( for @xmath234 ) is unbounded as well .
supposing that @xmath22 has exponential moments of all orders , we can calculate as follows using iterated expectations conditioning on the filtration @xmath235 generated by the paths of @xmath236 , for @xmath234 : @xmath237&=&\lambda(t){\mathbb{e}}_{\theta,\eta } \biggl [ { \mathbb{e}}_{\theta,\eta } \biggl[\exp \biggl(\int_{-\infty}^tg(t , s ) \omega_{s- } \,\mathrm{d}l_s \biggr ) \big| \mathcal{g}_t \biggr ] \biggr ] \\ & = & \lambda(t){\mathbb{e}}_{\eta } \biggl[\exp \biggl(\int_{-\infty}^0 \phi_l\bigl(g(t , s)\omega_s\bigr ) \,\mathrm{d}s \biggr ) \exp
\biggl(\int_{0}^t\phi^{\theta}_l \bigl(g(t , s)\omega_s\bigr ) \,\mathrm{d}s \biggr ) \biggr ] .\end{aligned}\ ] ] to have that @xmath238 , the two integrals must be finite .
this puts additional restrictions on the choice of @xmath239 and the specifications of @xmath240 and @xmath241 .
we note that when applying the esscher transform , we must require that @xmath22 has exponential moments of all orders , a rather strong restriction on the possible class of driving lvy processes . in our empirical study ,
however , we will later see that the empirically relevant cases are either that @xmath22 is a brownian motion or that @xmath22 is a generalised hyperbolic lvy process , which possess exponential moments of all orders .
we are now ready to price forwards under the esscher transform .
[ propforward - generalambit ] suppose that @xmath238 .
then , the forward price for @xmath242 is given by @xmath243 .\ ] ] as a special case , consider @xmath38 , where @xmath244 is a two - sided standard brownian motion under @xmath179 . in this case
we apply the girsanov transform rather than the generalised esscher transform , and it turns out that a rescaling of the transform parameter function @xmath182 by the volatility @xmath245 is convenient for pricing of forwards . to this end , consider the girsanov transform @xmath246 that is , we set @xmath247 for @xmath18 . supposing that the novikov condition @xmath248<\infty,\ ] ] holds , we know that @xmath249 is a brownian motion for @xmath250 under a probability @xmath251 having density process @xmath252 suppose that there exists a measurable function @xmath253 such that @xmath254 for all @xmath255 , with @xmath256 furthermore , suppose the moment generating function of @xmath257 exists on the interval @xmath258 .
then , for all @xmath182 such that @xmath259 , the novikov condition is satisfied , since by the subordinator property of @xmath260 ( restricting our attention to @xmath90 ) @xmath261 and therefore @xmath248\leq { \mathbb{e}}\biggl[\exp \biggl(\frac12 \int_{0}^{t^*}\frac{\theta^2(s)}{j(s ) } \,\mathrm{d}s \omega_0^{-2 } \biggr ) \biggr]<\infty.\ ] ] specifying @xmath262 , we have that @xmath263 , and condition ( [ suff - novikov ] ) holds with equality .
we discuss the integrability of @xmath264 with respect to @xmath265 . by double conditioning with respect to the filtration generated by the paths of @xmath245 ,
we find @xmath237&=&\lambda(t)\exp \biggl(\int _ { 0}^{t}g(t , s)\theta(s ) \,\mathrm{d}s \biggr ) { \mathbb{e}}_{\theta,\eta } \biggl[{\mathbb{e}}_{\theta,\eta } \biggl[\exp \biggl(\int _
{ -\infty}^tg(t , s)\omega_{s- } \
, \mathrm{d}w_s \biggr )
\big| \mathcal{g}_t \biggr ] \biggr ] \\ &
= & \lambda(t)\exp \biggl(\int_{0}^{t}g(t , s ) \theta(s ) \,\mathrm{d}s \biggr){\mathbb{e}}_{\eta } \biggl[\exp \biggl(\frac12\int _ { -\infty}^tg^2(t ,
s)\omega^2_s \,\mathrm{d}s \biggr ) \biggr ] .\end{aligned}\ ] ] from collecting the conditions on @xmath266 and @xmath239 for verifying all the steps above , we find that if @xmath267 is integrable on @xmath268 ( recall that @xmath269 for @xmath270 ) and @xmath271 is integrable on @xmath272 for all @xmath273 , then @xmath238 as long as @xmath274 we assume these conditions to hold .
we state the forward price for the case @xmath38 and the girsanov change of measure discussed above .
[ propforward - bmambit ] suppose that @xmath38 and that @xmath251 is defined by the girsanov transform in ( [ girsanovtransf ] ) .
then , for @xmath275 , @xmath276 let us consider an example . in the bns stochastic volatility model
, we have @xmath277 . hence , @xmath278
which yields , @xmath279 this implies from proposition [ propforward - bmambit ] that the forward price is affine in @xmath96 , the ( square of the ) stochastic volatility .
the stochastic volatility model studied in benth @xcite is recovered by choosing @xmath280 .
suppose for a moment that the stochastic volatility process @xmath245 is identical to one ( i.e. , that we do not have any stochastic volatility in the model ) . in this case
, the forward price becomes @xmath281 where @xmath282 for @xmath18 .
hence , the logarithmic forward ( log - forward ) price is @xmath283 with @xmath284 for @xmath285 . note that @xmath286 , for @xmath90 , is a @xmath179-martingale with the property ( for @xmath287 ) @xmath288 in the classical ornstein uhlenbeck case , with @xmath30 , @xmath69 for @xmath57 , we easily compute that @xmath289 and the forward price is explicitly dependent on the current spot price . in the general case ,
this does not hold true .
we have that @xmath290 , not unexpectedly , since the forward price converges to the spot at maturity ( at least theoretically ) .
however , apart from the special time point @xmath291 , the forward price will in general _ not _ be a function of the current spot , but a function of the process @xmath292 .
thus , at time @xmath293 , the forward price will depend on @xmath294 whereas the spot price depends on @xmath295 the two stochastic integrals can be pathwise interpreted ( they are both wiener integrals since the integrands are deterministic functions ) , and both @xmath79 and @xmath292 are generated by integrating over the same paths of a brownian motion .
however , the paths are scaled by two different functions @xmath296 and @xmath240 .
this allows for an additional degree of flexibility when creating forward curves compared to affine structures . in the classical ornstein uhlenbeck case ,
the forward curve as a function of time to maturity @xmath297 will simply be a discounting of today s spot price , discounted by the speed of mean reversion of the spot ( in addition comes deterministic scaling by the seasonality and market price of risk ) . to highlight the additional flexibility in our modelling framework of semistationary processes , suppose for the sake of illustration that @xmath298 .
then @xmath299 if furthermore @xmath300 , we are in a situation where the long end ( i.e. , @xmath301 large ) of the forward curve is not a constant . in fact
, we find for @xmath6 that @xmath302 since @xmath303 is random , we will have a randomly fluctuating long end of the forward curve .
this is very different from the situation with a classical mean - reverting spot dynamics , which implies a deterministic forward price in the long end ( dependent on the seasonality and market price of risk only ) .
various shapes of the forward curve @xmath304 can also be modelled via different specifications of @xmath42 .
for instance , if @xmath305 is a decreasing function , we obtain the contango and backwardation situations depending on the spot price being above or below the mean .
if @xmath306 has a hump , we will also observe a hump in the forward curve .
for general specifications of @xmath42 we can have a high degree of flexibility in matching desirable shapes of the forward curve .
observe that the time - dynamics of the forward price can be considered as correlated with the spot rather than directly depending on the spot . in the ornstein uhlenbeck
situation , the log - forward price can be considered as a linear regression on the current spot price , with time - dependent coefficients .
this is not the case for general specifications .
however , we have that @xmath292 and @xmath60 are both normally distributed random variables ( recall that we are still restricting our attention to @xmath38 ) , and the correlation between the two is @xmath307 obviously , for @xmath308 , the correlation is 1 . in conclusion
, we can obtain a weaker stochastic dependency between the spot and forward price than in the classical mean - reversion case by a different specification of the function @xmath42 . in the discussion above
, we saw that the choice @xmath298 yielded a forward price expressible in terms of @xmath79 . in the next proposition , we prove that this is the only choice of @xmath42 yielding an affine structure . the result is slightly generalising the analysis of carverhill @xcite .
[ profforwardaffine ] the forward price in proposition [ propforward - bmambit ] is affine in @xmath60 and @xmath309 if there exist functions @xmath310 and @xmath311 such that @xmath298 and @xmath312 .
conversely , if the forward price is affine in @xmath60 and @xmath309 , and @xmath42 and @xmath95 are strictly positive and continuously differentiable in the first argument , then there exists functions @xmath310 and @xmath311 such that @xmath298 and @xmath312 . obviously , the choice of @xmath42 and @xmath95 coming from ou - models , @xmath313 satisfy the conditions in the proposition above .
in fact , appealing to similar arguments as in the proof of proposition [ profforwardaffine ] above , one can show that this is the _ only _ choice ( modulo multiplication by a constant ) which is stationary and gives an affine structure in the spot and volatility for the forward price dynamics .
in particular , the specification @xmath314 considered in example [ ex - bjerksund - g ] gives a stationary spot price dynamics , but not an affine structure in the spot for the forward price .
next , we turn our attention to the risk - neutral dynamics of the forward price .
[ propriskneutraldynamicsf ] assume that the assumptions of proposition [ propforward - bmambit ] hold and that @xmath315 is given by the ( simple ) esscher transform .
then the risk - neutral dynamics of the forward price @xmath172 is given by @xmath316 where @xmath317 .
moreover @xmath318 is a @xmath319-martingale , where @xmath320 denotes the poisson random measure associated with @xmath73 , and @xmath321 is the lvy measure of @xmath73 under @xmath206 .
we observe that the dynamics will jump according to the changes in volatility given by the process @xmath260 .
as expected , the integrand in the jump expression tends to zero when @xmath322 , since the forward price must ( at least theoretically ) converge to the spot when time to maturity goes to zero .
the forward dynamics will have a stochastic volatility given by @xmath323 .
hence , whenever @xmath324 exists , and @xmath325 , we have @xmath326 , @xmath327 when passing to the limit , we have implicitly supposed that we work with the version of @xmath328 having left - continuous paths with right - limits . by the definition of our integral in @xmath329 , where the integrand is supposed predictable , this can be done .
thus , we find that the forward volatility converges to the spot volatility as time to maturity tends to zero , which is known as the samuelson effect .
contrary to the classical situation where this convergence goes exponentially , we may have many different shapes of the volatility term structure resulting from our general modelling framework . in bjerksund , rasmussen and stensland @xcite , a forward price dynamics for electricity contracts
is proposed to follow @xmath330 where @xmath331 and @xmath332 are positive constants .
they argue that in electricity markets , the samuelson effect is stronger close to maturity than what is observed in other commodity markets , and they suggest to capture this by letting it increase by the rate @xmath333 close to maturity of the contracts .
this is in contrast to the common choice of volatility being @xmath334 , resulting from using the schwartz model for the spot price dynamics .
there is no reference to any spot model in the bjerksund , rasmussen and stensland @xcite model .
the constant @xmath37 comes from a non - stationary behaviour , which can be incorporated in the @xmath0 framework .
however , here we focus on the stationary case and choose @xmath335 .
then we see that we can model the spot price by the @xmath39 process @xmath336 thus , after doing a girsanov transform , we recover the risk - neutral forward dynamics of bjerksund , rasmussen and stensland @xcite .
it is interesting to note that with this spot price dynamics , the forward dynamics is not affine in the spot .
hence , the bjerksund , rasmussen and stensland @xcite model is an example of a non - affine forward dynamics .
whenever @xmath337 , we do not have that @xmath338 , and thus the bjerksund , rasmussen and stensland @xcite model does not satisfy the samuelson effect , either .
we end this section with a discussion of option pricing .
let us assume that we have given an option with exercise time @xmath339 on a forward with maturity at time @xmath340 .
the option pays @xmath341 , and we are interested in finding the price at time @xmath342 , denoted @xmath343 . from arbitrage theory , it holds that @xmath344 , \ ] ] where @xmath163 is the risk - neutral probability . choosing @xmath345 as coming from the esscher
transform above , we can derive option prices explicitly in terms of the characteristic function of @xmath73 by fourier transformation .
[ optionprice ] let @xmath346 be the probability measure obtained from the esscher transform .
let @xmath347 , and suppose that @xmath348 . by applying the definitions of fourier transforms and their inverses in folland @xcite
, we have that @xmath349 , with @xmath350 is the fourier transform of @xmath351 defined by @xmath352 suppose that @xmath353
. then the option price is given by @xmath354 where @xmath355 and @xmath356 one can calculate option prices by applying the fast fourier transform as long @xmath357 is known .
if @xmath104 is not integrable ( as is the case for a call option ) , one may introduce a damping function to regularize it , see carr and madan @xcite for details .
let us consider the arithmetic spot price model , @xmath358 we analyse the forward price for this case , and discuss the affinity .
the results and discussions are reasonably parallel to the geometric case , and we refrain from going into details but focus on some main results . under a natural integrability condition of the spot price with respect to the esscher transform measure @xmath229 , we find the following forward price for the arithmetic model .
[ propforward - arithmetic ] suppose that @xmath238 .
then , the forward price is given as @xmath359 \int_t^tg(t , s)\mathbb{e}_{\eta } [ \omega_{s } | \mathcal { f}_t ] \,\mathrm{d}s \biggr\}.\ ] ] the price is reasonably explicit , except for the conditional expectation of the stochastic volatility @xmath236 . by the same arguments as in proposition [ profforwardaffine ] , the forward price becomes affine in the spot ( or in @xmath329 ) if and only if @xmath298 for sufficiently regular functions @xmath360 and @xmath361 .
in the case @xmath38 , we can obtain an explicit forward price when using the girsanov transform as in ( [ girsanovtransf ] ) .
we easily compute that the forward price becomes @xmath362 we note that there is no explicit dependence of the spot volatility @xmath236 except indirectly in the stochastic integral .
this is in contrast to the lvy case with esscher transform .
the dynamics of the forward price becomes @xmath363 if we furthermore let @xmath298 for some sufficiently regular functions @xmath360 and @xmath361 , we find that @xmath364 hence , the forward curve moves stochastically as the deseasonalised spot price , whereas the shape of the curve is deterministically given by @xmath365 .
this shape is scaled stochastically by the deseasonalised spot price .
in addition , there is a deterministic term which is derived from the market price of risk @xmath197 .
we finally remark that also in the arithmetic case one may derive expressions for the prices of options that are computable by fast fourier techniques .
in this section , we will show the practical relevance of our new model class for modelling empirical energy spot prices . here
we will focus on electricity spot prices and we will illustrate that they can be modelled by @xmath1 processes an important subclass of @xmath366 processes . note that the data analysis is exploratory in nature since the estimation theory for @xmath366 or @xmath1 processes has not been fully established yet .
we study electricity spot prices from the european energy exchange ( eex ) .
we work with the daily phelix peak load data ( i.e. , the daily averages of the hourly spot prices for electricity delivered during the 12 hours between 8 am and 8 pm ) with delivery days from 01.01.2002 to 21.10.2008 .
note that peak load data do not include weekends , and in total we have 1775 observations .
the daily data , their returns and the corresponding autocorrelation functions are depicted in figure [ eldata ] . before analysing the data
, we have deseasonalised the spot prices . here
, we have worked with a geometric model , that is , @xmath367 .
then @xmath368 where , as suggested in , for example , klppelberg , meyer - brandis and schmidt @xcite , @xmath369 which takes weakly and yearly effects and a linear trend into account .
in order to ensure that the spikes do not have a big impact on parameter estimation , we have worked with a robust estimation technique based on iterated reweighted least squares .
we have then subtracted the estimated seasonal function from the logarithmic spot prices from the time series and have worked with the deseasonalised data for the remaining part of the section .
figure [ gdeseas ] depicts the deseasonalised logarithmic prices and the corresponding returns .
the class of @xmath0 processes is very rich and hence in a first step we checked whether we can restrict it to a smaller class in our empirical work .
we have carried out unit root tests , more precisely the augmented dickey fuller test ( where the null hypothesis is that a unit root is present in the time series versus the alternative of a stationary time series ) ; we obtained a @xmath104-value which is smaller than 0.01 and , hence , clearly reject the unit root hypothesis at a high significance level . also the phillips
perron test led to the same conclusion .
hence , in the following , we assume that @xmath370 is an @xmath46 process . next , we study the question which distribution describes the stationary distribution of @xmath36 appropriately . we know that in the absence of stochastic volatility an @xmath1 process is a moving average process driven by a lvy process and hence the integral is itself infinite divisible
we are hence dealing with a stationary infinitely divisible stochastic process , see rajput and rosiski @xcite , sato @xcite , barndorff - nielsen @xcite for more details .
the literature on spot price modelling suggest to use semi - heavy and , in some cases , even heavy - tailed distributions in order to account for the extreme spikes in electricity spot prices , see , for example , klppelberg , meyer - brandis and schmidt @xcite and benth _
et al . _
@xcite who suggested to use the stable distribution for modelling electricity returns .
here we focus on a mixture of a normal distribution in the sense of mean - variance mixtures , see barndorff - nielsen , kent and srensen @xcite .
in particular , we will focus on the generalised hyperbolic ( gh ) distribution , see barndorff - nielsen and halgreen @xcite , barndorff - nielsen @xcite , barndorff - nielsen @xcite , which turns out to provide a good fit to the deseasonalised logarithmic spot prices as we will see in the following .
a detailed review of the generalised hyperbolic distribution can be found in , for example , mcneil , frey and embrechts @xcite and details on the corresponding implementation in r based on the ghyp package is provided in breymann and lthi @xcite .
let @xmath371 and let @xmath372 denote a @xmath373-dimensional random vector .
@xmath372 is said to have multivariate generalised hyperbolic ( gh ) distribution if @xmath374 where @xmath375 , @xmath376 , @xmath377 .
further , @xmath378 is a one - dimensional random variable , independent of @xmath379 and with generalised inverse gaussian ( gig ) distribution , that is , @xmath380 .
the density of the gig distribution with parameters @xmath381 is given by @xmath382 where @xmath383 denotes the modified bessel function of the third kind , and the parameters have to satisfy one of the following three restrictions @xmath384 typically , we refer to @xmath385 as the location parameter , to @xmath386 as the dispersion matrix and to @xmath387 as the symmetry parameter ( sometimes also called skewness parameter ) .
the parameters @xmath388 of the gig distribution determine the shape of the gh distribution .
the parametrisation described above is the so - called @xmath389-parametrisation of the gh distribution . however , for estimation purposes this parametrisation causes an identifiability problem and hence we worked with the so - called @xmath390-parametrisation in our empirical study .
note that the @xmath389-parametrisation can be obtained by from @xmath390-parametrisation by setting @xmath391 and @xmath392 remain the same , see breymann and lthi @xcite for more details . in our empirical study
, we work with the one - dimensional gh distribution .
that is , @xmath393 and @xmath394 and @xmath395 are scalars rather than a matrix and vectors , respectively .
we have fitted 11 distributions within the gh class to the deseasonalised log - spot prices using quasi - maximum likelihood estimation : the asymmetric and symmetric versions of the * generalised hyperbolic distribution ( ghyp ) : @xmath396 , ( @xmath397 ) , * normal inverse gaussian ( nig ) distribution : @xmath398 , ( @xmath399 ) , * student-@xmath293 distribution ( with @xmath400 degrees of freedom ) : @xmath401 , ( @xmath402 ) , * hyperbolic distribution ( hyp ) : @xmath403 , ( @xmath404 ) , * variance gamma distribution ( vg ) : @xmath405 , ( @xmath406 ) , and the gaussian distribution .
we have compared these distributions using the akaike information criterion , see table [ aicfit ] , which suggests that the symmetric nig distribution is the preferred choice for the stationary distribution of the deseasonalised logarithmic spot prices .
the diagnostic plots of the empirical and fitted logarithmic densities and the quantile
quantile plots of the fitted symmetric nig distribution are depicted in figure [ gsymnig ] .
we see that the fit is reasonable .
@lld2.3d1.3d2.3cd2.3cd4.2@ & symmetric & & & & & & & + nig&true&-0.5&0.431&-0.001&0.395&0&1313.14&-653.57 + ghyp&true&-0.183&0.438&-0.001&0.392&0&1314.13&-653.06 + nig&false&-0.5&0.431&-0.003&0.395&0.002&1315.10&-653.55 + ghyp&false&-0.184&0.438&-0.002&0.392&0.002&1316.10&-653.05 + student - t&true&-1.366&0&-0.001&0.458&0&1327.28&-660.64 + student - t&false&-1.365&0&-0.002&0.458&0.002&1329.26&-660.63 + hyp&true&1&0.150&0.000&0.375&0&1331.38&-662.69 + hyp&false&1&0.147&0.003&0.375&-0.003&1333.33&-662.66 + vg&true&0.975&0&0.003&0.379&0&1333.85&-663.92 + vg&false&0.970&0&0.007&0.379&-0.007&1335.42&-663.71 + gaussian&true&&&-0.000&0.395&0&1742.94&-869.47 + in our empirical study , we have seen that the symmetric normal inverse gaussian distribution fits the marginal distribution of the deseasonalised logarithmic electricity prices well .
hence , it is natural to ask whether there is a stationary @xmath39 or @xmath1 process with marginal normal inverse gaussian or , more generally , generalised hyperbolic distribution ?
the answer is yes , as we will show in the following .
note that the following investigation extends the study of barndorff - nielsen and shephard @xcite , where the background driving process of an ornstein uhlenbeck process was specified , given a marginal infinitely divisible distribution .
let us focus on a particular @xmath39 process given by @xmath407 for constants @xmath408 and for stationary @xmath28 and a standard brownian motion @xmath244 independent of @xmath28 .
note that we have introduced a drift term in the @xmath39 process again in order to derive the general theoretical result . for our empirical example
, however , it would be sufficient to set @xmath409 as suggested by our estimation results above .
the conditional law of @xmath410 given @xmath28 is normal:@xmath411 now suppose that @xmath412 follows an @xmath1 process given by@xmath413 where @xmath73 is a subordinator . then , by a stochastic fubini theorem we find@xmath414 where @xmath415 , the convolution of @xmath416 and @xmath98 .
similarly , @xmath417 with @xmath418 .
let @xmath419 denote the gamma density with parameters @xmath420 and @xmath421 , that is,@xmath422 now we define @xmath423 for @xmath424 , which ensures the existence of the integral ( [ specialbss ] ) ; then we have @xmath425 hence , if , for @xmath426 , @xmath427 and if , moreover,@xmath428 we obtain@xmath429 in other words,@xmath430 where@xmath431 we define the subordinator @xmath432 with lvy measure @xmath433 by @xmath434
. then @xmath435 then one can easily show that the marginal distribution of @xmath436 does not depend on @xmath437 , and the parameter @xmath437 determines the autocorrelation structure of @xmath436 .
it follows that if the subordinator @xmath432 is such that @xmath438 has the generalised inverse gaussian law @xmath439 then the law of @xmath410 is the generalised hyperbolic @xmath440 .
is there such a subordinator ?
the answer is yes . to see this ,
let @xmath441 and note that @xmath438 is infinitely divisible with kumulant function @xmath442 on the other hand , the subordinator @xmath432 ( here assumed to have no drift ) has kumulant function @xmath443 where @xmath433 is the lvy measure of @xmath432 .
combining we find@xmath444 that is , the lvy measure @xmath445 of @xmath438 is @xmath446 thus , the question is : does there exist a lvy measure @xmath433 on @xmath447 such that @xmath445 given by ( [ wv ] ) is the lvy measure of the @xmath448 law .
that , in fact , is the case since the @xmath449 laws are self - decomposable , cf .
halgreen @xcite and jurek and vervaat @xcite .
next , we focus on the autocorrelation structure implied by the choice of the kernel functions which lead to a marginal gh distribution of the @xmath39 process .
[ propcorgammakernel ] let @xmath36 be the @xmath39 process defined in the previous subsection with kernel function @xmath43 as defined in ( [ gfct ] ) .
in the case when @xmath450 and @xmath451 , we have @xmath452 where @xmath453 and @xmath454 denotes the modified bessel function of the third kind .
we have estimated the parameters @xmath455 and @xmath437 using a linear least squares estimate based on the empirical and the theoretical autocorrelation function using the first @xmath456 lags .
we obtain @xmath457 and @xmath458 .
figure [ gammaacf ] shows the empirical and the corresponding fitted autocorrelation function . and @xmath459 .
] note that the estimate @xmath458 implies that the corresponding @xmath39 process is not a semimartingale , see , for example , barndorff - nielsen and schmiegel @xcite for details . in the context of electricity prices , this does not need to be a concern since the electricity spot price is not tradeable .
we observe that the autocorrelation function induced by the gamma - kernel mimics the behaviour of the empirical autocorrelation function adequately .
however , it does not fit the first 10 lags as well as , for example , the carma - kernel which we have fitted in the following subsection , but performs noticeably better for higher lags
. the fit could be further improved by choosing @xmath460 to be a gig supou process rather than a gig ou process .
then one obtains an even more flexible autocorrelation structure .
the recent literature on modelling electricity spot prices has advocated the use of linear models , that is , carma models , as described in detail in section [ sectunifying ] .
since carma models are special cases of our general modelling framework , we briefly demonstrate their empirical performance as well .
it is well known , see , for example , brockwell , davis and yang @xcite , that a discretely sampled @xmath461 process ( for @xmath100 ) has a weak arma(@xmath104 , @xmath110 ) representation .
an automatic model selection using the akaike information criterion within the class of ( discrete - time arima ) models suggests that an arma(2,1 ) model is the best choice for our data .
we take that result as an indication that a @xmath122 process ( which has a weak arma(2,1 ) representation ) might be a good choice .
however , it should be noted that the relation between model selection in discrete and continuous time still needs to be explored in detail . we have estimated the parameters of the kernel function @xmath43 which corresponds to a @xmath122 process using quasi - maximum - likelihood estimation based on the weak arma(2,1 ) representation .
diagnostic plots for the estimated @xmath122 model are provided in figure [ gcarma ] .
first , we compare the empirical and the estimated autocorrelation function , see figure [ gcarma](a ) .
recall that the autocorrelation of @xmath36 is given by ( [ cor ] ) and it simplifies to @xmath462 if either the driving lvy process has zero mean or if the stochastic volatility process has zero autocorrelation .
after deseasonalising ( which also includes detrending ) the data , we have obtained data which have approximately zero mean . model . ]
the empirical and the estimated autocorrelation function implied by a @xmath122 kernel function @xmath43 match very well for the first 12 lags .
higher lags were however slightly better fitted by the gamma kernel used in the previous subsection .
figure [ gcarma](b ) depicts the corresponding residuals from the weak arma(2,1 ) representation and figures [ gcarma](c ) and [ gcarma](d ) show the autocorrelation functions of the corresponding residuals and squared residuals .
overall , we see that the fit provided by the @xmath122 kernel function is acceptable .
note that in addition to estimating the parameters of the @xmath43 function coming from a carma process one can also recover the driving lvy process of a carma process based on recent findings by brockwell , davis and yang @xcite .
this will make it possible to also address the question of whether stochastic volatility is needed to model electricity spot prices or not .
see veraart and veraart @xcite for empirical work along those lines in the context of electricity spot prices , whose results suggest that stochastic volatility is indeed important for modelling electricity spot prices .
this paper has focused on _ volatility modulated lvy - driven volterra _ ( @xmath0 ) processes as the building block for modelling energy spot prices . in particular , we have introduced the class of _ lvy semistationary _ ( @xmath1 ) processes as an important subclass of @xmath0 processes , which reflect the stylised facts of empirical energy spot prices well .
this modelling framework is built on four principles .
first , deseasonalised spot prices can be modelled directly in stationarity to reflect the empirical fact that spot prices are equilibrium prices determined by supply and demand and , hence , tend to mean - revert ( in a weak sense ) to a long - term mean .
second , stochastic volatility is regarded as a key factor for modelling ( energy ) spot prices .
third , our new modelling framework allows for the possibility of jumps and extreme spikes .
fourth , we have seen that @xmath366 and , in particular , @xmath46 processes feature great flexibility in terms of modelling the autocorrelation function and the samuelson effect .
we have demonstrated that @xmath0 processes are highly analytically tractable ; we have derived explicit formulae for the energy forward prices based on our new spot price models , and we have shown how the kernel function determines the samuelson effect in our model .
in addition , we have discussed option pricing based on transform - based methods .
an exploratory data analysis on electricity spot prices shows the potential our new approach has and more detailed empirical work is left for future research .
also , we plan to address the question of model estimation and inference . it will be important to study efficient estimation schemes for fully parametric specifications of @xmath0- and , in particular , @xmath1-based models .
proof of proposition [ thmsm ] in order to prove the semimartingale conditions suppose for the moment that @xmath36 is a semimartingale , so that the stochastic differential of @xmath36 exists .
then , calculating formally , we find @xmath463 \\[-8pt ] \nonumber & & { } + \int_{-\infty}^tq'(t - s)a_s \,\mathrm{d}s \,\mathrm{d}t,\end{aligned}\ ] ] which indicates that @xmath36 can be represented , for @xmath6 , as @xmath464 clearly , under the conditions formulated in proposition [ thmsm ] , the above integrals are well defined , and @xmath36 , defined by ( [ smrep ] ) , is a semimartingale , and @xmath465 exists and satisfies equation ( [ dy ] ) .
a direct rewrite now shows that ( [ dy ] ) agrees with the defining equation ( [ lss ] ) of @xmath36 , and we can then deduce that @xmath36 is a semimartingale .
proof of proposition [ thmqv ] the result follows directly from the representation ( [ smrep ] ) and from properties of the quadratic variation process , see , for example , protter @xcite .
proof of proposition [ propforward - generalambit ] first , write @xmath466 and observe that the first integral on the right - hand side is @xmath467-measurable
. the result follows by using double conditioning , first with respect to the @xmath468-algebra @xmath469 generated by the paths of @xmath470 and @xmath471 , and next with respect to @xmath471 .
proof of proposition [ propforward - bmambit ] by the girsanov change of measure , we have @xmath472 where we set @xmath473 for @xmath270 . by following the argumentation in the proof of proposition [ propforward - generalambit ] ,
we are led to calculate the expectation @xmath474.\ ] ] but , by the stochastic fubini theorem , see , for example , barndorff - nielsen and basse - oconnor @xcite , @xmath475 using the adaptedness to @xmath471 of the first integral and the independence from @xmath471 of the second , we find the desired result .
proof of proposition [ profforwardaffine ] if @xmath298 it holds that @xmath476 similarly , if @xmath312 , @xmath477 and affinity holds in both the volatility and the spot price .
opposite , to have affinity in @xmath60 we must have that @xmath478 for some function @xmath479 , which means that the ratio @xmath480 is independent of @xmath78 .
@xmath479 is differentiable in @xmath301 as long as @xmath42 is .
furthermore , @xmath481 by definition .
thus , by first differentiating @xmath482 with respect to @xmath301 and next letting @xmath483 , it holds that @xmath484 where we use the notation @xmath485 and @xmath486 for the corresponding partial derivatives with respect to the first argument .
hence , we must have that @xmath487 and the separation property holds . likewise , to have affinity in the volatility @xmath488
, we must have that @xmath489 must be independent of @xmath78 .
denote the ratio by @xmath479 , and differentiate with respect to @xmath301 to obtain @xmath490 hence , @xmath491 for @xmath492 and @xmath493 .
differentiating with respect to @xmath301 , and next letting @xmath483 gives @xmath494 whence , @xmath495 and the separation property holds for @xmath95 .
the proposition is proved .
proof of proposition [ propriskneutraldynamicsf ] let @xmath496 . from proposition [ propforward - bmambit ]
, we have that @xmath497 for a deterministic function @xmath498 given by @xmath499 note that the process @xmath500 is a ( local ) @xmath229-martingale for @xmath285 . moreover , from the stochastic fubini theorem it holds that @xmath501 where we note that @xmath502 hence , @xmath503 the result is then a direct consequence of the it formula for semimartingales , see , for example , protter @xcite .
proof of proposition [ optionprice ] from proposition [ propforward - bmambit ] , we know that we can write the forward price as @xmath504 let now @xmath347 , and suppose that @xmath505 . recall that @xmath349 , with @xmath350 is the fourier transform of @xmath351 defined by @xmath506 .
suppose that @xmath353 .
hence , we find @xmath507 next , by commuting integration and expectation using dominated convergence and @xmath471-adaptedness , we obtain @xmath508 \,\mathrm{d}y , \end{aligned}\ ] ] which holds by the stochastic fubini theorem . using the independent increment property of @xmath73 and double conditioning , we reach @xmath509 \\ & = & { \mathbb{e}}_{\eta } \biggl[\exp \biggl(\mathrm{i}y \int_{t}^{\tau } \frac12 h_t(s , s)\,\mathrm{d}u_s-\frac12 y^2 \int_t^{\tau}g^2(t , s ) \omega_{s}^2 \,\mathrm{d}s - \mathrm{i}y \int_t^{\tau } \frac12 g^2(t , s)\omega_s^2\,\mathrm{d}s \biggr ) \big| \mathcal{f}_t \biggr ] \\ & = & { \mathbb{e}}_{\eta } \biggl[\exp \biggl(\mathrm{i}y \int_{t}^{\tau } \frac12 h_t(s , s)\,\mathrm{d}u_s+a \int _
t^{\tau}g^2(t , s)\int _ { -\infty}^s i(s , v)\,\mathrm{d}u_v \ , \mathrm{d}s \biggr )
\big| \mathcal{f}_t \biggr],\end{aligned}\ ] ] where we define @xmath510 . using the stochastic fubini theorem again ,
we get @xmath511 \\ & = & { \mathbb{e}}_{\eta}\biggl[\exp\biggl(\int_{t}^{\tau } \mathrm{i}y\frac12 h_t(s , s)\,\mathrm{d}u_s \\ & & \hspace*{36pt}{}+a \int_{-\infty}^{t}\int_{v}^{\tau } g^2(t , s ) i(s , v ) \,\mathrm{d}s
\,\mathrm{d}u_v + a \int _ { t}^{\tau}\int_{v}^{\tau } g^2(t , s ) i(s , v ) \,\mathrm{d}s \,\mathrm{d}u_v\biggr ) \big| \mathcal{f}_t\biggr ] \\ &
= & \exp \biggl ( a \int_{-\infty}^{t}\int _ { v}^{\tau } g^2(t , s ) i(s , v ) \ , \mathrm{d}s \,\mathrm{d}u_v \biggr ) \\ & & { } \times { \mathbb{e}}_{\eta } \biggl[\exp \biggl(\mathrm{i}y \int_{t}^{\tau } \frac12 h_t(s , s)\,\mathrm{d}u_s + a \int _
{ t}^{\tau}\int_{v}^{\tau } g^2(t , s ) i(s , v ) \,\mathrm{d}s
\,\mathrm{d}u_v \biggr)\big | \mathcal{f}_t \biggr ] \\ & = & \exp \biggl ( a \int_{-\infty}^{t}\int _ { v}^{\tau } g^2(t , s ) i(s , v ) \ , \mathrm{d}s \,\mathrm{d}u_v \biggr ) \\ & & { } \times{\mathbb{e}}_{\eta } \biggl[\exp \biggl(\int_{t}^{\tau } \biggl\ { \mathrm{i}y \frac12 h_t(v , v ) + a \int_{v}^{\tau } g^2(t , s ) i(s , v ) \,\mathrm{d}s \biggr\ } \,\mathrm{d}u_v \biggr ) \big| \mathcal{f}_t \biggr].\end{aligned}\ ] ] altogether , we obtain @xmath512 the above expression can be further simplified by noting that @xmath513 then @xmath514 proof of proposition [ propforward - arithmetic ] observe that @xmath515=(-\mathrm{i})\frac{\mathrm{d}}{\mathrm{d}x } { \mathbb{e}}_{\theta,\eta } \biggl[\exp \biggl ( \mathrm{i}x\int_{-\infty } ^tg(t , s)\omega_{s- } \,\mathrm{d}l_s \biggr ) \big|
\mathcal{f}_t \biggr]_{x=0 } .\ ] ] we then proceed as in the proof of proposition [ propforward - generalambit ] , and finally we perform the differentiation and let @xmath516 .
proof of proposition [ propcorgammakernel ] we have @xmath517 which is a probability density and hence @xmath518 .
now we derive the explicit formula for the autocorrelation function .
@xmath519 note that according to gradshteyn and ryzhik @xcite , formula 3.383.8 @xmath520 for @xmath521 and @xmath522 , where @xmath454 is the modified bessel function of the third kind .
hence , @xmath523 now we apply gradshteyn and ryzhik @xcite , formula 8.335.1 , to obtain @xmath524 then @xmath525 since @xmath526 according to gradshteyn and ryzhik @xcite , formula 8.486.16 , the result follows .
we would like to thank andreas basse - oconnor and jan pedersen for helpful discussions and constructive comments . also , we are a grateful to the valuable comments by two anonymous referees and by the editor .
benth is grateful for the financial support from the project `` energy markets : modelling , optimization and simulation ( emmos ) '' funded by the norwegian research council under grant evita/205328 .
financial support by the center for research in econometric analysis of time series , creates , funded by the danish national research foundation is gratefully acknowledged by a.e.d . veraart . | this paper introduces the class of _ volatility modulated lvy - driven volterra _ ( @xmath0 ) processes and their important subclass of _ lvy semistationary _ ( @xmath1 ) processes as a new framework for modelling energy spot prices .
the main modelling idea consists of four principles : first , deseasonalised spot prices can be modelled directly in stationarity .
second , stochastic volatility is regarded as a key factor for modelling energy spot prices .
third , the model allows for the possibility of jumps and extreme spikes and , lastly , it features great flexibility in terms of modelling the autocorrelation structure and the samuelson effect . we provide a detailed analysis of the probabilistic properties of @xmath0 processes and show how they can capture many stylised facts of energy markets .
further , we derive forward prices based on our new spot price models and discuss option pricing .
an empirical example based on electricity spot prices from the european energy exchange confirms the practical relevance of our new modelling framework . , |
the baikal neutrino telescope nt-200 is operated in lake baikal , siberia , at a depth of .
a description of the detector as well as physics results from data collected in 1996 and 1998 ( 70 and 234 live days , respectively ) have been presented elsewhere @xcite . here
we present new limits including data taken in 1999 ( 268 live days ) .
we note that the year 2003 marks the tenth anniversary of the deployment of nt-36 , the pioneering first stationary underwater array @xcite .
fig.1 shows a textbook neutrino event ( an upward moving muon track ) recorded with the early 4-string configuration of 1996 @xcite .
the survey for high energy neutrinos is focused to bright cascades produced at the neutrino interaction vertex in a large volume around the neutrino telescope .
lack of significant light scattering allows to monitor a volume exceeding the geometrical volume by an order of magnitude .
this results in sensitivities of nt-200 comparable to those of the much larger amanda - b10 detector .
the background to this search are bright bremsstrahlung flashes along muons passing far outside the array ( see @xcite for details ) .
candidate events do not show a statistically significant excess of hit multiplicites compared to the simulated background from atmospheric muons . assuming an @xmath10 shape of the neutrino spectrum and a flavor ratio @xmath11 , the new , preliminary 90% c.l .
upper limit with respect to the flux of all three flavors is @xmath12 , about twice below previous results @xcite .
the preliminary limit on @xmath13 at the w - resonance energy is : @xmath14 fig .
2 shows the experimental upper limits obtained by baikal ( this work ) , and amanda @xcite , as well as the projected sensitivity for nt-200 + ( see below ) .
limits are compared to ssds @xcite and mpr @xcite predictions .
the slanted lines at the left side represent the atmospheric neutrino fluxes ( dashed for @xmath15 , solid for @xmath16 ) .
we have searched for high energy cascades coincident with 772 gamma ray bursts ( grb ) recorded between april 1998 and february 2000 by the batse detector and falling into on - line periods of nt-200 ( 386 triggered grb , 336 non - triggered from stern catalogue ) .
after cuts for background reduction , we are left with one event within one of the @xmath17100-second windows , where 0.47 events would have been expected from accidental coincidences .
precise limits on the fluence are being derived at present . with an effective volume in the megaton range ,
nt-200 is the largest northern detector for high energy neutrinos from grbs .
the search for wimps with the baikal neutrino telescope is based on a possible signal of nearly vertically upward going muons , exceeding the flux of atmospheric neutrinos ( see @xcite ) .
note that the threshold of @xmath0 10 gev for this analysis is lower than that of @xmath0 15 gev for atmospheric neutrinos spread across the full lower hemisphere .
3 ( top ) demonstrates that the angular distribution of events passing a special filter for events close to the vertical is well described by simulations including the effect of neutrino oscillations ( assuming @xmath18 ev@xmath4 ) . with no significant excess observed ,
we derive improved upper limits on the flux of muons from the direction of the center of earth related to wimp annihilation .
3 ( bottom ) compares our new limits to those obtained by other experiments ( see @xcite for references ) .
nt-200@xmath1 is an upgrade of nt-200 by three sparsely instrumented distant outer strings ( see fig . 5 ) .
the fenced volume is a few dozen mtons .
a prototype string of 140 m length with 12 optical modules was deployed in march 2003 , and electronics , data acquisition and calibration systems for nt-200@xmath1 have been tested .
the 3 strings allow for dramatically better vertex reconstruction of high energy cascades than with nt-200 alone ( see fig .
this , in turn , allows a much better determination of the energy and makes nt-200 + a true discovery detector with respect to reactions within the mentioned fiducial volume .
( this is in contrast to nt-200 which is an excellent exclusion experiment but would have certain difficulties to prove for a single event outside its geometrical volume that the signal is indeed due to a far high - energy event and not to a closer medium - energy event . )
nt-200 + will be installed in 2004 and 2005 .
first discussions have started on an even larger detector of the order of a cubic kilometer in lake baikal .
the configuration of such a detector will be consequently tailored to certain classes of events and should complement mediterranean arrays .
a possible design was sketched in @xcite . given the drawbacks of stronger absorbtion and shallower depth compared to mediterranean sites , this idea rests on the proven , stable deployment procedures , lower - cost considerations and possible national funding strategies . however , even in the case of strong national funding , such a project would need international participation .
we finally summarize some lessons which may be of use for all underwater arrays .
the early history of baikal could have been written as a history of leaking feed - throughs , connectors and pressure housings .
after having designed a special connector tailored to depths up to @xmath0 1.6 km ( the maximum depth in lake baikal ) , this phase was overcome .
nt-36 did not suffer from leaks . in the mean time , more than 1000 connectors and feed - throughs are in operation at 1.1 km depth , and the failure rate is less than 1 per year .
clearly , detectors much deeper , like the mediterranean ones , have to make use of different technologies .
still , is seems now obvious that instruments with hundreds of pressure spheres and more than thousand connectors can work in deep water over many years , with tolerable leakage rates a statement which ten years ago would have been questioned by many oceanographers .
sedimentation and bio - fouling are a concern in lake baikal .
the performance of the up - down symmetric nt-36 detector , with the same number of optical modules facing upward and downward , was particularly affected by sedimentation @xcite .
the sensitivity of the upward looking modules with respect to atmospheric muons decreased to 35% after one year @xcite . at present , only two of the 12 layers of nt-200 have upward facing modules .
the glass spheres are dressed with a tapering hat of plastic foil .
much of the sediments glide along the foil and do not stick to the hat .
the effect of sedimentation , i.e. a progressively decreasing sensitivity of the upward looking modules , is weakened , and an acceptable functionality may be kept by cleaning the hats once every two years . this is important since the outer strings of nt-200 + will be up - down symmetric .
we conclude that symmetric configurations can be operated in lake baikal , but only if the optical moudules can be easily hauled up , inspected and cleaned . the same limitation should hold for those mediterannean sites with strong biological activity .
the possibility to haul up the array and to repair connectors was a key for nt-200 s operation . in the very beginning , about 40% of the optical modules turned non - operational after one year of operation due to failures in electronics .
clearly this was not tolerable , be it with or without the possiblity to repair failed components . in the mean time the failure rate reduced to less than 5% per year .
this lower rate , combined with the possibility to repair failed components results in a tolerable regime of operation and maintenance . from our experience
we would recommend designs which give a not too complicated access to deployed parts of the detector . with the reliability of the detector itself improving from year to year
, we were faced by another , non - expected problem : we observed electro - corrosion along the cables to shore . actually electro - corrosion was thought to be negligible since the fresh water of lake baikal is nearly free of minerals and salts and is a bad conductor of electricity .
much of the return current was thought to flow along the outer metal jacket of the shore cable ( white arrows in figure 7 ) and not through the water . during the last year it became clear that the cable arming showed damages at the points indicated in the figure
. the reason seems to be that close to the shore the current leaks into ground and propagates there with a low resistance compared to the cable , and that also current flows trough the water at the depth of the big metallic , umbrella - like support frame of nt-200 ( black arrows ) . at these points ,
strong electro - corrosion is observed .
it led to the loss of two of the four shore cables in the last year .
the cables will be replaced by new cables in 2004 , and the distance between the vertical part of the shore cable and nt-200 will be increased .
electro - corrosion should be an even stronger issue for salt water detectors , as shown by the dumand history .
the baikal experience can certainly not be translated to different designs of power distribution and return current .
still it may suggest that one should be prepared to unpleasant surprises and necessary design changes .
a staged approach seems mandatory when building a large instrument in a new and challenging environment like deep water .
the idea that one can already in the design phase account for all possible problems has been disqualified not only by the baikal experience but also by dumand , nestor and antares .
the operation of prototype detectors and the readiness for technological modifications is of key importance .
underwater detectors do not suffer from strong light scattering as does ice .
therefore the timing information from distant light sources is not strongly dispersed .
this opens the possibility to monitor a huge volume around the detector .
a few strings far outside , like those planned for nt-200 + , can dramatically improve the sensitivity to pev processes .
due to best background rejection , the most efficient region is that _ below _ the geometrical volume of the detector .
apart from stronger bottom water currents and slighlty worse water quality close to ground , this may be another argument not to place the detector as deep as possible , but a few hundred meters above ground .
99 zh .- a .
dzhilkibaev for the baikal collaboration , proc .
workshop on neutrino telescopes ( venice 2001 ) ; and astro - ph/0105269 .
domogatsky for the baikal collaboration , proc .
xxth int .
conf . on neutrino physics and astrophysics ( munich 2002 ) 353 ; and astro - ph/0211571 .
i. belolaptikov et al . , astropart .
phys . 7 ( 1997 ) 263 .
underwater : the tide turns , cern courier 33/9 ( 1993 ) 36 .
i. belolaptikov et al . , astropart .
. 12 ( 1999 ) 75 .
balkanov et al .
, astropart .
phys . 14 ( 2000 ) 61 .
see the talk of e. bernardini in these proceedings .
stecker et al . , phys.rev.lett . 66 ( 1991 ) 2697 .
k. mannheim , r. protheroe , j. rachen , phys.rev .
d 63 ( 2001 ) 023003 .
v. balkanov et al .
, phys . atom .
nucl . 66 ( 2003 ) 503 .
bagduev et al . ,
a420 ( 1999 ) 138 . | we review recent results on the search for high energy extraterrestrial neutrinos , neutrinos induced by wimp annihilation and neutrinos coincident with gamma ray bursts as obtained with the baikal neutrino telescope nt-200 .
we describe the moderate upgrade of nt-200 towards a @xmath010 mton scale detector nt-200@xmath1 .
we finally draw a few lessons from our experience which may be of use for other underwater experiments .
> 0 _ the baikal collaboration : + _
v. aynutdinov@xmath2 , v. balkanov@xmath2 , i. belolaptikov@xmath3 , l. bezrukov@xmath2 , n. budnev@xmath4 , a. chensky@xmath4 , d. chernov@xmath5 , i. danilchenko@xmath2 , zh .- a .
dzhilkibaev@xmath2 , g. domogatsky@xmath2 , a. dyachok@xmath4 , o. gaponenko@xmath2 , o. gress@xmath4 , t. gress@xmath4 , a. klabukov@xmath2 , a. klimov@xmath6 , s. klimushin@xmath2 , k. konischev@xmath3 , a. koshechkin@xmath2 , v. kulepov@xmath7 , l. kuzmichev@xmath5 , vy .
kuznetzov@xmath2 , b. lubsandorzhiev@xmath2 , s. mikheyev@xmath2 , m. milenin@xmath7 , r. mirgazov@xmath4 , n. moseiko@xmath5 , e. osipova@xmath5 , a. panfilov@xmath2 , g. pankov@xmath4 , l. pankov@xmath4 , yu .
parfenov@xmath4 , a. pavlov@xmath4 , e. pliskovsky@xmath3 , p. pokhil@xmath2 , v. polecshuk@xmath2 , e. popova@xmath5 , v. prosin@xmath5 , m. rosanov@xmath8 , v. rubtzov@xmath4 , yu . semeney@xmath4 , b. shaibonov@xmath2 , c. spiering@xmath9 , b. tarashansky@xmath4 , r. vasiliev@xmath3 , e. vyatchin@xmath2 , r. wischnewski@xmath9 , i. yashin@xmath5 , v. zhukov@xmath2 + _ ( 1 ) institute for nuclear research , moscow , russia , ( 2 ) irkutsk state university , irkutsk , russia , ( 3 ) skobeltsyn institute of nuclear physics msu , moscow , russia , ( 4 ) joint institute for nuclear research , dubna , russia , ( 5 ) desy zeuthen , zeuthen , germany , ( 6 ) nizhni novgorod state technical university , ( 7 ) st.peterburg state marine university , st.peterburg , russia , ( 8) kurchatov institute , moscow , russia _ + |
the isoscalar giant monopole resonance ( isgmr ) , known as the so - called `` breathing mode '' , is one of the collective modes of nuclei . in the past decades
, gmr was extensively studied both theoretically @xcite and experimentally @xcite . especially , a strong correlation between the peak energy of gmr and the nuclear incompressibility @xmath4 at the nuclear saturation density was found @xcite .
the studies from both relativistic and non - relativistic models have reached a consensus on the value of the nuclear incompressibility at @xmath5 mev @xcite .
recently , gmr along the sn isotopic chain was studied experimentally @xcite . from the analysis based on the gmr data for sn isotopes ,
the asymmetry term of the nuclear incompressibility was constrained , i.e. , @xmath6 mev @xcite .
similar analysis on the gmr data in the cd isotopes gave a preliminary value of @xmath7 mev @xcite .
comparison between the experimental data and the theoretical results has indicated that models which can reproduce the peak energies of gmr in @xmath2zr , @xmath8sm , and @xmath3pb overestimate those in sn isotopes .
this realization leaves a puzzling question : `` why is tin so soft ? '' @xcite . in ref .
@xcite , the effect of pairing correlations on the peak energy of gmr was considered . however , the result that the peak energies of gmr in sn isotopes are shifted by about 100 - 150 kev compared to the case without pairing correlations was insufficient to explain the experimental data . in ref .
@xcite , a hybrid model with a small nuclear incompressibility of @xmath9 mev as fsugold and a stiff symmetry energy as nl3 was built .
although the improvement in the description for the experimental data of sn isotopes was significant and unquestionable , the hybrid model still underestimated the peak energy of gmr in @xmath3pb by almost 1 mev .
the authors of ref .
@xcite also suggested that the rapid softening with neutron excess predicted by the hybrid model might be unrealistic .
more details of the discussion on this anomaly can be found in refs .
@xcite . in the previous works ,
our group have applied the isospin - dependent quantum molecular dynamics ( iqmd ) model to study the dynamical dipole emission in fusion reactions @xcite and giant dipole resonances ( gdr ) as well as pygmy dipole resonances ( pdr ) in ni isotopes by coulomb excitations @xcite . in the present work
, we will investigate gmr in sn isotopes within a similar framework .
we will show that using a new function of the gaussian wave - packet width , which takes the isospin dependence into account , we are able to reproduce very well the gmr peak energies for @xmath1ni , @xmath2zr , and @xmath3pb as well as sn isotopes . the paper is organized as follows .
section [ sec-2 ] gives a brief introduction of the iqmd model as well as the formalism for gmr in the iqmd framework .
results and discussions are presented in sec .
[ sec-3 ] , where effects from the impact parameter , the incident energy , the equation of state ( eos ) , the symmetry energy , and the width of the gaussian wave packet used in the iqmd model on gmr are investigated .
a summary is given in sec .
[ sec-4 ] .
the iqmd model , which is based on the qmd model , is a kind of monte - carlo transport model @xcite .
the wave function of each nucleon is represented by a gaussian form : @xmath10.\ ] ] in the above , @xmath0 is the width parameter for the gaussian wave - packet , and its value depends on the size of the reacting system to keep some quantum effect of nucleons .
we will see in the following that the influence of the width on gmr should be treated carefully to reproduce the experimental results . @xmath11 and @xmath12 are the position and momentum coordinates of the @xmath13th nucleon . after performing variation method , the equations of motion , i.e.
, the time evolution of the mean position @xmath11 and momentum @xmath12 , are found to be @xmath14 @xmath15 is the total hamiltonian of the system @xmath16 where the @xmath17 is the kinetic contribution , and @xmath18 is the potential contribution @xmath19 in the above , the wigner distribution function @xmath20 , which is the phase - space density of the @xmath13th nucleon , is obtained by applying the wigner transformation on the single nucleon wave function @xmath21,\ ] ] and @xmath22 the two - body interaction including the contact skyrme - type interaction , the finite - range yukawa potential , the momentum - dependent interaction ( mdi ) , the isospin - dependent interaction , and the coulomb interaction @xmath23 } { |\vec{r}-\vec{r}^\prime|/\mu } + t_6 \rho^{\gamma-1}(\vec{r } ) t_{3i } t_{3j }
\delta(\vec{r } - \vec{r}^\prime ) \nonumber \\ & + & t_4\hbox{ln}^2 [ 1+t_5(\vec{p}-\vec{p}^\prime)^2 ] \delta ( \vec{r } -\vec{r}^\prime ) + \frac{z_i z_j e^2}{|\vec{r}-\vec{r}^\prime|},\end{aligned}\ ] ] where @xmath24 is the charge of the nucleon , and @xmath25 and @xmath26 are the parameters to fit the empirical properties of nuclear matter as well as nuclei . in the following , we will give the expressions for the skyrme potential , the momentum - dependent potential , and the symmetry potential to ease discussions for the gmr results .
the skyrme potential is @xmath27 where @xmath28 is the reduced density with @xmath29 @xmath30 being the nuclear saturation density and @xmath31 with @xmath32 being the interaction density of the @xmath13th nucleon .
we will in the following denote @xmath33 as @xmath34 for simplicity .
@xmath35 , @xmath36 , and @xmath37 are the skyrme parameters related to the isoscalar eos of bulk nuclear matter .
the momentum dependent potential , which is optional in the iqmd model , can be expressed as @xmath38,\ ] ] where @xmath39 $ ] is the momentum density distribution function of nucleon , @xmath40 = 1.57 mev and @xmath41 = 500 ( gev / c)@xmath42 are taken from the measured energy dependence of the proton - nucleus optical potential @xcite .
the isospin asymmetry potential can be also calculated from eq.([meanfield ] ) @xmath43 , \label{eq04}\end{aligned}\ ] ] where @xmath44 is the local isospin asymmetry from the contribution of all the other nucleons , and the symbol `` @xmath45 '' is for neutrons ( protons ) , @xmath46 is the stiffness parameter of the symmetry potential ( energy ) , and @xmath47 is the potential contribution of the symmetry energy at saturation density .
the nuclear incompressibility is calculated from the second - order derivative of the binding energy per nucleon @xmath48 table [ table1 ] gives different parameter sets for the skyrme potential with and without the momentum - dependent potential , leading to the nuclear incompressibility of @xmath49 mev and @xmath50 mev .
as mentioned in the introduction , although the latest experimental analysis leads to @xmath4 @xmath51 240 mev , we will use these extreme values to illustrate the sensitivity of the gmr peak energy on the nuclear incompressibility as well as the momentum dependence of the nuclear potential based on the iqmd model .
.the parameters @xmath35 , @xmath36 , and @xmath37 for different eoss .
[ cols= " < ,
< , < , < , < " , ] in the present framework , we first pick up a stable initial density distribution for a concerned nucleus , e.g. , @xmath52sn , as is done for most qmd model studies .
this density distribution is generally not the ground state for the nuclear interaction used , so the nucleus suffers from collective oscillation in its excited state @xcite , among them is the gmr mode . as gmr is a compression mode in radial direction , we take the root - mean - square ( rms ) radius of the nucleus as its monopole moment @xmath53 at each time step @xcite , as is shown by the solid line in fig .
one sees that the rms radius shows good oscillation structure , i.e. , the gmr mode .
however , the oscillation damps quickly due to the dissipation effect from both the mean - field potential and the nucleon - nucleon scatterings @xcite .
it is noteworthy that the period and the decay of the gmr oscillation leads respectively to the peak energy and the width of the gmr spectrum , which can still be hardly reproduced simultaneously @xcite . in the present study ,
we use @xmath3pb as a target and the concerned nucleus as the projectile . in this way ,
the strength , period , and damping of the gmr oscillation are thus modified by the coulomb interaction between the projectile and the target , as shown in fig .
it is seen that with coulomb interaction , the rms radii have a larger amplitude and a little longer oscillation period with the increasing of the beam energy and become larger on average in the later stage .
overall , a good oscillation behavior of the rms radius in relative long time scale is the key to form a gmr mode .
of course , considering the stability of time evolution of the rms radius in the present model , we calculate the gmr spectra by 200 fm / c . by applying the fourier transformation to the second - order derivative of @xmath53 with respect to time @xmath54
one can get the spectrum of probability for energy @xmath55 as follows @xmath56 as mentioned above , we set the stopping time of the monopole moment ( @xmath57 ) as 200 fm / c in the present calculation . from a gaussian fit to the spectrum of gmr , one can get the peak energy @xmath58 , the fwhm @xmath59 , and the strength @xmath60 of gmr .
the gmr spectra for @xmath52sn and @xmath61sn from our calculations are compared with the experimental data from ref .
@xcite in fig .
[ sn - exp - cal ] .
note that there are one major different point which we should mention in this comparison between our calculation and the data . in ref .
@xcite , the gmr data are taken from the excited sn nucleus by inelastic scattering of 400-mev @xmath35 particles at extremely forward angles . however , in our calculation , the gmr comes from the excited oscillation of sn nucleus as the initial density distribution is not the ground state of the nuclear interaction used .
the reason is based on the following consideration : ( 1 ) it is not an easy task to treat inelastic scattering of @xmath35-particles in our iqmd model even though the data is available ; ( 2 ) the peak energy and the fwhm are determined by the intrinsic properties of the nucleus and independent of how the gmr mode is excited . consequently , we have to take a compromise for the comparison , i.e. , taking the gmr for the same excited nucleus but with different reaction mechanism , with the strength @xmath60 of gmr from our calculation scaled by that from the experimental data . in this background
, we introduce the parameters used in our calculations as follows : incident energy @xmath62 mev / nucleon , impact parameter @xmath63 fm , the soft eos with mdi , @xmath64 mev , and @xmath65 .
although the condition of our calculation is different from that of the experiment in ref . @xcite as mentioned above , one can see that our results from the oscillation of excited nuclei modified by the coulomb interaction show a reasonable agreement with the inelastic @xmath35 scattering experimental data , i.e. , giving a similar peak energy but a slightly larger fwhm .
the calculated result of @xmath52sn shows a better agreement with the experimental data than that of @xmath61sn .
the results indicate that the iqmd model is suitable for the study of gmr by considering the rms radius as its monopole moment .
the sensitivities of the peak energy , the fwhm , and the strength obtained by a gaussian fit to the gmr spectrum have been explored .
the sensitivity of these quantities to the impact parameter for @xmath52sn is given in fig .
[ b - dep ] .
it shows that the gmr results do not change much with the increasing impact parameter .
this is understandable as the effect of coulomb interaction is not affected by much when the impact parameter changes from 25 fm to 40 fm .
figure [ e - dep ] shows the incident energy dependence of gmr results for @xmath52sn . with the increase of the incident energy , the peak energy of gmr decreases , while the fwhm and the strength of gmr increase .
this behavior can be understood by the oscillation of rms radii at different beam energies as shown in fig .
. as we know , the longer oscillation period corresponds to the lower frequency , i.e. lower energy , while the higher amplitude corresponds to the larger strength .
[ rms ] tells us that with the increasing of beam energy , gmr monopole moment has a little longer period but a larger amplitude , which results in a decreasing peak energy and an increasing strength of gmr as shown in fig .
[ e - dep ] . from the above discussions
, it is seen that although the gmr oscillation is already there for a nucleus alone , it can be slightly modified by the coulomb interaction with different incident energies and impact parameters .
many previous works indicated that the eos associated with the nuclear incompressibility @xmath66 has an important influence on gmr , i.e. , the peak energy of gmr can be used to constrain @xmath66 . by adjusting the parameters of eos in table .
[ table1 ] used in the iqmd model , the eos dependence of gmr parameters for @xmath52sn can be explored .
the sensitivity of the gmr results to the eos are illustrated in fig .
[ eos - dep ] , where a significant dependence of the gmr peak energy on the eos can be seen .
in general , the hard eos gives a higher peak energy than the soft one , and so does the eos with mdi . in this sense ,
a soft eos with mdi can reproduce the results from a hard eos without mdi .
similar results on giant or pygmy dipole resonance are seen within the same model @xcite . by comparing the calculated results with the experimental data @xcite
, one can see that the soft eos with mdi shows the best agreement with the experimental data .
the gmr results may also be affected by the nuclear symmetry energy which is important in understanding the structure of neutron- or proton - rich nuclei and the reaction dynamics of heavy - ion collisions @xcite .
again we use the extreme values of @xmath47 and @xmath46 to illustrate the sensitivity of the gmr results to the nuclear symmetry energy .
the dependence of the gmr results on the parameter @xmath47 is shown in fig .
[ csym - dep ] .
when @xmath47 changes from 16 mev to 64 mev , the peak energy of gmr shows a decreasing behavior , the fwhm of gmr increases , and the strength of gmr slightly decreases . the @xmath46 dependence of the gmr results is shown in fig .
[ gamma - dep ] .
when @xmath46 changes from 0.5 to 2 , the peak energy of gmr also decreases , while the fwhm and the strength of gmr show a non - monotonical behavior . to understand the dependence of the gmr peak energy on the symmetry energy ,
we express it in the form of @xcite : @xmath67 where @xmath68 is the nucleon mass , @xmath69 is the ground - state mean square radius , and @xmath70 , which is the incompressibility of a nucleus with mass a , can be written as @xcite : @xmath71 ^ 2 \nonumber \\ & & + k_cz^2a^{-4/3 } , \label{eq - ka}\end{aligned}\ ] ] with @xmath72 and @xmath24 the neutron and proton number , and @xmath73 as well as @xmath74 , @xmath75 , and @xmath76 the coefficients for the volume , asymmetry , and coulomb contributions , respectively . in the analysis of the symmetry energy effects on the gmr results where all the other parameters have been fixed
, @xmath70 increases with the increasing asymmetry incompressibility @xmath77 , with the latter expressed as @xcite @xmath78 the slope parameter @xmath79 and the curve parameter @xmath80 at saturation density can be calculated from the expression of the symmetry energy as @xmath81 @xmath4 and @xmath82 are related to the isoscalar part of the equation of state @xmath83 in the form of @xmath84 the dependence of @xmath85 on @xmath47 and @xmath46 in figs .
[ csym - dep ] and [ gamma - dep ] can be understood from the above formulae . the gaussian wave - packet width @xmath0 ( in eq .
( [ eq : gauspack ] ) ) for nucleons is a parameter in the iqmd model indicating the interaction range between nucleons .
previous studies have shown that the value @xmath0 has a large effect on gmr results @xcite . in many previous qmd model calculations ,
@xmath0 is set to be a constant .
however , there were also some discussions related to the influence of @xmath0 on the dynamical results , e.g. , flow , multifragmentation , and pion and kaon production , etc . @xcite . since in a finite system nucleons
are localized within a potential well , it is reasonable to make @xmath0 related to the size of a nucleus .
for example , the width of the gaussian wave packet is taken to be @xmath86 fm for ca + ca system and @xmath87 fm for au+au system in ref .
@xcite , while a system - size - dependent @xmath0 was presented in ref .
actually , in some models like an extended quantum molecular dynamics ( eqmd ) model @xcite and a fermionic molecular dynamics ( fmd ) model @xcite , @xmath0 is treated as a dynamical variable . in our iqmd model ,
we use a similar mass - number dependence of @xmath0 as in ref .
@xcite @xmath88 where @xmath89 is the mass number of the system ( projectile , target , or compound system ) . in heavy - ion reactions , @xmath0 for the compound system is set to be the mean value of that for the projectile and the target given by eq .
( [ eq - sigma ] ) . in the present study
, we just use @xmath0 for the concerned nucleus , i.e , the projectile .
this is similar to the treatment in ref .
@xcite , where the projectile and target have their own @xmath0 before they contact in the heavy - ion reaction process , and after the contact , the projectile and the target gradually melt into one system , and consequently all the particles have a universal @xmath0 .
figure [ del - dep ] displays the @xmath0 dependence of the gmr results . as @xmath0 increases ,
the peak energy and the fwhm of gmr show a clear decreasing trend , while the strength of gmr is not largely affected .
the influence of different forms of @xmath0 on the peak energy of gmr has also been studied .
figure [ del - com ] shows the calculated peak energies of gmr in sn isotopes with the fixed @xmath87 fm or the variational @xmath0 given by eq .
( [ eq - sigma ] ) together with the experimental data @xcite , respectively .
one can see that the mass - number dependent @xmath0 gives a larger peak energy and shows a stronger mass number dependence of the gmr peak energy than the constant @xmath0 .
however , it still gives a weaker mass - number dependence of the gmr peak energy in comparison with the experimental data as seen from the right panel of fig .
[ del - com ] .
results from other theoretical studies @xcite are also plotted for comparison .
one can see that they overestimate the peak energies of gmr by about @xmath90 mev in sn isotopes , and the mass - number dependence of the gmr peak energy seems also weaker than the experimental data . from the above discussions
, it is seen that introducing only the mass - number dependence to the width of the gaussian wave packet for nucleons in the iqmd model is not sufficient to reproduce the experimental results of the gmr peak energies for sn isotopes . however , considering the correlation between @xmath0 and the peak energy of gmr , we can fit the values of @xmath0 with the experimental data for different sn isotopes .
figure [ snx - dep ] shows such a fit of @xmath0 for @xmath52sn , @xmath91sn , @xmath92sn , and @xmath61sn using the experimental results of gmr peak energies from ref .
it is seen that with the increasing mass and isospin asymmetry along the sn isotope line , the gmr peak energy decreases while @xmath0 from fitting the gmr peak energy increases .
the above study gives us some hints that to reproduce reasonably well the experimental data of the gmr peak energies for sn isotopes , @xmath0 may depend not only on the mass number but also on the isospin asymmetry of the nucleus . on the other hand , since the gmr peak energy is strongly correlated with the nucleus incompressibility @xmath70 @xcite , it is reasonable to assume that @xmath0 has a functional form similar to that of @xmath70 , i.e. , @xmath93 ^ 2+cz^2a^{-4/3}+d.\ ] ] by fitting the experimental data of the gmr peak energies for sn isotopes @xcite with the above functional form , we obtain the following expression for @xmath0 @xmath94 ^ 2 \nonumber \\ & & + 1.6\times10^{-8}z^2a^{-4/3}+2.01~{\rm ( fm)}. \label{eq - del - iso}\end{aligned}\ ] ] the small coefficient for the charge number dependence indicates that the optimized choice of @xmath0 might be a quadratic function of the nucleus isospin asymmetry . the new function of @xmath0 is compared with the old one ( eq . ( [ eq - sigma ] ) ) used in the qmd model in fig .
[ del - iso ] , and it is seen that they are quite different especially for isotope chains .
equation ( [ eq - del - iso ] ) has also been extended in the calculation of gmr peak energies for @xmath95ca , @xmath1ni , and @xmath2zr as well as @xmath3pb and the overall results are compared with those from other theoretical models @xcite as well as the experimental data @xcite in fig .
[ a - dep - iso ] .
it is found that after introducing the isospin dependence to the gaussian wave - packet width fitted by the gmr peak energies for sn isotopes of intermediate nucleus mass , our results agree with the experimental data for light or heavy nuclei such as @xmath1ni , and @xmath2zr as well as @xmath3pb much better than others in the literature , which overestimate the gmr peak energies . for the even lighter nucleus such as @xmath95ca , although our result follows the same trend of mass - number dependence for intermediate and heavy nuclei , it overestimates the gmr peak energy compared with the experimental data .
different from the cases of intermediate and heavy nuclei , the gmr peak energies from the experimental studies fluctuate with mass number for light nuclei .
this is due to the increasing effect of the shell structure as well as the paring correlation for smaller nuclei , which has already been beyond the limit of the framework of the iqmd model and the other transport models with only mean - field potentials .
in this work , we have applied the iqmd model to investigate the isoscalar giant monopole resonance ( isgmr ) in sn isotopes and other nuclei .
the collective oscillation , including the gmr mode , appears for an initial density distribution for the concerned nucleus as it is generally not the ground state for the nuclear interaction used .
this oscillation is further modified by the coulomb interaction when we take the concerned nucleus as the projectile and the @xmath3pb nucleus as the target .
we took the rms radius of the concerned nucleus as the monopole moment of gmr and calculated the spectrum of gmr . using a gaussian fit to the spectrum
, we calculated the peak energy , the fwhm , and the strength of gmr .
the sensitivity of these gmr results to the parameters used in the iqmd model was discussed .
the gmr peak energy is found to slightly decrease with increasing incident energy , while it is almost independent of the impact parameter .
it seems difficult to extract the information of the symmetry energy from the present study of gmr , as we found that gmr is also sensitive to other parameters such as the isoscalar part of the eos and the gaussian wave - packet width @xmath0 .
as observed previously , the eos associated with the nuclear incompressibility has an important influence on gmr .
comparing our results with the experimental data , it is found that the soft eos with mdi can give a better fit to the experimental data .
the studies of the systematic evolution for sn isotopes have shown that a widely used mass - number dependent @xmath0 overestimates the peak energies of gmr by about 1 mev for sn isotopes and leads to a weaker mass - number dependence of the gmr peak energies compared with the experimental data . by fitting the experimental data of the gmr peak energies for sn isotopes using the functional form of the nucleus incompressibility
, we obtain a new function of @xmath0 with isospin dependence in addition to mass - number dependence . applying this new form of @xmath0 to the calculation of the gmr peak energies leads to a good agreement with the experimental data for @xmath1ni , @xmath2zr , and @xmath3pb , although it still overestimates the gmr peak energy of @xmath95ca due to the lacking of effects from shell structure as well as pairing correlation in the iqmd framework
. it will be interesting to check whether the width of the gaussian wave packet for nucleons obtained from the present study can give better explanations for other observables from iqmd model calculations in the future studies .
this work was supported in part by the national natural science foundation of china under contract nos .
11035009 , 11220101005 , 10979074 , 11175231,11205230 the major state basic research development program in china under contract no .
2013cb834405 , the knowledge innovation project of chinese academy of sciences under grant no .
kjcx2-ew - n01 , and the `` 100-talent plan '' of shanghai institute of applied physics under grant y290061011 from the chinese academy of sciences .
99 o. civitarese , a. g. dumrauf , and m. reboiro , http://dx.doi.org/10.1103/physrevc.43.2622[phys . rev . *
c43 * , 2622 ( 1991 ) . ] b. k. agrawal , s. shlomo , and v. kim au , http://dx.doi.org/10.1103/physrevc.68.031304[phys . rev . *
c68 * , 031304 ( 2003 ) . ]
g. col @xmath96 , http://dx.doi.org/10.1103/physrevc.70.024307[phys . rev . *
c70 * , 024307 ( 2004 ) . ] b. g. todd - rutel and j. piekarewicz , http://dx.doi.org/10.1103/physrevlett.95.122501[phys .
lett . * 95 * , 122501 ( 2005 ) . ]
j. piekarewicz , http://dx.doi.org/10.1103/physrevc.76.031301[phys . rev . *
c76 * , 031301(r ) ( 2007 ) . ]
j. piekarewicz and m. centelles , http://dx.doi.org/10.1103/physrevc.79.054311[phys . rev . * c79 * , 054311 ( 2009 ) . ]
v. tselyaev @xmath96 , http://dx.doi.org/10.1103/physrevc.79.034309[phys . rev . *
c79 * , 034309 ( 2009 ) . ] e. khan , http://dx.doi.org/10.1103/physrevc.80.011307[phys . rev . *
c80 * , 011307 ( 2009 ) . ] e. khan , http://dx.doi.org/10.1103/physrevc.80.057302[phys . rev . *
c80 * , 057302 ( 2009 ) . ] e. khan , j. margueron , and i. vidaa , http://dx.doi.org/10.1103/physrevlett.109.092501[phys .
lett . * 109 * , 092501 ( 2012 ) . ] s. shlomo and d. h. youngblood , http://dx.doi.org/10.1103/physrevc.47.529[phys . rev . *
c47 * , 529 ( 1993 ) , ] and references therein . t.
li @xmath96 , http://dx.doi.org/10.1103/physrevlett.99.162503[phys .
lett . * 99 * , 162503 ( 2007 ) . ]
u. garg @xmath96 , http://dx.doi.org/10.1016/j.nuclphysa.2007.01.046[nucl . phys . *
a788 * , 36 ( 2007 ) . ]
t. li @xmath96 , http://dx.doi.org/10.1103/physrevc.81.034309[phys . rev . *
c81 * , 034309 ( 2010 ) . ]
u. garg , http://dx.doi.org/10.5506/aphyspolb.42.659[acta phys . polon . *
b42 * , 659 - 666 ( 2011 ) . ]
j. blaizot @xmath96 , http://dx.doi.org/10.1016/0375-9474(95)00294-b[nucl . phys . *
a591 * , 435 ( 1995 ) . ] d. patel _ et al .
b * 718 * , 447 ( 2012 ) .
h. l. wu @xmath96 , http://dx.doi.org/10.1103/physrevc.81.047602[phys . rev . * c81 * , 047602 ( 2010 ) . ] c. tao @xmath96 , http://dx.doi.org/10.1103/physrevc.87.014621[phys . rev . *
c87 * , 014621 ( 2013 ) ] ; http://www.j.sinap.ac.cn/nst/en/article/downloadarticlefile.do?attachtype=pdf&id=14[nucl .
sci . tech .
* 24 * , 030502 ( 2013 ) . ]
j. aichelin , http://dx.doi.org/10.1016/0370-1573(91)90094-3[phys . rep . * 202 * , 233 ( 1991 ) . ] c. hartnack @xmath96 , http://dx.doi.org/10.1007/s100500050045[eur .
j. * a1 * , 151 ( 1998 ) .
] c. hartnack @xmath96 , http://dx.doi.org/10.1016/0375-9474(89)90328-x[nucl . phys .
* a495 * , 303c ( 1989 ) . ]
y. g. ma _ et al .
_ , http://dx.doi.org/10.1103/physrevc.73.014604[phys .
rev . * c73 * , 014604 ( 2006 ) . ] sanjeev kumar , suneel kumar , and r. k. puri , http://dx.doi.org/10.1103/physrevc.81.014601[phys . rev . * c81 * , 014601 ( 2010 ) ; ] http://dx.doi.org/10.1103/physrevc.81.014611[*c81 * , 014611 ( 2010 ) ; ] s. gautam and r. k. puri , http://dx.doi.org/10.1103/physrevc.85.067601[ibid . * c85 * , 067601 ( 2012 ) . ]
g. q. zhang @xmath96 , http://link.aps.org/doi/10.1103/physrevc.84.034612[phys . rev . *
c84 * , 034612 ( 2011 ) . ] j. wang @xmath96 , http://www.j.sinap.ac.cn/nst/en/article/downloadarticlefile.do?attachtype=pdf&id=13[nucl .
sci . tech .
* 24 * , 030501 ( 2013 ) . ] v. baran @xmath96 , http://dx.doi.org/10.1016/s0375-9474(00)00365-1[nucl .
a679 * , 373 ( 2001 ) . ]
t. gaitanos @xmath96 , http://dx.doi.org/10.1103/physrevc.81.054316[phys .
* c81 * , 054316 ( 2010 ) .
] d. h. youngblood @xmath96 , http://dx.doi.org/10.1103/physrevc.69.034315[phys . rev . *
c69 * , 034315 ( 2004 ) .
] y. w. lui @xmath96 , http://dx.doi.org/10.1103/physrevc.70.014307[phys . rev .
* c70 * , 014307 ( 2004 ) . ]
b. a. li , http://dx.doi.org/10.1103/physrevlett.85.4221[phys .
lett . * 85 * , 4221 ( 2000 ) . ]
w. j. xie , f. s. zhang , http://www.j.sinap.ac.cn/nst/en/abstract/abstract50.shtml[nucl .
sci . tech . * 24 * , 050502 ( 2013)];w .
z. jiang , http://www.j.sinap.ac.cn/nst/en/abstract/abstract55.shtml[ibid , 050507 ( 2013 ) ] ; c. w. ma _
/ nst / en / abstract / abstract58.shtml [ ibid , 050510 ( 2013 ) ] .
s. stringari , http://dx.doi.org/10.1016/0370-2693(82)91182-0[phys .
b108 * , 232 ( 1982 ) . ]
l. w. chen @xmath96 , http://link.aps.org/doi/10.1103/physrevc.80.014322[phys .
* c80 * , 014322 ( 2009 ) . ] t. furuta @xmath96 , http://dx.doi.org/10.1103/physrevc.82.034307[phys . rev . * c82 * , 034307 ( 2010 ) . ] s. gautam @xmath96 , http://dx.doi.org/10.1088/0954-3899/37/8/085102[j . phys .
* g37 * , 085102 ( 2010 ) . ]
t. maruyama , k. niita , and a. iwamoto , http://dx.doi.org/10.1103/physrevc.53.297[phys .
* c53 * , 297 - 304 ( 1996 ) . ] h. feldmeier , http://dx.doi.org/10.1016/0375-9474(90)90328-j[nucl . phys .
* a515 * , 147 ( 1990 ) . ] h. feldmeier , k. bieler , and j. schnack , http://dx.doi.org/10.1016/0375-9474(94)00792-l[nucl .
* a586 * , 493 ( 1995 ) . ]
n. wang , z. x. li , and x. z. wu , http://dx.doi.org/10.1103/physrevc.65.064608[phys . rev . *
c65 * , 064608 ( 2002 ) . ] | the isoscalar giant monopole resonance ( isgmr ) in sn isotopes and other nuclei is investigated in the framework of the isospin - dependent quantum molecular dynamics ( iqmd ) model .
the spectrum of gmr is calculated by taking the root - mean - square ( rms ) radius of a nucleus as its monopole moment .
the peak energy , the full width at half maximum ( fwhm ) , and the strength of gmr extracted by a gaussian fit to the spectrum have been studied .
the gmr peak energies for sn isotopes from the calculations using a mass - number dependent gaussian wave - packet width @xmath0 for nucleons are found to be overestimated and show a weak dependence on the mass number compared with the experimental data .
however , it is found that experimental data of the gmr peak energies for @xmath1ni , @xmath2zr , and @xmath3pb as well as sn isotopes can be nicely reproduced after taking into account the isospin dependence in isotope chains in addition to the mass number dependence of @xmath0 for nucleons in the iqmd model calculation . |
from last two decades or more , the exploration of neutron - rich nuclei is an important branch in nuclear physics research .
it is a source of observance of new phenomena and dynamics .
this is possible due to the development of accelerator techniques for radioactive ion beams ( ribs ) in various laboratories around the globe .
experimental methods and theoretical analysis have been widely used to collect information about the structure information , such as nuclear size , valence nucleon distribution and halo structure of these exotic nuclei .
the measurement of various reaction observables like total reaction cross - section @xmath3 , one- and two- nucleon removal cross - section ( @xmath2 , @xmath4 ) and the longitudinal momentum distribution @xmath5 are some of the established quantities for such studies . the relativistic mean field ( rmf ) or the effective field theory motivated rmf ,
i.e. , the e - rmf provides the internal structure or sub - structure information of the nuclei through the density distributions , using as an input while calculating the observables in conjunction with glauber model @xcite .
a systematic study of one- and two - neutron knockout data for @xmath6c , explained beautifully while using shell model , which gives a consistent structure information not only for the stable nuclei but also neutrons at the boundary @xcite .
it is well known that narrow fragment momentum distribution reflects large space distribution of the valence nucleon and there is a correlation of the magnitude of the @xmath2 with the width of the @xmath5 , in approaching the nucleon ( neutron or proton ) drip - lines .
however , one - neutron removal reaction cross - sections provide important nuclear structure information complementary to that obtained from @xmath5 . in the present paper ,
our aim is to calculate the @xmath3 and @xmath2 by using densities obtained from the rmf and e - rmf formalisms @xcite in conjunction with glauber model .
it will be shown for the nuclei near the drip - line which possess a halo - structure fails to be explained by the standard evaluation of @xmath2 .
contrary to this estimation , the difference in total reaction cross - section between two consecutive neighboring nuclei in an isotopic chain better matches with the experimental data .
the use of rmf and e - rmf formalisms for finite nuclei as well as infinite nuclear matter are well documented and details can be found in @xcite and @xcite respectively .
the working expressions for density profiles and other related quantities are available in @xcite .
the details to calculate @xmath3 using glauber approach has been given by r. j. glauber @xcite .
this model is based on the independent , individual nucleon - nucleon ( @xmath7 ) collisions in the overlapping zone of the colliding nuclei , and has been used extensively to explain the observed total nuclear reaction cross - sections for various systems at high energies .
the standard glauber form for the total reaction cross - sections at high energies is expressed as @xcite : @xmath8db \;,\ ] ] where @xmath9 is the transparency function with impact parameter @xmath10 .
the function @xmath9 is calculated in the overlap region between the projectile and the target assuming the interaction is formed from a single @xmath7 collision .
it is given by @xmath11 \;.\ ] ] the summation indices @xmath12 and @xmath13 run over proton and nucleon and subscripts @xmath14 and @xmath15 referred to projectile and target , respectively .
the experimental nucleon - nucleon reaction cross - section @xmath16 varies with energy .
the @xmath17-integrated densities @xmath18 are defined as @xmath19 with @xmath20 .
the argument of @xmath9 in eq .
( 2 ) is @xmath21 , which stands for the impact parameter between the @xmath22 and @xmath23 nucleons .
the original glauber model was designed for high energy approximation .
however , it was found to work reasonably well for both the nucleus - nucleus reaction and the differential elastic scattering cross - sections over a broad energy range @xcite . to include the low energy effects of @xmath7 interaction , the glauber model
is modified to take care of the finite range effects in the profile function and coulomb modified trajectories@xcite .
the modified @xmath9 is given by @xcite , @xmath24 \overline{\rho}_{pi } \left ( \vec{t } \right ) \overline{\rho } _ { tj } \left ( \vec{s}\right ) d\vec{s}d\vec{t } \right ] .
\label{eq:2}\ ] ] the profile function @xmath25 is defined as @xcite @xmath26 with @xmath27 , @xmath28 and @xmath29 are the dummy variables for integration over the @xmath17-integrated target and projectile densities . the parameters @xmath30 , @xmath31 , and @xmath32 are usually case - dependent ( proton - proton , enutron - neutron or proton - neutron ) , but we have used the appropriate average values from refs .
the deformed or spherical nuclear densities obtained from the rmf and e - rmf models are fitted to a sum of two gaussian functions with suitable co - efficients @xmath33 and ranges @xmath34 chosen for the respective nuclei which is expressed as @xmath35.\ ] ] then , the glauber model is used to calculate the total reaction cross - section for both the stable and unstable nuclei considered in the present study . the expression for one nucleon removal reaction cross - section @xmath36 is given by @xcite @xmath37 where @xmath38 are the possible final states @xmath39 . in the present formalism , it is considered that the projectile nucleus breaks up into a core and the removed nucleon .
the core @xmath40 has an internal wave function @xmath41 and the one - nucleon , i.e. , the departed nucleon has an asymptotic momentum @xmath42 in the continuum state with respect to the core .
the core is considered to be in the ground state ( @xmath43 ) at the time of the collision .
the total @xmath36 can be separated to an elastic ( @xmath44 ) with @xmath45 and inelastic ( @xmath46 ) part having @xmath47 as non - zero .
the @xmath44 and @xmath46 is expressed as @xcite @xmath48 @xmath49 here @xmath50 is the phase shift function and @xmath51 is the valence wave function ( the wavefunction of the removed nucleon ) .
the notation and the numerical procedure of calculation of one - nucleon removal reaction cross - section are followed from ref .
we obtain the field equations for nucleons and mesons from the rmf and e - rmf lagrangian . for deformed case ( rmf only ) , these equations are solved by expanding the upper and lower components of the dirac spinners and the boson fields in an axially deformed harmonic oscillator basis .
the set of coupled equations are solved numerically by a self - consistent iteration method taking different inputs of the initial deformation @xmath52 @xcite . for spherical densities both for rmf and e - rmf models ,
we follow the numerical procedure of refs . @xcite . in our calculation , the constant gap bcs pairing is used to add the pairing effects for open shell nuclei . the centre - of - mass motion ( c.m . )
energy correction is estimated by the usual harmonic oscillator formula @xmath53 .
the spherical proton ( @xmath54 ) and neutron ( @xmath55 ) density obtained from rmf ( nl3 ) parameter set for various isotopes of ( a ) carbon and ( b ) boron . ] comparing the binding energy ( be ) of the calculated solutions , the maximum be and the corresponding densities [ @xmath54 ( proton ) and @xmath55 ( neutron ) ] are for the ground state . all other solutions are the excited intrinsic state including the spherical one . since the main input in the glauber model estimation is the rmf or e - rmf densities , it is important to have an information of these quantities .
we have plotted the spherical @xmath54 and @xmath55 for both proton and neutron of carbon and boron isotopes in figure 1 using rmf ( nl3 ) parameter set @xcite . as expected , we find an extended density distribution for proton compared to neutron in case of @xmath56c and @xmath57b due to the proton - rich nature of these two nuclei .
the value of @xmath55 and @xmath54 are almost similar for @xmath1c which can be seen from figure 1 .
extension of @xmath55 is much more than @xmath54 for rest of the nuclei .
it is maximum for @xmath58c and @xmath59b in carbon and boron isotopic chains , respectively , because of the high neutron to proton ratio for these cases . [
cols="^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^ " , ] [ table 2 ] unlike to the total reaction cross - section , the @xmath2 obtained from the glauber model , depends very much on the structure information of the projectile and target nuclei , i.e. , input densities of these systems .
the nuclear single - particle energy @xmath60 for the last occupied orbit is very important for a reaction process .
thus , it is worthwhile to analyse the @xmath60 of the valence nucleon of the projectile and target nuclei . for simplicity ,
the spherical single - particle energy for the last occupied orbit for proton @xmath61 and neutron @xmath62 with rmf ( nl3 ) and e - rmf ( g2 ) are compared .
as expected , a small variant in @xmath61 or @xmath62 makes a remarkable change in @xmath63 [ eq .
( 10 ) ] for many cases . for example , the one neutron removal reaction cross - section @xmath63 , for @xmath64 are 55 and 39 mb for rmf ( nl3 ) and e - rmf ( g2 ) with their single - particle energies @xmath65 , @xmath66 and @xmath67 , @xmath68 mev , respectively .
this discrepancy is minimum in the calculation of @xmath36 [ eq . ( 7 ) ] with a lone exception for @xmath69 system .
as the valence @xmath60 plays a major role to determine the reaction observables , one needs to reproduce these values with the experimental observation .
this can be achieved by a small adjustment of the parameters in the relativistic mean field formalisms .
however , the philosophy of rmf or e - rmf of single set of parametrization for the entire domain of nuclear landscape goes against this parameter fiddling .
keeping this in mind , the quality of the results is compromised slightly using with the original values of nl3 , nl - sh or g2 sets . apart from the single - particle energy , the structure effect of the participating nuclei is crucial for a reaction study . in this context
, it is interesting enough to see this effect ( deformation effect ) on @xmath2 .
we repeat the calculations for @xmath36 and @xmath63 with the deformed densities ( rmf only ) as input in the glauber model @xcite .
we obtain spherical equivalent of the axially deformed densities using equations ( 3 ) and ( 6 ) following the prescription of refs . @xcite .
the nl - sh parameter set @xcite for this purpose is used and the results are listed in table 2 .
the reason to change the nl3 to nl - sh is the unavailability of converged ground state deformed solution with nl3 for very light mass nuclei , which is a situation in the present study @xcite .
also , the nl - sh parametrization is reasonably a better parameter set and we expect similar outcome from this force parameter .
due to similar reason as mentioned for the spherical nuclei , the deformed densities for some selected cases are imperative to analyse .
our earlier work on density study supply us enough signature about the complicated sub - structure @xcite .
the clustering and sub - structure of these deformed neutron and proton density distributions are demonstrated in figure 2 .
the density contours presented are in boxes of width and height 6 fm .
a uniform contour spacing of 0.01 @xmath70 is used for proton and neutron densities .
the z - axis is chosen as the symmetry axis , the densities are evaluated in the @xmath71 plane , where @xmath72 . in ref .
@xcite , it is noticed that @xmath1c possesses a @xmath73cluster with a tetrahedral configuration .
the same structure is reproduced in the present study with an oblate shape .
the structure of the neutron deficient @xmath74c nucleus has a prolate ground state and that of the neutron - rich @xmath58c has an oblate ground state deformation . in all the three cases ,
the density plot informs that the central part of the nucleus is a compact core , which is surrounded by a thin layer of nucleons .
the structure of the internal core for both proton and neutron have different density distribution from @xmath56c to @xmath58c .
the shape of @xmath58c proton density distribution looks like a perfect dumb - bell .
thus , it has a maximum probability to to have the structural effects on neutron removal reaction . on the other hand ,
the total nuclear reaction cross - section is less influenced by deformation , may be because of the averaging in input density in the glauber model calculations .
the results obtained from the deformed densities are tabulated ( see table 2 ) .
table 2 shows that most of the @xmath36 [ obtained from eq .
( 7 ) ] matches quite well with the experimental data of @xcite and few of them do not agree .
on the other hand @xmath2 evaluated from eq .
( 10 ) coincide with only @xmath75 experimental data @xcite . of particular interest amongst the nuclei investigated here are @xmath76 and @xmath77 which based on the relatively weak binding of the valence neutrons and measurements of one neutron removal cross - sections have suggested to be one neutron halo systems @xcite .
the single - particle energy for proton @xmath61 and neutron @xmath62 for the last occupied orbit are given in the 11th and 13th column of table 2 .
the last proton for @xmath78 and @xmath56c and the outer most neutron for @xmath79 and @xmath80 are loosely bound which are the possible candidates for either proton - halo ( or skin ) or neutron - halo ( skin ) .
going back to the analysis of figure 2 , which gives us enough indication for the absence of halo - like structure in @xmath81c .
contrary to the case of @xmath82 and @xmath83 a thin - layer of neutron distribution spread spatially to a large extent in case of @xmath77 , which looks like a halo - nucleus .
this behavior is also reflected in the one neutron removal reaction cross - section . in this particular case of @xmath58c , the calculated result @xmath63 = 263 mb is more closure to the experimental value of @xmath84 mb than the @xmath2 obtained by using eqn .
( 7 ) . if we recall the statement of abu - ibrahim et al .
@xcite , then @xmath77 is a halo - nucleus . on the other hand , the measured one neutron removal
reaction cross - section of @xmath76 is larger than its neighbors suggest the weak binding of the last neutron and extended valence density distribution .
summarising the whole discussions of tables 1 and 2 , in general , one can say it very briefly that except few cases like @xmath85c+@xmath1c and @xmath86b+@xmath1c the spherical density used from rmf ( nl3 ) and e - rmf ( g2 ) fails to reproduce the data .
when we use the deformed densities to evaluate the one neutron cross - section , the predicted @xmath36 matches reasonably well with the experimental measurement . in this case , only the result of the system @xmath58c+@xmath1c deviate from the observation .
however , the results predicted by eq . ( 10 ) disagree largely with the experiments irrespective of the densities used .
complementating to eq .
( 7 ) , the @xmath87 matches with the lone case @xmath58c+@xmath1c agreeing with the prediction of ref .
@xcite . the axially deformed density distribution for @xmath88c with rmf ( nl - sh ) parameter set . ] the energy dependence of the neutron removal cross - section for @xmath77 + @xmath83 system using spherical densities of rmf ( nl3 ) , rmf ( nl - sh ) and e - rmf ( g2 ) parameter sets for both elastic and inelastic processes .
the result obtained by deformed rmf ( nl - sh ) densities is also given for the comparison . ] in figure 3 , we have presented the @xmath36 with various incident energies for @xmath75 using the spherical nl3 , nl - sh and g2 densities in the glauber model calculation .
we also compare our results with the deformed nl - sh densities obtained from the axially deformed rmf .
all the spherical densities reproduce similar elastic @xmath89 and inelastic @xmath90 one neutron removal reaction cross - section .
the deformed nl - sh densities has a large impact on the evaluation of @xmath2 unlike to the total nuclear reaction cross - section @xmath3 , which is evident from figure 3 and consistent with @xmath36 .
the deformed @xmath46 value constantly over - estimate the spherical @xmath46 starting from low to very high incident energy of the projectile .
in summary , one neutron removal reaction cross - sections for the neutron - rich isotopes have been calculated using the densities obtained from rmf ( nl3 ) and e - rmf ( g2 ) formalisms for spherical and deformed nl - sh parameter sets .
the dependence of @xmath2 on single - particle energy of the last occupied nucleon is seen in our present calculations .
that means , although the total nuclear reaction cross - section does not show a significant difference , the @xmath2 values differ from each other depending on the nl3 or g2 parameter set .
the @xmath2 are in good agreement with the experiments , when we consider the deformation effect in the densities .
the glauber model fails for halo systems and in this case @xmath58c+@xmath1c is a typical example . in such case ,
the difference between the total reaction cross - section from the consecutive nuclei is applicable to evaluate @xmath63 .
it is also concluded in the present paper that the deformation effects for one neutron removal cross - section is very much crucial unlike to the total reaction cross - section @xmath3 . in other words ,
the glauber model reproduce the experimental data pretty well while considering the deformed densities for stable nuclei as projectile . on the other hand ,
when we estimate the difference of reaction cross - section of nuclei with mass number a and that of a-1 in an isotopic chain , we get good agreement with the experimental data for halo cases .
we are thankful to dr . birbikram singh and mr .
m. bhuyan for the fruitful discussions .
this work is supported in part by council of scientific @xmath91 industrial research ( no.03 ( 1060 ) 06/emr - ii ) , as well as the department of science and technology , govt . of india ,
project no .
sr / s2/hep-16/2005 .
99 a. shukla , b. k. sharma , r. chandra , p. arumugam and s. k. patra , phys . rev .
* c 76 * , 034601 ( 2007 ) ; b. k. sharma , s. k. patra , raj k. gupta , a. shukla , p. arumugam , p. d. stevenson and greiner walter , j. phys .
* g32 * , 2089 ( 2006 ) .
s. k. patra and r. n. panda , arxiv:0906.3797 ; s. k. patra , r. n. panda , p. arumugam and raj k. gupta , phys .
* c 80 * , 064602 ( 2009 ) .
e. c. simpson and j. a. tostevin , phys . rev .
* c 79 * , 024616 ( 2009 ) .
s. k. patra and c. r. praharaj , phys . rev .
* c 44 , * 2552 ( 1991 ) ; b. d. serot and j. d. walecka , adv .
* 16 * , 1 ( 1986 ) ; p. ring , prog .
. phys . * 37 * , 193 ( 1996 ) .
g. a. lalazissis , j. knig and p. ring , phys . rev .
* c 55 * , 540 ( 1997 ) .
r. j. furnstahl , b. d. serot and h. b. tang , nucl . phys . * a615 * , 441 ( 1997 ) .
m. del estal , m. centelles , x. vias and s. k. patra , phys . rev .
* c 63 * , 044321 ( 2001 ) ; s. k. patra , m. del estal , m. centelles , x. vias , phys . rev .
* c 63 * , 024311 ( 2001 ) ; b. d. serot and j. d. walecka , int . j. mod .
phys . * e6 * , 515 ( 1997 ) .
r. j. glauber , _ lectures on theoretical physics _
, edited : w. e. brittin and l. c. dunham ( interscience , new york ) , * vol.1 * , p.315 ( 1959 ) .
p. j. karol , phys . rev .
* c 11 * , 1203 ( 1975 ) .
j. chauvin , d. lubrun , a. lounis and m. buenerd , phys . rev .
* c 28 * , 1970 ( 1983 ) ; m. buenerd , a. lounis , j. chauvin , d. lebrun , p. martin , g. duhamel , j. c. gondrand and p. d. saintignon , nucl . phys . * a 424 * , 313 ( 1984 ) .
p. shukla , phys . rev . * c 67 * , 054607 ( 2003 ) . b. abu - ibrahim , y. ogawa , y. suzuki and i. tanihata , comp .
. comm . * 151 * , 369 ( 2003 ) .
a. bhagwat and y. k. gambhir , phys . rev . * c 77 * , 027602 ( 2008 ) ; _ ibid _ * c 73 * , 054601 ( 2006 ) ; _ ibid _ * c 73 * , 024604 ( 2006 ) ; _ ibid _ * c 69 * , 01315 ( 2004 ) ; _ ibid _ * c 68 * , 044301 ( 2003 ) .
s. k. charagi and s. k. gupta , phys .
* c 41 * , 1610 ( 1990 ) ; s. k. charagi , phys . rev . * c 48 * , 452 ( 1993 ) ; s. k. charagi and s. k. gupta , phys .
* c 46 * , 1982 ( 1992 ) ; s. k. charagi and s. k. gupta , phys . rev . * c 56 * , 1171 ( 1997 ) .
y. k. gambhir , p. ring and a. thimet , ann .
( n.y ) * 198 * , 132 ( 1990 ) .
d. cortina - gil d et al . 2001
j. * a10 * , 49 ( 2001 ) .
e. sauvan et al . , phys . rev . * c 69 * , 044603 ( 2004 ) ; e. sauvan
et al . , phys . lett . * b1 * , 491 ( 2000 ) .
wu cuie et al . , j. phys .
* g31 * , 47 ( 2005 ) .
d. l. olson et al . , phys . rev . * c 28 * , 1602 ( 1983 ) . m. m. sharma , m. a. nagarajan and p. ring , phys . lett .
* b312 * , 377 ( 1993 ) .
reinhard , rep .
52 * , 439 ( 1989 ) .
arumugam , b. k. sharma , s. k. patra and raj k. gupta , phys . rev .
* c 71 * , 064308 ( 2005 ) .
d. bazin et al . , phys . rev .
* c 57 * , 2156 ( 1998 ) . | we calculate the one neutron removal reaction cross - section @xmath0 for few stable and neutron - rich halo nuclei with @xmath1c as target , using relativistic mean field ( rmf ) densities , in the frame work of glauber model .
the results are compared with the experimental data .
study of the stable nuclei with the deformed densities have shown a good agreement with the data , however , it differs significantly for the halo nuclei .
we observe that while estimating the @xmath2value from the difference of reaction cross - section of two neighboring nuclei with mass number a and that of a-1 in an isotopic chain , we get good agreement with the known experimental data for the halo cases . |
g. acknowledges suppoort of the austrian science fund through the grant p19168-n16 . | at a critical finite chemical potential and low temperature qcd undergoes the chiral restoration phase transition .
the folklore tradition is that simultaneously hadrons are deconfined and there appears the quark matter .
we demonstrate that it is possible to have confined but chirally symmetric hadrons at a finite chemical potential and hence beyond the chiral restoration point at a finite chemical potential and low temperature there could exist a chirally symmetric matter consisting of chirally symmetric but confined hadrons .
if it does happen in qcd , then the qcd phase diagram should be reconsidered with obvious implications for heavy ion programs and astrophysics .
it is generally believed that chiral and deconfinement phase transitions in qcd coincide and hence beyond the semi - circle in the @xmath0 plane one obtains a deconfining and chirally restored matter . at zero temperature and density the t hooft anomaly matching conditions
require that in the confining phase chiral symmetry must be broken in the vacuum .
however , there is no such a restriction at a finite chemical potential .
typically models that give us information about the phase structure in qcd are of the nambu and jona - lasinio type .
these models are not confining , however , and consequently there is no basis to conclude that above the chiral restoration point at finite chemical potential one obtains a chirally symmetric and deconfining quark matter .
recently mclerran and pisarski presented qualitative large @xmath1 arguments showing that at reasonably large chemical potential and low temperature there might exist a confining but chirally symmetric phase @xcite .
this suggestion is in conflict with the naive intuition that once the hadrons are confined chiral symmetry should be broken .
here we demonstrate that it is not so and that it is possible to have confined but chirally symmetric hadrons at finite density and low temperature @xcite .
there exists only one known manifestly chirally - symmetric and confining model in four dimensions that is solvable @xcite .
this model can be considered as a generalization of the 1 + 1 dimensional t hooft model , that is qcd in the large @xmath1 limit .
once the gauge is properly chosen in 1 + 1 dimensions the coulomb interaction becomes a linear confining potential .
then this potential properly represents gluonic degrees of freedom in 1 + 1 dimensions .
it is postulated that there exists a linear confining potential of the coulomb type in four dimensions either .
this model represents a simplification of large @xmath1 qcd in four dimensions .
the model is exactly solvable .
consider first chiral symmetry breaking in the vacuum .
the dirac operator for the dressed quark is @xmath2 where @xmath3 is the bare dirac operator . parametrising the self - energy operator in the form @xmath4 , \label{se}\ ] ] where functions @xmath5 and @xmath6 are yet to be found , the schwinger - dyson equation for the self - energy operator in the rainbow approximation , which is valid in the large @xmath1 limit for the instantaneous interaction , is reduced to the nonlinear gap equation for the chiral angle @xmath7 , @xmath8 where @xmath9 here @xmath10 is the fourier transform of the linear confining potential with the string tension @xmath11 with a proper infrared regularisation .
the functions @xmath12 i.e. the quark self - energy , are divergent in the infrared limit , which implies that the single quark can not be observed and the system is confined .
however , the infrared divergence cancels exactly in the gap equation so this equation can be solved directly in the infrared limit .
the chiral symmetry breaking is signalled by the nonzero quark condensate and by the dynamical momentum - dependent mass of quarks @xmath13 the dynamical mass is finite at small momenta and vanishes at large momenta .
both these quantities were first obtained in ref .
@xcite and repetedly reconfirmed in all subsequent works on this model .
the numerical value of the quark condensate is @xmath14 .
now we are in position to include into this model a finite quark chemical potential @xmath15 at zero temperature . denoting the fermi momentum of quarks as @xmath16 one has to remove from the integration in the gap equation all quark momenta below @xmath16
since they are pauli - blocked .
this is similar to what is done within the njl type models or within the t hooft model @xcite . at
the critical quark chemical potential one observes a chiral restoration phase transition , as depicted in fig . 1 .
what crucially distinguishes this model from the njl model is that the system is still confined , even in the chirally restored phase .
this is because the self - energy integral @xmath6 is still infrared - divergent , even when the chiral angle @xmath7 and dynamical mass @xmath17 identically vanish .
hence the single quark is removed from the spectrum at any chemical potential . to demonstrate this explicitly
one has to solve the bethe - salpeter equation for the quark - antiquark bound states applying the quark green function obtained from the gap equation .
the infrared divergence of the single - quark green function cancels exactly in the color - singlet quark - antiquark system @xcite and the bound state mesons are finite and well defined quantities .
the spectrum below the critical chemical potential is similar to the one previously obtained in the vacuum @xcite .
the spectrum exhibits approximate restoration of the chiral symmetry in excited hadrons , for details we refer to @xcite and for a review to ref . @xcite . the spectrum above the critical chemical potential , i.e. for @xmath18 , is shown in fig .
2 .
this spectrum is qualitatively different .
all the states are in _ exact _ chiral multiplets . though the chiral symmetry is manifestly restored in the vacuum , one observes _
finite - energy well defined hadrons_. obviously the mass generation mechanism in these hadrons has nothing to do with the chiral symmetry breaking in the vacuum and is not related with the quark condensate .
the mass generation mechanism for these chirally symmetric hadrons is similar to the high - lying states in the chiral symmetry broken phase .
we have demonstrated that it is possible to have a confining but chirally symmetric matter consisting of chirally symmetric hadrons at finite density .
whether such a chirally symmetric but confining phase exists in qcd or not is still an open question , but if it does , then it will imply dramatic modifications of the qcd phase diagram .
it will also have significant implications for astrophysics : the interactions of these chirally - symmetric hadrons can be only of short - range as they decouple from the goldstone bosons and their weak decay rate is quite different since their axial charge vanishes @xcite . |
light quark masses are important unknown parameters of the standard model , and a number of lattice qcd calculations have been carried out to evaluate quark masses employing the wilson , clover or kogut - susskind ( ks ) fermion action @xcite . among the results ,
those with the ks action appear more accurate than others because of small lattice discretization errors and small statistical errors . a worry with the ks result
, however , has been that the employed one - loop renormalization factor takes a large value of @xmath8 in the range of @xmath9 studied , calling into question the viability of perturbation theory . in this article
we report a study to circumvent this problem : we calculate the renormalization factor of bilinear quark operators for the ks action non - perturbatively using the method of ref .
@xcite developed for the wilson / clover actions .
this calculation is carried out in quenched qcd at @xmath0 , @xmath1 , and @xmath2 on an @xmath10 lattice .
the results , combined with our previous calculation of bare quark masses @xcite , lead to a non - perturbative determination of the light quark mass .
the renormalization factor of a bilinear operator @xmath11 is obtained from the amputated green function , @xmath12 where the quark two - point function is defined by @xmath13 .
the quark field @xmath14 with momentum @xmath15 is defined from the original one - component field @xmath16 by @xmath17 , where @xmath18 , @xmath19 ( @xmath20 ) and @xmath21 $ ] . the renormalization condition imposed on @xmath22
is given by @xmath23 where @xmath24 is the amputated green function at tree level and @xmath25 is the wave function renormalization factor which can be calculated by the condition @xmath26 for the conserved vector current .
the relation between the bare operator on the lattice and the renormalized operator in the continuum takes the form , @xmath27 where @xmath28 is the renormalization - group running factor , and @xmath29 is the matching factor from the @xmath30 scheme defined by ( [ eq : ri ] ) to the @xmath31 scheme , calculated perturbatively in the continuum .
for the light quark mass we apply relation ( [ zzz ] ) in the scalar channel in the chiral limit .
we use a source in momentum eigenstate to evaluate quark propagators .
this results in very small statistical errors of @xmath32 in the green functions .
the external momentum @xmath15 should be taken in the range @xmath33 in order to keep under control higher order effects in continuum perturbation theory , non - perturbative hadronization effect on the lattice , and discretization errors on the lattice . in this work
we choose @xmath34 momenta in the range @xmath35 for all values of @xmath9 .
= 7.0 cm = 7.0 cm
in fig . [ fig1 ] we compare the scalar renormalization factor @xmath36 with that for pseudoscalar @xmath37 for three values of bare quark mass @xmath38 at @xmath0 . from chiral symmetry of ks fermions
, we expect naively @xmath39 for all momenta @xmath15 in the chiral limit
. clearly this relation does not hold with our result toward small momenta , where @xmath37 rapidly increases as @xmath40 , while @xmath41 does not show such a trend . to understand this result
, we note that chiral symmetry of ks fermion leads to the following identities : @xmath42 with @xmath43 $ ] . in fig .
[ fig5 ] @xmath44 in the chiral limit obtained by a linear extrapolation in @xmath45 is plotted .
it rapidly dumps for large momenta , but takes large values in the small momentum region .
combined with ( [ zp ] ) this implies that @xmath37 diverges in the chiral limit for small momenta , which is consistent with the result in fig .
[ fig1 ] .
the function @xmath44 is related to chiral condensate as follows : @xmath46 = \sum_p \frac { m(p ) } { c_\mu(p)^2 + m(p)^2 } \ ] ] where @xmath47/ \cos(p_\mu a)$ ] .
a non - vanishing value of @xmath44 for small momenta would lead to a non - zero value of the condensate .
therefore the divergence of @xmath48 near the chiral limit is a manifestation of spontaneous symmetry breakdown of chiral symmetry ; it is a non - perturbative effect arising from the presence of massless goldstone boson .
we expect this non - perturbative effect to affect the scalar renormalization factor @xmath41 much less , since the scalar operator can not interact directly with the pseudoscalar meson .
indeed the quark mass dependence is quite small as we have seen in fig .
[ fig1 ] . in fig .
[ fig8 ] we show the momentum dependence of the ratio @xmath49 calculated in the chiral limit where we set @xmath50 and use the three - loop formula @xcite for @xmath51 and @xmath52 .
while the ratio should be independent of the quark momentum @xmath15 , our results show a large momentum dependence which is almost linear in @xmath53 for @xmath54 .
a natural origin of the linear dependence on @xmath53 is the lattice discretization error of the scalar operator , which differs by terms of @xmath55 from that of the continuum for the ks fermion .
we then remove this error from the renormalization factor by a linear extrapolation in @xmath53 to @xmath56 . in fig .
[ fig8 ] the fitting lines are plotted , where filled data points are used for the linear extrapolation . for comparison ,
the ratio calculated with the one - loop value equals @xmath57 , @xmath58 , and @xmath59 for @xmath0 , @xmath1 and @xmath2 at @xmath5 .
hence one - loop perturbation theory underestimates the ratio by @xmath60 to @xmath61 .
our final results for the averaged up and down quark mass at @xmath50 are shown in fig .
[ fig9 ] by filled symbols .
here we use the jlqcd results for bare quark mass @xcite .
the values are substantially larger than those obtained with one - loop perturbation theory ( open circles for @xmath5 and squares for @xmath7 ) .
furthermore they exhibit a significant @xmath62 dependence , which we ascribe to the discretization error of the quark mass itself . making a linear extrapolation in @xmath62 , our final result in the continuum limit
is given by @xmath63 this value is @xmath61 larger than the perturbative estimates : @xmath64 for @xmath5 and @xmath65 for @xmath7 .
this work is supported by the supercomputer project no.32 ( fy1998 ) of high energy accelerator research organization ( kek ) , and also in part by the grants - in - aid of the ministry of education ( nos . 08640404 , 09304029 , 10640246 , 10640248 , 10740107 , 10740125 ) .
s.k . and s.t .
are supported by the jsps research fellowship .
9 r. gupta and t. bhattacharya , nucl .
phys . * b*(proc .
suppl.)*63 * ( 1998 ) 95 .
g. martinelli _ et.al .
b445 * ( 1995 ) 81 .
e. franco and v. lubicz , hep - ph/9803491 ; t. van ritbergen _ et.al . _ , phys .
* b400 * ( 1997 ) 379 ; j.a , m , vermaseren _ et.al .
_ , phys . lett .
* b405 * ( 1997 ) 327 ; k.g .
chetyrkin , phys .
b404 * ( 1997 ) 161 ; jlqcd collaboration , nucl .
* b*(proc .
* 53 * ( 1997 ) 209 . | light quark masses are computed for kogut - susskind fermions by evaluating non - perturbatively the renormalization factor for bilinear quark operators .
calculations are carried out in the quenched approximation at @xmath0 , @xmath1 , and @xmath2 .
for the average up and down quark mass we find @xmath3 in the continuum limit , which is significantly larger than @xmath4 ( @xmath5 ) or @xmath6 ( @xmath7 ) obtained with the one - loop perturbative renormalization factor . |
low - mass m stars are among the most numerous stellar type in the galaxy and also among the most long lived .
this would make them attractive candidates in the search for life beyond earth were it not for the possibility of atmospheric loss due to synchronous rotation . the ongoing discovery of extrasolar planets around low - mass stars @xcite has renewed discussion of the habitability of planets orbiting m type stars @xcite and provides additional motivation for understanding the atmospheric dynamics of such planets .
the observation of extrasolar planets around stars of all types has increased interest in characterizing planetary atmospheres through climate modeling .
these include general circulation model ( gcm ) simulations of atmospheric circulation on large , jupiter - sized planets in close orbit to their parent star known as hot jupiters " @xcite , as well as consideration of smaller , neptune - sized gas giants and earth - sized terrestrial planets @xcite .
the discovery of planets orbiting the m type star gliese 581 also prompted an assessment of the habitability of these new worlds that include radiative climate calculations of gliese 581c @xcite .
( note that radiative and gcm studies also were conducted for the suspected planets gliese 581d @xcite and gliese 581 g @xcite , which have now been determined to be artifacts due to stellar activity @xcite . )
these modeling studies help to identify prominent circulation features in the atmospheres of extrasolar planets and characterize the types of atmospheres that could exist in orbit around distant stars one way to characterize planetary habitability is the liquid water habitable zone @xcite , which describes the region around a star where a planet could sustain surface liquid water .
this habitable zone defines a region where the negative climate feedback of the carbonate - silicate cycle @xcite can keep a planet above freezing .
it is bounded on the inner edge by the loss of water through initiation of a runaway greenhouse effect @xcite and on the outer edge by the condensation of carbon dioxide @xcite .
although this habitable zone can not be generalized to all extrasolar planets , it can be used to characterize the habitability of any terrestrial planet with water and a carbonate - silicate cycle .
( some authors have extended consideration of habitability to scenarios such as a dry planet with limited water @xcite or a planet rich in atmospheric hydrogen @xcite . ) for low - mass m type stars , this liquid water habitable zone falls within the tidal locking orbital distance of its parent star , so any potentially habitable planets around these stars ( according to this definition of habitability ) are at risk of atmospheric loss and may in fact be uninhabitable .
the problem of atmospheric stability for planets orbiting low - mass stars has been investigated by @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite @xcite , @xcite , and @xcite using gcms to assess the degree to which atmospheric circulations can maintain planetary habitability .
these simulations show that a range of planetary atmospheres can remain stable under synchronous rotation , either by a sufficiently dense carbon dioxide atmosphere @xcite or in regimes where atmospheric dynamics provides sufficient heat transport to the night side @xcite .
these gcm studies demonstrate that extrasolar planets orbiting low - mass stars can indeed provide habitable conditions at locations on the surface , even with the problem of synchronous rotation .
in this study we use an idealized gcm to investigate the influence of geothermal heating on atmospheric dynamics .
we present a set of experiments that include fast and slow rotators to show that geothermal heating amplifies asymmetries in atmospheric circulation between the northern and southern hemispheres , which can cause transient regions of warming near the anti - stellar point .
we show that the dynamical structure of the atmosphere changes on either side of the sub - stellar point , and we find that a cross - polar circulation provides transport from the day to night side in all experiments .
these results provide qualitative descriptions of the expected meridional , zonal , and cross - polar circulation patterns that can help to guide further gcm studies of synchronously rotating planets .
a planet can fall into synchronous rotation around its parent star when tidal dissipation slows down the planet s rotation rate .
not all planets orbiting at this tidal locking distance from the parent star will necessarily become synchronous rotators , such as those with eccentric orbits or with close neighbors so that they fall into spin - orbit resonances .
for example , the planet mercury is within the tidal locking distance of our sun , but it has fallen into a 3:2 spin - orbit resonance that remains stable because of its nonzero eccentricity .
planets with more circular orbits , however , may evolve into synchronous rotators so that a single side is fixed at perpetual daylight , while the opposing side experiences perpetual night .
the tidal migration history of any particular planet depends on factors such as the initial rate of rotation , eccentricity , and semi - major axis @xcite , which complicates efforts to determine a particular orbital distance at which a planet should be expected to fall into a synchronously rotating state .
no single analytic equation can describe the tidal locking radius for all planetary systems , and empirical formulae may be limited in their predictive power . in this section
we use such an empirical equation to demonstrate limits on stellar mass and rotation rate for these systems ; however , we admit that this approach is simplistic and neglects many key factors that contribute to the actual tidal history of a planetary system . the tidal locking distance @xmath0 can be described empirically @xcite according to the formula @xmath1 where @xmath2 is the mass of the parent star , @xmath3 is the original rotation period of the planet , @xmath4 is the time period since planetary formation , and @xmath5 is a parameter describing the inverse of the specific dissipation function for the entire planet .
we restrict our focus in this study to synchronously rotating earth - like planets , which allows us to specify the parameters in eq .
( [ eq : tidallocking ] ) .
we assume early earth rotated more rapidly than today so that @xmath6 @xcite , and we likewise use the 4.5 billion year formation period of earth as the value of @xmath4 .
the parameter @xmath5 has a relatively low value on earth today ( @xmath713 ) because most tidal dissipation occurs from gravitational interactions with the sun and moon along oceanic shorelines @xcite .
however , this present - day value may be unusually large @xcite , and different continental configurations on early earth such as a larger ocean @xcite could have reduced the tidal dissipation rate .
we will therefore use a value of @xmath8 as characteristic of early earth @xcite to describe tidal forces on a synchronously rotating earth - like planet .
the value of the tidal locking distance @xmath0 in eq .
( [ eq : tidallocking ] ) is only weakly sensitive to changes in the parameters @xmath3 , @xmath4 , and @xmath5 , so the primary contributor is the mass @xmath2 of stellar types that could induce synchronous rotation for orbiting earth - like planets .
we consider an earth - like planet to be a terrestrial planet that receives the same incident flux of stellar radiation as earth .
the stellar flux @xmath9 reaching a planet at an orbital radius @xmath10 is related to the luminosity of its star @xmath11 according to @xmath12
. we can examine planets in orbit around different stellar types by defining the earth equivalent distance of a planet @xmath13 as the orbital distance at which a planet receives an equal magnitude of radiation from its parent star as earth does from the sun @xcite .
at the earth equivalent distance , the flux equals that on earth , so we can write @xmath14 where @xmath15 is the luminosity of the sun , and @xmath16 is the orbital radius of earth .
if we use the general approximation that stellar luminosity for low - mass stars is proportional to mass to the fourth power @xmath17 , then we can combine eqs .
( [ eq : tidallocking ] ) and ( [ eq : eed ] ) by setting @xmath18 and solving for stellar mass .
letting @xmath19 be the mass of the sun , we find that @xmath20^{3/5}. \label{eq : starmass}\ ] ] substituting in values for the parameters in eq .
( [ eq : starmass ] ) shows that the earth equivalent distance @xmath13 and the tidal locking distance @xmath0 intersect at @xmath21 ; planets in orbit around stars less massive than this will be in synchronous rotation at the earth equivalent distance .
this means that f , g , and k type stars are unlikely to harbor earth - like planets in synchronous rotation , but smaller m type stars are ideal candidates to find such planets .
we use this mass limit to calculate the maximum rotation rate for a synchronously rotating earth - like planet .
kepler s third law gives the relation @xmath22 , where @xmath23 and @xmath24 are the orbital period at the earth equivalent distance and at earth s present distance , respectively .
this corresponds to an upper limit on orbital period ( equivalently , rotation rate ) of @xmath25 for synchronously rotating earth - like planets around m type stars ( c.f .
planets rotating faster than this limit may still be prone to synchronous rotation , while planets rotating slower than this limit will be beyond the tidal locking distance and outside the liquid water habitable zone @xcite . in this study
, we consider synchronously rotating planets with both 1-day ( `` fast '' ) and 230-day ( `` slow '' ) rotation rates to show the similarities and differences that arise in these two dynamical regimes .
we use a general circulation model ( gcm ) to simulate the atmospheres of synchronously rotating earth - like planets and investigate the influence of geothermal heating .
this gcm uses a hydrostatic spectral dynamical core ( with t42 resolution ) , two - stream gray radiative transfer , a diffusive boundary layer scheme , and a surface slab ocean @xcite .
we set obliquity to zero and adjust the incoming solar radiation profile so the sub - stellar point is fixed on the intersection of the equator ( 0@xmath26 latitude ) and international date line ( 180@xmath26 longitude ) .
we also include a geothermal heat flux in the gcm by adding a constant term to the surface temperature tendency equation .
this model is still highly idealized , as the model atmosphere is transparent to incoming stellar radiation and absorbs only infrared radiation emitted upward from the surface .
water vapor is assumed to be radiatively neutral , so this gcm does not include any water vapor feedback , and the model experiments also do not include any topography or cloud formation .
although some of these processes will likely be significant in determining specific features of climate , these idealizations help to highlight large - scale dynamical patterns and make the model an ideal tool for investigating the prominent circulation patterns that arise on synchronously rotating planets .
we consider both dry and moist regimes for our gcm experiments .
we force dry gcm conditions by setting the saturation vapor pressure equal to zero and using dry adiabatic adjustment .
the moist regime includes latent heat release through large - scale condensation , and a simplified shallow penetrative adjustment also known as betts - miller adjustment @xcite . for non - synchronously rotating earth - like planets
, the presence of moisture in this gcm acts to increase static stability and reduce circulation strength @xcite .
these idealized dry and moist configurations are designed to emphasize the key differences in atmospheric dynamics that arise from the presence or absence of water vapor .
we conduct a series of gcm experiments using fast ( 1-day ) and slow ( 230-day ) rotation rates with 25 vertical levels and physical properties identical to those used by @xcite .
we initialize dry and moist control states at each rotation rate with a constant temperature at all grid points and run the gcm for a 3000 day spin - up period to reach a statistically steady state .
geothermal heating is then enabled ( or kept disabled ) , and the model is allowed to continue adjusting for another 500 days .
we then use the subsequent 1000 days of model output in the analysis that follows .
running this gcm using a linux workstation ( 6-core intel e5 - 2620 xeon 2.0 ghz ) requires approximately 8 hours in real - time to calculate 1000 model days .
table [ table : experiments ] lists the eight gcm experiments we consider in this study .
we use the designators `` fast '' and `` slow '' in reference to 1-day and 230-day rotation periods , respectively , and we likewise indicate `` dry '' and `` moist '' cases accordingly .
we use `` control '' to refer to experiments without any geothermal flux , and `` geothermal '' indicates the presence of a 2.0 w m@xmath27 geothermal heat flux across the entire surface .
this set of eight gcm experiments provide an overview of expected large - scale dynamical structures and the resulting perturbations that arise from geothermal heating . [ cols="<,<,<,<,<",options="header " , ]
terrestrial planets in synchronous rotation around low - mass stars experience tidal heating from gravitational interactions with the host star @xcite . within the solar system ,
io exhibits strong tidal interactions as it orbits jupiter , which drives constant volcanism across the moon s surface .
this form of tidal heating also should occur for extrasolar planets in synchronous rotation @xcite and even for some exomoons @xcite .
the magnitude of tidal heating for any particular planet depends on its orbital geometry and tidal migration history @xcite , but some degree of tidal heating should be expected to arise on most synchronously rotating planets .
we consider the effects of tidal heating in the gcm as a geothermal heat flux applied uniformly across all surface grid points .
actual profiles of tidal heating on synchronously rotating planets likely will show latitudinal and longitudinal variations , but we choose this idealized representation as a first - order estimation of the influence of a geothermal heat flux .
we choose a value of 2.0 w m@xmath27 for the geothermal heat flux in our model , which approximately corresponds to the tidal heat flux on io @xcite .
we do not represent any additional internal heat fluxes , such as from radiogenic decay , so these calculations can be considered as an upper limit on the influence of tidal heating on surface temperature and atmospheric circulation .
our gcm experiments include control cases ( no geothermal heat flux ) and geothermal cases ( 2.0 w m@xmath27 geothermal heat flux ) run under dry and moist conditions at both fast and slow rotation rates ( table [ table : experiments ] ) .
[ fig : surftemp ] shows the time average of surface temperature and horizontal wind @xmath28 for the full set of experiments . note that here and elsewhere we show surface temperature as the deviation from the freezing point of water . the top row of fig .
[ fig : surftemp ] shows the control cases , the middle row shows the geothermal cases , and the bottom row shows the difference between the geothermal and control cases .
note that each panel is centered on the anti - stellar point ( at 0@xmath26 latitude and 0@xmath26 longitude ) to emphasize the changes in temperature and wind that occur across the anti - stellar hemisphere .
a noticeable shift in wind patterns occurs between fast and slow rotators , which is an expected regime transition @xcite that will be discussed further in section [ section : dynamics ] .
asymmetries also persist between the northern and southern hemisphere in all cases , which has been noted by others @xcite and also will be discussed further in section [ section : dynamics ] .
the presence of geothermal heating ( fig .
[ fig : surftemp ] , bottom row ) provides warming across the surface as expected ; however , the distribution of warming yields regions near the anti - stellar point with asymmetric warm or cold features between the northern and southern hemispheres that persist even after averaging over 1000 days .
these effects are most pronounced in the fast dry cases , with deviations of @xmath295 k and asymmetric wind patterns around the anti - stellar point .
the presence of moisture appears to reduce the asymmetric influence of geothermal heating , although some asymmetric features between the northern and southern hemispheres still persist .
the slow rotators show the most uniform effect from geothermal heating , with only slight asymmetries in the time average of surface temperature and horizontal wind .
the asymmetric features between the northern and southern hemispheres observed in the time average quantities shown in fig . [
fig : surftemp ] are the result of transient variations in atmospheric dynamics across the anti - stellar hemisphere . the transient evolution of surface temperature and horizontal wind difference ( geothermal minus control ) is shown in fig .
[ fig : surftempevoall ] at 100-day intervals for the full set of experiments .
( animations of transient evolution for all gcm experiments are available online . ) in all cases , geothermal heating shows stronger regions of warming and cooling that are asymmetrically distributed between the northern and southern hemispheres near the anti - stellar point and fluctuate with time .
the fast dry case shows the strongest deviations of @xmath2920 k or more , while the presence of moisture reduces these changes by about a factor of two .
asymmetric features are also prominent in the slow dry case , with surface temperature changes of about @xmath295 k or so and slight reductions in the corresponding fast moist case .
these experiments all suggest that geothermal heating acts to create or amplify asymmetric circulation features between the northern and southern hemispheres that cause transient regions of warming and cooling most prominent near the anti - stellar point .
geothermal heating enhances the persistence of asymmetric features by affecting the formation of large - scale dynamical structures responsible for energy transport . even in a dry atmosphere ,
strong thermally - direct circulations ( discussed further in section [ section : dynamics ] ) are sensitive to changes in the surface energy flux , which alters the transport of mass and energy from the sub - stellar to anti - stellar point and generates a slightly different pattern in the general circulation .
the net effect is that the added energy from uniform geothermal heating is distributed unevenly across the surface through dynamical energy transport , which allows some additional cycles of warming and cooling to persist between the northern and southern hemispheres near the anti - stellar point . even without geothermal heating , asymmetric features between the northern and southern hemispheres
have been observed in other gcm simulations @xcite and appear to be a robust phenomenon of synchronous rotation .
we even attempted to initialize the gcm into a fully symmetric synchronously rotating state without any baroclinic eddies ( following the approach of @xcite for non - synchronous rotation ) , but the model immediately lost its symmetry and reverted to the states presented here .
these asymmetric features seem to persist regardless of the averaging period or sampling variability and do not appear sensitive to other changes in the model configuration .
these results all suggest that the asymmetric features seen in gcms are not artifacts of averaging but are dynamically - induced features of the atmosphere that affect the distribution of surface temperature between the northern and southern hemispheres , particularly near the anti - stellar point .
the atmospheric circulation on synchronously rotating planets is known to exhibit unique three - dimensional behavior that is markedly different from the prominent circulations on earth , such as equatorial superrotation and heat transport from the anti - stellar to sub - stellar point @xcite .
some of this behavior is captured by typical meridional and zonal mass streamfunctions , but we show below that careful consideration of the wind field is required to accurately describe the most prominent means of transport .
we describe the mean meridional circulation ( mmc ) with a mass streamfunction @xmath30 that calculates the northward mass flux above a particular pressure level @xmath31 as @xmath32 this streamfunction traces out the familiar patterns of the mmc when applied to earth models or time - averaged observations , showing direct ( i.e. , hadley ) and indirect ( i.e. , ferrel ) circulation cells . on earth , and other planets not in synchronous rotation ,
the mmc is a primary driver of meridional heat transport from the tropics to poles and results in the formation of a subtropical jet and polar front jet along the edge of the circulation cells @xcite . for planets in synchronous rotation ,
the mmc still acts as a means of meridional heat transport , but the geometry of the circulation depends strongly upon rotation rate . fig .
[ fig : umeanmmc ] shows the mmc @xmath30 and zonal mean zonal wind @xmath33 for the full set of experiments .
the fast rotators all show similar patterns of a hadley , ferrel , and polar cell similar to earth , but the upper atmosphere features strong superrotation in a single eastward equatorial jet with westward flow only at the very top of the model .
the slow rotators show an entirely different mmc configuration , with direct cells elevated above the surface and indirect cells circulating from equator to pole .
the zonal wind for the slow rotators appears to be zero , which we show below results from a change in circulation across hemispheres .
note that the difference between geothermal and control cases ( fig .
[ fig : umeanmmc ] , bottom row ) results in asymmetric structures that are weak in strength but emphasize the role of geothermal heating in creating small asymmetries in circulation between the northern and southern hemispheres .
the strength of the mmc is also stronger in dry cases compared to their moist counterparts , which occurs because the presence of water vapor increases static stability and decreases the magnitude of overturning circulations @xcite .
the mmc shown in fig .
[ fig : umeanmmc ] can be misleading , however , as the circulation pattern changes depending on position relative to the sub - stellar point .
this observation was made by @xcite in application to a simplified model of the walker circulation , where it was noticed that the direction of the mmc changes on either side of the heating source . on earth ,
the walker circulation is a transient phenomenon , but on synchronously rotating planets the star shines fixed on a single location at all times , so this theoretical prediction may be even more relevant .
we therefore may expect that the patterns of circulation in our gcm experiments should change between the hemispheres east and west of the sub - stellar point .
[ fig : umeanmmchemi ] shows the mmc @xmath30 and zonal mean zonal wind @xmath33 separated over the eastern hemisphere and western hemisphere from the sub - stellar point for the set control of experiments .
the fast rotators all show a change in direction of the mmc , with the eastern hemisphere showing circulation in the same direction as in fig .
[ fig : umeanmmc ] and the western hemisphere showing circulation in the opposite direction
. the atmosphere is still superrotating for the fast rotators , although the equatorial jet is much stronger in the eastern hemisphere , and notable differences persist between hemispheres in the geometry of the mmc cells .
by contrast , the slow rotators do not show a change in circulation direction , but instead the wind direction changes from eastward in the eastern hemisphere to westward in the western hemisphere .
slight asymmetries are apparent in the mmc between the northern and southern hemispheres , but the slow rotators appear to maintain a relatively consistent geometry across the sub - stellar point .
the contrast between the zonally averaged quantities in fig .
[ fig : umeanmmc ] and the hemispheric separation made in fig .
[ fig : umeanmmchemi ] provides one example of the circulation features that contribute to persistent atmospheric asymmetries . the zonal circulation , also known as the walker circulation , provides another possible mechanism for transporting energy from the sub - stellar to anti - stellar point . on earth , the walker circulation is characterized by rising motion at the heating source , eastward ( and westward ) flow aloft as air descends , and return flow along the surface back toward the heating source @xcite .
we describe the mean zonal circulation ( mzc ) with a mass streamfunction @xmath34 that calculates the eastward mass flux above a particular pressure level @xmath31 as @xmath35 where @xmath36 is the zonal average value minus the time - average component of zonal wind .
the extent of this circulation is limited by the propagation of equatorially trapped rossby waves and equatorial kelvin waves @xcite , which provides an explanation for the regime transition in wind patterns observed between fast and slow rotators in fig .
[ fig : surftemp ] .
@xcite describes a change in patterns of winds that occurs when the rossby deformation radius ( which is inversely proportional to rotation rate ) begins to approach the planetary radius at a rotation rate of about 4.2 days .
this regime shift has also been noted by others @xcite and is evident in many different gcms @xcite .
at least for the slow rotators , we expect that the zonal circulation should provide a circulation that reaches from the sub - stellar to the anti - stellar point
[ fig : omegamzc ] shows the mzc @xmath34 and vertical wind @xmath37 for the full set of experiments .
the slow rotators do indeed show a characteristic walker - like circulation with rising motion at the sub - stellar point , eastward ( and westward ) motion aloft , sinking motion near the sub - stellar point , and return flow along the surface .
the fast rotators show the opposite sense : sinking motion in the vicinity near the sub - stellar point , and rising motion within the anti - stellar hemisphere . because the rossby deformation radius is small compared to the planetary radius ,
the mzc cells for the fast rotators span a shorter longitudinal extent and does not fully reach the anti - stellar point .
conversely , the slow rotators have a larger rossby deformation radius that allows the mzc to traverse a full hemisphere .
geothermal heating appears to have relatively negligible effects on the mzc in all cases , while variations in vertical wind due to geothermal heating appear only in the fast rotators .
the mmc and mzc provide a partial description of the circulation on synchronously rotating planets , but the three - dimensional features noted by others @xcite include a cross - polar circulation that appears in all our gcm experiments and acts as an additional means of transport from the sub - stellar to anti - stellar point .
[ fig : polarheatfast ] shows a polar stereographic plot of eddy heat flux @xmath38 and horizontal wind for the set of fast rotators , where @xmath39 is potential temperature and the primes denote deviations from the zonal mean . following @xcite , the zonal mean also has been removed from all zonal wind vectors .
the first and third columns of fig .
[ fig : polarheatfast ] show the 950 hpa surface just above the ground , while the second and fourth columns show the 150 hpa surface at the top of the troposphere . in both dry and moist cases ,
a cross - polar circulation is evident with flow from the anti - stellar point to sub - stellar point along the surface , rising motion at the heating source , and flow aloft from the sub - stellar point back across the pole to the anti - stellar point .
likewise , fig .
[ fig : polarheatslow ] shows a polar stereographic plot of eddy heat flux and horizontal wind for the set of slow rotators , which shows the same cross - polar circulation pattern with more pronounced patterns of superrotation .
geothermal heating helps to increase the magnitude of superrotation in all these experiments , which increases the transport of mass and energy to the anti - stellar point by enhancing the cross - polar flow .
the cross - polar circulation appears to be evident as a means of transport from the sub - stellar to anti - stellar point in all these experiments .
the mmc and mzc also provide prominent circulations that maintain transient asymmetries between the northern and southern hemispheres , although only in the slow rotators where the rossby deformation radius is large does the mzc also contribute to this transport .
[ fig : geofluxcartoon ] shows a schematic diagram of the cross - polar circulation that contributes to energy transport on synchronously rotating planets .
this polar flow is present in all our experiments , regardless of rotation rate or geothermal heating , and should be a robust circulation feature in the atmospheres of synchronously rotating earth - like planets .
gcm studies have shown that synchronously rotating planets should be stable against atmospheric collapse as long as the atmosphere is sufficiently thick @xcite .
the dry and moist gcm calculations in this study all specify an earth - like value of reference surface pressure ( 1000 hpa ) , which should be resilient to collapse for any value of rotation rate ( or at least until the planet becomes close enough to its host star to trigger a runaway greenhouse @xcite ) .
the interpretation of these results therefore should be robust across any intermediate rotation rates between the fast and slow cases considered here , and similar patterns of large - scale dynamics should be evident in other gcms as well .
the presence of a geothermal heat flux due to tidal heating has the greatest impact on the fast rotators .
the cross - polar circulation provides enhanced warming on the anti - stellar side of the planet , which creates transient regions of warming that can reach or exceed 20 k at times .
this may be sufficient warming to initiate ice melt and provide warmer conditions on a transient basis , and in extreme cases this could even result in the permanent presence of standing liquid water on the anti - stellar side of the planet .
the degree to which this enhanced heat flux could melt ice will depend on the physical properties of each particular planet , and representing this complex ice flow is beyond the scope of this study .
nevertheless , the presence of a geothermal heat flux on fast rotators may contribute to preventing hard glaciation of the anti - stellar hemisphere and increase the available habitable surface area .
slow rotators should have greatest habitability potential along the terminator .
transient effects due to geothermal heating provide only a few degrees of warming or cooling , but this still might be sufficient to cause patterns in ice that forms on the anti - stellar side of such planets . whether or not this will result in permanent ice melt will depend on physical properties of the planet , but the strong superrotation on slow rotators tends to dominate the large - scale dynamical patterns .
the cross - polar circulation on slow rotators complements the zonal circulation so that surface winds everywhere are drawn into the sub - stellar point and carried aloft toward the anti - stellar point .
in general , the presence of the cross - polar circulation provides an additional mechanism for transporting energy and mass from the sub - stellar to anti - stellar point .
linearized theory based on the analysis of the walker circulation on earth suggests analogous structures for the circulation on synchronous rotators @xcite , but these theoretical expectations of a zonal circulation reaching from the sub - stellar to anti - stellar point only seem to apply to slow rotators where the rossby deformation radius is large .
faster rotators , with a smaller rossby deformation radius , do not exhibit significant zonal transport , but the pattern of the cross - polar circulation perhaps is more consistent with theory .
this suggests that any gcm simulations ( and , someday , observations ) of synchronously rotating planets should consider the mass transport by the cross - polar circulation , rather than the meridional or zonal circulations exclusively , as a means of characterizing the dynamics of climate .
the climate of a synchronously rotating planet is sometimes described using the term `` eyeball earth '' to denote open water beneath the sub - stellar point , relatively clement conditions along the surface in a surrounding ring , and ice - covered conditions everywhere else all the way to the anti - stellar hemisphere @xcite .
if geothermal heating can cause regions of transient warming through the enhancement of large - scale circulations , then regions of melting could create cratered pockets of ice , slowly flowing glaciers , or even standing liquid water around the anti - stellar point .
this suggests that `` eyeball earth '' planets might include a soft spot on the side farthest from the heat source .
further investigation with gcm s coupled to ice sheet models will help to constrain the range of climates on which this could occur .
low mass m - type stars provide some of the closest and most numerous targets for current exoplanet surveys , and the discovery of synchronously rotating planets within the habitable zones of these stars will increase interest in the atmospheric characterization of such worlds .
we have considered a set of experiments that include slow rotators ( 1 day ) toward the inner edge of the habitable zone and fast rotators ( 230 days ) toward the outer edge of the habitable zone . between
these fast and slow rotators is a regime transition where superrotation dominates , but the general patterns discussed from this limited set of experiments should extend appropriately to other rotation rates within the habitable zone .
tidal heating can produce a geothermal heat flux that warms the surface of a synchronously rotating planet , which can cause transient regions of warming and cooling of up to @xmath2920 k or more on the anti - stellar hemisphere .
this enhanced warming could contribute to increased ice melt along the anti - stellar point and extend the habitable surface area of the planet .
this effect is stronger on fast rotators , and the presence of moisture also tends to reduce the magnitude of anomalous regions .
the presence of geothermal heating acts to enhance asymmetries between the northern and southern hemispheres that alter the structure of large - scale circulation features such as the mmc , the mzc , and the cross - polar circulation .
the dynamics of synchronous rotators include different patterns on either side of the sub - stellar point .
the fast rotators exhibit a change in the direction of the hadley circulation from the eastern to western hemisphere , while the slow rotators show a change in the direction of zonal wind .
the walker circulation provides transport of mass and energy from the sub - stellar to anti - stellar point only for slow rotators , while fast rotators show a walker circulation with the opposite sense . in all cases
, a cross - polar circulation provides an additional mechanism for transport of mass and energy from the sub - stellar to anti - stellar point .
fast rotators maintain asymmetric features such as large - scale jets and vorticies at midlatitudes , while slow rotators have transitioned to a superrotation regime , but all experiments show a cross - polar circulation that provides or enhances the transport from day side to night side .
typical considerations of the hadley and walker circulation therefore may be insufficient for synchronously rotating planets , with the cross - polar circulation and hemispheric meridional circulation providing more accurate descriptions of atmospheric transport .
our set of experiments is designed to highlight important features in the large - scale dynamics of synchronously rotating planets .
the idealized gcm used in this study can not provide precise orbital limits to habitability , but the simplified nature of this model allows the most fundamental atmospheric processes to be more apparent .
more accurate representation of clouds and radiative transfer will improve the model s ability to predict particular climates for specific planetary configurations , but the general patterns of large - scale dynamics should remain consistent across the parameter space described here . as astronomers continue the hunt for alien worlds and turn their eyes to low mass stars nearby , atmospheric modelers should continue the quest to understand circulation patterns on worlds very much unlike our own .
funding for this research was provided by the nasa astrobiology institute s virtual planetary laboratory under awards nnx11ac95g , s03 ( j.h . ) and nnh05zda001c ( r.k.k . ) . | earth - like planets within the liquid water habitable zone of m type stars may evolve into synchronous rotators . on these planets , the sub - stellar hemisphere experiences perpetual daylight while the opposing anti - stellar hemisphere experiences perpetual darkness .
because the night - side hemisphere has no direct source of energy , the air over this side of the planet is prone to freeze out and deposit on the surface , which could result in atmospheric collapse .
however , general circulation models ( gcms ) have shown that atmospheric dynamics can counteract this problem and provide sufficient energy transport to the anti - stellar side . here
we use an idealized gcm to consider the impact of geothermal heating on the habitability of synchronously rotating planets .
geothermal heating may be expected due to tidal interactions with the host star , and the effects of geothermal heating provide additional habitable surface area and may help to induce melting of ice on the anti - stellar hemisphere .
we also explore the persistence of atmospheric asymmetries between the northern and southern hemispheres , and we find that the direction of the meridional circulation ( for rapidly rotating planets ) or the direction of zonal wind ( for slowly rotating planets ) reverses on either side of the sub - stellar point . we show that the zonal circulation approaches a theoretical state similar to a walker circulation only for slowly rotating planets , while rapidly rotating planets show a zonal circulation with the opposite direction .
we find that a cross - polar circulation is present in all cases and provides an additional mechanism of mass and energy transport from the sub - stellar to anti - stellar point .
characterization of the atmospheres of synchronously rotating planets should include consideration of hemispheric differences in meridional circulation and examination of transport due to cross - polar flow .
[ firstpage ] planets and satellites : terrestrial planets
stars : low - mass planets and satellites : atmospheres hydrodynamics astrobiology . |
the origin and nature of dark energy @xcite is one of the most difficult challenges facing physicists and cosmologists now . among all the proposed models to tackle this problem ,
a scalar field is perhaps the most popular one up to now .
the scalar field , denoted by @xmath4 , might only interact with other matter species through gravity , or have a coupling to normal matter and therefore producing a fifth force on matter particles .
the latter idea has seen a lot of interests in recent years , in the light that such a coupling could alleviate the coincidence problem of dark energy and that it is commonly predicted by low energy effective theories from a fundamental theory .
nevertheless , if there is a coupling between the scalar field and baryonic particles , then stringent experimental constraints might be placed on the fifth force on the latter provided that the scalar field mass is very light ( which is needed for the dark energy ) .
such constraints severely limit the viable parameter space of the model .
different ways out of the problem have been proposed , of which the simplest one is to have the scalar field coupling to dark matter only but not to standard model particles , therefore evading those constraints entirely .
this is certainly possible , especially because both dark matter and dark energy are unknown to us and they may well have a common origin .
another interesting possibility is to have the chameleon mechanism @xcite , by virtue of which the scalar field acquires a large mass in high density regions and thus the fifth force becomes undetectablly short - ranged , and so also evades the constraints .
study of the cosmological effect of a chameleon scalar field shows that the fifth force is so short - ranged that it has negligible effect in the large scale structure formation @xcite for certain choices of the scalar field potential .
but it is possible that the scalar field has a large enough mass in the solar system to pass any constraints , and at the same time has a low enough mass ( thus long range forces ) on cosmological scales , producing interesting phenomenon in the structure formation .
this is the case of some @xmath5 gravity models @xcite , which survives solar system tests thanks again to the chameleon effect @xcite .
note that the @xmath5 gravity model is mathematically equivalent to a scalar field model with matter coupling . no matter whether the scalar field couples with dark matter only or with all matter species , it is of general interests to study its effects in cosmology , especially in the large scale structure formation . indeed , at the linear perturbation level there have been a lot of studies about the coupled scalar field and @xmath5 gravity models which enable us to have a much clearer picture about their behaviors now .
but linear perturbation studies do not conclude the whole story , because it is well known that the matter distribution at late times becomes nonlinear , making the behavior of the scalar field more complex and the linear analysis insufficient to produce accurate results to confront with observations .
for the latter purpose the best way is to perform full n - body simulations @xcite to evolve the individual particles step by step .
n - body simulations for scalar field and relevant models have been performed before @xcite .
for example , in @xcite the simulation is about a specific coupled scalar field model .
this study however does not obtain a full solution to the spatial configuration of the scalar field , but instead simplifies the simulation by assuming that the scalar field s effect is to change the value of the gravitational constant , and presenting an justifying argument for such an approximation . as we will see below , this approximation is only good in certain parameter spaces and for certain choices of the scalar field potential , and therefore we believe fuller simulations than the one performed in @xcite is needed to study the scalar field behavior most rigorously .
note also the structure formation with a coupled chameleon scalar field has also been studied using non b - body techniques previously @xcite .
recently there have also appeared full simulations of the @xmath5 gravity model @xcite , which do solve the scalar degree of freedom explicitly .
however , embedded in the @xmath5 framework there are some limitations in the generality of these works . as a first thing ,
@xmath5 gravity model ( no matter what the form @xmath6 is ) only corresponds to the couple scalar field models for a specific value of coupling strength @xcite .
second , in @xmath5 models the correction to standard general relativity is through the modification to the poission equation and thus to the gravitational potential as a whole @xcite , while in the coupled scalar field models we could clearly separate the scalar fifth force from gravity and analyze the former directly . also , in @xmath5 models the coupling between matter and the scalar field is universal ( the same to dark matter and baryons ) , while in the couple scalar field models it is straightforward to switch on / off the coupling to baryons and study the effects on baryonic and dark matter clusterings respectively . and
finally , the general framework of n - body simulations in couple scalar field models could also handle the situation where the chameleon effect is absent and scalar field only couples to dark matter ( which is of no interests for @xmath5 people ) . in this article we present the general formulae and algorithm of full n - body simulations in coupled scalar field models and consider as an explicit example the results for a chameleon scalar field .
unlike in @xcite , here we shall calculate the spatial distribution of the scalar field directly without making simplifications such as a rescaled gravitational constant . neither
shall we use the concept of varying - mass dark matter particles as in @xcite , but instead we treat the system as constant - mass dark matter particles under the action of a fifth force .
the article is organized as follows : in [ sect : model ] we review the general equations of motion for the coupled scalar field model and introduce our specific choices of the coupling function and scalar field potential .
next we analyze in [ sect : lin ] the general linear perturbation equations in the @xmath7 formalism @xcite and integrate them into the numerical boltzmann code camb @xcite to study the effects on the linear structure formation . then in [ sect : nonlin ] we turn to our main focus , introducing the formulae and algorithm of the n - body simulation .
we also study the chosen model explicitly , display the preliminary results and discuss on them .
finally we conclude in [ sect : disc ] .
in this section we briefly introduce the model considered here and present the equations that will be analyzed in the following sections .
let us start by looking at the general field equations for a scalar field coupled to dark matter .
the lagrangian for our coupled scalar field model is
@xmath8 + v(\varphi ) - c(\varphi)\mathcal{l}_{\mathrm{cdm } } + \mathcal{l}_{\mathrm{s}}\ \\end{aligned}\ ] ] where @xmath9 is the ricci scalar , @xmath10 with @xmath11 newton s constant , @xmath4 is the scalar field , @xmath12 is its potential energy and @xmath13 its coupling to dark matter , which is assumed to be cold and described by the lagrangian @xmath14 .
@xmath15 includes all the terms for photons , neutrinos and baryons , and these will be considered only when we calculate the large scale structure formation in the next section .
the dark matter lagrangian , for a point - like particle with ( bare ) mass @xmath16 , is @xmath17 where @xmath18 is the coordinate and @xmath19 is the coordinate of the centre of the particle . from this equation
it can be easily derived that @xmath20 also , because @xmath21 where @xmath22 is the four velocity of the dark matter particle , so the lagrangian could be rewritten as @xmath23 which will be used below .
( [ eq : dmemt_particle ] ) is the energy momentum tensor for a single dark matter particle . for a fluid with many particles
the energy momentum tensor will be @xmath24 in which @xmath25 is a volume microscopically large and macroscopically small , and we have extend the 3-dimensional @xmath26 function to a 4-dimensional one by adding a time component .
meanwhile , using @xmath27 it is straightforward to show that the energy momentum tensor for the scalar field is given by @xmath28.\end{aligned}\ ] ] so the total energy momentum tensor is @xmath29\nonumber\\ & & + c(\varphi)t^{\mathrm{cdm}}_{ab } + t^{\mathrm{s}}_{ab}\end{aligned}\ ] ] where @xmath30 , @xmath31 is the energy momentum tensor for standard model particles , and the einstein equation is @xmath32 where @xmath33 is the einstein tensor . note that due to the coupling between the scalar field @xmath4 and the dark matter , the energy momentum tensors for either will not be conserved , and we have @xmath34 where throughout this paper we shall use a @xmath35 to denote the derivative with respect to @xmath4 .
finally , the scalar field equation of motion ( eom ) from the given lagrangian is @xmath36 where @xmath37 . using eq .
( [ eq : dmlagrangian2 ] ) it can be rewritten as @xmath38 eqs .
( [ eq : emt_tot ] , [ eq : einsteineq ] , [ eq : dm_energy_conservation ] , [ eq : phieom ] ) summarize all the physics that will be used in our analysis .
we will consider a special form for the scalar field potential , @xmath39^{\mu}},\end{aligned}\ ] ] where we have fixed the coefficient in front of @xmath4 to be @xmath40 without loss of generality , since we can always rescale @xmath4 to achieve this ; @xmath41 is a dimensionless constant and @xmath42 is a constant with mass dimension @xmath43 .
as will be discussed below , we need @xmath44 to evade local observational constraints . meanwhile , the coupling between the scalar field and dark matter particle is chosen as @xmath45 where @xmath46 is another dimensionless constant . as will be explained below
, the two dimensionless parameters @xmath41 and @xmath47 have clear physical meanings : _ roughly speaking _
, @xmath41 controls the time when the effect of the scalar field becomes important in cosmology while @xmath47 determines how important it would ultimately be . indeed , the potential given in eq .
( [ eq : potential ] ) is motivated by the @xmath5 cosmology @xcite , in which the extra degree of freedom behaves as a coupled scalar field in the einstein frame . as we could see from eq .
( [ eq : potential ] ) , the potential @xmath48 when @xmath49 while @xmath50 when @xmath51 . in the latter case , however , @xmath52 , so that the effective total potential @xmath53 has a global minimum at some finite @xmath4 .
if the total potential @xmath54 is steep enough around this minimum , then the scalar field becomes very heavy and thus follows its minimum dynamically , as is in the case of the chameleon cosmology .
if @xmath55 is not steep enough at the minimum , however , the scalar field will have a more complicated evolution .
these two different cases can be obtained by choosing appropriate values of @xmath56 : if @xmath47 is very large or @xmath41 is small then we run into the former situation and if @xmath47 is small and @xmath41 is large we have the latter . in fig .
[ fig : figure1 ] we present a schematic plot of the two situations : the significant difference is that at late times when @xmath57 becomes small , the effective potential becomes flat around its minimum if @xmath47 is not very large and @xmath41 not very small , and as a result the true value of @xmath4 will lag behind that corresponding to the minimum of @xmath55 .
of course if @xmath42 is chosen appropriately the scalar field can act as a dynamical dark energy in this slow - roll regime .
the complexity of the two cases also makes them phenomenologically rich , and it is of our interests to study how such a setup will affect the cosmology in background , linear perturbation and in particular nonlinear structure formation regimes . in the next two sections we shall consider these questions . as a function of @xmath58 .
the parameters are the same as in fig .
[ fig : figure2 ] and the values of @xmath41 are given beneath the curves . ]
our first task is to linearize the model .
since this involves many equations , we refer the interested readers to the appendices [ appen : perteqn ] and [ appen : k - eqn ] for the complete first - order perturbation equations and their representations in @xmath59-space .
we shall proceed then to study their effects on the large scale structure at linear perturbation level in the universe by putting these linearize equations into the numerical boltzmann code . in what follows
we will consider the limit of small @xmath41 , @xmath60 .
it is then useful to define the effective mass of the scalar field by the taylor expansion @xmath61 where @xmath62 is at the minimum of @xmath55 , given by @xmath63 in which we have used the facts that @xmath64 as well as @xmath65 so that @xmath66 and @xmath67 . thus the effective mass @xmath68^{1+\mu } } \left [ 1 + { ( 1+\mu ) y\over 1-y } \right ] + \gamma^{2}\kappa\rho_{\mathrm cdm } y^{-\gamma } \nonumber\\ & \approx & \frac { ( \gamma \kappa \rho_{\mathrm cdm})^2 } { \kappa v_{0 } } \left[{1+\mu \over \mu } + \ln { \mu v_0 \over \gamma \rho_{\mathrm cdm } } + { ( 1 + 3\gamma ) v_0 \over \gamma \rho_{\mathrm cdm}}\right ] , \nonumber\\ & \approx & \frac { ( \gamma \kappa \rho_{\mathrm cdm})^2 } { \mu \kappa v_{0}}\end{aligned}\ ] ] to see that the scalar field is heavy , _
i.e. _ , @xmath69 , note that @xmath70 , true for @xmath71 [ or @xmath72 but in the early universe , remember that we always assume @xmath73 .
this means that for @xmath44 ( or in the early times ) the scalar field has a very heavy mass and tends to settle at @xmath74 ( quickly oscillating there ) .
then as a good approximation we have @xmath75 in which we have used the facts that @xmath76 in this regime . at the same time , because @xmath77 with the use of eq .
( [ eq : phi_min ] ) , so the kinetic energy of the scalar field is negligible .
this shows that for small @xmath41 we do expect the scalar field to behave like a cosmological constant in background cosmology .
this analysis has been confirmed by numerical calculations . in fig .
[ fig : figure2 ] we show the fractional energy densities of the radiation , dust ( baryon plus cdm ) and dark energy ( the scalar field ) components for @xmath78 and several values of @xmath41 .
it can be seen there the evolutions of these fractional energy densities are not sensitive to @xmath41 . to see this more clearly we have also plotted in fig . [
fig : figure3 ] the total equation of state of all matter species . as @xmath79
the curves quickly become indistinguishable from each other ( and indistinguishable from the @xmath1cdm prediction ) .
to study the perturbation evolution and the effect of the scalar field on large scale structure formation , we just need to implement the perturbation equations listed in appendix [ appen : k - eqn ] into a boltzmann code .
as mentioned earlier , here we work in the @xmath80 frame , in which the cdm peculiar velocity @xmath81 must be dynamically evolved according to eq .
( [ eq : pb_v_cdm ] ) with @xmath80 .
also note that the tight coupling approximation is not affected as compared with gr+@xmath1cdm .
the initial condition for @xmath81 could be set to zero . in this way
the observer is comoving with the dark matter particles initially ( when the @xmath82 term in eq .
( [ eq : bg_rho_cdm ] ) is extremely small ) and only deviate from the dark matter particle world lines when this @xmath82 term becomes significant . as for the initial conditions of @xmath82 and @xmath83 , they are also very small at early times , and as we shall see below , there is good reason to set them to zero at those times .
thus we have adopted @xmath84 as the initial conditions of the new variables which need to be evolved in the code . in fig .
[ fig : figure4 ] we show a collect of the cmb power spectra for the model with @xmath85 . in fig .
[ fig : figure5 ] the matter power spectra of the corresponding parameter choices are given .
we can see that the effect of the scalar field on the cmb spectrum is most significant for low-@xmath86 ( large scales ) , and even there the effect is only to slightly reduce the power .
this behavior is quite similar to the one found in @xcite and is due to the less decay of the gravitational potential on large scales at late times , which decreases the integrated sachs - wolfe ( isw ) effect .
such small deviations of the cmb spectrum from the @xmath1cdm result are obviously not of much help in placing constraints on the model parameters @xmath41 and @xmath47 , especially because of the cosmic variance on large scales .
thus to distinguish this model from others we necessarily need to use other observables such as the linear ( on large scales ) and nonlinear ( on smaller scales ) matter power spectra .
to explain the matter power spectrum , let us consider the evolution of cold dark matter density contrast @xmath87 ( a similar analysis is first given in @xcite for @xmath5 model and later more generally in @xcite ) . for simplicity we shall assume a universe filled with only radiation and cdm particles .
taking the ( conformal ) time derivative of the equation @xmath88 we have @xmath89 taking the spatial derivative of eq .
( [ eq : raychaudhrui ] ) , collecting the coefficients of harmonic expansions and using the frame choice @xmath90 we obtain @xmath91 substituting eq .
( [ eq : zevolution ] ) into eq .
( [ eq : deltaevolution2 ] ) and using eqs .
( [ eq : pb_v_cdm ] , [ eq : deltaevolution1 ] ) we arrive at @xmath92 where the last term is equal to @xmath93 with @xmath82 being substituted using the propagation equation of @xmath94 ( remember that @xmath80 by choice of the frame ) . now according to eq .
( [ eq : sfpert_eom ] ) the evolution of @xmath82 is governed by @xmath95 for small scales ( and late times ) @xmath96 and the term @xmath97 dominates over other terms proportional to @xmath82 in the above equations dominates so that the equation is approximately @xmath98 where we choose to eliminate @xmath99 by using eq .
( [ eq : deltaevolution1 ] ) . substituting eq .
( [ eq : sf_pert ] ) into eq .
( [ eq : deltaevolution3 ] ) to eliminate @xmath82 and @xmath81 , we get @xmath100 during the matter dominated era we can neglect the contribution from radiation , and the above equation for the growth of the overdensity reduces to @xmath101 where @xmath102 where @xmath103 as before , and @xmath104 for our coupling function of @xmath105 as given by eq .
( [ eq : coupling_function ] ) .
thus we could see that the growth of cold dark matter density contrast is ( on small scales ) scale - independent , which explains the matter power spectrum at large @xmath59 values ( small scales ) in fig .
[ fig : figure5 ] . in particular , note that for our choice of parameters we have @xmath106 and @xmath107 , and so this equation is approximately the same as that in the @xmath108 paradigm but with an effective gravitational constant @xmath109 .
this indicates that the parameter @xmath47 characterizes the strength of the scalar field fifth force relative to gravity ; the larger @xmath47 is , the stronger the fifth force will _ ultimately _ be . .
the red , blue , green and black curves correspond to @xmath110 with chameleon perturbations and @xmath111 _ without _ chameleon perturbations , respectively .
curves for @xmath112 are indistinguishable from the @xmath1cdm result . ] .
the curves with @xmath112 are very close to the @xmath1cdm result . ]
meanwhile , from the derivation process of eq .
( [ eq : deltaevolution4 ] ) we see that the assumption that @xmath113 [ cf .
[ eq : sfpert_eom2 ] ] has been used .
this is of course true for our choices of @xmath41 , but eq .
( [ eq : v_phiphi ] ) tells us that this is not necessarily the case for @xmath44 , due to the strong nonlinearity of @xmath12 .
let s suppose now @xmath114 for some small enough value of @xmath41 ( say @xmath115 ) , then eq . ( [ eq : sfpert_eom2 ] )
should be approximated as @xmath116 rather than as eq .
( [ eq : sf_appox1 ] ) , so that in eq
. ( [ eq : deltaevolution3 ] ) the @xmath97 term could be neglected , implying that the scalar field simply has negligible effects on the evolution of @xmath87 .
thus we conclude that the effect of the parameter @xmath41 is to control the time when the fifth force becomes important as compared with gravity ; the smaller @xmath41 is , the later will this time be .
so with very strong chameleon effects ( very nonlinear potentials ) the fifth force could be greatly suppressed all through the cosmic history up to now , and it is therefore possible to get a cosmology very close to the @xmath1cdm paradigm in every aspect . obviously , at the very early times the fifth force is also suppressed even in cases of @xmath117 , because in this regime again we have @xmath114 .
indeed this is the reason why we could set the initial condition as in eq .
( [ eq : initcond ] ) .
now the matter power spectra as seen in fig .
[ fig : figure5 ] could be well explained : when @xmath41 decreases , the time when the fifth force could fully show its power becomes later and as a result the growth of structure does not reach its maximum potential this explains why the @xmath118 curves increasingly get far away from the black one .
but because all these values of @xmath41 are large enough ( compared with , say , @xmath115 ) , the fifth force in all these cases has _ almost _ realized its _
full _ power ( which is @xmath2 times stronger than gravity according to eq .
( [ eq : deltaevolution4 ] ) ) this explains why all the colored curves are fairly far from the black one . finally , eq . ( [ eq : deltaevolution4 ] ) explains why on small scales the colored @xmath119 curves are almost parallel to the black one .
given the complexity of the scalar field behavior , especially in regions with highly nonlinear matter distribution , we must say that the linear analysis performed above is only qualitatively correct and lacks the high precision to compare with various cosmological data sets .
in particular , it may well be that the effect of the fifth force is not significant on linear scales because the fifth force is not long - range enough , but much more important on smaller scales which are beyond the linear regime .
thus in the next section we shall consider the scalar field model in the context of n - body cosmological simulations , and try to study its effects in a more accurate manner .
in [ sect : lin ] we have considered the background evolution and linear large structure formation in the couple scalar field model . for the former , the coupling between @xmath4 and cdm prevents @xmath4 from rolling to infinity and instead tends to keep @xmath4 and @xmath12 constant ; here we find that for a value of @xmath60 this effect is significant enough to make the background cosmology similar to that of @xmath1cdm .
for the latter , the cmb spectrum is also not very sensitive to the value of @xmath41 or @xmath47 such that @xmath60 is difficult to be distinguished from @xmath1cdm .
these results suggest that large scale observables are not particularly well suited in studying the new features of this class of coupled scalar field models . on smaller but still linear scales
, we already see that the scalar field coupling effectively increases the gravitational force by a factor of @xmath2 , leading to a significant increase in the small scale power of @xmath119 for gravitational - strength couplings [ @xmath73 .
one could of course argue that the observed matter power spectrum is indeed for the luminous matter only and can not be applied to dark matter navely due to the bias between the two .
but the enhancement of dark matter clustering in this model is abstract and can be tested by observations such as weak lensing without much ambiguity . if this is done ,
it is unlikely that there remains much space for such a strongly coupled model .
however , two things must be taken into account before we arrive at any definite conclusions about the fate of the model .
first , on small scales where the scalar field effect becomes important , the distribution of matter is already beyond the linear perturbation regime and in some regions could be very nonlinear .
this means that a linear analysis as presented in the above section is no longer sufficient and we must consider the nonlinear effect , which is most precisely taken in account by large n - body simulations
. this will be the very topic of this section .
second , as we argued in
[ sect : lin ] , the behavior of the model is not only controlled by @xmath47 , but also by parameter @xmath41 : for @xmath44 , the epoch when the scalar field fifth force starts to realizes its full power as @xmath2 times of gravity could be greatly postponed .
it is then possible to prevent too powerful a structure formation . like in the background cosmology , this has something to do with nonlinear ( chameleon ) effects , but here things become much more complicated . in a homogeneous universe , the spatial gradient of the scalar field vanishes and the scalar field @xmath4 takes the same value anywhere , largely simplifying the analysis . for the real universe with nonlinear matter distributions ,
the spatial gradient of @xmath4 is normally much larger than the time derivative so that the configuration of the scalar field relies on the underlying matter distribution sensitively in a nonlinear way , and at the same time strongly affects the latter via the action of fifth force . quantifying such complex couplings
also calls for the use of n - body simulations . in this paper
we shall set up the basic framework of studying coupled scalar field models using n - body technique , putting much emphasis on the working mechanism and the fifth force effect .
we shall start with a derivation of all the relevant equations of motion , describe how to integrate them into the numerical code and present some preliminary results .
detailed analysis will be postponed to companion papers .
we will not provide an introduction to general n - body simulation techniques here , and interested readers are referred to the relevant literature @xcite .
our first step is to simplify the relevant equations of motion in the appropriate limit to get a set of equations which can be directly applied to the numerical code .
we know that the existence of the scalar field and its coupling to standard cold dark matter particles make the following changes to the @xmath1cdm model : first , the energy momentum tensor has a new piece of contribution from the scalar field ; second , the energy density of dark matter is multiplied by a function @xmath13 , which is because the coupling to scalar field essentially renormalizes the mass of dark matter particles ; third and most important , dark matter particles will not follow geodesics in their motions as in @xmath1cdm , rather , the total force on them has a contribution from the scalar field .
these imply that the following things need to be modified or added : 1 .
the scalar field @xmath4 equation of motion , which determines the value of the scalar field at any given time and position .
2 . the poisson equation , which determines the gravitational potential ( and thus gravity ) at any given time and position , according to the local energy density and pressure , which include the contribution from the scalar field ( as obtained from @xmath4 equation of motion ) now .
the total force on the dark matter particles , which is determined by the spatial configuration of @xmath4 , just like gravity is determined by the spatial configuration of the gravitational potential .
we shall describe these one by one now . for the scalar field equation of motion ,
we denote @xmath120 as the background value of @xmath4 and @xmath121 as the scalar field perturbation .
then eq .
( [ eq : phieom ] ) could be rewritten as @xmath122 by subtracting the corresponding background equation from it . here
@xmath123 is the covariant spatial derivative with respect to the physical coordinate @xmath124 with @xmath125 the conformal coordinate , and @xmath126 .
@xmath123 is essentially the @xmath127 , but because here we are working in the weak field limit we approximate it as @xmath128 by assuming a flat background ; the minus sign is because our metric convention is @xmath129 instead of @xmath130 . for the simulation here we will also work in the quasi - static limit , assuming that the spatial gradient is much larger than the time derivative , @xmath131 ( which will be justified below ) .
thus the above equation can be further simplified as @xmath132,\nonumber\end{aligned}\ ] ] in which @xmath133 is with respect to the conformal coordinate @xmath125 so that @xmath134 , and we have restored the factor @xmath135 in front of @xmath136 ( the @xmath4 here and in the remaining of this paper is @xmath137 times the @xmath4 in the original lagrangian unless otherwise stated ) .
note that here @xmath25 and @xmath57 both have the dimension of__mass _ _ density rather than _ energy _ density .
next look at the poisson equation , which is obtained from the einstein equation in weak - field and slow - motion limits . here
the metric could be written as @xmath138 from which we find that the time - time component of the ricci curvature tensor @xmath139 , and then the einstein equation @xmath140 gives @xmath141 where @xmath142 and @xmath143 are respectively the total energy density and pressure .
the quantity @xmath144 can be expressed in terms of the comoving coordinate @xmath125 as @xmath145 where we have defined a new newtonian potential @xmath146 and used @xmath147 .
thus @xmath148\nonumber\end{aligned}\ ] ] where in the second step we have used eq .
( [ eq : einsteineqn ] ) and the raychaudhrui equation , and an overbar means the background value of a quantity . because the energy momentum tensor for the scalar field is given by eq .
( [ eq : phiemt ] ) , it is easy to show that @xmath149 $ ] and so @xmath150\right\}\nonumber\\ & & + 4\pi ga^{3 } \left\{\bar{\rho}_{\mathrm{cdm}}c(\bar{\varphi})+ 2\left[\dot{\bar{\varphi}}^{2}-v(\bar{\varphi})\right]\right\}.\nonumber\end{aligned}\ ] ] now in this equation @xmath151 in the quasi - static limit and so could be dropped safely .
so we finally have @xmath152\nonumber\\ & & - 8\pi ga^{3}\left[v(\varphi)-v(\bar{\varphi})\right].\end{aligned}\ ] ] finally , for the equation of motion of the dark matter particle , consider eq .
( [ eq : dm_energy_conservation ] ) . using eqs .
( [ eq : dmemt_particle ] , [ eq : dmlagrangian2 ] ) , this can be reduced to @xmath153 obviously the left hand side is the conventional geodesic equation and the right hand side is the new fifth force due to the coupling to the scalar field . before going on further , note that the fifth force @xmath154 is perpendicular to the 4-velocity @xmath155 ; this means that the energy density of cdm will be conserved and only the particle trajectories are modified as mentioned above . now in the non - relativistic limit the spatial components of eq .
( [ eq : dmeom ] ) can be written as @xmath156 where @xmath157 is the physical time coordinate .
if we instead use the comoving coordinate @xmath125 , then this becomes @xmath158 where we have used eq .
( [ eq : newphi ] ) .
the canonical momentum conjugate to @xmath125 is @xmath159 so we have now @xmath160 eqs .
( [ eq : wfphieom ] , [ eq : wfpoisson ] , [ eq : wfdxdtcomov ] , [ eq : wfdpdtcomov ] ) will be used in the code to evaluate the forces on the dark matter particles and evolve their positions and momenta in time . in our numerical simulation
we use a modified version of mlapm ( @xcite , see [ subsect : n - body ] ) , and we must change our above equations in accordance with the internal units used in that code . here
we briefly summarize the main features .
mlapm code uses the following internal units ( with subscript @xmath161 ) : @xmath162 in which @xmath163 is the present size of the simulation box and @xmath164 is the present hubble constant . using these newly - defined quantities , it is easy to check that eqs .
( [ eq : wfdxdtcomov ] , [ eq : wfdpdtcomov ] , [ eq : wfpoisson ] , [ eq : wfphieom ] ) could be rewritten as @xmath165 and @xmath166 where @xmath167 is the present cdm fractional energy density , we have again restored the factor @xmath135 and again the @xmath4 is @xmath137 times the @xmath4 in the original lagrangian .
also note that from here on we shall use @xmath168 unless otherwise stated , for simplicity .
we also define @xmath169 to be used below .
making discretized version of the above equations for n - body simulations is non - trivial task .
for example , the use of variable @xmath170 instead of @xmath4 ( see below ) helps to prevent @xmath171 , which is unphysical , but numerically possible due to discretization .
we refer the interested readers to appendix [ appen : discret ] to the whole treatment , with which we can now proceed to do n - body runs .
the full name of mlapm is multi - level adaptive particle mesh code . as the name has suggested , this code uses multilevel grids @xcite to accelerate the convergence of the ( nonlinear ) gauss - seidel relaxation method @xcite in solving boundary value partial differential equations .
but more than this , the code is also adaptive , always refining the grid in regions where the mass / particle density exceeds a certain threshold .
each refinement level form a finer grid which the particles will be then ( re)linked onto and where the field equations will be solved ( with a smaller time step ) .
thus mlapm has two kinds of grids : the domain grid which is fixed at the beginning of a simulation , and refined grids which are generated according to the particle distribution and which are destroyed after a complete time step .
one benefit of such a setup is that in low density regions where the resolution requirement is not high , less time steps are needed , while the majority of computing sources could be used in those few high density regions where high resolution is needed to ensure precision .
some technical issues must be taken care of however .
for example , once a refined grid is created , the particles in that region will be linked onto it and densities on it are calculated , then the coarse - grid values of the gravitational potential are interpolated to obtain the corresponding values on the finer grid . when the gauss - seidel iteration is performed on refined grids , the gravitational potential on the boundary nodes
are kept constant and only those on the interior nodes are updated according to eq .
( [ eq : gs ] ) : just to ensure consistency between coarse and refined grids .
this point is also important in the scalar field simulation because , like the gravitational potential , the scalar field value is also evaluated on and communicated between multi - grids ( note in particular that different boundary conditions lead to different solutions to the scalar field equation of motion ) .
in our simulation the domain grid ( the finest grid that is not a refined grid ) has @xmath172 nodes , and there are a ladder of coarser grids with @xmath173 , @xmath174 , @xmath175 , @xmath176 , @xmath177 nodes respectively .
these grids are used for the multi - grid acceleration of convergence : for the gauss - seidel relaxation method , the convergence rate is high upon the first several iterations , but quickly becomes very slow then ; this is because the convergence is only efficient for the high frequency ( short - range ) fourier modes , while for low frequency ( long - range ) modes more iterations just do not help much . to accelerate the solution process
, one then switches to the next coarser grid for which the low frequency modes of the finer grid are actually high frequency ones and thus converge fast .
the mlapm solver adopts the self - adaptive scheme : if convergence is achieved on a grid , then interpolate the relevant quantities back to the finer grid ( provided that the latter is not on the refinements ) and solve the equation there again ; if convergence becomes slow on a grid , then go to the next coarser grid . this way it goes indefinitely except when converged solution on the domain grid
is obtained or when one arrives at the coarsest grid ( normally with @xmath178 nodes ) on which the equations can be solved exactly using other techniques . for our scalar field model ,
the equations are difficult to solve anyway , and so we truncate the coarser - grid series at the @xmath177-node one , on which we simply iterate until convergence is achieved . for the refined grids the method is different : here one
just iterate eq . ( [ eq : gs ] ) until convergence , without resorting to coarser grids for acceleration .
as is normal in the gauss - seidel relaxation method , convergence is deemed to be achieved when the numerical solution @xmath179 after @xmath180 iterations on grid @xmath59 satisfies that the norm @xmath181 ( mean or maximum value on a grid ) of the residual @xmath182 is smaller than the norm of the truncation error @xmath183\end{aligned}\ ] ] by a certain amount .
note here @xmath184 is the discretization of the differential operator eq .
( [ eq : diffop ] ) on grid @xmath59 and @xmath185 a similar discretization on grid @xmath186 , @xmath187 is the source term , @xmath188 is the restriction operator to interpolate values from the grid @xmath59 to the grid @xmath186 . in the modified code
we have used the full - weighting restriction for @xmath188 .
correspondingly there is a prolongation operator @xmath189 to obtain values from grid @xmath186 to grid @xmath59 , and we use a bilinear interpolation for it . for more details see @xcite .
mlapm calculates the gravitational forces on particles by centered difference of the potential @xmath190 and propagate the forces to locations of particles by the so - called triangular - shaped - cloud ( tsc ) scheme to ensure momentum conservation on all grids .
the tsc scheme is also used in the density assignment given the particle distribution .
the main modifications to the mlapm code for our model are : 1 . we have added a parallel solver for the scalar field based on eq .
( [ eq : u_phi_eom ] ) .
the solver uses a similar nonlinear gauss - seidel method and same criterion for convergence as the poisson solver .
2 . the solved value of @xmath170 is then used to calculate local mass density and thus the source term for the poisson equation , which is solved using fast fourier transform . 3 .
the fifth force is obtained by differentiating the @xmath170 just like the calculation of gravity .
4 . the momenta and positions of particles
are then updated taking in account of both gravity and the fifth force .
there are a lot of additions and modifications to ensure smooth interface and the newly added data structures . for the output , as there are multilevel grids all of which host particles , the composite grid is inhomogeneous and thus we choose to output the positions , momenta of the particles , plus the gravity , fifth force and scalar field value _ at the positions _ of these particles .
we can of course easily read these data into the code , calculate the corresponding quantities on each grid and output them if needed . in this subsection
we shall present some preliminary results of several runs and give a sense about the qualitative behaviors of the coupled scalar field model . also we will not make detailed and quantitative analysis in this paper , which will be shown in forthcoming papers .
we have performed 8 runs of the modified code with parameters @xmath191 and @xmath192 respectively .
for all these runs there are @xmath172 dark matter particles , the simulation box has a size @xmath193 in which @xmath194 and @xmath195 domain grid cells in each direction .
we assume a @xmath1cdm background cosmology which is a very good approximation for @xmath44 as we mentioned in [ sect : lin ] ; the current fractional energy densities of dark matter and dark energy are @xmath196 and @xmath197 ( note that this is a dark - matter - only simulation and baryons will be added in a later work to study the bias effect caused by the dark matter coupling ) .
given these parameters , the mass resolution of the simulation is @xmath198 with @xmath199 the solar mass .
the spatial resolution is @xmath200 on the finest refined grids and @xmath201 on the domain grid .
the high resolution in high density regions is actually necessary to ensure precision in those regions because the fifth force is generally short - ranged there .
all the simulations start at redshift @xmath202 . in principle ,
modified initial conditions ( initial displacements and velocities of particles which is obtained given a linear matter power spectrum ) need to be generated for the coupled scalar field model , because the zeldovich approximation @xcite is also affected by the scalar field coupling .
in practice , however , we find that the effect on the linear matter power spectrum is negligible ( @xmath203 ) for our choices of parameters @xmath56 .
thus we simply use the @xmath1cdm initial displacements / velocities for the particles in these simulations , which are generated using grafic @xcite again with @xmath196 and @xmath197 .
@xmath204 at present day . in figs .
[ fig : figurex1 ] , [ fig : figurex2 ] we have shown the results for the runs with @xmath205 and @xmath206 .
these choices of @xmath41 are such that for @xmath207 ( the lower two rows ) the chameleon effect is pretty strong while for @xmath208 ( the upper two rows ) it is much weaker ; the choices of @xmath47 are to see the effects of different full strengths of the fifth force .
the three panels in the first and third rows display the particle distributions at three output redshifts @xmath209 from right to left , and the three panels in the second and fourth rows show the correlation between the fifth force and gravity at these redshifts . for clearness we have only plotted a thin slice ( along the @xmath210 direction ) of the full 3-dimensional particle distribution .
first let s have a look at the first and third rows .
because we have output the scalar field value at the position of each particle together with other parameters of the particle , we also include this information in the plots
. on top of each panel the range of the value @xmath13 at the positions of all the shown particles is shown , and color is used to illustrate the amplitude of @xmath13 ( going from black to red from the minimum to the maximum value of @xmath13 ) . note that @xmath211 for @xmath65 , thus the color also indicates the value of the scalar field indirectly . also bear in mind that the same color may denote different values of @xmath13 at different redshifts .
the scalar field or @xmath13 grows with time in general .
next look at the second and fourth rows , which display the logarithmic of the magnitude of the fifth force versus that of gravity .
the color here has the same meaning as above .
furthermore , each point ( particle ) now is engaged with a square box centered on it , which denotes the size of the angle ( from @xmath212 to @xmath213 ) between the two forces ; the size of the box increases linearly with the angle , from a minimum size @xmath212 to a maximum size comparable to the largest box size shown in the entire figure . as we have mentioned above , the strength of the fifth force , if not suppressed by the chameleon mechanism , is @xmath2 times that of gravity , so we also plot the functions @xmath214 where @xmath215 are respectively the magnitudes of the fifth force and gravity , as the black solid lines in these figure , to compare with the simulation data
. we can understand these results qualitatively as follows by taking fig .
[ fig : figurex1 ] as example .
the chameleon effect is generally stronger at earlier times when matter density is high and the scalar field value is small .
as is shown in the panels of the first row , at redshift @xmath216 there is a strong contrast of the value @xmath13 in high density ( _ blue _ ) regions ( clusters hereafter ) compared to @xmath13 in the low density ( _ green , yellow and red _ ) regions ( voids hereafter ) , which is a direct reflection of the nonlinearity in the scalar field .
as time evolves and the background value of the scalar field increases , at redshift @xmath217 the chameleon effect gets suppressed ; this is manifested by the facts that ( 1 ) in the regions of small clusters the scalar field values are no longer significantly different from the background value , both of which are yellow and orange - colored , ( 2 ) even in the largest clusters the contrast between the scalar field values inside and outside ( green / light blue versus yellow / orange ) is not so strong compared with the result at redshift @xmath216 ( purple / dark blue versus green / yellow ) .
these show that the size of nonlinear regions is shrinking and the thin shells in the clusters are thickening , together leading to less nonlinear behaviors of the scalar field . at redshift
@xmath218 this tendency just becomes more obvious , leaving only a small portion of the space with chameleon effect and thin shells .
the strength of the chameleon effect is determined by several factors . in principle , the larger the effective mass of scalar field ( @xmath219 ) at a position is , the shorter - ranged the fifth force is and the less sensitive the scalar field value at this position will be to the matter distribution around it : this in turn corresponds to a stronger chameleon effect because the scalar field value is mainly determined by the local matter density . on the other hand , @xmath219 depends on @xmath41 ( smaller @xmath41 implies larger @xmath219 within a given cluster and thus stronger chameleon effect ) , @xmath47 and local @xmath57 ( larger values for these two parameters also imply larger @xmath219 and stronger chameleon effect ) and also the background value of the scalar field ( this is because the interior solution of the scalar field inside a cluster should only be solved given the boundary conditions outside the cluster , due to the differential nature of the scalar field eom . of course for very small @xmath41 and/or very large @xmath47 , the scalar field eom indeed behaves as an algebraic equation and then the influence of the background value of the scalar field is not important , but for our choices of @xmath220 that influence is significant ) .
these analysis agree well with what we have seen in the scalar field configuration from the above n - body simulation results ( first and third rows of fig .
[ fig : figurex1 ] ) . also note that the fifth force is much weaker in regions where chameleon effect is strong , because a particle there can only feel the fifth force from those particles that are very close to it : in this sense we say that the chameleon effect could suppress the fifth force .
a comparison between gravity and the fifth force can illustrate this more clearly .
remember that the chameleon effect could strongly suppress the fifth force . at low redshift ( @xmath218 )
the chameleon effect is not significant and so the fifth force is not suppressed ; in this case we find that there is a strong correlation between both the magnitudes and directions of these two forces , and the simulation results agree with the prediction eq .
( [ eq : fgcurve ] ) to very high degrees .
both forces are stronger in clusters and weaker in voids as expected .
going backwards in time to the redshift @xmath217 , there is still a good correlation but the simulation results begin to scatter over the predicted line eq .
( [ eq : fgcurve ] ) in the void regions due to the chameleon effect ( the scalar field potential term in eq .
( [ eq : intphieom ] ) becomes comparable with the matter coupling term ) .
finally , at very early times ( @xmath216 ) the scalar field value becomes so small that the potential term in eq .
( [ eq : intphieom ] ) is indeed much more important than the matter coupling term so that the latter can be neglected : we then have a complete mismatch between the numerical results and the prediction eq .
( [ eq : fgcurve ] ) as almost all data points are significantly below the black solid line .
the above mismatch in the strongly chameleon regime can be understood schematically as follows : 1 . when there is no potential for the scalar field but just a matter coupling , the fifth force is indeed long ranged and can probe the same region as gravity does . then a comparison of eqs .
( [ eq : intdpdtcomov ] , [ eq : intpoisson ] , [ eq : intphieom ] ) shows that the fifth force is exactly @xmath2 times of gravity .
this is indeed what we have observed for @xmath218 when the potential term in eq .
( [ eq : intphieom ] ) is negligible .
when the scalar field has a potential , it acquires an effective mass which is well - known to be inversely proportional to the range of the scalar fifth force . as a result
the fifth force is no longer as long range as gravity .
meanwhile , eq . ( [ eq : intdpdtcomov ] ) makes it clear that the scalar field @xmath4 ( or equivalently @xmath221 $ ] ) acts as a potential for the fifth force , and eq .
( [ eq : intphieom ] ) shows that this potential depends on the underlying matter distribution in a different ( nonlinear ) way from what the gravitational potential @xmath190 does .
+ so we could see that there will be differences between both the ranges and the magnitudes of gravity and fifth force . in high density regions these differences will be dominated over by the competing effects that ( _ i _ ) the majority part of the ( either gravitational or fifth ) force on a particle is contributed by nearby particles and there are so many particles nearby that contribution from distant particles is negligible , ( _ ii _ ) in eq .
( [ eq : intphieom ] ) the potential term is much less important than the matter coupling term , together making the fifth force behavior similar to that of gravity again . in the void regions
the number of nearby particles is small so that contribution from distant particles should be taken into account , and the matter coupling term in eq .
( [ eq : intphieom ] ) is much smaller , then the difference between gravity and the fifth force becomes manifesting .
these effects can be observed in the @xmath217 panel .
3 . at even higher redshifts
the potential of the scalar field makes its effective mass very large and thus its range very short compared with that of gravity .
meanwhile eq . ( [ eq : intphieom ] ) becomes very nonlinear due to the smallness of the scalar field value , and thus the fifth force potential depends on matter distribution very differently from the gravitational potential .
+ as a result , the fifth force reflects the matter distribution in a very small region around a given particle , while gravity probes that in a much larger region .
even in that small region where both forces exist , their magnitude can generally be very different .
so it is not surprising that the two forces look so different as in the @xmath216 panel .
note that here the fifth force is also much weaker than gravity because its strength is suppressed by the smallness of the scalar field value .
result for the @xmath115 case ( the lower two rows ) is qualitatively the same as that of the @xmath222 case above , but here because @xmath41 is much smaller , so the chameleon effect exists until much more recent than in the @xmath222 case . indeed , even at low redshifts @xmath217 and @xmath223 there is still strong contrast between the scalar field values inside and outside the clusters ( purple / dark blue versus yellow / orange ) . the correlation between the forces also follow our above analysis , but here there are some new features .
the first feature is that even today the fitting to eq .
( [ eq : fgcurve ] ) is far from perfect ; this is easy to understand , because the scalar field potential is so nonlinear that the scalar field potential term in eq .
( [ eq : intphieom ] ) is important up to now .
the second feature is that in the @xmath218 ( also @xmath217 ) panel we could find that in high density regions the fifth force does not obey eq .
( [ eq : fgcurve ] ) as in the @xmath222 case , but becomes much smaller than gravity actually , the stronger gravity is , the weaker the fifth force will be !
this is again due to the strong chameleon effect in the clusters , which makes the scalar field value very small and thus suppresses the fifth force there .
note that this feature is desirable because if we also couple the scalar field to baryonic matter then we definitely want the fifth force to be suppressed to evade solar system tests . if we choose @xmath224 as in fig .
[ fig : figurex2 ] , then all the qualitative results we have obtained above should still apply .
but the stronger coupling between matter and the scalar field enhances the chameleon effect .
this implies that a coupling whose strength is significantly larger than that of gravity could probably produce the correct amount of large scale structure as observed by suppressing the fifth force on all cosmological epochs and scales of interests to us .
the analysis above clearly shows the complexity of the fifth force and its possible effects on the nonlinear structure formation .
though in certain ( no chameleon ) limits the fifth force is greatly simplified and one can assume a modified gravitational constant in the n - body simulations as an approximation , this is evidently not the case if the scalar field potential is too much nonlinear .
of course , we have no _ a priori _ knowledge about when the chameleon effect becomes important , and full numerical simulations like the one presented here are therefore necessary to precisely study the effects of ( coupled ) scalar fields in the large scale structure . in forthcoming works
we shall analyze the nonlinear matter power spectrum , halo profile , scalar field configuration within clusters , as well as the development and disappearance of thin shells in a quantitative manner .
one may wonder if a more complete simulation should keep the time derivatives of the scalar field in eq .
( [ eq : wfphieom ] ) .
this is certainly true , yet these terms indeed have negligible effect @xcite , which could be understood in the following way : when there is no ( or very weak ) chameleon effect , the scalar field potential term in eq .
( [ eq : intphieom ] ) is negligible and so this equation has the same form as the poisson equation eq .
( [ eq : intpoisson ] ) ; as a result the quasi - static approximation works as well for the scalar field as for the gravitational potential ( which is what any n - body code relies on ) . on the other hand ,
if there is strong chameleon effect , the scalar field value tends to be much smaller which means that the time derivative of the scalar field also gets smaller ; and at the same time the spatial gradient of the scalar field becomes larger : these together indicate that the quasi - static approximation should be good here too .
to conclude , in this paper we have presented the general frameworks to study the linear and nonlinear structure formations in coupled scalar field models , and given some preliminary numerical results for both in the context of a specific coupling function and scalar field potential . for the linear large scale structure ,
we write down the perturbed field equations using the @xmath7 decomposition , which can be directly applied into numerical boltzmann codes such as camb to generate the cmb and matter power spectra . for the chosen coupling function with parameter @xmath47 and potential with parameter @xmath41 , we find that @xmath47 roughly controls the strength of the fifth force ( which is due to the propagation of the scalar field ) while @xmath41 controls how much the effect of the fifth force is suppressed by the chameleon mechanism . with @xmath60
the chameleon effect makes the model behave like @xmath1cdm on very large scales , but this is far from enough to significantly decrease the effects of the fifth force on the small scale density perturbation growth . on those small scales
, however , nonlinearity becomes an important issue , which leads us to the n - body simulation in [ sect : nonlin ] .
previous n - body simulations with scalar fields are generally simplified by certain approximations such as treating the scalar coupling effect as a simple change of gravitational constant @xmath11 , or assuming a yukawa - type force with a certain range .
these approximations do not work well in the present model , and so here we set up the formula needed for a more precise simulation , putting much emphasis on the calculation of the fifth force and its action on particles .
the equations derived and the algorithm described in [ sect : nonlin ] are general enough and should be directly applicable to other coupled scalar field models .
we integrate the equations into a modified version of the n - body code mlapm and performed several runs with different combinations of @xmath56 .
some results are displayed in figs .
[ fig : figurex1 ] and [ fig : figurex2 ] .
it is confirmed that when the chameleon effect is not important , the fifth force is parallel to gravity and is @xmath2 times stronger , meaning that simply using a different gravitational constant which is @xmath225 times as large as the bare one in the simulation should be an acceptable approximation .
there are some caveats however .
for one thing , the correlation between gravity and the fifth force is only good for a portion of the parameter space @xmath226 , and for @xmath44 the potential is so nonlinear as to destroy this correlation .
more importantly , even the correlation is perfect now it is possible that at earlier times the nonlinear effects modify it dramatically ( cf . fig .
[ fig : figurex2 ] with @xmath222 at @xmath227 ) .
this means that the above - mentioned approximation is unlikely to be correct consistently and full simulations as the ones in this paper are needed .
all in all , from the results we can spot the trend that , with the same value of @xmath41 increasing @xmath47 simply enhances the power of structure growth , while with the same value of @xmath47 decreasing @xmath41 has the effects of reducing that power . also , for small @xmath41 the configuration of the scalar field becomes very nonlinear and highly sensitive to the underlying matter distribution , while for large @xmath41 this becomes much more smooth and stiff .
all these observations ( and others as explained in [ subsect : numresult ] ) agree with the chameleon analysis .
our simulations also provides a test bed for scalar field theories . among the applications
, one could look for high - speed encounters of halos to see the probability of generating a bullet - cluster - like encounters @xcite . since this paper is only served to set up the general framework of linear and nonlinear simulations , these applications of the simulation results , as well as the matter power spectrum and scalar field profile / evolution , will be presented in forthcoming works .
the authors thank alexander knebe , claudio llinares and xufen wu for their helps in technical problems about the n - body code and its implementation , john barrow , carsten van der bruck , anne davis , george efstathiou , david f. mota and douglas shaw for encouragements to finish this work and discussions in the process , and luca amendola for listening about the work and helpful information at earlier stages .
b. li acknowledges supports from overseas research studentship , cambridge overseas trust , damtp and queens college .
we are also indebted to the hpc - europa transnational access visit programme for its support and the lorentz center and leiden observatory for hospitality when part of this work is undertaken .
the n - body simulations are performed on the sara supercomputer in the netherlands .
natexlab bibnamefont bibfnamefont citenamefont url urlprefix [ 2]#2 [ 2][]#2 , and , * * , ( ) . and , * * , ( ) . and , *
* , ( ) . and , * * , ( ) . and , * * , ( ) . , , , and , * * , ( ) . , , and , * * , ( ) . and , * * , ( ) . and , * * , ( )
. and , * * , ( ) . and , * * , ( ) . , , and , * * , ( ) . , *
* , ( ) . and , * * , ( ) .
, , and , * * , ( ) . and , * * , ( ) . and , * * , ( ) ; * * , ( ) and , * * , ( ) . , and ( ) , arxiv:0902.3452 [ astro - ph ]
. , , , and
, * * , ( ) .
, , and ( ) , arxiv:0812.3901 [ astro - ph ] . , , and , * * , ( ) . , * * , ( ) . , and , * * , ( ) . , and , * * , ( ) . and , * * , ( ) . and , * * , ( ) . , and , * * , ( ) . , * * , ( ) . , and , * * , ( ) . , * * , ( ) . , , and , _ numerical recipes in c. the art of scientific computing _
( cambridge university press , cambridge 1992 ) , second ed . , and , _ a multigrid tutorial _
( society for industrial and applied mathematics , philadelphia 2000 ) , second ed . , * * , ( ) . , , and , * * , ( ) . ) ,
arxiv : astro - ph/9506070 . , and , * * , ( ) .
consider the decomposition of the stress energy tensor @xmath228 @xmath229 where @xmath155 is the 4-velocity of an observer with respect to which @xmath7 space - time splitting is made , @xmath230 the projection tensor used to obtain covariant tensors perpendicular to @xmath155 . here
@xmath231 is the projected symmetric tracefree anisotropic stress , @xmath232 is the vector heat flux and @xmath233 and @xmath234 respectively the energy density and isotropic pressure .
these quantities could be obtained from @xmath228 through the relations @xmath235 based on these , the components of the stress energy tensor ( up to first order ) in this model is summarized in the following table : [ cols="^,^,^,^,^",options="header " , ] throughout this paper an overdot denotes the derivative with respect to the cosmic time @xmath157 and @xmath127 is the covariant spatial derivative perpendicular to @xmath155 ( up to first order in perturbation ) . note that it is the _ coupled _ cdm quantities which appear in the ( background and perturbed ) einstein equations , and also note that we have included the massless neutrinos into the model here .
the five constraint equations in the model are given as @xmath236 \epsilon_{b)ec}^{\ \ \ \ d}u^{e } ; \\ \hat{\nabla}^{b}\mathcal{e}_{ab } & = & \frac{1}{2}\kappa \left [ \hat{\nabla}^{b}\pi_{ab}+ \frac{2}{3}\theta q_{a}+\frac{2}{3}\hat{\nabla}_{a}\rho\right ] ; \\ \hat{\nabla}^{b}\mathcal{b}_{ab } & = & \frac{1}{2}\kappa \left [ \hat{\nabla}_{c}q_{d}+(\rho+p)\varpi_{cd}\right]\epsilon_{ab}^{\ \ cd}u^{b}.\end{aligned}\ ] ] here , @xmath237 is the covariant permutation tensor , @xmath238 and @xmath239 are respectively the electric and magnetic parts of the weyl tensor @xmath240 , given respectively through @xmath241 and @xmath242 . @xmath243
come from the decomposition of the covariant derivative of 4-velocity @xmath244 with @xmath245 being the acceleration , @xmath246 the expansion scalar , @xmath247}$ ] and @xmath248 the shear .
note that @xmath249 in the above section has a completely different meaning .
in addition , the seven propagation equations are : @xmath250 } & = & 0 ; \\ \frac{1}{2}\kappa \left[\dot{\pi}_{ab } + \frac{1}{3}\theta\pi_{ab}\right ] - \frac{1}{2}\kappa \left[(\rho + p)\sigma_{ab}\ + \hat{\nabla}_{\langle a}q_{b\rangle } \right ] \nonumber \\ -\left [ \dot{\mathcal{e}}_{ab } + \theta \mathcal{e}_{ab}-\hat{\nabla}^{c}\mathcal{b}_{d(a}\epsilon_{b)ec}^{\ \ \ \ d}u^{e}\right ] & = & 0 ; \\ \dot{\mathcal{b}}_{ab } + \theta\mathcal{b}_{ab}+\hat{\nabla}^{c}\mathcal{e}_{d(a}\epsilon_{b)ec}^{\ \ \ \ d}u^{e } + \frac{1}{2}\kappa \hat{\nabla}^{c}\mathcal{\pi}_{d(a}\epsilon_{b)ec}^{\ \ \ \ d}u^{e } & = & 0.\ \ \ \\end{aligned}\]]where the angle bracket means taking the trace - free part of a quantity . note that the first of eq .
( [ eq : q_dm ] ) is for normal matter while the second is for the coupled dark matter . besides the above equations , it is useful to express the projected ricci scalar @xmath251 into the hypersurfaces orthogonal to @xmath22 as @xmath252the spatial derivative of the projected ricci scalar , @xmath253 , is given as @xmath254 and its propagation equation @xmath255 as we are considering a spatially flat universe , the spatial curvature must vanish on large scales which means that @xmath256 . thus , from eq .
( [ eq : projected_ricci_scalar ] ) , we can obtain the friedman equation @xmath257 note that in all the above equations @xmath258 and @xmath231 are all the total quantities contributed by all matter species : @xmath259 finally , there is the perturbed scalar field eom @xmath260
to put the above perturbation equations into numerical calculation , we need to write them in the @xmath59-space .
as we want to present the complete framework to study the structure formation in coupled scalar field models , here we also list those equations .
first of all , the change into @xmath59-space is accomplished with the aid of following harmonic expansions : @xmath261 where @xmath262 with @xmath263 being the zero order eigenfunctions of the comoving laplacian @xmath264 ( @xmath265 ) and @xmath266 .
note that we only consider scalar mode perturbations in the present paper and so shall neglect the second order quantities such as @xmath239 and @xmath267 . with these we list the equations that will be used in the numerical calculation of the linear large scale structure .
@xmath268 note that because our definition of the dark matter energy density includes only the mass density , but not the contribution from its coupling to the scalar field , so @xmath57 evolves as in @xmath1cdm .
one can also understand this as follows : the coupling to the scalar field produces a fifth force on the dark matter particles , making their trajectories be non - geodesic , but as we shall see below , the fifth force is spatial ( perpendicular to the worldline of the dark matter particle ) and can not change the mass of dark matter particles .
consequently the averaged dark matter mass density is the same as in @xmath1cdm .
one can of course define the energy momentum tensor for dark matter as including @xmath13 , and in this case eq .
( [ eq : bg_rho_cdm ] ) will no longer be valid and we end up with varying mass dark matter particles .
@xmath269 @xmath270 @xmath271a^{2 } & = & 0,\\ k\eta ' + 2\frac{a'}{a}ka + \kappa qa^{2 } & = & 0,\end{aligned}\ ] ] in which we have used @xmath272 and defined the density contrast @xmath273 for matter species @xmath274 .
@xmath275a^{2 } & = & 0,\\ k^{2}\eta
+ 2\frac{a'}{a}k\mathcal{z } - \kappa\mathcal{x}a^{2 } & = & 0.\end{aligned}\ ] ] @xmath276 @xmath277 when it comes to perturbation calculations , we need to fix a gauge , _
i.e. _ , choose a @xmath155 .
one possibility is to use the 4-velocity of dark matter particles as our @xmath155 . in this case
the dark matter heat flux is zero and according to eq .
( [ eq : pb_v_cdm ] ) we will simply have @xmath278 , and this relation can then be used to replace the @xmath245 s appearing in all the above equations .
another possibility is to choose @xmath155 such that @xmath80 : in this case @xmath81 will become nonzero and we need to dynamically evolve it .
below in the numerical calculations we shall adopt the second possibility .
in the mlapm code the partial differential equation eq .
( [ eq : intpoisson ] ) is ( and in our modified code eq .
( [ eq : intphieom ] ) will also be ) solved on discretized grid points , and as such we must develop the discretized versions of eqs .
( [ eq : intdxdtcomov ] - [ eq : intphieom ] ) to be implemented into the code . but before going on to the discretization , we need to address a technical issue .
as the potential is highly nonlinear , in the high density regime the value of the scalar field @xmath279 will be very close to 0 , and this is potentially a disaster as during the numerical solution process the value of @xmath279 might easily go into the forbidden region @xmath171 @xcite .
one way of solving this problem is to define @xmath280 in which @xmath281 is the background value of @xmath282 , as in @xcite .
then the new variable @xmath170 takes value in @xmath283 so that @xmath284 is positive definite which ensures that @xmath285 .
however , since there are already exponentials of @xmath282 in the potential , this substitution will result terms involving @xmath286 $ ] , which could potentially magnify any numerical error in @xmath170 . instead
, we can define a new variable @xmath170 according to @xmath287 by this , @xmath170 still takes value in @xmath283 , @xmath288 and thus @xmath289 which ensures that @xmath282 is positive definite in numerical solutions .
besides , @xmath290^{\beta}$ ] so that there will be no exponential - of - exponential terms , and the only exponential is what we have for the potential itself .
@xmath291 above .
then the poisson equation becomes @xmath292\nonumber\\ & & - \frac{3\omega_{v_{0}}a^{3}}{\left[1-\left(1+e^{u}\right)^{\beta}\right]^{\mu } } + 3\bar{\omega}_{v}a^{3},\end{aligned}\ ] ] where we have defined @xmath293 which is determined by background cosmology , the quantity @xmath294 is also determined solely by background cosmology .
these background quantities should not bother us here .
the scalar field eom becomes @xmath295^{\mu+1}}\nonumber\\ & & - 3\gamma\omega_{\mathrm{cdm}}e^{\gamma\sqrt{\kappa}\bar{\varphi } } - \frac{3\mu\beta\omega_{v_{0}}a^{3}e^{\beta\sqrt{\kappa}\bar{\varphi } } } { \left[1-e^{\beta\sqrt{\kappa}\bar{\varphi}}\right]^{m+1}}\end{aligned}\ ] ] in which we have used the fact that @xmath296 , and moved all terms depending only on background cosmology ( the source terms ) to the right hand side .
so , in terms of the new variable @xmath170 , the set of equations used in the n - body code should be @xmath297 plus eqs .
( [ eq : u_poisson ] , [ eq : u_phi_eom ] ) .
these equations will ultimately be used in the code . among them , eqs .
( [ eq : u_poisson ] , [ eq : u_dpdt ] ) will use the value of @xmath170 while eq .
( [ eq : u_phi_eom ] ) solves for @xmath170 . in order that these equations can be integrated into mlapm
, we need to discretize eq .
( [ eq : u_phi_eom ] ) for the application of newton - gauss - seidel iterations . to discretize eq .
( [ eq : u_phi_eom ] ) , let us define @xmath298 .
the discretization involves writing down a discretion version of this equation on a uniform grid with grid spacing @xmath299 .
suppose we require second order precision as is in the standard poisson solver of mlapm , then @xmath300 in one dimension can be written as @xmath301 where a subscript @xmath302 means that the quantity is evaluated on the @xmath303-th point .
of course the generalization to three dimensions is straightforward .
@xmath309\nonumber\\ & & + \frac{1}{h^{2}}\left[b_{i , j+\frac{1}{2},k}u_{i , j+1,k } - u_{i , j , k}\left(b_{i , j+\frac{1}{2},k}+b_{i , j-\frac{1}{2},k}\right ) + b_{i , j-\frac{1}{2},k}u_{i , j-1,k}\right]\nonumber\\ & & + \frac{1}{h^{2}}\left[b_{i , j , k+\frac{1}{2}}u_{i , j , k+1 } - u_{i , j , k}\left(b_{i , j , k+\frac{1}{2}}+b_{i , j , k-\frac{1}{2}}\right ) + b_{i , j , k-\frac{1}{2}}u_{i , j , k-1}\right].\end{aligned}\ ] ] @xmath311\nonumber\\ & & + \frac{1}{h^{2}}\left[b_{i , j+\frac{1}{2},k}u_{i , j+1,k } - u_{i , j , k}\left(b_{i , j+\frac{1}{2},k}+b_{i , j-\frac{1}{2},k}\right ) + b_{i , j-\frac{1}{2},k}u_{i , j-1,k}\right]\nonumber\\ & & + \frac{1}{h^{2}}\left[b_{i , j , k+\frac{1}{2}}u_{i , j , k+1 } - u_{i , j , k}\left(b_{i , j , k+\frac{1}{2}}+b_{i , j , k-\frac{1}{2}}\right ) + b_{i , j , k-\frac{1}{2}}u_{i , j , k-1}\right]\nonumber\\ & & -\frac{\left(h_{0}b\right)^{2}}{ac^{2}}\left[3\gamma\omega_{\mathrm{cdm}}\rho_{c , i , j , k } \left(1+e^{u_{i , j , k}}\right)^{\gamma } + \frac{3\mu\beta\omega_{v_{0}}a^{3}\left(1+e^{u_{i , j , k}}\right)^{\beta } } { \left[1-\left(1+e^{u_{i , j , k}}\right)^{\beta}\right]^{\mu+1}}\right]\nonumber\\ & & + \frac{\left(h_{0}b\right)^{2}}{ac^{2}}\left [ 3\gamma\omega_{\mathrm{cdm}}e^{\gamma\sqrt{\kappa}\bar{\varphi } } + \frac{3\mu\beta\omega_{v_{0}}a^{3}e^{\beta\sqrt{\kappa}\bar{\varphi } } } { \left[1-e^{\beta\sqrt{\kappa}\bar{\varphi}}\right]^{\mu+1}}\right].\end{aligned}\ ] ] then the newton - gauss - seidel iteration says that we can obtain a new ( and often more accurate ) solution of @xmath170 , @xmath312 , using our knowledge about the old ( and less accurate ) solution @xmath313 as @xmath314 the old solution will be replaced by the new solution to @xmath315 once the new solution is ready , using the red - black gauss - seidel sweeping scheme .
note that @xmath316\nonumber\\ & & -\frac{1}{2h^{2}}\left[b_{i+1,j , k}+b_{i-1,j , k}+b_{i , j+1,k } + b_{i , j-1,k}+b_{i , j , k+1}+b_{i , j , k-1}+6b_{i , j , k}\right]\nonumber\\ & & -\frac{\left(h_{0}b\right)^{2}}{ac^{2}}3\gamma^{2}\omega_{\mathrm{cdm}}\rho_{c , i , j , k } \left(1+e^{u_{i , j , k}}\right)^{\gamma}b_{i , j , k}\nonumber\\ & & -\frac{\left(h_{0}b\right)^{2}}{ac^{2 } } \frac{3\mu\beta^{2}\omega_{v_{0}}a^{3}\left(1+e^{u_{i , j , k}}\right)^{\beta } } { \left[1-\left(1+e^{u_{i , j , k}}\right)^{\beta}\right]^{\mu+1}}b_{i , j , k } \left[1+(\mu+1)\frac{\left(1+e^{u_{i , j , k}}\right)^{\beta}}{1-\left(1+e^{u_{i , j , k}}\right)^{\beta}}\right].\end{aligned}\ ] ] in principle , if we start from a high redshift , then the initial guess of @xmath315 could be such that the initial value of @xmath282 in all the space is equal to the background value @xmath281 , because anyway at this time we expect this to be approximately true . for subsequent time steps we could use the solution for @xmath315 at the previous time step as our initial guess ; if the time step is small enough then we do not expect @xmath170 to change significantly between consecutive times so that such a guess will be good enough for the iteration to converge fast . | we lay out the frameworks to numerically study the structure formation in both linear and nonlinear regimes in general dark - matter - coupled scalar field models , and give an explicit example where the scalar field serves as a dynamical dark energy . adopting parameters of the scalar field which yield a realistic cmb spectrum
, we generate the initial conditions for our n - body simulations , which follow the spatial distributions of the dark matter and the scalar field by solving their equations of motion using the multilevel adaptive grid technique .
we show that the spatial configuration of the scalar field tracks well the voids and clusters of dark matter .
indeed , the propagation of scalar degree of freedom effectively acts a fifth force on dark matter particles , whose range and magnitude are determined by the two model parameters @xmath0 , local dark matter density as well as the background value for the scalar field .
the model behaves like the @xmath1cdm paradigm on scales relevant to the cmb spectrum , which are well beyond the probe of the local fifth force and thus not significantly affected by the matter - scalar coupling . on scales comparable or shorter than the range of the local fifth force ,
the fifth force is perfectly parallel to gravity and their strengths have a fixed ratio @xmath2 determined by the matter - scalar coupling , provided that the chameleon effect is weak ; if on the other hand there is a strong chameleon effect ( _ i.e. _ , the scalar field almost resides at its effective potential minimum everywhere in the space ) , the fifth force indeed has suppressed effects in high density regions and shows no obvious correlation with gravity , which means that the dark - matter - scalar - field coupling is not simply equivalent to a rescaling of the gravitational constant or the mass of the dark matter particles .
we show these spatial distributions and ( lack of ) correlations at typical redshifts ( @xmath3 ) in our multi - grid million - particle simulations .
the viable parameters for the scalar field can be inferred on intermediate or small scales at late times from , _
e.g. _ , weak lensing and phase space properties , while the predicted hubble expansion and linearly simulated cmb spectrum are virtually indistinguishable from the standard @xmath1cdm predictions . |
magnetic fields play an important role in the present day universe .
they pervade entire galaxies including our own milky way ( e.g. * ? ? ?
* ; * ? ? ?
* ) as well as the intergalactic medium ( e.g. * ? ? ?
* ; * ? ? ?
recent observations of tev blazars with fermi ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) provide hints about the strength of the intergalactic magnetic field .
the question , however , where these fields come from is still not conclusively answered , in particular how were the magnetic fields initially generated , and how amplified to the strength observed nowadays .
the first seed magnetic fields in the early universe could have been generated by a couple of different mechanisms .
one mechanism is the generation of magnetic fields in the phase of inflation in the very early universe resulting in a field strength of @xmath7 @xmath8
g on scales of about 1 mpc @xcite .
another possibility is the generation via first order phase transitions , with field strengths of @xmath9 g from the electroweak phase transition and @xmath10 g from the qcd phase transition on scales of 10 mpc @xcite .
the actual coherence length , however , depends crucially on the properties of the field , like e.g. the magnetic helicity and the spectral index of the magnetic power spectrum on large scales ( e.g. * ? ? ?
another process is the so - called `` biermann battery '' @xcite where due to the different behaviour of electrons and ions in a plasma magnetic fields with @xmath11 g can be generated .
more recently , @xcite showed that plasma fluctuations can generate magnetic fields of the order of @xmath8 g even in a completely unmagnetized , non - relativistic medium .
for comparison , in present day galaxies kpc - scale magnetic fields with strengths of a few @xmath12 g are found ( e.g. * ? ? ?
* ; * ? ? ?
for this reason at some point a significant amplification of primordial magnetic fields on large scales has to take place .
one way to accomplish this is the magnetic dynamo which amplifies the magnetic field on various scales , which is why the effects on large and small scales are usually considered separately . on large scales
the so - called mean - field dynamo describes the amplification in a differentially rotating fluid in the presence of helical turbulent motions ( e.g. * ? ? ?
* ; * ? ? ?
* ) . on scales smaller than the driving scale of turbulence
the small - scale dynamo is responsible for the amplification of the magnetic field by converting the turbulent kinetic energy into magnetic energy by stretching and twisting magnetic field lines @xcite . in the so - called kinematic phase of the small - scale dynamo the magnetic energy experiences an exponential growth mainly at the resistivity scale . once the magnetic field has grown to a significant strength , the non - linear phase sets in , in which the magnetic energy is transported to larger and larger scales @xcite .
the amplification ends once the magnetic field has reached saturation , i.e. when the magnetic energy is of the order of 10% of the turbulent kinetic energy on the driving scale of the turbulence .
the formation of the first stars , so called population iii stars ( hereafter pop iii stars ) , in the universe has taken place in dark matter halos with typical masses of 10@xmath2 m@xmath4 ( e.g. * ? ? ? * ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . by means of an analytical consideration
, @xcite have shown that the small - scale dynamo can amplify an initially very weak magnetic field in such a halo to strengths of @xmath13 g within short timescales , a mechanism which could also work during the formation of the first galaxies @xcite .
these analytical predictions were recently confirmed by a number of numerical simulations of pop iii star formation @xcite .
the authors showed that when using a high enough spatial resolution the magnetic field in the center of the halo indeed can reach saturation .
these magnetic fields are usually small - scaled , strongly tangled , and localized in the center of the star forming halos . for this reason
, the field can not contribute to the large - scale , volume - filling magnetic field observed nowadays unless it is removed from the center and transported to larger scales .
as shown by @xcite , pop iii stars with masses between 10 and 50 m@xmath4 will end their lives with a core collapse supernova and those with masses from 140 to 260 m@xmath4 with an up to 100 times more energetic pair instability supernova .
the effect of supernova explosions on primordial halos was studies numerically by a number of authors @xcite .
a common result of these simulations is that whether a halo is disrupted by a supernova or not crucially depends on the mass of the halo , the injection energy of the supernova , and prior to that on the ability of the progenitor star to ionize the halo and drive out gas in its inner region . even in the case that a hii regions was created , a too small injection energy or a too high halo mass would cause the expansion of the supernova bubble to stall and the gas to start to fall back towards the center . in case
the supernova is able to disrupt the halo , however , material from the center of the halo is ejected and transported to distances of the order of kpc .
so far , in these simulations the effect on magnetic fields was not studied .
we note , however , that @xcite considered the effect of a single supernova on a magnetized medium with homogeneous density in the present day interstellar medium on scales of @xmath14 100 pc thus significantly smaller than the scale of interest here .
the authors show that the magnetic field is swept up in the supernova shell , which can also be expected for primordial halos affected by supernovae .
how exactly the magnetic field on large ( kpc ) scales is affected in a primordial halo will be the main focus of the work presented here .
the paper is organized as follows : in section [ sec : techniques ] we describe the numerical techniques used for the simulations . in section [ sec : ic ] we describe the initial conditions used for the simulated primordial halos .
the results of the simulations are presented in section [ sec : results ] , where we first give a general overview of the simulations before we consider the properties of the magnetic field and the effect of varying conditions .
next , the results are discussed in a broader context in section [ sec : discussion ] before we summarize our work in section [ sec : conclusion ] .
the simulations were preformed with the versatile astrophysical code flash in version 4 @xcite .
we use the unsplit , staggered mesh solver , which comes along with the public flash version 4 @xcite , in order to solve the magnetohydrodynamical equations under the condition of ideal magnetohydrodynamics .
the solver implicitly satisfies the divergence - free constraint of the magnetic field up to the accuracy of machine round - off errors .
we used a third order hydro scheme and a hybrid riemann solver using the hll scheme at shocks and the hlld scheme otherwise in order to guarantee the best stability and accuracy . in all simulations the adiabatic index
is set to @xmath15 .
we do not use self - gravity in this work in order to speed up the simulations .
we instead model the gravitational effect of baryonic and dark matter by assuming a spherically symmetric gravitational potential which remains constant over time . despite the fact that the ( baryonic ) density distribution changes over time we consider it as a reasonable approximation for two reasons : firstly , on the timescales considered the dark matter distribution is not expected to change significantly .
secondly , at the scales of interest ( @xmath16 10 pc 1 kpc ) the gravitational potential is dominated by the dark matter which is why the change in the baryonic matter distribution would only have a minor effect on the overall gravitational potential .
we point out that for a mean dark matter density in the halos of @xmath17 ( see equation [ eq : mdm2 ] further below ) the free - fall time is about 30 myr thus significantly longer than the timescales considered here ( @xmath14 10 myr ) , which also justifies the use of a fixed gravitational potential .
we emphasize that a time - independent gravitational potential was also used by a number of other authors @xcite despite the fact that moderate changes in the dark matter profile could be introduced by supernova explosions @xcite .
we performed two kinds of simulations , simulations with an adiabatic behaviour ( @xmath15 ) , where no additional cooling is applied , and simulations using a tabulated cooling function appropriate for primordial gas with zero metallicity as a sink of thermal energy after the hydrodynamical timestep .
for the latter simulations we use the cooling rates @xmath18 given by @xcite .
these include collisional excitation and ionization cooling for h and he , recombination cooling for h and he , and bremsstrahlung .
we note that we are missing an accurate description for compton cooling for a redshift of z = 20 , which is why we most likely somewhat underestimate the cooling efficiency ( see also section [ sec : caveats ] ) . after each hydrodynamical timestep
we update the internal energy by @xmath19 where @xmath20 is the hydrogen number density and d@xmath21 the hydro timestep . by using a subcycling scheme for the cooling routine
we guarantee that the internal energy does not change by more than 20% within one cooling timestep . in case
more that one cooling timestep is required ( since the internal energy would change by more than 20% otherwise ) the cooling rate @xmath18 , which depends on the temperature of the gas , is updated accordingly .
we now describe the initial conditions used in our simulations .
we performed simulations with different masses of the dark matter halo , varying magnetic field strengths , and two different scaling relations between the magnetic field and the gas density .
moreover , we simulated the effect of supernova explosions with different amounts of energy . finally , as mentioned already before we repeated some selected simulations without cooling , i.e. with a purely adiabatic behaviour in order to investigate the effect of the cooling . all simulations and their corresponding parameters are listed in table [ tab : models ] .
the simulations where performed in a cubic box with a side length of 1024 pc using a uniform resolution of 512@xmath22 grid points , i.e. a spatial resolution of 2 pc . in order to cover a larger physical scale we have repeated one of the simulations ( m3_b1_pisn ) with a two times larger simulation box , i.e. a cube with a side length of 2048 pc .
the inner region , a cube with a side length of 1024 pc , has the same spatial resolution of 2 pc as in the original run whereas in the outer parts we have a two times lower resolution of 4 pc .
hence , the grid structure is that of a nested grid with a 512@xmath22 grid placed in the center of second 512@xmath22 grid with a two times lower resolution .
we point out that we used highly idealized simulations described in the following to probe the influence of supernovae on the magnetic field evolution .
this idealization was necessary in order to be able to cover the aforementioned parameter space with a reasonable number of simulations and to be able to use the high resolution required for a proper analysis of the magnetic field while simultaneously keeping the computational costs at a reasonable level . [ cols="<,^,^,^,^,^,^,^ " , ] next , we investigate the effect of the halo mass . since we have seen in section [ sec : bvary ] that changing the magnetic field strength over a wide range does not significantly alter the results , we concentrate on a single value for the magnetic field . in the left panel of fig . [
fig : mvary ] we plot the spectra of the runs m1_b1_sn , m1_b1_pisn , m2_b1_sn , m2_b1_pisn , and m3_b1_pisn at the end of each simulation .
we point out that for run m3_b1_sn showing recollapse the spectrum as well as the autocorrelation function are not shown . for the runs m1_b1_sn , m1_b1_pisn , m2_b1_pisn , and m3_b1_pisn , where the remnant is able to disrupt the halo and escape from the gravitational potential well , the spectra peak at low @xmath5-values of the order of 5 10 in agreement with the results of the previous sections . for run
m2_b1_sn is less clear . here
the supernova remnant has already started to recollapse in the central regions , whereas in the outer regions the remnant keeps expanding even after 10 myr .
hence , we see a bimodal spectrum with one peak coming from the ongoing expansion and the peak at small scales from the beginning recollapse .
the autocorrelation functions for the runs m1_b1_sn , m1_b1_pisn , m2_b1_sn , m2_b1_pisn , and m3_b1_pisn are shown in the right panel of fig .
[ fig : mvary ] .
for the all runs except run m2_b1_sn the correlation length increases by a factor @xmath16 10 reaching values of the order of 100 200 pc .
however , it can be seen that for the runs with the lower halo masses the correlation seems to be less strong on larger scales compared to run m3_b1_pisn ( black line ) as indicated by the steeper decline in @xmath23 .
this nicely agrees with the finding that the spectrum of run m3_b1_pisn is the one with the most magnetic energy on large scales , i.e. that it peaks at the smallest @xmath5-value ( see left panel of fig .
[ fig : mvary ] ) . for run m2_b1_sn ( violet line )
the decline is even stronger in agreement with the double - peaked spectrum .
recently , @xcite have shown that the cooling ability of primordial gas can be significantly reduced in regions close to a source emitting large amounts of ionizing radiation .
for this reason , we have simulated an extreme case in which the cooling is completely turned off , i.e. the gas behaves purely adiabatic .
we first point out that in run m2_b1_sn_ad and m3_b1_sn_ad the core collapse supernova manages to disrupt the halo in contrast to the runs m2_b1_sn and m3_b1_sn with cooling .
this can be understood on the basis of a simple energy argument : the gravitational binding energies of the baryonic gas for the runs with a total halo mass of @xmath24 and @xmath25 m@xmath4 are @xmath26 and @xmath27 erg , respectively .
since both values are smaller than the explosion energy of the supernova ( e@xmath28 erg ) and since no energy is lost by cooling processes , the supernova can disrupt the halo .
we now show the magnetic field spectra for the runs m2_b1_sn_ad , m3_b1_sn_ad , m3_b1_pisn_ad , and m3_b4_pisn_ad at the end of each run in fig .
[ fig : cool ] . .
] the autocorrelation function for those runs ( not shown here ) agree well with the results of the spectrum analysis .
as can be seen in fig .
[ fig : cool ] , the spectra of the runs without cooling in general peak a somewhat higher @xmath5-values than the corresponding runs with cooling .
the higher value of k@xmath29 is similar to the findings in the runs with a low halo mass ( the runs m1_b1_sn , m1_b1_pisn , and m2_b1_pisn ) as well to the findings of run m3_b1_5pisn , which all peak at somewhat smaller @xmath5-values than run m3_b1_pisn .
we attribute this behaviour to the same physical origin : for the lower - mass halos the cooling ability of the gas is reduced due to the overall lower densities ( see equation [ eq : cooling ] ) and the gas requires a longer time to cool .
similarly , for run m3_b1_5pisn a longer cooling time is required due to the higher explosion energy . in consequence
, in both cases the cooling time scale is prolonged when compared to the typical expansion timescale of the supernova .
for this reason , the evolution of the gas is more similar to an adiabatic evolution like in the runs in which cooling is turned off completely .
hence , for these runs as well as the runs without cooling a rather thick shell develops in contrast to the thin shock fronts seen in the other runs .
this is demonstrated exemplarily for run m3_b1_pisn_ad in fig .
[ fig : nocool ] where the density and magnetic energy at a slice through the center are shown .
the high - density region behind the shock front is much thicker than for the case with cooling ( see fig . [
fig : slicessn53 ] ) and contains significant , small - scale magnetic field fluctuations not observed in run m3_b1_pisn .
these small - scale fluctuations cause the spectra to peak at a higher values of k. similar results are also found for the runs m1_b1_sn , m1_b1_pisn , m2_b1_pisn , and m3_b1_5pisn , hence for all the runs in which the cooling time is increased relative to dynamical timescale .
we consider the final fate of the supernova remnant in dependence of the halo mass and the injection energy and compare the results with those from other numerical studies . to remind the reader we find that a core collapse supernova ( @xmath30 erg ) can not disperse halos as massive as @xmath31 m@xmath4 or above .
also for the halo with @xmath24 m@xmath4 ( run m2_b1_sn ) we see signs of recollapse in the central region although the outer regions keep expanding . hence , this mass seems to be roughly the borderline between expansion and recollapse .
we note , however , that we would have to include a more detailed treatment of the thermodynamics as well as radiation transfer in order to make a definite statement about the final state of the supernova remnant . in general , however , our findings are in agreement with recent works of @xcite and @xcite modelling the evolution of a core collapse supernova in a halo of @xmath16 10@xmath2 m@xmath4 .
two further parameter studies on the final fate of primordial supernova without magnetic fields were performed by @xcite and @xcite .
we combine our results with the results of these authors in fig . [
fig : phase ] showing the final fate of the supernova in dependence of the halo mass and the explosion energy . as can be seen
, our results agree reasonably well with the results of @xcite and @xcite .
the combination of the different results suggest that supernovae with an explosion energy of @xmath32 erg are able to disrupt halos with masses up to @xmath16 @xmath33 m@xmath4 . for pisn
, our results agree with that of @xcite and @xcite .
the apparent contradiction between the result of @xcite and @xcite might be due to the fact that the latter authors initialized their explosions with purely thermal energy , which might be radiated away too quickly @xcite .
a fraction of the explosion energy should rather be inserted as kinetic energy as it was done in this work .
hence , this would imply a mass limit of the of about @xmath34 m@xmath4 , up to which halos can be dispersed by a pisn .
our findings that the core collapse supernova is not able to disrupt the halo with a mass of @xmath25 m@xmath4 despite the explosion energy being somewhat larger than the gravitational binding energy of the baryons ( see section [ sec : nocool ] ) agrees with the findings of @xcite : for massive halos explosion energies significantly larger than the gravitational binding energy are required to disrupt the halo .
this is due to the efficient radiative cooling at high densities removing energy before it can contribute to the disruption of the halo .
overall , we find that despite the strong simplifications made to model our initial conditions and the highly idealized treatment of the thermodynamical gas properties our results concerning the final fate of the halo are in reasonable agreement with simulations including a much more sophisticated treatment of the physical processes involved @xcite
. from the magnetic field spectra discussed before we can infer the time evolution of the coherence length @xmath35 of the magnetic field over time . for the sake of simplicity
we define the coherence length as @xmath36 where @xmath37 is the ( dimensionless ) @xmath5-value of the peak of the spectra from our simulations . since the expansion of the supernova remnant is not linear in time ( see fig .
[ fig : rshell ] ) and since we are more interested in how the coherence length relates to the radius of the supernova shell @xmath38 , in fig .
[ fig : lb ] we plot @xmath35 against the @xmath38 for several selected simulations
. of the magnetic field plotted against the radius of the supernova shell for the runs m1_b1_sn , m1_b1_pisn , m3_b1_pisn , m3_b1_5pisn , and m3_b4_pisn .
for all cases the coherence length reveals an approximately linear scaling with @xmath38 . to guide the readers eye
, we show a linear relation of @xmath39 and @xmath40 , respectively ( black lines ) . ]
as can be seen , the coherence length @xmath35 keeps on increasing with increasing shell radius for all cases considered .
the two black lines show an assumed linear scaling of the coherence length @xmath35 as @xmath41 and @xmath42 .
it can be seen that the coherence length of the magnetic field inside a supernova bubble reaches values of the order of 100 300 pc ( see also table [ tab : results ] ) and thus of the order of 20% 50% of the radius of the shell .
for late times in run m3_b1_pisn the coherence length even reaches values of up to @xmath16 60% of the shell radius .
we note that in order to improve the readability of the plot we do not show the results of the runs m3_b2_pisn and m3_b3_pisn , in particular since qualitatively they do not differ too much from run m3_b1_pisn .
@xcite simulate the evolution of a pisn from redshift 20 to 12 , i.e. over 200 myr until the shock front stalls . by this time
the shock front has reached a distance of about 2.5 3 kpc from the progenitor star .
hence , under the assumption that the approximately linear relation between the coherence length of the magnetic field and the supernova shell radius found in our work holds over the _ entire _ evolution of the remnant , this results in ordered magnetic fields on scales as large as @xmath16 1.5 kpc .
moreover , as pointed out in the sections [ sec : spectra ] and [ sec : bvary ] , the amount of magnetic energy on large scales remains approximately constant or even increase over time at the cost of the small - scale fluctuations .
hence , for a region affected by a supernova one could expect a significant magnetic field to be present on large scales even at late times .
depending on the efficiency of the small - scale dynamo during the lifecycle of the pop iii stars , we find a large - scale ( @xmath43 100 pc ) magnetic field component with a strength of up to 10@xmath44 g , at least for an initially saturated field .
such a large - scale magnetic field as discussed before could have a significant effect on the possible recollapse of the bubble interior , on the ensuing potential formation of supermassive black holes @xcite , and on the overall fragmentation properties of the gas and the ability to form protostellar disks @xcite .
the latter authors showed that for primordial minihalos with a well - ordered magnetic field and a mass - to - flux ratio @xcite below @xmath45 fragmentation is largely reduced . in order to make some predictions about the expected fragmentation properties during a potential second collapse in our case , we approximate the mass - to - flux - ratio by @xmath46 with the critical mass - to - flux ratio @xmath47 and the gravitational constant g @xcite , the coherence length @xmath35 of the magnetic field ranging from 100 500 pc ( see fig . [
fig : lb ] ) , @xmath48 g @xmath49 ( see bottom left panel of fig . [
fig : slicessn53 ] ) , and the field strength @xmath50 from @xmath51 g @xmath52 g ( see table [ tab : results ] ) . with these values
we obtain mass - to - flux ratios in the range of 0.02 1500 . hence , under the reasonable assumption that during a subsequent recollapse the mass - to - flux ratio is more or less preserved and based on the results of @xcite , we would expect a relatively low degree of fragmentation in such a minihalo undergoing a second round of star formation .
as shown here , in regions with a high halo density the occurrence of several supernovae might lead to the generation of a volume filling , relatively well - ordered magnetic field of significant strength .
in particular , the bottom - up formation of the first galaxies is believed to occur in regions with a significantly higher halo density than predicted under the assumption of a homogeneous distribution .
@xcite find a mean separation between potentially star forming halos of the order of @xmath53 kpc .
since this separation is of the order of or even somewhat smaller than the typical coherence length of the magnetic field inferred in this work ( @xmath54 kpc ) , one could indeed expect a volume filling , large - scale magnetic field in regions where the first galaxies have formed . as shown by @xcite , only stars with masses between @xmath16 10
50 m@xmath4 and between 140 260 m@xmath4 will end their lives with a core collapse supernova or a pisn , respectively . in this context
, we point out that despite the expected fragmentation in primordial halos and the formation some low - mass pop iii stars @xcite , recent simulations have shown that star - forming halos should host pop iii stars with typical masses ranging from 10 m@xmath4 up to about 50 m@xmath4 limited by radiative feedback @xcite but also a good fraction with masses in excess of 100 m@xmath4 @xcite .
hence , one can expect that masses of pop iii stars indeed lie in the range suitable for a core collapse supernova or a pisn and that each star forming halo is affected by one or more supernovae .
moreover , around a redshift @xmath55 the press - schechter formalism @xcite predicts only few halos to have virial masses significantly above @xmath33 m@xmath4 .
therefore it is reasonable to assume that despite the fact that a core collapse supernova can only disrupt halos up to a few 10@xmath2 m@xmath4 as discussed in section [ sec : disruption ] most of the times a halo is forming pop iii stars , it will be disrupted by a subsequent supernova resulting in the build - up of a large - scale magnetic field .
hence , the volume affected by supernovae during the assembly of first galaxies seems to depend crucially on the exact number of star forming halos . @xcite and @xcite simulating the formation of a first galaxy including stellar feedback
find that of the order of 10 halos form pop iii stars prior to the formation of the galaxy .
the simulations presented here thus suggest that a significant fraction of the volume of a region forming a protogalaxy might indeed be pervaded by a large - scale magnetic field of considerable strength .
such a magnetic field configuration would possibly also affect the formation of the first galaxies and in particular their early evolution : it is usually assumed that the initial magnetic field in protogalaxies is rather small - scaled .
subsequently , a kpc - scale magnetic field of considerable strength builds up within about 10@xmath56 yr or less due to the small - scale dynamo @xcite .
this field serves as a seed for the mean - field dynamo responsible for the large - scale , well - ordered magnetic field observed in present day galaxies ( e.g. * ? ? ?
hence , a large - scale magnetic field from the very beginning as suggested here could possibly allow the mean - field dynamo to act at an even earlier phase thus potentially affecting the later evolution of the galaxy .
finally , we note that the emergence of highly energetic pair instability supernovae in lower - mass halos could also contribute to the magnetization of the intergalactic medium ( igm ) .
@xcite showed that supernova - driven winds in dwarf galaxies can magnetize the igm by a substantial degree starting around z = 10 .
we tentatively suggest that depending on the frequencies of pair instabilities supernovae in primordial halos the magnetization of the igm could have started already around z = 20 . a magnetization on kpc scale with field strengths of 10@xmath57 10@xmath44 g found in this work ( see table [ tab : results ] ) would be in good agreement with recent faraday rotation measurements of @xcite as well as with fermi observations , revealing lower bounds of the magnetic field on kpc scales of the order of @xmath58 g ( e.g. * ? ? ?
* see their figure 2 . ) .
we emphasize that in order to properly analyse the properties of the magnetic field the main focus of this work we had to use a high and homogeneous spatial resolution of 512@xmath22 grid points in our simulations . on the other hand , simulations making use of the adaptive mesh refinement algorithm and refining e.g. for the gas density ,
usually reveal a well resolved region in the center of the halo but rather poorly resolved outer parts .
such a grid structure , however , would be highly disadvantageous for analysing the magnetic field structure by means of a fourier transformation as done in section [ sec : spectra ] .
moreover , due to the large number of runs required to explore the parameter space consisting of the halo mass , the magnetic field strength and the explosion energy a fully self - consistent treatment of all relevant physical processes like self gravity , chemistry , cooling and radiation transport is currently not possible .
for this reason we have not applied self gravity in the simulations but used a time - independent gravitational potential .
furthermore , we modelled the thermodynamical behaviour of the gas by means of a tabulated cooling curve @xcite making use of a subcycling scheme for updating the thermal energy .
it was argued that such a subcycling scheme might be too inaccurate and that the global minimum of the courant timestep and the cooling timestep should rather be used to update the hydrodynamics ( e.g. * ? ? ? * in particular their section 2.4 ) .
however , since we are mainly interested in how the supernova globally reorganises the magnetic field structure , for our purposes we consider our approach as sufficient .
we furthermore note that due to the lack of the compton cooling process in our cooling routine , we somewhat underestimate the real cooling efficiency .
this , in turn , probably causes the final radii of our remnants ( fig .
[ fig : lb ] ) to be somewhat too large .
however , since our results concerning the radial expansion agree reasonably well with that of more sophisticated simulations ( see , e.g. fig . 7 of * ? ? ?
* and fig .
10 of @xcite ) , the actual effect of the reduced cooling efficiency on the remnant size is hard to assess
. finally , we note that the cooling efficiency in our work as well as the works mentioned before might generally be too low since cosmic ray acceleration and subsequent emission due to second order fermi processes in the shock front have not been considered @xcite .
this effect could again result in somewhat smaller final radii of the supernova remnants .
we suppose that a more sophisticated inclusion of all effects mentioned before would cause at least 10 most likely even more times higher computational costs . however , since we made a great effort to generate proper initial conditions in particular for the magnetic field and since we find a good agreement with more sophisticated simulations ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , we consider our simplified approach as justified .
we note that setting the initial density in the supernova remnant and the hii region to a constant value ( see fig .
[ fig : ic ] ) , is a rather crude approximation .
however , we would like to point out that due to the lack of spatial resolution we were not able to self - consistently model the earliest evolutionary phase of the supernova .
moreover , at later times density and velocity profiles in our simulations are in rough agreement with other simulations ( e.g. * ? ? ?
* ; * ? ? ?
for this reason , and since we are mainly interested in the late stages of the supernova evolution , we consider the approximations made in our initial setup as sufficient . however , as shown by @xcite , a reverse shock , occurring once the supernova shock front hits the edge of the hii region , might trap a fraction of the magnetic energy in the center thus possibly reducing the strength of the large - scale magnetic field , a process we are clearly lacking in our setup . hence , in combination with the possible overestimation of the final size of the remnants ( due to a simplified cooling description as discussed above ) we suggest that the values concerning the coherence length and strength of the large - scale magnetic field given in this work ( table [ tab : results ] ) should rather be considered as upper limits .
the simulations presented here can also serve as a valuable guide for future and more sophisticated simulations including a better treatment of processes like gravity , chemistry , and/or radiation transfer .
in particular , it is planned to repeat some of the selected simulations with self gravity and a proper treatment of the chemistry making use of a new primordial chemistry module implemented in flash @xcite . by doing
so we can test to what extent the magnetic field properties depend on the proper treatment of these processes .
moreover , as a next future step it is also planned to take the initial conditions directly from cosmological simulations ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we have presented simulations of core collapse supernovae and pair instability supernovae going off in the center of magnetized dark matter halos .
we have varied the degree of magnetization as well as the halo mass ( @xmath59 @xmath60 m@xmath4 ) in order to infer systematical effects on the evolution of the magnetic field .
the simulations were performed using a tabulated cooling function in order to model the thermodynamical behaviour of the gas . despite the simplified physical treatment ,
our simulations agree well with results from other authors and thus serve as a useful guide for more realistic simulations .
our simulations suggest that supernovae with a total explosion energy of 10@xmath0/10@xmath1 erg are not able to disrupt halos more massive than 10@xmath2 m@xmath4/10@xmath3 m@xmath4 in agreement with other studies ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . in case
the supernova manages to disrupt the halo , a thin shock layer develops separating the post - shock from the pre - shock material .
the initially small - scale , strongly disordered magnetic field is transported with the expanding gas resulting in a significant increase in the coherence length of the gas .
the magnetic field spectra show a shift of the peak down to @xmath5-values of about 4 at the end of the simulations corresponding to a length scale of up to 250 pc .
the typical correlation length of the magnetic field determined from autocorrelation functions at the end of the simulations is of the order of 150 200 pc in good agreement with the magnetic field spectra .
we note that the maximum reachable coherence length in this work was limited by the size of the simulation domain . for gas with a reduced cooling efficiency a thicker post - shock region builds up and the spectra peak at larger @xmath5-values .
in general we find that for halos being disrupted by a supernova the large - scale component of the magnetic field has a strength ranging from about @xmath61 g to @xmath62 g strongly depending on the initial magnetic field strength .
moreover , the coherence length @xmath35 of the magnetic field scales approximately linear with the extension of the supernova bubble being about 20% 50% of the supernova shell radius .
extrapolating this relation for later times we find that for a typical pisn reaching an extension of up to 3 kpc @xcite well - ordered magnetic fields with coherence lengths of up to 1.5 kpc can be expected .
we note that due to the simplified treatment of the gas cooling and the initial modelling of the hii region and supernova blast wave , the above values should rather be considered as upper limits .
we discussed a number of implications of our findings for several subsequent processes .
we suggest that magnetic fields with this strength and coherence length can have significant implications for a subsequent recollapse of the halo in particular by suppressing fragmentation @xcite .
furthermore , such fields could potentially affect regions of a high halo density where several star forming halos with supernova explosions are expected , e.g. in the regions where the first galaxies are believed to form .
our simulations suggest that these regions might have a significant large - scale magnetic field component from the very beginning
. this could affect the formation of first galaxies and the onset of the mean - field dynamo responsible for the large - scale magnetic field observed in galaxies nowadays .
finally , the occurrence of supernovae might have also contributed to the magnetization of the intergalactic medium already around a redshift of z = 20 .
the authors like to thank the anonymous referee for the very thorough report which helped to significantly improve the paper .
the authors also like to thank s. bovino and m. latif for stimulating discussions .
we acknowledges funding by the deutsche forschungsgemeinschaft via grants ba 3706/3 - 1 and schl 1964/1 - 1 within the spp _ the interstellar medium_. drgs also thanks for funding from the german science foundation via the sfb 963/1 `` astrophysical flow instabilities and turbulence '' ( project a12 ) .
the simulations presented here were performed on the supercomputer juqueen at the supercomputing centre in jlich .
the flash code was partly developed by the doe - supported alliances center for astrophysical thermonuclear flashes ( asc ) at the university of chicago . ,
a. , fisher , r. , graziani , c. , et al .
2008 , in astronomical society of the pacific conference series , vol .
385 , numerical modeling of space plasma flows , ed .
n. v. pogorelov , e. audit , & g. p. zank , 145 | the first supernova explosions are potentially relevant sources for the production of the first large - scale magnetic fields . for this reason we present a set of high resolution simulations studying the effect of supernova explosions on magnetized , primordial halos .
we focus on the evolution of an initially small - scale magnetic field formed during the collapse of the halo .
we vary the degree of magnetization , the halo mass , and the amount of explosion energy in order to account for expected variations as well as to infer systematical dependencies of the results on initial conditions .
our simulations suggest that core collapse supernovae with an explosion energy of 10@xmath0 erg and more violent pair instability supernovae with 10@xmath1 erg are able to disrupt halos with masses up to a about 10@xmath2 and 10@xmath3 m@xmath4 , respectively .
the peak of the magnetic field spectra shows a continuous shift towards smaller @xmath5-values , i.e. larger length scales , over time reaching values as low as @xmath5 = 4 . on small scales the magnetic energy decreases at the cost of the energy on large scales resulting in a well - ordered magnetic field with a strength up to @xmath6 g depending on the initial conditions .
the coherence length of the magnetic field inferred from the spectra reaches values up to 250 pc in agreement with those obtained from autocorrelation functions .
we find the coherence length to be as large as 50% of the radius of the supernova bubble .
extrapolating this relation to later stages we suggest that significantly strong magnetic fields with coherence lengths as large as 1.5 kpc could be created .
we discuss possible implications of our results on processes like recollapse of the halo , first galaxy formation , and the magnetization of the intergalactic medium .
[ firstpage ] mhd methods : numerical supernovae : general dark ages , reionization , first stars |
three dimensional ( 3d ) topological insulators ( tis ) @xcite are realizations of quantum matter that manifest protected surface states with dirac - like energy dispersion
. the novel features of ti electrodynamics are expressed in the axion " term in the lagrangian , @xmath7 , where @xmath8 has value the @xmath9 ( modulo @xmath10 ) in the case of a strong 3d ti @xcite , and @xmath11 is the fine structure constant . when @xmath8 is uniform in space the axion term has no observable effects because @xmath12 is a total space - time derivative and does not modify the equations of motion . however , observable consequences occur at the interface between topologically trivial and non - trivial media , where a 2d dirac metal is found .
if the dirac spectrum is gapped by a time - reversal symmetry ( trs ) breaking perturbation , half - integral hall conductivity , @xmath13 , is expected when the chemical potential lies in the gap @xcite . in practice the half - integral quantum hall effect ( or equivalently , the quantized magnetoelectric effect )
is difficult to observe because of parallel current paths that exist either through the bulk or side facets of the crystal that remain gapless despite the breaking of trs . as a result
, epitaxially - grown thin films are playing an increasing role in exhibiting the electromagnetic phenomena associated with 3d ti @xcite .
particularly interesting physics arises in the transition from 3d to 2d , as realized for example in films in which the thickness can be tuned with atomic precision . at finite thickness ,
the top and bottom surfaces are coupled , resulting in two degenerate copies of a massive dirac spectrum with an energy gap , @xmath14 , equal to twice the hopping matrix element , @xmath15 .
the development of this gap has been observed directly by _ in situ _ angle - resolved photo - emission spectroscopy ( arpes ) @xcite .
a striking prediction is that , in certain ranges of film thickness , breaking of trs by ferromagnetic ( fm ) order will give rise to a quantum anomalous hall effect ( qahe ) , that is , integer quantization in the absence of an applied magnetic field @xcite .
several recent reports have confirmed integer quantization with increasing accuracy @xcite , although under stringent conditions of temperature , chemical doping , and thickness .
for example , quantization in tis doped with the magnetic impurity cr is observed only for temperatures in the milli - kelvin range , despite the fact that spontaneous trs breaking appears at 15 - 20 k. to better understand the fragility of the qahe in magnetic - impurity doped epitaxial films , it is important to probe the strength of the fm order as a function of film thickness @xmath16 , temperature @xmath17 , and magnetic field @xmath18 . in this work
we measure the polarization rotation of light on reflection at normal incidence ( kerr effect ) , which vanishes in the presence of trs @xcite and is proportional to the out - of - plane component of the magnetization , @xmath19 , when trs is broken .
we report measurements of the real and imaginary parts of the complex kerr angle , @xmath1 , from near - ir through visible wavelengths ( photon energies in the range 0.8 ev@xmath20 3.0 ev ) .
although , in principle , detecting @xmath3 on the energy scale of gap induced by fm order , @xmath21 , might be preferable in terms of direct contact with theoretical predictions @xcite , in practice the present limits on sensitivity in this spectral range make this approach difficult .
the magnitude of the kerr angle predicted theoretically is @xmath22 mrad , where @xmath23 is the bulk ( or substrate in the case of an ultrathin film ) dielectric constant .
thus , for @xmath1 to be an effective probe of @xmath21 , particularly the manner in which it vanishes in the limit that @xmath24 , requires @xmath2510 - 100 @xmath26rad sensitivity in the frequency regime of 2 - 20 thz .
this can be compared with the current state of the art of terahertz spectroscopy , where sensitivity @xmath250.1 mrad in a frequency range up to @xmath25 2 thz has been demonstrated in measurements on bi@xmath27se@xmath28 films @xcite . in the the near - infrared through visible range of our spectrometer , we observe a peak in the real part of @xmath3 and a zero crossing in the imaginary part at approximately 1.7 ev that we attribute to optical transitions involving states near a spin - orbit avoided crossing in the bandstructure .
the enhanced kerr rotation on resonance allows measurement of @xmath1 in the ultrathin film limit , @xmath5 quintuple layers .
we find that the fm order decreases rapidly with decreasing thickness , which we attribute to the combined effects of disorder inherent in a doped magnetic semiconductor as well as competing anisotropic exchange interactions resulting from strong spin - orbit coupling .
finally , additional evidence for disordered magnetism is found in the significant difference in @xmath29 observed on warming in zero as compared with 100 mt applied magnetic field .
the samples used for this study are thin films of @xmath0 prepared using molecular - beam epitaxy on semi - insulating gaas ( 111 ) substrates .
these samples are not gated , so their fermi level is determined by film stoichiometry .
this is not perfectly controlled during growth , and can further change with exposure to air , yielding a range of carrier concentrations for the different samples from @xmath30 ( i.e. , hole doped ) to @xmath31 @xmath32 ( electron doped ) , as determined from hall measurements at temperatures above the magnetic transition @xcite . in a subset of the samples studied the anomalous hall resistance @xmath33 approaches the quantized value @xmath34k@xmath35 at dilution refrigerator temperatures @xcite , while other samples grown under nominally the same conditions display @xmath33 in the range @xmath36 to @xmath37 k@xmath35 . despite these variations in @xmath33 , we find that the trends in @xmath38 are reproducible .
this suggests that the kerr effect in the near - ir , in contrast to low temperature transport , does not originate principally from surface states . in this case reproducibility of @xmath1 requires only that the chemical potential remain in the bulk gap , which is a much less stringent condition than required for observation of the qahe .
the kerr spectra reported here were measured with a w - filament light source and grating monochromator , and single wavelength temperature and magnetic field scans were performed with a hene laser source at 633 nm ( 1.96 ev ) . in a polar kerr measurement ,
linearly polarized light is normally incident on a sample and the reflected light emerges in an elliptical polarization state whose major axis is rotated with respect to the initial polarization direction .
the real and imaginary parts of @xmath1 correspond to rotation and induced ellipticity , respectively . to measure the full complex @xmath1 we used a technique based on a photo - elastic optical phase modulator ( pem ) @xcite .
synchronous detection of the reflected light at the second and third harmonic of the pem frequency enabled precise discrimination between the real and imaginary parts of @xmath1 , respectively , as well as detection with 10 @xmath26rad sensitivity .
k for @xmath39 ql sample .
( a ) @xmath1 at at 6 k in @xmath40 mt ( r1 ) , and @xmath41 mt ( r2 ) , and at @xmath42 k in field @xmath41 mt ( r3 ) and @xmath40 mt ( r4 ) .
( b ) @xmath1 at @xmath42 k subtracted from @xmath1 at @xmath43 k in field @xmath40 mt ( d1 ) and @xmath41 mt ( d2).,width=316 ] by itself , this method does not distinguish between rotation induced by trs breaking and strain birefingence of the sample and cryostat window .
in addition , in the presence of an applied magnetic field there is a contribution to the measured rotation from the faraday effect of the window . to isolate the magneto - optic kerr effect of the sample we use the following procedure , which is illustrated in fig .
we first measure the spectrum of @xmath1 , in both positive and negative magnetic fields , at a temperature such as 50 k where the magnetic response is well below our detection limit .
we then measure @xmath1 with the sample cooled to the desired temperature , for example @xmath43 k , again for both positive and negative magnetic fields and calculate the difference spectra , @xmath44 and @xmath45 ( shown in fig .
[ fig1]b ) .
since the sample is cooled independently of its environment , this step eliminates contributions from faraday rotation and strain birefringence of the cryostat window ( which remains at room temperature ) .
finally , we calculate the component of the spectra in fig .
[ fig1]b that is odd in applied magnetic field to obtain the magneto - optic kerr spectra shown in succeeding figures of this paper .
[ fig2]a shows the real , @xmath46 , and imaginary , @xmath47 ( not to be confused with the dielectric constant @xmath23 ) , parts of @xmath1 in the spectral range from 1.1 - 2.9 ev for a 20 quintuple - layer ( ql ) sample measured at @xmath48 k using the procedure described above .
we find a resonance in the kerr spectrum , characterized by a peak in the real part and a zero crossing in the imaginary part . in the following
we describe the extraction of the off - diagonal component of the complex dielectric tensor , @xmath49 from the measured @xmath3 .
( a ) real @xmath46 and imaginary @xmath47 parts of the kerr spectrum measured at 6 k with magnetic field of 100 mt applied perpendicular to the film plane , ( b ) off - diagonal component of the dielectric function , @xmath49 , as determined by analysis of the kerr spectra shown in panel ( a).,width=336 ] because the optical penetration depth is less than 10 ql over most of the spectral range of our measurement , we model the 20 ql sample as a semi - infinite medium and assume that @xmath52 is uniform throughout the film . in this case , _
xy()=_k ( ) , where @xmath53 . in the analysis
we assume that @xmath54 and use the diagonal component of the dielectric function as reported in ref . .
[ fig2]b shows the real and imaginary parts of @xmath49 obtained by analysis of @xmath3 .
the appearance of a peak in @xmath55 $ ] and dispersive lineshape of @xmath56 $ ] are clear evidence for a resonance at approximately 1.6 ev .
note that the appearance of a peak in the imaginary part and dispersive lineshape in the real part of @xmath49 is reversed compared with the usual structure of @xmath57 in the neighborhood of a resonance .
this reversal is seen as well in spectra obtained from the magnetically doped 3d semiconductor mn : gaas @xcite and is consistent with what is expected for kerr spectra in dilute magnetic semiconductors . as a consequence of trs breaking the energy of transitions between valence and conduction bands
is slightly different for left and right circularly polarized light .
thus the difference spectrum , @xmath58 ( where @xmath59 are the imaginary parts of the dielectric function in the circular polarized basis ) will resemble the derivative of an absorption peak .
this quantity , the _ imaginary _ part of @xmath60 , is directly proportional to the _ real _ part of @xmath49 because @xmath61/2i$ ] on transforming back to the cartesian basis .
thus the derivative - like lineshape of @xmath62 that we observe is consistent with expectations for interband transitions in 3d dilute magnetic semiconductors . .
ss , vb , and cb denote surface state surface , valence band , and conduction band , respectively .
the shaded region in cb and vb indicate the energy range of two - dimensional subbands .
the ss gap is induced by fm order .
the green line illustrates the first ss to second ss optical transition , while the red line illustrates transitions between subband states that evolve to the `` bulk - to - bulk '' ( bb ) transition with increasing film thickness.,width=163 ] two - photon arpes measurements @xcite as well as bandstructure calculations @xcite , suggest that the resonance at 1.6 ev is associated with transitions from states near the fermi energy , @xmath63 , to states near a spin - orbit avoided band - crossing situated at the @xmath64 point at approximately @xmath65-@xmath66 ev above @xmath63 .
transitions resonant with spin - orbit avoided band crossings are known to make a large contribution to @xmath49 @xcite . a schematic bandstructure diagram for a film of thickness 20 nm , showing the avoided crossings at the @xmath64 point
is shown in fig .
the upper avoided crossing is a higher energy replica of the band inversion that takes place near the fermi energy , and is associated as well with a symmetry predicted dirac surface - state band .
the shaded regions in the schematic illustrate the range of energies associated with confinement in the direction normal to the plane of the film and the vertical arrow shows a typical transition from subband states of the first valence band ( vb1 ) to the second conduction band ( cb2 ) .
we use the sensitivity of the resonantly enhanced kerr effect to measure the kerr angle as a function of film thickness , @xmath16 , down to the ultrathin limit .
[ fig3]a shows the real and imaginary parts of @xmath1 at @xmath48 k as a function of @xmath16 , with fixed cr concentration , sampled at their peak photon energies of 1.7 and 1.5 ev , respectively .
[ fig4]b shows the imaginary part of @xmath3 for the same set of samples .
we observe that the frequency dependence of the kerr resonance remains unchanged even as its amplitude is significantly reduced with decreasing @xmath16 .
the insensitivity of the spectrum to thickness suggests that the width of the resonance is largely determined by the dispersion of energies with respect to the in - plane wavevector , rather than the energies associated with subband splittings .
this conclusion is consistent with bandstructure calculations that indicate subband splitting are no larger than 0.2 ev for films as thin as 4 nm @xcite , which can be compared with the characteristic width of the resonance in @xmath3 , which is approximately 0.8 ev . and @xmath47 at their respective peaks of 1.7 ev and 1.5 ev as a function of sample thickness at 6 k and 100 mt .
plotted for comparison ( dashed - dot line ) is @xmath67 calculated under the assumption of @xmath16-independent @xmath49 , ( b ) imaginary part of @xmath3 for various thicknesses in the range from 2 to 10 ql.,width=340 ] the dependence of the amplitude of the kerr angle on thickness is highly structured , with steep changes in amplitude taking place just below 8 and 4 ql and relatively small changes in between these thicknesses .
these features can be contrasted with the thickness dependence that would obtain if @xmath49 was independent of sample thickness . to make this comparison , we fitted @xmath49 as measured from the 20 ql film using a lorentzian function , and then input this fit into the formula for the reflection tensor of a thin film on a dielectric substrate @xcite .
the result of this simulation is shown as a dashed - dot line in fig .
not surprisingly , the assumption of thickness independent @xmath49 predicts that the kerr rotation decreases smoothly to zero once @xmath16 becomes smaller than the optical penetration depth .
thus the sharp features in the observed thickness dependence can be attributed to structure in @xmath49 vs. @xmath16 , and ultimately to structure in @xmath68 . at photon energy 1.96 ev as a function of temperature for a 20 ql film .
the larger amplitude curve was measured upon cooling in a 100 mt magnetic field and warming with the field left on , the smaller amplitude curve was recorded on warming with field turned off at 6 k , and is a measure of the remnant magnetization , ( b ) remanent @xmath1 at 6 k as a function of film thickness shown as solid circles(left axis ) ; triangles ( right axis ) show the temperature t@xmath69 above which the remanent signal vanishes .
dashed line indicates our experimental upper bound on @xmath70.,width=316 ] further evidence for structure in @xmath68 emerges from the @xmath17 dependence of the kerr rotation .
[ fig5]a shows @xmath71 of a 20 ql film at photon energy 1.96 ev measured with two different protocols for application of the @xmath18 field . in both measurements the sample
was cooled to 6 k with @xmath72 mt applied in the direction normal to the film surface .
the data shown in red were recorded after switching the field to zero at 6k and subsequently warming and are therefore proportional to the remanant magnetization .
we define the temperature at which @xmath71 warmed in zero field extrapolates to zero as the remnant temperature , @xmath70 , which in this case is @xmath25 20k .
the data shown in blue were recorded after again cooling in @xmath73 mt , but with the field left on during the warming cycle .
the different results for the two measurement procedures are quite striking and not characteristic of a conventional ferromagnet .
not only is @xmath71 in field much larger , but it persists to much higher temperature , @xmath25 50k .
thus there is a broad temperature regime , from @xmath74 k , in which the magnetization in zero - field is absent but a magnetization comparable to the remanent value is induced by magnetic fields in the 100 mt range . in fig .
[ fig5]b we illustrate the thickness dependence of the remanent magnetization by plotting both @xmath70 and the remanent value of @xmath75 _
vs_. @xmath16 .
the structured dependence on @xmath16 is even more clear in the remanent properties than those measured in 100 mt magnetic field .
for example , @xmath75 suddenly becomes too small to measure for sample thicknesses less than 8 ql .
we note that samples such as these may exhibit remnant magnetization when cooled below 6 k , for example into the milikelvin range where the qahe has been observed .
below we comment on possible origins for such structure in @xmath68 and @xmath76 .
evidence for a crossover to a regime of strong coupling between top and bottom surface states with decreasing film thickness was reported in a time - domain terahertz study @xcite of the system ( bi@xmath77in@xmath78se@xmath28 .
analysis of the real and imaginary parts of the drude conductivity indicated a crossover thickness , @xmath79 , below which the carrier scattering rate , @xmath80 increased rapidly .
the crossover was associated with entering a regime where @xmath81 , which is the condition for loss of topological protection against backscattering . for @xmath82 ,
@xmath83 is @xmath25 6 ql , which is comparable to , but somewhat smaller than , the thicknesses at which the remanent @xmath1 changes abruptly in our somewhat similar but magnetically - doped films . that the crossover to strong coupling between top and bottom surfaces , as heralded by @xmath1 dropping to zero , occurs at a larger thickness in the magnetic film is consistent with the expected reduction spin - orbit coupling strength that results from cr - doping @xcite .
the smaller spin - orbit coupling reduces the inverted bulk band gap , consequently increasing the penetration depth into the bulk of dirac surface states . in the earlier study , below @xmath83
the terahertz conductance decreases smoothly but extrapolates to zero at @xmath84 ql .
the absence of conductance at 2 ql was associated with either the theoretically expected transition to a topologically trivial state , or strong disorder in the ultrathin limit . as shown in fig . 2
, @xmath3 shows an analogous transition to zero response at this same thickness .
the pronounced structure that we observe in @xmath85 likely reflects a strong dependence of spin interactions on film thickness .
this dependence , in turn , is a consequence of the complicated spin interactions that exist even in the thick film regime .
theoretical studies @xcite have found that the effective interaction between impurity spins , as mediated by ti surface state electrons , is highly anisotropic because of the spin - momentum locking of the dbs . the interaction is characterized by an ising term , @xmath86 , that favors fm alignment in the direction normal to the film plane as is experimentally observed in our thicker films , plus contributions that depend on the relative position , @xmath87 , of the spins .
one such contribution favors fm alignment in the direction parallel to @xmath87 and the other favors anti - fm ( afm ) alignment in the direction perpendicular to @xmath87 .
if the position - dependent terms are sufficiently weak , the ground state has fm order along @xmath88 .
however , if the position - dependent interactions are strong enough the ground state will be a spin glass @xcite .
two recent theoretical studies have considered the dependence of these competing interactions on film thickness @xcite . as pointed out in ref .
, the ising fm interaction arises mainly from excitations across the gap , and therefore decreases monotonically with @xmath16 due to increased @xmath89 . on the other hand , the afm interactions , which derive from superexchange ,
depend only weakly on the gap , and consequently on thickness .
therefore , as films become sufficiently thin , the afm interaction will eventually dominate , in which case the random distribution of magnetic impurities @xcite will frustrate an afm ground state and lead to paramagnetic or spin glass behavior .
a similar conclusion was reached in ref . ,
in which the dependence of the effective interaction on the distance of the impurities from the surface was included . in that work
it was reported that , while the ising interaction mediated by a single surface state is fm , in agreement with previous work @xcite , the interaction between spins situated near opposite surfaces is afm . for a film that is sufficiently thin , interactions that favor alignment along @xmath88
can be frustrated by positional disorder and the net weakening of the ising contribution to the exchange energy can again lead to spin glass behavior .
our observation that as opposite surfaces approach each other the remanent magnetization vanishes , while the film retains a large susceptibility to external fields , is consistent with both the above theoretical pictures . in summary
, we have measured the complex magneto - optic kerr effect in @xmath0 in visible and near - infrared wavelengths as a function of temperature , photon energy and magnetic field .
we observed a resonance in the kerr effect at 1.7 ev , corresponding to the energy of a spin - orbit avoided band crossing with respect to the fermi energy .
the resonant enhancement allowed us to explore the effective out - of - plane component of the magnetization in the ultrathin film limit .
at our base temperature of 6 k , we observed a transition to zero remanent magnetization below a thickness of 8 ql , and to essentially zero magnetism at 2 ql . the dependence of @xmath19 on @xmath16 is consistent with theories of thickness dependence of magnetic impurity spin interaction in tis .
research primarily supported by the u.s .
department of energy , office of science , basic energy sciences , under contract no .
de - ac02 - 76sf00515 .
j. g. acknowledges a scholarship from the materials science and active surface program at ecole polytechnique , palaiseau , france ( chaire x - espci - saint - gobain ) for support .
k. l. w. acknowledges the support of the raytheon endorsement .
a. j. b. acknowledges support from a benchmark stanford graduate fellowship .
e. j. f. acknowledges support from a doe office of science graduate fellowship .
materials growth , surface characterization , preliminary electrical characterization , and electronic instrumentation were supported by the darpa meso program under contracts no .
n66001 - 12 - 1 - 4034 and no .
n66001 - 11 - 1 - 4105 .
infrastructure and cryostat were funded in part by the gordon and betty moore foundation through grant gbmf3429 to d. g .-
g . 36ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo
secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ] + 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevlett.114.187201 [ * * , ( ) ] link:\doibase 10.1103/physrevb.45.5504 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevb.84.205327 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.105.057401 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevlett.108.087403 [ * * , ( ) ] @noop @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevb.83.125206 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevb.81.041307 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1126/science.1230905 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevlett.106.136802 [ * * , ( ) ] link:\doibase 10.1103/physrevb.89.235316 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.115.036805 [ * * , ( ) ] @noop * * , ( ) | we report measurements of the polar kerr effect , proportional to the out - of - plane component of the magnetization , in thin films of the magnetically doped topological insulator @xmath0 .
measurements of the complex kerr angle , @xmath1 , were performed as a function of photon energy in the range @xmath2 .
we observed a peak in the real part of @xmath3 and zero crossing in the imaginary part that we attribute to resonant interaction with a spin - orbit avoided crossing located @xmath4 1.6 ev above the fermi energy .
the resonant enhancement allows measurement of the temperature and magnetic field dependence of @xmath1 in the ultrathin film limit , @xmath5 quintuple layers .
we find a sharp transition to zero remanent magnetization at 6 k for @xmath6 ql , consistent with theories of the dependence of impurity spin interactions on film thickness and their location relative to topological insulator surfaces . |
let @xmath10=\{1,\ldots , n\}$ ] be a set of size @xmath11 .
we denote the power set of @xmath12 by @xmath13 .
a family @xmath14 is a collection of sets in @xmath15 .
@xmath14 is said to be an intersecting family if @xmath16 imply @xmath17 .
an intersecting @xmath3-uniform hypergraph is an intersecting family where all sets have cardinality @xmath3 .
the problem of finding how large an intersecting family can be is trivial : an intersecting family can have size at most @xmath18 with @xmath19 being one of the extremal families .
if we consider this problem for intersecting @xmath3-uniform hypergraphs , we see that the problem is trivial for @xmath20 because the set of all @xmath3-sets in @xmath12 , denoted by @xmath21 , is intersecting for @xmath22 , and if @xmath23 , every family contains exactly one of any two complimentary sets , so the maximum size is at most @xmath24 if @xmath25 , then the problem is solved by the erds - ko - rado theorem @xcite , one of the seminal results in extremal set theory .
( erds - ko - rado theorem @xcite)[ekrt ] let @xmath26 and let @xmath27 be an intersecting hypergraph . then @xmath28 with equality iff @xmath29 for some @xmath30 .
there have been generalizations of the theorem in different directions .
deza and frankl @xcite give a very nice survey of the ekr - type results proved in the 1960s , 70 s and 80 s .
in this paper , we concern ourselves with the generalization for graphs , formulated by holroyd , spencer and talbot in @xcite .
the erds - ko - rado property for graphs is defined in the following manner . for a graph @xmath0 ,
vertex @xmath31 and some integer @xmath2 , denote the family of independent @xmath3-sets of @xmath4 by @xmath5 and the subfamily @xmath6 by @xmath7 , called a star .
then , @xmath0 is said to be @xmath3-ekr if no intersecting subfamily of @xmath5 is larger than the largest star in @xmath5 .
if every maximum sized intersecting subfamily of @xmath5 is a star , then @xmath0 is said to be strictly @xmath3-ekr .
this can be viewed as the erds - ko - rado property on a ground set , but with additional structure on this ground set .
in fact , the erds - ko - rado theorem can be restated in these terms as follows .
( erds - ko - rado theorem @xcite ) the graph on @xmath11 vertices with no edges is @xmath3-ekr if @xmath32 and strictly @xmath3-ekr if @xmath33 there are some results giving ekr - type theorems for different types of graphs .
the following theorem was originally proved by berge @xcite , with livingston @xcite characterizing the extremal case .
( berge @xcite , livingston @xcite ) [ berth ] if @xmath2 , @xmath34 and @xmath0 is the disjoint union of @xmath3 copies of @xmath35 , then @xmath0 is @xmath3-ekr and strictly so unless @xmath36
. other proofs of this result were given by gronau @xcite and moon @xcite .
berge @xcite proved a stronger result .
( berge @xcite)[berth2 ] if @xmath0 is the disjoint union of @xmath3 complete graphs each of order at least @xmath37 , then @xmath0 is @xmath3-ekr .
a generalization of theorem [ berth ] was first stated by meyer @xcite and proved by deza and frankl @xcite .
( meyer @xcite , deza and frankl @xcite ) [ dfth ] if @xmath2 , @xmath34 and @xmath0 is a disjoint union of @xmath38 copies of @xmath35 , then @xmath0 is @xmath3-ekr and strictly so unless @xmath36 and @xmath39 . in the paper which introduced the notion of the @xmath3-ekr property for graphs , holroyd , spencer and talbot @xcite
prove a generalization of theorems [ berth2 ] and [ dfth ] .
( holroyd et al .
@xcite ) if @xmath0 is a disjoint union of @xmath38 complete graphs each of order at least @xmath37 , then @xmath0 is @xmath3-ekr . the compression technique used in @xcite , which is equivalent to contracting an edge in a graph ,
was employed by talbot@xcite to prove a theorem for the @xmath40 power of a cycle .
the @xmath40 power of a cycle @xmath41 is a graph with vertex set @xmath42 $ ] and edges between @xmath43 $ ] iff @xmath44 .
( talbot @xcite ) if @xmath45 , then @xmath41 is @xmath3-ekr and strictly so unless @xmath46 and @xmath47 . an analogous theorem for the @xmath40 power of a path is also proved in @xcite .
the @xmath40 power of a path @xmath48 is a graph with vertex set @xmath42 $ ] and edges between @xmath43 $ ] iff @xmath49 .
( holroyd et al .
@xcite ) if @xmath45 , then @xmath48 is @xmath3-ekr . it can be observed here that the condition @xmath50 is not required for the graphs @xmath41 and @xmath48 because for each of the two graphs , there is no independent set of size greater than @xmath51 , so the @xmath3-ekr property holds vacuously if @xmath52 . the compression proof technique is also employed to prove a result for a larger class of graphs .
( holroyd et al .
@xcite)[mix ] if @xmath0 is a disjoint union of @xmath32 complete graphs , cycles and paths , including an isolated singleton , then @xmath0 is @xmath3-ekr . the problem of finding if a graph @xmath0 is @xmath37-ekr is addressed by holroyd and talbot in @xcite .
( holroyd and talbot @xcite ) [ 2ekr ] let @xmath0 be a non - complete graph of order @xmath11 with minimum degree @xmath53 and independence number @xmath54 . 1 . if @xmath55 , then @xmath0 is strictly @xmath37-ekr .
if @xmath56 , then @xmath0 is @xmath37-ekr if and only if @xmath57 and strictly so if and only if @xmath58 , the star centers being the vertices of minimum degree .
holroyd and talbot also present an interesting conjecture in @xcite .
the minimum size of a maximal independent vertex set of a graph @xmath0 is the minimax independent number , denoted by @xmath9 . it can be noted here that @xmath59 , where @xmath60 is the independent domination number .
let @xmath0 be any graph and let @xmath61 ; then @xmath0 is @xmath3-ekr(and is strictly so if @xmath62 ) .
this conjecture seems hard to prove or disprove ; however , restricting attention to certain classes of graphs makes the problem easier to tackle .
borg and holroyd @xcite prove the conjecture for a large class of graphs , which contain a singleton as a component .
( borg , holroyd @xcite)[mnd ] for a monotonic non - decreasing ( _ _ ) sequence @xmath63 of non - negative integers , let @xmath64 be the graph such that @xmath65 and for @xmath66 with @xmath67 , @xmath68 iff @xmath69 .
let @xmath70 be the subgraph of @xmath71 induced by the subset @xmath72\}$ ] of @xmath73 .
call @xmath74 an _ _ graph .
( borg , holroyd @xcite ) for @xmath75 , @xmath76 , @xmath77 , let @xmath78 be the graph with vertex set @xmath79\}$ ] and edge set @xmath80 . if @xmath81 , call @xmath78 a modified @xmath40 power of a cycle .
borg and holroyd @xcite prove the following theorem .
[ bhth ] conjecture [ minmax ] is true if @xmath0 is a disjoint union of complete multipartite graphs , copies of _ _ graphs , powers of cycles , modified powers of cycles , trees , and at least one singleton .
one of our main results in this paper extends the class of graphs which satisfy conjecture [ minmax ] by proving the conjecture for all chordal graphs which contain a singleton .
it can be noted that the graphs in theorem [ bhth ] are chordal .
we also define a special class of chordal graphs , and prove a stronger ekr result for these graphs .
finally , we consider similar problems for two classes of bipartite graphs , trees and ladder graphs .
a graph @xmath0 is a chordal graph if every cycle of length at least @xmath82 has a chord .
it is easy to observe that if @xmath0 is chordal , then every induced subgraph of @xmath0 is also chordal .
a vertex @xmath83 is called _ simplicial _ in a graph @xmath0 if its neighborhood is a clique in @xmath0 .
consider a graph @xmath0 on @xmath11 vertices , and let @xmath84 $ ] be an ordering of the vertices of @xmath0 .
let the graph @xmath85 be the subgraph obtained by removing the vertex set @xmath86 from @xmath0 .
then @xmath87 is called a _ simplicial elimination ordering _ if @xmath88 is simplicial in the graph @xmath85 , for each @xmath89 .
we state a well known characterization for chordal graphs , due to dirac @xcite .
[ dirc ] a graph @xmath0 is a chordal graph if and only if it has a simplicial elimination ordering .
it is easy to see , using this characterization of chordal graphs , that the graphs of definition [ mnd ] are chordal .
if @xmath74 is an _ _ graph on @xmath11 vertices , @xmath74 is chordal . it can be seen that ordering the vertices of @xmath74 , according to the corresponding degree sequence @xmath90 , as stated in definition [ mnd ] , gives a simplicial elimination ordering .
note that , with or without the non - decreasing condition on the sequence @xmath90 , the resulting graph is an interval graph use the interval @xmath91 $ ] for vertex @xmath92 which is chordal regardless .
we prove the non - strict part of conjecture [ minmax ] for disjoint unions of chordal graphs , containing at least one singleton .
[ mainthm ] if @xmath0 is a disjoint union of chordal graphs , including at least one singleton , and if @xmath93 , then @xmath0 is @xmath3-ekr .
we also consider graphs which do not have singletons .
consider a class of chordal graphs constructed as follows .
let @xmath94 be a path on @xmath11 edges with @xmath95 .
label the edge @xmath96 as @xmath97 , for each @xmath89 .
chain _ of complete graphs , of length @xmath11 , is obtained from @xmath94 by replacing each edge of @xmath94 by a complete graph of order at least @xmath37 in the following manner : to convert edge @xmath97 of @xmath94 into @xmath98 , introduce a complete graph @xmath99 and connect @xmath88 and @xmath100 to each of the @xmath101 vertices of the complete graph .
call the resulting complete graph @xmath85 , and call each @xmath85 a link of the chain .
we call @xmath88 and @xmath100 the _ connecting vertices _ of this complete graph , with the exception of @xmath102 and @xmath103 , which have only one connecting vertex each ( the ones shared with @xmath104 and @xmath105 respectively ) . in general , for each @xmath106 , call @xmath88 the @xmath107 connecting vertex of @xmath0 . unless otherwise specified , we will refer to a chain of complete graphs as just a chain .
we will call an isolated vertex a _ trivial chain _ ( of length @xmath108 ) , while a complete graph is simply a chain of length @xmath109 .
call a chain of length @xmath11 _ special _ if @xmath110 or if @xmath111 and the following conditions hold : 1 .
@xmath112 for each @xmath113 ; 2 .
@xmath114 .
we prove the following results for special chains .
[ mthm ] if @xmath0 is a _ special _ chain , then @xmath0 is @xmath3-ekr for all @xmath2 .
[ mthm2 ] if @xmath0 is a disjoint union of @xmath37 _ special _ chains , then @xmath0 is @xmath3-ekr for all @xmath2 .
we will also consider similar problems for bipartite graphs . a basic observation about complete bipartite graphs , and its obvious generalization for complete multipartite graphs , are mentioned below .
* if @xmath115 and @xmath116 , then @xmath0 is @xmath3-ekr for all @xmath117 . * if @xmath118 , with @xmath119 , then @xmath0 is @xmath3-ekr for all @xmath120 .
it is easy to see why these hold .
if @xmath121 is intersecting , then each @xmath122 lies in the same partite set . clearly , if @xmath123 , then @xmath0 is @xmath3-ekr by theorem [ ekrt ] .
a similar argument works for complete multipartite graphs as well .
holroyd and talbot @xcite proved conjecture [ minmax ] for a disjoint union of two complete multipartite graphs .
if we consider non - complete bipartite graphs with high minimum degree , it seems that they usually have low @xmath124 ( always at most @xmath125 ) .
instead , in this paper , we consider bipartite graphs with low maximum degree in order to have higher values of @xmath124 ( always at least @xmath126 ) .
in particular , we look at trees and ladder graphs , two such classes of sparse bipartite graphs .
one of the difficult problems in dealing with graphs without singletons is that of finding centers of maximum stars .
we consider this problem for trees , and conjecture that there is a maximum star in a tree that is centered at a leaf .
[ stree ] for any tree @xmath127 on @xmath11 vertices , there exists a leaf @xmath128 such that for any @xmath129 , @xmath130 .
we prove this conjecture for @xmath131 .
[ tstar ] let @xmath132 .
then , a maximum sized star of @xmath3-independent vertex sets of @xmath127 is centered at a leaf .
we will also prove that the ladder graph is @xmath133-ekr .
the ladder graph @xmath134 with @xmath11 rungs can be defined as the cartesian product of @xmath135 and @xmath136 .
it is not hard to see that , for @xmath134 , @xmath137 .
in fact , we show that equality holds .
@xmath138 the result is trivial if @xmath139 , so let @xmath140 .
suppose @xmath141 and let @xmath142 be a maximal independent set of size @xmath143 .
then , there exist two consecutive rungs , say the @xmath144 and @xmath145 in @xmath134 , with endpoints @xmath146 and @xmath147 respectively , such that @xmath148 and @xmath149 .
let @xmath150 , @xmath151 and @xmath152 if @xmath153 , otherwise , let @xmath154 , @xmath155 and @xmath156 .
@xmath157 is not independent , since @xmath142 is maximal .
then , @xmath158 and @xmath159 is independent , a contradiction .
[ ladder ] the graph @xmath134 is @xmath133-ekr for all @xmath160 .
the rest of the paper is organized as follows : in section [ s3 ] , we give a proof of theorem [ mainthm ] , in section [ s4 ] , we give proofs of theorems [ mthm ] and [ mthm2 ] , and in section [ s5 ] , we give proofs of theorems [ ladder ] and [ tstar ] .
we begin by fixing some notation . for a graph @xmath0 and a vertex @xmath31 ,
let @xmath161 be the graph obtained from @xmath0 by removing vertex @xmath83 .
also , let @xmath162 denote the graph obtained by removing @xmath83 and its set of neighbors from @xmath0 .
we note that if @xmath0 is a disjoint union of chordal graphs and if @xmath1 , the graphs @xmath161 and @xmath162 are also disjoint unions of chordal graphs .
we state and prove a series of lemmas , which we will use in the proof of theorem [ mainthm ] . [ l1 ] let @xmath0 be a graph containing an isolated vertex @xmath128 . then , for any vertex @xmath31 , @xmath163 .
let @xmath31 , @xmath164 .
we define a function @xmath165 as follows .
@xmath166 it is easy to see that the function is injective , and this completes the proof .
[ l2 ] let @xmath0 be a graph , and let @xmath167 be vertices such that @xmath168\subseteq n[v_2]$ ] .
then , the following inequalities hold : 1 . @xmath169 ; 2 .
@xmath170 .
we begin by noting that the condition @xmath168\subseteq n[v_2]$ ] implies that @xmath171 . 1
. we will show that if @xmath172 is a maximal independent set in @xmath173 , then @xmath172 is maximally independent in @xmath0
. suppose @xmath172 is not a maximal independent set in @xmath0 .
then , @xmath174 is an independent set in @xmath0 .
thus , for any @xmath175 $ ] , @xmath176 .
in particular , for any @xmath177 $ ] , @xmath176 .
thus , @xmath178 is an independent set in @xmath173 .
this is a contradiction .
thus , @xmath172 is a maximal independent set in @xmath0 . + taking @xmath172 to be the smallest maximal independent set in @xmath173 , we get @xmath179 .
2 . we will show that if @xmath172 is a maximal independent set in @xmath180 , then @xmath174 is a maximal independent set in @xmath0 .
of course , @xmath174 is independent , so suppose it is not maximal .
then , for some vertex @xmath181 and @xmath182 , @xmath183 is an independent set .
thus , @xmath184 is an independent set in @xmath180 , a contradiction .
+ taking @xmath172 to be the smallest maximal independent set in @xmath180 , we get @xmath185 . [ cor1 ] let @xmath0 be a graph , and let @xmath167 be vertices such that @xmath168\subseteq n[v_2]$ ] . then
, the following statements hold : 1 .
if @xmath93 , then @xmath186 ; 2 . if @xmath93 , then @xmath187 . 1 . this follows trivially from the first part of lemma [ l2 ] .
2 . to prove this part , we use the second part of lemma [ l2 ] to show @xmath188 let @xmath189 be a component of @xmath0 , so @xmath189 is a chordal graph on @xmath190 vertices , @xmath191 . let @xmath192 be a simplicial elimination ordering of @xmath189 and let @xmath193 for some @xmath194 . let @xmath195 be an intersecting family .
we define a compression operation @xmath196 for the family @xmath14 .
before we give the definition , we note that if @xmath142 is an independent set and if @xmath197 , then @xmath198 is also independent .
@xmath199 then , we define the family @xmath200 by @xmath201 it is not hard to see that @xmath202 . next , we define the families @xmath203 @xmath204
@xmath205 then we have @xmath206 we prove the following lemma about these families .
[ l3 ] 1 .
2 . @xmath208 .
@xmath209 is intersecting .
@xmath210 is intersecting .
it follows from the definitions of the families that @xmath211 and @xmath208 .
so , we only prove that the two families are intersecting .
consider @xmath212 .
if @xmath213 and @xmath214 , we are done .
if @xmath215 and @xmath216 , then @xmath217 and hence @xmath218 .
so , suppose @xmath215 and @xmath214 . then , @xmath219 . also , either @xmath220 , in which case we are done or @xmath221 .
then , @xmath222 .
finally , consider @xmath223 . since @xmath224 , @xmath225 and @xmath226 .
a similar argument works for @xmath227 .
thus , @xmath228 and hence , @xmath229 .
the final lemma we prove is regarding the star family @xmath230 , where @xmath128 is an isolated vertex .
[ l4 ] let @xmath0 be a graph containing an isolated vertex @xmath128 and let @xmath31 , @xmath164 .
then , we have @xmath231 partition the family @xmath230 into two parts .
let the first part contain all sets containing @xmath83 , say @xmath232 , and let the second part contain all sets which do not contain @xmath83 , say @xmath233 . then @xmath234 and @xmath235 .
we proceed to a proof of theorem [ mainthm ] .
the theorem trivially holds for @xmath236 , so suppose @xmath237 .
let @xmath0 be a disjoint union of chordal graphs , including at least one singleton , and let @xmath238 .
we do induction on @xmath239 .
if @xmath240 , then @xmath241 , and we are done by the erds - ko - rado theorem .
so , suppose @xmath242 , and there is one component , say @xmath189 , which is a chordal graph having @xmath190 vertices , @xmath191 .
let @xmath192 be a simplicial ordering of @xmath189 and suppose @xmath193 for some @xmath194 .
since the neighborhood of @xmath243 is a clique , we have @xmath168\subseteq n[v_i]$ ] .
also , let @xmath128 be an isolated vertex in @xmath0 .
let @xmath195 be intersecting . define the compression operation @xmath196 and the families @xmath209 and @xmath210 as before .
using equation [ eq1 ] , lemmas [ l1 ] , [ l2 ] , [ l3 ] , [ l4 ] , corollary [ cor1 ] and the induction hypothesis , we have @xmath244
the main technique we use to prove theorem [ mthm ] is a compression operation that is equivalent to compressing a clique to a single vertex . in a sense
, it is a more general version of the technique used in @xcite .
we begin by stating and proving a technical lemma , similar to the one proved in @xcite .
we will then use it to prove theorem [ mthm ] by induction .
let @xmath245 with @xmath246 .
let @xmath247 be the graph obtained by contracting the subgraph @xmath189 to a single vertex .
the contraction function @xmath248 is defined as follows .
@xmath249 when we contract @xmath189 to @xmath243 , the edges which have both endpoints in @xmath189 are lost and if there is an edge @xmath250 such that @xmath251 , then there is an edge @xmath252 .
duplicate edges are disregarded .
also , let @xmath253 be the ( possibly disconnected ) graph obtained from @xmath0 by removing all vertices in @xmath189 .
[ tl1 ] let @xmath254 be a graph and let @xmath255 be an intersecting family of maximum size . if @xmath189 is a subgraph of @xmath0 with vertex set @xmath256 , and if @xmath189 is isomorphic to @xmath98 , then there exist families @xmath257 , @xmath258 , @xmath259 , @xmath260 satisfying : 1 .
@xmath261 ; 2 .
@xmath262 is intersecting ; and 3 .
for each @xmath263 , 1 .
@xmath264 is intersecting , 2 .
@xmath265 , and 3 .
@xmath266 . to prove lemma [ tl1 ]
, we will need a claim , which we state and prove below .
[ cl ] let @xmath245 be isomorphic to @xmath98 , @xmath267 let @xmath195 be an intersecting family of maximum size .
suppose @xmath268 for some @xmath269 and @xmath270 then @xmath271 since we have @xmath272 , @xmath273 suppose @xmath274 . since @xmath14 is an intersecting family of maximum size , @xmath275 is not an intersecting family .
so , there exists a @xmath276 such that @xmath277 so , we have @xmath278 and @xmath279 . thus , @xmath280 this is a contradiction since @xmath88 and @xmath281 are adjacent to each other .
( proof of lemma [ tl1 ] ) define the following families : 1 .
@xmath282 ; and 2 . for each @xmath263
@xmath283 , 2 .
@xmath265 , and 3 .
@xmath266 . if @xmath217 and @xmath284 , then @xmath285 iff @xmath286 for some @xmath287 using this and claim [ cl ] ( if @xmath288 ) , we have @xmath289 also , if @xmath219 , then @xmath290 iff @xmath291 thus , we have @xmath292 . by the definition of the @xmath293 s ,
@xmath294 is a disjoint union , so we have @xmath295 it is obvious to show that @xmath257 is intersecting since @xmath14 is .
let @xmath296 to see that @xmath297 is intersecting , suppose @xmath298 and @xmath299 .
but @xmath300 and @xmath301 are in @xmath14 and hence , are intersecting .
this is a contradiction .
before we move to the proof of theorem [ mthm ] , we will prove one final claim regarding maximum sized star families in @xmath0 .
[ clmstr ] if @xmath0 is _ special _ chain of length @xmath11 , then a maximum sized star is centered at an internal vertex of @xmath102 . first note that for any @xmath97 , there is a trivial injection from a star centered at a connecting vertex of @xmath85 to a star centered at an internal vertex of @xmath85 , which replaces the star center by that internal vertex in every set of the family .
so suppose @xmath302 is a star centered at a internal vertex @xmath303 of any of the graphs @xmath85 , @xmath304 .
let @xmath305 . consider the following cases . 1 .
suppose @xmath303 is in @xmath104 . in this case
, define an arbitrary bijection between the @xmath306 internal vertices of @xmath102 and any @xmath306 internal vertices of @xmath104 containing @xmath303 , such that @xmath303 corresponds to an internal vertex of @xmath102 , say @xmath83 ( note that this can always be done , since if @xmath307 , then @xmath308 , with one connecting vertex , while if @xmath140 , then @xmath309 , with two connecting vertices ) .
2 . suppose @xmath303 is in some @xmath85 such that @xmath310 . then , define an arbitrary bijection between the @xmath190 vertices of @xmath102 and any @xmath190 internal vertices of @xmath85 including @xmath303 such that @xmath303 corresponds to an internal vertex of @xmath102 , say @xmath83 .
next , consider any set in @xmath302 .
if it contains a vertex @xmath311 in @xmath102 , replace that vertex by @xmath312 and replace @xmath303 by the vertex in @xmath85 corresponding to @xmath311 .
if it does not contain a vertex in @xmath102 , replace @xmath303 by @xmath83 .
this defines the injection from @xmath302 to a star centered at @xmath83 .
we now give a proof of theorem [ mthm ] .
let @xmath313 be a maximum sized star family in @xmath0 , where @xmath109 is an internal vertex of @xmath102 .
we do induction on @xmath3 .
the result is trivial for @xmath236 .
let @xmath314 .
we do induction on @xmath11 ( @xmath11 is the number of links ) . for @xmath315 ,
result is vacuously true .
if @xmath307 , then for @xmath316 , we use theorem [ 2ekr ] to conclude that @xmath0 is @xmath37-ekr while the result is vacuously true for @xmath317 .
so , let @xmath140 .
let @xmath195 be an intersecting family of maximum cardinality .
let the vertices of @xmath318 be labeled from @xmath319 to @xmath320 ( let @xmath319 be the connecting vertex which also belongs to @xmath105 ) .
define the compression operation @xmath248 on @xmath0 and the clique @xmath98 as before .
let the families @xmath257 , @xmath258 , @xmath259 , @xmath260 be defined as in lemma [ tl1 ] . clearly , for @xmath0 , @xmath321 for each @xmath322 so , by lemma [ tl1 ] , @xmath323 let @xmath324 let the vertices of @xmath105 be labeled from @xmath325 to @xmath326(@xmath327 ) , with @xmath328 . for every @xmath329 and @xmath330 define a set @xmath331 of families by @xmath332 we note that @xmath333 for each @xmath330 , and since each of the @xmath331 s are also disjoint , we have @xmath334 now , consider a complete bipartite graph @xmath335 .
label the vertices in part @xmath109 from @xmath325 to @xmath336 and vertices in part @xmath37 from @xmath337 to @xmath338 .
partition the edges of the bipartite graph @xmath335 into @xmath339 matchings , each of size @xmath340 . for each matching @xmath341 ( @xmath342 ) ,
define the family @xmath343 where a family @xmath344 is obtained from @xmath345 by removing @xmath346 from all its sets .
then of course latexmath:[\[\sum_{1\leq i\leq
t-1,2\leq j\leq s } for each @xmath342 , @xmath348 is a disjoint union and is intersecting .
the intersecting property is obvious if both sets are in the same @xmath349 since they contain @xmath350 .
if in different such sets , adding distinct elements which were removed ( during the above operation ) gives sets in the original family which are intersecting . finally , if we consider families @xmath351 for @xmath263 , each such family is a disjoint union .
it is also intersecting since for @xmath352 and @xmath353 , @xmath354 and @xmath355 for some @xmath356 gives us sets in @xmath14 .
so , we get @xmath357 the last inequality is obtained by partitioning the star based on whether or not it contains one of @xmath358 we do induction on @xmath3 .
since the case @xmath236 is trivial , let @xmath237 .
let @xmath0 be a disjoint union of @xmath37 _ special _
chains @xmath359 and @xmath360 , with lengths @xmath319 and @xmath337 respectively .
we will do induction on @xmath361 . if @xmath362 , the result holds trivially if @xmath316 and vacuously if @xmath317 .
so , let @xmath160 . if @xmath315 or if @xmath363 , then @xmath364 . in this case
, @xmath0 is vacuously @xmath3-ekr for @xmath317 .
also , if @xmath316 , then we are done by theorem [ 2ekr ] .
so , without loss of generality , we assume that @xmath102 has length at least @xmath37 .
we can now proceed as in the proof of theorem [ mthm ] .
the statement is trivial for @xmath236 .
if @xmath316 , we use the fact that for any vertex @xmath83 , @xmath365 where @xmath366 is the degree of vertex @xmath83 , and thus it will be maximum when @xmath83 is a leaf .
let @xmath367 let @xmath83 be an internal vertex(@xmath368 and let @xmath369 be the star centered at @xmath83 . consider @xmath127 as a tree rooted at @xmath83 .
we find an injection @xmath370 from @xmath371 to a star centered at some leaf .
let @xmath243 and @xmath372 be any two neighbors of @xmath83 and let @xmath303 be a leaf with neighbor @xmath311 .
let @xmath373 . 1 . if @xmath374 , then let @xmath375 .
2 . if @xmath376 , then we consider two cases . 1 . if @xmath377 , let @xmath378 .
2 . if @xmath379 , then @xmath380 . we consider the following two cases separately .
* @xmath381 + let @xmath382 .
we know that @xmath128 can not be connected to both @xmath243 and @xmath372 since that would result in a cycle . without loss of generality ,
suppose that @xmath383 .
then , let @xmath384 .
* @xmath385 + let @xmath386 .
we first note that if there is a leaf at distance two from @xmath83 , then by using @xmath109 and @xmath387 above , we can show that the size of the star at this leaf is at least as much as the given star .
we again consider two cases . *
* suppose that @xmath388 .
by symmetry , suppose @xmath389 . in this case ,
let @xmath390 . *
* suppose that @xmath391 .
label so that @xmath392 for @xmath393 ( in particular , @xmath88 is the parent of @xmath394 ) .
since neither @xmath395 nor @xmath396 is a leaf , they have at least one child , say @xmath397 and @xmath398 , respectively . in this case , let @xmath399 .
for this case , injection is less obvious .
we show it by contradiction as follows .
let @xmath400 .
we may assume that @xmath401 and let @xmath402 be the child of @xmath88 and @xmath403 be the child of @xmath402 ; then certainly @xmath404 gives a cycle in @xmath127 , a contradiction .
we believe that conjecture [ stree ] holds true for all @xmath3 .
however , it is harder to prove because it is not true that every leaf centered star is bigger than every non - leaf centered star ; an example is illustrated in figure [ figtree ] .
for each vertex , the first number denotes the label , while the second number denotes the size of the star centered at that vertex .
we note that @xmath407 , while @xmath408 .
however , we note that the maximum sized stars are still centered at leaves @xmath409 and @xmath405 .
fix positive integers @xmath414 and @xmath415 with @xmath416 .
let @xmath417 be the graph obtained from the complete bipartite graph @xmath418 and @xmath419 by identifying one endpoint of @xmath419 to be a vertex in @xmath418 lying in the bipartition of size @xmath37 .
let @xmath128 be the other endpoint of the path , and let @xmath420 be a vertex in @xmath418 lying in the bipartition of size @xmath414 , of degree @xmath37 .
an example is shown in figure [ treex2 ] .
we show that a similar construction acts as a counterexample for all @xmath425 .
given @xmath425 , consider the graph @xmath426 , @xmath427 .
let @xmath128 and @xmath420 be as defined before , with @xmath428 and @xmath429 .
let @xmath430 and @xmath431 .
we have @xmath432 and @xmath433 .
it follows that , for @xmath427 , @xmath434 .
if we consider trees , it can be seen that the tree in figure [ figtree ] satisfies this property .
it is also not hard to show that the path @xmath136 satisfies this property , since for all @xmath2 , @xmath435 holds for each @xmath113 .
observe that the vertex labeled @xmath443 , with degree @xmath37 , and the vertex labeled @xmath82 , with degree @xmath133 , lie in the same partite set , but we have @xmath444 and @xmath445 .
note that , in this example , @xmath446 .
another counterexample , with @xmath447 and @xmath406 , is shown in figure [ ct2 ] . in this section ,
we give a proof of theorem [ ladder ] , which states that the ladder graph @xmath134 is @xmath133-ekr for all @xmath160 .
first , we state and prove a claim about maximum star families in @xmath134 .
we prove the claim for @xmath455 .
the claim is obvious if @xmath139 , so suppose @xmath140 .
let @xmath371 be a star centered at some @xmath456 . without loss of generality , we assume that @xmath457 for some @xmath458 .
we now construct an injection from @xmath371 to @xmath459 .
define functions @xmath370 and @xmath460 as follows .
@xmath461 consider the function @xmath463 .
for every @xmath373 , define @xmath464 and similarly for @xmath460 .
we define a function @xmath465 as follows .
@xmath466 clearly , @xmath467 for every @xmath468 we will show that @xmath469 is an injection .
suppose @xmath470 and @xmath284 .
we show that @xmath471 . if both @xmath142 and @xmath227 are in the same category(out of the three mentioned in the definition of @xmath469 ) , then it is obvious .
so , suppose not . if @xmath472 and @xmath473 , then @xmath474 , but @xmath475 .
then , let @xmath142 be in either of the first two categories , and let @xmath227 be in the third category . then ,
@xmath476 , but @xmath477 .
this holds because otherwise , we would have @xmath478 , a contradiction
. we do induction on the number of rungs .
if @xmath315 , we have @xmath479 , which is trivially @xmath3-ekr for @xmath236 and vacuously true for @xmath316 and @xmath381 . similarly , for @xmath307 , @xmath480
, so it is trivially @xmath3-ekr for each @xmath481 and vacuously true for @xmath381 .
so , let @xmath140 .
the case @xmath236 is trivial .
if @xmath316 , since @xmath482 and @xmath483 , we can use theorem [ 2ekr ] to conclude that @xmath0 is @xmath37-ekr .
so consider @xmath0 such that @xmath140 and @xmath381 . if @xmath484 , the maximum size of an intersecting family of independent sets of size @xmath133 is @xmath109 , so @xmath133-ekr again holds trivially .
so , suppose @xmath485 .
let @xmath486 , @xmath487 .
also , let @xmath488 .
define a function @xmath248 as follows .
@xmath489 let @xmath490 be intersecting .
we prove the proposition for @xmath504 .
the proof for @xmath505 follows similarly .
each @xmath510 contains @xmath511 .
however , no member in @xmath506 contains @xmath511 .
thus , @xmath504 is a disjoint union . to show that it is intersecting ,
observe that @xmath506 is intersecting since for any @xmath512 , @xmath513 and @xmath514 are intersecting .
also , @xmath507 is intersecting since each member of the family contains @xmath511 .
so , suppose @xmath515 and @xmath510
. then , @xmath516 and @xmath517 are intersecting .
[ p2lad ] if @xmath450 , where @xmath485 , then we have @xmath518 is also a member of @xmath519 , containing neither @xmath520 nor @xmath521 .
each @xmath522 contributes two members to @xmath519 , @xmath523 and @xmath524 .
also , @xmath525 .
this completes the argument .
* @xmath527 and @xmath528 .
+ in this case , we must have @xmath529 and @xmath530 for some @xmath531 and hence , @xmath532 .
also @xmath533 .
so , using equation [ eqlad ] , propositions [ plad ] and [ p2lad ] and the induction hypothesis , we have @xmath534 * without loss of generality , we suppose that @xmath535 if @xmath536 , then @xmath537 , so we are done by proposition [ p2lad ] .
so , suppose @xmath538 .
we again consider two cases . 1 .
suppose @xmath539 and @xmath540 .
+ we note that at most one out of @xmath541 and @xmath542 can be nonempty .
we also note that @xmath543 and @xmath544 .
so , using proposition [ p2lad ] @xmath545 2 .
suppose that either @xmath546 or @xmath547 let @xmath548 and @xmath549 .
we have @xmath550 .
so , we have @xmath551 or @xmath552 so , @xmath553 . if @xmath554 , then @xmath555 , so @xmath556 also , @xmath557 since @xmath558 is nonempty .
if @xmath559 , then @xmath560 . thus , in either case , @xmath561
thus , using equation [ eqlad ] and proposition [ p2lad ] , we are done .
a similar argument works if @xmath562 is nonempty .
j. c. meyer , quelques problmes concernant les cliques des hypergraphes @xmath415-complets et @xmath563-parti @xmath469-complets , hypergraph seminar , columbus , ohio , 1972 , springer , new york , 1974 , pp .
127 - 139 . | one of the more recent generalizations of the erds - ko - rado theorem , formulated by holroyd , spencer and talbot @xcite , defines the erds - ko - rado property for graphs in the following manner : for a graph @xmath0 , vertex @xmath1 and some integer @xmath2 , denote the family of independent @xmath3-sets of @xmath4 by @xmath5 and the subfamily @xmath6 by @xmath7 , called a star .
then , @xmath0 is said to be @xmath3-ekr if no intersecting subfamily of @xmath5 is larger than the largest star in @xmath5 . in this paper
, we prove that if @xmath0 is a disjoint union of chordal graphs , including at least one singleton , then @xmath0 is @xmath3-ekr if @xmath8 , where @xmath9 is the minimum size of a maximal independent set .
we will also prove erds - ko - rado results for chains of complete graphs , which are a class of chordal graphs obtained by blowing up edges of a path into complete graphs .
we also consider similar problems for ladder graphs and trees , and prove preliminary results for these graphs .
* key words . *
intersecting family , star , independent sets , chordal graphs , trees |
the fundamental role played by thermally pulsing asymptotic giant branch ( tp - agb ) stars in the chemical evolution has been early recognized @xcite .
these stars are responsible for the nucleosynthesis of the main and the strong components of the _ s_-process @xcite and contribute to the synthesis of various light elements , such as li , c , n and f. these stars are bright red giants ( @xmath18 ranging from @xmath19 to @xmath20 ) . their compact co cores , whose mass ranges between 0.5 and 1 m@xmath21 , are progressively cooled by plasma neutrino emission . owing to the efficient heat conduction provided by degenerate electrons ,
the cores are maintained in a quasi isothermal state .
outside , there is an extended h - rich envelope , largely convective and progressively eroded by an intense mass loss ( from @xmath22 up to @xmath23 m@xmath21/yr ) .
the most interesting nucleosynthesis processes occur in a thin he - rich zone ( @xmath24 m@xmath21 ) , located between the h and the he burning shells ( hereafter he intershell ) .
the energy irradiated by these stars is mainly provided by the hydrogen burning .
he burning is off during most of the agb lifetime , so that the mass of the he intershell increases as a consequence of the advancing h burning shell .
then , periodic thermonuclear runaways ( _ thermal pulses _ , tps ) are driven by violent he ignitions occurring when the he buffer attains a critical value @xcite .
owing to the sudden release of energy by the @xmath25 reactions , the he intershell becomes dynamically unstable against convection , while the external layers expand and cool down . as a consequence of the convective mixing ,
the whole he intershell becomes enriched in c ( the main product of the partial he burning ) and heavy elements , as produced by the s process ( slow neutron captures ) . after a short period ( 10 to 200 yr ) , the thermonuclear runaway ceases , a stationary he burning takes place and convection disappears within the he intershell .
meanwhile , at the base of the h - rich envelope , the outgoing nuclear energy flux and the dropping temperature cause the radiative gradient to exceed the adiabatic gradient at positions more and more close to the he intershell .
finally , after a few hundred years , the convective envelope penetrates the he intershell zone ( third dredge up , tdu ) .
@xcite showed that all tp - agb stars may undergo recurrent third dredge up episodes , provided their envelope mass is sufficiently large .
the minimum envelope mass for the occurrence of the tdu is of a few tenths of m@xmath21 , but the precise value depends on the core mass and on the envelope composition .
the occurrence of many third dredge up episodes in thermally pulsing low mass agb stars has two important consequences .
first of all it is responsible for the chemical modification of the envelope , which becomes progressively enriched in primary c and in the products of the s process .
this provides a simple explanation for the observed spectroscopic sequence of agb stars , from the m to the c type , through the ms and s stars @xcite .
the carbon excess , mixed within the envelope by previous third dredge up episodes , causes a significant increase of the radiative opacity in the envelope , so that the effective temperature sinks , the stellar radius increases and the mass loss rate rises . the second important consequence of the penetration of the convective envelope into the he- and c - rich he intershell concerns the formation of the so called @xmath3c pocket , where the main and the strong components of the classical s process are built up ( see * ? ? ?
indeed , when the convective envelope recedes after a third dredge up episode , it leaves a small region ( of the order of a few @xmath23 m@xmath21 ) characterized by an increasing proton profile embedded in a bath of c and he ( about 20 and 80 % , by mass fractions , respectively ) . during the relatively long period that elapses between two subsequent thermal pulses ( interpulse )
, this region heats up ; then , @xmath3c is firstly produced through the @xmath26c@xmath27p@xmath28n(@xmath29)@xmath3c and , later on ( at @xmath30 k ) , the @xmath3c@xmath31n)@xmath4o provides the slow neutron flux required to activate the s process nucleosynthesis @xcite .
this pocket , strongly enriched in heavy elements , is engulfed into the convective zone generated by the subsequent thermal pulse .
when the temperature at the base of this convective zone exceeds @xmath32 k , a small neutron burst is powered by the @xmath14ne(@xmath0,n)@xmath16 mg reaction .
in low mass agb stars of nearly solar metallicity , this second neutron source is marginally activated @xcite .
several observations confirm such a theoretical expectation ( see e.g. * ? ? ?
* ; * ? ? ?
* ) . however , this second neutron source plays a more relevant role at low metallicity , even in low mass agb stars .
our knowledge of the physical properties of agb stars and of their nucleosynthesis is mainly based on observations and models for intermediate age stellar populations of the milky way .
the recent availability of optical and near ir spectroscopy of agb stars belonging to nearby galaxies , dwarf spheroidals as well as the magellanic clouds , allow us to extend our investigation to metal poor stellar populations with age of the order 1 - 2 gyr @xcite .
in addition to that , the growing number of studies on c - enhanced metal - poor stars enriched in s - elements ( cemps ) , as stimulated by dedicated surveys @xcite provide new hints on the nucleosynthesis occurred in the now extinct agb - halo population , as due to stars whose initial masses were in the range 1 to 7 m@xmath21 .
the aim of this paper is to investigate how the theoretical scenario changes with decreasing metallicity , from the solar value to that of the metal poor stars belonging to the galactic halo or to the nearby galaxies .
extant stellar models have been computed by adopting limited networks .
for example , @xcite include all the relevant processes and isotopes from h to si , plus a few fe - group isotopes and an artificial neutron sink to account for heavy element neutron captures . in this way
, they can provide a reasonable estimate of the light element yields .
our approach is different .
we compute evolutionary models , followed from the pre - main sequence up to the agb tip , of stars with initial mass 2 m@xmath1 and different metallicities ( namely , @xmath2=0.0138 , 0.006 , 0.003 , 0.001 and 0.0001 ) , by adopting a full nuclear network that includes all the stable and the relevant unstable isotopes from hydrogen to bismuth . in our previous works @xcite ,
full nucleosynthesis calculations were obtained by means of a post - process code : in such a case , the main physical parameters were obtained by means of stellar models calculated with a restricted nuclear network involving key isotopes and reactions @xcite .
then , an ad hoc ( average ) @xmath3c pocket was imposed . in the present work stellar evolution and nucleosynthesis
are coupled .
if the faster post - process calculation allow us to investigate a large area in the parameter space , the present , fully coupled , calculations shed light on the possible feedbacks between chemical and physical evolution .
these new models include several improvements in the input physics and in the computational algorithms ( see * ? ? ?
* thereinafter sgc06 ) : the most important for the agb evolution are illustrated in
the calibration of some free parameters used by the mixing algorithm are discussed in [ beta ] .
results are presented in [ models ] , [ varied ] and [ nsynt ] .
a final discussion follows in [ concl ] .
the stellar models here presented have been computed by means of the franec code ( * ? ? ? * and references therein )
. an updated description of the various input physics adopted in the agb computations can be found in sgc06 .
owing to the particular relevance for the present paper , let us describe the radiative opacity , some key nuclear reaction rates , the mixing algorithm and the mass loss rate .
as discussed by @xcite , the increase of the radiative opacity caused by the c dredged up induces important changes in the physical structure of the outermost layers of an agb star , with important consequences on the effective temperature , the radius and the mass loss rate .
in addition to that , when the cn cycle takes place in the innermost part of the convective envelope or immediately below it , an important fraction of the primary c that is dredged up to the surface may be converted into n. this is a common phenomenon in massive agbs ( the so called hot bottom burning , hbb ) @xcite , even if there are several observational evidence proving that deep - mixing processes can also occur in low mass agbs ( cool bottom process , cbp ) ( see @xcite ) . as a consequence of the carbon dredged up
, new molecular species may form in the cool atmosphere of an agb star , like cn , hcn or c@xmath33 .
note that since the amount of primary c synthesized by the 3@xmath0 reaction and mixed by the third dredge up is of the same order of magnitude at any metallicity , the resulting overabundance of this element ( and eventually n ) is very large at low metallicity .
whilst at solar metallicity the final c enhancement in the envelope may be a factor of 2 to 4 , at @xmath2=0.0001 it rapidly attains a factor of 1000 .
for this reason , if the adoption of a proper radiative opacity ( including enhancements of c and n ) may only introduce limited quantitative changes in the overall picture of agb stars with solar metallicity , its use at low metallicity is absolutely mandatory , because it leads to substantial modifications of the whole theoretical scenario ( see * ? ? ? * ) .
therefore , below @xmath34 k we use new opacity tables derived by means of the coma code @xcite , which include all the molecular and atomic species relevant for agb atmospheres . at higher temperatures ,
the opacity tables have been calculated by means of the the opal web facility ( http://www-phys.llnl.gov/research/opal/opal.html ) .
table [ tab1 ] lists the c and the n enhancement factors ( with respect to the solar scaled values ) adopted for the opacity tables at different metallicities .
all the elements ( except h , he , c and n ) are assumed to be solar scaled ( namely [ m / fe]=0 ) .
the enhancement of the @xmath0 elements ( o , ne , mg , si , s , ca ) , more appropriated for low metallicity halo stars , is not considered here , given that it only adds minor effects on the rosseland mean opacity . the full database providing low temperature rosseland opacities for a wide range of metallicities with varying carbon and nitrogen abundances is described in @xcite .
the nuclear network is essentially the same already described in section 6 of sgc06 : it includes about 500 isotopes ( from h to bi ) and more than 700 nuclear reactions ( charged particle reactions , neutron captures and @xmath35-decays ) .
however , few reaction rates have been changed or updated . in particular , concerning reactions among charged nuclei , the @xmath36n(@xmath0,@xmath37)@xmath38f rate is now taken from @xcite , whilst for the @xmath14ne(@xmath0,n)@xmath16 mg rate we follow @xcite . then , we have updated the @xmath39sm(n,@xmath37)@xmath40sm reaction rate , following the prescriptions reported by @xcite , and the cesium isotopes neutron captures rates , as suggested by @xcite .
concerning the @xmath41lu unthermalized isomeric state ( see section 7.5 of sgc06 ) , we adopt an isomeric ratio ir=0.2 for temperatures lower than @xmath42 k and ir=0.25 for higher temperatures . for a more detailed analysis of this important branching
see @xcite . mixing caused by convection
is obtained by means of a time - dependent mixing scheme in which the degree of mixing between two mesh points , whose separation is @xmath43r , is linearly dependent on the ratio between the time step and the mixing time scale . ] .
an exhaustive description of the mixing algorithm can be found in section 8 of sgc06 ( see also * ? ? ?
let us here recall the main concepts .
the average convective velocity ( @xmath44 ) is calculated according to the mixing - length theory @xcite , whereas the boundaries of the convective regions are located as given by the schwarzschild criterion . in order to handle the instability of the convective boundary taking place during the third dredge up episodes , when the h - rich envelope penetrates the h - depleted core , we assume that the convective velocity drops to 0 following an exponential decline , namely : @xmath45 where _ d _ is the distance from the formal convective boundary ( as defined by the schwarzschild criterion ) , @xmath46 is the velocity at the formal convective boundary , @xmath47 is the pressure scale height and @xmath35 is a free parameter .
all the evolutionary sequences presented in [ models ] and [ varied ] of the present paper have been obtained assuming @xmath48 . in the next section we will describe the procedure followed to fix the value of the @xmath35 parameter . a similar mathematical profile has been already applied to the calculation of agb models by @xcite ( see also * ? ? ?
however , @xcite used an exponential decline equation to calculate the diffusion coefficient , instead of the convective velocity , and applied this prescription to all the convective boundaries . in this way
, they always obtain overshoot at the top and at the bottom of any convective zone , while in our models eq .
[ param ] works only when @xmath49 , as it happens in the case of the bottom of the convective envelope at the time of the third dredge up .
a second important difference concerns the mixing scheme , which determines the chemical coupling among mesh points within the same convective zone . in the diffusive mixing adopted by @xcite ,
the variation of the composition during a given time step depends on the second derivative of the composition with @xmath50 ( the radial coordinate ) , whilst in our algorithm this relationship is linear .
these differences directly affect the profile of protons in the transition zone between the fully - convective envelope and the radiative core at the time of the third dredge up and the temperature attained in the he - rich intershell during a thermal pulse .
we will discuss the consequences on the nucleosynthesis in the next section .
concerning mass loss , we adopt a reimers formula ( @xmath51 ) for the pre - agb evolution , whilst for the agb , we follow a procedure similar to the one described by @xcite , but revising the mass loss - period and the period - luminosity relations , taking in to account more recent infrared observations of agb stars ( see sgc06 for the references ) .
a comparison among our mass loss rate and those obtained by means of the reimers formula and the vassiliadis and wood prescriptions can be found in section 5 of sgc06 .
an important remark concerns the dependence of the mass loss rate on the metallicity .
the mass loss - period relation we use is based on data relative to galactic agb stars having nearly solar iron . in a recent paper
, @xcite showed that the same relation applies also to carbon stars in the magellanic clouds , whose metallicity is lower , on average , than that of the galactic c - stars .
similar indications come from spitzer observations of mass loss of carbon stars belonging to dwarf spheroidal galaxies ( see * ? ? ?
* ) . in
[ varied ] and [ vario ] we discuss the effect of different mass loss prescriptions on the evolution and nucleosynthesis of the @xmath2=0.0001 model .
in this section we describe the models obtained by changing the value of the @xmath35 parameter introduced in the previous section to treat the instability of the convective boundary layer , arising when the h - rich convective envelope penetrates the he - rich zone . during a third - dredge - up episode , a chemical discontinuity forms at the convective boundary , causing an abrupt change of the radiative opacity . if a bare schwarzschild criterion is adopted to limit the mixing , the average convective velocity drops from about @xmath52 cm / s to 0 in 2 adjacent mesh points . in this condition
, any perturbation causing mixing of material across the boundary layer leads to an increase of the opacity in the underlying stable zone , which immediately becomes unstable against convection .
as it is well known , a similar condition occurs at the external border of the convective core during the central he - burning phase @xcite .
the most important consequence of this instability is the propagation of the convective instability ; if in the case of the central - he burning such an occurrence implies the growth of the convective core , in the case of the tdu a deeper penetration of the convective envelope into the h - exhausted core is expected .
actually , it should exist a transition region , between the fully convective envelope and the radiatively stable h - exhausted core , where the convective velocity smoothly decreases to 0 and , therefore , a partial mixing takes place . in such a way ,
after any dredge - up episode , a zone with a smoothed h profile is left behind .
since the amount of @xmath26c is rather large ( about 20 % by mass fraction ) in this transition zone , at h re - ignition those few protons are captured by the abundant @xmath26c and form a @xmath3c pocket .
later on , the s process nucleosynthesis is activated , through the @xmath3c@xmath31n)@xmath4o reaction .
note that a @xmath36n pocket and a @xmath53na pocket form , partially overlapped to the @xmath3c pocket ( see next section ) .
the mixing algorithm described in the previous section mimics the formation of such a transition zone .
the free parameter ( @xmath35 ) , used to tune the decline of the average convective velocity at the base of the convective envelope , determines the extension of such a transition zone and , in turn , the amount of @xmath3c available for the neutron capture nucleosynthesis . to calibrate such a parameter ,
we have repeated the calculation of the same sequence tp - interpulse - tp for @xmath54 .
this test has been performed at two metallicities ( @xmath2=0.0138 and @xmath2=0.0001 ) in order to test the sensitivity of our calibration procedure when changing the metal content .
results are illustrated in fig .
[ fig1a ] , fig .
[ fig1b ] , fig .
[ fig2a ] and fig .
[ fig2b ] . left panels refer to the epoch of the maximum penetration of the convective envelope during the third dredge up . in the right panels ,
we show the same region , but after the h re - ignition , once the convective envelope recedes and the @xmath3c pocket is fully developed .
the results of these test models are quantitatively summarized in table [ tab2 ] , where in columns 1 to 7 we report : the value of @xmath35 , the mass @xmath55 of the h - depleted material that is dredged up , the ratio @xmath56 ( being @xmath57 the mass of the material that has been burned by the h - shell during the previous interpulse period ) , the mass @xmath58 of the zone where the mass fraction of _ effective _ @xmath3c is larger than @xmath59 , the product @xmath60 at the epoch of the maximum penetration of the convective envelope , the total mass @xmath61h of hydrogen left below the formal convective boundary at the same epoch and the total mass @xmath62 of _ effective _ @xmath3c within the pocketc in a given mesh point is defined as @xmath63 , where @xmath64 and @xmath65 are the mass fractions of @xmath3c and @xmath36n , respectively . owing to its large neutron capture cross section , @xmath36n is the most efficient neutron poison , so that the s - process nucleosynthesis is significantly depressed where it becomes comparable to the @xmath3c .
for this reason , we define the mass of the @xmath3c pocket as the mass of the zone where @xmath66 ( see fig . [ figdef ] ) . ] .
[ figdef ] illustrates the definitions of the various quantities reported in table [ tab2 ] . @xmath55 and the masses of the effective @xmath3c for both metallicities
are also plotted , as a function of @xmath35 , in fig .
[ fig3 ] and fig .
[ fig4 ] , respectively .
the third dredge up occurs at lower core masses when a velocity profile is applied at the bottom of the convective envelope .
moreover , larger @xmath35 values imply deeper tdu . on the contrary , the total amount of @xmath3c in the pocket shows a maximum at @xmath48 .
note that the pocket practically disappears for the largest @xmath35 value , reflecting the fact that when enough h is mixed into the transition zone , the cn cycle goes to the equilibrium and @xmath36n , rather than @xmath3c , is produced . for @xmath48 ,
the effective @xmath3c masses are 7.4 @xmath67 and 5.5 @xmath67 m@xmath21 , at @xmath2=0.0138 and @xmath2=0.0001 , respectively .
these values are close to the standard one adopted by @xcite , namely 4 @xmath67 [email protected] is often referred as the st ( _ standard _ ) case for low mass agb nucleosynthesis models . ]
@xcite showed how this choice of the @xmath3c - pocket allows the buildup of the main and the strong components , _
i.e. _ the distribution of s - elements from the sr - y - zr peak up to the pb - bi peak in the solar system , whilst other works @xcite argue that a certain spread of the effective @xmath3c mass is required in order to reproduce the heavy elements overabundances in a large sample of galactic c(n type ) stars .
more recently , @xcite confirm the need for such a spread , extending their analysis to cemps stars .
it is worth to note that whilst @xcite assume a constant effective @xmath3c mass , in our models we find a characteristic evolution of the @xmath3c pocket ( see fig .
[ fig5 ] ) .
in particular , a maximum mass of @xmath3c is attained after very few thermal pulses ( 2 or 3 ) ; later on , the pocket progressively shrinks , in mass , until it disappears during the late part of the agb evolution ( see next section ) . nevertheless , the success of the post - process calculations in reproducing a large amount of observational data concerning agb stars @xcite , their progeny on the post - agb @xcite , galactic evolution of s process elements @xcite and the isotopic composition of c - rich pre - solar grains ( sic ) @xcite suggests to adopt @xmath35 values allowing the formation of @xmath3c pocket whose average size is similar to the st case . on the other hand , as shown in fig .
[ fig4 ] , only values of @xmath35 close to 0.1 provide enough _ effective _ @xmath3c .
for this reason , in all the computation here presented we have adopted @xmath48 ( but see [ concl ] for further remarks ) .
let us come back to the differences between our mixing scheme and the one adopted by @xcite .
basing on a few additional models computed by using a diffusion algorithm instead of our linear mixing , we have verified that in this case the extension in mass of the zone where the proton abundance left by the tdu is @xmath68 can not be larger than about @xmath69 m@xmath21 ( for any value of the @xmath35 parameter)c . if @xmath70 , @xmath36n rather than @xmath3c would be produced , while for @xmath71 too few @xmath3c is synthesized . ] . instead , as shown in fig .
[ figdef ] , when our mixing algorithm is adopted , the same region extends for a few @xmath23 m@xmath21 .
such a difference affects the total mass of the effective @xmath3c within the pocket , which may be about 20 times larger in the case of a linear mixing scheme ( for comparisons , see figure 4 in * ? ? ?
as a consequence , the resulting s - process yields are significantly reduced if a diffusive mixing is adopted .
a further difference concerns the overshoot at the bottom of the convective zone generated by a thermal pulse . in our models , the exponential decline described by eq .
[ param ] switches off when , as in this case , the radiative gradient is equal to the adiabatic gradient at the convective border , so that @xmath46=0 . on the contrary ,
@xcite force a certain overshoot below the base of this convective zone , in spite of the strong negative difference between the radiative and adiabatic gradient and the huge entropy barrier generated by the shell - he burning .
this implies stronger thermal pulses and larger temperatures at the base of the he - rich intershell , thus increasing the efficiency of the s - process nucleosynthesis powered by the @xmath14ne@xmath31n)@xmath16 mg reaction ( see * ? ? ?
* ) . a reliable ( hydrodynamical ) model of convection capable to describe the mixing across the boundaries of the major convective zones should provides a more realistic description of the agb evolution and nucleosynthesis ( for a recent attempt , see * ?
on the other hand , comparisons between the predicted agb nucleosynthesis and the abundances observed in agb stars may be used to constrain the efficiency of convection .
this is the approach we follow in the present paper .
in this section we present and compare 5 evolutionary sequences of agb models having the same initial mass ( 2 m@xmath1 ) , but different initial composition , namely : ( @xmath2 ; @xmath72)= ( 0.0138 ; 0.269 ) , ( 0.006 ; 0.260 ) , ( 0.003 ; 0.260 ) , ( 0.001 ; 0.245 ) , ( 0.0001 ; 0.245 ) .
the mass is representative of low mass agb stars and the five metallicities almost span the entire metal distribution of our galaxy and of the extragalactic resolved stellar populations , like the magellanic clouds , m31 and several dwarf spheroidal galaxies .
the most metal - rich model corresponds to the composition of the pre - solar nebula , as derived by means of an up - to - date standard solar model obtained by adopting the latest compilation of solar abundance ratios ( @xcite ; see @xcite for more details ) .
the 5 evolutionary tracks , from the pre - main sequence up to the end of the agb , are reported in fig .
the tp - agb phase is characterized by large oscillations in luminosity , associated to the expansions and contractions powered by thermal pulses , and an evident red excursion that is the direct consequence of the formation of new molecular species taking place after the transition from o- to c - rich atmosphere . in fig .
[ fig7 ] we show the evolution , during the tp - agb phase , of the positions , in mass coordinates , of the inner border of the convective envelope , of the location of the maximum energy production within the h - burning shell and of the location of the maximum energy production within the he - intershell , top to bottom line , respectively .
the last 2 lines separate 3 main regions within the star : the innermost zone is the degenerate co core , the intermediate one is the he intershell and the more external is the h - rich envelope .
the evolution of the 3 lines mark the most important events occurring during the tp - agb phase . between two thermal pulses ( interpulse phase ) ,
the h - burning shell advances in mass , while the he burning is practically off .
then , when the he intershell attains a critical mass , the he flash starts .
after a few years , during which the thermonuclear runaway occurs , a quiescent he burning settles in , while the h burning dies down . later on , within about 1000 yr , the convective envelope penetrates the h - he discontinuity ( third dredge up ) .
note that for a certain time during the interpulse , the line representing the maximum energy production within the he - intershell moves suddenly upwards . at that epoch ,
the 3@xmath0 reactions are practically extinguished in the whole he intershell , but the temperature in the region occupied by the newly formed @xmath3c pocket is large enough ( @xmath73 k ) to ignite the @xmath3c(@xmath0,n)@xmath4o reaction and the consequent s process nucleosynthesis .
indeed , the @xmath3c burning releases about 5 - 10 mev per reaction , depending on the energy of the consequent neutron capture on a seed nucleus .
note that this additional energy release has a negligible effect on the further physical evolution .
+ table [ tab3 ] reports some properties characterizing the 5 evolutionary sequences , namely ( from left to right ) : the progressive number of the thermal pulse ( @xmath74 ) , the total mass at the time of the onset of the thermal pulse ( @xmath75 ) , the corresponding mass of the h - exhausted core ( @xmath76 ) . ] , the mass of the h - depleted material dredged up ( @xmath55 ) , the maximum mass of the convective zone generated by the tp ( @xmath77 ) , the mass burnt by the h - shell during the previous interpulse period ( @xmath78 ) , the overlap factor _ r _
( defined as the fraction of @xmath77 already included in the convective zone generated by the previous tp ) , the @xmath79 factor ( defined as the ratio between @xmath55 and @xmath57 ) , the average number @xmath80 of neutrons captured per initial iron seed nucleus during the radiative @xmath3c pocket burning , the duration of the interpulse period preceding the current tp ( @xmath81 ) , the maximum temperature attained at the bottom of the convective zone generated by the tp ( @xmath82 ) , the surface metallicity after the dredge up ( @xmath83 ) and the corresponding c / o ratio .
finally , the total mass of the material cumulatively dredged up during the whole evolutionary sequence is reported in the last row ( @xmath84 ) .
as already discussed in @xcite , the tdu starts when the mass of the h - exhausted core exceeds a critical value , whereas it ceases when the envelope mass , eroded by the mass loss , is reduced down to a critical value .
in the five evolutionary sequences presented here , the first tdu episode takes place when ( from @xmath9 to @xmath85 ) m@xmath86 is 0.560 , 0.569 , 0.576 , 0.604 , 0.641 m@xmath21 .
note that at the lowest metallicities ( @xmath87 and 0.0001 ) a 2 m@xmath21 star already develops rather large core mass during the early - agb , so that the third dredge up occurs almost soon after the beginning of the tp - agb phase . at those metallicities , stars with smaller
mass may experience tdu episodes at smaller @xmath88 .
the last tdu episode takes place when @xmath89 is 0.534 , 0.183 , 0.207 , 0.292 , 0.242 m@xmath21 , for @xmath9 , 0.006 , 0.003 , 0.001 and 0.0001 , respectively .
+ as discussed in [ beta ] , the introduction of a smoothed profile of the convective velocities at the bottom of the envelope favors the occurrence of tdu at smaller core masses and enhances its efficiency , with respect to models where the bare schwarzschild criterion is adopted @xcite . as an example , in the 2 m@xmath21 model with @xmath90 of @xcite , the first tdu takes place when the core mass is 0.61 m@xmath21 and has a reduced efficiency with respect to the model presented here .
a second reason at the base of these differences is the rate of the @xmath36n(p,@xmath37)@xmath91o reaction : we are currently using the most recent low - energy laboratory measurement of this rate by @xcite , which is a factor of two lower with respect to the rate proposed by @xcite and @xcite .
being the @xmath36n(p,@xmath37)@xmath91o reaction the bottleneck of the cno cycle , a lower rate implies a reduced h - burning efficiency , stronger tps and larger tdus , as already verified by @xcite and confirmed by @xcite .
an important quantity affecting the nucleosynthesis is the maximum temperature attained at the base of the convective zone generated by a tp . during the agb evolution
, this value progressively increases , reaches a maximum and , then , slightly decreases toward the last tp ( fig .
[ fig8 ] , upper panel ) . in the most metal - rich model ,
@xmath82 remains always below 3.0@xmath92 k : this is due to the fact the h burning shell is efficient and , therefore , the tps are weaker with respect to lower metallicities ( see * ? ? ?
for that reason , in these models the @xmath14ne@xmath31n)@xmath16 mg reaction is only marginally activated . on the contrary , in the two most metal - poor models
, @xmath82 attains higher values ( up to 3.2@xmath93 k ) , so that the @xmath14ne@xmath31n)@xmath16 mg adds a significant contribution to the s process nucleosynthesis , in particular for the treatment of the branchings along the s - path .
we will come back on this issue in the next section .
concerning @xmath55 , it depends non - linearly on the metallicity , the envelope mass and the core mass ( see formula 3 in * ? ? ?
fixing the values of the other two parameters , it is generally larger at lower metallicity .
it is also larger at larger envelope mass or / and core mass . during the agb evolution , the core mass increases , while the envelope mass decreases , so that @xmath55 initially increases , following the growth of @xmath88 , but then decreases , when the envelope erosion by mass loss is stronger .
this is clearly showed in fig .
[ fig8 ] ( lower panel ) .
looking at the total mass of the material that is cumulatively dredged up , it has a minimum in the case of the most metal - rich evolutionary sequence ( @xmath94 m@xmath21 ) , whilst the maximum is attained at @xmath950.001 ( @xmath96 m@xmath21 ) . in [ varied ]
we will show how the assumed mass loss rate affects this quantity . the mass of the convective zone generated by a tp shrinks during the agb evolution .
in all of the 5 sequences , the last @xmath97 is about a factor of two smaller than the first .
it is also a factor of two smaller at the lowest metallicity with respect to the largest one . on the contrary ,
the @xmath79 factor is smaller at larger metallicities , whilst the overlap ( @xmath50 factor ) is similar in all the five sequences .
the @xmath2=0.0138 sequence attains the c - star condition ( c / o@xmath981 ) after 7 dredge up episodes and the final c / o is 1.88 .
more metal poor models becomes c - stars sooner and develop larger final c / o ratios : in the @xmath2=0.0001 case , for example , the final c / o is 53 .
this is mainly due to the lower initial amount of o , rather than to the deeper dredge up .
in this section we will show how the present theoretical scenario is affected by the assumed treatment of hydrodynamical phenomena , not explicitly included in the hydrostatic equations , such as mass loss and convection , whose efficiency in the models relies on some free parameters . for this purpose
, we have calculated 2 additional evolutionary sequences , both at @xmath2=0.0001 . in the first one
we have adopted a classical reimers formula ( @xmath99 ) , up to the tip of the agb , instead of that based on the calibrated mass loss - period - luminosity relation ( see [ code ] ) .
the second evolutionary sequence has been obtained by changing the value of the mixing length parameter ( @xmath100 , where @xmath101 is the mixing length ) from 2.15 , as required by the standard solar model ( see * ? ? ?
* ) , to 1.8 .
the evolutions of @xmath82 and @xmath55 for these two additional models are compared with the reference model in fig .
either the mass loss rate and the mixing length parameter are usually calibrated on stars with solar or nearly solar composition and little is known about the validity of these calibrations at low metallicity ; for a discussion on the possible calibration of the mixing length at different @xmath2 see @xcite and @xcite .
the early tp - agb evolution of the reimers model , which is mainly controlled by the growth of the h - exhausted core , is similar to the one obtained in the case of the reference model .
however , when the envelope erosion becomes important , the two sequences depart from each other .
in particular , as a consequence of the higher mass loss rate , the reference model terminates sooner . as a result , the reimers model experiences a larger number of tps and third dredge - up episodes .
the total mass cumulatively dredged up is 1.6 @xmath102 m@xmath21 , to be compared with @xmath103 m@xmath21 in the reference model ( see table [ tab4 ] ) .
the main effect of reducing the mixing length parameter is a smaller extension of the convective regions . as a consequence ,
the third dredge up is weaker and the temperature at the bottom of the convective zone powered by thermal pulses is lower ( see fig . [ fig9 ] ) . in this case
the total mass cumulatively dredged up is just @xmath104 m@xmath21 ( see table [ tab5 ] ) .
the corresponding effects on the nucleosynthesis are discussed in [ vario ] .
let us start with the solar metallicity model . as noted in [ models ] , owing to the low temperature within the convective zones generated by the various tps , the @xmath14ne@xmath31n)@xmath16 mg reaction plays a marginal role as a neutron source at high metallicities , so that the most important nucleosynthesis site is the transition region between the core and the envelope , where the @xmath3c pocket forms . in fig .
[ fig10 ] we report the mass fractions of selected isotopes in this zone , after the formation of the third pocket . actually , three different pockets can be distinguished .
the most internal one is the @xmath3c pocket ( solid line ) , which is partially overlapped to a @xmath36n pocket ( long dashed line ) .
note that the relevant s process nucleosynthesis occurs in the left ( more internal ) tail of the @xmath3c pocket , where the amount of @xmath36n , the major neutron poison , is low .
the third and smallest pocket is made of @xmath53na ( short - long - dashed line ) .
the latter forms where the abundance of @xmath14ne ( short - dashed line ) is comparable to the @xmath26c abundance ( dotted line ) , so that the @xmath14ne(p,@xmath37)@xmath53na reaction competes with the @xmath26c(p,@xmath37)@xmath3n in the proton capture game . as firstly suggested by @xcite , this @xmath53na pocket may provide a significant contribution to the synthesis of sodium in agb stars . the most interesting result obtained by coupling stellar structure evolution and nucleosynthesis concerns the variations of the size of the 3 sub - pockets when the star climbs the asymptotic giant branch . fig .
[ fig11a ] and fig .
[ fig11b ] show , for the solar metallicity model , the physical conditions in the core - envelope transition region at the time of the maximum penetration of the 3@xmath105 and of the 11@xmath106 ( _
i.e. _ the last ) tdu , respectively .
solid and dashed lines represent the hydrogen and the @xmath26c profiles , respectively .
the abundance curves in both figures have been shifted upward in order to match the pressure scale axis .
the slanting dashed area shows the region of the fully convective envelope ( as defined by the schwarzschild criterion ) , whilst the horizontal and the vertical dashed areas mark the regions where the @xmath3c and the @xmath53na pockets will form later ( the @xmath36n pocket , which overlaps with both the aforementioned pockets , is not reported in the plot for graphical reasons ) .
pressure and pressure scale height are also shown .
note as the pressure jump between the dense core and the loose envelope is definitely steeper in the case of the last tdu episode .
in fact , when comparing the two physical structures , it comes out that that the pressure of the h - exhausted core remains of the same order of magnitude , whilst the base of the envelope at the moment of the last tdu episode is more expanded with respect to the 3@xmath105 one , showing a pressure difference of about five orders of magnitude . in both cases ,
external convection penetrates down to a layer where the pressure is about 10@xmath107 dyne and @xmath47 is about 10@xmath108 cm , but due to the steeper pressure gradient , the extension of the zone with a smoothed h profile is significantly reduced in the last tdu episode . in this case , the resulting pocket is definitely smaller .
table [ tab6 ] lists the effective mass of @xmath3c for all the pockets produced in the @xmath2=0.0138 and @xmath2=0.0001 models . in the table , we report ( left to right ) : the mass of the h - exhausted core ( @xmath88 ) , the mass of effective @xmath3c in the pocket ( @xmath109c@xmath110 ) and the mass extension of the pocket ( @xmath111 ) . note how the last pocket is more than one order of magnitude smaller than the first one .
the shrinkage of the @xmath3c pocket implies a progressive decrease of the s process efficiency as the star evolves along the agb . for a long time
, it has been assumed that the synthesis of the main component of the s process would require the partial superposition of different neutron exposures : @xcite showed that an exponential distribution of these exposures could reproduce the observed @xmath112 curve ( see also * ? ? ?
@xcite noted that if the s process takes place at the base of the convective zone generated by a thermal pulse , the partial overlap of the recurrent convective zones provides , in a natural way , an exponential distribution of neutron exposures , thus reinforcing clayton s original suggestion .
however , as argued by @xcite ( see also @xcite ; @xcite ) , when the s process occurs during the interpulse in the @xmath3c pocket , the scenario becomes more complex and can not be described by a simple analytical function . beside its relatively small thickness , the stratification of the @xmath3c ( and of the neutron poisons , as @xmath36n ) within the pocket leads to stratified neutrons irradiations with different intensities .
in addition to that , owing to the variation of the overall amount of @xmath3c available in the pocket , as we find in self - consistent evolutionary models of agb stars , this stratification changes pulse by pulse .
in particular , the first pockets , the largest ones , give a major contribution to the overall agb nucleosynthesis compared with that of the smallest late pockets .
this is illustrated in fig .
[ fig12 ] , where we report , for the solar metallicity model , the production factors , within the he intershell at the time of the 1@xmath113 , 3@xmath105 , 6@xmath106 and 11@xmath106 tdu episodes , of nuclei whose synthesis is mainly ascribed to the s process ( namely , nuclei whose s process contribution is larger than 80% ) . , where @xmath114 is the abundance by number and @xmath115 is the corresponding solar scaled abundance . ]
the s process nucleosynthesis occurs in the radiative @xmath3c pocket and leaves a thin layer highly enriched with heavy elements that is located in the middle of the he intershell .
when the convection powered by the thermal pulse takes place , this material is spread out in the whole he intershell . in spite of the large dilution caused by the convective mixing , m@xmath21 , while the convective zones attain about @xmath116 m@xmath21 . ]
the s - element overabundances in the he intershell rapidly increase during the first tps .
a maximum is attained at the 6@xmath106 tp . at that time , in the he intershell the abundances of the s - only nuclei are on the average a factor of 500 larger than the initial ones ( 700 for the ls element and 300 for the hs elements ) .
later on , the production factors of the various s elements in the he intershell decrease .
this is due to the shrinkage of the late pockets , which are no more able to compensate the convective dilution .
nevertheless , quite large overabundances are found in the he intershell up to the agb tip , because of the partial overlap of the recurrent convective zones generated by the various tps .
the variations of the surface composition after selected tdu episodes of the solar metallicity model are reported in fig .
[ fig13 ] .
looking at the elements beyond fe ( z=26 ) , the most abundant species are those corresponding to isotopes with particularly small neutron capture cross sections .
in particular , the peaks corresponding to magic neutron numbers ( n=50 , 82 and 126 ) clearly emerge from the bulk production of the s process . in general ,
magic neutron nuclides act as bottlenecks of the s process . at solar metallicity ,
owing to the relatively large amount of iron seed , compared with the number of neutrons released by the radiative @xmath3c burning , the productions of the light , ls ( sr - y - zr ) , elements and the heavy , hs ( ba - la - ce - nd - sm ) , elements , namely the first and the second s - peaks , are favored with respect to the lead production ( the third s - peak ) . at the end of the agb evolution , the surface ratios [ hs / ls ] and [ pb / hs ] are @xmath117 and @xmath118 , respectively . this values are attained after a few tdu episodes , those following the first and largest @xmath3c pockets . later on , the overabundances continue to grow , as a consequence of the tdu , but these ratios remain constant . the final surface composition of the five models with different metallicity are shown in fig .
[ fig14 ] .
surface abundances after selected tdus are shown in table [ tab7 ] , table [ tab8 ] , table [ tab9 ] , table [ tab10 ] and table [ tab11 ] .
only key elements are tabulated . moreover , a selection of final elemental abundances , isotopic ratios and spectroscopic indexes are listed in table [ tab12 ] .
the @xmath3c(@xmath0,n)@xmath4o reaction is primary - like ( _ i.e. _ not directly affected by the metallicity of the pristine material ) , whilst the iron seeds scale with the metallicity .
thus , by lowering the metallicity , the number of neutron per seed nuclei progressively increases : this trend clearly results from column 9 of table [ tab3 ] . within the @xmath3c pockets , in fact ,
the average number @xmath80 of neutrons captured per initial iron seed nucleus increases with the metallicity , starting from about 40 at solar metallicity up to 4000 at @xmath85 involves neutron captures on both light elements and heavy elements .
neutron captures on light elements act as neutron poisons , thus decreasing the number of neutrons available for the s - process nucleosynthesis .
the major neutron poison , represented by the @xmath36n(n , p)@xmath36c reaction , has already been considered in the estimate of @xmath80 . ] .
therefore , the bulk of the s process nucleosynthesis first moves from ls elements to hs elements and , at low metallicities , it shifts directly to @xmath119pb , at the termination point of the s path .
this behavior is well reproduced by our models ( see fig .
[ fig14 ] ) . a second important change occurring at low z concerns the second neutron source , the @xmath14ne@xmath31n)@xmath16 mg . in low metallicity models , owing to the larger temperature attained at the inner border of the convective zone generated by the thermal pulse ,
a second neutron burst takes places , causing interesting changes in the he intershell composition . at variance with the nucleosynthesis in the @xmath3c pocket , since the maximum temperature in the convective shell generated by a tp increases pulse by pulse
( see fig .
[ fig8 ] ) , it is in the late part of the agb that the major effects induced by the second neutron burst become important . among them , we recall the overproduction of some neutron - rich isotopes . owing to the larger neutron density ( 10@xmath107 @xmath8 instead of 10@xmath120 obtained in the case of the radiative @xmath3c pocket ) ,
several interesting branchings along the s process path are activated .
thus , neutron - rich isotopes , whose production is otherwise prevented by @xmath35 decays of lighter isotopes , can be produced .
table [ tab13 ] reports some characteristic isotopic ratios sensitive to the neutron density .
the tabulated ratios relate to the branchings occurring at @xmath121kr , @xmath122rb , @xmath123zr , @xmath124xe and @xmath125ce , respectively . as a comparison , solar ratios are also reported .
note how they increase when the metallicity decreases .
in particular , the @xmath126rb is underproduced at large metallicity with respect to the lighter @xmath121rb , whereas it is overproduced at low z. this is a direct consequence of the opening of the @xmath121kr and @xmath122rb branchings , taking place when the neutron density exceeds @xmath127 neutron/@xmath128 ( see * ? ? ? * ) .
note that , since @xmath126rb is a magic neutron nucleus , the overall production of rb is significantly enhanced at low z. as a result , the ratio of rb and ls increases at low z. for example , log(rb / sr ) is 0.33 , 0.42 , 0.53 , 0.77 and 0.85 at @xmath2=0.0138 , 0.006 , 0.003 , 0.001 and 0.0001 , respectively . in fig .
[ ro_n ] we report the maximum neutron densities attained in the models at various metallicities . as stressed before , the lower the metallicity is , the higher the neutron density attained during the tp is .
the most striking consequence of the third dredge up is the surface carbon enhancement . by reducing the metallicity , the maximum enhancement with respect to iron increases from [ c / fe]=0.6 ( @xmath2=0.0138 model ) to [ c / fe]=3 ( @xmath2=0.0001 model ) .
obviously , this implies that the c / o ratio increases with decreasing the metallicity , passing from 1.87 at solar metallicity to 53 at @xmath2=0.0001 ( see table [ tab12 ] ) .
our models do not show evidences for hbb , therefore at low metallicities we obtain large c / n ratios and large @xmath26c/@xmath3c isotopic ratios .
note that possible reduction of these ratios due to deep - mixing processes ( as the cbp ) is not explicitly considered in our models . small enhancements of nitrogen are found , due to the dredge up of the thin region of incomplete h burning .
almost all the @xmath36n left by the h burning in the he intershell is converted into @xmath14ne during the thermal pulse phase , through the chain @xmath36n@xmath129f@xmath130o@xmath131ne .
the @xmath14ne plays a fundamental role , particularly in metal - poor stars .
it acts as a neutron source , as a poison and , at very low z , as a seed capable to counterbalance the scarcity of iron ( see * ? ?
note that at low z , most of the @xmath14ne is primary : at @xmath2=0.0001 , in fact , the amount of carbon in the envelope largely exceeds the pristine c+n+o since the first tdu episode . as a result , at the end of the agb the amount of @xmath14ne in the he intershell is comparable with that found in the same region of the solar metallicity model . correspondingly , the surface abundance of neon is significantly enhanced ( see table [ tab12 ] ) . despite the very low neutron capture cross section ,
the large abundance of @xmath14ne allows the production of light isotopes like na , mg and al .
sodium is further enhanced , because of the already mentioned formation of the @xmath53na pocket , due to protons capture in the most external layer of the core - envelope transition zone . at solar metallicity , the majority of the na overproduction is due to the ingestion of the @xmath53na pocket at the time of the tdu . at @xmath2=0.0001 , in addition to that , neutron captures on @xmath14ne taking place in the radiative @xmath3c pocket and during the thermal pulse account for 13% and 35% of the total na production , respectively @xcite . concerning o ,
it is only marginally affected by the internal nucleosynthesis at large z. in low metallicity models , the oxygen enhancements results larger ( [ o / fe]=1 at @xmath2=0.0001 ) because of the reduced initial @xmath4o abundance .
we recall that we have assumed a solar scaled initial composition , so that [ o / fe ] is nearly 0 at the beginning of the agb for all the models here presented .
thus , in order to compare theoretical o overabundances to those measured in c - enhanced stars belonging to the galactic halo ( cemp ) , one has to add an initial overabundance of , at least , @xmath132 dex . among the light elements , @xmath133f
is produced at all metallicities , with a significant production at the lowest one .
fluorine is synthesized by the @xmath91n@xmath134f reaction in the convective zone generated by a thermal pulse .
the @xmath91n production is due to the @xmath38o(p,@xmath135n reaction : therefore , the @xmath133f production requires the simultaneous presence of @xmath38o and protons ( see , e.g. , * ? ? ?
* ) . in the he
intershell , protons are available in the radiative @xmath3c pocket and at the beginning of a thermal pulse ( see , e.g. , * ? ? ?
* ) , released by the @xmath36n(n , p)@xmath36c reaction : neutrons are therefore required for a consistent production of @xmath91n .
+ in the radiative @xmath3c pocket neutrons are produced by the @xmath3c(@xmath0,n)@xmath4o reaction , while @xmath38o is synthesized by means of the @xmath36c(@xmath0,@xmath37)@xmath38f(@xmath136o chain .
+ few neutrons are also available at the beginning of a thermal pulse : they come from the burning of the @xmath3c left by the h - burning shell in the upper zone of the he intershell .
this @xmath3c remains unburned during the interpulse and is engulfed in to the convective zone generated by the tp .
the @xmath3c burns in a convective environment producing neutrons , which are almost totally captured by the ( still abundant ) @xmath36n , giving rise to a certain release of protons . at that time
, the temperature at the bottom of the convective shell is not high enough for a complete @xmath38o depletion , therefore @xmath91n is efficiently produced . the @xmath14ne@xmath31n)@xmath16 mg reaction ( which is the main source of neutrons during a tp ) does not contribute to the fluorine nucleosynthesis because , when it takes place , the temperature is so high that no @xmath36n and too few @xmath38o survive in the convective shell . in summary
, we have two main channels for the @xmath91n production in the he intershell . at solar metallicity
, the two sources equally contribute to the fluorine production ; at low metallicities , the @xmath91n accumulated in the radiative @xmath3c pocket is the main source of fluorine , because the @xmath3c left by the full cno burning is rather low for the major part of the agb lifetime ( but see [ vario ] ) .
one of the aims of this paper is to provide a complete and uniform set of agb yields , comprehensive of all chemical species , starting from hydrogen up to the pb - bi s process ending point .
yields of selected isotopes at different metallicities are reported in table [ tab14 ] : being the number of considered isotopes too large to provide it in a paper format , a constantly updated database of agb yields is available on the web and in the electronic version of this paper . according to @xcite ,
the yield is : @xmath137\frac{dm}{dt}dt\ ] ] where @xmath138 is the mass loss rate , while @xmath139 and @xmath140 stand for the current and the initial mass fraction of the isotope @xmath141 , respectively .
yields are given in solar mass units .
we recall that our calculation have been stopped after the last tdu episode , when the residual envelope mass is of the order of ( 0.2@xmath142 ) m@xmath21 . except for the natural decay of the eventually surviving unstable isotopes ,
the envelope composition is freezed after this moment . then , in computing the yields , we assume that the star loses the whole residual envelope through a single mass loss episode . in this section
we discuss the synthesis of a few short - lived radioactive isotopes , namely @xmath143al , @xmath144cl , @xmath145ca , @xmath17fe , @xmath146pd and @xmath147pb .
the present theoretical predictions for low mass agb stars may be used to interpret the evidence for the presence of these radioactive isotopes in the early solar system ( ess , see @xcite for a review on short - lived radioactive from agb stars ) and in pre - solar dust grains @xcite .
moreover , @xmath143al and @xmath17fe are particularly important in the field of @xmath37-ray astronomy @xcite : their detection from selected agb sources could be possible in the next future by means of @xmath37-rays instruments mounted on high - energy astronomical satellites .
let us start from these two short - lived isotopes .
the ground state of @xmath143al has a terrestrial half - life of @xmath148 yr , which is comparable to the duration of the whole tp - agb phase of a low mass star .
@xmath143al is produced in the h - burning shell by proton captures on @xmath16 mg .
within the h - depleted ( and @xmath143al - rich ) region , we need to distinguish two zones : the upper one , which extends from the tip of the convective shell generated by a tp up to the h - shell , and the lower one , which is engulfed in the convective shell . in the former region
@xmath143al survives and is dredged up to the surface , whilst in the latter one , owing to its large neutron - capture cross section , it is easily destroyed by neutron captures .
this happens in radiative conditions within the @xmath3c - pockets and , more importantly , within the convective tps . as a matter of fact , when the temperature at the bottom of the convective shell exceeds @xmath149 k , @xmath143al is easily destroyed . in our solar metallicity model ,
this condition is attained toward the end of the tp - agb phase , so that part of the @xmath143al engulfed in the convective shells is preserved .
this is not the case of the low metallicity models , where the @xmath14ne neutron source is efficient since the first tp ( see table [ tab14 ] ) . in that case , only the @xmath143al in the upper part of the he intershell ( not engulfed in the convective shell ) survives and is dredged up to the surface .
the synthesis of @xmath17fe during the radiative @xmath3c burning is limited , owing to the relatively fast @xmath35 decay of @xmath150fe ( half - life 45 days ) .
however , the higher neutron density generated by the convective @xmath14ne burning allows to open the @xmath150fe branching , leading to the production of @xmath17fe .
this explains the relatively large @xmath17fe yield obtained in low metallicity models . the yield is higher at @xmath2=0.001 than at @xmath2=0.0001 , because of the larger total dredged up mass .
a different production channel is active in the metal - rich models ( @xmath9 , @xmath10 and @xmath11 ) .
the appearance at the surface of @xmath17fe occurs after the first tdu episode .
this reflects a peculiarity of the first @xmath3c pocket of these models : the temperature developed in the pocket during the interpulse is not large enough to fully consume @xmath3c before the onset of the following thermal pulse @xcite .
then , a certain amount of unburned @xmath3c is engulfed into the convective zone generated by the tp and is rapidly burned at higher temperature .
the resulting neutron densities are rather large ( more than @xmath13 @xmath8 ) . in fig .
[ ro_n ] , we mark with an arrow the neutron peak corresponding to the ingestion of the first @xmath3c pocket into the convective shell generated by the following tp ( this fact occurring for the metal - rich models only )
. although the effect of this anomalous convective @xmath3c burning on the overall s process nucleosynthesis is negligible , some neutron - rich isotopes as , for example , @xmath17fe , preserve the signature of such a peculiar event .
note that the surface isotopic ratio @xmath17fe/@xmath151fe we obtain in the solar metallicity model ( namely @xmath152 ) is about 10 times higher then the value reported by @xcite , as obtained by means of a post - process calculation , where the effect of the peculiar first @xmath3c pocket was not considered .
+ let us finally discuss the production of the short - lived radioactive isotopes @xmath144cl , @xmath145ca , @xmath146pd and @xmath147pb , whose traces have been found in the ess and whose origin can be ascribed to an agb star . concerning @xmath145ca , an equilibrium value ( @xmath145ca@xmath153ca@xmath154 )
is rapidly attained whenever a neutron source is activated : this value is inversely proportional to the ratio of the corresponding neutron capture cross sections . @xmath146pd and @xmath147pb are mainly synthesized during the standard radiative @xmath3c burning at any metallicity .
the variations of the total yields of these two short - lived nuclei with the metallicity mainly reflect the increase of the neutron exposure when the metallicity decreases ( because the number of neutrons per iron seed increases ) .
this is only partially counterbalanced by the larger dredged up mass . as a general remark ,
since the decay rates depend on the environmental conditions ( temperature and electron density ) , the use of stellar models , which follow the temporal evolution of these conditions in detail and in all the layers where radioactive isotopes are stored after their production , is mandatory .
such a warning should be seriously considered , particularly , when these isotopes are engulfed in convective zones where the temperature and the density vary considerably from the the bottom to the top . as an example
, @xmath147pb has a rather long terrestrial half - life ( 1.5@xmath155 yr ) , but at the temperature of the he intershell of an agb star its half - life is several order of magnitude shorter @xcite .
the faster decays occurring in radiative conditions during the time elapsed between the thermal pulse and the tdu ( about @xmath156 yr ) significantly affects the resulting surface abundances of @xmath147pb ( see also * ? ? ?
+ in table [ tab15 ] , we list the final surface abundances ( by mass fraction ) of the aforementioned shot - lived radioactive isotopes ( with the corresponding stable isotopes ) , at different metallicities . in
[ varied ] we have shown how the physical parameters of the agb evolutions depends on the choices of the mass loss rate and the mixing length efficiency . in this section
we discuss the corresponding effects on the nucleosynthesis .
the results are illustrated in fig .
[ fig15 ] and table [ tab16 ] .
the final overabundances ( with respect to iron ) of the reimers model are generally larger than those found in the reference model .
this is a consequence of the larger duration of the agb , as obtained when the mass loss rate is lower , so that the total dredged up mass is larger . on the average , the abundances of the s - elements increase by a factor of two , while for the light elements we found a variation in the range between + 0.2 dex ( carbon ) and + 0.5 dex ( magnesium ) . concerning the fluorine production , we note that the delayed end of the agb phase favors the second fluorine source even in low @xmath2 models ( see [ pippo ] ) .
the primary @xmath3c in the ashes of the h burning becomes , in the late tp - agb phase of low @xmath2 models , as large as the @xmath3c found in the reference model with solar metallicity .
as explained before , neutrons released by the burning of this @xmath3c at the beginning of each tp provide an additional channel for the production of @xmath91n . as a result , a longer agb phase , as obtained by reducing the mass loss rate , favors this fluorine source , at any @xmath2 .
although the absolute abundances depend on the mass loss rate , the abundance ratios are less sensitive to the agb duration .
it occurs because the bulk of the s process nucleosynthesis has to be ascribed to the first @xmath3c pockets , the largest ones , so that a freeze out of the abundance ratios takes place after very few tps in the he intershell material ( see [ sproc ] ) . for this reason , the [ hs / ls ] and [ pb / hs ] obtained in the case of the reimers model are very similar to those of the reference model .
interestingly , the final c / n and @xmath26c/@xmath3c isotopic ratios of the reimers model decrease with respect to the reference one ( where monotonic trends are found ) , even if hbb is not at work .
this different behavior , which appears in the late agb phase , is due to an increase with the core mass of the temperature at the bottom of the convective envelope during tdu episodes , which leads to a partial h - burning . during these phases , therefore , mixing and burning simultaneously occur .
this phenomenon , already found by @xcite in a model with initial mass @xmath157= 3 m@xmath21 and z=0.0001 , enhances the @xmath3c and the @xmath36n in the envelope more rapidly than @xmath26c ( which , in any case , increases after each tdu episode ) . when applying a velocity profile at the base of the convective envelope , in fact ,
protons are mixed to higher temperatures with respect to the ones attained when using the bare schwarzschild criterion .
a more detailed analysis of this phenomenon , based on models with different initial masses , will be presented in a forthcoming paper .
finally , we evaluated the effects of varying the mixing efficiency by reducing the mixing - length parameter ( case @xmath158 ) . in that case , the lower cumulative dredged up mass leads to smaller overabundances .
as in the reimers model , the elemental ratios , which are sensitive to the metallicity and to the @xmath3c mass in the pocket , are less affected by the structural model change .
this paper reports the first systematic calculation of low mass agb models at different metallicities , in which the physical evolution of the star is coupled to a full nuclear network , from h to bi .
the major input physics , such as nuclear reaction rates , radiative opacity , mass loss rate , have been revised , in order to provide a reliable set of theoretical stellar yields .
the [ hs / ls ] and the [ pb / hs ] ratios at different @xmath2 provide important hints about the nucleosynthesis occurring at different metallicities and represent useful tools to verify the goodness of our theoretical models , when compared with observational data .
as shown in table [ tab12 ] , the [ hs / ls ] is rather low at large @xmath2 , because the low number of neutron per seed limits the production of heavy s elements .
when decreasing the metallicity , this ratio increases , achieving a sort of saturation for @xmath159 0.001 . below this threshold , the scarcity of seeds ( essentially fe ) is partially compensated by a reduction of the mass of the @xmath3c pocket ( see below ) .
[ pb / hs ] provides a more sensitive spectroscopic index to test low metallicity models .
it is expected to monotonically increase from high to low @xmath2 , indeed .
let us stress that , since the [ hs / ls ] and the [ pb / fe ] are practically frozen after a few dredge up episodes , these indexes are almost independent of the assumed mass loss rate and on the efficiency of the tdu , which are two of the most uncertain quantities in agb modelling .
in practice , as noted by @xcite ( see also * ? ? ? * ; * ? ? ? * ) , they essentially depend on the effective @xmath3c mass within the @xmath3c pockets . in this context , by comparing the theoretical predictions of these indexes with spectroscopic ratios at different @xmath2 , we may have a direct check of the validity of our choice of the @xmath35 parameter . in fig .
[ fig16 ] we have reported the observed [ hs / ls ] indexes , as measured in a sample of galactic and extragalactic c - stars with different metallicities @xcite .
spectroscopic and photometric studies indicate that these stars are intrinsic ( n type ) c - stars , _ i.e. _ low - mass agb stars undergoing the third dredge up .
the agreement with our theoretical predictions is comfortable .
in particular , data confirm the increase of [ hs / ls ] , when the metallicity is reduced below the solar value , up to a plateau attained at intermediate @xmath2 .
only the most metal - poor c(n ) star of the sample , namely alw - c7 , a carbon star of the carina dwarf galaxy , shows a too large [ hs / ls ] compared with the low @xmath2 plateau .
however , as suggested by @xcite , this star shows a particularly low abundance of zr compared with other light s elements .
moreover , the fitting process in determining the ls abundances in alw - c7 resulted more difficult with respect to alw - c6 , making the determination of the [ hs / ls ] even more uncertain ( c. abia , personal communication ) . a second interesting check for the new theoretical scenario concerns the comparison with nucleosynthesis models based on post - processes calculations @xcite . in the post - process calculations ,
the relevant stellar parameters are derived according to older stellar evolutionary models @xcite and , where models were not available , by using the interpolation formulas provided by @xcite . since these models have been computed by assuming a reimers mass loss rate , the duration of the agb and , in turn
, the total mass of h - depleted and s - enriched material dredged up is larger than that found in the present computations .
although this difference affects the predicted overabundance of a single element , it has negligible effects on the [ hs / ls ] or the [ pb / hs ] indexes .
a second important difference of the post process calculations concerns the mass of the @xmath3c pocket , which is fixed to a constant value for the full agb evolution ; according to @xcite , the value of the @xmath3c mass is a free parameter of the nucleosynthesis model and the st ( standard ) case corresponds to @xmath160c@xmath161 m@xmath21 ( see the discussion in [ beta ] ) .
the third important difference concerns the chemical profiles of @xmath3c and @xmath36n within the pocket ( for the post process calculation see fig . 1 of @xcite ) .
[ fig17 ] , the various lines represent the results of the post - process calculations @xcite , whilst the filled squares are our new predictions . in the plot
, st refers to the standard case , whilst the cases labelled with st*k or st / k correspond to @xmath3c pockets whose mass is k and 1/k times the standard one , respectively . at large metallicities , our predictions for [ hs / ls ] and
[ pb / hs ] are in good agreement with those of the st case ( dotted lines ) .
when the metallicity is reduced , however , the two spectroscopic indexes resulting from the new calculation move progressively toward the lines corresponding to smaller @xmath3c masses .
this behavior is likely due to the smaller @xmath3c pockets we found at low @xmath2 ( see fig . [ fig5 ] and tab .
[ tab6 ] ) , whose effect is to limit the maximum neutron exposure . let us conclude with a comment on the question of the spread of the @xmath3c pocket .
as discussed by @xcite , a certain spread is required to explain the variety of [ hs / ls ] observed in the available sample of carbon - enhanced metal - poor and s - rich stars ( cemps ) , whose overabundance of heavy elements is supposed to be originating from mass transfer from an ancient agb companion ( extrinsic c - stars ) .
similar conclusions arise from the abundance analysis of barium stars , which may be considered the cemps homologous at intermediate @xmath2 @xcite . by limiting our consideration to the galactic ( n type ) carbon stars , those belonging to the disk of the milky way and having a nearly - solar metallicity , for which a quite large sample of [ hs / ls ] is now available
, we note that the dispersion around the average value , namely @xmath162[hs / ls]@xmath163 , is of the same order of magnitude of the error bar of a single measurement ( @xmath164 ) .
in addition to that , it should be noted that the prediction of our solar metallicity model , namely [ hs / ls]=@xmath117 , is in excellent agreement with the measured average value .
it should be recalled that the present models refer to only one agb mass ( 2 m@xmath21 ) .
the natural variation of the stellar parameters among stars with similar metallicity , as for example the mass or the initial he , c , n or o , may account for a certain spread in the observed [ hs / ls ] and [ pb / hs ] ratios .
in the next future , we will extend the theoretical database to investigate such a possibility .
gallino , r. , bisterzo , s. , husti , l. , kppeler , f. , cristallo , s. , & straniero , o. , 2006 , proceedings of the `` _ international symposium on nuclear astrophysics - nuclei in the cosmos - ix _ '' , proceedings of science , pos 100 ccccccc + 0 & 0.00e+00 & & & & & + 0.01 & 0.00e+00 & & & & & + 0.05 & 3.00e-04 & 0.04 & 9.40e-05 & 5.62e+08 & 4.37e-05 & 8.32e-07 + 0.075 & 1.15e-03 & 0.14 & 3.47e-04 & 8.82e+08 & 8.49e-05 & 3.12e-06 + 0.09 & 1.70e-03 & 0.20 & 7.66e-04 & 1.06e+09 & 9.56e-05 & 6.41e-06 + 0.10 & 2.50e-03 & 0.28 & 5.54e-04 & 1.22e+09 & 1.18e-04 & 7.38e-06 + 0.11 & 2.90e-03 & 0.32 & 2.54e-04 & 1.36e+09 & 1.29e-04 & 4.40e-06 + 0.125 & 4.10e-03 & 0.43 & 2.00e-05 & 1.61e+09 & 1.76e-04 & 8.80e-07 + 0.15 & 5.80e-03 & 0.58 & 1.00e-05 & 2.00e+09 & 2.04e-04 & 1.42e-07 + 0.20 & 7.90e-03 & 0.75 & 0.00e+00 & 2.73e+09 & 1.96e-04 & 0.00e+00 + + 0 & 1.90e-03 & 0.29 & & & & + 0.01 & 2.80e-03 & 0.39 & 7.00e-06 & 5.16e+07 & 2.01e-05 & 5.55e-08 + 0.05 & 3.32e-03 & 0.45 & 6.50e-05 & 3.10e+08 & 2.35e-05 & 6.07e-07 + 0.075 & 3.63e-03 & 0.45 & 3.14e-04 & 5.28e+08 & 4.01e-05 & 2.31e-06 + 0.09 & 3.81e-03 & 0.45 & 6.10e-04 & 5.44e+08 & 5.48e-05 & 4.65e-06 + 0.10 & 3.90e-03 & 0.51 & 5.44e-04 & 6.19e+08 & 6.63e-05 & 5.48e-06 + 0.125 & 4.59e-03 & 0.57 & 8.40e-05 & 1.05e+09 & 1.49e-04 & 5.61e-07 + 0.15 & 5.15e-03 & 0.65 & 0.00e+00 & 1.37e+09 & 1.72e-04 & 0.00e+00 + 0.20 & 6.12e-03 & 0.75 & 0.00e+00 & 1.72e+09 & 2.07e-04 & 0.00e+00 + [ tab2 ] ccccccccccccc -2 & 1.91e+00 & 5.48e-01 & 0.00e+00 & 3.25e-02 & 4.20e-03 & 8.99e-01 & 0.00e+00 & & 14.0 & 2.20 & 1.39e-02 & 0.31 + -1 & 1.90e+00 & 5.53e-01 & 0.00e+00 & 3.07e-02 & 4.80e-03 & 8.66e-01 & 0.00e+00 & & 16.5 & 2.27 & 1.39e-02 & 0.31 + 1 & 1.89e+00 & 5.60e-01 & 2.00e-04 & 3.12e-02 & 7.00e-03 & 7.90e-01 & 2.90e-02 & 4.2@xmath165 & 18.6 & 2.46 & 1.39e-02 & 0.31 + 2 & 1.88e+00 & 5.68e-01 & 1.20e-03 & 2.94e-02 & 7.90e-03 & 7.48e-01 & 1.52e-01 & 4.6@xmath165 & 18.6 & 2.53 & 1.41e-02 & 0.34 + 3 & 1.86e+00 & 5.76e-01 & 2.50e-03 & 2.81e-02 & 9.00e-03 & 6.95e-01 & 2.78e-01 & 4.3@xmath165 & 17.8 & 2.62 & 1.46e-02 & 0.43 + 4 & 1.84e+00 & 5.83e-01 & 3.70e-03 & 2.69e-02 & 9.90e-03 & 6.46e-01 & 3.74e-01 & 4.3@xmath165 & 17.0 & 2.69 & 1.53e-02 & 0.57 + 5 & 1.82e+00 & 5.90e-01 & 4.50e-03 & 2.58e-02 & 1.07e-02 & 6.00e-01 & 4.21e-01 & 3.9@xmath165 & 16.2 & 2.75 & 1.62e-02 & 0.75 + 6 & 1.78e+00 & 5.96e-01 & 4.90e-03 & 2.48e-02 & 1.10e-02 & 5.67e-01 & 4.45e-01 & 3.6@xmath165 & 15.3 & 2.80 & 1.72e-02 & 0.95 + 7 & 1.72e+00 & 6.03e-01 & 4.90e-03 & 2.37e-02 & 1.11e-02 & 5.40e-01 & 4.41e-01 & 3.2@xmath165 & 14.3 & 2.84 & 1.82e-02 & 1.15 + 8 & 1.63e+00 & 6.09e-01 & 4.80e-03 & 2.25e-02 & 1.10e-02 & 5.23e-01 & 4.36e-01 & 3.2@xmath165 & 13.2 & 2.87 & 1.92e-02 & 1.35 + 9 & 1.51e+00 & 6.15e-01 & 4.40e-03 & 2.14e-02 & 1.07e-02 & 5.09e-01 & 4.11e-01 & 3.0@xmath165 & 12.1 & 2.90 & 2.03e-02 & 1.57 + 10 & 1.35e+00 & 6.21e-01 & 3.80e-03 & 2.03e-02 & 1.03e-02 & 5.02e-01 & 3.69e-01 & 2.9@xmath165 & 11.0 & 2.91 & 2.14e-02 & 1.78 + 11 & 1.16e+00 & 6.26e-01 & 1.25e-03 & 1.91e-02 & 9.70e-03 & 5.00e-01 & 1.29e-01 & 2.4@xmath165 & 9.91 & 2.92 & 2.19e-02 & 1.88 + 12 & 9.80e-01 & 6.33e-01 & 0.00e+00 & 1.71e-02 & 8.05e-03 & 5.31e-01 & 0.00e+00 & 2.2@xmath165 & 8.30 & 2.87 & 2.19e-02 & 1.88 + & * 3.62e-02 * & & & & & & & & & + + -2 & 1.97e+00 & 5.55e-01 & 0.00e+00 & 3.32e-02 & 4.70e-03 & 8.82e-01 & 0.00e+00 & & 1 5.9 & 2.27 & 6.07e-03 & 0.29 + -1 & 1.97e+00 & 5.61e-01 & 0.00e+00 & 3.20e-02 & 6.30e-03 & 8.21e-01 & 0.00e+00 & & 2 0.3 & 2.42 & 6.07e-03 & 0.29 + 1 & 1.96e+00 & 5.69e-01 & 1.60e-03 & 3.06e-02 & 7.90e-03 & 7.58e-01 & 2.03e-01 & 8.6@xmath165 & 2 0.9 & 2.55 & 6.29e-03 & 0.40 + 2 & 1.95e+00 & 5.77e-01 & 3.40e-03 & 2.91e-02 & 9.50e-03 & 6.87e-01 & 3.58e-01 & 8.6@xmath165 & 2 0.4 & 2.65 & 6.91e-03 & 0.69 + 3 & 1.93e+00 & 5.85e-01 & 5.30e-03 & 2.80e-02 & 1.09e-02 & 6.25e-01 & 4.86e-01 & 8.5@xmath165 & 1 9.6 & 2.74 & 7.91e-03 & 1.15 + 4 & 1.91e+00 & 5.91e-01 & 6.70e-03 & 2.70e-02 & 1.20e-02 & 5.68e-01 & 5.58e-01 & 8.3@xmath165 & 1 8.8 & 2.82 & 9.13e-03 & 1.70 + 5 & 1.89e+00 & 5.97e-01 & 7.30e-03 & 2.60e-02 & 1.28e-02 & 5.19e-01 & 5.70e-01 & 8.3@xmath165 & 17.7 & 2.87 & 1.04e-02 & 2.27 + 6 & 1.86e+00 & 6.03e-01 & 7.60e-03 & 2.50e-02 & 1.29e-02 & 4.93e-01 & 5.89e-01 & 8.0@xmath165 & 16.5 & 2.90 & 1.18e-02 & 2.85 + 7 & 1.81e+00 & 6.08e-01 & 7.60e-03 & 2.40e-02 & 1.29e-02 & 4.71e-01 & 5.89e-01 & 7.8@xmath165 & 15.3 & 2.93 & 1.31e-02 & 3.42 + 8 & 1.75e+00 & 6.13e-01 & 7.40e-03 & 2.30e-02 & 1.27e-02 & 4.59e-01 & 5.83e-01 & 7.5@xmath165 & 14.1 & 2.96 & 1.45e-02 & 3.99 + 9 & 1.66e+00 & 6.18e-01 & 7.10e-03 & 2.20e-02 & 1.22e-02 & 4.58e-01 & 5.82e-01 & 7.2@xmath165 & 12.9 & 2.99 & 1.58e-02 & 4.56 + 10 & 1.55e+00 & 6.23e-01 & 6.50e-03 & 2.10e-02 & 1.18e-02 & 4.45e-01 & 5.51e-01 & 7.1@xmath165 & 11.9 & 3.00 & 1.72e-02 & 5.13 + 11 & 1.40e+00 & 6.28e-01 & 5.50e-03 & 1.98e-02 & 1.14e-02 & 4.34e-01 & 4.82e-01 & 6.8@xmath165 & 10.8 & 3.00 & 1.87e-02 & 5.71 + 12 & 1.21e+00 & 6.33e-01 & 4.50e-03 & 1.87e-02 & 1.06e-02 & 4.45e-01 & 4.25e-01 & 6.6@xmath165 & 9.68 & 3.00 & 2.02e-02 & 6.35 + 13 & 1.00e+00 & 6.38e-01 & 1.50e-03 & 1.75e-02 & 9.60e-03 & 4.60e-01 & 1.56e-01 & 6.6@xmath165 & 8.47 & 2.99 & 2.11e-02 & 6.67 + 14 & 8.27e-01 & 6.44e-01 & 7.00e-04 & 1.54e-02 & 7.80e-03 & 5.04e-01 & 9.00e-02 & 5.3@xmath165 & 6.91 & 2.92 & 2.17e-02 & 6.91 + & * 7.27e-02 * & & & & & & & & & + + -2 & 1.98e+00 & 5.63e-01 & 0.00e+00 & 2.92e-02 & 2.30e-03 & 9.62e-01 & 0.00e+00 & & 10.4 & 2.09 & 3.04e-03 & 0.28 + -1 & 1.98e+00 & 5.69e-01 & 0.00e+00 & 3.16e-02 & 5.80e-03 & 8.33e-01 & 0.00e+00 & & 17.7 & 2.41 & 3.04e-03 & 0.28 + 1 & 1.97e+00 & 5.76e-01 & 1.30e-03 & 2.97e-02 & 7.20e-03 & 7.67e-01 & 1.81e-01 & 1.6@xmath166 & 20.1 & 2.54 & 3.21e-03 & 0.45 + 2 & 1.96e+00 & 5.84e-01 & 3.60e-03 & 2.82e-02 & 9.10e-03 & 6.86e-01 & 3.96e-01 & 1.7@xmath166 & 19.8 & 2.66 & 3.88e-03 & 1.09 + 3 & 1.94e+00 & 5.91e-01 & 5.20e-03 & 2.72e-02 & 1.06e-02 & 6.13e-01 & 4.91e-01 & 1.7@xmath166 & 19.1 & 2.77 & 4.92e-03 & 2.02 + 4 & 1.92e+00 & 5.98e-01 & 6.80e-03 & 2.62e-02 & 1.19e-02 & 5.56e-01 & 5.71e-01 & 1.6@xmath166 & 18.2 & 2.85 & 6.17e-03 & 3.09 + 5 & 1.90e+00 & 6.04e-01 & 7.40e-03 & 2.53e-02 & 1.25e-02 & 5.07e-01 & 5.92e-01 & 1.6@xmath166 & 17.1 & 2.89 & 7.51e-03 & 4.18 + 6 & 1.88e+00 & 6.09e-01 & 7.80e-03 & 2.43e-02 & 1.29e-02 & 4.72e-01 & 6.05e-01 & 1.4@xmath166 & 15.9 & 2.94 & 8.87e-03 & 5.25 + 7 & 1.85e+00 & 6.14e-01 & 8.00e-03 & 2.34e-02 & 1.30e-02 & 4.52e-01 & 6.15e-01 & 1.4@xmath166 & 14.7 & 2.97 & 1.02e-02 & 6.29 + 8 & 1.82e+00 & 6.19e-01 & 7.90e-03 & 2.24e-02 & 1.29e-02 & 4.37e-01 & 6.12e-01 & 1.4@xmath166 & 13.6 & 3.00 & 1.16e-02 & 7.30 + 9 & 1.77e+00 & 6.24e-01 & 7.50e-03 & 2.15e-02 & 1.23e-02 & 4.30e-01 & 6.10e-01 & 1.4@xmath166 & 12.5 & 3.03 & 1.29e-02 & 8.28 + 10 & 1.71e+00 & 6.28e-01 & 7.30e-03 & 2.06e-02 & 1.20e-02 & 4.23e-01 & 6.08e-01 & 1.3@xmath166 & 11.6 & 3.04 & 1.43e-02 & 9.24 + 11 & 1.62e+00 & 6.33e-01 & 7.00e-03 & 1.98e-02 & 1.18e-02 & 4.14e-01 & 5.93e-01 & 1.3@xmath166 & 10.7 & 3.05 & 1.57e-02 & 10.2 + 12 & 1.52e+00 & 6.37e-01 & 6.40e-03 & 1.89e-02 & 1.13e-02 & 4.12e-01 & 5.66e-01 & 1.3@xmath166 & 9.86 & 3.05 & 1.71e-02 & 11.2 + 13 & 1.38e+00 & 6.41e-01 & 5.40e-03 & 1.80e-02 & 1.07e-02 & 4.14e-01 & 5.05e-01 & 1.3@xmath166 & 9.05 & 3.06 & 1.86e-02 & 12.1 + 14 & 1.22e+00 & 6.46e-01 & 4.30e-03 & 1.70e-02 & 9.90e-03 & 4.22e-01 & 4.34e-01 & 1.3@xmath166 & 8.16 & 3.06 & 2.01e-02 & 13.1 + 15 & 1.03e+00 & 6.51e-01 & 3.00e-03 & 1.59e-02 & 9.20e-03 & 4.32e-01 & 3.26e-01 & 1.2@xmath166 & 7.29 & 3.04 & 2.17e-02 & 14.0 + 16 & 8.63e-01 & 6.56e-01 & 6.00e-04 & 1.46e-02 & 8.00e-03 & 4.58e-01 & 7.50e-02 & 1.2@xmath166 & 6.15 & 3.00 & 2.22e-02 & 14.4 + & * 8.95e-02 * & & & & & & & & & + + -1 & 1.94e+00 & 6.00e-01 & 0.00e+00 & 2.36e-02 & 2.20e-03 & 9.55e-01 & 0.00e+00 & & 6.48 & 2.13 & 1.01e-03 & 0.25 + 1 & 1.93e+00 & 6.04e-01 & 2.00e-04 & 2.53e-02 & 4.70e-03 & 8.23e-01 & 4.30e-02 & 4.3@xmath166 & 12.6 & 2.47 & 1.03e-03 & 0.31 + 2 & 1.92e+00 & 6.11e-01 & 3.00e-03 & 2.41e-02 & 6.70e-03 & 7.31e-01 & 4.48e-01 & 4.9@xmath166 & 15.1 & 2.64 & 1.56e-03 & 1.85 + 3 & 1.91e+00 & 6.17e-01 & 5.40e-03 & 2.33e-02 & 8.80e-03 & 6.27e-01 & 6.14e-01 & 4.6@xmath166 & 15.0 & 2.79 & 2.63e-03 & 4.56 + 4 & 1.89e+00 & 6.22e-01 & 7.20e-03 & 2.28e-02 & 1.06e-02 & 5.36e-01 & 6.79e-01 & 4.6@xmath166 & 14.7 & 2.90 & 4.05e-03 & 7.64 + 5 & 1.88e+00 & 6.27e-01 & 8.50e-03 & 2.23e-02 & 1.19e-02 & 4.72e-01 & 7.14e-01 & 4.6@xmath166 & 14.2 & 2.98 & 5.64e-03 & 10.6 + 6 & 1.85e+00 & 6.31e-01 & 9.30e-03 & 2.19e-02 & 1.26e-02 & 4.27e-01 & 7.38e-01 & 4.6@xmath166 & 13.6 & 3.03 & 7.31e-03 & 13.3 + 7 & 1.83e+00 & 6.35e-01 & 9.70e-03 & 2.14e-02 & 1.31e-02 & 3.94e-01 & 7.40e-01 & 4.5@xmath166 & 13.0 & 3.08 & 9.01e-03 & 15.9 + 8 & 1.81e+00 & 6.38e-01 & 9.70e-03 & 2.09e-02 & 1.31e-02 & 3.78e-01 & 7.40e-01 & 4.2@xmath166 & 12.2 & 3.11 & 1.07e-02 & 18.3 + 9 & 1.78e+00 & 6.41e-01 & 9.70e-03 & 2.03e-02 & 1.29e-02 & 3.70e-01 & 7.52e-01 & 3.9@xmath166 & 11.5 & 3.15 & 1.24e-02 & 20.5 + 10 & 1.75e+00 & 6.44e-01 & 9.50e-03 & 1.97e-02 & 1.27e-02 & 3.58e-01 & 7.48e-01 & 3.9@xmath166 & 10.8 & 3.16 & 1.41e-02 & 22.5 + 11 & 1.71e+00 & 6.47e-01 & 9.20e-03 & 1.91e-02 & 1.25e-02 & 3.51e-01 & 7.36e-01 & 3.8@xmath166 & 10.1 & 3.18 & 1.58e-02 & 24.4 + 12 & 1.66e+00 & 6.50e-01 & 8.80e-03 & 1.85e-02 & 1.20e-02 & 3.51e-01 & 7.33e-01 & 3.8@xmath166 & 9.46 & 3.18 & 1.74e-02 & 26.2 + 13 & 1.61e+00 & 6.53e-01 & 8.40e-03 & 1.79e-02 & 1.17e-02 & 3.48e-01 & 7.18e-01 & 3.8@xmath166 & 8.83 & 3.18 & 1.91e-02 & 27.8 + 14 & 1.53e+00 & 6.56e-01 & 8.00e-03 & 1.72e-02 & 1.13e-02 & 3.50e-01 & 7.08e-01 & 3.8@xmath166 & 8.24 & 3.19 & 2.08e-02 & 29.4 + 15 & 1.45e+00 & 6.59e-01 & 7.30e-03 & 1.66e-02 & 1.08e-02 & 3.50e-01 & 6.76e-01 & 3.9@xmath166 & 7.65 & 3.18 & 2.26e-02 & 30.9 + 16 & 1.34e+00 & 6.61e-01 & 6.40e-03 & 1.58e-02 & 1.02e-02 & 3.56e-01 & 6.27e-01 & 3.9@xmath166 & 7.06 & 3.17 & 2.44e-02 & 32.3 + 17 & 1.22e+00 & 6.65e-01 & 5.30e-03 & 1.50e-02 & 9.50e-03 & 3.68e-01 & 5.58e-01 & 3.8@xmath166 & 6.46 & 3.17 & 2.62e-02 & 33.7 + 18 & 1.09e+00 & 6.68e-01 & 3.70e-03 & 1.41e-02 & 8.70e-03 & 3.82e-01 & 4.25e-01 & 3.8@xmath166 & 5.83 & 3.13 & 2.80e-02 & 34.9 + 19 & 9.64e-01 & 6.72e-01 & 2.00e-03 & 1.31e-02 & 7.50e-03 & 4.11e-01 & 2.67e-01 & 3.8@xmath166 & 5.14 & 3.10 & 2.96e-02 & 36.0 + & * 1.31e-01 * & & & & & & & & & + + -1 & 1.98e+00 & 6.34e-01 & 0.00e+00 & 1.83e-02 & 2.50e-03 & 9.19e-01 & 0.00e+00 & & 4.21 & 2.19 & 1.01e-04 & 0.25 + 1 & 1.98e+00 & 6.41e-01 & 1.80e-03 & 1.96e-02 & 4.30e-03 & 7.97e-01 & 4.19e-01 & 4.6@xmath167 & 9.50 & 2.55 & 3.36e-04 & 7.76 + 2 & 1.97e+00 & 6.45e-01 & 3.90e-03 & 1.90e-02 & 6.30e-03 & 6.78e-01 & 6.19e-01 & 5.1@xmath167 & 10.0 & 2.72 & 1.02e-03 & 20.4 + 3 & 1.96e+00 & 6.49e-01 & 5.40e-03 & 1.86e-02 & 8.00e-03 & 5.77e-01 & 6.75e-01 & 5.4@xmath167 & 9.75 & 2.85 & 2.06e-03 & 28.1 + 4 & 1.95e+00 & 6.53e-01 & 6.70e-03 & 1.83e-02 & 9.20e-03 & 5.08e-01 & 7.28e-01 & 4.9@xmath167 & 9.58 & 2.95 & 3.31e-03 & 32.1 + 5 & 1.93e+00 & 6.56e-01 & 7.60e-03 & 1.80e-02 & 1.01e-02 & 4.51e-01 & 7.52e-01 & 4.6@xmath167 & 9.35 & 3.03 & 4.67e-03 & 35.0 + 6 & 1.92e+00 & 6.59e-01 & 8.00e-03 & 1.77e-02 & 1.06e-02 & 4.09e-01 & 7.55e-01 & 4.4@xmath167 & 9.06 & 3.07 & 6.08e-03 & 37.5 + 7 & 1.90e+00 & 6.62e-01 & 8.20e-03 & 1.74e-02 & 1.08e-02 & 3.84e-01 & 7.59e-01 & 4.2@xmath167 & 8.74 & 3.12 & 7.52e-03 & 40.0 + 8 & 1.88e+00 & 6.65e-01 & 8.50e-03 & 1.70e-02 & 1.09e-02 & 3.67e-01 & 7.80e-01 & 3.9@xmath167 & 8.37 & 3.16 & 8.96e-03 & 42.2 + 9 & 1.85e+00 & 6.67e-01 & 8.60e-03 & 1.67e-02 & 1.11e-02 & 3.45e-01 & 7.75e-01 & 3.9@xmath167 & 8.01 & 3.19 & 1.04e-02 & 44.3 + 10 & 1.80e+00 & 6.70e-01 & 8.50e-03 & 1.63e-02 & 1.10e-02 & 3.32e-01 & 7.73e-01 & 3.9@xmath167 & 7.62 & 3.22 & 1.19e-02 & 46.2 + 11 & 1.73e+00 & 6.72e-01 & 8.10e-03 & 1.58e-02 & 1.09e-02 & 3.21e-01 & 7.43e-01 & 3.9@xmath167 & 7.22 & 3.23 & 1.33e-02 & 47.9 + 12 & 1.62e+00 & 6.75e-01 & 7.50e-03 & 1.53e-02 & 1.04e-02 & 3.28e-01 & 7.21e-01 & 3.7@xmath167 & 6.82 & 3.24 & 1.49e-02 & 49.5 + 13 & 1.46e+00 & 6.77e-01 & 6.90e-03 & 1.48e-02 & 1.00e-02 & 3.34e-01 & 6.90e-01 & 3.7@xmath167 & 6.39 & 3.24 & 1.65e-02 & 51.1 + 14 & 1.20e+00 & 6.80e-01 & 4.90e-03 & 1.41e-02 & 9.50e-03 & 3.38e-01 & 5.16e-01 & 3.2@xmath167 & 5.92 & 3.23 & 1.83e-02 & 52.5 + 15 & 9.25e-01 & 6.83e-01 & 7.50e-04 & 1.28e-02 & 8.15e-03 & 3.64e-01 & 9.20e-02 & 2.9@xmath167 & 5.11 & 3.17 & 1.90e-02 & 53.1 + & * 9.54e-02 * & & & & & & & & & + + + + + + [ tab3 ] ccccccccccccc -1 & 1.98e+00 & 6.34e-01 & 0.00e+00 & 1.83e-02 & 2.50e-03 & 9.19e-01 & 0.00e+00 & & 4.21 & 2.19 & 1.01e-04 & 0.25 + 1 & 1.98e+00 & 6.41e-01 & 1.80e-03 & 1.96e-02 & 4.30e-03 & 7.97e-01 & 4.19e-01 & 4.6@xmath167 & 9.50 & 2.55 & 3.36e-04 & 7.76 + 5 & 1.93e+00 & 6.56e-01 & 7.60e-03 & 1.80e-02 & 1.01e-02 & 4.51e-01 & 7.52e-01 & 4.0@xmath167 & 9.35 & 3.03 & 4.67e-03 & 35.0 + 9 & 1.84e+00 & 6.70e-01 & 8.30e-03 & 1.63e-02 & 1.06e-02 & 3.38e-01 & 7.83e-01 & 3.7@xmath167 & 7.60 & 3.22 & 1.18e-02 & 46.1 + 14 & 1.75e+00 & 6.79e-01 & 7.00e-03 & 1.46e-02 & 9.60e-03 & 3.26e-01 & 7.29e-01 & 3.7@xmath167 & 5.92 & 3.25 & 1.71e-02 & 51.5 + 18 & 1.66e+00 & 6.88e-01 & 5.60e-03 & 1.29e-02 & 8.30e-03 & 3.39e-01 & 6.75e-01 & 3.6@xmath167 & 4.79 & 3.25 & 2.18e-02 & 54.7 + 25 & 1.50e+00 & 7.07e-01 & 3.20e-03 & 1.02e-02 & 6.50e-03 & 3.70e-01 & 4.92e-01 & 3.5@xmath167 & 3.45 & 3.19 & 2.84e-02 & 57.7 + 32 & 1.35e+00 & 7.31e-01 & 3.00e-04 & 7.74e-03 & 4.30e-03 & 4.28e-01 & 7.00e-02 & 3.0@xmath167 & 2.19 & 3.08 & 3.14e-02 & 58.6 + & * 1.60e-01 * & + + + + + + [ tab4 ] ccccccccccccc -1 & 1.98e+00 & 6.35e-01 & 0.00e+00 & 1.81e-02 & 2.10e-03 & 9.52e-01 & 0.00e+00 & & 3.67 & 2.13 & 1.01e-04 & 0.25 + 1 & 1.98e+00 & 6.39e-01 & 9.00e-04 & 1.97e-02 & 4.20e-03 & 8.15e-01 & 2.14e-01 & 4.6@xmath167 & 8.94 & 2.50 & 1.63e-04 & 2.37 + 2 & 1.97e+00 & 6.44e-01 & 2.60e-03 & 1.88e-02 & 5.80e-03 & 7.08e-01 & 4.48e-01 & 4.7@xmath167 & 9.62 & 2.66 & 5.86e-04 & 13.0 + 3 & 1.96e+00 & 6.49e-01 & 3.95e-03 & 1.81e-02 & 7.10e-03 & 6.24e-01 & 5.56e-01 & 5.1@xmath167 & 8.70 & 2.77 & 1.30e-03 & 22.0 + 4 & 1.95e+00 & 6.53e-01 & 4.70e-03 & 1.76e-02 & 8.05e-03 & 5.54e-01 & 5.84e-01 & 4.9@xmath167 & 8.19 & 2.86 & 2.20e-03 & 27.3 + 5 & 1.94e+00 & 6.57e-01 & 5.40e-03 & 1.71e-02 & 8.60e-03 & 5.10e-01 & 6.28e-01 & 4.6@xmath167 & 7.80 & 2.92 & 3.21e-03 & 30.7 + 6 & 1.93e+00 & 6.60e-01 & 5.70e-03 & 1.66e-02 & 8.90e-03 & 4.72e-01 & 6.40e-01 & 4.6@xmath167 & 7.44 & 2.98 & 4.28e-03 & 33.3 + 7 & 1.92e+00 & 6.63e-01 & 5.90e-03 & 1.61e-02 & 9.10e-03 & 4.44e-01 & 6.48e-01 & 4.6@xmath167 & 7.09 & 3.02 & 5.38e-03 & 35.6 + 8 & 1.89e+00 & 6.67e-01 & 6.00e-03 & 1.57e-02 & 9.10e-03 & 4.28e-01 & 6.59e-01 & 4.5@xmath167 & 6.75 & 3.05 & 6.49e-03 & 37.7 + 9 & 1.86e+00 & 6.70e-01 & 5.90e-03 & 1.52e-02 & 9.10e-03 & 4.10e-01 & 6.48e-01 & 4.5@xmath167 & 6.41 & 3.08 & 7.61e-03 & 39.6 + 10 & 1.80e+00 & 6.73e-01 & 5.80e-03 & 1.48e-02 & 9.00e-03 & 4.03e-01 & 6.44e-01 & 4.5@xmath167 & 6.08 & 3.10 & 8.74e-03 & 41.4 + 11 & 1.71e+00 & 6.76e-01 & 5.50e-03 & 1.43e-02 & 8.80e-03 & 3.97e-01 & 6.25e-01 & 4.5@xmath167 & 5.75 & 3.12 & 9.89e-03 & 43.0 + 12 & 1.59e+00 & 6.79e-01 & 4.70e-03 & 1.38e-02 & 8.50e-03 & 3.92e-01 & 5.53e-01 & 3.9@xmath167 & 5.42 & 3.13 & 1.10e-02 & 44.5 + 13 & 1.41e+00 & 6.82e-01 & 3.10e-03 & 1.32e-02 & 8.00e-03 & 4.01e-01 & 3.87e-01 & 3.9@xmath167 & 5.04 & 3.12 & 1.15e-02 & 45.1 + 14 & 1.19e+00 & 6.86e-01 & 1.10e-03 & 1.21e-02 & 7.10e-03 & 4.21e-01 & 1.55e-01 & 3.5@xmath167 & 4.54 & 3.08 & 1.19e-02 & 45.6 + & * 6.13e-02 * & & & & & & & & & + + + + + + [ tab5 ] c|ccc|ccc & & + 1 & 0.5600 & 8.19e-06 & 5.96 & 0.6405 & 6.35e-06 & 5.79 + 2 & 0.5677 & 7.81e-06 & 6.04 & 0.6450 & 6.84e-06 & 5.44 + 3 & 0.5755 & 7.38e-06 & 5.54 & 0.6491 & 6.06e-06 & 4.58 + 4 & 0.5829 & 6.41e-06 & 5.40 & 0.6529 & 5.33e-06 & 4.45 + 5 & 0.5899 & 5.56e-06 & 5.09 & 0.6563 & 4.70e-06 & 4.35 + 6 & 0.5964 & 4.77e-06 & 4.33 & 0.6593 & 4.12e-06 & 3.96 + 7 & 0.6026 & 4.10e-06 & 4.01 & 0.6621 & 3.79e-06 & 3.79 + 8 & 0.6087 & 3.35e-06 & 3.35 & 0.6648 & 3.44e-06 & 3.74 + 9 & 0.6146 & 2.73e-06 & 2.81 & 0.6674 & 3.17e-06 & 3.29 + 10 & 0.6205 & 2.00e-06 & 1.99 & 0.6698 & 2.81e-06 & 2.82 + 11 & 0.6264 & 4.95e-07 & 0.71 & 0.6722 & 2.54e-06 & 2.60 + 12 & & & & 0.6745 & 2.27e-06 & 2.51 + 13 & & & & 0.6770 & 1.81e-06 & 1.88 + 14 & & & & 0.6796 & 1.13e-06 & 1.31 + 15 & & & & 0.6829 & 5.96e-07 & 0.73 + * tot .
* & & * 5.28e-05 * & & & * 5.50e-05 * & + + + + [ tab6 ] ccccc c & -2.00e-01 & 6.72e-02 & 3.65e-01 & 5.75e-01 + n & 3.72e-01 & 3.71e-01 & 3.68e-01 & 3.65e-01 + o & -5.16e-04 & 6.71e-04 & 2.43e-03 & 1.36e-03 + f & -7.81e-03 & 7.03e-02 & 2.72e-01 & 4.88e-01 + ne & -2.55e-03 & 2.77e-02 & 1.03e-01 & 2.02e-01 + na & 8.62e-02 & 9.87e-02 & 1.29e-01 & 1.72e-01 + mg & -1.92e-07 & 6.82e-04 & 3.59e-03 & 1.06e-02 + al & 1.72e-04 & 1.76e-03 & 5.68e-03 & 1.06e-02 + sr & 1.27e-06 & 4.13e-01 & 8.80e-01 & 1.11e+00 + y & -3.40e-07 & 4.05e-01 & 8.51e-01 & 1.08e+00 + zr & 1.75e-08 & 3.95e-01 & 8.51e-01 & 1.07e+00 + ba & 1.34e-06 & 3.95e-01 & 9.05e-01 & 1.10e+00 + la & 3.04e-06 & 3.42e-01 & 8.35e-01 & 1.02e+00 + ce & 1.27e-06 & 2.87e-01 & 7.66e-01 & 9.51e-01 + pr & 3.04e-07 & 1.72e-01 & 5.41e-01 & 7.14e-01 + nd & 1.85e-06 & 1.90e-01 & 5.95e-01 & 7.69e-01 + sm & 3.19e-07 & 9.80e-02 & 3.83e-01 & 5.30e-01 + eu & -1.83e-06 & 1.67e-02 & 8.93e-02 & 1.43e-01 + pb & -1.19e-07 & 5.36e-02 & 3.61e-01 & 5.43e-01 + [ tab7 ] c & -9.26e-02 & 6.70e-01 & 1.03e+00 & 1.17e+00 + n & 3.94e-01 & 3.90e-01 & 3.83e-01 & 3.97e-01 + o & -4.65e-04 & 1.50e-02 & 2.62e-02 & 2.90e-02 + f & -8.65e-03 & 3.87e-01 & 8.51e-01 & 1.06e+00 + ne & 4.94e-04 & 1.45e-01 & 4.09e-01 & 5.58e-01 + na & 1.14e-01 & 1.52e-01 & 2.53e-01 & 3.58e-01 + mg & 1.58e-04 & 6.57e-03 & 3.47e-02 & 6.77e-02 + al & 5.90e-04 & 8.78e-03 & 2.04e-02 & 2.72e-02 + sr & 4.65e-02 & 8.60e-01 & 1.22e+00 & 1.33e+00 + y & 5.73e-02 & 8.72e-01 & 1.24e+00 & 1.36e+00 + zr & 5.21e-02 & 9.18e-01 & 1.28e+00 & 1.39e+00 + ba & 1.06e-02 & 1.40e+00 & 1.78e+00 & 1.87e+00 + la & 5.61e-03 & 1.43e+00 & 1.79e+00 & 1.88e+00 + ce & 2.97e-03 & 1.48e+00 & 1.83e+00 & 1.91e+00 + pr & 1.96e-03 & 1.23e+00 & 1.60e+00 & 1.69e+00 + nd & 1.45e-03 & 1.34e+00 & 1.68e+00 & 1.76e+00 + sm & 6.49e-04 & 1.09e+00 & 1.41e+00 & 1.49e+00 + eu & -1.52e-04 & 4.73e-01 & 7.39e-01 & 8.04e-01 + pb & 3.05e-04 & 1.21e+00 & 1.54e+00 & 1.59e+00 + [ tab8 ] ccccc c & -3.51e-02 & 1.06e+00 & 1.38e+00 & 1.54e+00 + n & 4.16e-01 & 4.12e-01 & 4.10e-01 & 4.38e-01 + o & -1.42e-03 & 4.34e-02 & 7.14e-02 & 8.97e-02 + f & -1.81e-02 & 7.27e-01 & 1.19e+00 & 1.47e+00 + ne & -8.52e-04 & 2.97e-01 & 6.36e-01 & 8.81e-01 + na & 1.37e-01 & 2.05e-01 & 3.61e-01 & 5.70e-01 + mg & 2.71e-04 & 1.72e-02 & 7.47e-02 & 1.70e-01 + al & 8.66e-04 & 1.81e-02 & 3.21e-02 & 4.41e-02 + sr & 4.20e-02 & 8.31e-01 & 1.12e+00 & 1.26e+00 + y & 5.65e-02 & 8.66e-01 & 1.17e+00 & 1.33e+00 + zr & 6.29e-02 & 9.17e-01 & 1.22e+00 & 1.38e+00 + ba & 3.10e-02 & 1.51e+00 & 1.86e+00 & 2.01e+00 + la & 2.13e-02 & 1.56e+00 & 1.90e+00 & 2.05e+00 + ce & 1.34e-02 & 1.67e+00 & 2.01e+00 & 2.15e+00 + pr & 8.80e-03 & 1.46e+00 & 1.82e+00 & 1.98e+00 + nd & 8.21e-03 & 1.57e+00 & 1.90e+00 & 2.04e+00 + sm & 4.89e-03 & 1.34e+00 & 1.66e+00 & 1.80e+00 + eu & 6.43e-04 & 6.79e-01 & 9.67e-01 & 1.10e+00 + pb & 4.89e-04 & 2.02e+00 & 2.25e+00 & 2.32e+00 + [ tab9 ] ccccc c & -2.13e-01 & 1.66e+00 & 1.95e+00 & 2.18e+00 + n & 4.59e-01 & 4.59e-01 & 4.72e-01 & 5.56e-01 + o & -9.12e-03 & 1.59e-01 & 2.40e-01 & 3.22e-01 + f & -3.85e-02 & 1.40e+00 & 1.83e+00 & 2.22e+00 + ne & -5.32e-03 & 7.21e-01 & 1.19e+00 & 1.60e+00 + na & 1.67e-01 & 4.06e-01 & 7.69e-01 & 1.28e+00 + mg & 9.40e-05 & 8.36e-02 & 3.27e-01 & 7.19e-01 + al & 2.91e-03 & 7.65e-02 & 1.12e-01 & 1.79e-01 + sr & 6.38e-06 & 7.94e-01 & 1.02e+00 & 1.26e+00 + y & 2.56e-06 & 8.47e-01 & 1.11e+00 & 1.37e+00 + zr & 2.56e-06 & 8.73e-01 & 1.14e+00 & 1.41e+00 + ba & 5.06e-06 & 1.39e+00 & 1.70e+00 & 1.99e+00 + la & 2.56e-06 & 1.43e+00 & 1.74e+00 & 2.03e+00 + ce & 6.55e-07 & 1.54e+00 & 1.87e+00 & 2.15e+00 + pr & -2.57e-06 & 1.38e+00 & 1.73e+00 & 2.02e+00 + nd & 2.56e-06 & 1.44e+00 & 1.75e+00 & 2.04e+00 + sm & -4.37e-06 & 1.24e+00 & 1.57e+00 & 1.86e+00 + eu & -2.54e-05 & 6.10e-01 & 9.05e-01 & 1.18e+00 + pb & 2.56e-06 & 2.54e+00 & 2.76e+00 & 2.87e+00 + [ tab10 ] cccccc c & 1.15e+00 & 2.41e+00 & 2.81e+00 & 3.00e+00 + n & 4.93e-01 & 5.23e-01 & 6.17e-01 & 8.58e-01 + o & -9.15e-04 & 5.95e-01 & 8.76e-01 & 1.01e+00 + f & -8.31e-02 & 2.13e+00 & 2.62e+00 & 2.90e+00 + ne & -3.48e-03 & 1.30e+00 & 1.91e+00 & 2.27e+00 + na & 2.22e-01 & 7.70e-01 & 1.40e+00 & 1.77e+00 + mg & -2.92e-04 & 2.31e-01 & 8.59e-01 & 1.30e+00 + al & 7.02e-03 & 1.94e-01 & 4.37e-01 & 6.37e-01 + sr & 1.19e-04 & 5.17e-01 & 8.64e-01 & 1.01e+00 + y & 8.69e-06 & 5.66e-01 & 9.60e-01 & 1.12e+00 + zr & 9.89e-06 & 5.75e-01 & 9.92e-01 & 1.17e+00 + ba & 6.13e-05 & 9.21e-01 & 1.36e+00 & 1.55e+00 + la & -4.48e-06 & 9.37e-01 & 1.38e+00 & 1.57e+00 + ce & 2.41e-06 & 1.04e+00 & 1.49e+00 & 1.69e+00 + pr & -1.07e-04 & 9.01e-01 & 1.35e+00 & 1.55e+00 + nd & -4.98e-06 & 9.27e-01 & 1.37e+00 & 1.57e+00 + sm & -1.09e-04 & 7.40e-01 & 1.20e+00 & 1.41e+00 + eu & -4.44e-04 & 2.66e-01 & 5.87e-01 & 7.65e-01 + pb & 9.70e-06 & 2.37e+00 & 2.74e+00 & 2.88e+00 + [ tab11 ] cccccc @xmath168 $ ] & 0.57 & 1.17 & 1.54 & 2.17 & 3.00 + @xmath169 $ ] & 0.37 & 0.40 & 0.44 & 0.56 & 0.86 + @xmath170 $ ] & 0.00 & 0.03 & 0.09 & 0.32 & 1.01 + @xmath171 $ ] & 0.49 & 1.06 & 1.46 & 2.22 & 2.90 + @xmath172 $ ] & 0.20 & 0.56 & 0.88 & 1.60 & 2.27 + @xmath173 $ ] & 0.17 & 0.36 & 0.57 & 1.28 & 1.77 + @xmath174 $ ] & 0.01 & 0.07 & 0.17 & 0.72 & 1.30 + @xmath175 $ ] & 0.01 & 0.03 & 0.04 & 0.18 & 0.64 + @xmath176 $ ] & 1.08 & 1.36 & 1.32 & 1.34 & 1.10 + @xmath177 $ ] & 0.87 & 1.78 & 2.00 & 2.01 & 1.56 + @xmath178 $ ] & 0.54 & 1.59 & 2.32 & 2.87 & 2.88 + @xmath179 $ ] & -0.21 & 0.42 & 0.68 & 0.67 & 0.46 + c / o & 1.87 & 6.92 & 14.4 & 36.0 & 53.0 + c / n & 5.84 & 23.0 & 46.1 & 151.0 & 406.0 + @xmath26c/@xmath3c & 1.42@xmath166 & 5.75@xmath166 & 1.37@xmath167 & 5.20@xmath167 & 1.72@xmath180 + [ tab12 ] cccccccc @xmath122kr/@xmath181kr & 0.31 & 0.26 & 0.37 & 0.52 & 0.68 & 0.72 + @xmath126rb/@xmath121rb & 0.39 & 0.40 & 0.58 & 0.94 & 1.32 & 1.30 + @xmath182zr/@xmath183zr & 0.16 & 0.03 & 0.07 & 0.16 & 0.44 & 0.61 + @xmath184xe/@xmath185xe & 0.37 & 0.07 & 0.06 & 0.13 & 0.47 & 0.71 + @xmath186ce/@xmath187ce & 0.13 & 0.01 & 0.01 & 0.04 & 0.14 & 0.20 + [ tab13 ] cccccc @xmath188h & -4.06e-02 & -7.05e-02 & -8.21e-02 & -1.16e-01 & -9.92e-02 + @xmath189h & -3.77e-05 & -3.79e-05 & -3.86e-05 & -5.42e-05 & -5.39e-05 + @xmath190he & 3.77e-04 & 3.35e-04 & 3.03e-04 & 2.59e-04 & 1.83e-04 + @xmath191he & 3.23e-02 & 5.14e-02 & 6.31e-02 & 8.85e-02 & 7.99e-02 + @xmath26c & 5.84e-03 & 1.28e-02 & 1.64e-02 & 2.23e-02 & 1.71e-02 + @xmath3c & 5.16e-05 & 2.03e-05 & 1.01e-05 & 3.63e-06 & 6.97e-07 + @xmath36n & 1.31e-03 & 6.03e-04 & 3.45e-04 & 1.52e-04 & 3.40e-05 + @xmath91n & -2.03e-06 & -9.21e-07 & -4.88e-07 & -1.65e-07 & -2.07e-08 + @xmath4o & -1.04e-05 & 1.70e-04 & 2.98e-04 & 4.47e-04 & 4.03e-04 + @xmath192o & 3.32e-05 & 2.33e-05 & 1.68e-05 & 7.09e-06 & 9.33e-07 + @xmath38o & -5.65e-06 & -2.78e-06 & -1.53e-06 & -5.08e-07 & -3.20e-08 + @xmath133f & 8.17e-07 & 1.61e-06 & 2.23e-06 & 3.97e-06 & 2.44e-06 + @xmath193ne & -2.34e-06 & 1.33e-06 & 6.48e-06 & 2.80e-05 & 1.17e-05 + @xmath194ne & 8.15e-08 & 2.56e-07 & 3.94e-07 & 1.00e-06 & 5.54e-07 + @xmath14ne & 7.06e-04 & 1.21e-03 & 1.54e-03 & 2.69e-03 & 1.41e-03 + @xmath53na & 1.95e-05 & 1.77e-05 & 1.77e-05 & 3.25e-05 & 1.38e-05 + @xmath195 mg & 9.94e-06 & 2.40e-05 & 3.32e-05 & 7.33e-05 & 2.64e-05 + @xmath16 mg & 7.10e-07 & 6.21e-06 & 1.28e-05 & 4.36e-05 & 2.55e-05 + @xmath143 mg & 3.16e-06 & 7.11e-06 & 8.17e-06 & 3.14e-05 & 3.21e-05 + @xmath143al & 3.06e-07 & 1.24e-07 & 6.62e-08 & 5.40e-08 & 5.33e-08 + @xmath144cl & 1.38e-09 & 1.42e-09 & 9.66e-10 & 4.34e-10 & 3.95e-11 + @xmath145ca & 3.90e-09 & 3.38e-09 & 2.05e-09 & 8.58e-10 & 7.11e-11 + @xmath17fe & 5.76e-08 & 5.04e-08 & 6.35e-08 & 1.26e-07 & 2.04e-08 + @xmath196y & 1.18e-07 & 1.01e-07 & 4.79e-08 & 1.55e-08 & 1.08e-09 + @xmath146pd & 1.56e-09 & 1.70e-09 & 7.88e-10 & 1.99e-10 & 1.03e-11 + @xmath197la & 1.56e-08 & 5.44e-08 & 3.96e-08 & 1.08e-08 & 4.80e-10 + @xmath147pb & 1.48e-10 & 1.05e-09 & 1.54e-09 & 3.94e-10 & 1.30e-11 + @xmath119pb & 1.88e-08 & 1.51e-07 & 5.50e-07 & 8.48e-07 & 1.03e-07 + [ tab14 ] cccccc @xmath143al & 3.07e-07 & 1.41e-07 & 7.80e-08 & 7.37e-08 & 5.33e-08 + @xmath198al & 6.14e-05 & 2.77e-05 & 1.43e-05 & 6.43e-06 & 1.98e-06 + @xmath199cl & 3.72e-06 & 1.61e-06 & 7.97e-07 & 2.59e-07 & 2.78e-08 + @xmath144cl & 1.70e-09 & 1.51e-09 & 9.29e-10 & 5.35e-10 & 3.50e-11 + @xmath200ca & 6.26e-05 & 2.70e-05 & 1.34e-05 & 4.35e-06 & 4.78e-07 + @xmath145ca & 5.13e-09 & 3.52e-09 & 1.90e-09 & 1.11e-09 & 7.11e-11 + @xmath151fe & 1.20e-03 & 5.17e-04 & 2.56e-04 & 8.26e-05 & 8.35e-06 + @xmath17fe & 4.78e-08 & 3.93e-08 & 5.70e-08 & 1.38e-07 & 2.04e-08 + @xmath201pb & 2.42e-09 & 1.24e-08 & 1.58e-08 & 4.80e-09 & 1.50e-10 + @xmath147pb & 1.58e-10 & 1.04e-09 & 1.55e-09 & 4.61e-10 & 1.31e-11 + [ tab15 ] cccc @xmath168 $ ] & 3.00 & 3.20 & 2.81 + @xmath169 $ ] & 0.86 & 1.15 & 0.68 + @xmath170 $ ] & 1.01 & 1.17 & 0.88 + @xmath171 $ ] & 2.90 & 3.27 & 2.66 + @xmath172 $ ] & 2.27 & 2.68 & 1.90 + @xmath173 $ ] & 1.77 & 2.26 & 1.36 + @xmath174 $ ] & 1.30 & 1.82 & 0.84 + @xmath175 $ ] & 0.64 & 0.97 & 0.47 + @xmath176 $ ] & 1.10 & 1.30 & 0.85 + @xmath177 $ ] & 1.56 & 1.85 & 1.33 + @xmath178 $ ] & 2.88 & 3.00 & 2.78 + @xmath179 $ ] & 0.46 & 0.55 & 0.48 + c / o & 53 & 52 & 46 + c / n & 406 & 325 & 388 + @xmath26c/@xmath3c & 1.72@xmath202 & 1.35@xmath202 & 1.72@xmath202 + [ tab16 ] c & -2.00e-01 & -1.58e-01 & -5.55e-02 & 6.72e-02 & 1.84e-01 & 2.83e-01 & 3.65e-01 & 4.36e-01 & 4.98e-01 & 5.54e-01 & 5.75e-01 + n & 3.72e-01 & 3.72e-01 & 3.72e-01 & 3.71e-01 & 3.71e-01 & 3.69e-01 & 3.68e-01 & 3.67e-01 & 3.66e-01 & 3.65e-01 & 3.65e-01 + o & -5.16e-04 & -5.27e-04 & -9.26e-05 & 6.71e-04 & 1.50e-03 & 2.13e-03 & 2.43e-03 & 2.44e-03 & 2.20e-03 & 1.73e-03 & 1.36e-03 + f & -7.81e-03 & -2.14e-03 & 2.30e-02 & 7.03e-02 & 1.34e-01 & 2.04e-01 & 2.72e-01 & 3.38e-01 & 4.02e-01 & 4.63e-01 & 4.88e-01 + ne & -2.55e-03 & 4.12e-04 & 1.05e-02 & 2.77e-02 & 5.05e-02 & 7.63e-02 & 1.03e-01 & 1.31e-01 & 1.60e-01 & 1.90e-01 & 2.02e-01 + na & 8.62e-02 & 8.78e-02 & 9.19e-02 & 9.87e-02 & 1.08e-01 & 1.18e-01 & 1.29e-01 & 1.40e-01 & 1.53e-01 & 1.66e-01 & 1.72e-01 + mg & -1.92e-07 & 4.70e-05 & 2.60e-04 & 6.82e-04 & 1.36e-03 & 2.34e-03 & 3.59e-03 & 5.16e-03 & 7.10e-03 & 9.44e-03 & 1.06e-02 + al & 1.72e-04 & 4.57e-04 & 9.36e-04 & 1.76e-03 & 2.92e-03 & 4.26e-03 & 5.68e-03 & 7.10e-03 & 8.54e-03 & 9.99e-03 & 1.06e-02 + si & 4.52e-07 & 2.82e-05 & 1.34e-04 & 2.72e-04 & 4.19e-04 & 5.65e-04 & 7.01e-04 & 8.25e-04 & 9.51e-04 & 1.07e-03 & 1.12e-03 + p & 1.25e-06 & 4.24e-04 & 2.16e-03 & 4.94e-03 & 8.24e-03 & 1.16e-02 & 1.47e-02 & 1.76e-02 & 2.02e-02 & 2.26e-02 & 2.35e-02 + s & -1.24e-06 & 6.36e-06 & 4.25e-05 & 9.76e-05 & 1.64e-04 & 2.32e-04 & 2.98e-04 & 3.55e-04 & 4.13e-04 & 4.63e-04 & 4.82e-04 + cl & 6.89e-07 & 2.86e-04 & 1.30e-03 & 2.79e-03 & 4.49e-03 & 6.28e-03 & 8.08e-03 & 9.91e-03 & 1.17e-02 & 1.36e-02 & 1.42e-02 + ar & 1.18e-06 & -3.61e-05 & -1.07e-04 & -1.84e-04 & -2.49e-04 & -2.95e-04 & -3.38e-04 & -3.88e-04 & -4.43e-04 & -5.10e-04 & -5.42e-04 + k & -2.58e-04 & -1.87e-05 & 7.40e-04 & 1.77e-03 & 2.89e-03 & 4.04e-03 & 5.19e-03 & 6.36e-03 & 7.58e-03 & 8.81e-03 & 9.53e-03 + ca & 1.41e-05 & -3.27e-05 & -1.32e-04 & -2.34e-04 & -3.19e-04 & -3.75e-04 & -4.26e-04 & -4.82e-04 & -5.42e-04 & -6.16e-04 & -6.66e-04 + sc & 1.16e-06 & 1.80e-03 & 5.79e-03 & 1.07e-02 & 1.59e-02 & 2.11e-02 & 2.64e-02 & 3.17e-02 & 3.73e-02 & 4.30e-02 & 4.53e-02 + ti & 9.27e-07 & 6.62e-05 & 4.86e-04 & 1.29e-03 & 2.36e-03 & 3.54e-03 & 4.69e-03 & 5.77e-03 & 6.78e-03 & 7.70e-03 & 8.03e-03 + v & 8.69e-07 & -1.25e-04 & -3.21e-04 & -4.23e-04 & -3.43e-04 & -7.78e-05 & 3.16e-04 & 7.88e-04 & 1.32e-03 & 1.88e-03 & 2.11e-03 + cr & 7.69e-07 & -1.55e-06 & 8.29e-06 & 2.82e-05 & 5.44e-05 & 8.85e-05 & 1.20e-04 & 1.46e-04 & 1.75e-04 & 2.03e-04 & 2.10e-04 + mn & 1.96e-06 & -9.30e-05 & -1.94e-04 & -1.44e-04 & 1.56e-04 & 7.38e-04 & 1.56e-03 & 2.58e-03 & 3.79e-03 & 5.13e-03 & 5.72e-03 + fe & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 + co & 1.22e-06 & 3.07e-03 & 1.08e-02 & 1.92e-02 & 2.65e-02 & 3.25e-02 & 3.76e-02 & 4.24e-02 & 4.74e-02 & 5.26e-02 & 5.67e-02 + ni & 2.51e-06 & 9.49e-05 & 5.58e-04 & 1.12e-03 & 1.63e-03 & 2.05e-03 & 2.35e-03 & 2.57e-03 & 2.72e-03 & 2.81e-03 & 2.74e-03 + cu & 2.07e-06 & 4.25e-03 & 1.59e-02 & 2.88e-02 & 4.05e-02 & 5.03e-02 & 5.85e-02 & 6.59e-02 & 7.25e-02 & 7.88e-02 & 8.12e-02 + zn & 6.38e-07 & 5.21e-04 & 3.68e-03 & 8.82e-03 & 1.48e-02 & 2.09e-02 & 2.68e-02 & 3.24e-02 & 3.76e-02 & 4.24e-02 & 4.41e-02 + ga & 1.31e-06 & 3.53e-03 & 1.76e-02 & 4.14e-02 & 6.95e-02 & 9.87e-02 & 1.26e-01 & 1.52e-01 & 1.75e-01 & 1.96e-01 & 2.03e-01 + ge & 1.09e-06 & 3.47e-03 & 1.72e-02 & 4.04e-02 & 6.82e-02 & 9.74e-02 & 1.25e-01 & 1.52e-01 & 1.75e-01 & 1.97e-01 & 2.04e-01 + as & -2.59e-06 & 2.25e-03 & 1.13e-02 & 2.71e-02 & 4.64e-02 & 6.71e-02 & 8.73e-02 & 1.07e-01 & 1.24e-01 & 1.41e-01 & 1.46e-01 + se & 2.66e-07 & 5.27e-03 & 2.65e-02 & 6.01e-02 & 9.81e-02 & 1.37e-01 & 1.72e-01 & 2.05e-01 & 2.35e-01 & 2.60e-01 & 2.69e-01 + br & -1.53e-06 & 2.29e-03 & 1.27e-02 & 3.29e-02 & 6.01e-02 & 8.93e-02 & 1.18e-01 & 1.45e-01 & 1.71e-01 & 1.94e-01 & 2.07e-01 + kr & -1.24e-06 & 7.41e-03 & 3.71e-02 & 8.45e-02 & 1.38e-01 & 1.90e-01 & 2.38e-01 & 2.79e-01 & 3.17e-01 & 3.49e-01 & 3.60e-01 + rb & -7.85e-07 & 1.84e-02 & 7.11e-02 & 1.26e-01 & 1.75e-01 & 2.19e-01 & 2.57e-01 & 2.92e-01 & 3.23e-01 & 3.52e-01 & 3.62e-01 + sr & 1.27e-06 & 2.02e-02 & 1.81e-01 & 4.13e-01 & 6.13e-01 & 7.67e-01 & 8.80e-01 & 9.67e-01 & 1.04e+00 & 1.09e+00 & 1.11e+00 + y & -3.40e-07 & 2.18e-02 & 1.85e-01 & 4.05e-01 & 5.95e-01 & 7.42e-01 & 8.51e-01 & 9.35e-01 & 1.00e+00 & 1.06e+00 & 1.08e+00 + zr & 1.75e-08 & 1.45e-02 & 1.68e-01 & 3.95e-01 & 5.93e-01 & 7.42e-01 & 8.51e-01 & 9.33e-01 & 9.98e-01 & 1.05e+00 & 1.07e+00 + nb & -1.34e-06 & -1.78e-04 & 4.65e-04 & 1.87e-02 & 6.66e-02 & 1.36e-01 & 2.10e-01 & 2.83e-01 & 3.50e-01 & 4.09e-01 & 4.61e-01 + mo & 1.11e-06 & 4.02e-03 & 9.07e-02 & 2.59e-01 & 4.29e-01 & 5.65e-01 & 6.66e-01 & 7.42e-01 & 8.03e-01 & 8.51e-01 & 8.67e-01 + ru & 6.44e-07 & 1.56e-03 & 6.41e-02 & 1.98e-01 & 3.47e-01 & 4.71e-01 & 5.66e-01 & 6.40e-01 & 6.98e-01 & 7.45e-01 & 7.62e-01 + rh & -8.37e-07 & 5.96e-04 & 3.23e-02 & 1.06e-01 & 1.97e-01 & 2.80e-01 & 3.48e-01 & 4.03e-01 & 4.48e-01 & 4.84e-01 & 4.96e-01 + pd & 1.88e-06 & 2.32e-03 & 9.79e-02 & 2.82e-01 & 4.64e-01 & 6.04e-01 & 7.07e-01 & 7.84e-01 & 8.45e-01 & 8.93e-01 & 9.08e-01 + ag & -1.06e-06 & 4.42e-04 & 2.80e-02 & 9.42e-02 & 1.77e-01 & 2.54e-01 & 3.18e-01 & 3.69e-01 & 4.12e-01 & 4.47e-01 & 4.60e-01 + cd & 1.54e-06 & 3.05e-03 & 1.20e-01 & 3.28e-01 & 5.24e-01 & 6.72e-01 & 7.79e-01 & 8.59e-01 & 9.22e-01 & 9.71e-01 & 9.87e-01 + in & 3.78e-07 & 1.78e-03 & 7.84e-02 & 2.32e-01 & 3.90e-01 & 5.18e-01 & 6.14e-01 & 6.86e-01 & 7.44e-01 & 7.89e-01 & 8.04e-01 + sn & -1.82e-07 & 2.85e-03 & 1.24e-01 & 3.42e-01 & 5.42e-01 & 6.92e-01 & 7.99e-01 & 8.78e-01 & 9.39e-01 & 9.87e-01 & 1.00e+00 + sb & -3.84e-07 & 1.01e-03 & 5.25e-02 & 1.57e-01 & 2.75e-01 & 3.76e-01 & 4.55e-01 & 5.17e-01 & 5.66e-01 & 6.06e-01 & 6.19e-01 + te & -1.11e-06 & 4.15e-04 & 3.77e-02 & 1.32e-01 & 2.45e-01 & 3.43e-01 & 4.19e-01 & 4.76e-01 & 5.22e-01 & 5.58e-01 & 5.70e-01 + i & -7.01e-07 & -6.82e-05 & 1.04e-02 & 3.99e-02 & 8.14e-02 & 1.23e-01 & 1.60e-01 & 1.90e-01 & 2.15e-01 & 2.36e-01 & 2.43e-01 + xe & 6.06e-07 & 2.84e-04 & 3.68e-02 & 1.29e-01 & 2.40e-01 & 3.37e-01 & 4.11e-01 & 4.69e-01 & 5.14e-01 & 5.50e-01 & 5.61e-01 + cs & -8.75e-07 & 6.77e-04 & 7.06e-02 & 1.69e-01 & 2.75e-01 & 3.68e-01 & 4.44e-01 & 5.04e-01 & 5.54e-01 & 5.94e-01 & 6.06e-01 + ba & 1.34e-06 & 8.19e-04 & 1.26e-01 & 3.95e-01 & 6.30e-01 & 7.94e-01 & 9.05e-01 & 9.83e-01 & 1.04e+00 & 1.08e+00 & 1.10e+00 + la & 3.04e-06 & 3.47e-04 & 9.84e-02 & 3.42e-01 & 5.67e-01 & 7.27e-01 & 8.35e-01 & 9.11e-01 & 9.66e-01 & 1.01e+00 & 1.02e+00 + ce & 1.27e-06 & 1.65e-04 & 7.25e-02 & 2.87e-01 & 5.00e-01 & 6.58e-01 & 7.66e-01 & 8.42e-01 & 8.98e-01 & 9.38e-01 & 9.51e-01 + pr & 3.04e-07 & 4.76e-05 & 4.42e-02 & 1.72e-01 & 3.22e-01 & 4.47e-01 & 5.41e-01 & 6.10e-01 & 6.62e-01 & 7.02e-01 & 7.14e-01 + nd & 1.85e-06 & 4.00e-05 & 4.21e-02 & 1.90e-01 & 3.60e-01 & 4.97e-01 & 5.95e-01 & 6.66e-01 & 7.19e-01 & 7.58e-01 & 7.69e-01 + sm & 3.19e-07 & -4.58e-05 & 1.86e-02 & 9.80e-02 & 2.06e-01 & 3.05e-01 & 3.83e-01 & 4.42e-01 & 4.87e-01 & 5.20e-01 & 5.30e-01 + eu & -1.83e-06 & -1.98e-04 & 2.64e-03 & 1.67e-02 & 3.98e-02 & 6.56e-02 & 8.93e-02 & 1.09e-01 & 1.26e-01 & 1.39e-01 & 1.43e-01 + gd & 5.50e-07 & -7.72e-05 & 8.97e-03 & 5.04e-02 & 1.14e-01 & 1.75e-01 & 2.26e-01 & 2.65e-01 & 2.96e-01 & 3.20e-01 & 3.27e-01 + tb & -3.14e-07 & -1.69e-04 & 4.25e-03 & 2.80e-02 & 6.68e-02 & 1.08e-01 & 1.44e-01 & 1.72e-01 & 1.95e-01 & 2.12e-01 & 2.18e-01 + dy & -1.13e-05 & -1.17e-04 & 7.88e-03 & 4.95e-02 & 1.15e-01 & 1.80e-01 & 2.34e-01 & 2.77e-01 & 3.10e-01 & 3.34e-01 & 3.43e-01 + ho & 4.64e-05 & -6.19e-05 & 5.41e-03 & 3.39e-02 & 7.14e-02 & 1.07e-01 & 1.39e-01 & 1.61e-01 & 1.78e-01 & 1.92e-01 & 1.90e-01 + er & 7.22e-07 & -7.44e-05 & 8.42e-03 & 5.15e-02 & 1.19e-01 & 1.88e-01 & 2.46e-01 & 2.91e-01 & 3.25e-01 & 3.52e-01 & 3.60e-01 + tm & 2.01e-07 & -1.37e-04 & 6.03e-03 & 4.16e-02 & 1.01e-01 & 1.64e-01 & 2.18e-01 & 2.62e-01 & 2.96e-01 & 3.22e-01 & 3.30e-01 + yb & 9.04e-07 & 5.43e-05 & 2.01e-02 & 1.07e-01 & 2.26e-01 & 3.35e-01 & 4.20e-01 & 4.84e-01 & 5.33e-01 & 5.69e-01 & 5.80e-01 + lu & -1.14e-03 & -1.24e-03 & 8.79e-03 & 5.78e-02 & 1.34e-01 & 2.10e-01 & 2.74e-01 & 3.25e-01 & 3.65e-01 & 3.95e-01 & 4.04e-01 + hf & 2.39e-04 & 5.17e-04 & 2.68e-02 & 1.35e-01 & 2.77e-01 & 4.02e-01 & 4.97e-01 & 5.67e-01 & 6.20e-01 & 6.59e-01 & 6.71e-01 + ta & -1.45e-06 & 1.28e-04 & 1.90e-02 & 1.03e-01 & 2.24e-01 & 3.37e-01 & 4.26e-01 & 4.94e-01 & 5.46e-01 & 5.85e-01 & 5.96e-01 + w & 6.97e-08 & 4.22e-04 & 2.55e-02 & 1.35e-01 & 2.79e-01 & 4.07e-01 & 5.05e-01 & 5.77e-01 & 6.32e-01 & 6.72e-01 & 6.85e-01 + re & -6.54e-02 & -6.54e-02 & -5.75e-02 & -2.72e-02 & 2.65e-02 & 8.61e-02 & 1.40e-01 & 1.86e-01 & 2.23e-01 & 2.53e-01 & 2.62e-01 + os & 4.68e-03 & 4.60e-03 & 7.73e-03 & 2.71e-02 & 6.02e-02 & 9.70e-02 & 1.30e-01 & 1.57e-01 & 1.79e-01 & 1.96e-01 & 2.02e-01 + ir & -2.11e-06 & -1.92e-04 & -1.67e-04 & 2.30e-03 & 7.32e-03 & 1.34e-02 & 1.91e-02 & 2.39e-02 & 2.77e-02 & 3.05e-02 & 3.12e-02 + pt & 7.70e-07 & -8.61e-05 & 1.81e-03 & 1.49e-02 & 3.84e-02 & 6.56e-02 & 9.08e-02 & 1.12e-01 & 1.30e-01 & 1.44e-01 & 1.48e-01 + au & 1.64e-06 & -1.43e-04 & 1.27e-03 & 1.11e-02 & 2.96e-02 & 5.22e-02 & 7.41e-02 & 9.34e-02 & 1.10e-01 & 1.23e-01 & 1.27e-01 + hg & 4.20e-06 & 3.68e-04 & 1.63e-02 & 9.73e-02 & 2.18e-01 & 3.34e-01 & 4.26e-01 & 4.96e-01 & 5.49e-01 & 5.90e-01 & 6.02e-01 + tl & 7.47e-07 & 9.63e-04 & 1.88e-02 & 1.23e-01 & 2.78e-01 & 4.20e-01 & 5.30e-01 & 6.11e-01 & 6.70e-01 & 7.14e-01 & 7.26e-01 + pb & -1.19e-07 & 1.55e-04 & 5.12e-03 & 5.36e-02 & 1.54e-01 & 2.66e-01 & 3.61e-01 & 4.35e-01 & 4.90e-01 & 5.31e-01 & 5.43e-01 + bi &
5.21e-07 & -2.18e-06 & 1.11e-04 & 2.10e-03 & 8.10e-03 & 1.76e-02 & 2.84e-02 & 3.91e-02 & 4.87e-02 & 5.68e-02 & 5.95e-02 + [ tab7_el ] c & -9.26e-02 & 1.47e-01 & 3.69e-01 & 5.41e-01 & 6.70e-01 & 7.71e-01 & 8.53e-01 & 9.22e-01 & 9.82e-01 & 1.03e+00 & 1.08e+00 & 1.13e+00 & 1.15e+00 & 1.17e+00 + n & 3.94e-01 & 3.94e-01 & 3.93e-01 & 3.92e-01 & 3.90e-01 & 3.89e-01 & 3.87e-01 & 3.86e-01 & 3.84e-01 & 3.83e-01 & 3.82e-01 & 3.82e-01 & 3.86e-01 & 3.97e-01 + o & -4.65e-04 & 2.02e-03 & 6.21e-03 & 1.09e-02 & 1.50e-02 & 1.84e-02 & 2.10e-02 & 2.31e-02 & 2.48e-02 & 2.62e-02 & 2.74e-02 & 2.86e-02 & 2.90e-02 & 2.90e-02 + f & -8.65e-03 & 4.79e-02 & 1.51e-01 & 2.70e-01 & 3.87e-01 & 4.96e-01 & 5.95e-01 & 6.87e-01 & 7.73e-01 & 8.51e-01 & 9.24e-01 & 9.98e-01 & 1.03e+00 & 1.06e+00 + ne & 4.94e-04 & 1.82e-02 & 5.16e-02 & 9.57e-02 & 1.45e-01 & 1.97e-01 & 2.51e-01 & 3.04e-01 & 3.57e-01 & 4.09e-01 & 4.60e-01 & 5.13e-01 & 5.39e-01 & 5.58e-01 + na & 1.14e-01 & 1.19e-01 & 1.27e-01 & 1.38e-01 & 1.52e-01 & 1.67e-01 & 1.85e-01 & 2.05e-01 & 2.28e-01 & 2.53e-01 & 2.82e-01 & 3.16e-01 & 3.37e-01 & 3.58e-01 + mg & 1.58e-04 & 7.25e-04 & 1.92e-03 & 3.86e-03 & 6.57e-03 & 1.01e-02 & 1.46e-02 & 2.01e-02 & 2.68e-02 & 3.47e-02 & 4.41e-02 & 5.55e-02 & 6.22e-02 & 6.77e-02 + al & 5.90e-04 & 1.63e-03 & 3.55e-03 & 6.10e-03 & 8.78e-03 & 1.15e-02 & 1.39e-02 & 1.63e-02 & 1.85e-02 & 2.04e-02 & 2.25e-02 & 2.46e-02 & 2.58e-02 & 2.72e-02 + si & 5.64e-05 & 2.36e-04 & 4.79e-04 & 7.27e-04 & 9.55e-04 & 1.17e-03 & 1.38e-03 & 1.59e-03 & 1.81e-03 & 2.04e-03 & 2.28e-03 & 2.55e-03 & 2.68e-03 & 2.76e-03 + p & 7.75e-04 & 4.24e-03 & 9.83e-03 & 1.63e-02 & 2.24e-02 & 2.81e-02 & 3.32e-02 & 3.79e-02 & 4.25e-02 & 4.68e-02 & 5.11e-02 & 5.54e-02 & 5.73e-02 & 5.85e-02 + s & 1.75e-05 & 8.30e-05 & 1.91e-04 & 3.07e-04 & 4.19e-04 & 5.18e-04 & 6.06e-04 & 6.82e-04 & 7.51e-04 & 8.15e-04 & 8.71e-04 & 9.21e-04 & 9.45e-04 & 9.58e-04 + cl & 3.74e-04 & 1.78e-03 & 3.74e-03 & 5.97e-03 & 8.20e-03 & 1.04e-02 & 1.27e-02 & 1.49e-02 & 1.72e-02 & 1.95e-02 & 2.18e-02 & 2.42e-02 & 2.54e-02 & 2.62e-02 + ar & -4.94e-05 & -1.54e-04 & -2.54e-04 & -3.38e-04 & -4.12e-04 & -4.93e-04 & -5.97e-04 & -7.20e-04 & -8.70e-04 & -1.03e-03 & -1.22e-03 & -1.45e-03 & -1.56e-03 & -1.65e-03 + k & 1.55e-04 & 1.32e-03 & 2.91e-03 & 4.51e-03 & 6.25e-03 & 7.86e-03 & 9.74e-03 & 1.14e-02 & 1.33e-02 & 1.55e-02 & 1.74e-02 & 1.96e-02 & 2.10e-02 & 2.18e-02 + ca & -5.53e-05 & -2.07e-04 & -3.46e-04 & -4.52e-04 & -5.51e-04 & -6.41e-04 & -7.67e-04 & -9.08e-04 & -1.08e-03 & -1.29e-03 & -1.52e-03 & -1.79e-03 & -1.96e-03 & -2.07e-03 + sc & 3.06e-03 & 9.71e-03 & 1.77e-02 & 2.59e-02 & 3.42e-02 & 4.28e-02 & 5.19e-02 & 6.15e-02 & 7.17e-02 & 8.21e-02 & 9.29e-02 & 1.04e-01 & 1.10e-01 & 1.14e-01 + ti & 1.30e-04 & 1.05e-03 & 2.81e-03 & 5.09e-03 & 7.46e-03 & 9.74e-03 & 1.19e-02 & 1.39e-02 & 1.58e-02 & 1.76e-02 & 1.93e-02 & 2.10e-02 & 2.17e-02 & 2.22e-02 + v & -1.69e-04 & -3.83e-04 & -3.61e-04 & 1.71e-05 & 6.93e-04 & 1.54e-03 & 2.44e-03 & 3.31e-03 & 4.13e-03 & 4.87e-03 & 5.53e-03 & 6.14e-03 & 6.42e-03 & 6.60e-03 + cr & 3.06e-07 & 2.54e-05 & 8.76e-05 & 1.71e-04 & 2.59e-04 & 3.45e-04 & 4.27e-04 & 5.01e-04 & 5.71e-04 & 6.39e-04 & 7.04e-04 & 7.72e-04 & 8.04e-04 & 8.20e-04 + mn & -1.45e-04 & -3.23e-04 & -2.21e-04 & 3.71e-04 & 1.48e-03 & 2.99e-03 & 4.74e-03 & 6.65e-03 & 8.62e-03 & 1.06e-02 & 1.26e-02 & 1.47e-02 & 1.57e-02 & 1.65e-02 + fe & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 + co & 5.40e-03 & 1.76e-02 & 3.03e-02 & 4.03e-02 & 4.95e-02 & 5.66e-02 & 6.59e-02 & 7.36e-02 & 8.27e-02 & 9.54e-02 & 1.06e-01 & 1.18e-01 & 1.29e-01 & 1.34e-01 + ni & 1.82e-04 & 1.05e-03 & 2.00e-03 & 2.77e-03 & 3.22e-03 & 3.49e-03 & 3.54e-03 & 3.61e-03 & 3.63e-03 & 3.50e-03 & 3.52e-03 & 3.56e-03 & 3.36e-03 & 3.35e-03 + cu & 7.33e-03 & 2.80e-02 & 4.88e-02 & 6.51e-02 & 7.71e-02 & 8.63e-02 & 9.44e-02 & 1.02e-01 & 1.10e-01 & 1.19e-01 & 1.28e-01 & 1.39e-01 & 1.44e-01 & 1.47e-01 + zn & 6.89e-04 & 5.33e-03 & 1.24e-02 & 1.98e-02 & 2.62e-02 & 3.17e-02 & 3.67e-02 & 4.14e-02 & 4.60e-02 & 5.06e-02 & 5.54e-02 & 6.04e-02 & 6.26e-02 & 6.43e-02 + ga & 4.39e-03 & 2.16e-02 & 4.89e-02 & 7.83e-02 & 1.04e-01 & 1.27e-01 & 1.48e-01 & 1.68e-01 & 1.86e-01 & 2.03e-01 & 2.20e-01 & 2.37e-01 & 2.44e-01 & 2.50e-01 + ge & 4.33e-03 & 2.02e-02 & 4.59e-02 & 7.46e-02 & 1.01e-01 & 1.25e-01 & 1.47e-01 & 1.67e-01 & 1.87e-01 & 2.06e-01 & 2.25e-01 & 2.44e-01 & 2.52e-01 & 2.57e-01 + as & 2.85e-03 & 1.31e-02 & 3.03e-02 & 5.01e-02 & 6.86e-02 & 8.59e-02 & 1.02e-01 & 1.18e-01 & 1.33e-01 & 1.47e-01 & 1.62e-01 & 1.77e-01 & 1.83e-01 & 1.87e-01 + se & 6.89e-03 & 3.08e-02 & 6.63e-02 & 1.05e-01 & 1.39e-01 & 1.69e-01 & 1.97e-01 & 2.23e-01 & 2.49e-01 & 2.72e-01 & 2.96e-01 & 3.19e-01 & 3.27e-01 & 3.34e-01 + br & 3.00e-03 & 1.48e-02 & 3.82e-02 & 6.43e-02 & 9.34e-02 & 1.17e-01 & 1.45e-01 & 1.65e-01 & 1.86e-01 & 2.10e-01 & 2.29e-01 & 2.47e-01 & 2.61e-01 & 2.68e-01 + kr & 1.13e-02 & 4.92e-02 & 1.05e-01 & 1.63e-01 & 2.13e-01 & 2.57e-01 & 2.96e-01 & 3.31e-01 & 3.64e-01 & 3.95e-01 & 4.25e-01 & 4.54e-01 & 4.66e-01 & 4.74e-01 + rb & 2.88e-02 & 9.82e-02 & 1.64e-01 & 2.19e-01 & 2.64e-01 & 3.04e-01 & 3.42e-01 & 3.79e-01 & 4.17e-01 & 4.55e-01 & 4.93e-01 & 5.31e-01 & 5.48e-01 & 5.57e-01 + sr & 4.65e-02 & 2.56e-01 & 5.13e-01 & 7.17e-01 & 8.60e-01 & 9.66e-01 & 1.05e+00 & 1.12e+00 & 1.17e+00 & 1.22e+00 & 1.27e+00 & 1.31e+00 & 1.32e+00 & 1.33e+00 + y & 5.73e-02 & 2.90e-01 & 5.38e-01 & 7.33e-01 & 8.72e-01 & 9.77e-01 & 1.06e+00 & 1.13e+00 & 1.19e+00 & 1.24e+00 & 1.29e+00 & 1.33e+00 & 1.35e+00 & 1.36e+00 + zr & 5.21e-02 & 2.99e-01 & 5.68e-01 & 7.74e-01 & 9.18e-01 & 1.02e+00 & 1.11e+00 & 1.18e+00 & 1.23e+00 & 1.28e+00 & 1.33e+00 & 1.37e+00 & 1.38e+00 & 1.39e+00 + nb & -2.33e-04 & 3.83e-03 & 5.46e-02 & 1.33e-01 & 2.56e-01 & 3.41e-01 & 4.51e-01 & 5.14e-01 & 5.77e-01 & 6.59e-01 & 7.03e-01 & 7.46e-01 & 7.98e-01 & 8.19e-01 + mo & 2.14e-02 & 1.90e-01 & 4.22e-01 & 6.18e-01 & 7.60e-01 & 8.66e-01 & 9.48e-01 & 1.01e+00 & 1.07e+00 & 1.12e+00 & 1.16e+00 & 1.20e+00 & 1.21e+00 & 1.22e+00 + ru & 9.51e-03 & 1.39e-01 & 3.41e-01 & 5.22e-01 & 6.62e-01 & 7.66e-01 & 8.50e-01 & 9.15e-01 & 9.71e-01 & 1.02e+00 & 1.06e+00 & 1.10e+00 & 1.11e+00 & 1.12e+00 + rh & 4.16e-03 & 7.20e-02 & 1.94e-01 & 3.20e-01 & 4.25e-01 & 5.08e-01 & 5.77e-01 & 6.33e-01 & 6.81e-01 & 7.23e-01 & 7.59e-01 & 7.92e-01 & 8.06e-01 & 8.13e-01 + pd & 1.38e-02 & 2.04e-01 & 4.61e-01 & 6.71e-01 & 8.19e-01 & 9.29e-01 & 1.01e+00 & 1.08e+00 & 1.14e+00 & 1.19e+00 & 1.23e+00 & 1.27e+00 & 1.28e+00 & 1.29e+00 + ag & 3.31e-03 & 6.31e-02 & 1.75e-01 & 2.94e-01 & 3.95e-01 & 4.76e-01 & 5.43e-01 & 5.98e-01 & 6.46e-01 & 6.88e-01 & 7.24e-01 & 7.57e-01 & 7.72e-01 & 7.79e-01 + cd & 1.88e-02 & 2.50e-01 & 5.34e-01 & 7.56e-01 & 9.12e-01 & 1.03e+00 & 1.12e+00 & 1.19e+00 & 1.25e+00 & 1.30e+00 & 1.34e+00 & 1.38e+00 & 1.40e+00 & 1.41e+00 + in & 1.13e-02 & 1.72e-01 & 4.01e-01 & 5.98e-01 & 7.42e-01 & 8.51e-01 & 9.37e-01 & 1.01e+00 & 1.07e+00 & 1.11e+00 & 1.16e+00 & 1.20e+00 & 1.21e+00 & 1.22e+00 + sn & 2.30e-02 & 2.88e-01 & 5.91e-01 & 8.19e-01 & 9.78e-01 & 1.09e+00 & 1.18e+00 & 1.25e+00 & 1.31e+00 & 1.36e+00 & 1.41e+00 & 1.45e+00 & 1.46e+00 & 1.47e+00 + sb & 1.06e-02 & 1.46e-01 & 3.32e-01 & 5.00e-01 & 6.31e-01 & 7.35e-01 & 8.19e-01 & 8.89e-01 & 9.49e-01 & 1.00e+00 & 1.04e+00 & 1.09e+00 & 1.10e+00 & 1.11e+00 + te & 5.64e-03 & 1.12e-01 & 2.92e-01 & 4.60e-01 & 5.88e-01 & 6.85e-01 & 7.61e-01 & 8.22e-01 & 8.73e-01 & 9.16e-01 & 9.54e-01 & 9.88e-01 & 1.00e+00 & 1.01e+00 + i & 1.13e-03 & 3.40e-02 & 1.03e-01 & 1.85e-01 & 2.59e-01 & 3.23e-01 & 3.76e-01 & 4.22e-01 & 4.61e-01 & 4.96e-01 & 5.26e-01 & 5.54e-01 & 5.66e-01 & 5.72e-01 + xe & 4.72e-03 & 1.22e-01 & 3.11e-01 & 4.85e-01 & 6.18e-01 & 7.19e-01 & 7.98e-01 & 8.62e-01 & 9.15e-01 & 9.61e-01 & 1.00e+00 & 1.04e+00 & 1.05e+00 & 1.06e+00 + cs & 9.88e-03 & 2.42e-01 & 4.34e-01 & 5.96e-01 & 7.28e-01 & 8.40e-01 & 9.31e-01 & 1.01e+00 & 1.08e+00 & 1.13e+00 & 1.18e+00 & 1.22e+00 & 1.24e+00 & 1.25e+00 + ba & 1.06e-02 & 5.27e-01 & 9.56e-01 & 1.23e+00 & 1.40e+00 & 1.53e+00 & 1.61e+00 & 1.68e+00 & 1.74e+00 & 1.78e+00 & 1.82e+00 & 1.85e+00 & 1.87e+00 & 1.87e+00 + la & 5.61e-03 & 5.39e-01 & 9.78e-01 & 1.25e+00 & 1.43e+00 & 1.55e+00 & 1.63e+00 & 1.70e+00 & 1.75e+00 & 1.79e+00 & 1.83e+00 & 1.86e+00 & 1.87e+00 & 1.88e+00 + ce & 2.97e-03 & 5.73e-01 & 1.03e+00 & 1.30e+00 & 1.48e+00 & 1.60e+00 & 1.68e+00 & 1.74e+00 & 1.79e+00 & 1.83e+00 & 1.86e+00 & 1.89e+00 & 1.90e+00 & 1.91e+00 + pr & 1.96e-03 & 4.29e-01 & 8.05e-01 & 1.06e+00 & 1.23e+00 & 1.35e+00 & 1.44e+00 & 1.50e+00 & 1.56e+00 & 1.60e+00 & 1.64e+00 & 1.67e+00 & 1.68e+00 & 1.69e+00 + nd & 1.45e-03 & 4.80e-01 & 9.03e-01 & 1.17e+00 & 1.34e+00 & 1.46e+00 & 1.54e+00 & 1.60e+00 & 1.64e+00 & 1.68e+00 & 1.71e+00 & 1.74e+00 & 1.75e+00 & 1.76e+00 + sm & 6.49e-04 & 3.15e-01 & 6.73e-01 & 9.24e-01 & 1.09e+00 & 1.20e+00 & 1.28e+00 & 1.33e+00 & 1.38e+00 & 1.41e+00 & 1.44e+00 & 1.47e+00 & 1.48e+00 & 1.49e+00 + eu & -1.52e-04 & 7.48e-02 & 2.14e-01 & 3.57e-01 & 4.73e-01 & 5.58e-01 & 6.21e-01 & 6.70e-01 & 7.08e-01 & 7.39e-01 & 7.65e-01 & 7.89e-01 & 7.98e-01 & 8.04e-01 + gd & 2.58e-04 & 1.91e-01 & 4.67e-01 & 6.73e-01 & 8.14e-01 & 9.11e-01 & 9.81e-01 & 1.04e+00 & 1.08e+00 & 1.11e+00 & 1.14e+00 & 1.17e+00 & 1.18e+00 & 1.19e+00 + tb & -5.67e-05 & 1.13e-01 & 3.19e-01 & 5.01e-01 & 6.31e-01 & 7.20e-01 & 7.83e-01 & 8.30e-01 & 8.67e-01 & 8.96e-01 & 9.21e-01 & 9.44e-01 & 9.53e-01 & 9.59e-01 + dy & 1.28e-04 & 1.78e-01 & 4.73e-01 & 6.87e-01 & 8.44e-01 & 9.42e-01 & 1.01e+00 & 1.06e+00 & 1.10e+00 & 1.13e+00 & 1.16e+00 & 1.18e+00 & 1.19e+00 & 1.20e+00 + ho & 7.99e-05 & 1.36e-01 & 2.87e-01 & 5.02e-01 & 5.79e-01 & 6.73e-01 & 7.33e-01 & 7.82e-01 & 8.29e-01 & 8.47e-01 & 8.73e-01 & 8.98e-01 & 9.03e-01 & 9.08e-01 + er & 1.51e-04 & 1.90e-01 & 4.87e-01 & 7.17e-01 & 8.73e-01 & 9.78e-01 & 1.05e+00 & 1.11e+00 & 1.15e+00 & 1.18e+00 & 1.21e+00 & 1.24e+00 & 1.25e+00 & 1.25e+00 + tm & -3.08e-05 & 1.46e-01 & 4.25e-01 & 6.61e-01 & 8.24e-01 & 9.32e-01 & 1.00e+00 & 1.06e+00 & 1.10e+00 & 1.13e+00 & 1.15e+00 & 1.17e+00 & 1.18e+00 & 1.19e+00 + yb & 5.28e-04 & 3.73e-01 & 7.62e-01 & 1.02e+00 & 1.19e+00 & 1.30e+00 & 1.38e+00 & 1.44e+00 & 1.48e+00 & 1.52e+00 & 1.55e+00 & 1.57e+00 & 1.58e+00 & 1.59e+00 + lu & -9.41e-04 & 2.33e-01 & 5.57e-01 & 7.97e-01 & 9.61e-01 & 1.07e+00 & 1.15e+00 & 1.20e+00 & 1.24e+00 & 1.28e+00 & 1.30e+00 & 1.33e+00 & 1.34e+00 & 1.34e+00 + hf & 1.03e-03 & 4.85e-01 & 8.95e-01 & 1.16e+00 & 1.33e+00 & 1.44e+00 & 1.52e+00 & 1.57e+00 & 1.62e+00 & 1.65e+00 & 1.68e+00 & 1.70e+00 & 1.71e+00 & 1.72e+00 + ta & 4.02e-04 & 3.99e-01 & 8.00e-01 & 1.07e+00 & 1.24e+00 & 1.36e+00 & 1.44e+00 & 1.49e+00 & 1.54e+00 & 1.57e+00 & 1.60e+00 & 1.63e+00 & 1.64e+00 & 1.64e+00 + w & 8.37e-04 & 4.92e-01 & 9.24e-01 & 1.19e+00 & 1.36e+00 & 1.48e+00 & 1.56e+00 & 1.61e+00 & 1.66e+00 & 1.69e+00 & 1.72e+00 & 1.75e+00 & 1.76e+00 & 1.76e+00 + re & -6.11e-02 & 1.64e-01 & 4.38e-01 & 6.62e-01 & 8.27e-01 & 9.43e-01 & 1.03e+00 & 1.09e+00 & 1.14e+00 & 1.17e+00 & 1.21e+00 & 1.23e+00 & 1.24e+00 & 1.25e+00 + os & 4.33e-03 & 1.23e-01 & 3.37e-01 & 5.20e-01 & 6.50e-01 & 7.40e-01 & 8.04e-01 & 8.51e-01 & 8.87e-01 & 9.16e-01 & 9.40e-01 & 9.62e-01 & 9.70e-01 & 9.75e-01 + ir & -2.65e-04 & 2.16e-02 & 7.54e-02 & 1.40e-01 & 1.98e-01 & 2.45e-01 & 2.81e-01 & 3.09e-01 & 3.32e-01 & 3.50e-01 & 3.66e-01 & 3.80e-01 & 3.86e-01 & 3.89e-01 + pt & -7.59e-05 & 8.15e-02 & 2.50e-01 & 4.09e-01 & 5.28e-01 & 6.11e-01 & 6.70e-01 & 7.13e-01 & 7.47e-01 & 7.73e-01 & 7.96e-01 & 8.16e-01 & 8.23e-01 & 8.28e-01 + au & -1.70e-04 & 6.77e-02 & 2.14e-01 & 3.63e-01 & 4.82e-01 & 5.67e-01 & 6.28e-01 & 6.73e-01 & 7.08e-01 & 7.36e-01 & 7.58e-01 & 7.78e-01 & 7.86e-01 & 7.91e-01 + hg & 6.03e-04 & 4.33e-01 & 8.53e-01 & 1.12e+00 & 1.29e+00 & 1.40e+00 & 1.47e+00 & 1.53e+00 & 1.57e+00 & 1.60e+00 & 1.62e+00 & 1.65e+00 & 1.66e+00 & 1.66e+00 + tl & 1.31e-03 & 5.27e-01 & 1.00e+00 & 1.28e+00 & 1.46e+00 & 1.57e+00 & 1.64e+00 & 1.69e+00 & 1.73e+00 & 1.76e+00 & 1.78e+00 & 1.80e+00 & 1.81e+00 & 1.81e+00 + pb & 3.05e-04 & 3.15e-01 & 7.41e-01 & 1.03e+00 & 1.21e+00 & 1.33e+00 & 1.41e+00 & 1.46e+00 & 1.50e+00 & 1.54e+00 & 1.56e+00 & 1.58e+00 & 1.59e+00 & 1.59e+00 + bi & 9.93e-06 & 1.68e-02 & 8.52e-02 & 1.93e-01 & 3.02e-01 & 3.91e-01 & 4.61e-01 & 5.15e-01 & 5.57e-01 & 5.91e-01 & 6.19e-01 & 6.43e-01 & 6.53e-01 & 6.58e-01 + [ tab8_el ] ccccccccccccccccc c & -3.51e-02 & 3.41e-01 & 6.18e-01 & 8.12e-01 & 9.53e-01 & 1.06e+00 & 1.15e+00 & 1.22e+00 & 1.28e+00 & 1.33e+00 & 1.38e+00 & 1.42e+00 & 1.46e+00 & 1.50e+00 & 1.53e+00 & 1.54e+00 + n & 4.16e-01 & 4.15e-01 & 4.15e-01 & 4.14e-01 & 4.13e-01 & 4.12e-01 & 4.11e-01 & 4.10e-01 & 4.10e-01 & 4.10e-01 & 4.10e-01 & 4.10e-01 & 4.11e-01 & 4.14e-01 & 4.20e-01 & 4.38e-01 + o & -1.42e-03 & 4.43e-03 & 1.41e-02 & 2.49e-02 & 3.49e-02 & 4.34e-02 & 5.05e-02 & 5.65e-02 & 6.19e-02 & 6.68e-02 & 7.14e-02 & 7.58e-02 & 8.01e-02 & 8.44e-02 & 8.87e-02 & 8.97e-02 + f & -1.81e-02 & 1.03e-01 & 2.82e-01 & 4.52e-01 & 5.99e-01 & 7.27e-01 & 8.39e-01 & 9.38e-01 & 1.03e+00 & 1.11e+00 & 1.19e+00 & 1.26e+00 & 1.32e+00 & 1.39e+00 & 1.45e+00 & 1.47e+00 + ne & -8.52e-04 & 2.64e-02 & 7.96e-02 & 1.48e-01 & 2.23e-01 & 2.97e-01 & 3.70e-01 & 4.41e-01 & 5.09e-01 & 5.74e-01 & 6.36e-01 & 6.97e-01 & 7.54e-01 & 8.10e-01 & 8.63e-01 & 8.81e-01 + na & 1.37e-01 & 1.41e-01 & 1.50e-01 & 1.65e-01 & 1.83e-01 & 2.05e-01 & 2.30e-01 & 2.58e-01 & 2.89e-01 & 3.23e-01 & 3.61e-01 & 4.03e-01 & 4.47e-01 & 4.95e-01 & 5.45e-01 & 5.70e-01 + mg & 2.71e-04 & 1.29e-03 & 3.36e-03 & 6.64e-03 & 1.12e-02 & 1.72e-02 & 2.48e-02 & 3.41e-02 & 4.55e-02 & 5.89e-02 & 7.47e-02 & 9.30e-02 & 1.13e-01 & 1.36e-01 & 1.60e-01 & 1.70e-01 + al & 8.66e-04 & 2.82e-03 & 6.22e-03 & 1.03e-02 & 1.44e-02 & 1.81e-02 & 2.13e-02 & 2.43e-02 & 2.70e-02 & 2.96e-02 & 3.21e-02 & 3.46e-02 & 3.70e-02 & 3.96e-02 & 4.24e-02 & 4.41e-02 + si & 6.20e-05 & 3.40e-04 & 6.68e-04 & 9.82e-04 & 1.28e-03 & 1.55e-03 & 1.83e-03 & 2.10e-03 & 2.39e-03 & 2.70e-03 & 3.04e-03 & 3.40e-03 & 3.77e-03 & 4.17e-03 & 4.58e-03 & 4.71e-03 + p & 9.60e-04 & 6.39e-03 & 1.43e-02 & 2.28e-02 & 3.08e-02 & 3.80e-02 & 4.46e-02 & 5.06e-02 & 5.64e-02 & 6.19e-02 & 6.74e-02 & 7.29e-02 & 7.83e-02 & 8.37e-02 & 8.88e-02 & 9.02e-02 + s & 1.76e-05 & 1.31e-04 & 2.79e-04 & 4.30e-04 & 5.65e-04 & 6.74e-04 & 7.67e-04 & 8.42e-04 & 9.05e-04 & 9.61e-04 & 1.01e-03 & 1.06e-03 & 1.10e-03 & 1.15e-03 & 1.19e-03 & 1.20e-03 + cl & 2.85e-04 & 1.72e-03 & 3.58e-03 & 5.64e-03 & 7.79e-03 & 9.94e-03 & 1.20e-02 & 1.42e-02 & 1.63e-02 & 1.84e-02 & 2.06e-02 & 2.28e-02 & 2.50e-02 & 2.72e-02 & 2.95e-02 & 3.02e-02 + ar & -4.74e-05 & -1.49e-04 & -2.47e-04 & -3.38e-04 & -4.40e-04 & -5.68e-04 & -7.19e-04 & -9.07e-04 & -1.12e-03 & -1.35e-03 & -1.61e-03 & -1.89e-03 & -2.19e-03 & -2.49e-03 & -2.81e-03 & -2.91e-03 + k & 2.05e-04 & 1.75e-03 & 3.54e-03 & 5.30e-03 & 7.09e-03 & 8.93e-03 & 1.11e-02 & 1.29e-02 & 1.51e-02 & 1.73e-02 & 1.95e-02 & 2.20e-02 & 2.46e-02 & 2.71e-02 & 2.97e-02 & 3.08e-02 + ca & -5.99e-05 & -2.25e-04 & -3.81e-04 & -5.09e-04 & -6.39e-04 & -7.95e-04 & -1.00e-03 & -1.22e-03 & -1.49e-03 & -1.80e-03 & -2.13e-03 & -2.50e-03 & -2.90e-03 & -3.30e-03 & -3.73e-03 & -3.88e-03 + sc & 3.53e-03 & 1.28e-02 & 2.27e-02 & 3.23e-02 & 4.21e-02 & 5.25e-02 & 6.35e-02 & 7.51e-02 & 8.73e-02 & 9.99e-02 & 1.13e-01 & 1.26e-01 & 1.40e-01 & 1.53e-01 & 1.66e-01 & 1.70e-01 + ti & 1.43e-04 & 1.42e-03 & 3.59e-03 & 6.22e-03 & 8.98e-03 & 1.16e-02 & 1.42e-02 & 1.65e-02 & 1.88e-02 & 2.09e-02 & 2.30e-02 & 2.50e-02 & 2.70e-02 & 2.89e-02 & 3.06e-02 & 3.11e-02 + v & -1.40e-04 & -2.92e-04 & -1.22e-04 & 4.82e-04 & 1.39e-03 & 2.38e-03 & 3.31e-03 & 4.12e-03 & 4.80e-03 & 5.34e-03 & 5.75e-03 & 6.04e-03 & 6.25e-03 & 6.39e-03 & 6.46e-03 & 6.48e-03 + cr & -3.04e-06 & 4.55e-05 & 1.41e-04 & 2.65e-04 & 3.99e-04 & 5.27e-04 & 6.55e-04 & 7.73e-04 & 8.87e-04 & 1.00e-03 & 1.11e-03 & 1.22e-03 & 1.34e-03 & 1.45e-03 & 1.56e-03 & 1.59e-03 + mn & -1.47e-04 & -3.69e-04 & -2.11e-04 & 5.92e-04 & 1.98e-03 & 3.70e-03 & 5.54e-03 & 7.35e-03 & 9.08e-03 & 1.07e-02 & 1.22e-02 & 1.36e-02 & 1.49e-02 & 1.62e-02 & 1.75e-02 & 1.80e-02 + fe & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 + co & 6.39e-03 & 2.21e-02 & 3.71e-02 & 4.86e-02 & 5.84e-02 & 6.82e-02 & 8.19e-02 & 9.18e-02 & 1.07e-01 & 1.22e-01 & 1.38e-01 & 1.56e-01 & 1.76e-01 & 1.95e-01 & 2.14e-01 & 2.22e-01 + ni & 2.63e-04 & 1.80e-03 & 3.33e-03 & 4.44e-03 & 5.10e-03 & 5.41e-03 & 5.42e-03 & 5.57e-03 & 5.55e-03 & 5.61e-03 & 5.81e-03 & 6.00e-03 & 6.26e-03 & 6.73e-03 & 7.20e-03 & 7.26e-03 + cu & 9.27e-03 & 4.61e-02 & 7.82e-02 & 1.01e-01 & 1.16e-01 & 1.27e-01 & 1.37e-01 & 1.46e-01 & 1.56e-01 & 1.67e-01 & 1.80e-01 & 1.94e-01 & 2.09e-01 & 2.25e-01 & 2.42e-01 & 2.47e-01 + zn & 8.14e-04 & 8.45e-03 & 1.85e-02 & 2.80e-02 & 3.55e-02 & 4.14e-02 & 4.62e-02 & 5.04e-02 & 5.44e-02 & 5.84e-02 & 6.26e-02 & 6.70e-02 & 7.18e-02 & 7.71e-02 & 8.24e-02 & 8.41e-02 + ga & 3.89e-03 & 2.92e-02 & 5.94e-02 & 8.71e-02 & 1.10e-01 & 1.29e-01 & 1.45e-01 & 1.60e-01 & 1.74e-01 & 1.87e-01 & 2.01e-01 & 2.14e-01 & 2.27e-01 & 2.41e-01 & 2.54e-01 & 2.58e-01 + ge & 3.49e-03 & 2.34e-02 & 4.84e-02 & 7.31e-02 & 9.54e-02 & 1.15e-01 & 1.32e-01 & 1.48e-01 & 1.63e-01 & 1.78e-01 & 1.92e-01 & 2.06e-01 & 2.20e-01 & 2.35e-01 & 2.48e-01 & 2.52e-01 + as & 2.20e-03 & 1.38e-02 & 2.92e-02 & 4.50e-02 & 5.98e-02 & 7.34e-02 & 8.58e-02 & 9.73e-02 & 1.08e-01 & 1.19e-01 & 1.30e-01 & 1.41e-01 & 1.51e-01 & 1.63e-01 & 1.73e-01 & 1.75e-01 + se & 5.29e-03 & 3.02e-02 & 6.00e-02 & 8.92e-02 & 1.16e-01 & 1.40e-01 & 1.62e-01 & 1.82e-01 & 2.02e-01 & 2.20e-01 & 2.39e-01 & 2.56e-01 & 2.73e-01 & 2.92e-01 & 3.07e-01 & 3.11e-01 + br & 2.37e-03 & 1.45e-02 & 3.28e-02 & 5.22e-02 & 7.24e-02 & 9.22e-02 & 1.15e-01 & 1.29e-01 & 1.48e-01 & 1.64e-01 & 1.78e-01 & 1.94e-01 & 2.10e-01 & 2.23e-01 & 2.36e-01 & 2.43e-01 + kr & 9.75e-03 & 5.06e-02 & 1.00e-01 & 1.46e-01 & 1.88e-01 & 2.24e-01 & 2.55e-01 & 2.84e-01 & 3.12e-01 & 3.38e-01 & 3.64e-01 & 3.88e-01 & 4.12e-01 & 4.37e-01 & 4.58e-01 & 4.64e-01 + rb & 2.78e-02 & 1.11e-01 & 1.78e-01 & 2.28e-01 & 2.71e-01 & 3.11e-01 & 3.51e-01 & 3.90e-01 & 4.30e-01 & 4.70e-01 & 5.10e-01 & 5.49e-01 & 5.87e-01 & 6.24e-01 & 6.56e-01 & 6.64e-01 + sr & 4.20e-02 & 2.33e-01 & 4.39e-01 & 6.04e-01 & 7.32e-01 & 8.31e-01 & 9.09e-01 & 9.72e-01 & 1.03e+00 & 1.08e+00 & 1.12e+00 & 1.16e+00 & 1.19e+00 & 1.23e+00 & 1.25e+00 & 1.26e+00 + y & 5.65e-02 & 2.76e-01 & 4.82e-01 & 6.43e-01 & 7.67e-01 & 8.66e-01 & 9.46e-01 & 1.01e+00 & 1.07e+00 & 1.13e+00 & 1.17e+00 & 1.22e+00 & 1.26e+00 & 1.29e+00 & 1.32e+00 & 1.33e+00 + zr & 6.29e-02 & 3.02e-01 & 5.24e-01 & 6.92e-01 & 8.19e-01 & 9.17e-01 & 9.96e-01 & 1.06e+00 & 1.12e+00 & 1.17e+00 & 1.22e+00 & 1.26e+00 & 1.30e+00 & 1.34e+00 & 1.37e+00 & 1.38e+00 + nb & 9.58e-04 & 9.42e-03 & 5.04e-02 & 1.12e-01 & 1.88e-01 & 2.68e-01 & 3.83e-01 & 4.21e-01 & 4.98e-01 & 5.57e-01 & 5.98e-01 & 6.51e-01 & 7.05e-01 & 7.37e-01 & 7.74e-01 & 8.03e-01 + mo & 3.27e-02 & 1.98e-01 & 3.88e-01 & 5.47e-01 & 6.70e-01 & 7.65e-01 & 8.41e-01 & 9.03e-01 & 9.57e-01 & 1.00e+00 & 1.05e+00 & 1.08e+00 & 1.12e+00 & 1.15e+00 & 1.18e+00 & 1.19e+00 + ru & 1.61e-02 & 1.43e-01 & 3.06e-01 & 4.52e-01 & 5.72e-01 & 6.67e-01 & 7.47e-01 & 8.08e-01 & 8.64e-01 & 9.11e-01 & 9.52e-01 & 9.90e-01 & 1.02e+00 & 1.06e+00 & 1.08e+00 & 1.09e+00 + rh & 6.95e-03 & 7.32e-02 & 1.70e-01 & 2.69e-01 & 3.56e-01 & 4.28e-01 & 4.89e-01 & 5.40e-01 & 5.85e-01 & 6.25e-01 & 6.61e-01 & 6.94e-01 & 7.23e-01 & 7.51e-01 & 7.75e-01 & 7.81e-01 + pd & 2.14e-02 & 2.10e-01 & 4.20e-01 & 5.91e-01 & 7.22e-01 & 8.23e-01 & 9.04e-01 & 9.70e-01 & 1.03e+00 & 1.08e+00 & 1.12e+00 & 1.16e+00 & 1.19e+00 & 1.23e+00 & 1.25e+00 & 1.26e+00 + ag & 5.08e-03 & 6.40e-02 & 1.53e-01 & 2.45e-01 & 3.28e-01 & 3.98e-01 & 4.58e-01 & 5.08e-01 & 5.53e-01 & 5.93e-01 & 6.28e-01 & 6.60e-01 & 6.90e-01 & 7.17e-01 & 7.41e-01 & 7.48e-01 + cd & 2.63e-02 & 2.50e-01 & 4.87e-01 & 6.74e-01 & 8.16e-01 & 9.25e-01 & 1.01e+00 & 1.08e+00 & 1.15e+00 & 1.20e+00 & 1.25e+00 & 1.29e+00 & 1.33e+00 & 1.36e+00 & 1.39e+00 & 1.40e+00 + in & 1.48e-02 & 1.66e-01 & 3.55e-01 & 5.19e-01 & 6.51e-01 & 7.55e-01 & 8.38e-01 & 9.08e-01 & 9.67e-01 & 1.02e+00 & 1.07e+00 & 1.11e+00 & 1.14e+00 & 1.18e+00 & 1.20e+00 & 1.21e+00 + sn & 3.68e-02 & 2.85e-01 & 5.38e-01 & 7.33e-01 & 8.79e-01 & 9.92e-01 & 1.08e+00 & 1.16e+00 & 1.22e+00 & 1.27e+00 & 1.32e+00 & 1.37e+00 & 1.41e+00 & 1.44e+00 & 1.48e+00 & 1.48e+00 + sb & 1.94e-02 & 1.49e-01 & 3.02e-01 & 4.45e-01 & 5.67e-01 & 6.69e-01 & 7.54e-01 & 8.28e-01 & 8.92e-01 & 9.50e-01 & 1.00e+00 & 1.05e+00 & 1.09e+00 & 1.13e+00 & 1.16e+00 & 1.17e+00 + te & 1.12e-02 & 1.11e-01 & 2.56e-01 & 3.93e-01 & 5.04e-01 & 5.93e-01 & 6.66e-01 & 7.26e-01 & 7.79e-01 & 8.25e-01 & 8.66e-01 & 9.04e-01 & 9.37e-01 & 9.69e-01 & 9.96e-01 & 1.00e+00 + i & 2.66e-03 & 3.35e-02 & 8.83e-02 & 1.52e-01 & 2.14e-01 & 2.70e-01 & 3.19e-01 & 3.62e-01 & 4.01e-01 & 4.35e-01 & 4.67e-01 & 4.96e-01 & 5.22e-01 & 5.47e-01 & 5.69e-01 & 5.75e-01 + xe & 1.10e-02 & 1.23e-01 & 2.77e-01 & 4.21e-01 & 5.38e-01 & 6.33e-01 & 7.11e-01 & 7.77e-01 & 8.34e-01 & 8.84e-01 & 9.29e-01 & 9.69e-01 & 1.01e+00 & 1.04e+00 & 1.07e+00 & 1.07e+00 + cs & 1.85e-02 & 2.12e-01 & 3.79e-01 & 5.32e-01 & 6.69e-01 & 7.86e-01 & 8.81e-01 & 9.64e-01 & 1.03e+00 & 1.09e+00 & 1.15e+00 & 1.19e+00 & 1.24e+00 & 1.27e+00 & 1.30e+00 & 1.31e+00 + ba & 3.10e-02 & 5.65e-01 & 9.56e-01 & 1.21e+00 & 1.38e+00 & 1.51e+00 & 1.61e+00 & 1.69e+00 & 1.75e+00 & 1.81e+00 & 1.86e+00 & 1.90e+00 & 1.94e+00 & 1.97e+00 & 2.00e+00 & 2.01e+00 + la & 2.13e-02 & 5.91e-01 & 9.99e-01 & 1.26e+00 & 1.43e+00 & 1.56e+00 & 1.66e+00 & 1.74e+00 & 1.80e+00 & 1.86e+00 & 1.90e+00 & 1.95e+00 & 1.98e+00 & 2.02e+00 & 2.04e+00 & 2.05e+00 + ce & 1.34e-02 & 6.71e-01 & 1.10e+00 & 1.37e+00 & 1.55e+00 & 1.67e+00 & 1.77e+00 & 1.85e+00 & 1.91e+00 & 1.97e+00 & 2.01e+00 & 2.05e+00 & 2.09e+00 & 2.12e+00 & 2.15e+00 & 2.15e+00 + pr & 8.80e-03 & 5.21e-01 & 8.91e-01 & 1.14e+00 & 1.32e+00 & 1.46e+00 & 1.56e+00 & 1.64e+00 & 1.71e+00 & 1.77e+00 & 1.82e+00 & 1.87e+00 & 1.91e+00 & 1.94e+00 & 1.97e+00 & 1.98e+00 + nd & 8.21e-03 & 5.95e-01 & 1.01e+00 & 1.27e+00 & 1.44e+00 & 1.57e+00 & 1.67e+00 & 1.74e+00 & 1.81e+00 & 1.86e+00 & 1.90e+00 & 1.94e+00 & 1.98e+00 & 2.01e+00 & 2.03e+00 & 2.04e+00 + sm & 4.89e-03 & 4.35e-01 & 8.03e-01 & 1.05e+00 & 1.22e+00 & 1.34e+00 & 1.43e+00 & 1.50e+00 & 1.56e+00 & 1.61e+00 & 1.66e+00 & 1.70e+00 & 1.73e+00 & 1.77e+00 & 1.79e+00 & 1.80e+00 + eu & 6.43e-04 & 1.17e-01 & 2.90e-01 & 4.51e-01 & 5.80e-01 & 6.79e-01 & 7.58e-01 & 8.23e-01 & 8.78e-01 & 9.25e-01 & 9.67e-01 & 1.01e+00 & 1.04e+00 & 1.07e+00 & 1.09e+00 & 1.10e+00 + gd & 2.55e-03 & 2.82e-01 & 5.76e-01 & 7.82e-01 & 9.31e-01 & 1.05e+00 & 1.14e+00 & 1.21e+00 & 1.28e+00 & 1.33e+00 & 1.38e+00 & 1.42e+00 & 1.46e+00 & 1.49e+00 & 1.52e+00 & 1.53e+00 + tb & 1.15e-03 & 1.77e-01 & 4.16e-01 & 6.02e-01 & 7.35e-01 & 8.34e-01 & 9.11e-01 & 9.75e-01 & 1.03e+00 & 1.08e+00 & 1.12e+00 & 1.16e+00 & 1.19e+00 & 1.22e+00 & 1.25e+00 & 1.26e+00 + dy & 2.24e-03 & 2.78e-01 & 5.88e-01 & 8.06e-01 & 9.58e-01 & 1.07e+00 & 1.15e+00 & 1.22e+00 & 1.27e+00 & 1.32e+00 & 1.36e+00 & 1.40e+00 & 1.43e+00 & 1.46e+00 & 1.48e+00 & 1.49e+00 + ho & 8.11e-04 & 1.51e-01 & 3.78e-01 & 5.89e-01 & 7.22e-01 & 8.17e-01 & 8.75e-01 & 9.42e-01 & 9.90e-01 & 1.03e+00 & 1.08e+00 & 1.11e+00 & 1.14e+00 & 1.17e+00 & 1.19e+00 & 1.20e+00 + er & 2.07e-03 & 2.85e-01 & 6.12e-01 & 8.46e-01 & 1.01e+00 & 1.13e+00 & 1.22e+00 & 1.29e+00 & 1.35e+00 & 1.40e+00 & 1.44e+00 & 1.48e+00 & 1.51e+00 & 1.54e+00 & 1.57e+00 & 1.57e+00 + tm & 1.23e-03 & 2.21e-01 & 5.48e-01 & 7.94e-01 & 9.57e-01 & 1.07e+00 & 1.15e+00 & 1.22e+00 & 1.27e+00 & 1.31e+00 & 1.35e+00 & 1.38e+00 & 1.41e+00 & 1.44e+00 & 1.46e+00 & 1.47e+00 + yb & 4.19e-03 & 5.06e-01 & 9.06e-01 & 1.17e+00 & 1.34e+00 & 1.47e+00 & 1.56e+00 & 1.64e+00 & 1.70e+00 & 1.75e+00 & 1.80e+00 & 1.84e+00 & 1.87e+00 & 1.90e+00 & 1.92e+00 & 1.93e+00 + lu & 7.71e-04 & 3.29e-01 & 6.86e-01 & 9.36e-01 & 1.11e+00 & 1.23e+00 & 1.32e+00 & 1.39e+00 & 1.45e+00 & 1.49e+00 & 1.53e+00 & 1.57e+00 & 1.60e+00 & 1.63e+00 & 1.65e+00 & 1.66e+00 + hf & 5.19e-03 & 6.55e-01 & 1.06e+00 & 1.32e+00 & 1.49e+00 & 1.62e+00 & 1.71e+00 & 1.79e+00 & 1.85e+00 & 1.90e+00 & 1.94e+00 & 1.98e+00 & 2.02e+00 & 2.05e+00 & 2.07e+00 & 2.08e+00 + ta & 2.99e-03 & 5.59e-01 & 9.75e-01 & 1.24e+00 & 1.42e+00 & 1.55e+00 & 1.64e+00 & 1.71e+00 & 1.77e+00 & 1.82e+00 & 1.86e+00 & 1.90e+00 & 1.93e+00 & 1.96e+00 & 1.98e+00 & 1.99e+00 + w & 3.89e-03 & 6.77e-01 & 1.11e+00 & 1.38e+00 & 1.55e+00 & 1.68e+00 & 1.78e+00 & 1.85e+00 & 1.91e+00 & 1.96e+00 & 2.00e+00 & 2.04e+00 & 2.07e+00 & 2.10e+00 & 2.12e+00 & 2.13e+00 + re & -5.86e-02 & 2.89e-01 & 6.10e-01 & 8.57e-01 & 1.04e+00 & 1.17e+00 & 1.27e+00 & 1.35e+00 & 1.42e+00 & 1.47e+00 & 1.51e+00 & 1.55e+00 & 1.58e+00 & 1.61e+00 & 1.64e+00 & 1.65e+00 + os & 4.65e-03 & 2.04e-01 & 4.63e-01 & 6.62e-01 & 8.06e-01 & 9.13e-01 & 9.97e-01 & 1.06e+00 & 1.12e+00 & 1.17e+00 & 1.21e+00 & 1.24e+00 & 1.28e+00 & 1.30e+00 & 1.33e+00 & 1.33e+00 + ir & -1.69e-04 & 4.08e-02 & 1.20e-01 & 2.08e-01 & 2.88e-01 & 3.56e-01 & 4.13e-01 & 4.62e-01 & 5.03e-01 & 5.39e-01 & 5.71e-01 & 6.00e-01 & 6.25e-01 & 6.48e-01 & 6.68e-01 & 6.73e-01 + pt & 1.92e-04 & 1.47e-01 & 3.67e-01 & 5.49e-01 & 6.84e-01 & 7.85e-01 & 8.64e-01 & 9.27e-01 & 9.80e-01 & 1.03e+00 & 1.06e+00 & 1.10e+00 & 1.13e+00 & 1.16e+00 & 1.18e+00 & 1.18e+00 + au & 1.09e-05 & 1.23e-01 & 3.25e-01 & 5.08e-01 & 6.49e-01 & 7.54e-01 & 8.34e-01 & 8.98e-01 & 9.49e-01 & 9.92e-01 & 1.03e+00 & 1.06e+00 & 1.09e+00 & 1.11e+00 & 1.13e+00 & 1.14e+00 + hg & 1.52e-03 & 6.41e-01 & 1.08e+00 & 1.34e+00 & 1.52e+00 & 1.65e+00 & 1.74e+00 & 1.81e+00 & 1.87e+00 & 1.92e+00 & 1.96e+00 & 2.00e+00 & 2.03e+00 & 2.05e+00 & 2.08e+00 & 2.08e+00 + tl & 1.75e-03 & 7.92e-01 & 1.27e+00 & 1.56e+00 & 1.74e+00 & 1.86e+00 & 1.95e+00 & 2.02e+00 & 2.08e+00 & 2.12e+00 & 2.16e+00 & 2.20e+00 & 2.22e+00 & 2.25e+00 & 2.27e+00 & 2.27e+00 + pb & 4.89e-04 & 9.40e-01 & 1.45e+00 & 1.74e+00 & 1.91e+00 & 2.02e+00 & 2.09e+00 & 2.15e+00 & 2.19e+00 & 2.22e+00 & 2.25e+00 & 2.27e+00 & 2.29e+00 & 2.31e+00 & 2.32e+00 & 2.32e+00 + bi & 1.98e-05 & 2.93e-01 & 6.44e-01 & 8.91e-01 & 1.05e+00 & 1.17e+00 & 1.24e+00 & 1.30e+00 & 1.34e+00 & 1.38e+00 & 1.41e+00 & 1.43e+00 & 1.46e+00 & 1.47e+00 & 1.49e+00 & 1.50e+00 + [ tab9_el ] ccccccccccccccccc c & -2.13e-01 & 5.67e-01 & 9.88e-01 & 1.25e+00 & 1.43e+00 & 1.56e+00 & 1.66e+00 & 1.74e+00 & 1.80e+00 & 1.86e+00 & 1.91e+00 & 1.95e+00 & 1.99e+00 & 2.06e+00 & 2.13e+00 & 2.18e+00 + n & 4.59e-01 & 4.59e-01 & 4.58e-01 & 4.58e-01 & 4.58e-01 & 4.59e-01 & 4.59e-01 & 4.61e-01 & 4.62e-01 & 4.65e-01 & 4.68e-01 & 4.72e-01 & 4.76e-01 & 4.88e-01 & 5.09e-01 & 5.56e-01 + o & -9.12e-03 & 4.42e-03 & 3.39e-02 & 7.04e-02 & 1.05e-01 & 1.35e-01 & 1.59e-01 & 1.79e-01 & 1.96e-01 & 2.12e-01 & 2.27e-01 & 2.40e-01 & 2.53e-01 & 2.78e-01 & 3.02e-01 & 3.22e-01 + f & -3.85e-02 & 3.32e-01 & 6.98e-01 & 9.50e-01 & 1.14e+00 & 1.28e+00 & 1.40e+00 & 1.51e+00 & 1.60e+00 & 1.68e+00 & 1.76e+00 & 1.83e+00 & 1.89e+00 & 2.02e+00 & 2.13e+00 & 2.22e+00 + ne & -5.32e-03 & 5.01e-02 & 1.75e-01 & 3.21e-01 & 4.64e-01 & 5.98e-01 & 7.21e-01 & 8.32e-01 & 9.32e-01 & 1.03e+00 & 1.11e+00 & 1.19e+00 & 1.26e+00 & 1.39e+00 & 1.51e+00 & 1.60e+00 + na & 1.67e-01 & 1.73e-01 & 1.94e-01 & 2.32e-01 & 2.82e-01 & 3.41e-01 & 4.06e-01 & 4.75e-01 & 5.46e-01 & 6.20e-01 & 6.94e-01 & 7.69e-01 & 8.43e-01 & 9.94e-01 & 1.15e+00 & 1.28e+00 + mg & 9.40e-05 & 1.12e-03 & 5.20e-03 & 1.47e-02 & 3.06e-02 & 5.35e-02 & 8.36e-02 & 1.21e-01 & 1.65e-01 & 2.16e-01 & 2.70e-01 & 3.27e-01 & 3.85e-01 & 5.04e-01 & 6.22e-01 & 7.19e-01 + al & 2.91e-03 & 1.16e-02 & 2.74e-02 & 4.38e-02 & 5.73e-02 & 6.79e-02 & 7.65e-02 & 8.40e-02 & 9.08e-02 & 9.79e-02 & 1.05e-01 & 1.12e-01 & 1.20e-01 & 1.38e-01 & 1.58e-01 & 1.79e-01 + si & 7.01e-06 & 4.65e-04 & 1.14e-03 & 1.88e-03 & 2.63e-03 & 3.37e-03 & 4.12e-03 & 4.92e-03 & 5.77e-03 & 6.71e-03 & 7.68e-03 & 8.69e-03 & 9.75e-03 & 1.20e-02 & 1.43e-02 & 1.63e-02 + p & 2.56e-06 & 7.25e-03 & 2.10e-02 & 3.81e-02 & 5.48e-02 & 6.93e-02 & 8.16e-02 & 9.28e-02 & 1.03e-01 & 1.14e-01 & 1.24e-01 & 1.34e-01 & 1.44e-01 & 1.63e-01 & 1.82e-01 & 1.96e-01 + s & 2.56e-06 & 1.44e-04 & 3.88e-04 & 6.56e-04 & 8.91e-04 & 1.07e-03 & 1.19e-03 & 1.29e-03 & 1.38e-03 & 1.47e-03 & 1.56e-03 & 1.66e-03 & 1.76e-03 & 1.95e-03 & 2.18e-03 & 2.38e-03 + cl & 1.53e-05 & 9.81e-04 & 2.56e-03 & 4.63e-03 & 6.96e-03 & 9.36e-03 & 1.19e-02 & 1.44e-02 & 1.69e-02 & 1.96e-02 & 2.22e-02 & 2.49e-02 & 2.76e-02 & 3.30e-02 & 3.87e-02 & 4.37e-02 + ar & 2.56e-06 & -6.86e-05 & -1.54e-04 & -2.75e-04 & -4.76e-04 & -7.66e-04 & -1.14e-03 & -1.57e-03 & -1.97e-03 & -2.46e-03 & -2.91e-03 & -3.37e-03 & -3.82e-03 & -4.73e-03 & -5.64e-03 & -6.41e-03 + k & -1.91e-04 & 1.52e-03 & 3.90e-03 & 6.37e-03 & 8.89e-03 & 1.18e-02 & 1.48e-02 & 1.83e-02 & 2.11e-02 & 2.59e-02 & 2.98e-02 & 3.38e-02 & 3.78e-02 & 4.60e-02 & 5.41e-02 & 6.09e-02 + ca & 1.37e-05 & -1.73e-04 & -3.89e-04 & -6.12e-04 & -8.98e-04 & -1.30e-03 & -1.80e-03 & -2.38e-03 & -2.99e-03 & -3.65e-03 & -4.31e-03 & -4.97e-03 & -5.64e-03 & -7.01e-03 & -8.36e-03 & -9.51e-03 + sc & 1.83e-05 & 1.49e-02 & 3.06e-02 & 4.56e-02 & 6.16e-02 & 7.93e-02 & 9.90e-02 & 1.20e-01 & 1.42e-01 & 1.65e-01 & 1.88e-01 & 2.10e-01 & 2.31e-01 & 2.71e-01 & 3.08e-01 & 3.35e-01 + ti & 2.56e-06 & 1.08e-03 & 3.21e-03 & 6.04e-03 & 9.05e-03 & 1.19e-02 & 1.45e-02 & 1.71e-02 & 1.95e-02 & 2.21e-02 & 2.47e-02 & 2.72e-02 & 2.97e-02 & 3.47e-02 & 4.00e-02 & 4.42e-02 + v & 7.19e-06 & -6.72e-05 & 2.34e-04 & 9.95e-04 & 1.82e-03 & 2.33e-03 & 2.44e-03 & 2.29e-03 & 1.99e-03 & 1.61e-03 & 1.20e-03 & 7.66e-04 & 3.29e-04 & -6.33e-04 & -1.49e-03 & -2.17e-03 + cr & 2.56e-06 & 3.27e-05 & 1.25e-04 & 2.73e-04 & 4.48e-04 & 6.42e-04 & 8.35e-04 & 1.04e-03 & 1.24e-03 & 1.46e-03 & 1.69e-03 & 1.91e-03 & 2.15e-03 & 2.61e-03 & 3.08e-03 & 3.46e-03 + mn & 1.48e-05 & -2.21e-04 & 3.09e-05 & 1.12e-03 & 2.68e-03 & 4.13e-03 & 5.16e-03 & 5.79e-03 & 6.12e-03 & 6.26e-03 & 6.28e-03 & 6.24e-03 & 6.16e-03 & 6.06e-03 & 6.23e-03 & 6.94e-03 + fe & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 + co & 7.54e-06 & 2.01e-02 & 4.16e-02 & 5.89e-02 & 7.57e-02 & 9.95e-02 & 1.28e-01 & 1.63e-01 & 1.99e-01 & 2.38e-01 & 2.74e-01 & 3.08e-01 & 3.41e-01 & 4.01e-01 & 4.53e-01 & 4.90e-01 + ni & 3.02e-06 & 2.25e-03 & 5.00e-03 & 7.13e-03 & 8.45e-03 & 9.24e-03 & 1.00e-02 & 1.10e-02 & 1.23e-02 & 1.39e-02 & 1.58e-02 & 1.78e-02 & 1.99e-02 & 2.45e-02 & 2.93e-02 & 3.35e-02 + cu & 1.11e-05 & 6.56e-02 & 1.24e-01 & 1.64e-01 & 1.92e-01 & 2.15e-01 & 2.39e-01 & 2.66e-01 & 2.95e-01 & 3.30e-01 & 3.64e-01 & 3.97e-01 & 4.29e-01 & 4.90e-01 & 5.44e-01 & 5.83e-01 + zn & 2.56e-06 & 1.15e-02 & 3.24e-02 & 5.29e-02 & 6.81e-02 & 7.90e-02 & 8.74e-02 & 9.49e-02 & 1.02e-01 & 1.11e-01 & 1.20e-01 & 1.30e-01 & 1.40e-01 & 1.61e-01 & 1.83e-01 & 2.02e-01 + ga & 8.70e-06 & 4.04e-02 & 9.26e-02 & 1.39e-01 & 1.74e-01 & 2.02e-01 & 2.27e-01 & 2.48e-01 & 2.69e-01 & 2.91e-01 & 3.13e-01 & 3.35e-01 & 3.58e-01 & 4.00e-01 & 4.43e-01 & 4.75e-01 + ge & 2.56e-06 & 2.95e-02 & 7.10e-02 & 1.13e-01 & 1.48e-01 & 1.75e-01 & 2.00e-01 & 2.21e-01 & 2.41e-01 & 2.62e-01 & 2.83e-01 & 3.04e-01 & 3.24e-01 & 3.63e-01 & 4.04e-01 & 4.35e-01 + as & -4.84e-06 & 1.53e-02 & 3.88e-02 & 6.59e-02 & 9.00e-02 & 1.09e-01 & 1.27e-01 & 1.43e-01 & 1.58e-01 & 1.75e-01 & 1.90e-01 & 2.06e-01 & 2.22e-01 & 2.51e-01 & 2.84e-01 & 3.09e-01 + se & 5.91e-06 & 2.58e-02 & 6.38e-02 & 1.08e-01 & 1.48e-01 & 1.79e-01 & 2.09e-01 & 2.35e-01 & 2.58e-01 & 2.84e-01 & 3.07e-01 & 3.30e-01 & 3.54e-01 & 3.95e-01 & 4.41e-01 & 4.73e-01 + br & -9.21e-06 & 9.82e-03 & 2.93e-02 & 5.49e-02 & 8.14e-02 & 1.09e-01 & 1.31e-01 & 1.54e-01 & 1.75e-01 & 1.98e-01 & 2.18e-01 & 2.38e-01 & 2.59e-01 & 2.96e-01 & 3.32e-01 & 3.61e-01 + kr & 6.63e-06 & 3.08e-02 & 8.16e-02 & 1.41e-01 & 1.93e-01 & 2.34e-01 & 2.71e-01 & 3.05e-01 & 3.35e-01 & 3.67e-01 & 3.97e-01 & 4.26e-01 & 4.56e-01 & 5.09e-01 & 5.67e-01 & 6.11e-01 + rb & 2.56e-06 & 7.17e-02 & 1.45e-01 & 2.23e-01 & 3.03e-01 & 3.77e-01 & 4.47e-01 & 5.10e-01 & 5.65e-01 & 6.21e-01 & 6.71e-01 & 7.17e-01 & 7.64e-01 & 8.42e-01 & 9.18e-01 & 9.74e-01 + sr & 6.38e-06 & 7.82e-02 & 2.70e-01 & 4.75e-01 & 6.27e-01 & 7.22e-01 & 7.94e-01 & 8.52e-01 & 9.00e-01 & 9.46e-01 & 9.86e-01 & 1.02e+00 & 1.06e+00 & 1.13e+00 & 1.20e+00 & 1.26e+00 + y & 2.56e-06 & 1.04e-01 & 2.99e-01 & 5.01e-01 & 6.60e-01 & 7.66e-01 & 8.47e-01 & 9.15e-01 & 9.72e-01 & 1.03e+00 & 1.07e+00 & 1.11e+00 & 1.16e+00 & 1.23e+00 & 1.31e+00 & 1.37e+00 + zr & 2.56e-06 & 1.38e-01 & 3.52e-01 & 5.48e-01 & 6.94e-01 & 7.95e-01 & 8.73e-01 & 9.40e-01 & 9.97e-01 & 1.05e+00 & 1.10e+00 & 1.14e+00 & 1.19e+00 & 1.27e+00 & 1.35e+00 & 1.41e+00 + nb & -1.21e-05 & 1.21e-05 & 1.47e-02 & 4.43e-02 & 8.74e-02 & 1.63e-01 & 2.14e-01 & 2.81e-01 & 3.42e-01 & 4.08e-01 & 4.61e-01 & 5.12e-01 & 5.64e-01 & 6.48e-01 & 7.16e-01 & 7.77e-01 + mo & 5.72e-06 & 9.16e-02 & 2.61e-01 & 4.23e-01 & 5.41e-01 & 6.22e-01 & 6.84e-01 & 7.39e-01 & 7.87e-01 & 8.34e-01 & 8.76e-01 & 9.17e-01 & 9.57e-01 & 1.03e+00 & 1.11e+00 & 1.16e+00 + ru & -2.34e-06 & 5.74e-02 & 1.93e-01 & 3.41e-01 & 4.53e-01 & 5.32e-01 & 5.87e-01 & 6.34e-01 & 6.75e-01 & 7.13e-01 & 7.48e-01 & 7.81e-01 & 8.14e-01 & 8.76e-01 & 9.47e-01 & 1.00e+00 + rh & -1.52e-05 & 2.72e-02 & 9.94e-02 & 1.90e-01 & 2.65e-01 & 3.19e-01 & 3.60e-01 & 3.96e-01 & 4.27e-01 & 4.57e-01 & 4.85e-01 & 5.13e-01 & 5.40e-01 & 5.93e-01 & 6.57e-01 & 7.08e-01 + pd & 1.19e-05 & 7.93e-02 & 2.60e-01 & 4.43e-01 & 5.76e-01 & 6.67e-01 & 7.32e-01 & 7.87e-01 & 8.35e-01 & 8.79e-01 & 9.21e-01 & 9.59e-01 & 9.97e-01 & 1.07e+00 & 1.15e+00 & 1.20e+00 + ag & -1.61e-05 & 2.03e-02 & 8.12e-02 & 1.62e-01 & 2.31e-01 & 2.84e-01 & 3.24e-01 & 3.61e-01 & 3.94e-01 & 4.26e-01 & 4.56e-01 & 4.85e-01 & 5.15e-01 & 5.70e-01 & 6.34e-01 & 6.81e-01 + cd & 8.25e-06 & 9.19e-02 & 3.01e-01 & 5.09e-01 & 6.61e-01 & 7.66e-01 & 8.40e-01 & 9.03e-01 & 9.56e-01 & 1.01e+00 & 1.05e+00 & 1.09e+00 & 1.13e+00 & 1.21e+00 & 1.29e+00 & 1.35e+00 + in & -8.46e-06 & 5.37e-02 & 2.04e-01 & 3.76e-01 & 5.09e-01 & 6.04e-01 & 6.71e-01 & 7.27e-01 & 7.74e-01 & 8.19e-01 & 8.59e-01 & 8.97e-01 & 9.34e-01 & 1.00e+00 & 1.08e+00 & 1.14e+00 + sn & 3.77e-06 & 1.19e-01 & 3.48e-01 & 5.63e-01 & 7.16e-01 & 8.29e-01 & 9.12e-01 & 9.82e-01 & 1.04e+00 & 1.10e+00 & 1.15e+00 & 1.20e+00 & 1.24e+00 & 1.32e+00 & 1.41e+00 & 1.47e+00 + sb & -1.09e-05 & 6.31e-02 & 1.88e-01 & 3.35e-01 & 4.58e-01 & 5.59e-01 & 6.38e-01 & 7.09e-01 & 7.71e-01 & 8.28e-01 & 8.82e-01 & 9.30e-01 & 9.77e-01 & 1.06e+00 & 1.15e+00 & 1.21e+00 + te & 1.16e-06 & 4.37e-02 & 1.50e-01 & 2.71e-01 & 3.65e-01 & 4.39e-01 & 4.96e-01 & 5.47e-01 & 5.92e-01 & 6.36e-01 & 6.77e-01 & 7.16e-01 & 7.55e-01 & 8.25e-01 & 9.02e-01 & 9.58e-01 + i & -1.97e-05 & 1.24e-02 & 4.85e-02 & 9.92e-02 & 1.44e-01 & 1.81e-01 & 2.10e-01 & 2.37e-01 & 2.62e-01 & 2.86e-01 & 3.11e-01 & 3.34e-01 & 3.59e-01 & 4.06e-01 & 4.62e-01 & 5.08e-01 + xe & 6.72e-06 & 5.53e-02 & 1.74e-01 & 3.08e-01 & 4.12e-01 & 4.94e-01 & 5.58e-01 & 6.14e-01 & 6.64e-01 & 7.12e-01 & 7.57e-01 & 7.99e-01 & 8.41e-01 & 9.16e-01 & 9.96e-01 & 1.05e+00 + cs & -2.35e-06 & 8.73e-02 & 2.37e-01 & 4.19e-01 & 5.65e-01 & 6.78e-01 & 7.62e-01 & 8.33e-01 & 8.93e-01 & 9.49e-01 & 1.00e+00 & 1.05e+00 & 1.09e+00 & 1.17e+00 & 1.26e+00 & 1.32e+00 + ba & 5.06e-06 & 3.00e-01 & 6.98e-01 & 9.91e-01 & 1.17e+00 & 1.30e+00 & 1.39e+00 & 1.47e+00 & 1.53e+00 & 1.59e+00 & 1.65e+00 & 1.70e+00 & 1.75e+00 & 1.84e+00 & 1.93e+00 & 1.99e+00 + la & 2.56e-06 & 3.10e-01 & 7.24e-01 & 1.03e+00 & 1.21e+00 & 1.34e+00 & 1.43e+00 & 1.51e+00 & 1.57e+00 & 1.63e+00 & 1.69e+00 & 1.74e+00 & 1.80e+00 & 1.89e+00 & 1.97e+00 & 2.03e+00 + ce & 6.55e-07 & 3.56e-01 & 8.03e-01 & 1.12e+00 & 1.31e+00 & 1.45e+00 & 1.54e+00 & 1.62e+00 & 1.69e+00 & 1.75e+00 & 1.81e+00 & 1.87e+00 & 1.92e+00 & 2.01e+00 & 2.10e+00 & 2.15e+00 + pr & -2.57e-06 & 2.41e-01 & 6.09e-01 & 9.20e-01 & 1.13e+00 & 1.28e+00 & 1.38e+00 & 1.47e+00 & 1.54e+00 & 1.61e+00 & 1.67e+00 & 1.73e+00 & 1.78e+00 & 1.88e+00 & 1.97e+00 & 2.02e+00 + nd & 2.56e-06 & 3.04e-01 & 7.18e-01 & 1.02e+00 & 1.22e+00 & 1.35e+00 & 1.44e+00 & 1.52e+00 & 1.58e+00 & 1.64e+00 & 1.70e+00 & 1.75e+00 & 1.80e+00 & 1.90e+00 & 1.98e+00 & 2.04e+00 + sm & -4.37e-06 & 2.59e-01 & 5.91e-01 & 8.46e-01 & 1.02e+00 & 1.14e+00 & 1.24e+00 & 1.32e+00 & 1.39e+00 & 1.45e+00 & 1.51e+00 & 1.57e+00 & 1.63e+00 & 1.72e+00 & 1.81e+00 & 1.86e+00 + eu & -2.54e-05 & 6.15e-02 & 1.86e-01 & 3.25e-01 & 4.39e-01 & 5.33e-01 & 6.10e-01 & 6.78e-01 & 7.36e-01 & 7.95e-01 & 8.51e-01 & 9.05e-01 & 9.56e-01 & 1.05e+00 & 1.13e+00 & 1.18e+00 + gd & 7.94e-06 & 1.65e-01 & 3.97e-01 & 6.11e-01 & 7.75e-01 & 9.01e-01 & 9.99e-01 & 1.08e+00 & 1.16e+00 & 1.23e+00 & 1.29e+00 & 1.35e+00 & 1.41e+00 & 1.51e+00 & 1.60e+00 & 1.65e+00 + tb & -1.26e-05 & 9.74e-02 & 2.67e-01 & 4.27e-01 & 5.57e-01 & 6.61e-01 & 7.46e-01 & 8.21e-01 & 8.87e-01 & 9.50e-01 & 1.01e+00 & 1.07e+00 & 1.13e+00 & 1.22e+00 & 1.31e+00 & 1.36e+00 + dy & -2.15e-05 & 1.66e-01 & 4.12e-01 & 6.12e-01 & 7.53e-01 & 8.64e-01 & 9.49e-01 & 1.03e+00 & 1.09e+00 & 1.16e+00 & 1.22e+00 & 1.28e+00 & 1.34e+00 & 1.44e+00 & 1.52e+00 & 1.58e+00 + ho & 9.25e-05 & 9.92e-02 & 2.57e-01 & 4.16e-01 & 5.45e-01 & 6.21e-01 & 7.14e-01 & 7.80e-01 & 8.54e-01 & 8.87e-01 & 9.51e-01 & 1.01e+00 & 1.05e+00 & 1.14e+00 & 1.23e+00 & 1.28e+00 + er & 1.21e-05 & 1.88e-01 & 4.57e-01 & 6.82e-01 & 8.34e-01 & 9.45e-01 & 1.03e+00 & 1.10e+00 & 1.17e+00 & 1.23e+00 & 1.29e+00 & 1.35e+00 & 1.41e+00 & 1.50e+00 & 1.59e+00 & 1.64e+00 + tm & -4.63e-06 & 1.45e-01 & 4.04e-01 & 6.11e-01 & 7.45e-01 & 8.42e-01 & 9.18e-01 & 9.86e-01 & 1.05e+00 & 1.10e+00 & 1.16e+00 & 1.22e+00 & 1.27e+00 & 1.37e+00 & 1.45e+00 & 1.51e+00 + yb & 1.32e-05 & 3.49e-01 & 7.20e-01 & 9.92e-01 & 1.17e+00 & 1.29e+00 & 1.38e+00 & 1.46e+00 & 1.52e+00 & 1.58e+00 & 1.64e+00 & 1.70e+00 & 1.76e+00 & 1.85e+00 & 1.94e+00 & 1.99e+00 + lu & -8.88e-04 & 2.09e-01 & 5.17e-01 & 7.62e-01 & 9.20e-01 & 1.03e+00 & 1.11e+00 & 1.18e+00 & 1.24e+00 & 1.30e+00 & 1.36e+00 & 1.42e+00 & 1.47e+00 & 1.56e+00 & 1.64e+00 & 1.70e+00 + hf & 1.92e-04 & 4.66e-01 & 8.57e-01 & 1.13e+00 & 1.32e+00 & 1.44e+00 & 1.53e+00 & 1.61e+00 & 1.68e+00 & 1.74e+00 & 1.81e+00 & 1.86e+00 & 1.92e+00 & 2.02e+00 & 2.10e+00 & 2.16e+00 + ta & -5.97e-06 & 3.81e-01 & 7.78e-01 & 1.05e+00 & 1.23e+00 & 1.35e+00 & 1.43e+00 & 1.51e+00 & 1.57e+00 & 1.63e+00 & 1.69e+00 & 1.75e+00 & 1.80e+00 & 1.90e+00 & 1.98e+00 & 2.04e+00 + w & 2.56e-06 & 4.80e-01 & 9.04e-01 & 1.19e+00 & 1.37e+00 & 1.49e+00 & 1.57e+00 & 1.64e+00 & 1.71e+00 & 1.77e+00 & 1.82e+00 & 1.88e+00 & 1.93e+00 & 2.03e+00 & 2.11e+00 & 2.16e+00 + re & -6.34e-02 & 1.63e-01 & 4.63e-01 & 7.22e-01 & 8.96e-01 & 1.01e+00 & 1.10e+00 & 1.17e+00 & 1.23e+00 & 1.29e+00 & 1.35e+00 & 1.40e+00 & 1.45e+00 & 1.55e+00 & 1.63e+00 & 1.69e+00 + os & 4.54e-03 & 1.26e-01 & 3.25e-01 & 5.12e-01 & 6.54e-01 & 7.58e-01 & 8.38e-01 & 9.07e-01 & 9.67e-01 & 1.03e+00 & 1.08e+00 & 1.14e+00 & 1.19e+00 & 1.28e+00 & 1.36e+00 & 1.41e+00 + ir & -2.16e-05 & 2.35e-02 & 7.51e-02 & 1.42e-01 & 2.07e-01 & 2.63e-01 & 3.11e-01 & 3.57e-01 & 3.98e-01 & 4.41e-01 & 4.84e-01 & 5.27e-01 & 5.68e-01 & 6.45e-01 & 7.13e-01 & 7.59e-01 + pt & 1.41e-05 & 9.52e-02 & 2.58e-01 & 4.15e-01 & 5.40e-01 & 6.37e-01 & 7.13e-01 & 7.81e-01 & 8.41e-01 & 9.00e-01 & 9.58e-01 & 1.01e+00 & 1.07e+00 & 1.16e+00 & 1.24e+00 & 1.29e+00 + au & -3.15e-06 & 7.59e-02 & 2.31e-01 & 3.89e-01 & 5.07e-01 & 5.93e-01 & 6.56e-01 & 7.12e-01 & 7.62e-01 & 8.11e-01 & 8.60e-01 & 9.08e-01 & 9.55e-01 & 1.04e+00 & 1.11e+00 & 1.17e+00 + hg & 2.34e-05 & 4.00e-01 & 8.35e-01 & 1.14e+00 & 1.33e+00 & 1.45e+00 & 1.54e+00 & 1.61e+00 & 1.67e+00 & 1.73e+00 & 1.79e+00 & 1.84e+00 & 1.89e+00 & 1.98e+00 & 2.06e+00 & 2.12e+00 + tl & 1.36e-06 & 4.53e-01 & 9.66e-01 & 1.29e+00 & 1.50e+00 & 1.63e+00 & 1.72e+00 & 1.80e+00 & 1.86e+00 & 1.92e+00 & 1.98e+00 & 2.03e+00 & 2.08e+00 & 2.17e+00 & 2.25e+00 & 2.30e+00 + pb & 2.56e-06 & 1.16e+00 & 1.78e+00 & 2.11e+00 & 2.32e+00 & 2.45e+00 & 2.54e+00 & 2.60e+00 & 2.65e+00 & 2.70e+00 & 2.73e+00 & 2.76e+00 & 2.78e+00 & 2.83e+00 & 2.86e+00 & 2.87e+00 + bi & 2.56e-06 & 8.43e-01 & 1.33e+00 & 1.63e+00 & 1.84e+00 & 1.98e+00 & 2.09e+00 & 2.16e+00 & 2.23e+00 & 2.28e+00 & 2.32e+00 & 2.35e+00 & 2.37e+00 & 2.42e+00 & 2.44e+00 & 2.46e+00 + [ tab10_el ] cccccccccccccccc c & 1.15e+00 & 1.72e+00 & 2.04e+00 & 2.26e+00 & 2.41e+00 & 2.52e+00 & 2.61e+00 & 2.69e+00 & 2.75e+00 & 2.81e+00 & 2.86e+00 & 2.90e+00 & 2.95e+00 & 2.99e+00 & 3.00e+00 + n & 4.93e-01 & 4.95e-01 & 5.01e-01 & 5.11e-01 & 5.23e-01 & 5.39e-01 & 5.56e-01 & 5.75e-01 & 5.95e-01 & 6.17e-01 & 6.41e-01 & 6.68e-01 & 7.01e-01 & 7.51e-01 & 8.58e-01 + o & -9.15e-04 & 1.49e-01 & 3.30e-01 & 4.82e-01 & 5.95e-01 & 6.79e-01 & 7.43e-01 & 7.95e-01 & 8.38e-01 & 8.76e-01 & 9.09e-01 & 9.40e-01 & 9.71e-01 & 1.00e+00 & 1.01e+00 + f & -8.31e-02 & 1.16e+00 & 1.65e+00 & 1.94e+00 & 2.13e+00 & 2.27e+00 & 2.37e+00 & 2.47e+00 & 2.54e+00 & 2.62e+00 & 2.68e+00 & 2.75e+00 & 2.82e+00 & 2.88e+00 & 2.90e+00 + ne & -3.48e-03 & 4.30e-01 & 8.19e-01 & 1.09e+00 & 1.30e+00 & 1.46e+00 & 1.60e+00 & 1.71e+00 & 1.82e+00 & 1.91e+00 & 2.00e+00 & 2.08e+00 & 2.17e+00 & 2.24e+00 & 2.27e+00 + na & 2.22e-01 & 2.38e-01 & 3.73e-01 & 5.77e-01 & 7.70e-01 & 9.34e-01 & 1.07e+00 & 1.19e+00 & 1.30e+00 & 1.40e+00 & 1.49e+00 & 1.58e+00 & 1.66e+00 & 1.74e+00 & 1.77e+00 + mg & -2.92e-04 & 2.39e-03 & 3.81e-02 & 1.18e-01 & 2.31e-01 & 3.60e-01 & 4.92e-01 & 6.21e-01 & 7.43e-01 & 8.59e-01 & 9.67e-01 & 1.07e+00 & 1.17e+00 & 1.27e+00 & 1.30e+00 + al & 7.02e-03 & 2.28e-02 & 6.55e-02 & 1.29e-01 & 1.94e-01 & 2.53e-01 & 3.04e-01 & 3.51e-01 & 3.95e-01 & 4.37e-01 & 4.80e-01 & 5.23e-01 & 5.69e-01 & 6.15e-01 & 6.37e-01 + si & 9.10e-04 & 2.55e-03 & 4.95e-03 & 8.07e-03 & 1.18e-02 & 1.61e-02 & 2.08e-02 & 2.58e-02 & 3.12e-02 & 3.69e-02 & 4.28e-02 & 4.91e-02 & 5.63e-02 & 6.35e-02 & 6.61e-02 + p & 2.27e-05 & 6.64e-03 & 2.90e-02 & 6.19e-02 & 9.92e-02 & 1.37e-01 & 1.72e-01 & 2.06e-01 & 2.36e-01 & 2.64e-01 & 2.90e-01 & 3.14e-01 & 3.39e-01 & 3.60e-01 & 3.67e-01 + s & 4.87e-07 & 8.39e-05 & 3.32e-04 & 7.02e-04 & 1.14e-03 & 1.60e-03 & 2.05e-03 & 2.48e-03 & 2.88e-03 & 3.26e-03 & 3.61e-03 & 3.93e-03 & 4.26e-03 & 4.55e-03 & 4.65e-03 + cl & 9.84e-05 & 5.73e-04 & 1.30e-03 & 2.14e-03 & 3.03e-03 & 4.03e-03 & 5.06e-03 & 6.14e-03 & 7.35e-03 & 8.57e-03 & 9.79e-03 & 1.10e-02 & 1.24e-02 & 1.37e-02 & 1.42e-02 + ar & -6.15e-06 & 9.16e-06 & 6.10e-05 & 1.65e-04 & 3.34e-04 & 5.73e-04 & 8.92e-04 & 1.24e-03 & 1.65e-03 & 2.09e-03 & 2.59e-03 & 3.16e-03 & 3.61e-03 & 4.17e-03 & 4.38e-03 + k & -1.40e-04 & 2.88e-04 & 1.14e-03 & 2.43e-03 & 4.16e-03 & 6.13e-03 & 8.29e-03 & 1.10e-02 & 1.40e-02 & 1.69e-02 & 1.96e-02 & 2.19e-02 & 2.64e-02 & 3.00e-02 & 3.14e-02 + ca & 5.81e-06 & -9.99e-05 & -3.36e-04 & -6.80e-04 & -1.12e-03 & -1.62e-03 & -2.18e-03 & -2.78e-03 & -3.39e-03 & -4.03e-03 & -4.69e-03 & -5.37e-03 & -6.13e-03 & -6.88e-03 & -7.16e-03 + sc & 4.96e-04 & 4.95e-03 & 1.34e-02 & 2.58e-02 & 4.22e-02 & 6.24e-02 & 8.56e-02 & 1.11e-01 & 1.38e-01 & 1.64e-01 & 1.90e-01 & 2.15e-01 & 2.41e-01 & 2.66e-01 & 2.75e-01 + ti & 5.91e-06 & 5.15e-04 & 1.45e-03 & 2.70e-03 & 4.14e-03 & 5.68e-03 & 7.29e-03 & 8.99e-03 & 1.09e-02 & 1.28e-02 & 1.47e-02 & 1.65e-02 & 1.87e-02 & 2.07e-02 & 2.15e-02 + v & 1.10e-04 & 3.24e-04 & 3.04e-04 & 7.60e-05 & -3.39e-04 & -9.21e-04 & -1.62e-03 & -2.36e-03 & -3.07e-03 & -3.81e-03 & -4.61e-03 & -5.46e-03 & -6.39e-03 & -7.36e-03 & -7.71e-03 + cr & -7.02e-07 & 2.27e-05 & 7.59e-05 & 1.56e-04 & 2.66e-04 & 3.96e-04 & 5.45e-04 & 7.10e-04 & 8.86e-04 & 1.07e-03 & 1.26e-03 & 1.46e-03 & 1.68e-03 & 1.91e-03 & 2.01e-03 + mn & 2.61e-04 & 1.09e-03 & 1.83e-03 & 2.28e-03 & 2.40e-03 & 2.21e-03 & 1.72e-03 & 1.04e-03 & 2.18e-04 & -6.77e-04 & -1.63e-03 & -2.62e-03 & -3.69e-03 & -4.70e-03 & -4.93e-03 + fe & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 + co & 7.18e-05 & 3.49e-03 & 1.60e-02 & 3.97e-02 & 7.34e-02 & 1.13e-01 & 1.58e-01 & 2.04e-01 & 2.49e-01 & 2.92e-01 & 3.33e-01 & 3.72e-01 & 4.12e-01 & 4.47e-01 & 4.60e-01 + ni & -1.88e-06 & 1.28e-05 & 1.46e-04 & 5.74e-04 & 1.46e-03 & 2.92e-03 & 4.89e-03 & 7.32e-03 & 1.02e-02 & 1.34e-02 & 1.68e-02 & 2.05e-02 & 2.45e-02 & 2.87e-02 & 3.04e-02 + cu & 2.20e-04 & 4.26e-03 & 1.63e-02 & 3.85e-02 & 7.31e-02 & 1.20e-01 & 1.74e-01 & 2.32e-01 & 2.90e-01 & 3.46e-01 & 3.98e-01 & 4.44e-01 & 4.97e-01 & 5.42e-01 & 5.57e-01 + zn & -1.06e-05 & 1.22e-03 & 4.09e-03 & 8.41e-03 & 1.42e-02 & 2.26e-02 & 3.34e-02 & 4.58e-02 & 6.21e-02 & 7.84e-02 & 9.55e-02 & 1.13e-01 & 1.31e-01 & 1.50e-01 & 1.57e-01 + ga & 1.60e-04 & 7.75e-03 & 2.09e-02 & 3.87e-02 & 5.94e-02 & 8.74e-02 & 1.22e-01 & 1.58e-01 & 2.06e-01 & 2.50e-01 & 2.92e-01 & 3.32e-01 & 3.72e-01 & 4.10e-01 & 4.22e-01 + ge & 2.74e-05 & 8.36e-03 & 2.23e-02 & 4.11e-02 & 6.06e-02 & 8.66e-02 & 1.17e-01 & 1.49e-01 & 1.94e-01 & 2.33e-01 & 2.71e-01 & 3.06e-01 & 3.41e-01 & 3.75e-01 & 3.87e-01 + as & -1.93e-04 & 4.57e-03 & 1.27e-02 & 2.43e-02 & 3.58e-02 & 5.13e-02 & 7.00e-02 & 8.90e-02 & 1.18e-01 & 1.43e-01 & 1.68e-01 & 1.91e-01 & 2.15e-01 & 2.38e-01 & 2.46e-01 + se & 3.48e-05 & 1.20e-02 & 2.97e-02 & 5.41e-02 & 7.68e-02 & 1.06e-01 & 1.39e-01 & 1.71e-01 & 2.19e-01 & 2.58e-01 & 2.93e-01 & 3.24e-01 & 3.56e-01 & 3.86e-01 & 3.95e-01 + br & -2.61e-04 & 4.72e-03 & 1.27e-02 & 2.52e-02 & 3.88e-02 & 5.56e-02 & 7.63e-02 & 9.80e-02 & 1.30e-01 & 1.58e-01 & 1.84e-01 & 2.09e-01 & 2.35e-01 & 2.61e-01 & 2.74e-01 + kr & 9.98e-05 & 1.76e-02 & 4.51e-02 & 8.36e-02 & 1.20e-01 & 1.61e-01 & 2.07e-01 & 2.52e-01 & 3.12e-01 & 3.60e-01 & 4.00e-01 & 4.36e-01 & 4.71e-01 & 5.03e-01 & 5.14e-01 + rb & 2.02e-05 & 2.40e-02 & 7.17e-02 & 1.48e-01 & 2.25e-01 & 3.09e-01 & 3.96e-01 & 4.78e-01 & 5.75e-01 & 6.51e-01 & 7.10e-01 & 7.59e-01 & 8.08e-01 & 8.49e-01 & 8.63e-01 + sr & 1.19e-04 & 1.41e-01 & 2.81e-01 & 4.21e-01 & 5.17e-01 & 6.00e-01 & 6.70e-01 & 7.38e-01 & 8.11e-01 & 8.64e-01 & 9.04e-01 & 9.37e-01 & 9.74e-01 & 1.00e+00 & 1.01e+00 + y & 8.69e-06 & 1.41e-01 & 2.96e-01 & 4.52e-01 & 5.66e-01 & 6.62e-01 & 7.41e-01 & 8.19e-01 & 9.00e-01 & 9.60e-01 & 1.00e+00 & 1.04e+00 & 1.08e+00 & 1.11e+00 & 1.12e+00 + zr & 9.89e-06 & 1.40e-01 & 2.97e-01 & 4.52e-01 & 5.75e-01 & 6.77e-01 & 7.59e-01 & 8.42e-01 & 9.28e-01 & 9.92e-01 & 1.04e+00 & 1.08e+00 & 1.12e+00 & 1.16e+00 & 1.17e+00 + nb & -3.65e-04 & -5.82e-04 & 4.29e-03 & 1.94e-02 & 5.12e-02 & 8.52e-02 & 1.31e-01 & 1.81e-01 & 2.36e-01 & 2.90e-01 & 3.42e-01 & 3.91e-01 & 4.39e-01 & 4.93e-01 & 5.30e-01 + mo & 7.02e-05 & 8.91e-02 & 1.98e-01 & 3.09e-01 & 4.07e-01 & 4.91e-01 & 5.61e-01 & 6.35e-01 & 7.13e-01 & 7.76e-01 & 8.25e-01 & 8.63e-01 & 9.04e-01 & 9.37e-01 & 9.47e-01 + ru & -1.36e-04 & 5.86e-02 & 1.35e-01 & 2.13e-01 & 2.84e-01 & 3.42e-01 & 3.90e-01 & 4.42e-01 & 4.99e-01 & 5.47e-01 & 5.88e-01 & 6.20e-01 & 6.55e-01 & 6.85e-01 & 6.96e-01 + rh & -3.11e-04 & 3.11e-02 & 7.38e-02 & 1.21e-01 & 1.65e-01 & 2.02e-01 & 2.33e-01 & 2.68e-01 & 3.06e-01 & 3.41e-01 & 3.70e-01 & 3.93e-01 & 4.19e-01 & 4.41e-01 & 4.50e-01 + pd & 1.96e-04 & 9.85e-02 & 2.16e-01 & 3.28e-01 & 4.23e-01 & 4.98e-01 & 5.56e-01 & 6.16e-01 & 6.79e-01 & 7.32e-01 & 7.75e-01 & 8.08e-01 & 8.44e-01 & 8.75e-01 & 8.86e-01 + ag & -3.48e-04 & 2.63e-02 & 6.37e-02 & 1.06e-01 & 1.47e-01 & 1.83e-01 & 2.14e-01 & 2.48e-01 & 2.87e-01 & 3.20e-01 & 3.48e-01 & 3.70e-01 & 3.95e-01 & 4.18e-01 & 4.26e-01 + cd & 1.47e-04 & 1.19e-01 & 2.64e-01 & 3.97e-01 & 5.11e-01 & 5.96e-01 & 6.62e-01 & 7.33e-01 & 8.05e-01 & 8.63e-01 & 9.09e-01 & 9.44e-01 & 9.84e-01 & 1.02e+00 & 1.03e+00 + in & -1.88e-04 & 7.72e-02 & 1.82e-01 & 2.85e-01 & 3.77e-01 & 4.47e-01 & 5.02e-01 & 5.61e-01 & 6.24e-01 & 6.75e-01 & 7.16e-01 & 7.46e-01 & 7.83e-01 & 8.12e-01 & 8.23e-01 + sn & 2.64e-05 & 1.25e-01 & 2.85e-01 & 4.33e-01 & 5.65e-01 & 6.62e-01 & 7.36e-01 & 8.14e-01 & 8.95e-01 & 9.59e-01 & 1.01e+00 & 1.05e+00 & 1.09e+00 & 1.12e+00 & 1.13e+00 + sb & -2.49e-04 & 5.89e-02 & 1.58e-01 & 2.67e-01 & 3.82e-01 & 4.71e-01 & 5.44e-01 & 6.21e-01 & 7.02e-01 & 7.68e-01 & 8.19e-01 & 8.56e-01 & 8.99e-01 & 9.34e-01 & 9.46e-01 + te & -3.36e-05 & 4.12e-02 & 1.04e-01 & 1.74e-01 & 2.50e-01 & 3.11e-01 & 3.63e-01 & 4.19e-01 & 4.81e-01 & 5.33e-01 & 5.74e-01 & 6.05e-01 & 6.41e-01 & 6.71e-01 & 6.80e-01 + i & -4.12e-04 & 1.30e-02 & 3.52e-02 & 6.10e-02 & 9.17e-02 & 1.17e-01 & 1.40e-01 & 1.67e-01 & 1.98e-01 & 2.27e-01 & 2.51e-01 & 2.69e-01 & 2.91e-01 & 3.11e-01 & 3.18e-01 + xe & 1.25e-04 & 5.94e-02 & 1.49e-01 & 2.42e-01 & 3.41e-01 & 4.15e-01 & 4.77e-01 & 5.42e-01 & 6.12e-01 & 6.73e-01 & 7.21e-01 & 7.57e-01 & 7.97e-01 & 8.31e-01 & 8.42e-01 + cs & -1.19e-04 & 7.15e-02 & 1.95e-01 & 3.20e-01 & 4.50e-01 & 5.42e-01 & 6.15e-01 & 6.90e-01 & 7.67e-01 & 8.32e-01 & 8.81e-01 & 9.17e-01 & 9.59e-01 & 9.93e-01 & 1.00e+00 + ba & 6.13e-05 & 2.58e-01 & 5.33e-01 & 7.36e-01 & 9.21e-01 & 1.04e+00 & 1.13e+00 & 1.21e+00 & 1.29e+00 & 1.36e+00 & 1.41e+00 & 1.45e+00 & 1.50e+00 & 1.54e+00 & 1.55e+00 + la & -4.48e-06 & 2.65e-01 & 5.44e-01 & 7.49e-01 & 9.37e-01 & 1.05e+00 & 1.15e+00 & 1.23e+00 & 1.31e+00 & 1.38e+00 & 1.43e+00 & 1.47e+00 & 1.52e+00 & 1.56e+00 & 1.57e+00 + ce & 2.41e-06 & 3.17e-01 & 6.26e-01 & 8.43e-01 & 1.04e+00 & 1.16e+00 & 1.26e+00 & 1.34e+00 & 1.42e+00 & 1.49e+00 & 1.54e+00 & 1.59e+00 & 1.63e+00 & 1.67e+00 & 1.69e+00 + pr & -1.07e-04 & 2.25e-01 & 4.98e-01 & 7.08e-01 & 9.01e-01 & 1.02e+00 & 1.12e+00 & 1.21e+00 & 1.29e+00 & 1.35e+00 & 1.40e+00 & 1.45e+00 & 1.50e+00 & 1.54e+00 & 1.55e+00 + nd & -4.98e-06 & 2.58e-01 & 5.38e-01 & 7.44e-01 & 9.27e-01 & 1.04e+00 & 1.14e+00 & 1.23e+00 & 1.30e+00 & 1.37e+00 & 1.42e+00 & 1.47e+00 & 1.52e+00 & 1.56e+00 & 1.57e+00 + sm & -1.09e-04 & 1.64e-01 & 3.78e-01 & 5.63e-01 & 7.40e-01 & 8.61e-01 & 9.64e-01 & 1.05e+00 & 1.13e+00 & 1.20e+00 & 1.26e+00 & 1.30e+00 & 1.35e+00 & 1.39e+00 & 1.41e+00 + eu & -4.44e-04 & 3.36e-02 & 9.75e-02 & 1.73e-01 & 2.66e-01 & 3.40e-01 & 4.09e-01 & 4.72e-01 & 5.34e-01 & 5.87e-01 & 6.33e-01 & 6.72e-01 & 7.15e-01 & 7.51e-01 & 7.65e-01 + gd & 6.74e-06 & 8.95e-02 & 2.40e-01 & 3.91e-01 & 5.51e-01 & 6.69e-01 & 7.71e-01 & 8.59e-01 & 9.40e-01 & 1.01e+00 & 1.06e+00 & 1.11e+00 & 1.16e+00 & 1.21e+00 & 1.22e+00 + tb & -2.43e-04 & 4.84e-02 & 1.37e-01 & 2.38e-01 & 3.56e-01 & 4.49e-01 & 5.34e-01 & 6.10e-01 & 6.81e-01 & 7.42e-01 & 7.94e-01 & 8.39e-01 & 8.85e-01 & 9.24e-01 & 9.39e-01 + dy & -6.82e-05 & 8.11e-02 & 2.08e-01 & 3.41e-01 & 4.90e-01 & 5.98e-01 & 6.98e-01 & 7.85e-01 & 8.67e-01 & 9.35e-01 & 9.91e-01 & 1.04e+00 & 1.09e+00 & 1.13e+00 & 1.15e+00 + ho & 1.65e-05 & 4.87e-02 & 1.46e-01 & 2.44e-01 & 3.30e-01 & 4.43e-01 & 5.27e-01 & 5.96e-01 & 6.53e-01 & 7.15e-01 & 7.72e-01 & 8.17e-01 & 8.52e-01 & 8.78e-01 & 8.89e-01 + er & 1.89e-04 & 1.05e-01 & 2.57e-01 & 4.03e-01 & 5.55e-01 & 6.71e-01 & 7.74e-01 & 8.64e-01 & 9.47e-01 & 1.02e+00 & 1.08e+00 & 1.13e+00 & 1.18e+00 & 1.22e+00 & 1.23e+00 + tm & 3.42e-05 & 8.02e-02 & 1.97e-01 & 3.15e-01 & 4.45e-01 & 5.49e-01 & 6.44e-01 & 7.29e-01 & 8.09e-01 & 8.77e-01 & 9.34e-01 & 9.85e-01 & 1.03e+00 & 1.08e+00 & 1.09e+00 + yb & 3.10e-04 & 2.18e-01 & 4.64e-01 & 6.58e-01 & 8.37e-01 & 9.65e-01 & 1.08e+00 & 1.17e+00 & 1.26e+00 & 1.33e+00 & 1.39e+00 & 1.44e+00 & 1.50e+00 & 1.54e+00 & 1.55e+00 + lu & -1.11e-03 & 1.25e-01 & 2.95e-01 & 4.44e-01 & 5.91e-01 & 7.00e-01 & 7.97e-01 & 8.81e-01 & 9.60e-01 & 1.03e+00 & 1.09e+00 & 1.14e+00 & 1.19e+00 & 1.23e+00 & 1.24e+00 + hf & 4.02e-04 & 3.03e-01 & 6.09e-01 & 8.28e-01 & 1.02e+00 & 1.16e+00 & 1.27e+00 & 1.36e+00 & 1.45e+00 & 1.52e+00 & 1.58e+00 & 1.64e+00 & 1.69e+00 & 1.73e+00 & 1.75e+00 + ta & -8.80e-05 & 2.41e-01 & 5.11e-01 & 7.12e-01 & 8.94e-01 & 1.02e+00 & 1.13e+00 & 1.22e+00 & 1.31e+00 & 1.38e+00 & 1.44e+00 & 1.49e+00 & 1.54e+00 & 1.59e+00 & 1.60e+00 + w & -1.47e-05 & 2.92e-01 & 5.88e-01 & 7.99e-01 & 9.81e-01 & 1.11e+00 & 1.22e+00 & 1.31e+00 & 1.39e+00 & 1.46e+00 & 1.52e+00 & 1.57e+00 & 1.63e+00 & 1.67e+00 & 1.68e+00 + re & -7.45e-02 & 5.86e-02 & 2.49e-01 & 4.11e-01 & 5.65e-01 & 6.78e-01 & 7.79e-01 & 8.64e-01 & 9.44e-01 & 1.01e+00 & 1.07e+00 & 1.12e+00 & 1.17e+00 & 1.21e+00 & 1.23e+00 + os & 4.98e-03 & 6.95e-02 & 1.78e-01 & 2.86e-01 & 4.01e-01 & 4.91e-01 & 5.76e-01 & 6.50e-01 & 7.20e-01 & 7.81e-01 & 8.34e-01 & 8.81e-01 & 9.28e-01 & 9.67e-01 & 9.81e-01 + ir & -4.27e-04 & 9.94e-03 & 3.32e-02 & 6.32e-02 & 1.02e-01 & 1.39e-01 & 1.77e-01 & 2.15e-01 & 2.54e-01 & 2.90e-01 & 3.23e-01 & 3.54e-01 & 3.85e-01 & 4.13e-01 & 4.23e-01 + pt & 1.81e-04 & 4.45e-02 & 1.25e-01 & 2.15e-01 & 3.17e-01 & 4.03e-01 & 4.85e-01 & 5.59e-01 & 6.29e-01 & 6.90e-01 & 7.43e-01 & 7.91e-01 & 8.38e-01 & 8.77e-01 & 8.90e-01 + au & -9.42e-06 & 4.07e-02 & 1.10e-01 & 1.85e-01 & 2.68e-01 & 3.39e-01 & 4.08e-01 & 4.71e-01 & 5.33e-01 & 5.86e-01 & 6.34e-01 & 6.78e-01 & 7.21e-01 & 7.58e-01 & 7.70e-01 + hg & 6.84e-04 & 3.19e-01 & 6.27e-01 & 8.43e-01 & 1.02e+00 & 1.14e+00 & 1.24e+00 & 1.32e+00 & 1.40e+00 & 1.46e+00 & 1.52e+00 & 1.57e+00 & 1.62e+00 & 1.66e+00 & 1.67e+00 + tl & -6.60e-06 & 3.96e-01 & 7.25e-01 & 9.47e-01 & 1.13e+00 & 1.25e+00 & 1.36e+00 & 1.44e+00 & 1.51e+00 & 1.57e+00 & 1.63e+00 & 1.67e+00 & 1.72e+00 & 1.76e+00 & 1.78e+00 + pb & 9.70e-06 & 1.46e+00 & 1.95e+00 & 2.21e+00 & 2.37e+00 & 2.49e+00 & 2.58e+00 & 2.64e+00 & 2.69e+00 & 2.74e+00 & 2.78e+00 & 2.81e+00 & 2.84e+00 & 2.87e+00 & 2.88e+00 + bi & -5.59e-07 & 9.88e-01 & 1.50e+00 & 1.80e+00 & 2.00e+00 & 2.14e+00 & 2.25e+00 & 2.33e+00 & 2.40e+00 & 2.46e+00 & 2.50e+00 & 2.55e+00 & 2.59e+00 & 2.62e+00 & 2.63e+00 + [ tab11_el ] c & 5.75e-01 & 1.17e+00 & 1.54e+00 & 2.18e+00 & 3.00e+00 + n & 3.65e-01 & 3.97e-01 & 4.38e-01 & 5.56e-01 & 8.58e-01 + o & 1.36e-03 & 2.90e-02 & 8.97e-02 & 3.22e-01 & 1.01e+00 + f & 4.88e-01 & 1.06e+00 & 1.47e+00 & 2.22e+00 & 2.90e+00 + ne & 2.02e-01 & 5.58e-01 & 8.81e-01 & 1.60e+00 & 2.27e+00 + na & 1.72e-01 & 3.58e-01 & 5.70e-01 & 1.28e+00 & 1.77e+00 + mg & 1.06e-02 & 6.77e-02 & 1.70e-01 & 7.19e-01 & 1.30e+00 + al & 1.06e-02 & 2.72e-02 & 4.41e-02 & 1.79e-01 & 6.37e-01 + si & 1.12e-03 & 2.76e-03 & 4.71e-03 & 1.63e-02 & 6.61e-02 + p & 2.35e-02 & 5.85e-02 & 9.02e-02 & 1.96e-01 & 3.67e-01 + s & 4.82e-04 & 9.58e-04 & 1.20e-03 & 2.38e-03 & 4.65e-03 + cl & 1.42e-02 & 2.62e-02 & 3.02e-02 & 4.37e-02 & 1.42e-02 + ar & -5.42e-04 & -1.65e-03 & -2.91e-03 & -6.41e-03 & 4.38e-03 + k & 9.53e-03 & 2.18e-02 & 3.08e-02 & 6.09e-02 & 3.14e-02 + ca & -6.66e-04 & -2.07e-03 & -3.88e-03 & -9.51e-03 & -7.16e-03 + sc & 4.53e-02 & 1.14e-01 & 1.70e-01 & 3.35e-01 & 2.75e-01 + ti & 8.03e-03 & 2.22e-02 & 3.11e-02 & 4.42e-02 & 2.15e-02 + v & 2.11e-03 & 6.60e-03 & 6.48e-03 & -2.17e-03 & -7.71e-03 + cr & 2.10e-04 & 8.20e-04 & 1.59e-03 & 3.46e-03 & 2.01e-03 + mn & 5.72e-03 & 1.65e-02 & 1.80e-02 & 6.94e-03 & -4.93e-03 + fe & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 & 0.00e+00 + co & 5.67e-02 & 1.34e-01 & 2.22e-01 & 4.90e-01 & 4.60e-01 + ni & 2.74e-03 & 3.35e-03 & 7.26e-03 & 3.35e-02 & 3.04e-02 + cu & 8.12e-02 & 1.47e-01 & 2.47e-01 & 5.83e-01 & 5.57e-01 + zn & 4.41e-02 & 6.43e-02 & 8.41e-02 & 2.02e-01 & 1.57e-01 + ga & 2.03e-01 & 2.50e-01 & 2.58e-01 & 4.75e-01 & 4.22e-01 + ge & 2.04e-01 & 2.57e-01 & 2.52e-01 & 4.35e-01 & 3.87e-01 + as & 1.46e-01 & 1.87e-01 & 1.75e-01 & 3.09e-01 & 2.46e-01 + se & 2.69e-01 & 3.34e-01 & 3.11e-01 & 4.73e-01 & 3.95e-01 + br & 2.07e-01 & 2.68e-01 & 2.43e-01 & 3.61e-01 & 2.74e-01 + kr & 3.60e-01 & 4.74e-01 & 4.64e-01 & 6.11e-01 & 5.14e-01 + rb & 3.62e-01 & 5.57e-01 & 6.64e-01 & 9.74e-01 & 8.63e-01 + sr & 1.11e+00 & 1.33e+00 & 1.26e+00 & 1.26e+00 & 1.01e+00 + y & 1.08e+00 & 1.36e+00 & 1.33e+00 & 1.37e+00 & 1.12e+00 + zr & 1.07e+00 & 1.39e+00 & 1.38e+00 & 1.41e+00 & 1.17e+00 + nb & 4.61e-01 & 8.19e-01 & 8.03e-01 & 7.77e-01 & 5.30e-01 + mo & 8.67e-01 & 1.22e+00 & 1.19e+00 & 1.16e+00 & 9.47e-01 + ru & 7.62e-01 & 1.12e+00 & 1.09e+00 & 1.00e+00 & 6.96e-01 + rh & 4.96e-01 & 8.13e-01 & 7.81e-01 & 7.08e-01 & 4.50e-01 + pd & 9.08e-01 & 1.29e+00 & 1.26e+00 & 1.20e+00 & 8.86e-01 + ag & 4.60e-01 & 7.79e-01 & 7.48e-01 & 6.81e-01 & 4.26e-01 + cd & 9.87e-01 & 1.41e+00 & 1.40e+00 & 1.35e+00 & 1.03e+00 + in & 8.04e-01 & 1.22e+00 & 1.21e+00 & 1.14e+00 & 8.23e-01 + sn & 1.00e+00 & 1.47e+00 & 1.48e+00 & 1.47e+00 & 1.13e+00 + sb & 6.19e-01 & 1.11e+00 & 1.17e+00 & 1.21e+00 & 9.46e-01 + te & 5.70e-01 & 1.01e+00 & 1.00e+00 & 9.58e-01 & 6.80e-01 + i & 2.43e-01 & 5.72e-01 & 5.75e-01 & 5.08e-01 & 3.18e-01 + xe & 5.61e-01 & 1.06e+00 & 1.07e+00 & 1.05e+00 & 8.42e-01 + cs & 6.06e-01 & 1.25e+00 & 1.31e+00 & 1.32e+00 & 1.00e+00 + ba & 1.10e+00 & 1.87e+00 & 2.01e+00 & 1.99e+00 & 1.55e+00 + la & 1.02e+00 & 1.88e+00 & 2.05e+00 & 2.03e+00 & 1.57e+00 + ce & 9.51e-01 & 1.91e+00 & 2.15e+00 & 2.15e+00 & 1.69e+00 + pr & 7.14e-01 & 1.69e+00 & 1.98e+00 & 2.02e+00 & 1.55e+00 + nd & 7.69e-01 & 1.76e+00 & 2.04e+00 & 2.04e+00 & 1.57e+00 + sm & 5.30e-01 & 1.49e+00 & 1.80e+00 & 1.86e+00 & 1.41e+00 + eu & 1.43e-01 & 8.04e-01 & 1.10e+00 & 1.18e+00 & 7.65e-01 + gd & 3.27e-01 & 1.19e+00 & 1.53e+00 & 1.65e+00 & 1.22e+00 + tb & 2.18e-01 & 9.59e-01 & 1.26e+00 & 1.36e+00 & 9.39e-01 + dy & 3.43e-01 & 1.20e+00 & 1.49e+00 & 1.58e+00 & 1.15e+00 + ho & 1.90e-01 & 9.08e-01 & 1.20e+00 & 1.28e+00 & 8.89e-01 + er & 3.60e-01 & 1.25e+00 & 1.57e+00 & 1.64e+00 & 1.23e+00 + tm & 3.30e-01 & 1.19e+00 & 1.47e+00 & 1.51e+00 & 1.09e+00 + yb & 5.80e-01 & 1.59e+00 & 1.93e+00 & 1.99e+00 & 1.55e+00 + lu & 4.04e-01 & 1.34e+00 & 1.66e+00 & 1.70e+00 & 1.24e+00 + hf & 6.71e-01 & 1.72e+00 & 2.08e+00 & 2.16e+00 & 1.75e+00 + ta & 5.96e-01 & 1.64e+00 & 1.99e+00 & 2.04e+00 & 1.60e+00 + w & 6.85e-01 & 1.76e+00 & 2.13e+00 & 2.16e+00 & 1.68e+00 + re & 2.62e-01 & 1.25e+00 & 1.65e+00 & 1.69e+00 & 1.23e+00 + os & 2.02e-01 & 9.75e-01 & 1.33e+00 & 1.41e+00 & 9.81e-01 + ir & 3.12e-02 & 3.89e-01 & 6.73e-01 & 7.59e-01 & 4.23e-01 + pt & 1.48e-01 & 8.28e-01 & 1.18e+00 & 1.29e+00 & 8.90e-01 + au & 1.27e-01 & 7.91e-01 & 1.14e+00 & 1.17e+00 & 7.70e-01 + hg & 6.02e-01 & 1.66e+00 & 2.08e+00 & 2.12e+00 & 1.67e+00 + tl & 7.26e-01 & 1.81e+00 & 2.27e+00 & 2.30e+00 & 1.78e+00 + pb & 5.43e-01 & 1.59e+00 & 2.32e+00 & 2.87e+00 & 2.88e+00 + bi & 5.95e-02 & 6.58e-01 & 1.50e+00 & 2.46e+00 & 2.63e+00 + [ tab12_el ] cccccc @xmath188h & -4.06e-02 & -7.05e-02 & -8.21e-02 & -1.16e-01 & -9.92e-02 + @xmath189h & -3.77e-05 & -3.79e-05 & -3.86e-05 & -5.42e-05 & -5.39e-05 + @xmath190he & 3.77e-04 & 3.35e-04 & 3.03e-04 & 2.59e-04 & 1.83e-04 + @xmath191he & 3.23e-02 & 5.14e-02 & 6.31e-02 & 8.85e-02 & 7.99e-02 + @xmath203li & -9.09e-10 & -3.87e-10 & -1.97e-10 & -6.38e-11 & -6.80e-12 + @xmath204li & -1.29e-08 & -5.43e-09 & -2.70e-09 & -7.86e-10 & -1.60e-11 + @xmath204be & 2.39e-17 & 6.06e-17 & 4.60e-16 & 1.47e-16 & 8.07e-17 + @xmath205be & -2.39e-10 & -1.02e-10 & -5.22e-11 & -1.69e-11 & -1.75e-12 + @xmath108b & -1.20e-09 & -5.15e-10 & -2.63e-10 & -8.59e-11 & -8.99e-12 + @xmath107b & -5.20e-09 & -2.23e-09 & -1.15e-09 & -3.75e-10 & -3.94e-11 + @xmath26c & 5.84e-03 & 1.28e-02 & 1.64e-02 & 2.23e-02 & 1.71e-02 + @xmath3c & 5.16e-05 & 2.03e-05 & 1.01e-05 & 3.63e-06 & 6.97e-07 + @xmath36c & 2.07e-10 & 1.02e-09 & 1.65e-09 & 3.83e-09 & 2.80e-09 + @xmath36n & 1.31e-03 & 6.03e-04 & 3.45e-04 & 1.52e-04 & 3.40e-05 + @xmath91n & -2.03e-06 & -9.21e-07 & -4.88e-07 & -1.65e-07 & -2.07e-08 + @xmath4o & -1.04e-05 & 1.70e-04 & 2.98e-04 & 4.47e-04 & 4.03e-04 + @xmath192o & 3.32e-05 & 2.33e-05 & 1.68e-05 & 7.09e-06 & 9.33e-07 + @xmath38o & -5.65e-06 & -2.78e-06 & -1.53e-06 & -5.08e-07 & -3.20e-08 + @xmath133f & 8.17e-07 & 1.61e-06 & 2.23e-06 & 3.97e-06 & 2.44e-06 + @xmath193ne & -2.34e-06 & 1.33e-06 & 6.48e-06 & 2.80e-05 & 1.17e-05 + @xmath194ne & 8.15e-08 & 2.56e-07 & 3.94e-07 & 1.00e-06 & 5.54e-07 + @xmath14ne & 7.06e-04 & 1.21e-03 & 1.54e-03 & 2.69e-03 & 1.41e-03 + @xmath53na & 1.95e-05 & 1.77e-05 & 1.77e-05 & 3.25e-05 & 1.38e-05 + @xmath195 mg & 9.94e-06 & 2.40e-05 & 3.32e-05 & 7.33e-05 & 2.64e-05 + @xmath16 mg & 7.10e-07 & 6.21e-06 & 1.28e-05 & 4.36e-05 & 2.55e-05 + @xmath143 mg & 3.16e-06 & 7.11e-06 & 8.17e-06 & 3.14e-05 & 3.21e-05 + @xmath143al & 3.06e-07 & 1.24e-07 & 6.62e-08 & 5.40e-08 & 3.18e-08 + @xmath198al & 1.00e-06 & 1.38e-06 & 1.25e-06 & 1.88e-06 & 1.43e-06 + @xmath206si & -1.01e-07 & -1.27e-07 & -1.08e-07 & 2.48e-07 & 6.81e-07 + @xmath207si & 8.53e-08 & 1.49e-07 & 1.69e-07 & 1.97e-07 & 6.12e-08 + @xmath208si & 5.24e-07 & 5.92e-07 & 5.88e-07 & 5.99e-07 & 9.89e-08 + @xmath209si & 4.12e-16 & 4.20e-15 & 1.24e-14 & 2.88e-14 & 3.42e-14 + @xmath210p & 3.53e-07 & 3.97e-07 & 3.23e-07 & 2.42e-07 & 8.53e-08 + @xmath209p & 1.22e-19 & 1.24e-18 & 3.68e-18 & 8.51e-18 & 1.01e-17 + @xmath209s & -6.76e-07 & -7.54e-07 & -6.74e-07 & -5.54e-07 & -2.33e-08 + @xmath211s & 8.49e-08 & 1.06e-07 & 7.94e-08 & 3.95e-08 & 3.63e-09 + @xmath212s & 2.53e-07 & 2.94e-07 & 2.84e-07 & 2.49e-07 & 1.31e-08 + @xmath144s & 2.05e-08 & 2.17e-08 & 1.92e-08 & 1.56e-08 & 1.67e-09 + @xmath199cl & -2.00e-08 & -2.00e-08 & -1.60e-08 & -1.08e-08 & -1.00e-09 + @xmath144cl & 1.38e-09 & 1.42e-09 & 9.66e-10 & 4.34e-10 & 3.95e-11 + @xmath213cl & 1.82e-07 & 1.45e-07 & 8.83e-08 & 4.05e-08 & 1.80e-09 + @xmath144ar & -4.64e-07 & -4.73e-07 & -3.92e-07 & -2.91e-07 & -1.07e-08 + @xmath214ar & 1.28e-07 & 1.25e-07 & 1.07e-07 & 8.60e-08 & 9.35e-09 + @xmath215ar & 3.59e-12 & 9.41e-12 & 1.10e-11 & 7.74e-12 & 9.88e-13 + @xmath200ar & 2.04e-08 & 2.10e-08 & 1.69e-08 & 1.28e-08 & 7.03e-10 + @xmath215k & 3.14e-08 & 2.50e-08 & 1.78e-08 & 1.40e-08 & 4.58e-10 + @xmath200k & 7.10e-09 & 9.23e-09 & 6.93e-09 & 3.53e-09 & 2.14e-10 + @xmath145k & 3.87e-08 & 3.94e-08 & 2.78e-08 & 1.49e-08 & 1.13e-09 + @xmath200ca & -2.85e-07 & -2.91e-07 & -2.41e-07 & -1.80e-07 & -1.68e-08 + @xmath145ca & 3.90e-09 & 3.38e-09 & 2.05e-09 & 8.58e-10 & 8.14e-11 + @xmath216ca & 2.43e-08 & 2.29e-08 & 1.79e-08 & 1.25e-08 & 1.10e-09 + @xmath217ca & 4.59e-09 & 4.30e-09 & 3.59e-09 & 3.00e-09 & 2.82e-10 + @xmath218ca & 2.20e-08 & 1.78e-08 & 1.00e-08 & 4.89e-09 & 3.65e-10 + @xmath219ca & 1.24e-09 & 1.24e-09 & 1.38e-09 & 1.87e-09 & 1.82e-10 + @xmath220ca & -1.60e-10 & -1.74e-10 & -1.41e-10 & -9.51e-11 & -8.90e-12 + @xmath221sc & 4.41e-09 & 4.89e-09 & 3.97e-09 & 2.97e-09 & 2.65e-10 + @xmath219ti & 7.54e-09 & 6.39e-09 & 3.35e-09 & 8.34e-10 & 3.63e-11 + @xmath222ti & -1.97e-10 & -3.36e-10 & -4.36e-10 & -4.38e-10 & -5.03e-11 + @xmath220ti & -1.70e-08 & -1.93e-08 & -1.70e-08 & -1.23e-08 & -1.10e-09 + @xmath223ti & 1.89e-08 & 2.20e-08 & 1.65e-08 & 8.60e-09 & 6.05e-10 + @xmath224ti & 4.42e-08 & 5.85e-08 & 4.61e-08 & 2.29e-08 & 1.45e-09 + @xmath224v & -1.30e-11 & -1.46e-11 & -1.16e-11 & -7.24e-12 & -5.98e-13 + @xmath225v & 7.92e-10 & 1.69e-09 & 8.58e-10 & -4.69e-10 & -8.96e-11 + @xmath224cr & -1.05e-08 & -1.17e-08 & -9.32e-09 & -5.78e-09 & -4.87e-10 + @xmath226cr & -5.40e-08 & -5.39e-08 & -4.57e-08 & -3.85e-08 & -3.87e-09 + @xmath227cr & -1.00e-08 & -7.17e-09 & -4.28e-09 & -2.74e-09 & -3.08e-10 + @xmath228cr & 4.85e-08 & 5.44e-08 & 4.87e-08 & 3.80e-08 & 3.51e-09 + @xmath229mn & 1.28e-07 & 1.69e-07 & 9.90e-08 & 1.90e-09 & -2.31e-09 + @xmath228fe & -7.89e-07 & -8.52e-07 & -6.95e-07 & -4.67e-07 & -3.98e-08 + @xmath229fe & 1.19e-12 & 1.62e-12 & 2.47e-12 & 3.97e-13 & 2.02e-14 + @xmath151fe & -5.39e-06 & -5.92e-06 & -5.15e-06 & -4.07e-06 & -3.78e-07 + @xmath230fe & 1.67e-06 & 2.24e-06 & 1.89e-06 & 1.13e-06 & 8.31e-08 + @xmath231fe & 1.95e-06 & 2.09e-06 & 2.08e-06 & 1.98e-06 & 1.92e-07 + @xmath17fe & 5.76e-08 & 5.04e-08 & 6.35e-08 & 1.27e-07 & 1.89e-08 + @xmath150co & 5.17e-07 & 5.17e-07 & 4.64e-07 & 4.55e-07 & 4.80e-08 + @xmath17co & 2.02e-13 & 1.77e-13 & 2.23e-13 & 4.46e-13 & 6.63e-14 + @xmath231ni & -5.46e-07 & -6.19e-07 & -5.16e-07 & -3.47e-07 & -3.02e-08 + @xmath150ni & 2.34e-08 & 2.60e-08 & 1.57e-08 & 4.77e-09 & 2.84e-10 + @xmath17ni & 2.23e-07 & 2.13e-07 & 2.15e-07 & 2.01e-07 & 1.74e-08 + @xmath232ni & 1.66e-07 & 1.70e-07 & 1.51e-07 & 1.25e-07 & 1.20e-08 + @xmath233ni & 2.86e-07 & 2.40e-07 & 2.21e-07 & 1.98e-07 & 1.90e-08 + @xmath234ni & 1.06e-12 & 3.29e-12 & 5.91e-12 & 6.71e-12 & 1.43e-12 + @xmath235ni & 2.75e-07 & 1.91e-07 & 1.48e-07 & 1.57e-07 & 1.56e-08 + @xmath234cu & 1.10e-07 & 1.01e-07 & 1.05e-07 & 1.23e-07 & 1.38e-08 + @xmath236cu & 8.27e-08 & 5.88e-08 & 4.28e-08 & 3.66e-08 & 2.95e-09 + @xmath235zn & 3.36e-08 & 1.88e-08 & 1.29e-08 & 5.32e-09 & -2.42e-10 + @xmath237zn & 8.34e-08 & 5.60e-08 & 4.03e-08 & 3.56e-08 & 2.89e-09 + @xmath238zn & 1.87e-08 & 1.28e-08 & 9.17e-09 & 7.98e-09 & 6.71e-10 + @xmath239zn & 1.03e-07 & 6.50e-08 & 4.23e-08 & 3.70e-08 & 3.45e-09 + @xmath240zn & -8.83e-11 & -8.92e-11 & -4.21e-11 & 1.86e-10 & 4.12e-11 + @xmath241ga & 1.92e-08 & 1.10e-08 & 6.48e-09 & 5.18e-09 & 4.83e-10 + @xmath242ga & 2.13e-08 & 1.18e-08 & 5.76e-09 & 3.64e-09 & 3.21e-10 + @xmath240ge & 3.50e-08 & 1.97e-08 & 1.05e-08 & 7.42e-09 & 6.77e-10 + @xmath243ge & 3.91e-08 & 2.27e-08 & 1.14e-08 & 7.22e-09 & 6.27e-10 + @xmath244ge & 1.09e-08 & 6.52e-09 & 3.31e-09 & 2.12e-09 & 1.84e-10 + @xmath245ge & 5.39e-08 & 3.11e-08 & 1.50e-08 & 9.21e-09 & 8.05e-10 + @xmath246ge & -1.45e-10 & -1.38e-10 & -7.56e-11 & 1.41e-10 & 2.98e-11 + @xmath247as & 4.80e-09 & 2.75e-09 & 1.30e-09 & 8.10e-10 & 7.06e-11 + @xmath246se & 1.94e-08 & 1.11e-08 & 5.20e-09 & 3.04e-09 & 2.66e-10 + @xmath248se & 7.78e-09 & 4.34e-09 & 1.99e-09 & 1.17e-09 & 1.04e-10 + @xmath249se & 2.94e-08 & 1.69e-08 & 7.66e-09 & 4.21e-09 & 3.75e-10 + @xmath250se & 3.54e-09 & 2.47e-09 & 1.24e-09 & 7.34e-10 & 9.01e-11 + @xmath251se & 5.93e-08 & 3.41e-08 & 1.57e-08 & 8.52e-09 & 7.82e-10 + @xmath252se & -2.50e-11 & -4.35e-11 & 7.66e-12 & 1.33e-10 & 1.73e-11 + @xmath250br & 4.80e-09 & 3.14e-09 & 1.57e-09 & 8.95e-10 & 5.66e-11 + @xmath253br & 1.02e-08 & 5.60e-09 & 2.36e-09 & 1.16e-09 & 1.04e-10 + @xmath251kr & 5.36e-09 & 1.90e-09 & 4.69e-10 & 9.30e-11 & 6.24e-12 + @xmath253kr & 8.75e-10 & 2.29e-10 & 3.67e-11 & 7.03e-12 & 8.97e-13 + @xmath252kr & 3.87e-08 & 2.28e-08 & 9.51e-09 & 3.91e-09 & 3.30e-10 + @xmath254kr & 1.27e-08 & 7.77e-09 & 3.32e-09 & 1.36e-09 & 1.12e-10 + @xmath181kr & 8.07e-08 & 5.01e-08 & 2.28e-08 & 1.04e-08 & 8.73e-10 + @xmath122kr & 1.91e-08 & 2.03e-08 & 1.47e-08 & 8.31e-09 & 8.22e-10 + @xmath121rb & 1.42e-08 & 9.85e-09 & 5.32e-09 & 3.19e-09 & 3.04e-10 + @xmath126rb & 5.99e-09 & 6.58e-09 & 6.05e-09 & 4.71e-09 & 4.60e-10 + @xmath122sr & 4.11e-08 & 2.25e-08 & 8.13e-09 & 2.28e-09 & 1.70e-10 + @xmath126sr & 2.67e-08 & 1.53e-08 & 5.61e-09 & 1.39e-09 & 8.75e-11 + @xmath255sr & 5.78e-07 & 4.46e-07 & 1.94e-07 & 5.64e-08 & 3.72e-09 + @xmath196y & 1.18e-07 & 1.01e-07 & 4.79e-08 & 1.55e-08 & 1.08e-09 + @xmath256zr & 1.29e-07 & 1.14e-07 & 5.07e-08 & 1.34e-08 & 8.43e-10 + @xmath257zr & 3.12e-08 & 3.12e-08 & 1.56e-08 & 4.79e-09 & 3.06e-10 + @xmath258zr & 5.12e-08 & 5.09e-08 & 2.55e-08 & 8.42e-09 & 5.64e-10 + @xmath259zr & 1.39e-08 & 1.30e-08 & 6.44e-09 & 2.21e-09 & 1.63e-10 + @xmath183zr & 7.15e-08 & 7.38e-08 & 3.51e-08 & 1.03e-08 & 6.84e-10 + @xmath182zr & 1.89e-09 & 4.66e-09 & 5.41e-09 & 4.52e-09 & 4.13e-10 + @xmath259nb & 3.57e-09 & 4.13e-09 & 2.01e-09 & 5.95e-10 & 2.82e-11 + @xmath258mo & -1.60e-11 & -1.80e-11 & -1.40e-11 & -8.27e-12 & -6.07e-13 + @xmath183mo & 5.69e-10 & 4.40e-10 & 1.18e-10 & 3.01e-12 & -2.69e-13 + @xmath123mo & 6.99e-09 & 7.93e-09 & 4.13e-09 & 1.32e-09 & 8.48e-11 + @xmath182mo & 1.49e-08 & 1.52e-08 & 6.88e-09 & 1.87e-09 & 1.25e-10 + @xmath260mo & 5.16e-09 & 5.27e-09 & 2.38e-09 & 6.90e-10 & 4.96e-11 + @xmath261mo & 1.86e-08 & 1.97e-08 & 8.84e-09 & 2.07e-09 & 1.33e-10 + @xmath262mo & 1.62e-10 & 6.08e-10 & 6.24e-10 & 3.01e-10 & 1.72e-11 + @xmath263tc & 7.25e-10 & 6.89e-10 & 3.06e-10 & 7.92e-11 & 6.94e-12 + @xmath264ru & -6.21e-12 & -6.46e-12 & -4.49e-12 & -2.26e-12 & -1.78e-13 + @xmath265ru & -1.01e-12 & -1.30e-12 & -1.14e-12 & -7.29e-13 & -6.02e-14 + @xmath266ru & 1.83e-09 & 2.04e-09 & 9.26e-10 & 2.03e-10 & 1.06e-11 + @xmath262ru & 9.32e-09 & 9.84e-09 & 4.39e-09 & 9.48e-10 & 5.43e-11 + @xmath267ru & 2.01e-09 & 2.16e-09 & 9.98e-10 & 2.30e-10 & 1.28e-11 + @xmath268ru & 1.03e-08 & 1.13e-08 & 5.22e-09 & 1.23e-09 & 6.54e-11 + @xmath269ru & 1.73e-10 & 7.03e-10 & 6.42e-10 & 2.57e-10 & 1.36e-11 + @xmath270rh & 2.14e-09 & 2.38e-09 & 1.11e-09 & 2.60e-10 & 1.36e-11 + @xmath269pd & 7.04e-09 & 7.44e-09 & 3.32e-09 & 7.58e-10 & 3.95e-11 + @xmath271pd & 1.81e-09 & 1.98e-09 & 9.26e-10 & 2.26e-10 & 1.17e-11 + @xmath272pd & 8.26e-09 & 9.16e-09 & 4.27e-09 & 1.07e-09 & 5.62e-11 + @xmath146pd & 1.56e-09 & 1.70e-09 & 7.88e-10 & 1.99e-10 & 1.08e-11 + @xmath273pd & 1.02e-08 & 1.17e-08 & 5.45e-09 & 1.33e-09 & 7.20e-11 + @xmath274pd & 1.57e-10 & 7.12e-10 & 6.98e-10 & 2.99e-10 & 1.68e-11 + @xmath146ag & 6.27e-11 & 9.60e-11 & 4.32e-11 & 6.92e-12 & -1.43e-14 + @xmath275ag & 2.64e-09 & 2.99e-09 & 1.39e-09 & 3.36e-10 & 1.82e-11 + @xmath273cd & 4.38e-11 & 2.62e-11 & 5.98e-12 & -2.82e-14 & -2.04e-14 + @xmath274cd & 8.68e-09 & 9.75e-09 & 4.42e-09 & 1.00e-09 & 5.52e-11 + @xmath276cd & 2.91e-09 & 3.36e-09 & 1.60e-09 & 3.87e-10 & 2.19e-11 + @xmath277cd & 1.07e-08 & 1.28e-08 & 6.16e-09 & 1.48e-09 & 8.43e-11 + @xmath278cd & 3.13e-09 & 3.74e-09 & 1.80e-09 & 4.30e-10 & 2.45e-11 + @xmath279cd & 1.53e-08 & 1.87e-08 & 9.13e-09 & 2.20e-09 & 1.22e-10 + @xmath280cd & 2.58e-10 & 1.15e-09 & 1.37e-09 & 8.39e-10 & 5.09e-11 + @xmath278 in & -7.30e-13 & -6.66e-13 & -4.32e-13 & -2.08e-13 & -1.61e-14 + @xmath281 in & 3.02e-09 & 3.76e-09 & 1.86e-09 & 4.52e-10 & 2.50e-11 + @xmath279sn & -1.16e-12 & -1.44e-12 & -1.13e-12 & -6.33e-13 & -5.09e-14 + @xmath281sn & -5.68e-13 & -7.01e-13 & -5.62e-13 & -3.25e-13 & -2.63e-14 + @xmath280sn & 2.10e-08 & 2.57e-08 & 1.21e-08 & 2.70e-09 & 1.42e-10 + @xmath282sn & 6.33e-09 & 7.87e-09 & 3.86e-09 & 9.84e-10 & 5.31e-11 + @xmath283sn & 2.75e-08 & 3.68e-08 & 1.87e-08 & 4.89e-09 & 2.64e-10 + @xmath284sn & 8.36e-09 & 1.11e-08 & 5.54e-09 & 1.44e-09 & 7.79e-11 + @xmath285sn & 4.27e-08 & 6.13e-08 & 3.16e-08 & 8.42e-09 & 4.57e-10 + @xmath286sn & 2.79e-10 & 1.45e-09 & 2.21e-09 & 2.20e-09 & 1.73e-10 + @xmath287sn & -3.97e-12 & 6.44e-12 & 3.90e-11 & 1.49e-10 & 2.14e-11 + @xmath288sb & 3.23e-09 & 4.68e-09 & 2.44e-09 & 6.62e-10 & 3.60e-11 + @xmath289sb & 1.86e-10 & 7.58e-10 & 6.64e-10 & 3.36e-10 & 2.21e-11 + @xmath286te & 5.49e-09 & 7.32e-09 & 3.57e-09 & 9.24e-10 & 5.03e-11 + @xmath289te & 1.99e-09 & 2.55e-09 & 1.20e-09 & 3.02e-10 & 1.65e-11 + @xmath287te & 1.08e-08 & 1.53e-08 & 7.46e-09 & 2.01e-09 & 1.18e-10 + @xmath290te & 3.64e-09 & 5.19e-09 & 2.50e-09 & 6.52e-10 & 3.86e-11 + @xmath291te & 1.98e-08 & 3.01e-08 & 1.50e-08 & 3.50e-09 & 1.92e-10 + @xmath292te & 2.18e-09 & 3.82e-09 & 1.91e-09 & 7.06e-10 & 6.07e-11 + @xmath293te & -3.95e-11 & -4.15e-11 & -3.47e-11 & -2.55e-11 & -2.30e-12 + @xmath294i & 2.53e-09 & 3.94e-09 & 1.99e-09 & 4.47e-10 & 2.35e-11 + @xmath295i & 3.35e-11 & 7.80e-11 & 4.92e-11 & 2.03e-11 & 1.75e-12 + @xmath292xe & 6.18e-09 & 9.33e-09 & 4.74e-09 & 1.09e-09 & 5.51e-11 + @xmath295xe & 2.43e-09 & 3.81e-09 & 1.91e-09 & 4.33e-10 & 2.30e-11 + @xmath293xe & 1.32e-08 & 2.07e-08 & 1.04e-08 & 2.37e-09 & 1.18e-10 + @xmath296xe & 3.56e-09 & 7.79e-09 & 2.94e-09 & 6.52e-10 & 4.25e-11 + @xmath185xe & 2.45e-08 & 4.16e-08 & 2.18e-08 & 5.13e-09 & 2.39e-10 + @xmath184xe & 2.28e-10 & 1.44e-09 & 2.33e-09 & 2.32e-09 & 1.69e-10 + @xmath297xe & -2.22e-12 & 1.23e-12 & 1.49e-11 & 7.25e-11 & 1.42e-11 + @xmath124cs & 3.22e-09 & 5.78e-09 & 3.12e-09 & 7.56e-10 & 3.47e-11 + @xmath298cs & 7.34e-10 & 3.34e-09 & 2.24e-09 & 8.42e-10 & 5.97e-11 + @xmath184ba & 9.02e-09 & 1.37e-08 & 6.40e-09 & 1.40e-09 & 6.57e-11 + @xmath298ba & 3.28e-09 & 5.14e-09 & 2.29e-09 & 4.97e-10 & 2.31e-11 + @xmath297ba & 2.32e-08 & 4.32e-08 & 2.23e-08 & 4.82e-09 & 2.39e-10 + @xmath299ba & 1.81e-08 & 3.57e-08 & 2.02e-08 & 5.95e-09 & 3.70e-10 + @xmath300ba & 1.38e-07 & 4.38e-07 & 3.10e-07 & 8.54e-08 & 3.80e-09 + @xmath197la & 1.56e-08 & 5.44e-08 & 3.96e-08 & 1.08e-08 & 4.80e-10 + @xmath187ce & 3.52e-08 & 1.57e-07 & 1.32e-07 & 3.44e-08 & 1.36e-09 + @xmath186ce & 9.99e-11 & 1.68e-09 & 4.10e-09 & 4.49e-09 & 2.65e-10 + @xmath125pr & 2.76e-09 & 1.40e-08 & 1.34e-08 & 4.26e-09 & 1.75e-10 + @xmath186nd & 7.89e-09 & 3.82e-08 & 3.28e-08 & 7.27e-09 & 2.70e-10 + @xmath301nd & 1.04e-09 & 5.57e-09 & 5.33e-09 & 1.66e-09 & 7.41e-11 + @xmath302nd & 3.20e-09 & 1.73e-08 & 1.71e-08 & 5.53e-09 & 2.56e-10 + @xmath303nd & 5.71e-10 & 3.17e-09 & 3.19e-09 & 1.03e-09 & 4.69e-11 + @xmath304nd & 3.00e-09 & 1.62e-08 & 1.59e-08 & 4.95e-09 & 2.16e-10 + @xmath305nd & 8.15e-11 & 1.44e-09 & 2.62e-09 & 1.23e-09 & 5.21e-11 + @xmath306nd & -3.82e-12 & 1.37e-11 & 8.29e-11 & 9.80e-11 & 4.04e-12 + @xmath302sm & -5.02e-13 & -5.68e-13 & -4.38e-13 & -2.52e-13 & -2.06e-14 + @xmath304sm & 7.07e-14 & 7.64e-14 & 3.84e-14 & 6.43e-15 & 2.42e-16 + @xmath307sm & 3.20e-10 & 1.86e-09 & 1.88e-09 & 5.84e-10 & 2.51e-11 + @xmath305sm & 9.75e-10 & 4.70e-09 & 4.05e-09 & 1.11e-09 & 4.77e-11 + @xmath308sm & 1.32e-10 & 7.46e-10 & 7.74e-10 & 2.62e-10 & 1.12e-11 + @xmath306sm & 5.94e-10 & 3.31e-09 & 3.35e-09 & 1.10e-09 & 4.72e-11 + @xmath39sm & 2.71e-15 & 1.51e-14 & 2.49e-14 & 1.06e-14 & 5.97e-16 + @xmath40sm & 4.29e-10 & 2.67e-09 & 2.84e-09 & 9.84e-10 & 4.21e-11 + @xmath309sm & 3.01e-11 & 6.48e-10 & 1.22e-09 & 5.95e-10 & 2.59e-11 + @xmath39eu & 7.69e-11 & 4.55e-10 & 4.67e-10 & 1.59e-10 & 6.71e-12 + @xmath310eu & 7.49e-11 & 4.89e-10 & 5.25e-10 & 1.84e-10 & 7.84e-12 + @xmath40gd & 1.53e-11 & 4.99e-11 & 3.59e-11 & 1.16e-11 & 5.07e-13 + @xmath309gd & 1.51e-10 & 7.84e-10 & 7.76e-10 & 2.65e-10 & 1.14e-11 + @xmath311gd & 9.43e-11 & 6.26e-10 & 7.37e-10 & 2.80e-10 & 1.21e-11 + @xmath312gd & 3.71e-10 & 2.10e-09 & 2.29e-09 & 8.58e-10 & 3.75e-11 + @xmath313gd & 1.76e-10 & 9.78e-10 & 1.05e-09 & 3.93e-10 & 1.73e-11 + @xmath314gd & 7.58e-10 & 4.11e-09 & 4.17e-09 & 1.51e-09 & 6.80e-11 + @xmath315gd & 1.60e-11 & 4.22e-10 & 8.65e-10 & 4.81e-10 & 2.23e-11 + @xmath316 tb & 1.65e-10 & 9.33e-10 & 9.42e-10 & 3.41e-10 & 1.53e-11 + @xmath315dy & 2.61e-10 & 1.37e-09 & 1.34e-09 & 4.71e-10 & 2.14e-11 + @xmath317dy & 1.14e-10 & 6.64e-10 & 6.95e-10 & 2.65e-10 & 1.21e-11 + @xmath318dy & 5.42e-10 & 2.98e-09 & 3.00e-09 & 1.09e-09 & 5.05e-11 + @xmath319dy & 1.73e-10 & 9.82e-10 & 9.69e-10 & 3.41e-10 & 1.56e-11 + @xmath320dy & 9.73e-10 & 5.46e-09 & 5.31e-09 & 1.69e-09 & 7.63e-11 + @xmath319ho & 3.97e-13 & 1.64e-12 & 2.02e-12 & 1.05e-12 & 6.58e-14 + @xmath321ho & 2.15e-10 & 1.28e-09 & 1.29e-09 & 4.35e-10 & 2.02e-11 + @xmath320er & 8.04e-11 & 4.74e-10 & 4.96e-10 & 1.91e-10 & 9.10e-12 + @xmath322er & 4.01e-10 & 2.05e-09 & 1.93e-09 & 6.17e-10 & 2.91e-11 + @xmath323er & 1.82e-10 & 9.55e-10 & 9.10e-10 & 2.93e-10 & 1.39e-11 + @xmath324er & 7.66e-10 & 4.21e-09 & 4.00e-09 & 1.20e-09 & 5.55e-11 + @xmath325er & 6.39e-11 & 1.16e-09 & 2.02e-09 & 7.77e-10 & 3.40e-11 + @xmath326tm & 1.92e-10 & 1.12e-09 & 1.08e-09 & 3.25e-10 & 1.51e-11 + @xmath325yb & 2.53e-10 & 1.35e-09 & 1.26e-09 & 3.61e-10 & 1.64e-11 + @xmath327yb & 2.57e-10 & 1.71e-09 & 1.91e-09 & 5.83e-10 & 2.56e-11 + @xmath328yb & 7.55e-10 & 4.32e-09 & 4.52e-09 & 1.38e-09 & 6.10e-11 + @xmath329yb & 3.47e-10 & 2.01e-09 & 2.09e-09 & 6.48e-10 & 2.85e-11 + @xmath330yb & 1.62e-09 & 9.76e-09 & 1.01e-08 & 3.07e-09 & 1.32e-10 + @xmath331yb & 3.71e-11 & 8.71e-10 & 1.94e-09 & 1.04e-09 & 4.44e-11 + @xmath41lu & 2.17e-10 & 1.36e-09 & 1.42e-09 & 4.35e-10 & 1.86e-11 + @xmath331lu & 3.84e-11 & 2.42e-10 & 2.52e-10 & 7.64e-11 & 3.24e-12 + @xmath331hf & 2.92e-10 & 1.73e-09 & 1.77e-09 & 5.30e-10 & 2.26e-11 + @xmath332hf & 1.67e-10 & 1.08e-09 & 1.20e-09 & 4.03e-10 & 1.72e-11 + @xmath333hf & 7.86e-10 & 4.89e-09 & 5.38e-09 & 1.80e-09 & 7.69e-11 + @xmath334hf & 2.67e-10 & 1.67e-09 & 1.83e-09 & 6.12e-10 & 2.61e-11 + @xmath335hf & 1.48e-09 & 9.38e-09 & 1.01e-08 & 3.20e-09 & 1.37e-10 + @xmath336hf & 1.59e-11 & 4.17e-10 & 1.14e-09 & 7.12e-10 & 3.01e-11 + @xmath335ta & 4.67e-14 & 3.03e-13 & 3.52e-13 & 1.31e-13 & 5.91e-15 + @xmath337ta & 2.98e-10 & 1.99e-09 & 2.19e-09 & 6.88e-10 & 2.93e-11 + @xmath335w & -2.18e-14 & -2.06e-14 & -1.35e-14 & -6.55e-15 & -5.74e-16 + @xmath336w & 7.09e-10 & 3.88e-09 & 3.73e-09 & 1.06e-09 & 4.48e-11 + @xmath338w & 4.27e-10 & 2.62e-09 & 2.76e-09 & 8.60e-10 & 3.63e-11 + @xmath339w & 1.10e-09 & 7.05e-09 & 7.55e-09 & 2.19e-09 & 9.07e-11 + @xmath340w & 1.93e-10 & 2.95e-09 & 4.95e-09 & 1.70e-09 & 6.78e-11 + @xmath341re & 2.06e-10 & 1.39e-09 & 1.51e-09 & 4.33e-10 & 1.80e-11 + @xmath342re & -6.45e-12 & 5.80e-10 & 9.89e-10 & 3.43e-10 & 1.35e-11 + @xmath340os & 4.36e-10 & 2.45e-09 & 2.45e-09 & 6.63e-10 & 2.73e-11 + @xmath342os & 1.83e-10 & 7.14e-10 & 6.59e-10 & 1.76e-10 & 7.71e-12 + @xmath343os & 5.34e-10 & 3.47e-09 & 4.19e-09 & 1.45e-09 & 5.98e-11 + @xmath344os & 1.33e-10 & 9.44e-10 & 1.15e-09 & 4.06e-10 & 1.66e-11 + @xmath345os & 7.13e-10 & 4.77e-09 & 5.62e-09 & 2.02e-09 & 8.50e-11 + @xmath346os & 3.31e-11 & 9.85e-10 & 1.64e-09 & 6.32e-10 & 2.65e-11 + @xmath347ir & 1.21e-10 & 1.01e-09 & 1.21e-09 & 4.31e-10 & 1.80e-11 + @xmath348ir & 1.23e-10 & 1.20e-09 & 1.54e-09 & 5.47e-10 & 2.31e-11 + @xmath346pt & 3.13e-10 & 1.59e-09 & 1.59e-09 & 5.29e-10 & 2.29e-11 + @xmath349pt & 8.69e-10 & 5.65e-09 & 6.85e-09 & 2.68e-09 & 1.20e-10 + @xmath350pt & 2.94e-10 & 2.21e-09 & 2.62e-09 & 9.28e-10 & 4.13e-11 + @xmath351pt & 1.37e-09 & 8.98e-09 & 1.06e-08 & 3.24e-09 & 1.43e-10 + @xmath352pt & 2.46e-12 & 2.82e-10 & 9.56e-10 & 7.60e-10 & 3.35e-11 + @xmath353au & 3.45e-10 & 2.46e-09 & 2.97e-09 & 8.86e-10 & 3.82e-11 + @xmath352hg & 1.13e-09 & 7.43e-09 & 8.73e-09 & 2.31e-09 & 9.67e-11 + @xmath354hg & 4.85e-10 & 3.43e-09 & 4.31e-09 & 1.25e-09 & 5.14e-11 + @xmath355hg & 1.64e-09 & 1.16e-08 & 1.50e-08 & 4.42e-09 & 1.70e-10 + @xmath356hg & 6.68e-10 & 4.85e-09 & 6.48e-09 & 1.95e-09 & 7.30e-11 + @xmath357hg & 2.74e-09 & 1.92e-08 & 2.55e-08 & 7.61e-09 & 2.77e-10 + @xmath201hg & 2.60e-12 & 1.32e-10 & 5.73e-10 & 7.91e-10 & 4.18e-11 + @xmath358tl & 1.21e-09 & 9.00e-09 & 1.29e-08 & 4.15e-09 & 1.51e-10 + @xmath147tl & 3.18e-09 & 2.18e-08 & 3.06e-08 & 8.74e-09 & 3.13e-10 + @xmath201pb & 2.06e-09 & 1.32e-08 & 1.68e-08 & 4.27e-09 & 1.51e-10 + @xmath147pb & 1.48e-10 & 1.05e-09 & 1.54e-09 & 3.95e-10 & 1.12e-11 + @xmath359pb & 1.05e-08 & 7.41e-08 & 1.38e-07 & 5.00e-08 & 2.17e-09 + @xmath360pb & 1.25e-08 & 8.01e-08 & 2.02e-07 & 8.74e-08 & 3.67e-09 + @xmath119pb & 1.88e-08 & 1.51e-07 & 5.50e-07 & 8.48e-07 & 1.03e-07 + @xmath361bi & 1.07e-10 & 1.21e-09 & 5.52e-09 & 1.63e-08 & 2.68e-09 + @xmath362po & 1.66e-24 & 1.21e-22 & 4.54e-21 & 1.37e-20 & 1.41e-20 + [ tab14_el ] | the envelope of thermally pulsing agb stars undergoing periodic third dredge - up episodes is enriched in both light and heavy elements , the ashes of a complex internal nucleosynthesis involving p , @xmath0 and n captures over hundreds of stable and unstable isotopes . in this paper , new models of low - mass agb stars ( 2 m@xmath1 ) , with metallicity ranging between @xmath2=0.0138 ( the solar one ) and @xmath2=0.0001 ,
are presented .
main features are : i ) a full nuclear network ( from h to bi ) coupled to the stellar evolution code , ii ) a mass loss - period - luminosity relation , based on available data for long period variables , and ii ) molecular and atomic opacities for c- and/or n - enhanced mixtures , appropriate for the chemical modifications of the envelope caused by the third dredge up . for each model a detailed description of the physical and chemical evolution is presented ; moreover , we present a uniform set of yields , comprehensive of all chemical species ( from hydrogen to bismuth ) .
the main nucleosynthesis site is the thin @xmath3c pocket , which forms in the core - envelope transition region after each third dredge up episode .
the formation of this @xmath3c pockets is the principal by - product of the introduction of a new algorithm , which shapes the velocity profile of convective elements at the inner border of the convective envelope : both the physical grounds and the calibration of the algorithm are discussed in detail .
we find that the pockets shrink ( in mass ) as the star climbs the agb , so that the first pockets , the largest ones , leave the major imprint on the overall nucleosynthesis .
neutrons are released by the @xmath3c(@xmath0,n)@xmath4o reaction during the interpulse phase in radiative conditions , when temperatures within the pockets attain @xmath5 k , with typical densities of ( @xmath6 ) neutrons @xmath7 @xmath8 .
exceptions are found , as in the case of the first pocket of the metal - rich models ( @xmath9 , @xmath10 and @xmath11 ) , where the @xmath3c is only partially burned during the interpulse : the surviving part is ingested in the convective zone generated by the subsequent thermal pulse and then burned at @xmath12 k , thus producing larger neutron densities ( up to @xmath13 neutrons @xmath7 @xmath8 ) .
an additional neutron exposure , caused by the @xmath14ne@xmath15,n)@xmath16 mg during the thermal pulses , is marginally activated at large z , but becomes an important nucleosynthesis source at low z , when most of the @xmath14ne is primary .
the final surface compositions of the various models reflect the differences in the initial iron - seed content and in the physical structure of agb stars belonging to different stellar populations .
thus , at large metallicities the nucleosynthesis of light s elements ( sr , y , zr ) is favored , whilst , decreasing the iron content , the overproduction of heavy s elements ( ba , la , ce , nd , sm ) and lead becomes progressively more important . at low metallicities ( @xmath2=0.0001 ) the main product is lead .
the agreement with the observed [ hs / ls ] index observed in intrinsic c stars at different [ fe / h ] is generally good .
for the solar metallicity model , we found an interesting overproduction of some radioactive isotopes , like @xmath17fe , as a consequence of the anomalous first @xmath3c pocket .
finally , light elements ( c , f , ne and na ) are enhanced at any metallicity . |
the cobalt arsenides @xmath6co@xmath7as@xmath7 ( @xmath6 = ca , sr , ba ) with the tetragonal thcr@xmath7si@xmath7 structure ( space group @xmath8 ) have received increased attention recently because of their close relationship to the parent compounds of the @xmath6fe@xmath7as@xmath7 superconductor family.@xcite the isostructural @xmath6fe@xmath7as@xmath7 parent compounds manifest structural and magnetic transitions at approximately the same temperature , from a high - temperature paramagnetic tetragonal phase to a low - temperature orthorhombic antiferromagnetic ( afm ) structure characterized by the magnetic propagation vector , * * q**@xmath9 = ( @xmath10 ) , the so - called stripe - like magnetic phase.@xcite various substitutions can be made for any of the three different atomic sites , which reduce the structural and magnetic transition temperatures .
doping @xmath6fe@xmath7as@xmath7 with k,@xcite co,@xcite
ni,@xcite rh,@xcite p,@xcite pd,@xcite ir,@xcite pt,@xcite or ru@xcite also induces superconductivity ( sc ) over a finite range in concentration .
for example , upon doping with co for fe in ba(fe@xmath11co@xmath12)@xmath7as@xmath7 , both the structural and afm order are suppressed to lower temperatures and split , with the structural transition found at a slightly higher temperature than the magnetic transition.@xcite sc emerges for co substitutions between @xmath13 = 0.03 and @xmath13 = 0.12 with superconducting transition temperatures ( @xmath14 ) up to @xmath1523 k. neutron diffraction studies have clearly established the coexistence of , and competition between , sc and afm order over a small range in co concentration.@xcite above @xmath13 @xmath15 0.06 , the afm order is completely suppressed , but correlated afm fluctuations are observed at * * q**@xmath9 .
co substitution levels above @xmath13 @xmath15 0.12 , lead to the suppression of both the correlated afm fluctuations and sc.@xcite cafe@xmath7as@xmath7 @xcite presents a particularly interesting case . at ambient pressure ,
the substitution of co or rh for fe @xcite results in the suppression of afm order and sc emerges with @xmath14 up to @xmath1520 k. under modest applied pressure cafe@xmath7as@xmath7 manifests fascinating new behavior including a transition to an isostructural volume - collapsed tetragonal ( ct ) phase@xcite that is nonmagnetic@xcite and nonsuperconducting.@xcite the ct phase in cafe@xmath7as@xmath7 is distinguished by a striking 9.5% reduction in the tetragonal _
c_-axis lattice parameter , with respect to the high - temperature ambient - pressure tetragonal ( t ) phase , along with the absence of the stripe - like magnetic order found for the low - temperature ambient - pressure orthorhombic phase and of the magnetic fluctuations found in the t phase . in some respects
the @xmath6co@xmath7as@xmath7 compounds appear to exhibit very different behavior as compared to the iron arsenides .
baco@xmath7as@xmath7 and srco@xmath7as@xmath7 are metals with enhanced paramagnetic susceptibilities and neither manifest magnetic order or sc down to 2 k.@xcite however , recent inelastic neutron scattering measurements have revealed that @xmath16 features afm fluctuations at the same wave vector , * * q**@xmath9 = ( @xmath10 ) , ubiquitous in the @xmath6fe@xmath7as@xmath7 parent compounds.@xcite this is surprising because , for srco@xmath7as@xmath7 , there is no clear nesting feature at this wave vector that would suggest an instabiltity toward magnetic ordering.@xcite @xmath0 is reported to have a ct @xmath1-type structure with a small @xmath17.@xcite two groups have independently reported the bulk magnetic and transport properties of single crystals of this compound.@xcite in contrast to the nonmagnetic state that characterizes the ct phase of cafe@xmath7as@xmath7 , cheng et al . found afm order in @xmath0 below @xmath18 k with two successive magnetic field - induced transitions at @xmath19 t and 4.7 t with a narrow hysteresis.@xcite on the other hand ying et al . observed afm order below @xmath20 k and only one spin - flop transition at 3.5 t. @xcite ying et al .
also reported that a 10% sr substitution for ca increases @xmath21 from 70 k to 90 k whereas the spin - flop field decreases from 3.5 t in @xmath0 to 1.5 t in @xmath22.@xcite both groups grew their single crystals using the coas self - flux growth technique and suggested that the afm structure is collinear a - type , in which the co moments within the @xmath23-plane are ferromagnetically aligned along the @xmath4-axis and the moments in adjacent layers are aligned antiferromagnetically .
however , the proposed a - type afm order has not previously been confirmed via diffraction techniques and is surprising in light of the inelastic neutron scattering measurements in srco@xmath7as@xmath7 mentioned above which show stripe - like afm fluctuations.@xcite here we report on our crystal and magnetic structure investigations of powders and single crystals of sn - flux grown @xmath0 by x - ray and neutron diffraction . the co - refined x - ray and neutron powder diffraction measurements confirm the essential elements of the structure determined previously.@xcite however , from two independent measurements , we show that there are 7(1)% vacancies on the co site , leading to a stoichiomentry of @xmath2 instead of the previously reported stoichiometry @xmath0 .
we also report on the temperature dependence of the lattice constants which exhibit conventional behavior in contrast to the negative thermal expansion found previously for the _ c_-axis of srco@xmath7as@xmath7.@xcite our single - crystal magnetic neutron diffraction measurements are consistent with collinear a - type afm order , as suggested by cheng _
et al _ and ying _ et al_.@xcite the @xmath24 k that we consistently observe is about 20 k lower than the recent reports@xcite for , as yet , unknown reasons .
single crystals of @xmath25 were prepared by solution growth using sn flux .
first , a polycrystalline sample was synthesized by the standard solid state reaction route starting with high purity ca ( 99.98% ) , co ( 99.998% ) and as ( 99.99999% ) from alfa aesar in the stoichiometric 1:2:2 ratio .
the polycrystalline @xmath0 and the sn flux ( 99.999% , alfa aesar ) were taken in a 1:20 molar ratio , placed in an alumina crucible , and sealed inside an evacuated silica tube .
the sealed tube was then slowly heated to 1150 @xmath26c at a rate of 100 @xmath26c / h and soaked for 30 h. the crystal growth was carried out by slow - cooling the solution to 800 @xmath26c at a rate of 3 @xmath26c / h .
shiny plate - like crystals of typical size @xmath27 mm@xmath28 were obtained by centrifuging the flux at 800 @xmath26c .
a jeol jxa-8200 electron probe microanalyzer was used to determine the chemical composition of the crystals using wavelength - dispersive x - ray spectroscopy ( wds ) analysis . a quantum design , inc . , superconducting quantum interference device magnetic properties measurement system ( mpms )
was used for the magnetic susceptibility , @xmath29 , measurements .
high - resolution x - ray powder diffraction data were collected at ambient temperature on beamline x16c at the national synchrotron light source .
the crushed - crystal samples were loaded into a 1 mm diameter glass capillary along with si powder which served as a standard for the refinement of the lattice parameters .
the x - ray wavelength , 0.6995 , was chosen using a si(111 ) double monochromator .
the powder diffraction pattern was collected in the 545@xmath26 @xmath30 range with a constant step size of 0.005@xmath26 , and a linearly varying counting time of 13 s / point .
the incident beam intensity was monitored with an ion chamber and the diffracted radiation was measured with a nai scintillation detector .
the axial and in - plane resolution of the diffractometer were set by slits and a ge(111 ) analyzer crystal , respectively .
diffraction data were obtained at temperatures of 295 k , 200 k , 100 k , 60 k and 20 k. neutron powder diffraction data were collected at the spallation neutron source ( powgen ) at oak ridge national laboratory .
a sample of mass 1.3 g was prepared by crushing single crystals and loaded into a vanadium sample can in the presence of he exchange gas to maintain thermal conductivity at low temperatures .
data were collected at temperatures of 300 k and 10 k using center wavelengths of 1.333 and 4.797 which allowed the acquisition of diffraction patterns for @xmath31 spacings of 0.45 to 8.5 . to search for evidence of magnetic order at low temperature ,
the longer - wavelength data at 10 k were collected for six times longer than for 300 k as the afm signal was expected to be weak because of the expected small ordered moment ( @xmath15 0.3 @xmath32/co).@xcite single - crystal neutron diffraction measurements were performed on a 34 mg crystal using the triax triple - axis spectrometer at the university of missouri research reactor .
the horizontal beam collimation before the monochromator , between the monochromator and sample , between the sample and analyzer , and between the analyzer and detector were 60@xmath33 - 40@xmath33 - 40@xmath33 - 80@xmath33 , respectively .
incident and scattered neutrons with an energy of 14.7 mev were selected by a pyrolitic graphite ( pg ) ( 002 ) monochromator and pg ( 002 ) analyzer , and two pg filters were employed to effectively eliminate higher harmonics in the incident beam .
the sample was aligned such that the ( @xmath34 0 @xmath35 ) reciprocal lattice plane was coincident with the scattering plane of the spectrometer , and was mounted on the cold finger of a helium closed - cycle refrigerator .
samples at 300 k. the data in ( a ) contain peaks of the si powder used as an internal lattice parameter standard .
both patterns contain peaks from the sn flux .
shown are the measured data ( open circles ) , the calculated intensity pattern ( red lines ) , and the corresponding residuals ( i.e. the difference between observed and calculated pattern ) by the curve below the fit .
the vertical lines indicate the positions of diffraction peaks corresponding to phases used in the refinements.,title="fig : " ] samples at 300 k. the data in ( a ) contain peaks of the si powder used as an internal lattice parameter standard .
both patterns contain peaks from the sn flux .
shown are the measured data ( open circles ) , the calculated intensity pattern ( red lines ) , and the corresponding residuals ( i.e. the difference between observed and calculated pattern ) by the curve below the fit .
the vertical lines indicate the positions of diffraction peaks corresponding to phases used in the refinements.,title="fig : " ] rietveld co - refinements of the ambient - temperature x - ray powder data and the neutron powder patterns at both 300 k and 10 k were done together using the general structure analysis system package ( gsas)@xcite and the graphical user interface ( expgui).@xcite the 300 k neutron and x - ray powder data and refinements are shown in fig .
[ fig : xrd_neut ] and the results of the co - refinement are given in tables [ table : crystaldata ] and [ tab : xrd2 ] .
for the neutron data , the absorption parameter at 10 k was kept to the same value as obtained from the 300 k data as this is highly correlated to the thermal parameters and is not expected to change as a function of temperature . for the x - ray data ,
the strongest si lines at low q were excluded from the refinements .
the refinement confirmed that @xmath25 crystallizes in the @xmath1-type body - centered tetragonal structure ( space group @xmath8 ) shown in fig .
[ fig : structure ] .
the lattice parameters and the @xmath36-coordinate of the as atom are in good agreement with the literature values.@xcite we obtain the @xmath37 ratio and the interlayer as as distance @xmath38 using the crystallographic data in table [ table : crystaldata ] , yielding @xmath39 and @xmath40 .
these values indicate that @xmath25 is in the ct phase .
indeed , the value of @xmath41 and @xmath37 for @xmath25 are close to the respective values observed for the pressure - induced ct phase of @xmath42.@xcite the major impurity phase found in the samples is the adventitious sn flux which comprises approximately 7% of the sample by weight , and was included in the refinement .
coas , at a level of @xmath43 wt% was also noted in the x - ray data .
a few very low - intensity diffraction peaks in the x - ray pattern could not be identified conclusively with other known phases .
both the x - ray and neutron refinements ( together and separately ) show that the compound is deficient in co , giving a composition of @xmath2 .
wds analysis was done on several crystals to obtain an independent estimate of the co vacancy concentration .
the average atomic ratios obtained from the wds analysis are ca : co : as = 20.2(2):38.3(3):40.8(4 ) which indicates the stoichiomentry to be @xmath44 , consistent with the value obtained from the refinement of the powder x - ray and neutron data .
although we see a significant amount of sn impurity in the powder data due to the adventitious sn flux , our wds analysis showed that about 0.5 mol% of sn is also present in the bulk of the crystals . [ cols="<,^,^,^,^,^,^",options="header " , ] ( space group @xmath8 ) .
the red arrows denote the ordered magnetic moment directions in the a - type antiferromagnetic structure determined from single - crystal neutron diffraction measurements . ] and @xmath4 lattice parameters for @xmath25 determined by neutron single - crystal diffraction ( open red circles ) and x - ray powder diffraction ( filled blue circles ) ] the temperature dependence of the @xmath3 and @xmath4 lattice parameters was determined from both single - crystal neutron and x - ray powder diffraction measurements , and the results are shown in fig .
[ fig : lattice_parameters ] .
these results are conventional , but the @xmath4-axis data qualitatively differ from the temperature dependence of @xmath4 for @xmath16 , where @xmath4 _ decreases _ monotonically as @xmath45 increases from 7 to 300 k.@xcite these x - ray and neutron data also show that @xmath46 remains constant as temperature is varied from ambient down to 10 k. we now turn to the single - crystal neutron diffraction determination of the magnetic order in @xmath25 . for temperatures above the nel temperature @xmath47 = 52(1 )
k only bragg peaks associated with the body - centered tetragonal chemical structure , with miller indices @xmath48 ( integer @xmath49 ) in the ( @xmath50 ) scattering plane , were observed . as shown in fig .
[ fig : neutron_scan ] , below @xmath47 new bragg peaks appear at positions @xmath51 , consistent with breaking the body - centered symmetry as expected for the predicted @xmath6-type antiferromagnetic ( afm ) structure.@xcite figure [ fig : ordered_moment](b ) plots the temperature evolution of the integrated intensity of the ( 2 0 1 ) magnetic bragg peak as a function of temperature on cooling and warming .
no hysteresis was observed and the determined value @xmath52 k is in excellent agreement with the value obtained from single - crystal @xmath29 measurements shown in fig . [
fig : ordered_moment](a ) .
the direction of the ordered moment can be determined through a comparison of the measured intensities of magnetic bragg peaks .
in particular , only the component of the ordered magnetic moment that is perpendicular to the scattering vector , * * q**@xmath53 , contributes to the magnetic peak with indices @xmath54 . for @xmath25 , below @xmath47 , we note that the ( @xmath55 ) magnetic peaks with @xmath56 odd are absent , demonstrating that the moments are directed along the tetragonal @xmath4-axis as illustrated in fig .
[ fig : structure ] . in principle , the magnitude of the magnetic moment can also be determined from a comparison of the magnetic and nuclear bragg peak intensities .
however , the small mass of sample and its geometry ( thin flat plate ) allow us to place only an upper limit of @xmath57/co on the low - temperature ordered moment from these measurements , consistent with the small moment value of @xmath58/co estimated from the @xmath59 measurements.@xcite we further note that our neutron powder diffraction measurements at 10 k failed to observe reflections at the magnetic bragg peak positions , again consistent with a small ordered moment estimated from the single - crystal neutron diffraction experiment .
= 3 k ( filled blue circles ) , below @xmath47 = 52(1 ) k , and @xmath45 = 64 k ( open red circles ) , above @xmath47 . ] ) for a field of 1 koe parallel ( solid line ) and perpendicular ( dashed line ) to the tetragonal @xmath4-axis , and ( b ) the integrated intensity of the ( 2 0 1 ) magnetic bragg peak on cooling ( red filled circles ) and warming ( blue filled circles ) .
the vertical green bar denotes @xmath60 . ]
in a ct compound , pnictogen ( _ pn _ ) bonding occurs between the _ pn _ atoms in adjacent transition metal-_pn _ layers , where the _ pn _ atoms in adjacent layers are directly above and below each other ( see fig . [
fig : structure ] ) .
pn _ bonding effect was first observed in @xmath61 in which the bonding between the p - p atoms results in a redistribution of electric charge and the oxidation state of p changes to p@xmath62 [ p
p]@xmath63 from its formal oxidation state of p@xmath64 in the t phase.@xcite as shown previously , the pressure - induced transition from the t to the ct phase has been found to quench the fe magnetic moment and associated afm fluctuations in @xmath42.@xcite in a chemical picture this can be viewed as a transition from the fe@xmath65 oxidation state in the t phase to a fe@xmath66 oxidation state in ct phase due to the appearance of as as bonding . thus there is a change in the formal oxidation state of the transition metal ( tm ) in a transition from the t to ct structure .
even though it is not clear whether the formation of a ct structure drives the change in oxidation state of the tm , or it is the stability of the oxidation state of tm that drives the formation of as - as bonding and the stability of the ct structure , it is quite evident from these observations that the formation of a ct structure and the formal oxidation state of tm are interrelated .
in particular , a change in formal oxidation state of the tm may strongly modify the magnetic character of 122-type pnictides .
co@xmath66 with a @xmath67 electronic configuration is expected to be magnetic .
indeed , there are several known ct 122-type co compounds that exhibit interesting magnetic behavior due to the co magnetism .
@xmath61 is one such example in which the co atoms are in a magnetic @xmath68 co@xmath66 state .
@xcite this compound exhibits a - type antiferromagnetism below @xmath69 k , where the co ordered moments are aligned ferromagnetically within the basal @xmath23 plane and antiferromagnetically along the @xmath4-axis.@xcite an interesting magnetic phase diagram driven by a t to ct transition with concurrent p
p bonding has been observed for the system sr@xmath11ca@xmath12co@xmath7p@xmath7 .
@xcite high - temperature co - sublattice afm transitions have been found in ct rare earth compounds @xmath70 ( @xmath71 k ) , @xmath72 ( @xmath73 k ) , @xmath74 ( @xmath75 k ) and @xmath76 ( @xmath77 k ) , whereas @xmath78 exhibits ferromagnetic ordering with a curie temperature @xmath79 k.@xcite the application of pressure has been found to result in co - moment magnetic ordering in @xmath80 that develops due to a pressure - induced isostructural phase transition from the t to the ct phase at a critical pressure @xmath81 gpa .
@xcite at ambient pressure , @xmath80 orders antiferromagnetically below @xmath82 k due to the ordering of localized eu@xmath65 moments with spin @xmath83 which is suppressed for pressures above @xmath84 with a simultaneous development of itinerant co - moment magnetic order at @xmath85 k in the ct phase .
@xcite a comparison of the values of @xmath37 and @xmath41 in the family of @xmath86 ( @xmath6 = ca , sr , ba ; @xmath87 = cr , mn , fe , co , ni , cu , and @xmath88 = p , as ) in ref .
demonstrates that @xmath25 has a ct structure . as a consequence of the formation of interlayer as as bonds in the ct structure , the as atoms have an unusual oxidation state of as@xmath62 [ as as]@xmath89/2 , whereas in the t structure of bafe@xmath7as@xmath7 and @xmath42 , the formal oxidation state of as is as@xmath64 .
therefore , the oxidation state of co in ct - type @xmath25 is co@xmath66 whereas that of the fe in the t phase of cafe@xmath7as@xmath7 is fe@xmath65 .
interestingly , our comparison of the magnetic properties of @xmath86 ( @xmath6 = ca , sr , ba ; @xmath87 = @xmath90 transition metal ; @xmath88 = p , as ) compounds in ref . revealed that all the magnetically ordered compounds have the t structure except for the ct compound @xmath61 which exhibits afm order .
this suggests that co@xmath66 in the ct phase is magnetic in nature , consistent with the occurrence of afm order in the ct - phase of @xmath0 .
using both x - ray and neutron diffraction measurements on powders and single crystals of sn - flux grown samples we have investigated the crystal and magnetic structures of @xmath25 .
the co - refined x - ray and neutron powder measurements confirm the essential elements of the structure determined previously by pfisterer and nagorsen.@xcite however , we find , from both our powder diffraction and wds measurements , that there are 7(1)% vacancies on the co site , leading to a stoichiomentry of @xmath2 instead of the previously reported stoichiometry of @xmath0 .
we have also reported on the temperature dependence of the lattice constants which exhibit conventional behavior in contrast to the negative thermal expansion found previously for the _
c_-axis lattice parameter of srco@xmath7as@xmath7.@xcite our single - crystal magnetic neutron diffraction measurements are consistent with collinear a - type afm order , as suggested by cheng _
et al _ and ying _ et al_,@xcite although the @xmath24 k that we consistently observe is about 20 k lower than these reports .
this research was supported by the u.s .
department of energy , office of basic energy sciences , division of materials sciences and engineering .
ames laboratory is operated for the u.s .
department of energy by iowa state university under contract no .
de - ac02 - 07ch11358 .
use of the national synchrotron light source , brookhaven national laboratory , was supported by the u.s .
department of energy , office of basic energy sciences , under contract no .
de - ac02 - 98ch10886 .
research at ornl s spallation neutron source was sponsored by the scientific user facilities division , office of basic energy sciences , u.s .
department of energy .
a. jesche , n. caroca - canales , h. rosner , h. borrmann , a. ormeci , d. kasinathan , h. h. klauss , h. luetkens , r. khasanov , a. amato , a. hoser , k. kaneko , c. krellner , and c. geibel , phys .
b * 78 * , 180504(r ) ( 2008 ) .
r. m. fernandes , d. k. pratt , wei tian , j. zarestky , a. kreyssig , s. nandi , m. g. kim , a. thaler , n. ni , p. c. canfield , r. j. mcqueeney , j. schmalian , and a. i. goldman , phys .
b * 81 * , 140501(r ) ( 2010 ) .
a. kreyssig , m. a. green , y. lee , g. d. samolyuk , p. zajdel , j. w. lynn , s. l. budko , m. s. torikachvili , n. ni , s. nandi , j. b. leo , s. j. poulton , d. n. argyriou , b. n. harmon , r. j. mcqueeney , p. c. canfield , and a. i. goldman , phys .
b * 78 * , 184517 ( 2008 ) . a. i. goldman , a. kreyssig , k. proke@xmath91 , d. k. pratt , d. n. argyriou , j. w. lynn , s. nandi , s. a. j. kimber , y. chen , y. b. lee , g. samolyuk , j. b. leo , s. j. poulton , s. l. budko , n. ni , p. c. canfield , b. n. harmon , and r. j. mcqueeney , phys . rev .
b * 79 * , 024513 ( 2009 )
. d. k. pratt , y. zhao , s. a. j. kimber , a. hiess , d. n. argyriou , c. broholm , a. kreyssig , s. nandi , s. l. budko , n. ni , p. c. canfield , r. j. mcqueeney , and a. i. goldman , phys . rev .
b * 79 * , 060510 ( 2009 ) .
a. pandey , d. g. quirinale , w. jayasekara , a. sapkota , m. g. kim , a. kreyssig , r. j. mcqueeney , a. i. goldman , t. w. heitmann , p. w. stephens , r. s. dhaka , a. kaminski , y. b. lee , b. n. harmon , y. furukawa , v. ogloblichev , and d. c. johnston , phys .
b * 88 * 014526 ( 2013 ) .
w. jayasekara , y. lee , a. pandey , g. s. tucker , a. sapkota , j. lamsal , s. calder , d. a. abernathy , j. niedziela , b. n. harmon , a. kreyssig , d. vaknin , d. c. johnston , a. i. goldman , and r. j. mcqueeney , phys .
rev . lett . * 111 * , 157001 ( 2013 ) . | neutron and x - ray diffraction measurements are presented for powders and single crystals of @xmath0 .
the crystal structure is a collapsed - tetragonal @xmath1-type structure as previously reported , but with 7(1)% vacancies on the co sites corresponding to the composition @xmath2 .
the thermal expansion coefficients for both the @xmath3 and @xmath4 axes are positive from 10 to 300 k. neutron diffraction measurements on single crystals demonstrate the onset of a - type collinear antiferromagnetic order below the nel temperature @xmath5(1 ) k with the ordered moments directed along the tetragonal @xmath4-axis , aligned ferromagnetically in the _ ab_-plane and antiferromagnetically stacked along the _ c_-axis . |
there are numerous scattering media that are either of natural origin - such as biological tissue , fog , cloud , and teeth - or man - made - such as paint , diffuser glass , and phosphor inclusion in white leds .
these materials multiply scatter waves in all directions , due to the random spatial variations of the refractive indices .
random interference occurs between the multiply scattered waves that is well - known as speckles @xcite . upon averaging over realizations of scatterers ,
the speckles average out and the resulting average energy density is well described by diffusion theory @xcite .
recently , novel wave - shaping methods such as feedback - based wavefront shaping @xcite , time reversal@xcite , phase conjugation@xcite , and transmission matrix - based control @xcite have demonstrated the control of transmission and reflection through scattering media .
these wave - shaping methods remarkably enhance the intensity in a single speckle as well as the total reflected and transmitted intensity .
the wave - shaping methods have led the way for exciting applications such as non - invasive biomedical imaging @xcite , advanced optics @xcite , and cryptography and secure communication@xcite . although these methods are useful tools in controlling the intensity at the interfaces of a scattering medium , the energy density distribution of the shaped waves inside the scattering medium is still unknown . in the case of light ,
the knowledge of the energy density distribution is important for applications such as enhanced energy conversion in white leds @xcite , efficient light harvesting in solar cells @xcite , and controlled illumination in biomedical imaging @xcite . to date , only numerical calculations of scalar waves @xcite and a single - realization elastic wave experiment @xcite have addressed the energy density distribution of shaped waves inside two - dimensional ( 2d ) scattering media . however , none of these studies provide an analytical model for the energy density distribution . for high - transmission channels , davy _
@xcite described the energy density using a parabolic function . in ref .
@xcite , our team reported the first measurement of the total energy density of light inside a three - dimensional ( 3d ) scattering medium and reported an enhanced total energy density for shaped waves .
the results were interpreted with a heuristic model that the total energy density of shaped waves can be described with only the fundamental eigenfunction of the diffusion equation , which agreed well with experimental observation .
however , the exact distribution of the spatial energy density for shaped waves still remains unknown . in this paper
, we study the energy density of shaped waves inside a scattering medium by using the eigenfunctions of the diffusion equation to reconstruct the energy density distribution . here
, we refer to shaped waves to mean perfect phase conjugation of the transmitted waves , which is equivalent to shaping to optimally focus light to a single speckle spot .
the diffusion eigenfunctions form a complete set and therefore describe the ensemble average energy density within the domain of the sample boundary . by numerical calculation of concatenated scattering matrices @xcite
, we obtain the intensity transmission coefficients @xmath1 of @xmath2th transmission channel and its energy density @xmath3 along the depth @xmath4 of a quasi-1d disordered waveguide .
a quasi-1d disordered waveguide provides a tractable platform to study the essential physics . by calculating the transmission - averaged energy density for all transmission channels
, we obtained the energy density of shaped waves @xcite .
we show that only the first seven diffusion eigenfunctions are sufficient to reconstruct the spatial energy density distribution of the shaped waves . taking into account only the fundamental eigenfunction of the diffusion equation ,
the total internal energy inside the sample is underestimated by only 2@xmath0 .
the spatial energy density is very similar to the fundamental eigenfunction , with a cosine distance of 0.01 .
furthermore , we are able to reconstruct the energy density distribution of both high- and low- transmission channels , using a few @xmath5 eigenfunctions of the diffusion equation , _ e.g. _
@xmath6 for channel with transmission @xmath7 .
the reconstruction of the distribution of the energy densities that we found here can be extended to three dimensional ( 3d ) scattering samples as well as to other forms of classical and quantum waves .
the diffusion equation describes the ensemble averaged energy density @xmath8 of multiply scattered waves as a function of position @xmath9 and time @xmath10 inside a scattering medium @xcite , @xmath11 where @xmath12 is the diffusion constant .
anticipating the fact that the directions @xmath13 are decoupled from the @xmath4-direction we decompose the energy density as @xmath14 since we are considering a slab that is a quasi-1d scattering medium , we assume translational invariance in the perpendicular directions @xmath13 .
the sample boundaries are at @xmath15 and @xmath16 , where @xmath17 is the sample thickness .
this symmetry allows us to solve @xmath18 by using its two - dimensional spatial fourier transform @xmath19 . solving the @xmath13-part of eq .
( [ eq : dif3dt ] ) gives @xmath20 next we solve the remaining 1d equation @xmath21 we will turn this into a sturm - liouville eigenvalue problem by introducing the ansatz @xmath22 @xmath23 as is well - known from linear algebra @xcite solutions of sturm - liouville equations are orthogonal and form a complete set . applying boundary condition will lead to a discrete set of eigenvalues @xmath24 .
the general eigenfunction for eigenvalue @xmath24 can be cast in the form @xmath25 as outlined in detail in ref .
@xcite the wavevector @xmath26 and phase @xmath27 will be determined from solving eq .
[ eq : sturm - liouville ] with the appropriate boundary conditions .
the paramater @xmath28 takes care of the normalization .
the boundary conditions are derived from the sample parameters : the mean free path @xmath29 , the left and right extrapolation lengths @xmath30 and the sample thickness @xmath17 .
the allowed wave vectors @xmath26 are obtained by solving the implicit equation @xmath31 following ref .
@xcite , the phase factor @xmath32 fulfills the following equation @xmath33 and the normalization factor @xmath28 is given by @xmath34 the relation between eigenvalue and wavevector is @xmath35 .
for the exact solution of eq .
[ eq : ks ] , see appendix .
the first ten wavevectors obtained from the solution are shown in table [ table : table1 ] for a particular set of boundary conditions . in fig .
[ fig : figure1 ] , we plot the first three eigenfunctions using eq .
[ eq : w_modes ] with the same set of boundary conditions .
.values of the wavevector @xmath26 for the first @xmath36 eigenfunctions of the diffusion equation .
the values were obtained by solving eq .
[ eq : ks ] .
the extrapolation lengths are @xmath37 ( for a 2d sample @xcite ) and l = 12.1@xmath29 . [ cols="^,^,^ " , ] .
the wave vectors @xmath26 are presented in table [ table : table1 ] and the used parameters can be found in the caption of that table . ] the diffusion equation is of the sturm - liouville type so its eigenfunctions form a complete , orthogonal , set @xcite .
therefore , any distribution of energy density within the domain of the sample boundaries can be decomposed into a sum over a finite number @xmath5 of eigenfunctions for a specific set of coefficients .
such a decomposition is only useful if the number of needed modes is small .
we express the ensemble averaged energy density @xmath38 inside the scattering medium in terms of the normalized eigenfunctions @xmath39 @xmath40 where @xmath41 are coefficients .
we calculate the overlap integral @xmath42 of the product of @xmath38 and @xmath43 @xmath44 and substitute eq .
[ eq : w_z ] into eq .
[ eq : i_def ] to obtain @xmath45 we use orthonormality of functions @xmath43 and @xmath46 , which is expressed as @xmath47 where @xmath48 is the kronecker delta .
therefore , eq . [ eq : i_1 ] becomes @xmath49 we now can write the decomposition ( [ eq : w_z ] ) as @xmath50 we obtained @xmath38 for shaped waves and different transmission eigenchannels , which we obtained from the simulation described in section [ sec : numsamples ] .
the simulation results are compared to the reconstructed energy density @xmath51 in section [ sec : results ] . in order to quantify the overlap between the reconstructed function @xmath52 and the numerical data @xmath38
, we use cosine distance @xmath53 ref [ ] , which is defined as @xmath54^{\frac{1}{2 } } [ \sum\limits_{i = 1}^{n_s } w^2(z_i)]^{\frac{1}{2 } } } \ , .\ ] ] @xmath55 is the number of points in the numerical data .
@xmath53 varies between 1 and 0 and tends towards 0 as the reconstructed function fully describes the numerical data . as a figure - of - merit of a good reconstruction
, we choose the eigenfunction , which has @xmath56 .
we perform simulations of transport of monochromatic scalar waves through a waveguide containing scatterers .
the waveguide is modeled as a quasi - one - dimensional ( quasi-1d ) system possessing @xmath57 transversal modes , with elastic scattering on bulk impurities .
this is implemented as a 2d waveguide with 100 transversal modes , containing scatterers at random positions with a density of 0.2 scatterer per @xmath58 , @xmath59 is the wavelength .
the distribution of the scattering potential @xmath60 of scatterers in one waveguide configuration is shown in fig .
[ fig : figure2 ] . in order to calculate the energy density inside the waveguide
we employ a recursive @xmath61-matrix formalism , where the waveguide is first divided into sub - wavelength slices , each containing only few scatterers . for each slice
we define the @xmath61-matrix in a mode representation as follows @xcite , @xmath62 where @xmath63 and @xmath64 are the left and right reflection matrix respectively of dimension @xmath65 , @xmath66 is the left - to - right transmission matrix .
the right - to - left transmission matrix in this reciprocal system is the transpose of @xmath66 .
the @xmath61-matrix of each slice , of dimensions @xmath67 , is calculated in the approximation that no recurrent scattering takes place within the thin slice .
next , the @xmath61-matrices of the slices are joined , including contributions of recurrent scattering , using the composition rule for @xmath61-matrices@xcite .
this numerically stable procedure yields the @xmath61-matrix of the entire waveguide and of subsections of it . from the transmission matrix
we extract the singular values and vectors corresponding to all channels by singular value decomposition
. the energy density inside the waveguide , for a given incident field , is calculated as follows : for a given axial coordinate @xmath68 the @xmath61 matrices @xmath69,@xmath70 of the waveguide sections @xmath71 and @xmath72 respectively are calculated .
the field in the @xmath73 plane is then found by calculating recurrent scattering diagrams : @xmath74 @xmath75 here , @xmath76 is the left reflection submatrix of @xmath70 , @xmath77 is the right reflection submatrix of @xmath69 , and @xmath78 is the left to right transmission submatrix of @xmath69 .
subsequently for every @xmath79 the time - averaged energy density @xmath80 is integrated over the cross section to obtain the projected energy density @xmath38 .
the calculation is repeated for 8000 independently generated random waveguides from which the transmission channels are extracted . to obtain plots of the average energy density as a function of transmission
, we average the energy densities of channels in a narrow band of transmission values . of point scatters inside one configuration of the numerical samples versus the reduced sample depth @xmath81 .
the scattering potential is in the unit of @xmath82 , @xmath83 is the speed of light , and h is the planck constant . ] . in ( a ) and ( b ) ,
we ensemble averaged over 8000 waveguides and the sample thickness is @xmath84 . ] in order to characterize the numerical samples , we plot in fig .
[ fig : figure3]a the probability distribution @xmath85 as a function of the transmission @xmath86 . the probability distribution @xmath85 obtained from the simulation is bi - modal : there is a high probability for transmitting channels with transmission @xmath86 close to zero and one . in fig .
[ fig : figure3]a , we compare the dorokhov - mello - pereyra - kumar ( dmpk ) distribution @xcite with our numerical result .
the probability distribution of transmission for a scattering medium is expected to follow the dmpk distribution .
the dmpk distribution agrees well with our numerical result .
furthermore , in order to confirm that our numerical samples are in the diffusive regime , we plot in fig .
[ fig : figure3]b the equally - weighted ensemble averaged energy density versus reduced sample depth .
the ensemble averaged energy density shows a linear decrease from the front surface towards the end surface of the sample , in agreement with the prediction of diffusion theory for a diffusive sample , see fig .
[ fig : figure3]b .
we obtained the energy density of shaped waves @xmath87 as the transmission - weighted ensemble average of the energy density of transmission channels @xmath3 , as shown in ref .
@xcite : @xmath88 where @xmath89 is the number of transmission channels . using eq .
[ eq : w_contr ] , we obtained the energy density of the shaped waves from the transmission - weighted averaged over the energy density of all transmission channels .
we show in fig .
[ fig : figure4]a the ensemble averaged energy density distribution of shaped waves .
the energy density distribution of the shaped waves increases from the interfaces of the sample and is maximum at @xmath90 , which is about 20@xmath0 off the center of the sample ( @xmath91 ) .
the ensemble averaged energy density of wavefront - shaped light was first obtain in ref .
@xcite by solving the maxwell equation using the finite - difference time - domain ( fdtd ) method .
despite the different numerical calculation methods , our numerical result is closely similar to the one obtained by choi _
et al _ @xcite .
the exact peak position of the energy density obtain in @xcite is difficult to estimate due to the noise in the data , however , the peak position is close to the center of the sample .
we reconstruct the distribution of the energy density of shaped waves from eigenfunctions of the diffusion equation . from the reconstruction
, we obtain the cosine distance @xmath53 which we plot in fig .
[ fig : figure4 ] . in case of shaped waves ,
a summation over the first @xmath92 eigenfunctions is remarkably sufficient ( @xmath56 ) to reconstruct the energy density , see fig .
[ fig : figure4]a
. only the diffusion fundamental eigensolution @xmath93 diffusion eigensolution has @xmath94 and is not sufficient to describe the energy density profile .
the fundamental eigensolution peaks at exactly the center of the sample , in contrast to the energy density of shaped waves .
we further investigated the contribution of the eigenfunctions to the total energy density integrated along the depth of the sample relative to the contribution of the first 100 eigenfunctions .
as a reference , we choose the first 100 eigenfunctions because the higher order eigenfunctions has a very negligible contribution . in fig .
[ fig : figure5 ] , we plot the contributions . for shaped waves , the @xmath93 contributes 98@xmath0 of the total energy density inside the sample .
remarkably , our finding about the contribution of @xmath93 to the total internal energy density agrees very well with the heuristic model in ref .
@xcite that wavefront shaping _
predominately _ couples light to the fundamental diffusion eigensolution .
the contribution of @xmath95 is negligible compared to @xmath93 since @xmath95 has negative energy density in almost half of the sample depth , which cancels out the positive energy density ( see fig .
[ fig : figure1 ] ) .
a similar effect also happens for other even index eigenfunction .
the ensemble averaged energy density of open transmitting eigenchannels with a transmission @xmath86 in the range @xmath96 is shown in fig .
[ fig : figure6 ] .
the energy density shows a symmetric profile , which peaks in the middle of the sample .
our numerical result is similar to the one obtained by davy _
et al _ @xcite .
interestingly , only the fundamental diffusion eigenfunction @xmath93 is sufficient to reconstruct the energy density of the open channel , see fig . [
fig : figure4]a .
the fundamental diffusion eigenfunction @xmath93 is expected to describe the open eigenchannels for the following reasons : firstly , as the open eigenchannel has the highest individual transmission , the fundamental eigenfunction @xmath97 as well contributes most to the total transmission as shown in ref.@xcite .
secondly , the fundamental eigenfunction is the only physical solution with a positive energy density along the sample depth @xmath4 , see fig .
[ fig : figure1 ] .
interestingly , a simplified mathematical model by davy _
@xcite described the energy density of open channels as a parabolic function .
we note that the deviations from a parabolic shape are obvious only near the edges and the shape of the energy density curve agrees better with the cosine shape of the fundamental eigenfunction , see fig .
[ fig : figure6](a ) .
the @xmath53 for the parabola is @xmath98 and @xmath99 for the fundamental eigenfunction .
for eigenfunctions @xmath100 and @xmath101 plotted with red circles , blue squares , green triangles , and black diamonds respectively . ] in order to describe the distribution of energy density of low - transmission channels , we also use the eigenfunctions of the diffusion equation to reconstruct the ensemble averaged energy densities of low - transmission channels obtained from simulation . in figs .
[ fig : figure6 ] ( b - d ) , we show the ensemble averaged energy densities for transmission eigenchannels with transmissions @xmath102 and @xmath103 . for the eigenchannel with transmission @xmath104 , only @xmath92 eigenfunctions of the diffusion equation are sufficient to reconstruct its energy density , @xmath6 eigenfunctions for @xmath7 , and @xmath105
is required for a closed channel @xmath106 . as the transmission @xmath86 decreases , the number of eigenfunctions sufficient to reconstruct the energy density increases .
the increase in the number @xmath5 of sufficient eigenfunctions is because the asymmetry of the distribution of energy density increases as the transmission @xmath86 reduces , and likewise , the asymmetry of the eigenfunctions increases with the eigensolution index .
therefore , the higher index eigenfunctions contribute more to an asymmetric function .
the contribution of the eigenfunctions with index @xmath107 = 1 , 2 , 3 and 4 relative to a summation over the first 100 eigenfunctions is shown in fig .
[ fig : figure7new ] . the relative contribution of the fundamental eigenfunction of the diffusion equation @xmath93 , which is a symmetric function increases as the transmission increases , while the relative contributions of the higher index eigenfunctions , which are asymmetric functions are higher for low - transmission channels . here
, we investigate the effect of the sample thickness on the number of eigenfunctions sufficient to reconstruct the energy density of shaped waves .
we therefore perform simulation of the energy density on different samples with various sample thicknesses : @xmath108 and @xmath109 .
we show in fig .
[ fig : figure7 ] the @xmath53 versus the eigensolution index for the 4 samples .
interesting , all the samples have a convergence to a value of @xmath110 at the seventh eigensolution .
this convergence of the @xmath53 shows that the summation over the first 7 eigenfunctions of the diffusion equation is sufficient to describe the energy density of shaped waves , irrespective of the sample thickness .
there is however a slight deviation for the thickest sample @xmath111 and this deviation is probably due to artefacts from the simulation due to the large number of waveguide slices . in the previous sections
, we have shown that the diffusion fundamental eigenfunction of the diffusion equation @xmath93 dominates the energy density of high - transmission channels .
we investigate if this result holds for the different samples thicknesses . in fig .
[ fig : figure7 ] , we show the coefficient of the @xmath93 mode relative to the sum of the first 100 eigenfunctions versus the transmission for the 4 different thicknesses .
the relative contribution of the @xmath93 mode is similar for all samples , irrespective of the thickness and this is quite remarkable .
the diffusion fundamental eigensolution dominates the energy density of high - transmission channels for all sample thicknesses .
we have shown that only a few eigenfunctions of the diffusion equation suffice to accurately reconstruct the distribution of the shaped energy density inside a quasi-1d scattering waveguides .
in particular , the fundamental eigenfunction of the diffusion equation is very similar to the distribution of energy density of shaped waves . to reconstruct the distribution of open channels
only the diffusion fundamental eigenfunction @xmath93 is sufficient .
in addition , we have shown that a few number of diffusion eigenfunctions reconstruct the energy density of low transmitting channels .
our results are relevant for applications that require the precise knowledge of distribution of the energy density inside scattering media .
such applications include as efficient light harvesting in solar cells especially in near infrared where silicon has low absorption ; for enhanced energy conversion in white leds , which serves to reduce the quantity of expensive phosphor ; for low threshold and higher output yield of random lasers ; as well as in homogeneous excitation of probes in biological tissues .
we thank pepijn pinkse , henri thyrrestrup , and hasan yilmaz for useful discussions .
this project is part of the research program of the `` stichting voor fundamenteel onderzoek der materie '' ( fom ) fom - program `` stirring of light ! '' , which is part of the `` nederlandse organisatie voor wetenschappelijk onderzoek '' ( nwo ) .
we acknowledge nwo - vici , darpa , and stw .
in eq . [ eq : ks ] , we show an implicit equation , which defines the allowed wave vectors of the diffusion eigenfunctions . the equation is again stated here : @xmath112 we define @xmath113 where @xmath114 and @xmath115 our goal is to find the wave vectors @xmath116 that fulfills the condition @xmath117 in fig .
[ fig : figurea1 ] , we plot the two functions @xmath118 and @xmath119 .
the condition in eq .
[ eq : rooteq ] is fulfilled by the wave vectors @xmath116 at the intersection of the two functions @xmath118 and @xmath119 , as indicated in fig .
[ fig : figurea1 ] .
we found the wave vectors @xmath116 by dividing function @xmath120 into domains with size @xmath121 .
in most of the domains , only one root is expected to be in each of the domains , expect the domain where function @xmath119 has a divergence , which has two roots . in each domain
, we use a simple bisection method to search the domain for wave vectors that fulfill the desired condition .
each domain is divided into two equal parts .
the function @xmath120 is first evaluated for left part of the domain and then check if there is any root present .
if there is no root present , the left domain is further divided into smaller subdomains and each subdomains is searched for the root .
we divide the domains until the size of the subdomain is @xmath122 .
if there is no root in all the subdomains of the left domain , then , the right domain is searched . this way , all the desired roots are found , iteratively . | we show that the spatial distribution of the energy density of optimally shaped waves inside a scattering medium can be described by considering only a few of the lowest eigenfunctions of the diffusion equation . taking into account only the fundamental eigenfunction ,
the total internal energy inside the sample is underestimated by only 2@xmath0 .
the spatial distribution of the shaped energy density is very similar to the fundamental eigenfunction , up to a cosine distance of about 0.01 .
we obtained the energy density inside a quasi-1d disordered waveguide by numerical calculation of the joined scattering matrix .
computing the transmission - averaged energy density over all transmission channels yields the ensemble averaged energy density of shaped waves . from the averaged energy density obtained , we reconstruct its spatial distribution using the eigenfunctions of the diffusion equation .
the results from our study have exciting applications in controlled biomedical imaging , efficient light harvesting in solar cells , enhanced energy conversion in solid - state lighting , and low threshold random lasers .
10 j. goodman , _ statistical optics_. wiley , new york , 2000 .
a. ishimaru , _ wave propagation and scattering in random media_. academic press , new york , 1978 .
m. c. w. van rossum and t. m. niewenhuizen , `` multiple scattering of classical waves , '' _ rev .
mod .
phys .
_ , vol .
71 , pp . 313371 , 1999 .
e. akkermans and g. montambaux , _ mesoscopic physics of electrons and photons_. cambridge university press , cambridge , 2007 .
i. freund , `` looking through walls and around corners , '' _ physica a : statistical mechanics and its applications _ , vol . 168 , no . 1 ,
pp .
49 65 , 1990 .
i. m. vellekoop and a. p. mosk , `` focusing coherent light through opaque strongly scattering media , '' _ opt .
lett .
_ , vol .
32 , pp . 23092311 , 2007 .
i. m. vellekoop and a. p. mosk , `` universal optimal transmission of light through disordered materials , '' _ phys .
rev .
lett .
_ , vol . 101 , p. 120601
, 2008 .
s. m. popoff , a. goetschy , s. f. liew , a. d. stone , and h. cao , `` coherent control of total transmission of light through disordered media , '' _ phys .
rev .
lett .
_ , vol . 112 , p. 133903
, apr 2014 .
m. davy , z. shi , and a. z. genack , `` focusing through random media : eigenchannel participation number and intensity correlation , '' _ phys .
rev .
b _ , vol . 85 , p. 035105
, jan 2012 .
a. p. mosk , a. lagendijk , g. lerosey , and m. fink , `` controlling waves in space and time for imaging and focusing in complex media , '' _ nature photon .
_ , vol . 6 , pp .
283292 , 2012 .
i. m. vellekoop , `` feedback - based wavefront shaping , '' _ opt .
express _
, vol .
23 , pp . 1218912206 , may 2015 .
a. derode , p. roux , and m. fink , `` robust acoustic time reversal with high - order multiple scattering , '' _ phys .
rev .
lett .
_ , vol . 75 , pp .
42064209 , dec 1995 .
g. lerosey , j. de rosny , a. tourin , a. derode , g. montaldo , and m. fink , `` time reversal of electromagnetic waves , '' _ phys .
rev .
lett .
_ , vol .
92 , no .
19 , p. 193904
, 2004 .
g. lerosey , j. de rosny , a. tourin , and m. fink , `` focusing beyond the diffraction limit with far - field time reversal , '' _ science _ , vol .
315 , pp . 11201122 , 2007 .
e. n. leith and j. upatnieks , `` holographic imagery through diffusing media , ''
_ j. opt .
soc .
am .
_ , vol .
56 , pp . 523523 , apr 1966 . d. r. dowling and d. r. jackson , `` narrow band performance of phase conjugate arrays in dynamic random media , '' _ the journal of the acoustical society of america _ , vol .
91 , no . 6 , 1992 .
z. yaqoob , d. psaltis , m. s. feld , and c. yang , `` optical phase conjugation for turbidity suppression in biological samples , '' _ nat .
photonics _ ,
vol . 2 , pp .
110115 , 2008 .
m. kim , y. choi , c. yoon , w. choi , j. kim , q .- h .
park , and w. choi , `` maximal energy transport through disordered media with the implementation of transmission eigenchannels , '' _ nat .
photonics _ , vol . 6 , p. 581
, 2012 .
s. m. popoff , g. lerosey , r. carminati , m. fink , a. c. boccara , and s. gigan , `` measuring the transmission matrix in optics : an approach to the study and control of light propagation in disordered media , '' _ phys .
rev .
lett .
_ , vol . 104 , p. 100601
, 2010 .
s. m. popoff , g. lerosey , m. fink , a. c. boccara , and s. gigan , `` controlling light through optical disordered media : transmission matrix approach , '' _ new journal of physics _
, vol . 13 , no . 12 , p. 123021
, 2011 .
y. m. wang , b. judkewitz , c. a. dimarzio , and c. yang , `` deep - tissue focal fluorescence imaging with digitally time - reversed ultrasound - encoded light , '' _ nat commun _ ,
vol . 3 , p. 928 , jun 2012 .
k. si , r. fiolka , and m. cui , `` fluorescence imaging beyond the ballistic regime by ultrasound - pulse - guided digital phase conjugation , '' _ nat photon _
, vol . 6 , pp .
657661 , oct 2012 .
j. bertolotti , e. g. van putten , c. blum , a. lagendijk , w. l. vos , and a. p. mosk , `` non - invasive imaging through opaque scattering layers , '' _ nature _ , vol .
491 , pp . 232234 , nov 2012 .
t. cizmar , m. mazilu , and k. dholakia , `` in situ wavefront correction and its application to micromanipulation , '' _ nat photon _ ,
vol . 4 , pp .
388394 , jun 2010 .
j .- h .
park , c. park , h. yu , y .- h .
cho , and y. park , `` dynamic active wave plate using random nanoparticles , '' _ opt .
express _
, vol .
20 , pp . 1701017016 , jul 2012 .
y. guan , o. katz , e. small , j. zhou , and y. silberberg , `` polarization control of multiply scattered light through random media by wavefront shaping , '' _ opt .
lett .
_ , vol .
37 , pp . 46634665 , nov 2012 .
j .- h .
park , c. park , h. yu , y .- h .
cho , and y. park , `` active spectral filtering through turbid media , '' _ opt .
lett .
_ , vol .
37 , pp . 32613263 , aug 2012 .
e. small , o. katz , y. guan , and y. silberberg , `` spectral control of broadband light through random media by wavefront shaping , '' _ opt
. lett .
_ , vol .
37 , pp . 34293431 , aug 2012 .
s. r. huisman , t. huisman , t. a. w. wolterink , a. p. mosk , and p. pinkse , `` programmable multiport optical circuits in opaque scattering materials , '' _ opt .
express _
, vol .
23 , p. 31023116
, 02/2015 2015 .
r. horstmeyer , b. judkewitz , i. m. vellekoop , s. assawaworrarit , and c. yang , `` physical key - protected one - time pad , '' _ scientific reports _ , vol . 3 , pp .
3543 ep , dec 2013 . article .
s. a. goorden , m. horstmann , a. p. mosk , b. skoric , and p. pinkse , `` quantum - secure authentication of a physical unclonable key , '' _ optica _ , vol . 1 ,
pp . 421424 , 2014 .
m. krames , o. shchekin , r. mueller - mach , g. o. mueller , l. zhou , g. harbers , and m. craford , `` status and future of high - power light - emitting diodes for solid - state lighting , '' _ display technology , journal of _ , vol . 3 , pp .
160175 , june 2007 .
j. phillips , m. coltrin , m. crawford , a. fischer , m. krames , r. mueller - mach , g. mueller , y. ohno , l. rohwer , j. simmons , and j. tsao , `` research challenges to ultra - efficient inorganic solid - state lighting , '' _ laser ` & ' photonics reviews _ , vol . 1 , no . 4 , pp .
307333 , 2007 .
v. y. f. leung , a. lagendijk , t. w. tukker , a. p. mosk , w. l. ijzerman , and w. l. vos , `` interplay between multiple scattering , emission , and absorption of light in the phosphor of a white light - emitting diode , '' _ opt .
express _
, vol .
22 , pp .
81908204 , 04/2014 2014 .
t. ogi , a. b. d. nandiyanto , k. okino , f. iskandar , w .- n .
wang , e. tanabe , and k. okuyama , `` towards better phosphor design : effect of sio2 nanoparticles on photoluminescence enhancement of yag : ce , '' _ ecs journal of solid state science and technology _ , vol . 2 , no . 5 , pp .
r91r95 , 2013 .
m. meretska , a. lagendijk , h. thyrrestrup , a. p. mosk , w. l. ijzerman , and w. l. vos , `` how to distinguish elastically scattered light from stokes shifted light for solid - state lighting ?
, ''
_ arxiv:1511.00467 _ , 2016 .
j. a. levitt and w. h. weber , `` materials for luminescent greenhouse solar collectors , '' _ appl .
opt .
_ , vol . 16 , pp .
26842689 , oct 1977 .
a. polman and h. atwater _ nat .
mater .
_ , vol . 11 , p. 174
, 2012 .
f. t. si , d. y. kim , r. santbergen , h. tan , r. a. c. m. m. van swaaij , a. h. m. smets , o. isabella , and m. zeman , `` quadruple - junction thin - film silicon - based solar cells with high open - circuit voltage , '' _ applied physics letters _
, vol .
105 , no . 6 , 2014 .
o. yizhar , l. e. fenno , t. j. davidson , m. mogri , and k. deisseroth , `` optogenetics in neural systems , '' _ neuron _ , vol . 71 , no . 1 ,
pp .
9 34 , 2011 .
w. choi , a. p. mosk , q .- h .
park , and w. choi , `` transmission eigenchannels in a disordered medium , '' _ phys .
rev .
b _ , vol . 83 , p. 134207
, apr 2011 .
m. davy , z. shi , j. park , c. tian , and a. z. genack _ nat .
communications .
_ , vol . 6 , p. 6893
, 2015 .
s. f. liew and h. cao , `` modification of light transmission channels by inhomogeneous absorption in random media , '' _ opt . express _
, vol .
23 , pp . 1104311053 , may 2015 .
b. grardin , j. laurent , a. derode , c. prada , and a. aubry , `` full transmission and reflection of waves propagating through a maze of disorder , '' _ phys .
rev .
lett .
_ , vol . 113 , p. 173901
, oct 2014 .
o. s. ojambati , h. yilmaz , a. lagendijk , a. p. mosk , and w. l. vos , `` selective coupling of optical energy into the fundamental diffusion mode of a scattering medium , '' _
arxiv:1505.08103v2_.
d. y. k. ko and j. c. inkson , `` matrix method for tunneling in heterostructures : resonant tunneling in multilayer systems , '' _ phys .
rev .
b _ , vol . 38 , pp . 99459951 , nov 1988 .
j. b. pendry and a. mackinnon , `` calculation of photon dispersion relations , '' _ phys .
rev .
lett .
_ , vol .
69 , pp . 27722775 , nov 1992 .
c. w. j. beenakker , `` random - matrix theory of quantum transport , '' _ rev .
mod .
phys .
_ , vol .
69 , pp . 731808 , 1997 . p.
m. morse and h. feshbach , _ methods of theoretical physics_. new york , mcgraw - hill . , 1953 .
a. lagendijk , r. vreeker , and p. de vries , `` influence of internal reflection on diffusive transport in strongly scattering media , '' _ phys .
lett .
a _ , vol . 136 , no .
12 , pp .
8188 , 1989 . |
in the last four years , observations by using integral field spectroscopy ( ifs ) of individual star - forming galaxies have revealed that galaxies at @xmath5 show a large variety of kinematic and dynamical properties .
recent studies at @xmath6 have confirmed that in addition of the evidence of organized rotation non - negligible random motions are also detected at this redshift .
in addition , it appears that large and massive disks with strong star formation already exist at @xmath5 @xcite .
however more numerous studies at @xmath7 have found an increase in non - circular motions relatives to lower redshift samples .
indeed most of these high - redshift galaxies have high velocity dispersions .
even when a large - scale velocity gradient throughout the galaxy is detected , the value of the ratio of @xmath8 is relatively low ( @xmath9 ) suggesting that it is unlikely for these galaxies to have a dynamically cold rotating disk of ionised gas similar to the local spiral galaxies ( @xmath10 @xcite ) .
most of these ` heated disks ' appear also to be extremely rich in gas with evidence of high turbulent star formation@xcite .
another significant difference in the properties of some of these high - redshift objects in comparison to the local universe is the increase in the irregular and asymmetric shape of these galaxies .
although some of these irregular objects may be associated with mergers , the morphology and kinematics of the majority of them are incompatible with being ongoing mergers @xcite and suggest that they are rotationally supported @xcite .
these so - called ` chain - galaxies ' have the appearance of giant highly clumpy disks .
the objects can be kpc wide and as massive as @xmath11 @xcite .
@xcite was the first to suggest that they could be the progenitor of @xmath12 spirals .
although the origin of the ` heated disks ' remains uncertain and highly debated , recent theoretical studies @xcite propose a bimodality in galaxy type by @xmath13 with clumpy star - forming disks and spheroid - dominated galaxies with low star formation rates ( sfrs ) . at @xmath14 ,
the disks should be stabilized by the dominant stellar disks and bulges , suggested that the redshift region of @xmath15 is a crucial step in the formation of massive disk galaxies .
nowadays theoretical studies try to understand the formation and evolution of extended turbulent rotating disks , and observational studies , as the ones presented in this paper , probing the kinematics properties ( like observed morphology , dynamical mass , etc ) , the stellar population properties ( age of the population , stellar masses , etc ) , star forming rates , and the properties of the ionized gas ( mass de gas and gas fraction ) of the @xmath15 disks are mandatory for setting constrains on galaxy formation and evolution .
the columns are as follows : ( 1 ) redshift estimated from optical spectrum obtained with vimos , ( 2 ) vvds-22h wide field ( @xmath16 ) and vvds-02h deep field ( @xmath17 ) , ( 3 ) median seeing estimated from the psf stars ( one taken per hour of observation ) .
[ runs ] at lower redshifts , the velocity fields of spiral / rotating disk galaxies have been used to place important constraints on total masses and hence on dark matter halo masses @xcite .
these rotating disks produce a tully - fisher relationship ( the scaling law between total dynamical mass and the luminous stellar mass ) which has apparently not evolved in slope and scatter since @xmath18 .
therefore the large scatter found in previously reported tully - fisher relationships at moderate redshifts is produced by galaxies with perturbed rotation or complex kinematics .
the importance of the role of the non - ordered motion through the gas velocity dispersion for galaxies showing a perturbed rotation was investigated by @xcite and @xcite .
they defined a new tracer of galaxy - dark halo potential , which combines dynamical support from the rotation with that of the non - ordered motion .
the new stellar mass tully - fisher relation shows no detectable evolution neither in slope or of its intercept up to [email protected] * * @xcite .
two - dimensional velocity fields allow rotation curves to be deduced in a more robust manner than slit spectra * * @xcite .
ifs allows pure rotationally - supported disks to be distinguished from other dynamically - disturbed galaxies , which include minor or major mergers , merger remnants and/or inflow / outflows . in this paper
, we explore the kinematic properties of the ionized gas and star formation rates of a sample of ten galaxies at @xmath19 selected in the vimos vlt deep survey ( vvds ) , using integral field unit ( ifu ) @xmath20-band and @xmath21-band spectroscopy with vlt / sinfoni .
the results presented here are part of a study to investigate kinematics and physical properties of unbiased samples representative of the global intermediate and high-@xmath22 population .
this is the second paper of three articles based on data obtained during two observing runs with sinfoni . from the two other companion papers , one diskusses the chemical properties of the intermediate redshift sample presented here and the other presents the results found on the @xmath23 galaxy sample @xcite .
in section [ obs ] we describe our sample , observational strategy and the data reduction techniques . in section [ sprectro_prop ]
we address the properties of the nebular emission and also the nature of the stellar population of the galaxies , using broad - band photometry to constrain properties such as stellar mass , age and sfrs . in section [ kine_prop ]
we investigate the kinematic structure and dynamical properties inferred from ifs . in section [ nature ] we explore the nature of these star - forming galaxies and in section [ relations ] we diskuss various relations between the properties of the gas and those of the stellar population and compare our findings with previous studies of the kinematics of intermediate and high - redshift galaxies .
finally in section [ summary ] we summarize our results and diskuss their implications .
we assume a cosmology with @xmath24 , @xmath25 and @xmath26 = 70 throughout , and all magnitudes are on the @xmath27 system @xcite .
the columns are as follows : ( 1 ) flux(@xmath28 ) , ( 2 ) luminosity(@xmath29 ) , ( 3 ) raw sfr derived from -line ( ) , ( 4 ) reddening suffered by the ionized gas , ( 5 ) dereddened sfr , ( 6 ) global velocity dispersion derived from -line ( km / s ) , ( 7 ) total mass of ionized gas ( @xmath30 ) , ( 8) gas mass fraction .
[ tabline ]
the 10 galaxies , presented in this paper , were selected in the vvds-02h `` deep '' field ( @xmath17 ) of the vvds and in the vvds-22h `` wide '' field ( @xmath16 ; ) .
the objects of this sample have spectroscopic redshifts , derived from vlt / vimos spectroscopy , between @xmath31 and @xmath32 ( see table[runs ] ) .
their redshifts are such that the expected wavelengths of the rest - optical h@xmath33 and [ nii ] emission lines would be clear of bright oh sky lines in the near - infrared ( nir ) .
these targets were mainly selected on the basis of their measured intensity of emission lines in the vimos spectra .
we chose objects showing the strongest [ oii]3727 emission lines ( rest - frame equivalent width ( ew ) @xmath34 and flux @xmath35 ergs@xmath2@xmath36 ) and having a photometric spectral energy distribution ( sed ) corresponding to late - type star - forming galaxies .
the nir spectroscopic observations were acquired with the 3d spectrograph sinfoni at eso - vlt during two 4-nights runs , on 2005 september 5 - 8 ( eso run 75.a-0318 ) and 2006 november 12 - 15 ( eso run 78.a-0177 ) .
sinfoni was used in its seeing - limited mode , with the 0.25 pixel scale leading to a field - of - view of 8@xmath378 @xmath38 , and the @xmath20 grism ( 1.447 - 1.847 @xmath39 ) providing a spectral resolution @xmath40 .
one galaxy , vvds-1235 , has been observed with the @xmath21 grism ( 1.098 - 1.399 @xmath39 ) , where the detection of the emission line is expected .
three galaxies ( vvds-1235 , vvds-4103 and vvds-4167 ) have also been observed with the @xmath21 grism ( 1.098 - 1.399 @xmath39 ) .
conditions were photometric and the median seeing for each objects , estimated on 2d imaging of point spread function ( psf ) stars taken each hour of object observation , is indicated in table[runs ] . observing strategy and data reduction
have been described in @xcite .
integrated rest - frame optical one - dimensional spectra were extracted from the sinfoni data cube for each object . from these 1d spectra , we measured line fluxes and linewidths of the integrated line emission ( see fig.[fig1d ] ) . using the iraf task ` splot ' , we derived the systemic redshift @xmath41 , the total flux and the global velocity dispersion @xmath42 ( corrected for the instrumental resolution of @xmath16.5 in the @xmath21 band and @xmath16.8 in the @xmath20 band ) .
table [ tabline ] list the properties of the nebular emission of the 10 galaxies of our sample .
it is interesting to note that all our targets , except vvds-6027 , have a value of @xmath43 km s@xmath2 .
the h@xmath33 nebular recombination line is a direct probe of the young , massive stellar population and therefore provides a nearly instantaneous ( i.e. averaged on the last ten million years ) measure of the sfr .
the h@xmath33 line is moreover not strongly affected by dust extinction .
we have calculated @xmath44 following the calibration , re - normalized to @xcite imf : @xmath45 table [ tabline ] list the star formation rates deduced from the emission for the 10 galaxies of our sample .
we found that the non - dust - corrected @xmath46 is between @xmath14 @xmath47 114 @xmath48 , with seven of our galaxies with a @xmath49 10 @xmath48 .
these high sfrs are expected , taking into account our selection criteria for the [ oii ] flux ( see section [ sample ] ) .
@xmath50 : magnitudes from the ukidss survey.[tabmag ] we derived the spectral energy distributions ( sed ) of our 10 galaxies from broad - band photometry .
the magnitudes in @xmath51 filters obtained with the cfh12k camera on cfht were extracted from the vvds cataloghttp://cencosw.oamp.fr . , and magnitudes in @xmath52 filters from the cfhtls t003 release .
finally , we have also used magnitudes in @xmath21 and @xmath53 bands from the ukidss survey @xcite .
all the photometric data are presented in table [ tabmag ] . [ cols="^,^,^,^,^,^,^",options="header " , ]
the columns are as follows : ( 1 ) @xmath54 ; ( 2 ) toomre parameter ( toomre 1964 ) ; ( 3 ) dynamical classification : rd for _ rotation - dominated _ rotating disk , dd for _ dispersion - dominated _ rotating disk , and f.l . for
_ featureless_. [ tabkine ]
here , we study and diskuss the relations between the kinematical and physical properties of the ionized gas and the physical properties of the stellar population for the 10 galaxies of our sample .
we calculated the recent - to - past star formation rate ratio ( @xmath55 ) , which is the ratio between the star formation rate estimated from the ionized gas ( using the emission ) and the one estimated from the best - model sed fitting ( both of which have been dereddened ) .
@xmath56 is averaged on a timescale ten times longer than the @xmath57 , which is an indication of the instantaneous star formation .
given these different timescales , the ratio between the two estimates gives information on the presence of recent burst of star formation above the standard continuous declining star formation .
we investigated , thus , the correlation between the recent - to - past star formation rate ratio ( @xmath55 ) and the gas mass fraction and the stellar mass .
+ in fig . [ sfr_mu_age ] , we plot the recent - to - past sfr ratio vs. the gas mass fraction for the 10 galaxies of our sample .
we found that only galaxies with a significant amount of gas ( @xmath580.6 ) , are undergoing episodes of strong star formation ( bursts ) . from fig .
[ sfr_mu_age ] , a correlation between the recent - to - past sfr ratio and the gas mass fraction , given that we divide the sample into two classes is therefore clearly seen .
the first one include the three objects which have the smallest recent - to - past sfr ratio .
these three objects are : the face - on / featureless vvds-6027 ( star symbol ) , the dd rotating galaxy vvds-7106 ( filled triangle ) , and the rd rotating galaxy vvds-4167 ( open cirle ) , and therefore are not showing evidence of bursts of star formation .
this first class seems to follow the relation ( least square fit ) : @xmath59 and the second class , all the remaining objects , seems to follow the relation : @xmath60 can these two classes be related to underlying kinematical properties ?
we find indeed that the ratio of the kinematical radius to the gas radius ( @xmath61 , see table [ tabkine ] ) has a mean value of @xmath62 for the galaxies with a high recent - to - past sfr ratio , and @xmath63 for the three others of the first class .
it appears , thus , that the rather intuitive relation between the amount of recent star formation and the amount of ionized gas depends strongly on kinematics . `
slow rotator ' ( intermediate redshift galaxies with a @xmath64 ) appear less efficient in converting a high mass fraction of ionized gas into a burst of star formation .
this will have to be confirmed and refined with better data and more objects . ] in fig .
[ sfr_mass ] , we plot the recent - to - past sfr ratio vs. the stellar mass for the 10 galaxies of our sample .
we found that the galaxies of our sample need higher stellar masses to reach higher recent star formation rates .
the recent star formation to stellar mass efficiency is moreover significantly stronger for half of our sample . from fig .
[ sfr_mass ] , we see two possible correlations between the recent - to - past sfr ratio and the stellar mass .
the first class includes objects dominated by recent star formation ( @xmath65 ) : the merger object and all dd galaxies , except vvds-6913 with the following relation ( least squares fit ) : @xmath66 @xmath67 the second class includes objects dominated by past star formation ( @xmath68 ) : the featureless object and all rd galaxies , except vvds-2331 , seems to follow the relation ( least squares fit ) @xmath69 @xmath70 the star formation to stellar mass efficiency is moreover stronger for objects experiencing a burst of recent star formation , and appears also to be related to the dominance of random motions in the gas .
we indeed notice that the ratio of the maximum velocity to the velocity dispersion ( @xmath71 ; see table [ tabkine ] ) is different for the rotating disks galaxies inside the two classes : it has a mean value of @xmath72 for the first class , and @xmath73 for the second one .
we thus propose to refine our classification : disk galaxies of the first family would be called `` hot / perturbed rotators '' , while the disk galaxies of the second family would be ` cold rotators ' . only hot rotators and objects dominate by random motions , like the merger seem to be able to experience a recent burst .
the separation between cold rotators and hot / perturbed rotators would then be @xmath74 .
this also will have to be confirmed and refined with better data and more objects .
[email protected] , using the maximum velocity ( top panel ) and using the @xmath76 index ( bottom panel ; @xmath77 ) . in the top panel ,
the solid line is the local relation from @xcite , while in the bottom panel it is the z@xmath11 relation from @xcite .
the symbols are _ open cirle _ : rd rotating disks and _ filled symbols _ : dd rotating disks with in particular _ filled triangle _ : vvds-7106 , and _ filled square _ : vvds-6913.,title="fig : " ] + [email protected] , using the maximum velocity ( top panel ) and using the @xmath76 index ( bottom panel ; @xmath77 ) . in the top panel ,
the solid line is the local relation from @xcite , while in the bottom panel it is the z@xmath11 relation from @xcite .
the symbols are _ open cirle _ : rd rotating disks and _ filled symbols _
: dd rotating disks with in particular _ filled triangle _ : vvds-7106 , and _ filled square _ : vvds-6913.,title="fig : " ] in fig .
[ fig : tully ] , we show the tully - fisher relations for the 1.2@[email protected] galaxies showing evidence of rotation . in the top panel
, we plotted the stellar mass versus the maximum velocity .
as can be seen , there is a clear separation between the rd rotating disks ( open circle ) from the dd rotating disks ( filled symbols ) .
however , both groups of objects suggest a good correlation between the stellar mass and the rotation maximum speed similar to that seen in the local universe ( sold line ) . by assuming the local slope of the tully - fisher relation from @xcite
, we found a best fitting zero point of @xmath78 for the rd galaxies and @xmath79 for the dd galaxies . in the bottom panel of fig .
[ fig : tully ] , we replace the maximum velocity by the @xmath76 index ( @xmath80 ) , which allows the disordered motions in the gas to be taken into account @xcite .
the separation between the rd rotating disks and the dd rotating disks is less obvious while adding the disordered motions .
it seems that we also detect an evolution of the relation since z@xmath11 ( which is shown as the solid line in fig .
[ fig : tully ] ) .
previous studies at @xmath81 have shown the presence of large quantity of gas in comparison to the stellar population in high - redshift galaxies @xcite .
we found that six out of then galaxies of our sample have a gas fraction higher than 0.50 .
the typical @xmath82 for our sample is similar to the one found for the sins galaxies ( @xmath83 ; @xcite ) , but more higher than the values found by @xcite ( @xmath84 ) and @xcite galaxies at @xmath85 ( @xmath86 ) .
@xcite found in their sample , 3 galaxies showing the presence of rotation .
fig [ weiner2006_fig19 ] plots the combined velocity scale against the 1d linewidth for our sample together with this three galaxies at @xmath87 .
it should give an idea of the evolution of the dynamical mass which take also into account the contribution from the random motions in the gas .
fig [ weiner2006_fig19 ] shows an increase of the dynamical mass for @xmath87 to @xmath88 for the dd disks .
line and the @xcite @xmath89 line ( @xmath90 is the circular velocity ) .
linewidth @xmath42 and are correlated ; the 0.5 pre - factor makes the combined velocity width a better estimate of velocity dispersion , so that the correlation is tighter and the galaxies closer to the @xmath91 line .
the symbols are _ open cirle _ : rd rotating disks and _ filled symbols _
: dd rotating disks with in particular _ filled triangle _ : vvds-7106 , and _ filled square _ : vvds-6913 ; _ asterisk _ : @xmath13 galaxies from @xcite ] despite the low resolution of the morphological data ( cfthls i - band images ) which shows that , except vvds-1235 , all the galaxies seems to be disk like objects with a central bulge , the ifu data obtained on the sample of the 10 star - forming galaxies at 1.0@[email protected] have confirmed the presence of a velocity gradient resembling that expected for an internally rotating system , but has also reveals the presence of non - negligible random motions in the gas .
recent studies at the same redshift have found similar results . have study a clumpy galaxy at @xmath92 , which was previously classified as an on - going merger .
their study reveals that the kinematics properties ( the large - scale rotation with @xmath93 km s@xmath2 ) , the stellar properties ( young age , stellar mass range ) and the properties of the gas are very similar to those of the galaxies of our sample ( in particular vvds-6913 ) .
their numerical model shows that a such turbulent rotation disk , results more from the evolution of an unstable gas - rich disk galaxy than from a merging event .
five of the rotating disk of our sample have a @xmath94 , which indicates the stability of their disk .
can this large stable disk systems evolve in present - day thin disks ?
we already note that the disk galaxies in our sample ( except the two galaxies vvds-1328 and vvds-5726 which already have similar properties to that of the present - day local spirals ) possess a large disk of gas and younger stars ( age < 1 gyr ) , which can further grow in mass by continued accretion .
the lack of rotation observed in this objects in comparison to the thin local disk is nowadays well known and taken into recent plausible formation and evolution of high - redshift thick disk models .
these scenarios can predict that on the largest scales a velocity gradient tracing a rotation can be observed with high velocity dispersion .
they also shown that before @xmath95 the hot mode of accretion dominates , but after @xmath85 , the galaxy has a thin and extended disk component with @xmath96 , which indicates that the disk is marginally stable .
they indicate also that after these earlier stages the galaxy enters a slow accretion phase and the disk evolves quiescently until @xmath97 @xcite .
the results presented in this paper support the hypothesis that stable gas - rich disks seen at intermediate and high - redshifts may internally evolve into present - day spirals .
we have presented the 2d kinematics and the physical properties of a sample of ten star - forming disk galaxies at 1.0@[email protected] . among these objects ,
three ( vvds-4252 , vvds-4103 and vvds-4167 ) are undergoing a strong burst of star formation .
we found mainly four kinematical types in our sample which are : one merger ( vvds-1235 ) ; one featureless ( or face - on ) galaxy ( vvds-6027 ) ; four dd rotating disks @xmath47 vvds-4252 , vvds-7106 , vvds-4103 and vvds-6913 which is also consistent of being the relic on a major merging event ( to which the disk system seems to have survived ) or a good candidate for being a clumpy galaxy ; and finally four other galaxies are rd disks @xmath47 vvds-2331 , vvds-4167 , including vvds-1328 and vvds-5726 which are pure rotationally supported disks .
these two rotating disks achieve a maximum velocity of @xmath1 180 - 290 km s@xmath2 km / s within @xmath1 0.5 - 1 kpc , similar to local spirals with thin disk @xcite .
regarding most of the dd rotating disk , they display a plateau velocity range of 105 - 257 km s@xmath2 km / s , certainly underestimated due to beam smearing .
however , their plateau radii ( 4.5 - 10.8 kpc ) derived from our rotating disk model are significantly higher than those derived for pure rotating disks and local spiral galaxies .
we did not find any trace of agn in the 10 objects .
the galaxies of our sample have a relatively young stellar population ( @xmath75 1 gyr ) and possess a range of stellar mass of 0.6 - 5 @xmath54 .
in addition , most of them have not yet converted the majority of their gas into stars ( six galaxies have their gas fraction @xmath4 50 per cent ) .
therefore , the galaxies which already have a stable disk ( six of them ) would have their final stellar mass similar to the present - day spirals , to which these rotating systems can be seen as precursors .
we also investigated the stellar mass tully - fisher relation at 1.2@[email protected] and found a scatter between the rd disk galaxies and the dd disk galaxies .
we also found for this two groups of objects a change of the zero point in comparison with the stellar mass tully - fisher relation in the local universe , but this is speculative considering the statistics of our sample . although , we have presented results on the dynamical structure and the physical properties of a small sample of star - forming galaxies at 1.0@[email protected] , we have been able to investigate the dynamical type and the physical properties of late - type objects at this intermediate redshifts .
increasing the samples at different redshift ranges for which near - infrared ifs data and multi - wavelength broad - band photometric data can be obtained , will definitely develop our understanding of the dynamical characteristics and nature of various systems ( mergers , rotating disks , etc . ) and therefore will lead us to understand how galaxies have evolved to match the present - day hubble sequence .
we would like to thank aprajita verma and andy bunker for helpful diskussions .
we are very grateful to the vlt observatory for accepting this programme .
the authors also thank markus hartung for help obtaining the observations during the observing runs and sebastian snchez for providing the code to create the kinematics maps from the 3d cube .
the anonymous referee is greatly acknowledged for providing useful and constructive comments .
the author wish to recognize and acknowledge the significant contribution the vvds collaboration has done by providing the targets . in particular thierry contini who helped greatly with the targets selection .
part of this work was supported by the marie curie research training network _
euro3d ; contract no .
hprn - ct-2002 - 00305_. m. lemoine - busserolle is supported by the gemini observatory , which 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 combined two - dimensional kinematic and morphology information on the emission , obtained using near - infrared integral field spectroscopy , with broad - band photometry to investigate the dynamical structure and the physical properties of a sample of ten late - type galaxies at 1.0@[email protected] .
their star formation rate ranges from @xmath14 to @xmath1400 with a mean value of @xmath180 .
we found that three of these objects are undergoing a strong burst of star formation .
the sample displays a range of kinematical types which include one merger , one face - on galaxy , and eight objects showing evidence of rotation . among these eight objects ,
half are rotation - dominated galaxies , while the rest are dispersion - dominated .
we found also that two galaxies out of the rotation - dominated galaxies are pure rotationally supported disks .
they achieve a maximum velocity of @xmath1 180 - 290 km s@xmath2 within @xmath1 0.5 - 1 kpc , similar to local spirals with thin disks .
regarding the perturbed rotation and the dispersion - dominated galaxies , they display a plateau velocity range of 105 - 257 km s@xmath2 , which is certainly underestimated due to beam smearing
. however , their plateau radii ( 4.5 - 10.8 kpc ) derived from our rotating disk model are significantly higher than those derived for pure rotating disks and local spiral galaxies .
the galaxies of our sample have relatively young stellar populations ( @xmath0 1.5 gyr ) and possess a range of stellar mass of 0.6 - 5 @xmath3 .
in addition , most of them have not yet converted the majority of their gas into stars ( six galaxies have their gas fraction @xmath450 per cent ) .
therefore , those of them which already have a stable disk will probably have their final stellar mass similar to the present - day spirals , to which these rotating systems can be seen as precursors .
we conclude our study by investigating the stellar mass tully - fisher relation at 1.2@[email protected] .
[ firstpage ] galaxies : evolution galaxies : formation galaxies : kinematics and dynamics galaxies : spiral galaxies : stellar content galaxies : starburst |
recently , the experiment of krasznahorkay _ et al_. studied the decays of a @xmath0 excited state to its ground state and found a bump in both the opening angle and invariant mass distributions of @xmath1 pairs produced in the transitions @xcite .
this 6.8@xmath2 deviation from the expectation can be fitted by the production of a new particle @xmath7 of mass around 17 mev in the transition @xmath8 and the subsequent decay of @xmath7 into electron - positron pair .
although the excess can be due to unknown nuclear effects , the existence of a new particle was investigated in @xcite from different aspects .
it is shown that the proton coupling to @xmath7 is suppressed , or the @xmath7 boson is @xmath9 , in order to alleviate various experimental constraints for a new light gauge boson @xcite . some anomaly free phenomenological models are also proposed in refs .
furthermore , simple models involving new scalar or pseudo - scalar particles are excluded as shown in ref .
@xcite . in this paper
, we take the new particle as a @xmath4 gauge boson corresponding to a new @xmath10 gauge symmetry . the charge
@xmath11 can be a completely hidden quantum number governing the dark sector or a certain combination of standard model ( sm ) quantum numbers .
hence the @xmath4 boson plays as a portal between the dark sector and visible sector .
a simple dark photon model containing two parameters @xmath12 has been ruled out by confronting various experiments both in quark sector and lepton sector @xcite .
the dark @xmath5 model , however , has one more parameter @xmath13 which is induced by the mass mixing between @xmath5 and @xmath4 .
this mixing allows interactions between @xmath4 and fermions including neutrinos .
then it is a natural and interesting topic to discuss the possibility of this more generic model . in the following sections
, we will show that the hidden gauge boson couples to proton ( neutron ) and electron ( neutrino ) in equal strengths . due to
the above coupling relations , we will present explicitly the incompatiblity between the measurement of krasznahorkay _ et al . _
@xcite and texono @xmath14electron scattering experiments @xcite .
since the generic hidden @xmath3 model is disfavored , we also consider non - hidden @xmath3 portal models for further phenomenological studies .
we shall demonstrate that such models are also disfavored by the existing dm direct search data for dm mass above 500 mev . to probe dm lighter than 500 mev in non - hidden models ,
we propose direct searches based upon dm - electron scatterings .. we organized the paper as follows : in section ii we describe the generic @xmath4 lagrangian and take the @xmath4 boson as the dm force mediator ; in section iii we derive the constraints for the model parameters from the @xmath0 transition and other experiments , in particular texono experiment excludes the simple dark @xmath5 model ; in section iv we investigate the implications of dm direct search with dm - nucleus scatterings and discuss the sensitivities of future si and ge detectors to dm - electron scattering in the framework of non - hidden @xmath3 portal model with the wimp mass between 20 mev and 500 mev ; we then conclude in section [ sec : summary ] .
we assume that the dark sector interacts among themselves and links the sm particles via a hidden abelian @xmath15 gauge symmetry .
the connection between visible and dark sector is established by the @xmath15 gauge boson @xmath4 through its kinetic mixing with the sm @xmath3 gauge boson @xmath16 , @xmath17 where @xmath18 characterizes the mixing .
the massive gauge bosons @xmath5 and @xmath4 obtain their masses after spontaneous symmetry breaking as well as a further rotation between them .
hereafter for simplicity we keep the same notation @xmath4 to denote the gauge boson of @xmath10 in mass eigenstate . without loss of generality ,
the interaction among @xmath4 and sm fermions can be described phenomenologically as @xmath19 where @xmath20 , and @xmath13 is some certain combination of @xmath18 and rotation angles , which also depends on uv complete theory and mass generation mechanism .
the electromagnetic current and weak neutral current are @xmath21 respectively , where @xmath22 stands for the fermions with corresponding electric charge @xmath23 , isospin @xmath24 .
more details about this model can be found in @xcite .
furthermore , we assume the interaction in dark sector is @xmath25 where the dm @xmath26 is assumed to be dirac fermion carrying a charge @xmath27 under the hidden gauge symmetry . in fitting to the measurement by krasznahorkay _
et al_. , one obtains the mass of the light boson to be @xcite @xmath30 therefore , the @xmath4 considered in our framework can only decay into @xmath1 and @xmath31 .
the corresponding decay widths are given by @xmath32 and @xmath33 respectively could be a tuning parameter . ] .
we have summed up three flavors of neutrino and the parameters @xmath34 and @xmath35 are defined as @xmath36 numerically , we have @xmath37 , @xmath38 and @xmath39 respectively .
the dark @xmath4 interaction with nucleon can be characterized as @xmath41 with @xmath42 and @xmath43 respectively .
only the vector current part contributes to the matrix element @xmath44 if parity is conserved .
it is interesting that the couplings for proton and neutron are identical to @xmath45 and @xmath46 except the sign flip by opposite electric charge and weak isospin .
thus , we deduce that @xmath47 according to refs .
@xcite , we have @xmath48 in the isospin symmetry limit and @xmath49 can be both isovector and isoscalar .
a isospin breaking derivation for the transition is presented in ref .
@xcite . however , the modification is only about 20% .
since our results are not much affected , we present the isospin symmetry limit of the transition rate for simplicity . ] , we have @xcite @xmath50^{3/2}.\ ] ] the branching ratio of @xmath51 in our scenario can be deduced from eq .
( [ gammae ] ) and eq .
( [ gammanu ] ) , @xmath52 where we have applied the relations @xmath53 , @xmath54 , which particularly hold in dark @xmath5 model .
in addition , as pointed in ref .
@xcite and constraints summarized in ref .
@xcite , two most severe bounds from the search of dark photon in the relevant mass range are obtained by na48/2 @xcite and e141 experiments @xcite .
na48/2 gives an upper bound requiring @xmath55 at 90% c.l .
we translate the constraint into our framework by interpretting the process as @xmath56 . again
the branching ratio modification is taken into account , @xmath57 e141 is a electron beam dump experiment at slac which searches for a dark photon bremsstrahlung resulting from electrons incident on a nuclear target @xcite .
the experiment sets a lower bound for the coupling strength in our scenario @xmath58 combining the condition of eq .
( [ constraint1 ] ) and two constraints , eq . ( [ constraint2 ] ) and eq .
( [ constraint3 ] ) , we plot the allowed parameter region for @xmath59 in figure [ fig : parameterspace ] .
the green band is the allowed region to fit the @xmath49 anomaly while the purple shaded and pink shaded areas are excluded by the beam dump and na48/2 experiments , respectively .
one observes that @xmath60 lies within a narrow region between @xmath61 and @xmath62 and @xmath63 is constrained in the range of @xmath64 .
the allowed coupling strength of @xmath4 to proton is relatively smaller than the coupling strength of @xmath4 to neutron .
hence a @xmath9 feature is suggested by the measurement of krasznahorkay _ et al _ .. for the hidden @xmath3 model
, the same constraints also apply to @xmath66 as shown in eq .
( [ ep = ae ] ) and eq .
( [ en = anu ] ) .
therefore , the constraints from short baseline accelerator and/or reactor neutrino - electron scattering experiments must be taken into account @xcite .
a global analysis on the nonstandard interactions that are deviated from the sm predictions is presented in ref .
we take the effective lagrangian approach by integrating the intermediate @xmath4 boson , which yields the following bounds @xmath67 using eqs .
( [ ep = ae ] ) and ( [ en = anu ] ) , these bounds can be translated into constraints for @xmath63 and @xmath60 .
these constraints are so stringent that they are incompatible with the experimental constraints we just derived in the framework of generic hidden @xmath3 model . the allowed parameter space on @xmath59 or @xmath66 plane in generic hidden @xmath3 model . e141 and
na48/2 exclusion regions are indicated .
the green band is the allowed parameter space resulting from @xmath49 anomaly with the error of @xmath4 mass taken into account .
the allowed narrow band is however incompatible with the texono @xmath65 scattering experiment @xcite.,scaledwidth=45.0% ]
to accommodate the new light gauge boson indicated in @xmath49 anomaly as well as @xmath3 portal scenario , we are led to consider models with non - hidden @xmath3 gauge symmetry and mev - scale dm mev are viable .
we concentrate our discussion on mev - scale dm in this paper . ] .
non - hidden @xmath3 charge suggests a certain linear combination of sm quantum number and/or other hidden charge .
phenomenologically , such models will include a new set of direct gauge - fermion couplings .
the interplay between these couplings with the gauge boson mixings will modify the relations among quark and lepton couplings such as eq .
( [ ep = ae ] ) and eq .
( [ en = anu ] ) .
there are various ways of model - building to impose such non - hidden @xmath3 gauge symmetry motivated by @xmath49 anomaly @xcite . in this paper
, we do not intend to study these models in detail but rather assume that the couplings of @xmath4 to various fermions are not correlated . in particular , we assume the severe constraint from @xmath65 scattering can be alleviated model with @xmath16 the baryon number .
in such a model the neutrino-@xmath4 coupling vanishes , thus the texono bounds can be evaded .
an anomaly free @xmath68 model is proposed in ref .
@xcite ] . in this section ,
we first discuss various constraints on those generic non - hidden @xmath3 models , including constraints from cosmology and constraints from dm direct search with dm - nucleus scattering .
we then discuss the sensitivities of future si and ge detectors to dm - electron scatterings for light dm in the mev mass range .
dm relic abundance requires the wimp annihilation cross section to be around @xmath69 . in our study
we shall consider other cosmological constraints on @xmath70 .
for cosmic microwave background ( cmb ) the additional injection of energy via the dm annihilations will increase the ionization fraction on the cmb anisotropy .
hence it will suppress the power spectrum at small angular scales due to the broadening of the last scattering surface , and also enhance the polarization power spectra at low multipoles due to the increasing probability of the thomson scattering .
the planck data puts strong bounds on s - wave annihilation cross section for the dm mass in a range of sub - mev to 100 gev @xcite .
in particular , the constraints are stringent if the fraction of electron final state is non - negligible .
we note that the @xmath49 anomaly indicates a substantial fraction of @xmath71 final state in @xmath4 decays .
in such a case the planck data sets the limits of s - wave @xmath72 for mev - scale dm @xcite .
therefore , we consider the process , @xmath73 , as the dm annihilation channel in this scenario . in the parity conservation limit , this process is mostly p - wave as pointed out in @xcite .
we include the sommerfeld enhancement factor @xcite and find @xmath74 , the analogous fine structure constant for @xmath15 gauge interaction , is about @xmath75 to satisfy the thermal relic abundance .
finally , in order to prevent the photodissociation to alter the light element abundances during the big bang nucleosynthesis , one requires the lifetime of @xmath76 to be less than 1 second in the early universe @xcite .
the mixing parameter is constrained to be @xmath77 @xcite . in this paper , we focus on the range that @xmath78 and make a threshold cut at @xmath79 for the study of dm - electron scattering process . in the limit of zero momentum transfer ( @xmath80 ) , the dm - nucleus scattering cross section is given by @xmath81 ^ 2\ ] ] where @xmath5 and @xmath82 are proton number and mass number , respectively , and @xmath83 is the dm - nucleus reduced mass . in this case ,
dm - neutron cross section @xmath84 is the main contribution to @xmath85 , while dm - proton cross section @xmath86 is suppressed by the factor @xmath87 .
additionally , the force mediator is light enough such that a suppression factor @xmath88 due to the momentum transfer should be included in the propagator . for a single nucleon , @xmath89 when @xmath90 from the halo dm . in our analysis
, we consider this momentum suppression in the calculation of @xmath91 .
we found that the predicted dm - nucleon cross section based upon the allowed range of @xmath59 is far beyond the current dm direct search constraints .
the theoretical predictions of @xmath84 from the allowed range of @xmath59 parameter space is shown .
the purple shaded regions are the predicted dm - neutron cross section .
the exclusion lines by cresst - ii(2015 ) @xcite , damic(2012 ) @xcite , cdmslite(2015 ) @xcite , and lux(2015 ) @xcite are presented for @xmath92 and @xmath93 , respectively , where the latter ratio is the protophobic scenario favored by @xmath49 experiment .
it can be seen that the dm direct search excludes dm down to @xmath94 for both @xmath95 ratios .
, scaledwidth=50.0% ] we show in fig .
[ fig : dm - nucleon_sigma ] the theoretical predictions of @xmath84 for dm mass between 0.5 - 10 gev .
the corresponding dm direct search bounds obtained by cresst - ii(2015 ) @xcite , damic(2012 ) @xcite , cdmslite(2015 ) @xcite , and lux @xcite are also shown . for dm mass higher than 10 gev ,
the direct searches give even more stringent bounds .
it is seen that the predicted @xmath84 based upon the allowed parameter region as indicated in fig .
[ fig : parameterspace ] is excluded .
therefore , the dm direct searches also disfavor the non - hidden @xmath4 model for the dm mass above 500 mev . in other words , for this range of dm mass , the current dm direct search bounds are incompatible with the parameter space derived from @xmath49 anomaly .
the red shaded regions are the theoretical predictions of @xmath96 with @xmath97 lying in the range of @xmath98 for @xmath99 .
the gray shaded region are the predictions with @xmath97 lying in the range of @xmath100 .
xenon 10 @xcite excludes the parameter space of @xmath101 for @xmath102 mev .
the projected sensitivities of si and ge detectors with threshold charges of @xmath6 are represented by the green solid line and the orange dashed line , respectively.,scaledwidth=50.0% ] the conventional dm direct search looks for the nuclear recoils .
however , the nuclear recoil energy , @xmath103 ev , is sub - ev for the mev - scale dm and is far below the threshold energies in current experiments .
is the dm velocity dispersion in the halo .
] instead of detecting the nuclear recoils , it was suggested to detect the dm - electron scattering as the dm signal @xcite . the dm - electron cross section @xmath96 is given by @xcite @xmath104 where @xmath105 is the dm - electron reduced mass .
we show the theoretical predictions on @xmath96 and xenon 10 exclusion region in fig .
[ fig : dm - electron_sigma ] .
the blue shaded area is excluded by xenon 10 based on the dm - electron scattering and the capability of charge threshold @xmath106 @xcite .
it can be seen that @xmath107 is excluded for @xmath102 mev .
the projected sensitivities for silicon and germanium targets with improved @xmath108 are presented as green solid line and orange dashed line , respectively @xcite .
the red shaded region is the theoretical predictions of @xmath96 corresponding to the range of @xmath109 .
@xmath110 is taken to be between @xmath111 and @xmath112 where 20 mev roughly corresponds to the dm mass lower bound for the annihilation @xmath113 to take place while 500 mev is the energy threshold for dm direct search with dm - nucleon scattering .
our calculation on @xmath96 shows that the mev dm under @xmath4 model can be tested by future experiments .
motivated by the possible existence of a new light boson from the experiment of krasznahorkay _
et al_. @xcite , we investigate the @xmath4-portal models .
the additional @xmath3 gauge symmetry may correspond to a hidden charge or a non - hidden charge .
the reactor neutrino - electron scattering sets a severe constraint which excludes the parameter space of generic hidden @xmath3 model .
since the generic hidden @xmath3 model is disfavored , we are led to considering non - hidden @xmath3 portal models . using the current dm direct search data
, we have shown that such models are also disfavored for dm mass larger than 500 mev . to probe dm lighter than 500 mev in non - hidden @xmath3 models ,
we propose direct searches based upon dm - electron scatterings .
the sensitivities of future si and ge detectors to @xmath96 are given in fig .
[ fig : dm - electron_sigma ] . hence the sensitivities of these detectors to the couplings strength @xmath97 can be determined accordingly .
csc is supported by ministry of science and technology ( most ) , taiwan , r.o.c . under grant no . 104 - 2112-m-001 - 042-my3 ;
gll and yhl are supported by ministry of science and technology , taiwan under grant no .
104 - 2112-m-009 - 021 .
frx is supported by nsfc under grant no .
11605076 as well as the fundamental research funds for the central universities in china under the grant no .
especially frx would like to acknowledge the hospitality of institute of physics , academia sinica , at which part of the work was done .
1 a. j. krasznahorkay _ et al . _ ,
* 116 * , no . 4 , 042501 ( 2016 ) doi:10.1103/physrevlett.116.042501 [ arxiv:1504.01527 [ nucl - ex ] ]
. j. l. feng , b. fornal , i. galon , s. gardner , j. smolinsky , t. m. p. tait and p. tanedo , phys .
lett . * 117 * , no . 7 , 071803 ( 2016 ) [ arxiv:1604.07411 [ hep - ph ] ] .
p. h. gu and x. g. he ,
arxiv:1606.05171 [ hep - ph ] . l. b. chen , y. liang and c. f. qiao , arxiv:1607.03970 [ hep - ph ] .
y. liang , l. b. chen and c. f. qiao , arxiv:1607.08309 [ hep - ph ] . c. h. chen and t. nomura , arxiv:1608.02311 [ hep - ph ] .
j. l. feng , b. fornal , i. galon , s. gardner , j. smolinsky , t. m. p. tait and p. tanedo , arxiv:1608.03591 [ hep - ph ] . l. b. jia and x. q. li , arxiv:1608.05443 [ hep - ph ] . t. kitahara and y. yamamoto , arxiv:1609.01605 [ hep - ph ] .
u. ellwanger and s. moretti , arxiv:1609.01669 [ hep - ph ] .
x. chu , c. garcia - cely and t. hambye , arxiv:1609.00399 [ hep - ph ] .
m. deniz _ et al . _
[ texono collaboration ] , phys . rev .
d * 81 * , 072001 ( 2010 ) [ arxiv:0911.1597 [ hep - ex ] ] .
h. davoudiasl , h. s. lee and w. j. marciano , phys .
* 109 * , 031802 ( 2012 ) [ arxiv:1205.2709 [ hep - ph ] ] .
f. xu , jhep * 1506 * , 170 ( 2015 ) [ arxiv:1504.07415 [ hep - ph ] ] .
j. alexander _ et al .
_ , arxiv:1608.08632 [ hep - ph ] . j. r. batley _ et al . _
[ na48/2 collaboration ] , phys .
b * 746 * , 178 ( 2015 ) [ arxiv:1504.00607 [ hep - ex ] ] .
e. m. riordan _ et al .
_ , phys .
lett . *
59 * , 755 ( 1987 ) .
l. b. auerbach _ et al . _
[ lsnd collaboration ] , phys .
d * 63 * , 112001 ( 2001 ) [ hep - ex/0101039 ] .
z. daraktchieva _ et al . _
[ munu collaboration ] , phys .
b * 615 * , 153 ( 2005 ) [ hep - ex/0502037 ] .
a. n. khan , phys .
d * 93 * , no .
9 , 093019 ( 2016 ) [ arxiv:1605.09284 [ hep - ph ] ] .
g. angloher _ et al . _
[ cresst collaboration ] , arxiv:1509.01515 [ astro-ph.co ] .
j. barreto _ et al .
_ [ damic collaboration ] , phys .
b * 711 * , 264 ( 2012 ) [ arxiv:1105.5191 [ astro-ph.im ] ] .
r. agnese _ et al . _
[ supercdms collaboration ] , [ arxiv:1509.02448 [ astro-ph.co ] ] .
d. s. akerib _ et al . _
[ lux collaboration ] , phys .
lett . * 116 * , 161301 ( 2016 ) [ arxiv:1512.03506 [ astro-ph.co ] ] .
t. r. slatyer , phys .
d * 93 * , no . 2 , 023527 ( 2016 ) [ arxiv:1506.03811 [ hep - ph ] ] . h. an , m. b. wise and y. zhang , arxiv:1606.02305 [ hep - ph ] .
r. iengo , jhep * 0905 * , 024 ( 2009 ) [ arxiv:0902.0688 [ hep - ph ] ] .
s. cassel , j. phys .
g * 37 * , 105009 ( 2010 ) [ arxiv:0903.5307 [ hep - ph ] ] .
e. holtmann , m. kawasaki , k. kohri and t. moroi , phys .
d * 60 * , 023506 ( 1999 ) [ hep - ph/9805405 ] .
t. lin , h. b. yu and k. m. zurek , phys .
d * 85 * , 063503 ( 2012 ) [ arxiv:1111.0293 [ hep - ph ] ] .
r. bernabei _ et al .
_ , phys . rev .
d * 77 * , 023506 ( 2008 ) [ arxiv:0712.0562 [ astro - ph ] ] .
a. dedes , i. giomataris , k. suxho and j. d. vergados , nucl .
b * 826 * , 148 ( 2010 ) [ arxiv:0907.0758 [ hep - ph ] ] .
j. kopp , v. niro , t. schwetz and j. zupan , phys .
d * 80 * , 083502 ( 2009 ) [ arxiv:0907.3159 [ hep - ph ] ] .
r. essig , j. mardon and t. volansky , phys .
d * 85 * , 076007 ( 2012 ) [ arxiv:1108.5383 [ hep - ph ] ] .
r. essig , a. manalaysay , j. mardon , p. sorensen and t. volansky , phys .
lett . * 109 * , 021301 ( 2012 ) [ arxiv:1206.2644 [ astro-ph.co ] ] .
r. essig , m. fernandez - serra , j. mardon , a. soto , t. volansky and t. t. yu , jhep * 1605 * , 046 ( 2016 ) [ arxiv:1509.01598 [ hep - ph ] ] . | the experiment of krasznahorkay _ et al _ observed the transition of a @xmath0 excited state to its ground state and accompanied by an emission of @xmath1 pair with 17 mev invariant mass .
this 6.8@xmath2 anomaly can be fitted by a new light gauge boson .
we consider the new particle as a @xmath3 gauge boson , @xmath4 , which plays as a portal linking dark sector and visible sector .
in particular , we study the new @xmath3 gauge symmetry as a hidden or non - hidden group separately . the generic hidden @xmath3 model , referred to as dark @xmath5 model , is excluded by imposing various experimental constraints . on the other hand , a non - hidden @xmath4 is allowed due to additional interactions between @xmath4 and standard model fermions .
we also study the implication of the dark matter direct search on such a scenario . we found the search for the dm - nucleon scattering excludes the range of dm mass above 500 mev .
however , the dm - electron scattering for mev - scale dm is still allowed by current constraints for non - hidden @xmath3 models .
it is possible to test the underlying @xmath3 portal model by the future si and ge detectors with @xmath6 threshold charges . |
relativistic wind models of gamma ray bursts ( grb ) are a recent development of the dissipative relativistic fireball scenario which is a natural consequence of observationally set requirements irrespective of the grb distance scale ( e.g. @xcite ) .
the observed bursts are expected to be produced in optically thin shocks in the later stages of the fireball expansion .
two types of shocks have been considered , based on different interpretations of the burst duration and variability .
first , the `` external '' deceleration shocks ( @xcite ; @xcite ) that develop at @xmath0 , due to the unavoidable interaction of the relativistic ejecta with the surrounding medium , give bursts whose duration is determined ( in the `` impulsive '' regime ) by the relativistic dynamic timescale @xmath1 , where @xmath2 is the lorentz factor of the expansion ( e.g. @xcite ; @xcite ; sari , narayan & piran 1996 ) .
second , the `` internal '' dissipation shocks in an unsteady relativistic wind outflow ( @xcite [ rm94 ] ; @xcite ) , where both the burst duration @xmath3 and time variability @xmath4 , are characteristic of the progenitor mechanism ( e.g. a disrupted disk accretion timescale , dynamic or turbulent timescales , etc . ) .
simple `` external '' shocks tend to produce fairly smooth light curves with a fast rise followed by an exponential decay ( fred ) , as observed in some bursts .
however , many bursts have multi - peaked , irregular light curves ; those may find a more natural explanation ( fenimore , madras & nayakshin 1996 ) in the relativistic wind scenario mentioned above . while the spectral properties of impulsive `` external '' shocks have been studied in some detail ( @xcite [ mrp94 ] ; @xcite ) ,
so far , only rough estimates have been made for wind `` internal '' shock spectra ( @xcite ; @xcite ) .
it is of great interest to explore the spectral properties of both types of shock scenarios , particularly in view of the increasingly sophisticated analyses of both grb light curves ( e.g. @xcite ; @xcite ) and spectral characteristics ( e.g. @xcite ) . in this letter
, we investigate the spectral properties of `` internal '' shocks for representative parameter values and discuss those properties in the light of anticipated results from hete .
in the unsteady relativistic wind model , the grb event is caused by the release of an amount of energy @xmath5 erg , inside a volume of typical dimension @xmath6 cm , over a timescale @xmath3 ( rm94 ) .
the average dimensionless entropy per particle @xmath7 determines the character of the overall flow .
the bulk lorentz factor of the flow increases as @xmath8 at first and later saturates to a value @xmath9 at @xmath10 , where @xmath11 for the values of @xmath12 that we consider here ( eq . [ [ eta_lim ] ] ) .
it is assumed that @xmath12 varies substantially , @xmath13 , on a timescale @xmath14 . as an example
, we consider shells of matter with two different @xmath12 values ejected at time intervals @xmath15 . after saturation these will coast with bulk lorentz factors @xmath16 and @xmath17 , where @xmath18 .
shells of different @xmath12 catch up with each other at a dissipation radius @xmath19 cm ( see rm94 , eq .
[ 4 ] ) , where @xmath20 and the subscript 2 ( 3 ) stands for quantities measured in units of @xmath21 ( @xmath22 ) .
this gives rise to a @xmath23 and a @xmath24 shock which are mirror images of each other in the contact discontinuity ( center of momentum , com ) frame .
the com moves in the lab frame with @xmath25 , for @xmath26 . in the com frame
the two shocks move in opposite directions with small lorentz factors @xmath27^{1/2}-1 \ge 10^{-2}$ ] , for @xmath28 , while in the lab frame both shocks move forward with lorentz factors @xmath29 and @xmath30 respectively .
the shocks compress the gas to a comoving frame ( primed quantities refer to the @xmath31 frame , which is the com frame ) density @xmath32 where @xmath33 is the normalized opening solid angle of the wind ( @xmath34 if spherical ) .
the bulk kinetic energy carried by the flow is randomized in the shocks with a mechanical efficiency @xmath35 ( eq . [ 1 ] in rm94 ) . in order for the shocks to produce a non - thermal spectrum and radiate a substantial portion of the wind energy ,
the dissipation must occur beyond both the wind photosphere ( see section 2.2 in rm94 ) and the saturation radius .
this restricts @xmath12 to the range @xmath36 the collisionless `` internal '' shocks accelerate particles to relativistic energies with a nonthermal distribution .
we parameterize the post - shock electron energy by a factor @xmath37 , so that the average random electron lorentz factor is @xmath38 ( with the protons being @xmath39 ) .
an injection fraction @xmath40 of the post - shock electrons is accelerated to a power - law distribution ( @xmath41 , for @xmath42 ) .
we take @xmath43 , so that most of the energy is carried by the low energy part of the electron spectrum ( @xmath44 ) . a measure of the efficiency of the transfer of energy between protons and electrons behind the shock
is given by @xmath45 .
here , we will use @xmath46 and @xmath47 ( but see @xcite ) and @xmath48 which is consistent with the fits to observed spectra ( @xcite ) .
the electrons get accelerated at the expense of the randomized proton energy behind the shocks , which , for an individual shock ( of which there are @xmath49 ) , represents a fraction @xmath50 of the total energy @xmath51 .
magnetic fields in the shocks can be due to a frozen - in component from the progenitor , or may build up by turbulence behind the shocks .
we parameterize the strength of the magnetic field in the shocks by the fraction @xmath52 of the magnetic energy density to the particle random energy density ( @xmath53 ) .
the comoving magnetic field therefore is @xmath54 the relativistic electrons will lose energy due to synchrotron radiation , and inverse compton ( ic ) scattering of this radiation .
the respective radiative efficiencies are @xmath55 , and @xmath56 .
the timescales are defined below .
the comoving synchrotron timescale is determined by the least energetic electrons : @xmath57 where the subscript _ sy _ ( _ ic _ ) refers to the synchrotron ( ic ) emitting shock . the ic cooling depends on , and competes with , synchrotron cooling .
ic cooling dominates if the magnetic field is weak , a large fraction of electrons are accelerated and share the protons momentum very efficiently ( i.e. @xmath58 ) ; while synchrotron cooling dominates in the opposite case .
the ic timescale ( @xmath59 ) for the two limiting cases is @xmath60 & \mbox{ms $ \;\;\;$if ic dominates}\\ \label{t_ic_ic } \\
4.95\ ; \displaystyle \;\frac{q^4}{\varepsilon_{sh } } \frac{\theta^2 t_w t_{var}^2}{e_{51}}\ ; \frac{\eta_2 \gamma_{s_2}^5 } { \sqrt{\zeta \lambda
\kappa_{ic_3 } < \gamma_3 ^ 2 > } } & \mbox{ms \quad if synchr .
dominates } \label{eq : t_ic_sy}. \end{array } \right.\ ] ] here @xmath61 is the electron lorenz factor that corresponds to the peak emission frequency , @xmath62 $ ] , @xmath63 is defined below equation ( [ v_abs ] ) and @xmath64 .
the lab frame shell width ( shocked region ) is @xmath65 , and the comoving crossing time @xmath66 s provides an estimate for the adiabatic loss time .
the spectrum of an `` internal '' shock burst consists of two synchrotron and four ic components ( coming from all shocks combinations ) .
the synchrotron components are characterized by up to three _ break frequencies _ , given below in the lab frame : \i ) the @xmath67 break frequency @xmath68 , which in the lab frame gets blue - shifted by @xmath29 ( @xmath69 ) for the forward ( reverse ) shock . using @xmath70 $ ] to refer to the shock , we have @xmath71 \ii ) the self absorption frequency @xmath72 .
if the electron power law starts at low enough energies , the radiation field becomes optically thick at a frequency determined by @xmath73 , and is obtained by solving a non - linear algebraic equation ; in the limit @xmath74 it is @xmath75 where @xmath76 .
\iii ) the frequency where the photon spectrum of the minimum energy electrons becomes optically thick ( @xmath77 ) , and below which it assumes a rayleigh - jeans spectrum slope .
this happens at @xmath78 when @xmath79 ; if @xmath80 it is @xmath81 \iv ) a cutoff is expected at @xmath82 , where @xmath83 is the electron energy in each shock at which the electron power law cuts off due to radiative or adiabatic losses .
it is determined by @xmath84 , where @xmath85 is the electron acceleration timescale in the shocks ( a multiple @xmath86 of the inverse gyro - synchrotron frequency ) , @xmath87 s , and @xmath88 is the minimum of all the radiation and adiabatic cooling timescales involved , @xmath89
. each synchrotron component can be characterized by three frequencies in ascending order , @xmath90 , where @xmath91 and @xmath92 for the reverse and @xmath93 for the forward shock .
similarly , the pure and combined ic spectra are characterized by the frequencies @xmath94 where , @xmath95 ( 1 corresponds to _ ic reverse _ , 2 to _ ic forward _
, 3 to _ ic reverse - forward _ and 4 to _ ic forward - reverse _ ) , and if j is odd , @xmath96 , otherwise @xmath97 .
the shape of each component depends on the relationship between the relevant @xmath98 and @xmath83 . in a power per logarithmic frequency interval plot ,
the fluence of each component exhibits a peak of @xmath99 , where @xmath100 for synchrotron ( ic ) , and @xmath101 is the luminosity distance corresponding to @xmath102 for a flat universe , with @xmath103 ( @xmath104 ) .
the spectrum is obtained by adding up these six components . in practice ,
the forward and reverse components have values very close to each other and they essentially merge .
the resultant spectrum is then checked for the effects of pair production ; the @xmath105 optical depth is calculated for each comoving frequency above @xmath106 using the number density of photons above the corresponding threshold , as obtained from the initial spectrum ; finally the spectrum is modified accordingly .
we note that most of the scattering in our spectra occurs in the thomson regime , unlike in @xcite , who consider large @xmath107 in the framework of impulsive shock models .
( klein - nishina corrections may become relevant in a few @xmath108 cases , but at frequencies which lie in the @xmath105 absorbed part of the spectrum ) .
we discuss here the properties of some representative spectra . we assume a total event energy of @xmath109 erg and a geometry of a spherical section , or jet , of opening angle @xmath110 ( the physics is the same as in a spherical wind , provided @xmath111 ) .
we have investigated a range of dynamic parameters ( @xmath112 and @xmath15 ) , and used @xmath113 , and @xmath114 . in figure [ fig_long ] we present spectra for a long burst ( @xmath115 s ) and in figure [ fig_short ] for a short one ( @xmath116 s ) , for different @xmath117 and @xmath37 values .
the sharper spectral features present would be smoothed by inhomogeneities in a real flow . the range of @xmath12 where a substantial fraction of the wind energy is radiated with nonthermal spectra is considerably restricted ( eq . [ [ eta_lim ] ] ) and implies lower values than in impulsive models .
larger @xmath12 may be difficult to produce in a natural way , while lower values lead to shocks below the photosphere that make bursts too dim to be observed .
the allowed volume of parameter space is not very large ( @xcite [ pm96 ] ) .
nonetheless , it allows for an appreciable variety of spectra , since models differing only slightly in @xmath12 can produce significantly different spectra .
this is due to the strong dependence of the photon number density on @xmath118 ( i.e. @xmath119 ) which determines pair production .
lower @xmath37 spectra are less affected by pair production , since they contain less energetic electrons and therefore fewer photons .
generally , other parameters being equal , the higher @xmath12 models tend to produce spectra that span over a wider range of frequencies . for spectra like the majority of the ones observed by batse ( ) values of @xmath120 and @xmath121 are excluded .
low @xmath37 spectra are either too dim or too soft ; high @xmath52 brings up the synchrotron component and may be appropriate for a small percentage of bursts with low frequency excess ( @xcite ) . in general , for a given @xmath37 value a wide range of @xmath52 values ( 3 - 6 orders of magnitude ) is allowed , the trend being that higher @xmath37 values must combine with lower @xmath52 values , fairly independently of @xmath12 ( within the allowed range of values ) .
this trend is due to the fact that lower magnetic fields require higher electron energies in order for the break to fall in the batse window ( e.g. eq . [ [ v_sy_min ] ] , [ [ v_abs ] ] ) .
high energy power laws ( like those reported in @xcite ) are common ( see fig .
[ fig_short ] and right column of fig .
[ fig_long ] ) . for a more complete discussion
see pm96 .
the effect of a longer @xmath3 is to push the pair cutoff to higher frequencies , because , for fixed @xmath51 , the flux and photon density are lower .
for the same @xmath12 , @xmath37 and @xmath15 , longer bursts require higher @xmath52 .
a greater @xmath15 with the rest of the parameters unchanged would again require higher @xmath52 in order to produce observed - like spectra .
the cases considered here were chosen with @xmath122 s.
an unsteady relativistic wind provides an attractive scenario for the generation of grb , since it requires smaller bulk lorentz factors than the impulsive models , i.e. it can accommodate higher baryon loads ( rm94 ) .
in addition , the lack of kinematic restrictions ( e.g. @xcite ) allows it , in principle , to describe events with arbitrarily complex light curves .
we have calculated spectra from optical through tev frequencies for bursts with a range of durations and variability timescales . at cosmological distances the total energies and photon densities implied by the model are likely to turn those spectra optically thick to @xmath105 pair production .
most of them may therefore be missed by , or show a high energy cut - off in , the egret window , but they would be prominent in the batse and hete gamma - ray windows .
low frequency ( down to 5 kev ) excess reported recently ( @xcite ) may be attributed to a pronounced synchrotron component due to a relatively high magnetic field , in a fraction of the bursts .
the continued propagation of an unsteady wind flow should eventually lead to its deceleration by the surrounding medium .
if the latter is of any appreciable density , it could lead to another burst with the `` external '' shock characteristics ( mrp94 , @xcite ) , provided that not all of the wind energy was radiated away by the `` internal '' ones ( @xmath123 ) .
the spectra of the `` internal '' shocks are different from those coming from the `` external '' ones , the main differences being that the former cover a narrower range of frequencies and most have high energy cutoffs due to pair production opacity ( pm96 ) .
if grb progenitors have escaped their parent galaxies and are in a low density intergalactic medium ( @xmath124 @xmath125 ) , the `` external '' impulsive shocks would be of long duration ( @xmath126 s ) and low intensity , and most would be totally missed , hence the grb could be entirely due to unsteady wind `` internal '' shocks .
if the grb occur in a denser medium ( e.g. the galactic ism ) , the last peak of multiple - peaked grb would be smooth and fred - like but the earlier peaks , due to unsteady wind `` internal '' shocks , may be arbitrarily complicated .
those might also be responsible for delayed gev emission ( @xcite ) .
a number of unknown parameters enter into the calculation of grb models , and hete may provide the information required to narrow the range of their allowed values , as well as a test for the general scenario of unsteady relativistic winds . for the hete sensitivities indicated in figures [ fig_long ] and [ fig_short ] , we expect that some bursts will show a simultaneous x - ray counterpart
. however , detection by the ultraviolet transient camera ( utc ) on hete should be rare for the majority of bursts ( note though that figures [ fig_long ] and [ fig_short ] refer to @xmath127 1 distances , and that at a few @xmath128 mpc a bright burst would be @xmath129 times brighter , thus increasing the likelihood of detection by the utc ) .
for the majority of faint bursts , a detection in the ultraviolet would be possible only for low values of @xmath37 , resulting in a very broad and flat spectrum with upper cutoffs , if any , only at the highest energies . | we calculate the spectra expected from unsteady relativistic wind models of gamma - ray bursts , suitable for events of arbitrary duration . the spectral energy distribution of the burst is calculated over photon energies spanning from ev to tev , for a range of event durations and variability timescales .
the relative strength of the emission at different wavelengths can provide valuable information on the particle acceleration , radiation mechanisms and the possible types of models .
psfig # 1#2.5pt |
positronium ( ps . ) , the positron - electron bound state , is the lightest known atomic system . it is bound and annihilates through the electromagnetic interaction alone . at the current level of experimental and theoretical precision this is the only interaction present in this system ( see _ e.g. _
this feature has made the positronium an ideal system for testing the accuracy of the qed calculations for bound states , in particular for the triplet ( @xmath5 ) state of ps .
, called orthopositronium ( @xmath6 ) . due to the odd - parity under c - transformation
the @xmath6 decays predominantly into three photons . as compared with the singlet ( @xmath7 ) state ( parapositronium ) decay rate , the @xmath6 decay rate , due to the phase - space and additional @xmath8 suppression factors , is about @xmath9 times smaller .
thus , it is more sensitive to potential admixtures of new interactions , which are not accommodated in the standard model .
the study of the @xmath6 system has a long history @xcite .
however , in spite of the substantial efforts devoted to the theoretical and experimental determination of the @xmath6 properties , there is a long - standing puzzle : the @xmath6 decay rate in vacuum measured by the ann arbor group , @xmath10 @xcite , has a @xmath11 discrepancy with respect to the predicted value @xmath12 @xcite ( see also @xcite ) .
this discrepancy has been recently confirmed by more precise calculations of adkins @xcite , including corrections of the order @xmath13 .
the result of the recent tokyo measurements of @xmath6 decay rate in low density sio@xmath14 powder corrected for matter effects @xcite agrees , within the errors , with the theoretical value of @xcite . however , the method to extract the value of the @xmath6 decay rate in vacuum from the measurement in matter is still under discussion @xcite .
thus , it is difficult to disagree with the adkins @xcite statement that : _ ... no conclusions can be drawn until the experimental situation is clarified_. various exotic @xmath6 decay modes have been investigated , with the hope that a relative contribution to the @xmath6 decay rate at the level of @xmath15 would solve the discrepancy at a ( 2@xmath16 ) level of @xmath17 . ] ( for review , see _ e.g. _
@xcite ) .
invisible decays of @xmath6 , such as , _
e.g. _ , @xmath18 , are excluded @xcite .
however , there is still a possible explanation of both the discrepancy and the tokyo results .
this is based on the existence of the @xmath19 decay in _ vacuum _ ( for more details see ref .
@xcite@xcite ) .
visible exotic decays of @xmath6 ( _ i.e. _ decays accompanied by at least one photon in the final state ) have been experimentally searched for : @xmath20 , @xmath21 and @xmath22 , where x is a new light particle and @xmath23 .
these decay modes have definitely been excluded @xcite@xcite as the cause of the @xmath6 lifetime discrepancy . the possibility of a new exotic three - body decay @xmath0 into a photon and a pair of new light weakly interacting particles had not yet been considered .
the photon energy spectrum from this decay mode has no peak , differently from the two - body decay @xmath20 .
thus , the sensitivity of previous searches based on the conventional peak - hunting technique is not sufficient to exclude this decay mode as the source of the discrepancy .
cosmological arguments place an indirect stringent constraint on the branching ratio of this decay mode @xcite .
however , one may argue that these arguments are depending on many assumptions .
the purpose of this experiment is to perform a direct search for the decay @xmath0 with the sensitivity needed to exclude this decay unambiguously .
the schematic illustration of the detector setup used in the experiment is shown in figure [ detector ] .
positrons from the @xmath24na source with an activity of 3.6 kbq are stopped in the sio@xmath14 aerogel target ( density @xmath25 0.1 g/@xmath26 , average grain size 50 - 100 ) where a fraction of them produces orthopositronium .
the source is prepared by sealing a drop of the @xmath24na solution between two 5 @xmath27 m mylar foils .
the source is sandwiched between two 120 @xmath27 m thick scintillators fabricated by squeezing 1 mm thick scintillator fibers .
the light produced by positrons crossing one of the scintillator fibers is delivered by the fibers to a pair of photomultipliers philips xp2020 ( pmt1 and pmt2 in figure [ detector ] ) .
the coincidence of the signals from pmt1 and pmt2 is used to tag the positron emission ( @xmath28 trigger ) and to define the time @xmath29 of positronium formation in the target .
the photons produced by the positronium annihilation are detected by a 4@xmath30 crystal calorimeter , schematically shown in figure [ detector ] .
the detector is composed of an inner and an outer ring of 8 and 14 bgo crystals , respectively , surrounding the target region .
two additional bgo crystals serve as endcaps .
each crystal has a hexagonal cross - section with an inner diameter of 55 mm and a length of 200 mm . for a more detailed description of the bgo crystals
see @xcite .
one of the endcap counters , hereafter called trigger counter , is also used for the measurement of the annihilation time of the positronium relative to @xmath29 .
when one of the annihilation photons is detected by the trigger counter , the recoil photons are detected by the other bgo counters , hereafter called the veto detector .
the detector is calibrated and monitored internally using the 511 kev annihilation photons and the 1.27 mev photon emitted by the @xmath24na source in association with the positron emission .
variations of the energy scale are within @xmath31 and are corrected on the basis of an internal calibration procedure . the particles @xmath32 and @xmath33 are assumed to be weakly interacting penetrating particles .
thus the experimental signature of the @xmath0 decay is the presence of energy deposition in the trigger counter , within a time interval consistent with the delayed annihilation of the @xmath6 in the target , and no energy deposition in the veto detector . in order to decrease the contribution of two photon events from collisional quenching of @xmath6 ,
the sio@xmath14 target is dehydrated in a vacuum of 10@xmath34 torr at a temperature about 200@xmath35c during two hours before installation inside the detector .
furthermore , during the data taking period high purity dry nitrogen is flowing through the target .
this procedure increases the lifetime of @xmath6 in the prepared sample of the sio@xmath14 aerogel from @xmath36 70 ns to 132 ns giving a two photon suppression factor of @xmath37 .
the signal of positron emission is also used to open a 3 @xmath27s gate for complete recording of the signals from the bgo counters and the scintillator pmts .
an event is recorded if an energy deposition @xmath38 40 kev is detected in the trigger counter within the 3 @xmath27s gate .
a camac - vme system interfaced to a personal computer is used for data acquisition . for each event
the following quantities are recorded : * the amplitudes , @xmath39 and @xmath40 , of the pulses from pmt1 and pmt2 and the time interval @xmath41 between them ; * the energy deposition @xmath42 in the trigger counter and the time interval @xmath43 between the trigger counter pulse and the @xmath28 trigger ; * the pulses from each of the 23 bgo crystals for the measurement of the total energy deposition in the veto detector .
the search for the @xmath0 decay described in this paper uses a data sample of @xmath44 recorded events .
candidate events for o - ps decays , with at least one photon in the final state , are identified by the following selection criteria , illustrated in figure [ cuts ] : * the pmt1 and pmt2 pulses from the scintillator fibers in the energy range @xmath45 kev .
this cut is applied to reject signals of fake positrons from accidental coincidences due to pmt noise ; * the time difference between pmt1 and pmt2 pulses @xmath46 ns ; * the time difference between the pmt pulses from the scintillator fibers and the trigger counter signal is required to be in the range @xmath47 ns .
the lower cut is used to eliminate background from the tail of prompt positron annihilation , shown in figure 2c ) .
the upper cut is chosen to eliminate a region where the background from accidental coincidences becomes dominant . * the presence in any single bgo crystal , except the two endcaps , of an energy deposition in the range @xmath48 kev .
this criterion is applied to select events with a 1.27 mev photon , emitted by the source at the same time as the positron . after imposing the above requirements 338786 candidates events
are found .
the distribution of these events in the scatter plot @xmath49 vs @xmath42 is shown in figure 3a ) . here @xmath49 is defined as the sum over all bgo crystals minus the energy deposited in the trigger crystal and the energy deposited by the 1.27 mev photon : @xmath50 .
as expected for positronium decay into photons , the events accumulate around the line @xmath49+@xmath42=2@xmath51 , @xmath51=511 kev being the electron or positron mass . in figure 3b ) the region of the scatter plot for @xmath49@xmath52 kev
the box for @xmath53 20 kev and @xmath54 400 kev defines the signal region .
the cut on the trigger energy @xmath55 400 kev is applied to reject events where two photons from the @xmath56 decay deposit energy in the trigger counter and to eliminate background from @xmath6 annihilation into two photons due to collisional quenching . the cut @xmath53 20 kev is chosen according to the measured width of the zero - energy peak in the veto detector . for this measurement , prompt positron annihilation events ( @xmath57 ns ) are selected with one 511 kev photon in the trigger counter . the spectrum measured in the veto counter for zero - energy signal , when the second photon escapes detection , shows that 95% of the zero - energy signal is collected in the region @xmath58 kev .
the width of the zero - energy signal spectrum is determined mostly by the overlap of close in time events , while the contribution from the adc pedestals fluctuation is found to be negligible .
the distribution of the veto energy for @xmath53 500 kev is shown in figure [ result]c ) .
the distribution of the events with @xmath59 kev can be extrapolated into the signal region to evalute the background contribution in this region .
a fit shown in figure [ result]c ) results in a background estimate of @xmath60 events , where the error is evaluated from the uncertainty related to the extrapolation procedure itself . as shown in figure [ result]b ) and [ result]c ) one event is found in the signal region .
this is consistent with the background evaluation described above .
hence , no evidence for the decay @xmath0 is found .
this result allows us to set an upper limit on @xmath1 from the 90% confidence level ( cl ) upper limit on the expected number of signal events , @xmath61 . because of the uncertainty on the background estimate we have chosen conservatively not to subtract the background .
using poisson statistics @xcite , for 1 event observed and 0 background event expected , the limit is @xmath61=3.8 events . given the measured number @xmath62 of @xmath6 decays in the target , the numbers @xmath63 and @xmath64 of @xmath0 and @xmath56 decays , respectively , detected in the trigger counter , are @xmath65 resulting in @xmath66 here @xmath67 and @xmath68 are the efficiencies of the trigger counter for @xmath0 and for @xmath69 decays , respectively .
these efficiencies are evaluated with a monte carlo simulation of the detector response to both decay modes . for the @xmath0 decay , phase space distribution of the photon and of the two particles @xmath32 and
@xmath33 are considered .
it is found that for the kinematically allowed region of masses @xmath70 kev the efficiency ratio varies in the range @xmath71 .
the lower value corresponds to @xmath67 calculated for @xmath72=0 kev .
the upper value corresponds to @xmath72=900 kev .
the efficiency for the @xmath0 is relatively high over this kinematically allowed region owing to our low trigger threshold of 40 kev . for the determination of the limit
we take conservatively the value @xmath73 .
the number of @xmath6 decays in the target is measured from the decay curve of figure 2c by fitting the distribution to the function @xmath74 ( @xmath75 is the accidental background ) starting from the time @xmath76 ns when @xmath6 is completely thermalized in the target . comparing the measured lifetime @xmath77 ns with the lifetime in vacuum ( 141.9 ns ) the probability of @xmath6 quenching in the target
is found to be 6.6% . correcting for this efficiency factor the total number of detected @xmath6 is determined to be 3.2@xmath78 . finally , our 90 % cl limit on @xmath1 for the photon energy range @xmath79 kev and masses @xmath80 kev is @xmath81 this limit is more than 20 times smaller than the value needed to explain the discrepancy in @xmath6 decay rate .
thus , the @xmath0 decay is definitely excluded as a possible origin of the discrepancy .
we thank the paul scherrer institute ( psi ) for providing us with the bgo crystals and w. fetscher for advice .
we gratefully acknowledge the help of g. roubaud for the @xmath24na source preparation .
we wish to thank n.v .
krasnikov for useful discussions .
+ part of this work was supported by the swiss national science foundation . | we report on a direct search for a three - body decay of the orthopositronium into a photon and two penetrating particles , @xmath0 .
the existence of this decay could explain the discrepancy between the measured and the predicted values of the orthopositronium decay rate . from the analysis of the collected data
a single candidate event is found , consistent with the expected background .
this allows to set an upper limit on the branching ratio @xmath1@xmath2 ( at the 90% confidence level ) , for the photon energy in the range from 40 kev @xmath3 400 kev and for mass values in the kinematical range @xmath4 900 kev .
this result unambiguously excludes the @xmath0 decay mode as the origin of the discrepancy . , , , , , , , , , orthopositronium decay , new particles , experimental tests |
recently , the role played by geometrical frustration in strongly correlated electron systems has attracted renewed interests.@xcite the pyrochlore lattice , namely , corner - sharing tetrahedra ( fig.1 ) , is one of the typical system in which the geometrical frustration is crucial to determine its properties . here
we investigate effects of geometrical frustration on itinerant electron systems based upon the pyrochlore hubbard model .
one of the purpose of this paper is to investigate the mott transition in the absence of magnetic long - range order . according to the study of the heisenberg model on the pyrochlore lattice ,
it has been pointed out that the ground state of the insulating state is quantum disordered spin liquid.@xcite the critical character of the mott transition without magnetic order is quite different from that accompanying antiferromagnetic one .
moreover the character of the quantum spin liquid has not yet been unveiled sufficiently .
it is expected that the study from electron systems will shed new light on this subject .
another purpose of this paper is to reveal how the geometrical frustration affects strong electron correlation effects in the metallic state .
some recent experiments on the pyrochlore itinerant electron systems have reported that remarkable heavy - fermion - like behaviors manifest in these systems implying the crucial role of the geometrical frustration.@xcite here we show that the heavy - fermion state is realized in the pyrochlore hubbard model in the vicinity of the half - filling .
we also show that the dynamical spin susceptibility obtained from our model away from the half - filling is consistent with recent neutron scattering experiments for itinerant pyrochlore systems.@xcite the organization of this paper is as follows . the model and the basic method
are given in sec .
ii . in sec .
iii , we discuss about the mott transition and the realization of the quantum disordered spin liquid in the half - filling case . in sec .
iv , we consider the hole - doped case focusing on the heavy fermion state and its magnetic properties .
= 5.5 cm fig .
1 . pyrochlore lattice .
the hamiltonian of the model is given by , @xmath0 2 here @xmath1 is an annihilation ( creation ) operator of electrons with spin @xmath2 on the @xmath3-th site of the @xmath4-th tetrahedron . @xmath5 .
the kinetic term is diagonalized as , @xmath6 , where @xmath7 , @xmath8 , @xmath9 with @xmath10 .
the basis in the diagonalized space is obtained from the canonical transformation , @xmath11 . using the abbreviation , @xmath12 etc .
, we write down @xmath13(k ) , @xmath14 where @xmath15 , and , @xmath16 for @xmath17 , where @xmath18^{\frac{1}{2}}$ ] , and , @xmath19 for @xmath20 .
the expressions of @xmath21 and @xmath22 are very complicated .
however in the following , we need only @xmath21 for small @xmath23 , which is given by , @xmath24 at the half - filling , in the absence of electron - electron interaction , @xmath25 and @xmath26 are filled completely , and the two degenerate flat bands are empty .
@xmath25 touches with the two flat bands at the @xmath27 point of the brillouin zone .
this state is a gapless band insulator .
we now consider the effect of electron interaction on this state .
since , in the case of the half - filling or in the vicinity of the half - filling , the band @xmath26 is sufficiently far from the fermi level , it does not affect low energy properties and is negligible in the following argument .
then , the single - particle green s functions of @xmath28-electron are determined by the following equation . @xmath29
g^{(\lambda\nu)}_k(\varepsilon)=\delta_{\mu\nu}.\end{aligned}\ ] ] we put the chemical potential @xmath30 . in general , the off - diagonal self - energy @xmath31(@xmath32 ) is not negligible . however , in the vicinity of the @xmath27 point , at which the most important scattering processes occur , the off - diagonal terms vanish because of the momentum dependence of @xmath33 .
thus in the following argument , we neglect the off - diagonal self - energy .
to investigate the electronic state at the half - filling , we calculate the self - energy .
recently , isoda and mori obtained the diagonal self - energy of this model up to the second order in @xmath34,@xcite which is given by , @xmath35 for @xmath36 .
however , as will be shown here , the higher order corrections to the self - energy give rise more singular contributions and a drastic change of the electronic state , namely , the mott transition . fig .
2 . the most singular diagrams of the self - energy up to the forth - order in @xmath34 . @xmath37 denotes @xmath38-electron .
@xmath39 denotes @xmath40 or @xmath41-electron .
the dotted line denotes the interaction @xmath34 . at the half - filling
, the @xmath25-band is completely filled , while the flat bands are empty .
hence , at zero temperature , particle - particle channel between @xmath38-electron and @xmath40 or @xmath41-electron vanishes . on the other hand ,
particle - hole channel gives non - vanishing contributions only for the pair of @xmath38-electron and @xmath40 or @xmath41-electron .
then the most singular diagrams up to the forth order in @xmath34 are those shown in fig.2 .
since the fermi level is at the position between the two flat bands and @xmath25-band , we shift the chemical potential infinitesimally , @xmath42 . after integrating over energy and momentum , we take the limit @xmath43 .
this procedure is required in this singular perturbative calculation to fix the density of electrons @xmath44 .
neglecting the momentum dependence of @xmath33 does not affect the leading singular behaviors .
then , we can carry out the calculation analytically.@xcite for @xmath36 , we obtain , @xmath45 where @xmath46 is a positive constant of order unity .
the real part of the self - energy for hole - like excitations is divergent .
the same divergent behavior appears in the imaginary part , however , for particle - like excitations .
higher order corrections also show stronger divergence .
this divergent behavior means that the unperturbed state is unstable in the presence of electron correlation and that the single particle energy gap is generated . to see this we take only up to the third order term of the self - energy .
hole - like excitations are possible only for @xmath25-band in the half - filling case at zero temperature .
thus the single particle energy of this band measured from the fermi level is changed to @xmath47 for @xmath48 .
@xmath25 band is pushed down to the lower energy , and the energy gap opens between the flat bands and @xmath25-band .
this gap generation due to electron correlation signifies that the systems is in the mott insulating state .
although the exact magnitude of the gap should be determined by taking into account higher order corrections in @xmath34 , the gapful state is the self - consistent solution for non - vanishing @xmath34 .
thus an infinitesimally small @xmath34 drives the system into the mott insulator at the half - filling . in the ground state
, the two bands below the gap , @xmath25 , @xmath26 , are filled with up and down spins .
hence the system is in the spin singlet state .
it should be stressed that , after the mass gap generation , the singularities of the perturbative expansion are eliminated , and as a result , the gapful state is stabilized even if one take into account higher order corrections , which just renormalize the magnitude of the gap . to see the magnetic property of the mott insulating state obtained in the previous section , we compute spin - spin correlation functions , @xmath49 , where @xmath50 .
we carry out perturbative expansion in terms of @xmath34 using the green s function of the gapful state , @xmath51^{-1}$ ] , as the unperturbed propagator . calculating the diagrams up to the third order in @xmath34
, we have , @xmath52 with @xmath53 a step function .
the spin gap exists between the spin singlet ground state and the triplet state .
thus the system is in the quantum disordered spin liquid state without magnetic long - range order .
the result may not be changed qualitatively by higher order corrections .
these results are consistent with the previous studies for the pyrochlore heisenberg model.@xcite the singularity of the self - energy ( [ reself ] ) is deeply related with the accidental degeneracy at the @xmath27 point in the momentum space , which may cause an instability of lattice structure
. this property may be relevant to the metal - insulator transition of @xmath54 , which accompanies a structure change.@xcite
in the hole - doped case , the singularities appeared in the half - filling case are eliminated by the presence of the cutoff which is the chemical potential @xmath55 measured from the half - filling level .
then the fermi liquid metallic state is realized .
however , in the vicinity of the half - filling , the effective mass is much enhanced by electron correlation . up to the third order in @xmath34 , the leading term of the mass enhancement factor
@xmath56 is , @xmath57 as the electron filling approaches the half - filling value @xmath58 , the mass enhancement factor shows the divergent behavior indicating the precursor of the mott transition .
such a large mass enhancement is actually observed in the specific heat measurement for some pyrochlore itinerant systems like @xmath59.@xcite = 5.5 cm fig .
3.the most singular diagrams of spin - spin correlation functions up to the third - order in @xmath34 .
= 6.5 cm fig .
a ) the structure factor @xmath60 plotted in the [ 001]-[110 ] plane . the black line denotes the zone boundary .
the brighter regions have the stronger intensity .
( b ) the neutron cross section .
the largely enhanced effective mass obtained in the previous section is a remarkable property of geometrically frustrated electron systems in which several modes of spin fluctuations compete with each other , and magnetic long - range order is suppressed .
the spin fluctuation is almost localized in real space as in the case of @xmath61-electron based heavy fermion systems .
this property is seen more clearly in the momentum dependence of the spin - spin correlation function .
we calculate it by the perturbative calculation .
since the magnetic frustration is so strong , any approximations which neglect the coupling between several modes of spin fluctuation like random phase approximation(rpa ) or fluctuation exchange approximation(flex ) are invalid for this system .
actually the most singular contributions up to the order of @xmath62 comes from the diagrams shown in fig.3 . to simplify the calculation we exploit the following approximation .
near the half - filling , the momenta running in @xmath63-line take small values in the vicinity of the fermi surface .
moreover for small @xmath23 , @xmath64 does not depend on @xmath65 .
thus we can separate the integral over @xmath65 and the angle for the momenta running in @xmath63-line . then , the spin - spin correlation function is decomposed into the momentum part and the energy part , @xmath66 .
such a factorization of the spin - spin correlation function is indeed observed in an inelastic neutron scattering experiment for @xmath59.@xcite the energy part is given by , @xmath67 for small @xmath68 . here
@xmath40 , @xmath41 are constants . thus in the vicinity of the half - filling , spin fluctuation is much enhanced .
the presence of the giant spin fluctuation is also consistent with several experiments for itinerant pyrochlore systems.@xcite it is noted that such a large enhancement of @xmath69 is not obtained by rpa or flex approximations for the appropriate value of @xmath34 .
we calculate the structure factor @xmath70 numerically . in fig.4(a ) , we plot the intensity of @xmath60 in the [ 110]-[001 ] plane of the reciprocal lattice .
@xmath60 is almost constant in the [ 001]-direction , indicating that the spin fluctuation is strongly localized in this direction .
it also has a small peak on the @xmath71 line showing the presence of small fluctuation toward a collinear magnetic order , as was pointed out in the study of the heisenberg model.@xcite however this mode competes with the other modes which exist in @xmath72 , @xmath73 , @xmath74 , @xmath75 , and thus such a magnetic order is suppressed . in fig.4(b ) , we show the calculated result of the neutron cross section , @xmath76 , in the [ 110]-[001 ] plane . here
@xmath77 is a reciprocal lattice vector .
the positions of the maximum are almost consistent with the neutron scattering experiment for @xmath59.@xcite note that this structure is very similar to that found for the heisenberg model.@xcite thus the itineracy of electron affect little the structure in the momentum space , though the dynamical properties between the mott insulating state and the metallic state are quite different ; namely , the former is gapful and the latter is gapless .
the momentum dependence of @xmath69 implies that the spin fluctuation is localized in a tetrahedron forming collective 4-spin singlet.@xcite such a localized character of spin fluctuation is very similar to magnetic properties of @xmath61-electron based heavy fermion systems .
however , in the latter systems , magnetic long - range order is suppressed by the presence of kondo temperature higher than magnetic ordering temperature . in our system ,
the geometrical frustration is crucial .
here we discuss the charge response of the heavy - fermion state near the half - filling . in general ,
the charge susceptibility @xmath78 in the vicinity of the mott metal - insulator transition point shows quite different behaviors depending on the density of states near the fermi level.@xcite in our systems , the bare density of states in the hole - doped case is non - singular , though four - point vertices are much enhanced by electron correlation .
the perturbative calculation of the charge susceptibility gives singular terms in the limit of @xmath79 , namely , @xmath80 .
thus it is highly non - trivial how the charge susceptibility behaves , as the electron filling approaches the half - filling value . to see this
we resum the most singular terms of the perturbative expansion in @xmath34 .
as was done before , we neglect the momentum dependence of @xmath81 .
we first consider the terms which contain only one irreducible four - point vertex .
the most singular terms of this type are diagrammatically expressed as fig.5(a ) these terms are calculated as , @xmath82 here @xmath83,@xmath84 , and @xmath85 are positive constants . on the other hand , the term which contains @xmath86 irreducible four - point vertices shown in fig.5(b ) gives the contribution @xmath87^n/\vert\mu\vert^{\frac{n-1}{2}}$ ] with @xmath46 a positive constant .
summing up all most singular terms , we have , @xmath88 thus as the electron filling approaches the half - filling value @xmath79 , the charge susceptibility decreases toward zero , @xmath89 , indicating that the system becomes incompressible .
diagrams of the charge susceptibility .
@xmath90 is the irreducible 4-point vertex .
in this paper , we investigate the mott transition , strong correlation effects and spin dynamics of the pyrochlore hubbard model .
we show that , at the half - filling , electron - electron interaction leads the system into the mott insulator in which the quantum disordered spin liquid state realizes . in the hole - doped case ,
we show that the effective mass is anomalously enhanced in the vicinity of the half - filling signifying the heavy fermion state .
we have also calculated the neutron cross sectin of this state which is consistent with the recent experiment for some itinerant pyrochlore systems like @xmath59 .
although we carried out the perturbative calculation up to the fourth order in @xmath34 , the singular divergent behavior of the self - energy shown in sec .
iii implies that we need to resum all order singular diagrams . since the most singular diagrams have the specific structure as discussed in sec .
iii , we can carry out this program in principle .
we would like to address this issue in the near future .
the author would like to thank k. yamada , k. ueda , n. kawakami , h. tsunetsugu , h. ikeda for invaluable discussions .
this work was partly supported by a grant - in - aid from the ministry of education , science , and culture , japan .
s. kondo , d. c. johnston , c. a. swenson , f. borsa , a. v. mahajan , l. l. miller , t. gu , a. i. goldman , m. b. maple , d. a. gajewski , e. j. freeman , n. r. dilley , r. p. dickey , j. merrin , k. kojima , g. m. luke , y. j. uemura , o. chmaissen , and j. d. jorgensen , phys .
lett . * 78 * , 3729 ( 1997 ) . | we investigate the interplay between geometrical frustration and strong electron correlation based upon the pyrochlore hubbard model . in the half - filling case , using the perturbative expansion in terms of electron correlation , we show that the self - energy shows a divergent behavior leading the system into the mott insulating state , in which quantum disordered spin liquid without magnetic long - range order realizes . in the hole - doped case ,
we obtain heavy - fermion - like fermi liquid state .
we also calculate the neutron cross section which is well consistent with recent neutron scattering experiments for itinerant pyrochlore systems .
= 10000 2 |
the investigation of observables having singular or highly oscillating behaviors at microscopic scales , which become smooth after appropriate coarse graining is a typical occurrence in many research areas of modern physics . a detailed theoretical investigation and description is therefore highly desirable and important .
quasiclassical green s functions have been introduced in the transport theory of spatially nonuniform superconductors and superconducting hybrid structures as an envelop
approximation for exact green s functions quickly oscillating at fermi wavelength scales , see ref .
@xcite for a review .
one result of this coarse - gaining of the order parameter is the landau - khalatnikov theory of second order phase transitions @xcite .
envelop approximations of highly oscillating solutions of hydrodynamic equations gave an opportunity to achieve analytical and numerical advances in hydrodynamics @xcite . in the study of granular superconducting materials , josephson junction arrays ( jja ) and strongly disordered superconducting films near the superconductor insulator transition ,
the current - voltage characteristics and the dynamic conductivity show singular behaviors described by weighted superpositions of delta - functions @xcite .
the imaginary part of the two - particle green s function in ultrasmall metallic granules , where the electron spectrum is essentially discrete , have the form of a series of delta - functions .
the transport characteristics of ultrasmall tunnel junctions exchanging energy with a quantum bath with discrete spectrum have a similar singular behavior @xcite .
an important and in general unsolved problem is the construction of the envelop approximation for an observable ( which is typically measured ) in case the theoretically derived expression is a countable superposition of delta - functions .
remarkably , this problem has a general analytical solution going beyond standard approximations that allows transforming the discrete series into integrals .
we found the envelop approximation for singular observables analytically , using the jacoby theta - function formalisms @xcite .
our calculations can be physically interpreted using the language of the landau - hopf turbulence @xcite , at least qualitatively . in this paper
we derive the envelop approximation for the real - valued function @xmath1 with real argument @xmath2 : @xmath3 here , @xmath4 , is a @xmath5-dimensional vector of quantum numbers , @xmath6 is a scalar real function of @xmath4 and the summation is performed over all configurations of @xmath4 with weights ( probabilities ) @xmath7 . in physical applications
@xmath2 is usually a variable with dimension of energy , @xmath6 is closely related to ( the difference of ) the energy levels of a quantum system , while @xmath8 is the gibbs distribution probability . in the most general situation
when all components of @xmath4 belong to a countable set , standard approximations like the euler - maclaurin asymptotic @xcite do not work anymore . in that case
one can not naively integrate out the @xmath0-functions in . our main result deals with this case .
we find the conditions when the density of the @xmath0-functions starts to increase similar to the development of landau - hopf chaos and find the envelope approximation for @xmath1 in that case .
we discuss the use of our results for solutions of specific physical problems mentioned above .
the case with linear @xmath6 and @xmath8 with bilinear exponent is one of the most relevant for applications .
we focus on the situation when @xmath9 here @xmath10 is a vector of integer numbers , @xmath11 is a ( fixed ) vector and @xmath12 is a symmetric positively defined @xmath13 matrix .
we will also show that our investigation is not crucially dependent on the specific form of @xmath8 in .
the case where @xmath8 decays quickly with @xmath14 ( all eigenvalues @xmath15 of the matrix @xmath12 are much larger than unity ) is the `` quantum '' limit .
then the set of numbers @xmath4 is essentially discrete .
moreover @xmath8 effectively restricts @xmath4 to a subset with small @xmath14 . in this case
the graph of @xmath1 looks like a sparse sequence of isolated @xmath0-function peaks .
rather straight - forward is the opposite `` classical '' case where @xmath8 depends only slightly on @xmath4 ( all eigenvalues @xmath15 of the matrix @xmath12 are much smaller than unity and @xmath5 is sufficiently large ) .
then the euler - maclaurin approximation is applicable and one can replace the discrete sum over @xmath4 in eq . by an integration over @xmath16 .
the result is @xmath17 , where @xmath18 is constant .
the most interesting case is the `` mesoscopic '' case when the eigenvalues @xmath15 of the matrix @xmath12 are slightly smaller or of the order of unity and @xmath5 is not too large .
this case is in between the quantum and the classical case .
then the restriction for the choice of @xmath4 is rather weak and there are a many @xmath4-vectors solving the inequality @xmath19 where @xmath20 is the width of the @xmath0-function .
[ @xmath0-functions in physical applications have always some small width , @xmath21 , due to , e.g. , interaction with a heat bath ( dissipation ) . ]
if the components of the vector @xmath22 are integer numbers ( commensurate ) then only if @xmath2 is close to an integer [ the fractional part of @xmath2 is smaller than @xmath21 ] there is vector @xmath4 the solution of inequality .
however , in case the components of vector @xmath22 are not commensurate as it is most natural in physical systems the formal solution of eq
. for arbitrary @xmath2 is only applicable in the classical regime where no ( integer ) restriction of @xmath14 exists . in the mesoscopic regime , @xmath24 , the set of @xmath2 for which eq .
has a solution is restricted .
but the effective measure of this set is much larger then in the case when @xmath22 has commensurable components .
this implies that the effective density of @xmath0-functions in eq
. strongly increases in a given @xmath2 interval , but is still discrete .
we will further refer to this observation as the @xmath0-function `` condensation '' . as the result
the singular part of @xmath1 becomes relatively small and @xmath1 has a smooth envelop approximation , @xmath25 .
we emphasize that the functional behavior of @xmath1 is different from the classical limit due to the nontrivial factor @xmath26 .
we find that @xmath27 is the one - dimensional jacoby theta - function , @xmath28 , where the real parameter @xmath29 depends on @xmath12 and @xmath11 and qualitatively shows to what degree the components of @xmath22 are incommensurate .
we explicitly calculate @xmath30 , which behaves in the classical limit as @xmath31 resulting in @xmath32 . in order to investigate @xmath33 in the regime where the @xmath0-functions condensate
, it is quite ineffective to use the @xmath0-function representation of @xmath33 , eq . , directly .
instead , we rewrite @xmath33 in terms of matrix @xmath34-function , @xmath35 [ generalizing the approach suggested in ref . @xcite ] , where the real symmetric matrix @xmath36 has one zero eigenvalue , while the other eigenvalues are positive .
this zero mode is the manifestation of @xmath0-functions in eq . .
we investigate the class of the @xmath34-functions with the ( nearly)degenerate @xmath36 ; it is an interesting problem itself .
using the matrix @xmath34-function representation of @xmath33 we can analytically and numerically investigate @xmath1 in the regime where the @xmath0-function condensation takes place . in a limiting case
we reproduce the results of ref .
we relate the nature of the strong increase of the @xmath0-function density to the chaotic behavior of quasi - periodic functions .
our results can help understanding the transport theory of josephson junction arrays and the superconductor - insulator transition .
finally , we discuss how stable our results related to the @xmath0-function condensation in eq .
are with respect to the choice of the shape of the weight functions @xmath8 other than given in eq . .
the structure of our paper is the following : in sec .
[ sec : theta ] we show how problem can be reformulated in terms of matrix theta - functions ; in sec . [ sec2 ] we investigate the properties of the matrix theta - function in the mesoscopic regime and in particular formulate the conditions for the delta - function condensation ; in appendix [ sec3 ] we give a numerical receipt for an efficient calculation of the matrix theta - function ; finally in the discussion section [ disc ] , we show that the problem we address in this paper has a number of important physical applications .
we rewrite the sum in eqs . - using the poisson s formula for summation @xmath37 where @xmath38 is a continuous integrable function .
function @xmath33 in eq . depends on the vector @xmath39 .
therefore , we have to introduce also @xmath40 and @xmath41 accordingly . approximating the @xmath42-function by a smeared gaussian function @xmath43 with @xmath44 we get after the integration over @xmath41 : @xmath45+\left[\frac{\omega}{\sigma^2}\right]^2e_ie_j\right)\frac 12\langle x_i x_j\rangle\right\}.\end{gathered}\ ] ] here
@xmath46_{ij}.\end{aligned}\ ] ] we simplify eq . by expanding it over @xmath21
. then we find , @xmath47 where the explicit analytical expressions for @xmath48 , @xmath49 $ ] and @xmath50 can be found in appendix , eqs . - .
introducing the riemann theta - function @xcite , @xmath51 we rewrite @xmath33 as follows : @xmath52 the proof that @xmath53 is given in appendix , eq . .
therefore , the @xmath34-function matrix parameter is degenerate resulting in the appearance of the @xmath0-functions in eq . .
it is worth noting that @xmath54 is generally not an eigenvector of @xmath36 ( such that @xmath55 ) .
an important role play the eigenvalues @xmath56 of matrix @xmath36 .
we will distinguish the cases when @xmath56 are smaller or larger than unity below .
without loss of the generality we can assume that the eigenvalue corresponding to the eigenvector @xmath22 has the index one : @xmath57 .
then for the other eigenvalues follows : @xmath58 .
we kept @xmath59 above which resulted in @xmath57 . in typical physical realizations ,
the @xmath0-function always have a finite width @xmath21 as was mentioned in the introduction .
if we preserve the sub - leading terms in @xmath60 in @xmath33 then @xmath61 .
perturbations of the other parameters in eq .
produce sub - leading corrections to the shape of @xmath1 .
.,title="fig : " ] + there are special cases when @xmath62 and the parameters @xmath18 , @xmath63 , and @xmath36 have very simple forms .
one case corresponds to a diagonal matrix @xmath12 as follows from the proof given in appendix [ apb ] .
the case , when @xmath12 has the additional structure : @xmath64 , where @xmath65 is a parameter and @xmath66 , is important for transport physics in jjas and dirty superconducting films near the superconductor - insulator transition @xcite .
we will refer to this case as to the `` minimal model '' .
it should be noted that @xmath22 in this case is not an eigenvector of @xmath12 in general .
this case is of theoretical interest since it is possible to find @xmath67 exactly and @xmath68 analytically in all orders of @xmath69 : @xmath70_{ij}\beta=-\frac{1}{\beta\gamma+\sum_{k}e_k}+\frac{\delta_{ij}}{e_i } , \\\label{eq : detg } \det g=\beta^n\frac{(\beta\gamma+\sum_i e_i)\prod_k e_k}{\beta\gamma}.\end{gathered}\ ] ] then we get provided @xmath71 : @xmath72 similarly we find using eq . that @xmath73 expressions eqs .
- generalize similar results in ref .
@xcite obtained for the limit where @xmath74 , @xmath75 , @xmath76 , and @xmath22 is close to the vector @xmath77 .
we took @xmath59 in eq .
such that @xmath57 .
if we take into account @xmath20 then we should correct all matrix elements , @xmath78 , by an additional term , @xmath79 $ ] corresponding to @xmath80 .
below we study how the properties of the function @xmath34 in eq .
depend on the parameters @xmath22 , @xmath12 , and especially @xmath5 .
we start our analysis with the simplest case , @xmath81 .
then we get from eq . that @xmath34 reduces to the usual jacobi theta - function :
@xmath82 for @xmath83 , @xmath84 and for @xmath85 , @xmath86 , see fig .
[ fig_theta3d ] for an illustration .
the case @xmath81 is marginal to some extent because then @xmath87 and so @xmath27 is always a set of @xmath0-functions ; formally there is no classical limit for @xmath81 . in the following we focus on the case
@xmath88 only . for @xmath89 . without loss of the generality
we assume that @xmath22 ( direction singled out by the green line ) is close to @xmath90 .
the dotted lines are parallel to @xmath91 .
the red points belong to the subset of @xmath92 that gives the leading contribution to @xmath34 in the mesoscopic regime , eq .
.,title="fig : " ] + the classical limit as follows from eq . can be found by setting @xmath93 in eqs . - . then @xmath94 and as follows from eq .
@xmath95 we recall that in the mesoscopic regime @xmath96 ( @xmath97 ) . in the mesoscopic ( quantum ) regime @xmath1 strongly differs from eq . because the prefactor , @xmath98 , in eq
. behaves in a non - trivial way .
the definition of the matrix @xmath34-function , eq . , includes the sum over @xmath92 .
the components of @xmath92 are integer numbers . so @xmath92 can be treated as edge vectors of the nodes of an effective cubic crystal in a @xmath5-dimensional space , see fig .
[ condition ] .
it follows from the definition of the @xmath34-function , eq .
, that the leading contribution to @xmath34 gives @xmath92 nearly collinear to @xmath22 .
we focus first on the case when @xmath22 is close to the crystallographic axis @xmath91 of the effective crystal
. then the subset of @xmath99 , @xmath100 , where @xmath101 , gives the leading contribution to @xmath34 : @xmath102 and @xmath103 where @xmath104 .
the vectors @xmath22 and @xmath91 should be close such that the following condition is fulfilled : @xmath105 which ensures a complete delta - function overlap ( we recall that the @xmath0-function width is @xmath106 , while @xmath107 ) . here
the integer @xmath108 $ ] , where @xmath109 is the integer part .
this condition indicates how far from the origin the distance between the points on the lines o@xmath22 and o@xmath91 becomes of the order of the effective crystal period , see fig .
[ condition ] . for @xmath110 [ black color ] .
the minimal model has been used . taking the vector from the set of the gaussian random numbers with zero average and the dispersion @xmath111 we generated @xmath112 , where @xmath113 .
the red line corresponds to @xmath28 the envelop function .
the inset shows that the @xmath0-function width @xmath21 is the minimal characteristic frequency of the matrix @xmath34-function.,title="fig : " ] + expression is in fact the envelop approximation for the matrix theta - function .
the @xmath34-function quickly oscillates as the function of @xmath2 on the scale of the order of @xmath21 on top of the envelop function , see fig .
[ fig_theta ] .
the amplitude of the oscillations does not exceed @xmath114 .
it is worth noting that condition is sufficient but not necessary . even beyond the limitations set by condition the envelop function eq .
usually still approximates @xmath34 quite well .
this case is illustrated in fig .
[ fig_theta ] .
numerical calculations @xcite show that the envelope of the @xmath34-function follows well eq . while @xmath115 , where @xmath116 is a constant of order unity .
this exponential factor is closely related to the density of the @xmath0-functions in eq .
, which have in the mesoscopic regime weights of the same order .
if @xmath117 then @xmath118 reduces to a superposition of @xmath0-functions , see eq . and fig .
[ fig_theta3d ] .
these small values of @xmath30 appear when @xmath119 , see eq . .
in case the components of @xmath22 are integer numbers ( they are commensurable ) then the @xmath0-function singularities of @xmath1 remain in the mesoscopic regime as well . however ,
if @xmath22 and @xmath91 are even slightly different , @xmath30 can easily become of order unity and @xmath27 becomes a smooth function of @xmath2 . returning to eq .
we will have a strong increase of the @xmath0-function density and their final overlap , when the components of @xmath22 become not commensurable . to conclude this section , we emphasize that eq .
is one of our main results .
it shows that in the mesoscopic regime @xmath1 is nearly continuous ( plus a relatively small and quickly oscillating background ) and the form of the function @xmath1 is nontrivial .
we find the envelop approximation of @xmath1 analytically : @xmath120 the accuracy of the prefactor in eq .
is illustrated in fig .
[ fig_theta ] .
the density graph [ fig3d ] shows the evolution of the matrix @xmath34-function in the minimal model when the parameter @xmath65 switches the model from the quantum to the mesoscopic regime .
the numerical calculations are briefly described in appendix [ sec3 ] . for @xmath110 .
the axes @xmath121 corresponds to @xmath2 while @xmath122 axes corresponds to @xmath123 .
when @xmath124 the theta function behaves as the discrete set of @xmath0-peaks . with decreasing @xmath65 , more peaks become visible , for @xmath125 the peaks start merging and for @xmath126 the matrix theta function becomes nearly smooth.,title="fig : " ] +
the time representation helps to understand the properties of @xmath1 from another point of view . for a special case it was shown in @xcite that the landau - hopf turbulence scenario can account for the merging of the @xmath0-functions in eq . in the mesoscopic regime . here
we apply the ideas developed in ref .
@xcite for the general case . to this end
we express @xmath1 , eq . , in time representation , @xmath127 , and find : @xmath128 function @xmath129 belongs to the set of _ quasi - periodic _
functions @xcite .
if we take @xmath130 as coordinates in an @xmath5-dimensional space then the trajectory @xmath131 is modeled by a curve on a torus @xmath132 that wraps around without ever exactly coming back on itself if the @xmath133 are _ incommensurate _ , see fig .
[ fig3 ] for @xmath89 .
the path covers the torus surface densely everywhere . if we return to eq .
then this property would mean that the @xmath0-function positions are densely distributed .
in the quantum regime only @xmath4 with @xmath134 contribute to @xmath129 such that not more than one frequency appears in the exponent in eq . .
the classical limit formally corresponds to a torus with an infinite number of dimensions . our consideration for the condensation of the @xmath0-functions in eq .
can be extended to the more general case of arbitrary weights @xmath135 which decay quickly with @xmath4 @xcite . in this case
, the time representation of @xmath33 would consist of quasi - periodic functions as well and the topological argument of the path covering a torus surface densely would be applicable again . in this more general case @xmath1 can not be reduced to the matrix theta - function .
however , the poisson transformation of @xmath33 with an appropriate @xmath21-expansion still allows to reduce @xmath33 to an `` easy to handle '' form for analytical and numerical investigation in the regime when the @xmath0-functions start to merge .
it is worth noting that expressions similar to eq .
appear in many applications .
for example , it describes the velocity field @xmath136 of a liquid when turbulence develops according to the landau - hopf scenario @xcite , @xmath137 . here
the developed turbulence corresponds to the case of large @xmath5 @xcite .
quasi - periodic functions describe quasi - periodic motion of mechanical systems @xcite and often appear in the theory of differential equations , see e.g. , ref . @xcite .
+ in the study of granular superconducting materials and josephson junction arrays ( jja ) transport properties are of central importance .
the most interesting case corresponds to the quantum limit when the josephson coupling @xmath138 between the granules is much smaller than the characteristic coulomb energy @xmath139 of the cooper pair located on a granule .
the transport problem was solved within linear response approaches , see e.g. , @xcite . it was shown that the conductivity @xmath140 of the granular superconductor is a singular function in the leading order over @xmath138 .
it behaves as the countable superposition of @xmath0-functions with the gibbs weights .
for example , the conductivity in a jja according to ref .
@xcite has the following form [ which can also be derived from eq . in appropriate limiting cases ] : @xmath141 where @xmath2 is the frequency , @xmath132 is the temperature , the indices @xmath142 , @xmath143 label the islands of the jja , @xmath144 is the inverse capacitance matrix of the jja and @xmath145 . here
@xmath146 is the charge of the cooper pair and @xmath147 is the electron charge [ we will use the system of units where @xmath148 . the integer numbers @xmath149 show the effective number of cooper pairs sitting on the island @xmath142 .
we assume here and below the einstein sum convention .
the exponential weights in eq .
come from the gibbs distribution .
the delta - functions ensure energy conservation during the processes of cooper pair tunneling from one granule of the jja to its neighbor .
they can not be easily integrated out in eq . since changes of the coulomb energy are essentially discrete because of charge quantization within granules .
the size of the jja arrays can be quite small , especially in the case of one - dimensional ( 1d ) arrays .
therefore , the number of junctions @xmath5 in the jja is typically finite , @xmath150 . on the other hand
the interaction matrix @xmath151 can be rather quickly decaying with @xmath152 .
generally , we can not use standard statistical physics approaches based on the thermodynamic limit @xmath153 trying to smear out the @xmath0-function singularities in the observables , as it was done in eq . .
an important question is to understand the nature of the @xmath0-function singularities in the observables and finding systematic ways of their regularization .
one way to overcome the difficulties related to the @xmath0-functions in transport observables of the jja was proposed in ref .
that calculation was based on the assumption of an energy band for cooper - pair tunneling .
however , this band can form in jja with nearly identical granules but should be suppressed by the disorder in typical disordered jjas @xcite . in refs .
@xcite an attempt was made to find the transport characteristics in the disordered jjas .
the model of refs .
@xcite leads to the conductivity behaving according to eq .
with @xmath154 and @xmath22 close to the vector @xmath77 ( a limiting case of our minimal model ) .
it was shown in ref .
@xcite that although @xmath22 has non - commensurate components , the delta - functions merge and the conductivity may become a smooth function of its parameters . in this paper
we do not restrict the choice of @xmath12 and @xmath22 like in ref .
@xcite .
the properties of the matrix-@xmath34-function strongly depend on the number @xmath155 of the nonzero components of @xmath54 .
this number can be treated as an effective dimension of the @xmath34-function .
generally @xmath156 .
without loss of the generality we can assume that the first @xmath155 components of @xmath54 are nonzero while the others take zero values .
we can apply a reduction procedure if @xmath157 : to take the sum over @xmath158 with @xmath159 in eq . .
we finally obtain a function very similar to the matrix - theta function but now with a summation over @xmath155-dimensional integer vectors .
for example , @xmath160 for the problem considered in ref .
@xcite and @xmath161 for ref .
@xcite .
the problem of the evolution of a quantum system interacting with a ( quantum ) environment of ( soft ) modes is being studied for more than 60 years , but more important to be solve than ever , see e.g. , refs .
usually it is implied that the environment has a continues spectrum of modes .
that condition is important since it typically avoids the the appearance of the @xmath0-functions like in eq . in observables of the quantum system .
a discrete environment , on the other hand , can not simply absorb arbitrary amounts of energy , but rather only discrete energy portions in quantums on the order of its level spacing .
in other words , the set of quantum modes can serve as a `` bath '' only when its spectrum is continuous @xcite , which we will clarify below . .
the wavy lines show schematically the exchange of the energy between the junction and the environment @xcite .
b ) the environment is not necessary described by the equilibrium density matrix @xcite .
here we sketched the environment interacting with a thermal bath and the quantum system .
, title="fig : " ] + an important problem where a quantum system interacts with an environment , is the problem of quantum transport through an ultrasmall tunnel junction , see fig .
[ figeh ] , and ref .
@xcite for a review .
if the contacts of the junction are superconducting , the supercurrent is given by @xmath162 where @xmath163 is the probability to exchange energy @xmath2 with the environment and @xmath138 is the josephson energy of the junction .
next , we concentrate on the shape of @xmath164 . according to ref .
@xcite @xmath165\exp[-ie^*\hat\phi(0)]\rangle e^{i\omega t},\ ] ] where @xmath166 is the `` charge transfer '' operator , @xmath167 for cooper pairs and @xmath168 denotes averaging over the density matrix @xmath169 of the environment .
[ we recall that we have chosen units where @xmath170 .
since the environment is isolated , it can be described by the hamiltonian @xmath171 with a set of energies @xmath172 and eigenfunctions @xmath173 .
in general we have @xmath174 . writing the phase operator in heisenberg representation
explicitly , we get @xmath175=\hat u^{-1}(t ) \exp[ie^*\hat\phi(0)]u(t)$ ] , where @xmath176 .
this way we find : @xmath177 thus @xmath163 is represented as the superposition of @xmath0-functions .
only if the spectrum of the environment is continuous we can introduce a continuous density of states and integrate out the @xmath0-functions in eq . .
as long as the spectrum of the environment is discrete , we can assume without loss of generality that the greek indices labeling the levels are a set of integer numbers .
if the environment is represented by oscillator modes ( or quantum rotators ) , @xmath178 is linear in the indices labeling the environment states like in eq . .
if , in addition , the environment is in local equilibrium , such that @xmath179 , where @xmath180 is the environment temperature , the structure of eq .
is the same as the structure of eq . .
to complete our investigation we focus on the quasi - particle current @xmath181 through the tunnel junction shown in fig .
[ figeh ] . according to refs .
@xcite we have @xmath182 where @xmath183 ( @xmath184 ) is the tunneling rate from the left ( right ) to the right ( left ) , and , for a single junction , @xmath185 where @xmath186 are the electronic distribution functions within the left ( right ) electrodes , and @xmath187 is the bare tunnel resistance , representing the interaction of electrons with the bath . here
@xmath163 is the probability for electron quasi - particle to lose the energy @xmath2 to the environment ; it is given by eq . with @xmath188 .
the backward scattering rate is given by @xmath189 .
[ if the contacts are superconducting , we can account for that by introducing the quasi - particle densities of the states in the contacts @xcite in eq .
. would then give the quasi - particle current .
] if the environment is absent and the relaxation is provided by a phonon bath , @xmath190 and eq .
reproduces the conventional ohm law .
it follows from eqs . -
that the integration over the energy removes the @xmath0-functions of the environment and the quasiparticle current is not as singular as the supercurrent when the environment has a discrete spectrum of modes . to summarize , the problem we solve in this paper is closely related to the problem of a discrete environment exchanging energy with a quantum conductor .
it should be emphasized that the environment should not be necessary located somewhere outside the quantum conductor . on the contrary , it can be part of the conductor itself .
such a situation is realized in jjas , see e.g. , ref .
@xcite , or in very dirty conductors where the transition between discrete and continuous `` built - in '' environments is closely related to many - body localization , refs .
we have shown that observables having a form as described by eq . can be rewritten in terms of the matrix theta - function .
we demonstrated that the @xmath0-functions typically condensate in the mesoscopic regime such that the observable @xmath33 we focus on becomes a nearly continuous function .
we found the envelop function for @xmath33 analytically and therefore , our results can help to understand and describe transport properties in a number of strongly correlated systems .
the authors thank a. petkovic and v. vinokur for active discussions in the initial stage of the work ; we also thank m. fistul for helpful comments and t. baturina for interest to our work .
this work was supported in part by the russian foundation for basic research ( grant no .
11 - 02 - 00341 ) and by the u.s .
department of energy office of science under the contract no .
de - ac02 - 06ch11357 .
the matrix @xmath12 can be diagonalized by the orthogonal transformation @xmath191 , such that @xmath192 , where @xmath193 is the diagonal matrix of the eigenvalues . then expanding the determinant of @xmath194 over @xmath69 we get the following relation : @xmath195 where @xmath196 .
similarly we can prove , using an induction procedure , that @xmath197 and @xmath198= \\
\frac{\det e}{g}=\left(\sum_{i } \frac{(\tilde e_{i})^2}{\lambda_{i}}\right)^{-1}. \ ] ] it should be noted that @xmath199 .
@xmath200 we will see that the vector @xmath54 is the characteristic direction of the @xmath27-function .
it is worth noting that generally @xmath22 and @xmath54 are not parallel in euclidean space .
the vector @xmath54 can be found explicitly : @xmath201 it follows from eq .
that @xmath202 finally , we find the inverse of the matrix @xmath194 at @xmath44 .
this is somewhat tricky since @xmath194 becomes singular while , following eq .
, @xmath67 becomes degenerate .
nevertheless there is a finite nontrivial limit for @xmath67 at @xmath44 : @xmath203 the diagonal elements have the structure : @xmath204 where @xmath205 .
we assume that @xmath22 is an eigenvector of matrix @xmath12 corresponding without loss of generality to eigenvalue @xmath206 .
then @xmath207 , and therefore @xmath208 where @xmath209 .
most important in this case is @xmath210 this identity can be proven in a diagonal representation of @xmath12 . without loss of the generality
@xmath211 corresponds to the eigenvalue @xmath206 of @xmath12 .
then we can take again @xmath212 and @xmath213 so , @xmath214 .
an important task is to verify the accuracy of eq . and the envelope approximation of the matrix @xmath34-function numerically
. the required calculations of the matrix @xmath34-function for @xmath88 requires care since one should sum up many quickly oscillating functions .
we have found an effective numerical algorithm explicitly relying on the existence of the soft mode @xmath22 of the matrix @xmath36 .
the matrix @xmath36 can be expressed explicitly through its eigenvalues @xmath215 and the corresponding eigenvectors @xmath22 and @xmath216 , @xmath217 [ @xmath218 @xmath219 the set of @xmath92 , see eq . , form a @xmath5-dimensional cubic crystal .
we construct a cylinder in this space with generating lines @xmath22 and an elliptical support with main directions @xmath216 . doing numerical calculations we take into account only points of the crystal , @xmath92 , in eq .
that belong to the cylinder volume , see fig . [ fig_tube ] .
the proportions of the cylinder depend on the calculation accuracy .
the ratio of the cylinder height to its characteristic diameter is of the order of @xmath220 .
the case @xmath97 is the most interesting since then the @xmath0-functions in eq .
are expected to condensate .
then only @xmath92 , the nearest neighbors to the line directed along @xmath22 , should be taken into account , see green diamonds in fig .
[ fig_tube ] and red dots in fig .
[ condition ] .
this property strongly reduces the numerical efforts compared to a direct calculation of @xmath221 using eq . . , in eq . ,
that belong to the `` cylinder '' volume and satisfy the condition @xmath222 . here
the solid line is directed along @xmath223 .
the green diamonds distinguish the points of the effective crystal that give the leading contribution to @xmath34 for @xmath224 ; the black points should be added if @xmath225 while the small blue points become important for @xmath226 . ,
title="fig : " ] + a typical graph of @xmath98 for @xmath110 is shown in fig .
[ fig_theta ] [ black line ] , the red line shows the envelop function represented according to eq . by the one dimensional ( jacoby ) @xmath34-function .
the inset shows that the @xmath0-function width @xmath21 is the minimal characteristic `` frequency '' of the matrix @xmath34-function .
it follows also that the matrix @xmath34-function can be approximated by the one - dimensional ( jacoby ) @xmath34-function with specially chosen parameters . for producing the graph [ fig_theta ] we have used @xmath36 from eq . .
this minimal model we have chosen for two reasons : 1 ) @xmath36 is parameterized by a minimal set of parameters , 2 ) this model has some relation to physical applications , see refs .
@xcite,@xcite .
the vector @xmath22 we expressed as @xmath227 , where we chose @xmath228 according to ref .
the components of @xmath229 are generated by a gaussian random number generator with zero average and the variance @xmath111 .
for simplicity we used the restriction , @xmath230 .
then we get @xmath231 $ ] that agrees with ref .
@xcite .
if @xmath119 , the components of @xmath22 are commensurate and @xmath232 behaves just as the superposition of @xmath0 functions shifted by a constant period of the order unity along the @xmath233-axis as follows from eq . .
but if @xmath234 then the components of @xmath22 are not commensurate , there is no periodicity in the @xmath0-function set , as it is illustrated in fig .
[ fig_theta ] for @xmath235 .
we implied nearly everywhere above that @xmath236 . if this is not the case and @xmath124 , @xmath237 in eq .
quickly decay with growing @xmath14 and we can disregard all @xmath4 with @xmath238 within acceptable error bars .
it is clear that in that case @xmath1 [ as well as @xmath27 ] behaves always as a set of @xmath0-functions with some finite distance from each other .
disorder in @xmath22 only slightly shifts the positions of the @xmath0-functions on the @xmath2-axis and there is no @xmath0-function condensation .
what happens when we go from @xmath236 to @xmath124 is illustrated in fig .
[ fig3d ] .
when @xmath124 the theta function behaves as a discrete set of the @xmath0-peaks . with decreasing @xmath65 more peaks
develop , for @xmath125 the peaks start merging and for @xmath126 the matrix theta - function becomes nearly smooth .
l. d. landau and i. m. khalatnikov , collected papers of l.d .
landau , d. ter haar , ed .
( new york : gordon and breach , 1965 ) ; l. d. landau and i. m. khalatnikov , dokladii academii nauk cccp 96 , 469 ( 1954 ) ; v.l .
ginzburg , and l. d. landau , zh .
fiz . * 20 * , 1064 ( 1950 ) .
the function @xmath240 is the quasi - periodic function with periods @xmath241 if @xmath242 where @xmath243 is periodic function of @xmath244 with the periods , @xmath241 , respectively .
all the periods @xmath241 are strictly positive , and their reciprocals @xmath245 are rationally linearly independent
. the quasiperiodic function can be represented by the series : @xmath246 .
d.v.averin and k.likharev , in _ quantum efects in small disordered systems _ , ed . by b.l.altshuler , p.a.lee , and r.a.webb ( elsevier , amsterdam , 1991 ) ; g.schn and a.d.zaikin , phys
. rep . * 198 * , 237 ( 1990 ) . | we investigate dynamic properties of inhomogeneous nano - materials , which appear in analytical descriptions typically as series of @xmath0-functions with corresponding gibbs weights .
we focus on observables relevant for transport theories of josephson junction arrays and granular systems near the superconductor insulator transition .
furthermore , our description applies to the theory of tunnel junctions exchanging energy with a `` bath '' , the latter having a discrete spectrum . using the matrix theta - function formalism we find an analytical expression for the transport characteristics capturing the complete temperature driven transition from the quantum to the classical regime . |
the mutual information is @xmath116 where @xmath27 is the entropy of the @xmath6 measurement outcomes and @xmath28 is the conditional entropy of the @xmath6 outcomes , which can be also written as @xmath117 where @xmath118 is the entropy of the probability distribution @xmath119 for fixed @xmath120 . by definition , separable states can be written as @xmath121 .
the conditional state @xmath122 when the result @xmath120 is obtained from a @xmath7 measurement on the second subsystem is @xmath123 in the above expression for @xmath124 the term in the denominator is @xmath125 , namely the probability of getting outcome @xmath120 when measuring @xmath7 on the second subsystem .
( note that , in contrast to entangled states , the components @xmath126 of the first subsystem have not changed , only the spectrum has changed . )
the concavity of the entropy gives @xmath127 where @xmath128 denotes the shannon entropy of a measurement of @xmath43 on the state @xmath41 .
the same reasoning for @xmath13 and @xmath15 yields @xmath129 now we use maassen and uffink s ( mu ) entropic uncertainty relation @xcite , which says that for any state @xmath41 we have @xmath130 with @xmath131 . for complementary observables , @xmath132 .
this means that @xmath133\nonumber\\ & & \geqslant\log_2 d , \labell{ineqab } \ ] ] where the first inequality is due to the concavity of the entropy , the second is the mu inequality . the above chain of inequalities and the fact that @xmath134 , @xmath135 imply that @xmath136 which concludes the proof . the proof given in the previous section employs three inequalities : ( a ) the concavity of the entropy , ( b ) the mu inequality , and ( c ) the upper bounds for the entropy @xmath134 and @xmath135 .
we want now to find the states @xmath41 for which the above three inequalities are equalities .
inequality ( b ) is an equality only if @xmath126 is a pure state and it is an eigenstate of either @xmath6 or @xmath13 @xcite .
we will first assume that all states @xmath126 are eigenstates of @xmath6 , namely @xmath137 .
moreover , the conditions ( c ) become equalities iff the state of the first subsystem ( obtained by tracing over the second ) is @xmath138 . therefore , in order to saturate the inequalities ( b ) and ( c ) , the separable state must be of the form @xmath139 , where @xmath140 are eigenstates of @xmath6 .
since the entropy is strictly concave , inequality ( a ) in eq . is saturated only if the states @xmath126 for all @xmath141 have the same entropy @xmath142 for the outcomes of measurements of @xmath6 . since the @xmath126s are all eigenstates of @xmath6 , also @xmath122 must be an eigenstate of @xmath6 , namely @xmath143 for some @xmath144 ( i.e. @xmath145 ) . from the definition of @xmath124 in eq .
( [ rhob ] ) , this corresponds to @xmath146 being an eigenstate of @xmath7 .
therefore , the form of the input state that fulfills the above requirements is @xmath147 for this state , eq .
( [ conccd ] ) is then automatically satisfied with equality because we can write it as @xmath148 } \otimes|b_l\>\<b_l| \\
= \tfrac1{d^2}\sum_{j}|c_j\>\<c_j|\otimes\sum_l |b_l\>\<b_l| + \sum_i\sum_{j\neq k}\cdots \labell{cc - state2}\;,\end{aligned}\ ] ] where @xmath149 is some phase factor and where in the second line we have separated the part diagonal in @xmath150 ( which is the only one that contributes to the conditional probabilities ) from the rest . in this form , it is clear that the probability distribution of the @xmath13 measurement is uniform and independent of the outcome of the measurement @xmath15 on the second subsystem .
therefore the concavity condition ( a ) , see eq .
( [ conccd ] ) , gives again an equality .
so , if @xmath126 are all eigenstates of @xmath6 , the state that satisfies ( a ) , ( b ) , and ( c ) with equality is the classically correlated state @xmath58 .
the same argument _ mutatis mutandis _ applies to the case in which all @xmath126 are eigenstates of @xmath13 . in order to complete the proof we have to exclude the case where both eigenstates of @xmath6 and @xmath13 appear in the input state @xmath41 , namely a state of the form @xmath151 ( which , by an appropriate choice of @xmath152 and @xmath153 may also describe states such as @xmath154 ) . by taking the partial trace over the second subsystem and then multiplying to the left by @xmath155 and to the right by @xmath156 for each @xmath157
, one sees that this state can not have identity as a marginal to saturate inequality ( c ) unless all terms in the first sum are nonzero .
analogously , using @xmath158 and @xmath159 one sees that also all terms in the second sum are nonzero .
now , in order to saturate the concavity ( a ) of eq .
( [ conc1 ] ) , all states @xmath126 must have the same entropy for measurement of @xmath6 as @xmath122 but , to saturate the mu inequality ( b ) , it must also be equal to a pure state ( either an eigenstate of @xmath6 or of @xmath13 ) .
this is impossible for a state of the form unless @xmath160 or 1 , in which case we go back to the case considered above , that leads to the optimal states @xmath58 of ( [ cc - state ] ) .
here we prove the results on maximal entanglement .
we start from the mutual information @xmath109 and the sum of conditional probabilities @xmath161 .
the case of the pearson coefficient is treated separately below .
start by proving that if @xmath30 or @xmath162 then the state is maximally entangled ( the converse will be proven below ) . from the definition of @xmath163 in eq .
( [ i_ab ] ) , it is upper bounded by @xmath35 and can achieve this bound only when the conditional entropy is null . since the conditional entropy is defined as @xmath164
it is null only when the conditional probabilities are @xmath165 or @xmath102 which implies the result for @xmath80 .
we first prove this for two qubits for simplicity , with the two complementary properties identified by projectors on the computational basis @xmath166 and on the fourier basis @xmath167 .
we prove that if @xmath168 , the state of the two systems is a maximally entangled state @xmath169 .
the conditional probability is @xmath170tr@xmath171 $ ] , where @xmath172 is the state conditioned on obtaining @xmath165 on the second system : @xmath173 } \rightarrow p(0|0 ) = \frac{\<00|\rho|00\ > } { \<00|\rho|00\>+\<10|\rho|10\ > } \labell{r}\;,\end{aligned}\ ] ] where @xmath41 is the initial state of the composite system . from the above expression
we have that @xmath174 only if @xmath175 . by the same argument for @xmath176 we can conclude that @xmath177 , so that @xmath41 must belong to the subspace spanned by @xmath178 .
note that @xmath179 implies that also @xmath180 ( see below ) .
this means that @xmath181 which is @xmath182 if and only if @xmath183 . the above relation holds if and only if @xmath41 is a pure state , and is equal to @xmath184 . in order to prove this we employ the inequality @xmath185 which is an equality only for states that are pure and balanced when restricted to the @xmath186 subspace , namely for states that in such subspace are @xmath187
this inequality can be easily proved by decomposing @xmath188 and using the same inequality for pure states : @xmath189 ( where @xmath190 , @xmath191 ) , which can be easily derived noticing that @xmath192 , where equality holds if and only if @xmath193 .
we will now extend the above proof to couples of systems in arbitrary finite dimension @xmath10 .
the matrix elements of the global state @xmath41 in the computational basis are denoted for brevity as @xmath194 .
the conditions that the conditional probabilities in the computational basis must be one are given by @xmath195 the above condition is satisfied iff @xmath196 , which implies that @xmath197 for @xmath198 .
the same argument holds for the conditional probability @xmath199 and we then have@xmath200 the above condition is satisfied iff @xmath201 for fixed @xmath202 , which implies that @xmath203 for @xmath204 . as a consequence
, the positivity of @xmath41 requires that @xmath205 for @xmath206 .
the above conditions then imply that the only nonvanishing elements of @xmath41 are of the form @xmath207 .
let us now denote as @xmath208 the fourier transform of the computational basis , namely@xmath209the matrix elements in the fourier transformed basis can be written as@xmath210 the conditional probability @xmath211 then takes the form @xmath212 by using the identity @xmath213 which means that the l.h.s . of eq .
( [ iden ] ) vanishes for all values of @xmath141 that are not multiples of the dimension @xmath10 , we can write the denominator in eq .
( [ cdf ] ) as @xmath214 by imposing that the probability in eq .
( [ cdf ] ) must be one we have that @xmath215 by the same reasoning as for the qubit case the above condition implies that @xmath41 is a projector onto the maximally entangled state @xmath216 .
to finish the proof for @xmath109 and @xmath111 , we have to prove the converses : if the state is maximally entangled , then there exist two complementary bases such that @xmath30 or @xmath162 .
any maximally entangled state @xmath217 is local - unitarily equivalent to @xmath218 .
namely , @xmath219 with @xmath220 and @xmath221 unitaries ( which transform local bases one into another ) .
the same result as for @xmath222 can then be achieved by starting from @xmath223 and considering as complementary bases the ones that are obtained by applying @xmath224 to the computational and the fourier bases .
finally , to prove the necessary and sufficient condition for the pearson coefficient @xmath110 , we note that we have just shown that a state has maximal correlation for two complementary bases if and only if it is local - unitarily equivalent to @xmath222 .
namely , in two complementary bases , the local outcomes match : @xmath225 .
it is known that the pearson coefficient takes its extremal values @xmath56 iff the two stochastic variables are linearly dependent and perfectly correlated ( i.e. in one - to - one correspondence ) . here
we prove that cc states can have maximal correlations only on one of two complementary variables and that cq states can not have maximal correlations on any variable .
the first statement is formalized above as : if @xmath88 then we must have @xmath226 , where @xmath90 are the results of the measurements of @xmath91 .
the second statement is formalized as : even if @xmath89 we still can not obtain @xmath227 . we can prove both statements at the same time by observing that cc and cq states can be written in the form @xmath228 which can have maximal correlation on some property only for cc states ( where the @xmath229 are orthogonal for different @xmath230 ) .
we can then show that the state @xmath231 which has some correlation on @xmath230 can not have any correlation on @xmath232 when @xmath232 is a complementary property , namely if @xmath233 .
actually , @xmath234 } \otimes\rho_a \\ = \sum_{c}\frac{1}d|c\>\<c|\otimes\sum_ap_a\rho_a+\sum_a\sum_{c\neq c'}\cdots \labell{lafs}\;,\end{aligned}\ ] ] where @xmath149 is some phase factor and where , as above , in the second line we have separated the part diagonal in @xmath232 ( which is the only one that contributes to the correlations for @xmath232 ) from the rest .
it is clear from the form of the state in that such a state does not have any correlations in @xmath232 : namely that a measurement of @xmath13 on the first system gives no information on the second .
following a suggestion by an anonymous referee , we prove here that the sufficient condition @xmath235 for entanglement of pairs of qubits can be made stronger by adding a third mub for each qubit , which we name @xmath236 and @xmath237 .
actually , for a separable two - qubit state the argument given in the first supplemental section until eq .
( [ conccd ] ) can be applied also for the third pair of bases @xmath238 .
we can then write we can then exploit a generalisation of the mu inequality to the case of three mubs @xmath240 for qubits provided in @xcite , which says that for any qubit state @xmath41 we have @xmath241 .
we can then conclude that separable states fulfill the condition @xmath242 the conjectured bound @xmath243 for separability in terms of pearson correlations can also be extended to three mubs , leading to @xmath244 the latter has been tested numerically ( see also the next section ) . in fig .
[ f : exampleextended ] we report the corresponding results for the mutual information and the pearson correlations with three mubs , to be compared with fig .
[ f : example ] , where only two mubs are considered . as we can see , the improvement is striking .
in particular note that the pearson correlations are now suitable to detect all werner states since they signature the presence of entanglement for @xmath245 . in this section
we study the effectiveness of our entanglement criterion based on complementary correlations to detect entanglement and compare it to known entanglement detection schemes .
we stress that our proposal is not aimed at introducing a new entanglement detection scheme but it is aimed at giving a new interpretation of entanglement in terms of correlations for complementary properties .
nonetheless , interestingly we see that these perform better than entanglement witnesses that employ the same measurements .
moreover , entanglement detection based on correlations for complementary properties allows to detect many entangled states that entanglement witnesses miss . in the following we
provide a comparison between our method and an entanglement detection scheme based on the use of witness operators @xcite . for simplicity
, we compare the case of two qubit states , where entanglement witnesses that involve measurements of three complementary properties of qubits take a simple form , namely @xmath246 if tr@xmath247<0
$ ] for at least one of @xmath248 for a two - qubit state @xmath41 , then the state is entangled .
the first four operators are witness operators optimised for the bell states @xcite , while @xmath249 was added in order to include also the detection scheme based on the inequality @xmath250 for separable states @xcite .
actually , testing for @xmath251 and @xmath249 allows to test the above condition .
we performed a comparison between these detection schemes and our criterion ( extended to all three complementary bases of qubits as shown in the previous section ) with a monte carlo simulation , by considering @xmath252 random states obtained by generating @xmath253 random matrices according to the method detailed in @xcite . whether each state is entangled or not can be evaluated with the ppt criterion @xcite , which is a necessary and sufficient condition in this case
: we found that @xmath254 of the states are entangled , which is consistent with known results @xcite .
the pearson coefficient conjecture detects @xmath255 of these entangled states , whereas the entanglement witnesses detect @xmath256 of them .
therefore , the pearson coefficient turns out to be more powerful for entanglement detection . in fig .
[ f : venn]a and the first panel of table [ ta : pearson ] we compare the performance of these witnesses against the pearson correlation conjecture .
note that @xmath257 of the detected entangled states are not detected by the entanglement witnesses and are seen only by the pearson coefficient conjecture ( this corresponds to a fraction @xmath258 of the entangled stated detected using the pearson coefficient : these are detected only by the pearson correlations ) .
the entanglement witnesses are less effective : they exclusively detect @xmath259 of the detected states ( this corresponds to a fraction @xmath260 of the states detected by the entanglement witnesses ) . in fig .
[ f : venn]b and the second panel of table [ ta : pearson ] we compare also the performance of the sum of conditional probabilities conjecture .
the mutual information does not perform well for entanglement detection , but interestingly most of the ( few ) entangled states it detects are distinct from the ones detected by the witnesses , as seen from the third panel in table [ ta : pearson ] .
.monte carlo simulation results .
the first panel compares the witnesses @xmath261 with the pearson conjecture , the second panel adds also the sum of conditional probabilities conjecture , the third panel adds the mutual information .
the first two tables are graphically depicted in fig . [
f : venn ] .
the percentages depicted refer to the fraction of the entangled states that were detected by at least one of the methods , which is a fraction of @xmath262 of the entangled states in all cases since the mutual information and the sum of conditional entropies detect very few new entangled states that are not detected by the other two schemes : on the order of @xmath263 ) .
we sampled @xmath252 random @xmath253 states and interpreted them as two - qubit states . [ cols="^,^",options="header " , ] we now present a comparison between the pearson conjecture and a relation ( proposed by an anonymous referee ) based on local uncertainty relations ( lur ) @xcite for qubits .
indeed the lur implies that all separable states have @xmath264 .
this is obtained from the uncertainty relation @xmath265 .
this condition is equivalent to saying that all separable states satisfy @xmath266 where @xmath267 is the reduced state over the first subsystem and @xmath268 over the second , and where @xmath269 .
the comparison between this entanglement detection condition and the pearson conjecture is presented in fig .
[ f : referee ] , whence it is evident that the pearson conjecture that uses three complementary observables identifies many more entangled states than , which can identify only very few states ( @xmath270 of the identified ones ) that are not seen by the pearson conjecture .
if one , instead , limits the pearson conjecture to the same two complementary observables employed in as in , then the former procedure results stronger , as it identifies all the entangled states seen . [ [ efficiency - of - complementary - correlations - as - entanglement - criterion ] ] efficiency of complementary correlations as entanglement criterion ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ we want to discuss here the efficiency of our proposal as an entanglement criterion .
it is clear from our proof above that our criterion could be immediately strengthened by considering just the entropic uncertainty relations @xcite instead of considering the mutual information : in fact , inequality is clearly stronger than .
a discussion on these grounds goes beyond our aims : we do not want to provide a strong entanglement criterion , but a different way to interpret entanglement .
nonetheless , the question of how strong is our proposal as entanglement criterion is a relevant one and we will explore it in this section .
let us first distinguish between entanglement detection and entanglement criteria .
the first refers to a measurement procedure , whose aim is to determine whether a given state is entangled or not using measurements that are independent of the state .
the second refers to a general characterization of entanglement : how well does some criterion ( in our case , the classical correlation among complementary observables ) characterize entanglement ?
it turns out that our proposal is quite effective in characterizing entanglement . for low dimensions
, we explore this numerically for mixed states . in fig .
[ f : optim ] we explore numerically the use of the pearson coefficient conjecture as an entanglement criterion by using only two mubs and by optimizing the complementary observables for each entangled state . for qubits ,
optimizing only over the second observable already yields an increase in the detected states : one goes from a fraction @xmath271 of detected entangled states up to @xmath272 .
an optimization over both observables achieves a further increase up to a fraction @xmath273 of entangled states detected .
note that the first number @xmath271 is lower than the one reported in the previous section , where we reported that the pearson correlation detects a much larger fraction of the entangled states , namely @xmath274 of the detected entangled states which are @xmath275 of the total states , namely @xmath276 of the total states are detected as entangled .
the reason that we achieve only @xmath271 here is because we only considered two instead of the three complementary observables of a qubit : that optimization would lead to the increase @xmath277 , as shown in fig .
[ f : optim2]a . as mentioned above , in these simulations we have only considered two complementary observables ( @xmath278 and @xmath279 ) .
further increases can be achieved if one optimizes over all three complementary observables of qubits : see fig .
[ f : optim2 ] . as regards to the percentage given in fig .
[ f : optim2]b , we have to note one should be able to further increase it by optimizing also over all three complementary observables , instead of only one .
preliminary investigations in this direction are promising but numerically very demanding .
it is difficult to provide an estimate of the statistical error since the fluctuations due to the monte carlo simulations are non gaussian in this case .
finally , we mention that our method becomes almost an optimal criterion for large dimensions and pure states according to an argument pointed out by k. zyczkowski @xcite .
unfortunately it is practically impossible to explore numerically this regime .
zyczkowski s argument @xcite is based on the fact that for @xmath280 almost all pure entangled states are close to maximally entangled , since the average entanglement of randomly chosen @xmath10-dimensional pure states goes as @xmath281 @xcite . for large @xmath10 ,
the term @xmath282 becomes irrelevant .
since our criterion gives a necessary - and - sufficient condition for maximally entangled states , it will then asymptotically detect almost all pure entangled states for @xmath280 .
this is only a heuristic argument , however : one must account for the fact that the convergence to maximal entanglement might be so slow that the above argument might only be meaningful in the limit @xmath280 . in this regime
our proof ( valid for finite @xmath10 ) does not hold , and discontinuities might be present @xcite .
t. durt , b .-
englert , i. bengtsson , k. zyczkowski , int .
j. quantum information * 8 * , 535 ( 2010 ) .
d. bru , j. math . phys . * 43 * , 4237 ( 2002 ) ; r. horodecki , p. horodecki , m. horodecki , k. horodecki , rev .
phys . * 81 * , 865 ( 2009 ) .
a. peres , phys .
lett . * 77 * , 1413 ( 1996 ) ; m. horodecki , p. horodecki , r. horodecki , phys .
a * 223 * , 1 ( 1996 ) .
g. vidal , r.f .
werner , phys .
a * 65 * , 032314 ( 2002 ) .
hofmann , s. takeuchi , phys .
a * 68 * , 032103 ( 2003 ) .
j. schlienz and g. mahler , phys .
a * 52 * , 4396 ( 1995 ) . c. kothe , g. bjrk , phys .
a * 75 * , 012336 ( 2007 ) ; i.s .
abascal , g. bjrk , phys .
a * 75 * , 062317 ( 2007 ) .
o. ghne , m. mechler , g. tth , p. adam , phys .
a * 74 * , 010301(r ) ( 2006 ) .
zhang , y .- s .
zhang , s. zhang , g .- c .
guo , phys .
a * 76 * , 012334 ( 2007 ) .
v. giovannetti , phys .
a * 70 * , 012102 ( 2004 ) .
o. ghne and m. lewenstein , phys .
a * 70 * , 022316 ( 2004 ) .
y. huang , phys .
a * 82 * , 012335 ( 2010 ) .
coles , m. piani , phys .
rev . a * 89 * , 022112 ( 2014 ) ; p.j .
coles , phys .
a * 85 * , 042103 ( 2012 ) .
terhal , linear algebra appl . * 323 * , 61 ( 2000 ) , arxiv : quant - ph/9810091 ; o. ghne , g. tth , physics reports * 474 * , 1 ( 2009 ) .
m. lewenstein , b. kraus , j.i .
cirac , p. horodecki , phys .
a * 62 * , 052310 ( 2000 ) .
g. tth , phys .
a * 71 * , 010301(r ) ( 2005 ) ; m.r .
dowling , a.c .
doherty , s.d .
bartlett , phys .
a * 70 * , 062113 ( 2004 ) . c. spengler , m. huber , s. brierley , t. adaktylos , b.c .
hiesmayr , phys .
a * 86 * , 022311 ( 2012 ) ; b.c .
hiesmayr , w. lffler , new j. phys . * 15 * , 083036 ( 2013 ) .
wootters , phys .
lett . * 80 * , 2245 ( 1998 ) .
o. rudolph , quantum inf . process .
* 4 * 219 ( 2005 ) , arxiv : quant - ph/0202121 .
o. ghne , phys .
lett . * 92 * , 117903 ( 2004 ) .
de vicente , quantum inf . comput .
* 7 * , 624 ( 2007 ) ; j.i .
de vicente , m. huber , phys .
a * 84 * , 062306 ( 2011 ) .
o. ghne , p. hyllus , o. gittsovich , j. eisert , phys .
lett . * 99 * , 130504 ( 2007 ) .
o. gittsovich , o. ghne , p. hyllus , j. eisert , phys .
a * 78 * , 052319 ( 2008 ) .
zhang , y .- s .
zhang , s. zhang and g .- c .
guo , phys .
a * 77 * , 060301(r ) ( 2008 ) .
s. wu , z. ma , z. chen , s. yu , sci .
rep . * 4 * , 4036 ( 2014 ) .
j. schneeloch , c.j .
broadbent , s.p .
walborn , e.g. cavalcanti , j.c .
howell , phys .
a * 87 * , 062103 ( 2013 ) ; j. schneeloch , c.j .
broadbent , j.c .
howell phys .
a * 90 * , 062119 ( 2014 ) .
h. maassen , j.b.m .
uffink , phys .
lett . * 60 * , 1103 ( 1988 ) .
e. schrdinger , sitzungsberichte der preussischen akademie der wissenschaften , physikalisch - mathematische klasse 14 : * 296 * , 303 ( 1930 ) .
k. zyczkowski , p. horodecki , a. sanpera , m. lewenstein , phys .
a * 58 * , 883 ( 1998 ) .
foster , p. grassberger , phys .
e * 83 * , 010101(r ) ( 2011 ) .
w. k. wootters , b. d. fields , ann .
phys . * 191 * , 363 ( 1989 ) .
nielsen , j. kempe , phys .
lett . * 86 * , 5184 ( 2001 ) .
a.o . pittenger , m.h .
rubin , phys .
a * 62 * , 032313 ( 2000 ) .
gonzales , j.a .
vaccaro and s.m .
barnett , phys .
a * 205 * , 247 ( 1995 ) .
j. sanchez , phys .
a * 173 * , 233 ( 1993 ) .
o. ghne , p. hyllus , d. bru , a. ekert , m. lewenstein , c. macchiavello and a. sanpera , phys .
a * 66 * , 062305 ( 2002 ) .
o. ghne , p. hyllus , d. bru , a. ekert , m. lewenstein , c. macchiavello and a. sanpera , j. mod . opt . * 50 * , 1079 ( 2003 ) .
k. zyczkowski , p. horodecki , a. sanpera , m. lewenstein , phys .
a * 58 * , 883 ( 1998 ) .
m. pozniak , k. zyczkowski , m. kus , j. phys . a : math
* 31 * , 1059 ( 1998 ) .
the venn diagrams are plotted with the app presented in l. micallef , p. rodgers plos one * 9 * , e101717 ( 2014 ) , http://www.eulerdiagrams.org/eulerape . k. zyczkowski , private communication ( during a memorable hike in the iranian mountains ) ( 2014 ) . k. zyczkowski , h .- j .
sommers , j. phys .
gen . * 34 * , 7111 ( 2001 ) , j. eisert , c. simon , m.b .
plenio , j. phys .
a : math . gen . * 35 * , 3911 ( 2002 ) . | we provide an interpretation of entanglement based on classical correlations between measurement outcomes of complementary properties : states that have correlations beyond a certain threshold are entangled .
the reverse is not true , however .
we also show that , surprisingly , all separable nonclassical states exhibit smaller correlations for complementary observables than some strictly classical states .
we use mutual information as a measure of classical correlations , but we conjecture that the first result holds also for other measures ( e.g. the pearson correlation coefficient or the sum of conditional probabilities ) . two properties of a quantum state are called complementary if they are such that , if one knows the value of one property , all possible values of the other property are equiprobable .
more rigorously , let @xmath0 represent the eigenstates corresponding to possible values of a nondegenerate property @xmath1 , and @xmath2 the eigenstates of a nondegenerate property @xmath3 ( with @xmath4 and @xmath5 arbitrary bijective functions ) .
then @xmath6 and @xmath7 are complementary properties if for all @xmath8 we have @xmath9 , @xmath10 being the hilbert space dimension .
clearly complementary properties with this definition identify two mutually unbiased bases ( mubs ) @xcite . here
we study what classical correlations in the measurements of these complementary properties tell us about the quantum correlations of the state of the system .
typically one discusses entanglement @xcite in terms of non - locality , bell inequality violations , monotones over locc , etc .
for example , previous literature on entanglement focused on time - reversal ( for the ppt criterion @xcite ) , local uncertainty relations @xcite , entropic uncertainty relations @xcite , entanglement witnesses @xcite , concurrence @xcite , the cross - norm criterion @xcite and the covariance matrix criterion @xcite ( the latter encompassing many of the former ) . in contrast to these studies , we focus specifically on classical correlations for complementary properties .
classical correlations are typically quantified in terms of the mutual information , which is the main quantity considered here .
we will also discuss the case of alternative measures such as the pearson correlations and the sum of conditional probabilities . in @xcite related approaches using specific measures of correlations ( different from the ones used here )
were proposed .
the outline of the paper follows .
we start by describing the general scenario we employ for correlation evaluation .
we then introduce different measures of correlations and state our results and our conjectures regarding entanglement and quantum correlations .
we provide some examples of applications .
the details of the proofs of our results are reported in the supplemental material .
= .8 consider two systems of finite dimension @xmath10 ( ) and two observables @xmath11 and @xmath12 ( fig .
[ f : illust ] ) where @xmath6 and @xmath13 are complementary on the first system ( namely @xmath14 for all eigenstates of @xmath6 and @xmath13 ) and @xmath7 and @xmath15 on the second .
for example , take the computational basis of the two systems as the eigenstates of @xmath6 and @xmath7 , and the fourier basis as the ones of @xmath13 and @xmath15 .
we can quantify the correlations between the results of measurements of @xmath6 and @xmath7 with some correlation measure @xmath16 and the correlations between @xmath13 and @xmath15 with @xmath17 .
as @xmath18 below we will define and investigate three possibilities : the mutual information @xmath19 , the sum of conditional probabilities @xmath20 , and the pearson correlation coefficient @xmath21 .
a measure of the overall correlation of the initial state , which we name the `` complementary correlations '' , can then be given as the sum of the absolute value of the two measures @xmath22 or as the product @xmath23 .
the latter is typically a weaker measure than the former , since an upper bound for the sum implies an upper bound for the product .
indeed , @xmath24 implies @xmath25 .
thus we will mainly consider the sum of correlations for complementary observables @xmath22 as a way to evaluate the complementary correlations .
we start considering the mutual information : @xmath26 , where @xmath27 is the shannon entropy of the probabilities of the measurement outcomes of the first system and @xmath28 is the conditional entropy of the outcomes of the first conditioned on the second .
the complementary correlations are then @xmath29 .
the relation of this quantity to the entanglement and the discord of the state of the system is illustrated by the following results : ( i ) the state of a bipartite composite quantum system is maximally entangled if and only if there exist two complementary measurement bases where @xmath30 ; ( ii ) if @xmath31 the state of the bipartite system is entangled ( see also @xcite ) ; ( iii ) the separable states that satisfy this inequality with equality ( i.e. @xmath32 ) , are the classically - correlated ( cc ) zero - discord states of the form @xmath33 with @xmath0 and @xmath34 eigenstates of @xmath6 and @xmath7 ( or the analogous state with a uniform convex combination of eigenstates of @xmath13 and @xmath15 ) .
some examples of @xmath29 for various families of states are plotted in fig .
[ f : example]a , where we emphasize the threshold @xmath35 above which all states are entangled .
the first result follows from the fact that each term in the sum is upper bounded by @xmath35 by definition .
the maximum value for the sum is then @xmath36 and is achievable if and only if there is maximal correlation both between @xmath6 and @xmath7 , and between @xmath13 and @xmath15 .
simple properties of the conditional probabilities ( see supplemental material ) imply that this can happen for a suitable choice of observables if and only if the state is maximally entangled .
the second result is a consequence of the concavity of the entropy and of maassen and uffink s entropic uncertainty relation @xcite ( see supplemental material for the details ) .
it gives a sufficient condition for entanglement that can be used for entanglement detection .
the third result is surprising : one might expect that the separable states at the boundary with the entangled region are highly quantum correlated , whereas we find that they only have classical correlations ( cc ) and no discord .
this means that quantum correlated states without entanglement do not have higher correlations for complementary properties than cc states .
this result is peculiar for the mutual information as a figure of merit , it is no longer true for the pearson correlation ( where a family of qq states sits on the border , as shown in fig .
[ f : example]b ) .
it can be proved by analyzing the conditions for equality of the concavity of the entropy and of maassen and uffink s inequality ( see supplemental material ) .
= 1 . the second measure of correlation we consider is the pearson correlation coefficient @xmath37 , defined as @xmath38 where , as before , @xmath6 and @xmath7 denote observables relative to the two systems , @xmath39tr@xmath40 $ ] is the expectation value on the quantum state @xmath41 and @xmath42 is the variance of the observable @xmath43 .
the above quantity can not be applied to eigenstates of @xmath6 or @xmath7 . clearly , @xmath44 for uncorrelated ( product ) states .
in contrast to the classical pearson correlation coefficient , the quantum one is , in general , complex if @xmath6 and @xmath7 do not commute , but as in the classical case , its modulus is upper - bounded by one : @xmath45\>+\<\{a , b\}\>}2-\<a\>\<b\>|^2 = \nonumber\\&&|\tfrac12{\<[a , b]\>}|^2 + \sigma_a^2\sigma_b^2 \labell{pearsbou}\;,\end{aligned}\ ] ] where @xmath46 $ ] and @xmath47 denote the commutator and anticommutator respectively , and where the final inequality is the schrdinger uncertainty relation @xcite .
we now use @xmath48 as a measure of complementary correlations to recover some entanglement properties of the system state .
the pearson coefficient gauges only the linear correlation of two stochastic variables , so it will not detect maximal correlation even for a maximally entangled state unless pairs of observables are linear in each others eigenvalues ( e.g. it would fail if @xmath49 and @xmath50 ) . however , if one restricts to linear observables , one can prove that a state is maximally entangled if and only if there exist two complementary bases such that @xmath51 , e.g. if one uses @xmath52 , @xmath53 where @xmath54 and @xmath55 are two complementary bases .
the proof follows from the properties of the conditional probabilities ( used to prove the analogous statement for the mutual information ) and from the fact that the pearson coefficient is @xmath56 if and only if there is a functional relation that connects the two stochastic variables ( details in supplemental material ) . instead , for non maximally entangled states we have two conjectures which are supported by numerical evidence : ( i ) if @xmath57 , the two systems are entangled .
as for the mutual information , the inequality is tight since @xmath58 is separable and has @xmath59 ; ( ii ) if @xmath60 , the two systems are entangled . also this inequality is tight : it is attained by the separable state @xmath61 , with @xmath2 eigenstates of @xmath13 . as argued above
, the conjecture with the product is weaker than the one with the sum : proving that all separable states have @xmath62 implies @xmath63 .
the proof of these conjectures is complicated by the fact that the convexity properties of @xmath37 are unknown .
nonetheless , they are natural conjectures that are easy to verify for large classes of states ( e.g. see fig . [
f : example]b ) . we have also performed extensive numerical checks by testing them on large sets of random states generated according to the prescription described in @xcite , and verifying that no state with non - positive partial transpose @xcite lies over the conjectured threshold .
note that the pearson correlation only measures linear correlation , whereas the mutual information measures all types of correlations .
so one could think that the latter is stronger and that these conjectures are implied by the mutual information results of the previous section .
surprisingly , this is false since there exist probability distributions that have maximal pearson correlation but negligible mutual information @xcite .
indeed , consider the family of entangled two - qubit states @xmath64 with @xmath65 $ ] .
if one uses @xmath66 and @xmath67 , for all @xmath68 such state has @xmath69 ( . ] ) , but @xmath70 clearly has negligible mutual information for @xmath71 . in other words ,
the pearson correlation identifies @xmath70 as entangled for all @xmath68 ( assuming the above conjectures ) , whereas the mutual information does not even identify it as classically correlated _ at all _ for @xmath71 . indeed
, numerical simulations suggest that pearson correlation is more effective at detecting entanglement in random states than mutual information .
the third measure of correlation we consider is the sum of conditional probabilities @xmath72 , defined as @xmath73 where @xmath74 is the probability of outcome @xmath75 on the first system conditioned on result @xmath76 on the second .
[ this is a somewhat limited measure of correlations as the correspondence @xmath77 among results is clearly arbitrary
. a more relevant measure of correlation should also maximize ( or minimize ) over the permutations of the measurement outcomes , but for the sake of simplicity we will avoid it . ] in @xcite a similar approach was used , but employing joint probabilities in place of conditional ones .
gauging complementary correlations with the sum @xmath78 we can again obtain information about entanglement and quantum correlations : ( i ) analogously to the case of the mutual information , the sum is optimized only for maximally entangled states : a state is maximally entangled if and only if there exist two complementary bases such that @xmath79 ; ( ii ) as for the pearson correlation , we have a conjecture for non - maximally entangled states : if @xmath80 has a value outside the interval @xmath81 $ ] , we conjecture that the two systems are entangled . as in the previous cases ,
the inequalities are tight since the upper bound is attained by the separable state @xmath58 and the lower bound by the separable state @xmath82 , with @xmath83 sum modulo @xmath10 .
let us analyse the case of separable states .
we remind the reader that classical - quantum ( cq ) and quantum - classical ( qc ) states have the form @xmath84 and @xmath85 respectively , where @xmath86 is a set of orthogonal states for one subsystem and @xmath87 is not an orthogonal set of states .
note that separable quantum - quantum ( qq ) states comprise all separable ones that are not cc , cq or qc .
for these states we can prove that : ( iii ) if cc states have maximal correlations on one of two complementary variables , they are uncorrelated on the other [ formally : if @xmath88 then we must have @xmath89 , where @xmath90 are the results of the measurements of @xmath91 with @xmath6 complementary to @xmath13 and @xmath7 to @xmath15 ] ; ( iv ) cq states can not have maximal correlations on any variable [ formally : we can not obtain @xmath93 , even when @xmath94 ; ( v ) qq states can have only _ partial _ correlation for each complementary property .
for example , the separable two - qubit state @xmath95 has partial correlation on both complementary variables , since @xmath96 .
given the properties ( iii ) and ( iv ) , one might suspect that separable states with non vanishing quantum correlations have always less complementary correlations , but this is not the case , as emphasized by ( v ) . summarizing
, cc states can have maximal correlation only on one property , cq states can not have maximal correlation in any property , and qq states can have some correlation on multiple properties , but you need pure , maximally entangled states to get maximal correlations on more than one property .
regarding the result ( i ) , the proof is a direct consequence of simple properties of conditional probabilities ( see supplemental material ) as for the cases seen previously .
the difficulty in proving the conjecture ( ii ) stems again from a lack of definite concavity properties of @xmath97 , but as for the previous conjecture we have extensively tested it numerically on random states .
one may ask whether the sum over all outcomes in the statement of the conjecture is necessary .
indeed it is : the statement that all separable states satisfy @xmath98 for some @xmath99 is false ( where the two bounds @xmath100 and @xmath101 give the bounds @xmath102 and @xmath103 we used above when the sum over @xmath99 is performed ) .
a counterexample is the separable state @xmath104 of two qubits for which @xmath105 .
if one uses joint probabilities in place of conditional ones , a sufficient condition for entanglement can indeed be proven @xcite .
the results ( iii ) and ( iv ) can be proved at the same time by using simple properties of cc and cq states when they are expressed in two complementary bases ( see supplemental material ) , whereas property ( v ) is a direct consequence of the example provided above . up to now
we have considered the correlations of the measurement outcomes of two complementary observables .
all systems have at least three complementary observables @xcite , and it is known that there are @xmath103 for @xmath10-dimensional systems if @xmath10 is a power of a prime @xcite .
our results can be immediately extended to an arbitrary number of complementary observables by calculating the correlations of all the known complementary observables and considering the sum of the two largest ones .
for example , for mutual information , we can extend the condition to conclude that the state is entangled if @xmath106 where max@xmath107 denotes the second largest term and where @xmath11 , @xmath12 , @xmath108 , etc .
are all observables complementary to each other .
the extensions of all other results and conjectures are analogous .
moreover , at least in the case of qubits the bound at point ( ii ) for the mutual information and the conjectured bound at point ( i ) for the pearson correlations can be made stronger by adding correlations for the third complementary observable .
this can improve significantly the efficiency of the present method if used for entanglement detection . for details and for a comparison with other known entanglement detection schemes based on measurements of mubs
see the supplemental material . in summary
, we have introduced an interpretation of entanglement based on classical correlations of the measurement outcomes of complementary observables .
we have studied different types of correlations ( mutual information @xmath109 , pearson coefficient @xmath110 , and sum of conditional probabilities @xmath111 ) for complementary observables of two systems .
we have shown how they provide information on the entanglement and quantum correlations of a bipartite system .
we have derived the following results and presented a few reasonable conjectures : ( i ) we proved necessary and sufficient conditions for maximal entanglement for @xmath109 , @xmath110 , @xmath111 , ( ii ) we proved sufficient conditions for entanglement based on @xmath109 and conjectured sufficient conditions based on @xmath110 and on @xmath111 ; ( iii ) when gauging complementary correlations using @xmath109 , we proved that the separable states on the boundary with the entangled - states region are strictly classically correlated , but the same result is false if one uses @xmath110 or @xmath111 ; moreover we have shown how @xmath111 provides insight on cc , cq , qc , and qq states , showing that ( iv ) without entanglement only classically correlated cc states can have maximal correlation on one variable ( but then they have no correlation on the complementary one ) , whereas ( v ) separable qq states can have only partial correlations on complementary variables
. one can ask if it is possible to give necessary and sufficient conditions based on correlations for complementary observables . the naive statement that entangled states _ always _ have larger correlations than separable states is false , since it is known that entangled states exist ( e.g. @xmath70 defined above ) that are arbitrarily close to separable pure states @xcite and to the maximally mixed state ( in the sense that for any distance @xmath112 one can choose a sufficiently large dimension @xmath10 such that an entangled state is within distance @xmath112 from the maximally mixed state @xcite ) .
these have vanishing correlations for most measures of correlation .
a notable exception , described above , is the pearson coefficient that is able to detect the entanglement of @xmath70 for all @xmath113 ( but it misses other types of entangled states ) .
we acknowledge useful feedback from b. kraus , k. zyczkowski , and an anonymous referee . |
atom interferometers are very sensitive to inertial effects @xcite and this property was used to build accelerometers @xcite and gyrometers @xcite . because of this large sensitivity , a high mechanical stability of the experiment is required and several experiments @xcite used an active control of the interferometer vibrations . in the present letter ,
we study the phase noise induced by mechanical vibrations in three - gratings mach - zehnder thermal atom interferometers and we show that the rapid decrease of the fringe visibility with the diffraction order is largely due to this phase noise . vibrations displace and bend the rail which holds the three diffraction gratings and we have developed a model based on elasticity theory to describe this dynamics .
we are thus able to understand the contributions of various frequencies and to make a detailed evaluation of this phase noise in the case of our setup : the result is in good agreement with the phase noise value deduced from fringe visibility measurements .
we have built a very stiff rail for our atom interferometer and this arrangement has revealed very efficient : the remaining phase noise is dominated by rotations of the rail , which should be reduced by a better rail suspension .
a phase noise @xmath0 has the effect of reducing the fringe visibility @xmath1 . assuming a gaussian distribution of @xmath2 , the visibility is given by @xcite @xmath3\ ] ] when the phase noise is due to inertial effects
, we prove below that it is proportional to the diffraction order @xmath4 , @xmath5 , where @xmath6 corresponds to the order @xmath4 .
equation ( [ n1 ] ) predicts a gaussian dependence of the fringe visibility @xmath7 with the diffraction order @xmath4 @xcite : @xmath8\ ] ] only two atom interferometers have been operated with several diffraction orders , by siu au lee and co - workers @xcite in 1995 and more recently by our group @xcite .
the observed fringe visibility is plotted as a function of the diffraction order @xmath4 in figure [ fig1 ] .
a gaussian fit , following equation ( [ n2 ] ) , represents very well the data in both cases and the quality of this fit suggests the importance of a phase noise from inertial origin .
we consider a three - grating mach - zehnder interferometer represented in figure [ mzschematic ] .
the inertial sensitivity of this type of interferometer is due to the existence of the diffraction phase which depends on the grating positions .
the resulting phase of the interference signal is given by @xcite : @xmath9\ ] ] here @xmath10 is the grating wavevector ( @xmath11 being the grating period ) ; @xmath4 is the diffraction order and @xmath12 the @xmath13-coordinate of a reference point of grating @xmath14 at time @xmath15 when it is crossed by the atomic wavepacket .
equation ( [ n3 ] ) can be simplified by introducing the atom time of flight @xmath16 from one grating to the next , with @xmath17 and @xmath18 being the atom velocity ( we will neglect its dispersion throughout the present paper ) .
we can express the time @xmath15 as a function of @xmath19 , which will be noted @xmath20 , and @xmath21 : @xmath22 and @xmath23 .
then , if we expand @xmath2 in powers of @xmath21 up second order , we get : @xmath24 t - \frac{\left [ a_{1x}(t ) + a_{3x}(t ) \right ] t^2}{2}\right]\ ] ] here @xmath25 while @xmath26 and @xmath27 are the @xmath13-components of the velocity and acceleration of grating @xmath14 measured by reference to a galilean frame . in equation ( [ n4 ] )
, the first term is due to the instantaneous bending @xmath28 of the rail , the second term represents sagnac effect , as the velocity difference is related to the rail angular velocity , and the third term describes the sensitivity to accelerations .
we want to relate the positions @xmath12 of the three gratings to the mechanical properties of the rail and to its coupling to the seismic noise . as the interferometer is sensitive only to the grating @xmath13-coordinates ,
we use a 1d model to describe the rail dynamics .
this model is based on elasticity theory @xcite in order to describe , in an unified way , the motion and the deformations of the rail . in this model ,
the rail of length @xmath29 along the @xmath30 direction can bend only in the @xmath13 direction .
the shape of its cross - section , assumed to be independent of the @xmath30-coordinate , is characterized by its area @xmath31 and by the moment @xmath32 , the @xmath13-origin being taken on the neutral line .
the rail material has a density @xmath33 and a young s modulus @xmath34 .
the neutral line is described by a function @xmath35 which measures the position of this line with respect to a galilean frame and which verifies @xcite : @xmath36 the rail is coupled to the laboratory by forces and torques exerted at @xmath37 ( @xmath38 ) by its supports .
the @xmath13-component of the force @xmath39 and the @xmath40-component of the torque are respectively related @xcite to the third and second derivatives of @xmath35 with respect to @xmath30 .
we assume that the torques vanish so that @xmath41 at @xmath37 and that the force is the sum of an elastic term proportional to the relative displacement and a damping term proportional to the relative velocity : @xmath42 - \mu \frac{\partial\left[x(\epsilon l , t ) -x_{\epsilon}(t ) \right]}{\partial t}\ ] ] where @xmath43 is the @xmath13-position of the support at @xmath44 and time @xmath20 .
we introduce the fourier transforms @xmath45 and @xmath46 of the functions @xmath35 and @xmath47 .
the general solution of equation ( [ elasticity1 ] ) is : @xmath48 @xmath11 , @xmath49 , @xmath50 and @xmath51 are the four @xmath52-dependent components of @xmath45 .
@xmath50 and @xmath51 are related to @xmath11 and @xmath49 , thanks to assumption of vanishing torques
. then @xmath11 and @xmath49 are linearly related to the source terms @xmath46 by equations ( [ elasticity3 ] ) . @xmath52 and @xmath53 are related by : @xmath54 when @xmath55 is small enough , these equations predict a series of resonances . the first resonance , when @xmath56 , describes an in - phase oscillation of the two ends of the rail while the second resonance , occurring when @xmath57 , describes a rotational oscillation of the rail around its center
. then , there is an infinite series of bending resonances occurring for @xmath58 verifying @xmath59 and @xmath60 related to the @xmath61 , by @xmath62 . in our model , the rail stiffness is described by one parameter only , namely @xmath61 .
the fourier component @xmath64 of the phase @xmath6 given by equation ( [ n3 ] ) can be expressed as a function of the amplitudes @xmath11 and @xmath49 .
we assume that the grating reference points are on the neutral line , at @xmath65 ( @xmath66 ) and we get : @xmath67\end{aligned}\ ] ] where the different lines correspond to the bending , the sagnac and the acceleration terms in this order .
this complicated equation can be given a very simple form by making expansions in powers of @xmath68 and @xmath69 ( assuming @xmath70 for further simplification ) : @xmath71 \frac{3i ( \omega t)}{\left(3-r\right ) } \right .
\nonumber \\ & + & \left .
\left[x_{+}(\omega ) + x_{-}(\omega ) \right ] \frac { 13.0 ( \omega/\omega_0)^2 + ( \omega t)^2 } { 2(1-r ) } \right]\end{aligned}\ ] ] where @xmath72 $ ] . equation ( [ excitation4 ] ) has a limited validity because of numerous approximations but it gives a very clear view of the various contributions .
the first term , proportional to @xmath73 $ ] and to the time of flight @xmath21 , describes the effect of the rotation of the rail excited by the out of phase motion of its two ends .
this term , which is independent of the stiffness of the rail , is sensitive to the rail suspension through the @xmath74 denominator .
the second term is the sum of the bending term , in @xmath75 , and the acceleration term , in @xmath76 .
both terms have the same sensitivity to the suspension of the rail , being sensitive to the first pendular resonance , when @xmath77 .
the bending term is small if the rail is very stiff , i.e. when @xmath61 is large .
when we built our interferometer , we knew that d. pritchard @xcite and siu au lee @xcite had been obliged to reduce @xmath28 in their atom interferometers by a servo - loop .
rather than using a servo - loop , we decided to improve the grating stability by building a very stiff rail .
we use aluminium alloy for its large @xmath78 ratio and we made the largest possible rail in the @xmath13 direction to get a large @xmath79 ratio . using equation ( [ solution11 ] ) , we estimate @xmath80 hz , in reasonable agreement with our measurement , @xmath81 hz , with a rather large q - factor , @xmath82 ( more details in @xcite ) .
the suspension of the rail is very simple , with rubber blocks made to support machine tools . from a rough estimate of their force constant @xmath83 and the rail mass ,
the first resonance is calculated to be at @xmath84 hz . following previous works @xcite , the instaneous bending @xmath28
is conveniently measured by a 3-grating mach - zehnder optical interferometer attached to the gratings of the atom interferometer .
we have built such an interferometer @xcite , with @xmath85 lines / mm gratings from paton hawksley @xcite ( @xmath86 m@xmath87 ) and an helium - neon laser at @xmath88 nm .
the phase @xmath89 of the signal of such an optical interferometer is also given by equation ( [ n4 ] ) simplified because the time of flight @xmath21 for light is negligible : @xmath90 .
in our experiment , the excitation of the rail by the environment gives very small signals , from which we deduce an upper limit of the bending @xmath91 nm . to evaluate the phase noise , we need to know the seismic noise spectrum .
a spectrum was recorded on our setup well before the operation of our interferometer and we use this measurement as a good estimate of the seismic noise .
we have replaced the recorded spectrum with several peaks appearing in the @xmath92 hz range by a smooth spectrum just larger than the measured one .
most of the peaks do not appear on a spectrum taken on the floor , because they are due due to resonances of the structure supporting the vacuum pipes and their exact frequency has probably changed because of modifications of the experiment since the recording .
the smoothed noise spectrum @xmath93 is plotted in figure [ noisespectra ] .
this figure also plots the calculated phase noise spectrum @xmath94 , using equation ( [ excitation1 ] ) , and the sagnac phase noise spectrum @xmath95 deduced from equation ( [ excitation1 ] ) by keeping only the term proportional to the @xmath11 amplitude : clearly , sagnac phase noise is dominant except near the in - phase pendular oscillation and the first bending resonance .
the contribution of the in - phase pendular oscillation depends strongly on its frequency and @xmath96-factor .
the bending resonance is in a region where the excitation amplitude is very low , and , even after amplification by the resonance @xmath96-factor , the contribution of the bending resonance to the total phase noise is fully negligible . in this calculation , we have not used our estimate of the first pendular resonance @xmath97 hz , because the predicted rms value of the bending @xmath98 was considerably larger than the measured upper limit .
we have used @xmath99 hz , with @xmath100 and the measured @xmath61 value , @xmath101 hz and @xmath102 s ( @xmath103 and @xmath104 m / s ) .
we assume that the two excitation terms @xmath105 have the same spectrum but no phase relation , so that we neglect the cross - term @xmath106 .
for very low frequencies up to a few hertz , we expect @xmath107 and the associated correction would cancel the sagnac term and this why we have not extended the @xmath94 curves below @xmath108 hz .
as soon as the frequency is larger than the lowest resonance frequency of the structure supporting the vacuum chambers ( near @xmath109 hz ) , the assumption that @xmath110 and @xmath111 have no phase relation should be good . by integrating the phase noise over the frequency from @xmath108 to @xmath112 hz , we get an estimate of the quadratic mean of the phase noise : @xmath113 this estimate compares well with the value @xmath114 , deduced from the fit of figure [ fig1 ] : we think that the agreement is convincing , if ones considers the large uncertainty on several parameters ( seismic noise , frequency and @xmath96 factors of the pendular resonances ) , .
this result is largely due to sagnac phase noise , as the same integration only on sagnac phase noise gives @xmath115 rad@xmath116 and as shown by equation ( [ excitation4 ] ) , this phase noise can be reduced only by a modification of the rail suspension .
the present paper has analyzed the phase noise induced in a mach - zehnder atom interferometer by mechanical vibrations ( more details in @xcite ) . starting from the well - known inertial sensitivity of atom interferometers ,
we have developed a simple @xmath117d model describing the dynamics of the rail holding the diffraction gratings .
this model provides an unified description of the low- and high - frequency dynamics , in which the rail behaves respectively as a solid object and an elastic object . in the low - frequency range , up to the frequency of the rotational resonance of the suspension , the out - of - phase vibrations of the two ends of the rail induce small rotations , which are converted into phase noise by sagnac effect , and this is the dominant cause of inertial phase noise in our interferometer .
we think that the present analysis is important as it gives access to a reduction of this phase noise in atom interferometers .
a better rail suspension should considerably reduce this phase noise .
then , we would be able to observe atom interference effects with an excellent fringe visibility , close to the fitted value @xmath118 % of figure [ fig1 ] , and we would also be able to work either with diffraction orders @xmath119 or with considerably slower atoms .
j. schmiedmayer , m. s. chapman , c. r. ekstrom , t. d. hammond , d. a. kokorowski , a. lenef , r.a .
rubinstein , e. t. smith and d. e. pritchard , in atom interferometry edited by p. r. berman ( academic press 1997 ) , p 1 | atom interferometers are very sensitive to accelerations and rotations .
this property , which has some very interesting applications , induces a deleterious phase noise due to the seismic noise of the laboratory and this phase noise is sufficiently large to reduce the fringe visibility in many experiments .
we develop a model calculation of this phase noise in the case of mach - zehnder atom interferometers and we apply this model to our thermal lithium interferometer .
we are able to explain the observed phase noise which has been detected through the rapid dependence of the fringe visibility with the diffraction order .
we think that the dynamical model developed in the present paper should be very useful to reduce the vibration induced phase noise in atom interferometers , making many new experiments feasible . |
with the advent of quantum chromodynamics ( qcd ) a serious effort was mounted to understand nuclei in terms of quarks .
early in the development of this line of inquiry , a nucleous with mass number @xmath6 was conceived as a system of @xmath7 quarks moving in a large bag and forming a shell structure @xcite , @xcite .
were this model experimentally successful , one would regard it as an effective theory to low energy qcd which covers both free hadrons and nuclei .
however , the predicted nuclear magnetic moments differ considerably from their observed values @xcite , @xcite .
furthermore , the relationship between the number of quarks @xmath8 that are contained in a stable bag and its size @xmath9 is given by @xmath10 which implies that the link with the liquid drop model , exhibiting the relationship @xmath11 , is lost in this approach .
this seems to be the origin of the discrepance stated by the authors of refs .
@xcite and @xcite .
we thus should proceed from another paradigm .
a direct way for clarifying the properties of a bound many - quark system in the strong - coupling regime is to use the feynman path integral machinery whereby all color degrees of freedom are integrated out , except for those of a single quark @xmath12 , so that this quark is affected by a mean field generated by all other constituents of the system .
a systematic implementation of this calculation program is still a good distance in the future . and yet integrating out the remaining degrees of freedom can be neatly approximated if we make plausible assumptions about the color relief over which the quark @xmath12 travels , and invoke the semiclassical approximation for exploring the behavior of the quark @xmath12 .
this strategy , proposed in @xcite , enables the quarkonium spectra to be found in good agreement with experiment . under a different boundary condition ,
the same procedure makes possible to describe some static properties of nuclei , in particular nuclear sizes characteristic of the liquid drop model .
we will show in the present paper that reasonably accurate values of magnetic moments for a rich variety of nuclei are also attainable in this framework .
the dynamics of the quark @xmath12 , specified by the dirac field @xmath13 , is assumed to be encoded by the lagrangian @xmath14\psi + g_s{\bar\psi}\psi\,\phi\ , , \label { qcd - lagrangian}\ ] ] where @xmath15 and @xmath16 are respectively the lorentz vector- and scalar potential of the mean color field , @xmath17 and @xmath18 their associated couplings , and @xmath19 the current - quark mass of the quark @xmath12 .
the second interaction term in ( [ qcd - lagrangian ] ) is absent from the qcd lagrangian because the scalar yukawa coupling is contrary to asymptotic freedom .
however , our concern is with the effective theory in the infrared region where the dynamics is anticipated to arrange itself into the form shown in eq .
( [ qcd - lagrangian ] ) .
the pseudoscalar coupling @xmath20 is a further fascinating candidate for the color relief , while keeping @xmath21 finite , we find that the dirac equation diagonalizes , so that the system with a pseudoscalar relief enjoys the property of supersymmetry @xcite . ]
whose analysis will be given elsewhere .
the semiclassical treatment implies that the extremal path contribution dominates the feynman path integral .
this is another way of stating that the wave function @xmath22 of the quark @xmath12 is a solution to the dirac equation in the background @xmath23 and @xmath24 because this equation results from the requirement that the action with the lagrangian ( [ qcd - lagrangian ] ) be extremal .
in fact , we restrict our attention to spherically symmetric static interactions , and assume that the contribution of the lorentz vector potential to the mean field is given by @xmath25 , that is , @xmath26 .
what this means is a particle with reduced mass @xmath27 orbits the center of mass being driven by central potentials @xmath28 and @xmath29 .
we are thus dealing with the eigenvalue problem for the dirac hamiltonian , @xmath30
\right\}\psi({\bf r})=\varepsilon\psi({\bf r})\ , , \label { dirac - genera}\ ] ] where @xmath31 , and @xmath32 . the distinction between free hadrons and nuclei is controlled by the phenomenological spin- and pseudospin symmetry conditions ( an extended discussion of spin and pseudospin symmetries can be found in refs .
@xcite and @xcite ) .
we ascribe the spin symmetry condition to free hadrons , and the pseudospin symmetry condition to nuclei .
we take the cornell potential @xcite @xmath33 as a phenomenological realization of both @xmath34 and @xmath35 , and put @xmath36 .
we separate variables in the two - row dirac eigenstates , @xmath37^{(j)}_m\cr ig_{n_r,\ell , j}(r)\left[y^{(\ell_j)}(\theta,\phi)\chi\right]^{(j)}_m\cr}. \label { dirac - eigenstate}\ ] ] here , @xmath38 and @xmath39 are the radial amplitudes , @xmath40 , @xmath41 , @xmath42 are the radial , orbital , and total angular momentum quantum numbers , respectively , @xmath43^{(j)}_m$ ] stands for the coupled amplitude @xmath44 , @xmath45 is the spherical harmonic of order @xmath41 , @xmath46 is given by @xmath47 and @xmath48 , @xmath49 is the spin function .
the operator @xmath50 commutes with the spherically symmetric dirac hamiltonian .
thus , the dirac eigenstates are the eigenstates of this operator with eigenvalues @xmath51 , with ` @xmath52 ' for aligned spin @xmath53 , and ` @xmath54 ' for unaligned spin @xmath55 , so that the quantum number @xmath56 is sufficient to label the orbitals @xcite . the radial part of eq .
( [ dirac - genera ] ) is @xmath57 @xmath58 @xmath59 @xmath60 where the prime stands for differentiation with respect to @xmath61 .
we use ( [ dirac_radia_f ] ) for expressing @xmath62 in terms of @xmath63 and substitute the result in ( [ dirac_radia ] ) .
our concern is with @xmath63 because it is @xmath63 that survives in the nonrelativistic free - particle limit .
we eliminate the first derivative of @xmath63 from the resulting second - order differential equation to obtain @xmath64 @xmath65 @xmath66 once all angular ( orbital and spin ) variables are integrated out , and the pertinent phenomenological condition is imposed on @xmath35 and @xmath34 , the function @xmath67 defined in ( [ k - df ] ) acts as the effective potential .
spin symmetry is inherent in free hadron states @xcite , @xcite .
this symmetry occurs when @xmath68 . with @xmath36 , eqs .
( [ k - df ] ) and ( [ b_1-df ] ) can be solved to give the effective potential @xmath69 . \label { u_eff}\ ] ] the form of the effective potential ( [ u_eff ] )
is shown in figure 1 . ) with the parameters @xmath70 gev , @xmath71 gev , @xmath72 , @xmath73 , @xmath74 . ]
[ spin ] in the nonrelativistic limit @xmath75 , @xmath67 becomes the sum of the centrifugal term and the cornell potential , and hence , the dirac equation reproduces results for the spectrum of quarkonia which were obtained through the use of the schrdinger equation @xcite , @xcite .
we now outline the procedure of looking for numerical solutions to eqs .
( [ dirac_radia_f])([b - df ] ) with imposing the conditions @xmath76 .
since our interest is with solutions that are regular at @xmath77 , we use the anstze @xmath78 and @xmath79 .
the functions @xmath80 and @xmath81 satisfy the following integral equations @xmath82v(s)\ , , \label { u_eq}\ ] ] @xmath83\!u(s)+ \frac{\alpha_s(\varepsilon+m)}{2|\kappa|}\ ! \left[\left(\frac{s}{r}\right)^{2|\kappa|}-1\right]\!v(s)\right\ } , \label { v_eq}\ ] ] where @xmath84 is the heaviside step function .
we write @xmath80 and @xmath81 as the liouville
neumann series , @xmath85 where @xmath86 , @xmath87 , @xmath88 @xmath89v(s)\ , , \label { phi12}\ ] ] @xmath90u(s)\ , , \label { phi21}\ ] ] @xmath91v(s)\ , .
\label { phi22}\ ] ] every term of these series is a polynomial .
the coefficient of a given monomial can be found recursively taking into account that @xmath92r^{k+1}\ , , \label { phi12-r}\ ] ] @xmath93r^{k+2}- ( \varepsilon - m)\left[\frac{\theta(\kappa)}{k+1}+\frac{\theta(-\kappa)}{k+1 + 2|\kappa| } \right]r^{k+1}\ , , \label { phi21-r}\ ] ] @xmath94 the agreement between the energy levels @xmath95 obtained by this means and the masses of well established @xmath96 and @xmath97 states is within @xmath98 for @xmath99 , @xmath73 , @xmath100 gev , @xmath101 gev , see table [ tabone ] which lists the experimental data from @xcite .
.quarkonium masses [ cols="^,^,^,^,^,^",options="header " , ]
the wavefunction classification stemming from pseudospin symmetry is important for both light and very heavy nuclei whose superdeformation appears already at low spin @xcite , @xcite . the pseudospin transformation @xmath102 acts on the angular - spin wave functions of the particle , without affecting the radial motion .
pseudospin degeneracy in heavy nuclei derives from the fact that nucleons in a nucleus move in an attractive scalar , @xmath103 , and repulsive vector , @xmath34 , mean fields , which are nearly equal in magnitude , @xmath104 .
this near equality of mean fields is likely a general feature of any relativistic model which fits nuclear binding energies @xcite .
one can go a step further and consider this symmetry as that arising from a quark @xmath12 moving in a mean field generated by other quarks in the nucleus @xcite .
we adopt the condition @xmath105 where @xmath106 is a constant .
the hamiltonian governing the quark @xmath12 becomes @xmath107 we thus see that @xmath27 is shifted , @xmath108 .
this shift may signal that the current - quark masses become the corresponding constituent - quark masses . in
what follows @xmath27 will be regarded as constituent - quark mass of the quark @xmath12 , and @xmath106 will be omitted .
we put @xmath36 , and solve eqs .
( [ k - df ] ) and ( [ b_1-df ] ) to give @xmath109 @xmath110 ^ 2}+ \frac{\alpha_s(\kappa+1)+\kappa\sigma r^2 } { \sigma r^2-\left(\varepsilon + { m}\right)r-\alpha_s } \biggr\}\ , .
\label { u_eff - pseudo}\ ] ] the terms in the first line of ( [ u_eff - pseudo ] ) closely resembles the respective terms involved in ( [ u_eff ] ) , except for changing the overall factor of the cornell potential , but the terms of the second line dramatically change the situation .
they are singular at @xmath111 which is the positive root of the equation @xmath112 , @xmath113 a plot of @xmath67 defined by ( [ u_eff - pseudo ] ) is depicted in fig .
[ pseudo - spin - potential ] .
[ c][c][0.7]@xmath61,gev@xmath114 [ c][c][0.7]@xmath67,gev ) with the parameters @xmath115 , @xmath116 , @xmath99 , @xmath73 , @xmath74 ] it is thus seen that the condition @xmath118 vastly enhances the interaction between spin degrees of freedom and the mean field to yield a spherical shell of radius @xmath119 on which @xmath67 is infinite .
the boundary of the spherical cavity of radius @xmath119 keeps colored objects in this cavity from escaping .
it is well known @xcite that the tunneling through a potential barrier of the form @xmath120 with @xmath121 is forbidden in one - dimensional quantum mechanics .
this condition is fulfilled by eq .
( [ u_eff - pseudo ] ) .
therefore , the boundary of the cavity sets up an impenetrable quantum - mechanical barrier to every colored object .
specifically , this implies that the wavefunction of the quark @xmath12 is zero outside the cavity .
a singular boundary arises whenever @xmath122 grows indefinitely with @xmath61 .
this is because in going from eqs .
( [ dirac_radia_f ] ) and ( [ dirac_radia ] ) to eq .
( [ 1d_schroedinger ] ) , we have to apply the factor @xmath123 which is infinite when @xmath124 .
such is not the case when we adopt the spin symmetry condition @xmath68 by which @xmath125 .
in contrast , the pseudospin symmetry condition @xmath126 implies that @xmath127 , and @xmath124 has a positive root provided that @xmath34 increases monotonically with @xmath61 beginning at @xmath77 where @xmath34 assumes a negative value .
however , no singular boundary arises when @xmath128 as @xmath129 , where @xmath130 is a constant which is less than @xmath131 .
this is the reason for the absence of confinement from systems with electromagnetic bindings . to verify that the effective potential @xmath67 defined by eq .
( [ u_eff - pseudo ] ) is indeed attributable to the description of nuclei , we solve numerically eqs .
( [ dirac_radia_f ] ) and ( [ dirac_radia ] ) using the parameters @xmath99 and @xmath132 ( borrowed from the description of quarkonia ) , and taking @xmath27 to be 0.33gev .
the procedure closely parallels that outlined in the previous section .
we seek solutions in the form @xmath78 and @xmath79 with @xmath80 and @xmath81 obeying the set of integral equations @xmath133\!u(s)+ ( \varepsilon+m-{\sigma}s)\!\left[\theta(-\kappa ) + \theta(\kappa ) \left(\frac{s}{r}\right)^{2|\kappa|}\right]\!v(s)\right\ } , \label { u_eq-}\ ] ] @xmath134u(s)\ , .
\label { v_eq-}\ ] ] the functions @xmath80 and @xmath81 can be expanded into the liouville
neumann series , @xmath135 with @xmath136 , @xmath137 , and @xmath138 @xmath139u(s)\ , , \label { phi11}\ ] ] @xmath140v(s)\ , , \label { phi12-}\ ] ] @xmath141u(s)\ , .
\label { phi21-}\ ] ] every term of these series is a polynomial .
the coefficient of a given monomial can be found recursively taking into account that @xmath142 @xmath143r^{k+1}- \sigma \left[\frac{\theta(-\kappa)}{k+2}+ \frac{\theta(\kappa)}{k+2 + 2|\kappa|}\right]r^{k+2}\ , , \label { phi12-r2}\ ] ] @xmath144r^{k+1}\ , .
\label { phi21-r-}\ ] ] we find the energy levels @xmath95 for @xmath145 and the corresponding sizes of the cavities @xmath146 . to a good approximation the enery levels @xmath95 turn out to be proportional to @xmath147
, we recast eqs .
( [ dirac_radia_f ] ) and ( [ dirac_radia ] ) as @xmath148g=0 $ ] , where @xmath149 .
since our concern is with large @xmath119 , we may safely omit the terms proportional to @xmath150 and @xmath151 .
this gives the airy equation .
regular solutions to this equation behave asymptotically as @xmath152^{3/2 } + \phi_0\right\}$ ] , which shows that the number of radial nodes @xmath153 is estimated at @xmath154\left[({\varepsilon+m})/{\sigma}\right]^{3/2}$ ] . therefore , @xmath155 . ] . by assuming that @xmath153 is proportional to @xmath156 , where @xmath6 is the nucleon mass number , we come to the relationship @xmath157 with @xmath158 fm for the chosen values of the @xmath159 and @xmath160 , which is characteristic of the liquid drop model . the conjecture that @xmath153 equals the integral part of @xmath156 has far - reaching implications , sending us in search of evidence for or against this conjecture . with this in mind , we compare the magnetic dipole of the quark @xmath12 and that of the nucleus in which this quark is incorporated .
note that the magnetic moment of a nucleon in the states with isospin @xmath161 and the total angular momentum @xmath162 and @xmath163 is given @xcite , respectively , by @xmath164 and @xmath165 .
\label { mu - j+1_2}\ ] ] here , @xmath166 is the orbital gyromagnetic ratio , @xmath167 is the anomalous magnetic moment , @xmath168 , @xmath169 , where @xmath170 is the nucleon magneton .
we reiterate mutatis mutandis the arguments of ref .
@xcite to conclude that ( [ mu - j-1_2 ] ) and ( [ mu - j+1_2 ] ) , in which the infinite limits of integration are replaced by @xmath119 , are well suited for the magnetic moment of neutron - odd nuclei with @xmath171 and @xmath172 to be represented by the magnetic moment of a constituent @xmath80 quark contained in these nuclei , with the understanding that @xmath153 is associated with @xmath173 $ ] .
we determine the quark wave functions @xmath174 and @xmath175 for @xmath99 , @xmath73 , @xmath115 . as to the anomalous magnetic moments of quarks ,
the present notion of their values is far from complete . among suggested values of @xmath176 for @xmath80 quarks @xcite , @xcite
, we adopt to test @xmath177 , and @xmath178 .
3(a , b ) shows that the observed magnetic moments of neutron - odd nuclei with @xmath171 and @xmath172 @xcite agree with the calculated magnetic moments of a @xmath80 quark involved in those nuclei to an accuracy of @xmath179 .
the exceptions are @xmath180 and @xmath181 whose magnetic moments are consistent with the results of our calculation to within @xmath182 .
@xmath180 and @xmath181 have respectively 50 and 52 protons .
50 is a magic number whereas 52 is not .
note also that other neutron - odd nuclei with magic numbers of protons ( @xmath183 and @xmath184 in the state with @xmath185 , and @xmath186 , @xmath187 and @xmath188 in the state with @xmath189 ) do not exhibit this discrepance .
similar calculations can be done for proton - odd nuclei with @xmath171 .
it is seen from fig .
3(c ) that the accuracy of @xmath179 between the results of our calculations and the data for the magnetic moments of such nuclei @xcite can be attained if we turn to the behavior of a @xmath80 quark , but adopt @xmath190 .
the `` wrong '' sign of @xmath176 suggests that we should speak of a hole instead of a real quark .
[ c][c][0.7]@xmath191 [ c][c][0.7]@xmath192 [ r][r][0.6]theory @xmath193 [ r][r][0.6]theory @xmath194 [ r][r][0.6]theory [ r][r][0.6]experiment , ( b ) neutron - odd @xmath195 , ( c ) proton - odd @xmath196 nuclei ] , ( b ) neutron - odd @xmath195 , ( c ) proton - odd @xmath196 nuclei ] , ( b ) neutron - odd @xmath195 , ( c ) proton - odd @xmath196 nuclei ] since a single proton may be regarded as both a free hadron and the lightest nucleus , @xmath197 , that is , a system subject to both spin- and pseudospin symmetry conditions , which are in conflict with each other , our argument is unsuitable for it . in this connection ,
the idea of spherical singular cavity is to be applied to light nuclei ( which fall within a transition region between the spin- and pseudospin symmetry coverages ) with care .
a further justification for the rendition of a nucleus as a singular cavity resulting from the superstrong interaction between spin degrees of freedom of a single quark @xmath12 and the mean field generated by other constituents of this nucleus is the consistency of the nodal wavefunction structure with the pseudospin symmetry condition .
it is well known ( see , e. g. , ref .
@xcite , sect .
15.2 ) that the relativistic harmonic oscillator in the pseudospin limit has positive energy bound states .
furthermore , the energy depends only on pseudo - oscillator quantum number , and the states with pseudo - angular momentum @xmath198 or 1 are all degenerate .
however , the eigenstates of the relativistic harmonic oscillator have the nodal structure @xmath199 which is characterisic of `` antiparticles '' , rather than the nodal structure @xmath200 which is peculiar to conventional positive - energy particles . here , @xmath201 and @xmath202 are the number of internal nodes of @xmath203 and @xmath204 , respectively . by contrast , the eigenstates of a quark @xmath12 executing a periodic motion in the discussed singular cavity have the true nodal structure shown in eq .
( [ nodal - particle ] ) .
this provides us with all essential prerequisites to a refined shell model which proceeds from quarks rather than nucleons .
two novel features of this model are the qcd inspired interquark potential @xmath205 in combination with the pauli exclusion principle for color - neutral three - quark clusters .
this is a challenging task .
indeed , the singular cavity arises from the quark dynamics rather than the nucleon dynamics because a rising potential is peculiar to the former and alien to the latter .
therefore , the radial nodes @xmath153 refer to energy levels of the quark @xmath12 .
the possibility to derive the static properties of a nucleus from the dynamics of a single quark suggests that the notion of nucleon structure of nuclei is somewhat smeared ; when a quark enters a nucleus , it liberates from its parent nucleon .
the evidence in support of this suggestion is that a neutron taken up by a nucleus loses its individuality which is apparent from the fact that the probability of @xmath206-decay of such neutrons may be arbitrarily small , and its life time arbitrarily long . on the other hand , the magic numbers @xmath207 bear on nucleons rather than quarks
. why does the pauli principle selects the three - quark cluster structures ? is the condition of color neutrality for the selected structures crucial ?
it is well known @xcite that the spin - statistics theorem stems from the following conditions : ( i ) the vacuum is the lowest energy state , ( ii ) field variables either commute or anti - commute at spacelike separations , ( iii ) the norm of every physical state is positive definite .
none of these conditions is dynamical .
how can this theorem concerned with `` elementary '' entities ( quarks ) be adapted to systems composed of three such entities ?
let us briefly summarize the main features of our approach to understanding nuclei in terms of quarks .
we integrate out the degrees of freedom of a single quark @xmath12 contained in some nucleus , based on the belief that all remaining degrees of freedom of this nucleus are already integrated out , and the result of this integration can be neatly approximated by using plausible assumptions about the color relief over which the quark @xmath12 moves .
we anticipate that the quark @xmath12 is governed by the lagrangian ( [ qcd - lagrangian ] ) , and invoke the semiclassical treatment which implies that the extremal path contribution dominates the feynman path integral , that is , the wave function of the quark @xmath12 is a solution to the dirac equation in the color background @xmath23 and @xmath24 .
we restrict our attention to spherically symmetric static interactions between the quark @xmath12 and the mean field due to all other constituents of the nucleus , and assume that the contribution of the lorentz vector potential @xmath208 to the mean field is given by @xmath25 , that is , @xmath209 . to control the fact that the quark @xmath12 is a constituent of some nucleus , the pseudospin symmetry condition is imposed on @xmath210 and @xmath32 , namely @xmath126 .
we take the cornell potential ( [ cornell ] ) as a phenomenological realization of both @xmath34 and @xmath35 , and put @xmath36 . as this
take place , the interaction between spin degrees of freedom of the quark @xmath12 and the mean color field becomes superstrong to yield a spherical shell of radius @xmath119 on which the effective potential @xmath211 is infinite .
the boundary of the spherical cavity of radius @xmath119 keeps colored objects in this cavity from escaping .
we solve numerically the dirac equation with the parameters of the cornell potential borrowed from the description of quarkonia , and obtain the energy levels @xmath95 for @xmath212-states and the corresponding sizes of the cavities @xmath119 .
the enery levels @xmath95 turn out to be proportional to @xmath147 . by assuming that @xmath153 is proportional to @xmath156 , where @xmath6 is the nucleon mass number , we come to the relationship @xmath157 , which is characteristic of the liquid drop model . to verify that the assumption @xmath213 $ ] is consistent with the experimental data , we compare the magnetic dipole of the quark @xmath12 and that of the nucleus in which this quark is incorporated .
the agreement between the calculated and observed values of @xmath214 is for the most part within @xmath179 which is better than expected when taken into account that the picture in which a single quark moving in a static spherically symmetric mean field applies to a rich variety of nuclei whose dynamical contents are highly tangled . | we show that the static properties of a nucleus could arise from a single quark moving in a mean field due to all other constituents of this nucleus .
the resulting model provides a way for determining nuclear sizes characteristic of the liquid drop model and reasonably accurate values of magnetic moments of different nuclei .
the static properties of nuclei can be deduced + from the dynamics of a single quark + b. p. kosyakov@xmath0 , e. yu . popov@xmath1 , and m. a. vronski@xmath2 + @xmath1russian federal nuclear center
vniief , sarov , 607190 nizhni novgorod region , russia ; + @xmath3moscow institute of physics & technology , dolgoprudny , 141700 moscow region ; russia ; @xmath4sarov institute of physics & technology , sarov , 607190 nizhni novgorod region , russia .
+ e - mail address : @xmath5 * keywords * : qcd phenomenology |
it is now widely accepted that most normal galaxies have central supermassive black holes ( smbhs ) with @xmath10 ( e.g.,@xcite ) .
high resolution observations of nearby galaxies have revealed a close correlation between the mass of smbhs and the bulge mass , or the velocity dispersion of the bulge ( e.g. , @xcite ) .
this implies that the formation of smbhs is strongly coupled with the formation of galaxies .
however , the formation and evolution of smbhs is still an open question and a hot topic in astrophysics .
the discovery of luminous quasars at redshift @xmath11 shows that smbhs of the order of billions of solar masses exist at the end of the re - ionization epoch ( e.g. , @xcite ) .
considering eddington - limited accretion , @xmath12 quasars may not form readily . unless the seed black hole mass is large , gas accretion with low radiative efficiency and/or bh mergers significantly contribute to the growth of smbhs ( see @xcite ) .
if super - eddington accretion is possible , the growth timescale of bhs can be much shorter than the eddington timescale ( e.g. , @xcite ) .
thus , super - eddington accretion might explain the formation of smbhs .
recently , @xcite predicted that super - eddington accretion is required for the formation of @xmath13 quasars with @xmath14 , because the final bh mass is greatly suppressed by star formation in a circumnuclear disc and agn outflow based on the coevolution model of smbh growth and a circumnuclear disc ( @xcite ; see also @xcite ) .
the occurrence of super - eddington accretion flow has not yet been accurately verified , although this issue has been investigated since the 1970s ( e.g. , @xcite . by using two - dimensional radiation hydrodynamic simulations
, @xcite demonstrated for the fist time that quasi - steady super - eddington disc accretion is possible , because both the radiation field and the mass accretion flow around a bh are non - spherical and the large number of photons generated in the disc is swallowed by the bh without being radiated away due to photon - trapping ( see @xcite ) .
this conclusion has recently been confirmed by two - dimensional radiation magnetohydrodynamic simulations @xcite .
however , it is still unclear whether super - eddington accretion can significantly contribute to the formation of smbhs , because the radiation pressure and/or disc wind associated with super - eddington accretion might play an important role in the gas accretion process in host galaxies ( e.g. , @xcite ) . in the local universe
, we have discovered a class of agns with a higher eddington ratio among the agn population , i.e. , narrow line seyfert 1 galaxies ( nls1s ) .
they have characteristic properties such as narrow balmer lines ( e.g. , @xcite ) , strong soft x - ray excess ( e.g. , @xcite ) and rapid variability ( e.g. , @xcite ) .
these properties indicate that they have a small black hole mass and high eddington ratio ( e.g. , @xcite ) .
the observed bolometric luminosity of nls1s saturates at a few times the eddington luminosity ( e.g. , @xcite ) , which is consistent with predictions of the super - eddington accretion disc ( e.g. , @xcite ) .
in addition , the star formation rate of nls1 hosts is higher than that of bls1s ( @xcite ) . thus , nls1s may be the early phase of rapid bh growth via super - eddington accretion ( e.g. , @xcite ) . in the high-@xmath15 universe , we may speculate that super - eddington agns are a more dominant population of agns because the lifetime of smbh growth is closer to the cosmic age if the formation of smbhs is mainly a gas accretion process . to confirm this ,
it is important to determine whether the fraction of super - eddington agns increases with redshift . some observations to date have indicated that the average eddington ratio increases slightly with redshift ( e.g. , @xcite ) . however , we have not observed super - eddington agns among high-@xmath15 quasars ( e.g. , @xcite ) and agns in high-@xmath15 massive galaxies ( e.g. , @xcite ) .
thus , it is essential to find candidate super - eddington agns before starting detailed observations . according to unification models of agns ,
the accretion disc is surrounded by a dusty structure ( e.g. , @xcite ) .
a significant fraction of the emitted ultraviolet and optical radiation of the accretion disc is absorbed by the dust and is re - emitted at ir wavelengths .
in particular , the hot dust is directly heated by the central engine and produces near - ir ( nir ) emission ( e.g. , @xcite ) . since it has been reported that the polar angle dependence of the radiation flux for super - eddington accretion flow deviates from that for sub - eddington flow ( e.g. , @xcite hereafter w05 ) , the reprocessed ir emission may be useful for exploring candidates of super - eddington accretion flow in agns . in this paper
, we investigate the properties of ir ( as well as nir ) emission from the inner edge of the dusty torus , employing radiative flux from the disc accretion flows by w05 , in which the radiation fluxes of sub- and super - eddington accretion discs are given as functions of the polar angle and the mass accretion rate .
we briefly summarize the relation between the radiation flux from the accretion disc and the mass accretion rate into an smbh .
for a small mass accretion rate , i.e. , @xmath16@xmath17 , where @xmath18 and @xmath19 are the mass accretion rate into an smbh and the eddington luminosity , respectively , if the accretion disc is geometrically thin , it is known as a standard disc @xcite . as the mass accretion rate increases ( @xmath20 ) ,
the disc becomes geometrically thick via strong radiation pressure , i.e. , a slim disc , as first introduced by @xcite .
the maximum thickness of the disc is about @xmath21 because the ratio of the scale height and the disc size is approximately unity ( e.g. , @xcite .
w05 examined the observed spectra for various mass accretion rates and included the effect of disc geometry .
they found that the radiation flux is proportional to @xmath22 for the sub - eddington regime , @xmath23@xmath17 , where @xmath24 is the polar angle ( see fig .
1 ) . the factor @xmath25 represents the change of the effective area via a change of @xmath24 , i.e. , the projection effect . for a super - eddington accretion disc with @xmath26 , although the radiation flux is also proportional to @xmath22 at @xmath27 , it decreases more significantly as @xmath24 increases in regions with @xmath28 .
the bolometric luminosity of an accretion disc is expressed as a function of @xmath29 @xcite as follows : @xmath30 based on the results of w05 , we can describe the radiation flux from the accretion disc , @xmath31 , as follows : for @xmath27 , @xmath32 for @xmath28 , @xmath33 where @xmath34 is the distance from the central bh .
the coefficients of the cosine functions ( see eqs . ( 2 ) and ( 3 ) ) are determined by setting the integral of @xmath31 to be equal to @xmath35 ( see eq .
( 1 ) ) and the continuity of the radiation flux @xmath31 at @xmath36 .
we should note that eqs .
( 2 ) and ( 3 ) well reproduce the results of w05 ( see fig . 3 in w05 ) .
note that we do not take the corona into account in the present study(but see @xmath37 ) .
although x - ray observations have shown the radiation from the disc to be comptonized at the corona above the disc , the size and geometries of the corona are still open questions ( e.g. , @xcite ) .
the optical thickness of the corona is thought to be around unity or less in the case of sub - eddington discs ( e.g. , @xcite ) .
hence , the @xmath24-dependence of the radiative flux should not change much .
global radiation hydrodynamic simulations of super - eddington flows have revealed that the radiative flux is highly collimated , although hot rarefied plasma ( corona ) appears above the disc @xcite .
recently , similar results have been found by global radiation magnetohydrodynamic simulations @xcite .
thus , we assume that the radiative flux of the sub - eddington disc is proportional to @xmath22 and employ the collimated radiative flux from the super - eddington disc in the present study .
is the bolometric luminosity of the accretion disc and @xmath24 is the polar angle .
@xmath2 , @xmath34 and @xmath38 are the half opening angle of the torus , the distance from the bh and the infrared luminosity re - emitted by the inner edge of the dusty torus , respectively .
@xmath39 is the inclination angle .
, height=170 ]
in order to examine the agn ir properties for various mass accretion rates , we model a dusty torus around an smbh plus an accretion disc system ( fig .
1 ) . assuming that the radiation of the accretion disc is absorbed by the dust and the agn ir radiation is re - emitted isotropically at the inner edge of the torus , the agn ir luminosity ( @xmath38 ) depends on only the half opening angle of the torus , @xmath2 , and
is independent of the structure of the inner edge of the torus . in other words ,
energy conservation holds at the inner edge of the torus .
then , @xmath40 can be given by @xmath41 where @xmath34 is the distance from the central bh .
note that eq .
( 4 ) is valid even in a clumpy model ( e.g. , @xcite ) since there are several optically thick clouds along the line of sight and thus all the radiation of the accretion disc is absorbed by clouds near a bh . substituting eqs .
( 2 ) and ( 3 ) into eq . ( 4 ) , the ratio of the agn ir luminosity and the bolometric luminosity is obtained as follows : for @xmath42 , @xmath43 for @xmath44 , @xmath45 these relations are understood in terms of the projection effect , self - occultation and the covering factor of the dusty torus . for @xmath23@xmath17
, the self - occultation effect does not occur because the disc is geometrically thin . since the radiation flux is reduced by the projection effect , @xmath46 ( see eqs .
( 2 ) and ( 3 ) ) , and since the covering factor of the torus is given by @xmath47 , the ir luminosity decreases with an increase of @xmath2 .
the radiation flux is more sensitive to the polar angle in super - eddington accretion discs with @xmath48 than in sub - eddington accretion discs via the self- occultation effect ( see eqs . ( 2 ) and ( 3 ) ) .
hence , the ir luminosity for @xmath49 and @xmath50 drastically decrease as the half opening angle increases when @xmath51 ( see eq .
( 6 ) ) . against @xmath2 for various @xmath29 .
the red line corresponds to @xmath52 , the green line shows the case for @xmath53 and the blue line represents the case for @xmath54 .
, height=245 ] in fig .
2 , we plot the ratio of the ir luminosity to the bolometric luminosity , @xmath55 , against @xmath2 for various @xmath29 .
as mentioned above , @xmath55 decreases with @xmath2 for any @xmath29 , and this trend is much stronger as @xmath29 becomes larger .
super - eddington agns with @xmath26 have much smaller ratios of the ir and bolometric luminosities than sub - eddington agns .
although the ratios of @xmath55 for @xmath56 and @xmath50 are only a few times smaller than that for @xmath57@xmath17 in the case of @xmath58 , they are one or two orders of magnitude lower than that for @xmath59 in a wide range of half opening angle , @xmath7 .
we also find that it is hard for sub - eddington agns to achieve a low @xmath55 , e.g. , @xmath60 , unless the obscuring structure is extremely geometrically thin , e.g. , @xmath61 .
in contrast , we find @xmath60 when @xmath62 for @xmath63 and when @xmath64 for @xmath65 .
thus , except for an extremely thin torus , a low @xmath66 is a signature of the occurrence of super - eddington accretion flow with @xmath26 .
these results are summarized as a schematic diagram in fig .
3 . the bottom left panel represents sub - eddington accretion with a thick torus . in this case , @xmath55 is maximal .
for super - eddington accretion ( @xmath26 ) with a thick torus ( top left panel ) , @xmath55 is slightly smaller than the bottom left panel due to the self - occultation effect . for sub - eddington accretion with a thin torus ( bottom right panel ) , @xmath55 becomes smaller than the case of the bottom left panel , because the radiation of the accretion disc is reprocessed at the small inner surface of the torus .
super - eddington accretion ( @xmath26 ) with a thin torus ( top right panel ) has the smallest @xmath55 in the four panels of fig . 3 due to the self - occultation effect of the disc along with the small covering factor of the torus . on the mass accretion rate into an smbh and the geometry of the dusty torus .
typical values of @xmath55 ( see fig .
2 ) are shown in each panel .
super - eddington agns " are expressed as agns with @xmath67 and sub - eddington agns " are expressed as agns with @xmath68 .
, height=245 ] we here mention the effect of the inclination angle , @xmath39 ( fig .
1 ) . the observed ratio @xmath69 depends on @xmath39 , but @xmath70 is independent of @xmath39 , where @xmath71 is the observed disc luminosity at the viewing angle of @xmath39 . however , the difference between @xmath72 ( face on ) and @xmath39=30@xmath73 is a factor of @xmath74 .
in contrast , in the case of 45 deg @xmath75 , the dependence of @xmath39 on @xmath76 is sensitive to @xmath29 .
for example , the difference between @xmath39=0@xmath73 ( face on ) and @xmath39=60 @xmath73 is a factor of @xmath77 , @xmath78 and @xmath79 for @xmath23@xmath17 , @xmath0 and @xmath80 , respectively .
thus , when @xmath81 deg , the ir luminosity of super - eddington agns could be similar to that of sub - eddington agns . in summary ,
the ir luminosity of super - eddington agns is very faint for @xmath82 , while it is not so faint for 45@xmath83 .
therefore , in order to examine the candidate super - eddington agns effectively by ir observations , it is worth constructing samples of pure type 1 agns ( i.e. , @xmath82 ) rather than type 1.5 and 1.8 agns . in this paper , we assume that the dusty torus is aligned with the accretion disc , such as in fig .
however , this is not a trivial assumption because it is unclear whether the dusty torus is a reservoir of the gas which accretes onto the accretion disc .
if the accretion disc is misaligned with the dusty torus , the self - occultation effect becomes less important as the angle between the rotation axis of the accretion disc and that of the torus increases ( see eqs .
( 2 ) and ( 3 ) ) , since a large number of photons from the accretion disc are absorbed and re - emitted at the dusty torus .
thus , the ratio @xmath55 for super - eddington agns would be comparable to that for sub - eddington agns .
lastly , we note the relation between the agn type and @xmath2 . it may be questioned whether our method is worthwhile if the agn types are mainly determined by @xmath2 and @xmath84 is a standard torus model for all types of agn .
however , recent observations on circum - nuclear regions of agns have revealed that type 2 seyferts are more frequently associated with starbursts than type 1 seyferts ( e.g. , @xcite ) .
in addition , the fraction of type 2 agns decreases with agn luminosity ( e.g. , @xcite ) .
these findings imply that the type1/type2 ratio is not simply determined by the opening angle ( @xmath2 ) .
moreover , the formation mechanism of the obscuring torus is still a hotly debated issue theoretically ( e.g. , @xcite ) .
thus , even if the effect we proposed becomes significant only for relatively large @xmath2 ( @xmath85 ) , our method is useful for exploring the candidates of super - eddington agns .
near - ir ( nir ) emission is produced by the hot dust ( @xmath86 ) which is directly heated by the central agns , though the total ir emission may be contaminated by the nuclear starburst .
thus , based on the results of fig .
2 , we investigate the dependence of the ratio @xmath6 on the mass accretion rate , @xmath29 .
the nir luminosity , @xmath87 , strongly depends on the shape of the inner edge of the torus , because the hot dust temperature , @xmath88 , is sensitive to the structure of the torus .
we here consider the following three typical models for the structure of the inner edge of the torus ( see fig .
first , we examine the case where the inner radius of the torus @xmath89 is determined by the sublimation radius @xmath90 ( fig .
4(i ) ) . here
@xmath90 is defined as the radius where the hot dust temperature is @xmath91 where @xmath92 is the silicate grain sublimation temperature ( e.g. , @xcite ) .
if this is the case , the sublimation radius could be elongated to extend to the central bh in the equatorial plane ( see @xcite ) , since the dust in the equatorial plane can survive because @xmath93 .
since such dust particles at @xmath94 can be heated to @xmath951500 k due to illumination and emit nir ( see fig . 4 ( i ) ) , the nir luminosity is equal to the ir luminosity , i.e. , @xmath96 for all @xmath29 .
thus , we can just replace @xmath97 assigned at the vertical axis in fig .
2 . with @xmath98 . second , we consider the case where the nir emission comes from only a small part of the inner edge ( fig .
4(ii ) ) . in this case
, we define the nir luminosity ( @xmath99 ) as the luminosity emitted by the hot dust ( @xmath100@xmath101 ) .
the nir luminosity is obtained as @xmath102 where @xmath103 is defined as @xmath104 , as seen in fig .
the contribution of nir emission decreases as the mass accretion rate is high , because for a large @xmath29 the hot region at the inner edge of the torus ( the red shaded region in fig .
4(ii ) ) becomes smaller due to self - occultation of the super - eddington accretion flow . as a result
, we find that for a given @xmath2 , the ratio @xmath105 for super - eddington agns with @xmath67 is much smaller than that for sub - eddington agns with @xmath23@xmath17 , as shown in table 1 .
third , we consider the case where the dust temperature of the whole inner edge of the torus becomes lower than @xmath92 ( fig .
4(iii ) ) .
if this is the case , it is expected that there will be no nir emission from the dusty torus for both sub- and super - eddington agns .
however , this prediction is in conflict with nir observations of sub - eddington agns ( e.g. , @xcite ) .
thus , only for nir - faint agns , the dust temperature at @xmath89 might be below the dust sublimation temperature , because the inner radius is larger than the expected sublimation radius . to investigate this possibility , for nir - faint agns ,
it is important to evaluate the time lag between the uv and nir light by photometric monitoring observations ( e.g. , @xcite ) .
dependence of @xmath106 on @xmath2 and @xmath29 + [ cols="^,^,^,^",options="header " , ] taking into account the above results , we can conclude that a relatively low @xmath6 , as well as @xmath107 , is a sign of super - eddington agns with @xmath67 , unless the half opening angle of the torus is small ( @xmath108 ) .
we can also confirm that the predicted @xmath6 with @xmath23@xmath17 ( fig .
2 , see also table 1 ) almost coincides with the observed results of sub - eddington quasars ( e.g. , @xcite ) , assuming @xmath109 where @xmath110 is the agn luminosity at @xmath111 @xcite .
finally , we discuss the evaporation of nearby dusty clouds irradiated by agns . in the steady state for a dusty torus , cloud evaporation at the inner radius of the torus should balance the mass supply of the clouds from the outer region ( see @xcite ) .
if the evaporation timescale , @xmath112 , is longer than the orbital period of a cloud around smbhs , @xmath113 , the inner radius is determined by the sublimation radius , @xmath114 . on the other hand , when @xmath115 , the evaporation could be determined by the location of the inner edge of the torus .
thus , the inner radius of the dusty torus becomes larger than the sublimation radius .
if this is the case , the nir luminosity could be overestimated because the dust temperature is lower .
to check this , we estimate the two timescales ( @xmath112 and @xmath113 ) for the present model . assuming @xmath116 , the orbital period of a cloud at @xmath114 is @xmath117 where @xmath118 is the rotational velocity and @xmath119 . following @xcite , for a cloud of mass @xmath120 and @xmath121 ,
the evaporation timescale is given by @xmath122 note that @xmath123 at @xmath114 is @xmath124 ( eqs.(2 ) and ( 3 ) ) .
comparing these two timescales , we find @xmath125 for sub - eddington agns with @xmath68(see @xcite in details ) . on the other hand , for super - eddington agns with @xmath126 and @xmath127 ,
the evaporation timescale is comparable to the orbital period of a cloud , i.e. , @xmath128 ( see eqs . ( 7 ) and ( 8) )
. if this is the case , evaporation of a dusty cloud would be fast enough to reduce the overall nir emission .
therefore , agns with high @xmath129 and @xmath29 , such as bright narrow - line quasars , might have a much smaller ratio @xmath130 , compared to the results of fig .
so far , we have not taken into account the ir emission from narrow - line region ( nlr ) clouds .
however , it is known that dust grains are present in the nlr of seyferts ( e.g. , @xcite ) .
thus , the dusty clouds in the nlr might be a source of ir emission and hence we must consider the contribution of ir emission from nlr clouds .
nlr clouds are assumed to be distributed in a geometrically thin spherical shell with radius @xmath131 in the range @xmath132 . assuming the covering factor of the nlr clouds , @xmath133 , and the optical depth of the nlr clouds , @xmath134 , the ir luminosity from nlr clouds can be given by @xmath135 where @xmath55 is a function of @xmath2 ( see eqs . ( 5 ) and ( 6 ) ) . in fig .
4 , we plot the ratio of the ir luminosity to the bolometric luminosity against @xmath2 for the nlr clouds , assuming @xmath136 and @xmath137 .
note that the dust extinction of nlr emission is small ( i.e. , @xmath138 ) for type 1 agns @xcite and the typical covering factor is about @xmath139 ( e.g. , @xcite ) . figure 4 shows that the ir emission from the nlr clouds dominates above @xmath140 for @xmath52 , while for @xmath53 and @xmath54 the ir emission from nlr clouds is important above @xmath141 and @xmath142 , respectively .
this suggests that for super - eddington agns with @xmath67 , the ir emission from nlr clouds can dominate the ir emission from the dusty torus .
this is because the radiative flux of super - eddington agns is highly collimated compared to sub - eddington agns ( see eqs .
( 2 ) and ( 3 ) ) .
however , we should mention that the nlr size , @xmath131 , is about @xmath0 times as large as the sublimation radius ( e.g. , @xcite ) and the dust temperature of nlr clouds is lower than the hottest dust temperature ( i.e. , 1500 k ) .
thus , the ir emission from nlr clouds would contribute mainly mir ( 1030 @xmath143 m ) emission ( see @xcite ) . in order to avoid contamination of the ir emission from nlr clouds
, the ratio @xmath6 can be used to probe an inflated disk due to super - eddington accretion , as mentioned in @xmath144 . against @xmath2 for various @xmath29 .
the solid lines show the ratio of the ir luminosity and bolometric luminosity for the nlr clouds , assuming @xmath136 and @xmath137 .
the dashed lines show @xmath145 for the dusty torus ( see fig .
2 ) . , height=245 ] we here discuss the contribution of x - ray emission from the corona on the ir emission . for the model of x - ray emission from the corona , we adopt a simple lamp - post geometry model , where the x - ray source is located above the smbh at some height @xmath146 ( e.g. , @xcite ) .
first , we examine the condition where the x - ray emission can reach the torus inner edge for super - eddington accretion when @xmath147 . if the light ray , which passes though the maximum thickness of the slim disc @xmath148 from the x - ray source located at @xmath146 , hits the torus inner edge , @xmath114 , a large amount of x - ray emission is absorbed by the dust torus .
assuming @xmath149 , this condition can be simply expressed by the following : @xmath150 where @xmath151 is the radius of the slim disc and @xmath152 because the half opening angle of the slim disc is @xmath21 ( see @xmath153 ) .
for example , the x - ray emission could produce the ir emission , if @xmath154 and @xmath155 for @xmath156 and @xmath157 , respectively .
according to the theory of slim discs , @xmath151 is about @xmath158 for @xmath53 .
note that for @xmath54 , @xmath151 is much larger than @xmath158 ( e.g. , @xcite ) .
this may imply that the contribution of x - rays from the corona is not negligible if @xmath146 is larger than @xmath159@xmath160 .
the lower limit of @xmath146 is comparable to @xmath161 , which is frequently used to discuss the iron - line reverberation in agns ( e.g. , @xcite ) .
we next investigate the contribution of x - rays from the corona on the ir emission when @xmath162@xmath163 .
the ratio of the radiation flux from the accretion disc , @xmath164 , and the corona , @xmath165 , at @xmath2 is expressed as @xmath166 here @xmath167 is the distance from the x - ray source to the inner radius of the dusty torus .
assuming that the x - ray emission is isotropic , it is found that as @xmath2 and @xmath29 increase , the x - ray emission from the corona begins to contribute to the ir emission from the dusty torus .
however , we should keep in mind that the dust opacity for x - ray radiation is lower than that for optical - uv radiation ( e.g. , @xcite ) .
thus , the evaluated contribution of x - rays from the corona could be overestimated . and
@xmath92 are the dust temperature and sublimation temperature , respectively .
note that @xmath168 is the sublimation radius at @xmath169 .
the red thick lines and shaded portions denote the region of nir emission ( i.e. , @xmath170 ) .
the blue shaded portions show the region of no nir emission ( i.e. , @xmath171 ) .
, height=188 ] in @xmath144 , we considered three typical models for the configuration of the inner edge of the dusty torus ( see fig .
we here discuss how these three types can be understood physically .
the following two physical quantities may be closely related to the shape of the inner part of the dusty torus .
one is the sublimation radius , @xmath90 , which determines the region where hot dust can survive . in order to maintain a geometrically thick structure ,
energy input from stars and/or supernovae in the dusty torus is necessary ( e.g. , @xcite ) .
this energy input may occur outside of the critical radius , @xmath172 , above which the torus is gravitationally unstable , if we adopt toomre s stability criterion .
thus , the magnitude relation between @xmath168 and @xmath172 may determine the structure of the inner part of the dusty torus , as in the following three cases .
( i ) if @xmath173 , the inner radius of the torus , @xmath89 , may be determined by @xmath90 .
this corresponds to the case of fig .
( ii ) when @xmath174 , the structure of the dusty torus is like fig .
as mentioned above , the dust can survive at even @xmath175 , though the structure within @xmath172 is geometrically thin because there is no energy source in this region .
thus , the nir emission from @xmath176 is negligible since the covering factor becomes small .
that is , @xmath89 is determined by the radius for which the dusty torus is gravitationally unstable .
( iii ) if @xmath177 , the dust temperature of the whole inner edge of the dusty torus becomes lower than @xmath92 because @xmath178 is much larger than @xmath90 .
if this is the case , the structure of the dusty torus is similar to fig .
by exploring the relation between the region of nuclear starbursts and that of the hot dust , we may find evidence revealing the formation of the dusty torus .
we now discuss the evolutionary tracks of smbh growth in the @xmath55@xmath2 diagram . as mentioned in @xmath180 , although the formation mechanism of the obscuring torus is still a hotly debated issue ( e.g. , @xcite ) , @xcite proposed a dusty torus supported by turbulent pressure , in which the turbulence is produced by sn explosions ( see also @xcite ) . in their model , the geometrical thickness of the torus , which is determined by the balance between the gravity of the central bh and the force due to turbulent pressure , is smaller for more massive bhs .
this tendency is consistent with observations in which the covering factor of the dusty torus decreases with increasing bh mass ( e.g. , @xcite ) . this model also implies that @xmath2 increases with time as the bh grows .
that is , the evolution proceeds from left to right in the @xmath55@xmath2 diagram ( see fig .
we consider two simple scenarios : ( i ) the super - eddington growth dominated scenario where most of the mass of the smbhs is supplied not through sub - eddington accretion discs but through super - eddington accretion discs .
( ii ) the sub - eddington growth dominated scenario where the sub - eddington accretion disc mainly feeds the smbh .
figure 6 shows the evolutionary tracks for these two scenarios , assuming an initial opening angle of the torus of @xmath181 and initially @xmath54 .
we find from this figure that the agns stay in a super - eddington phase ( @xmath182 ) for a long time ( a wide range of @xmath2 ) in case ( i ) , whereas in case ( ii ) the agns shift to a sub - eddington phase when the torus is relatively thick .
the two distinct scenarios can be clearly understood from fig . 3
as follows : for case ( i ) the evolution proceeds from top left @xmath183 top right @xmath183 bottom right , whereas for case
( ii ) it goes from top left @xmath183 bottom left @xmath183 bottom right .
that is , the agns never undergo a phase of extremely low @xmath55 for case ( ii ) . in contrast , in case ( i ) , many agns can be identified as ir faint objects .
the nir luminosity of such agns is also very small .
the dispersion of @xmath55 as well as @xmath6 for case ( i ) is significantly larger than that for case ( ii ) .
if the majority of high-@xmath15 quasars ( @xmath184 ) are nir - faint , super - eddington growth may be inevitable for smbh growth in the early universe . in order to compare the predictions with observations in detail
, we must elucidate the evolution of @xmath6 based on the coevolution model of smbh growth and a circumnuclear disc such as @xcite .
this is a subject for future work .
increases ( see text ) .
we assume that the initial opening angle of the torus is @xmath36 and @xmath54 .
, height=245 ] in this paper , we predicted that super - eddington agns with @xmath67 have relatively low @xmath8 .
so far , more than 30 quasars have been discovered at @xmath185 ( e.g. , @xcite ) .
@xcite found that two are unusually nir - faint quasars , i.e. , their ratio of @xmath186 is one order of magnitude smaller than the average for ordinary quasars , although their properties are similar to those of low-@xmath15 quasars in the rest - frame ultraviolet / optical and x - ray bands ( e.g. , @xcite ) .
more interestingly , the two quasars have the smallest bh mass ( @xmath187 ) and the highest eddington ratios ( @xmath188 ) of @xmath9 quasar samples .
these features can be nicely explained by our interpretation , i.e. , the faint nir emission for high eddington agns .
we should note that since nir - faint quasars can be explained by sub - eddington agns with extremely thin tori ( i.e. , @xmath189 ) , we need to measure the thickness of a dusty torus to confirm whether nir - faint quasars are really super - eddington agns ( @xmath190 ) . based on the kinematics of cold molecular gas in a torus , it would be possible to evaluate the thickness of a torus by alma , because the ratio of velocity dispersion and rotational velocity indicates the scale height of the torus ( see @xcite ) . on the other hand , @xcite interpreted these two objects as dust - free quasars .
however , this seems to contradict the observations , which suggests that nir - faint quasars possess super - metallicity as do other @xmath9 quasars and high-@xmath15 agns ( e.g. , @xcite ) .
also , @xcite recently reported quasars with unusually faint nir emission at @xmath191@xmath192 , when the universe was 24 gyr old .
thus , it seems likely that nir - faint quasars are exclusively dust - free agns . as an alternative scenario ,
a large fraction of gas in a dusty torus would be ejected from host galaxies as a result of strong radiation pressure from the brightest agns , such as quasars @xcite . in this case , nir - faint quasars may be explained as being quasars without dusty tori around the smbh .
this possibility could be confirmed by investigating the presence of dusty tori in nir - faint quasars using alma . in the local universe , nls1s
are thought to be agns of a high @xmath193 system .
one nls1 ( ark 564 ) shows no nir emission @xcite , which is consistent with our predictions . however , most nls1s tend to have nir emission similar to or stronger than that of ordinary seyfert galaxies ( e.g. @xcite ) .
one of the reasons for the discrepancy between our predictions and the observations is that nls1s with @xmath67 may be absent ( or their fraction is very small ) .
however , some nls1s have a high mass accretion rate with @xmath67 ( e.g. , @xcite ) .
another possibility is that these discrepancies may suggest misalignment between the accretion discs and the dusty torus , as mentioned in @xmath180 .
lastly , we note the possibility of contamination of hosts and circumnuclear starbursts in the nir band , if the host luminosity is comparable to or exceeds the agn luminosity , as in seyfert galaxies . although it may be hard to find many nls1s with low @xmath6 in current observations , in future studies it is worth examining this issue by nir ( rest frame ) observations with high spatial resolution .
in this work , we propose a new method to search for candidate galaxies in which super - eddington growth occurs .
in particular , taking account of the dependence of the polar angle on the radiation flux of accretion flows , we investigate the properties of reprocessed ir emission ( as well as nir emission ) from the inner edge of the dusty torus . as a result
, we find that a relatively low ratio of agn ir ( as well as nir ) luminosity and disc luminosity is a genuine property that can be used for exploring for candidates of super - eddington agns with @xmath67 .
this method is especially powerful for searching for high-@xmath15 super - eddington agns because the direct measurement of @xmath29 is difficult for high-@xmath15 agns compared with nearby agns .
the following are the main results of the present paper : 1 .
we find that the ratio of the agn ir luminosity , @xmath38 , and the disc bolometric luminosity , @xmath35 , decreases as the half opening angle of the dusty torus , @xmath2 , increases and the mass accretion rate normalized by the eddington rate , @xmath29 , increases .
the dependence of the ratio on @xmath29 is caused by the self - occultation effect , which attenuates the illumination of the torus via the self - absorption of the emission from the inner disc surface at the outer region of super - eddington accretion discs .
the ratio in super - eddington agns with @xmath67 is more than one order of magnitude lower than that in sub - eddington agns with @xmath23@xmath17 for @xmath7 ( @xmath194 ) and for @xmath195 ( @xmath196 ) .
a small value of @xmath197 could be a signature of super - eddington accretion with @xmath67 , since it can not be realized in sub - eddington agns , unless the torus has quite a wide opening angle , @xmath4 .
we also examined the properties of the agn near - ir luminosity ( @xmath87 ) radiated from hot dust at @xmath198 , because it may be hard to detect pure agn ir emission from the inner edge of a torus because of contamination of the nuclear starburst . in this case , @xmath199 of super - eddington agns with @xmath67 is at least one order of magnitude smaller than that for sub - eddington agns over a wide range of half opening angle ( @xmath200 ) .
this is true regardless of which of the three typical models of the dusty torus is used .
we appreciate the useful comments of the anonymous referees that reviewed this paper .
we are grateful to m. umemura and t. nagao for useful comments and discussions .
this work is supported in part by the ministry of education , culture , sports , science , and technology ( mext ) research activity start - up 2284007 ( nk ) and young scientist ( b ) 20740115 ( ko ) .
abramowicz m. a. , czerny b. , lasota j. p. , szuszkiewicz e. , 1988 , apj , 332 , 646 antonucci r. , 1993 , ara&a , 31 , 473 barvainis r. , 1987 , apj , 320 , 537 baskin a. , laor a. , 2005 , mnras , 358 , 1043 begelman m. c. , 1978 , mnras , 184 , 53 bennert n. , falcke h. , schulz h. , wilson a. s. , wills b. j. , 2002 , apj , 574 , l105 boller th . , brandt w. n. , fink h. h. , 1996 , a&a , 305 , 53 boller th . , fabian a. c. , sunyaev r. , et al . , 2002 , mnras , 329 , l1 boller th . , tanaka y. , fabian a. c. , et al . , 2003 , mnras , 343 , l89 brandt n. , boller t. , 1998 , astron .
nach . , 319 , 163 collin s. , kawaguchi t. , 2004 , a&a , 426 , 797 collin s. , kawaguchi t. , peterson b. m. , vestergaard m. , 2006 , a&a , 456 , 75 dahari o. , de robertis m. m. , 1988 , apj , 331 , 727 draine b. t. , lee h. m. , 1984 , apj , 285 , 89 elitzur m. , 2008 , new astronomy review , 52 , 274 elvis m. , et al .
, 1994 , apjs , 95 , 1 fabian a. c. , 1999 , mnras , 308 , l9 fan x. , et al . , 2001 , aj , 122 , 2833 , 2004 , aj , 128 , 515 , 2006 ,
aj , 131 , 1203 ferrarese l. , merritt d. , 2000 , apjl , 539 , l9 freudling w. , corbin m. r. , korista k. t. , 2003 , apj , 587 , l67 fukue j. , 2000 , pasj , 52 , 829 gallo l. c. , boller th . , tanaka y. , et al .
2004 , mnras , 347 , 269 gillessen s. , eisenhauer f. , trippe s. , alexander t. , genzel r. , martins f. , ott t. , 2009 , apj , 692,1075 george i. m. , fabian a. c. , 1991 , mnras , 249 , 352 [ glass i. s. , 2004 , mnras , 350 , 1049 glikman e. , helfand d. j. , white r. l. , 2006 , apj , 640 , 579 gonzlez delgado r. m. , heckman t. , leitherer c. , 2001 , apj , 546 , 845 goto t. , 2006 , mnras , 371 , 769 haas m. , et al . , 2003 , a&a , 402 , 87 haba y. , terashima y. , kunieda h. , ohsuga k. , 2008 , pasj , 60 , 487 hamann f. , ferland g. , 1993 , apj , 418 , 11 hao h. , et al . , 2010 , apjl , 724 , l59 hring n. , rix h .- w . , 2004 , apjl , 604 , l89 hasinger g. , 2008 , a&a , 490 , 905 hayashida k. , 2000 , new astron . rev . , 44 , 419 heckman t. m. , blitz l. , wilson a. s. , armus l. , miley , g. k. , 1989 , apj , 342 , 735 hnig s. f. , beckert t. , ohnaka k. , weigelt g. , 2006 , a&a , 452 , 459 houck j. c. , chevalier r. a. , 1991 , apj , 376 , 234 hunt l. k. , malkan m. a. , salvati m. , mandolesi n. , palazzi e. , wade r. , 1997 , apjs , 108 , 229 jiang l. , et al . , 2010 , nature , 464 , 380 jiang l. , et al . , 2006 , aj , 131 , 2788 juarez y. , maiolino r. , mujica r. , pedani m. , marinoni s. , nagao t. , marconi a. , oliva , e. , 2009 , a&a , 494 , l25 kaspi s. , smith p. s. , netzer h. , maoz d. , jannuzi b. t. , giveon u. , 2000 , apj , 533 , 631 kato s. , fukue j. , mineshige s. , 1998 , black - hole accretion discs ( kyoto : kyoto univ . press ) kato s. , fukue j. , mineshige s. , 2008 , black - hole accretion discs towards a new paradigm ( kyoto : kyoto univ .
press ) kawakatu n. , wada k. , 2008 , apj , 681 , 73 kawakatu n. , wada k. , 2009 , apj , 706 , 676 kawaguchi t. , aoki k. , ohta k. , collin s. , 2004 , a&a , 420 , l23 kawaguchi t. , mori m. , 2010 , apjl , 724 , l183 kawanaka n. , kato y. , mineshige s. , 2008 , pasj , 60 , 399 king a. , 2003 , apj , 596 , l27 kollmeier j. a. , et al . , 2006 , apj , 706 , 676 kormendy j. , richstone d. , 1995 , ara&a , 33 , 581 krolik j. h. , begelman m. c. , 1988 , apj , 329 , 702 kubota a. , done c. , 2004 , mnras , 353 , 980 lamastra a. , bianchi s. , matt g. , perola g. c. , barcons x. , carrera f. j. , 2009 , a&a , 504 , 73 laor a. , draine b. t. , 1993 , apj , 402 , 441 leighly k. m. , 1999 , apjs , 125 , 297 maiolino r. , ruiz m. , rieke g. h. , papadopoulos p. , 1997 ,
apj , 485 , 552 maiolino r. , shemmer o. , imanishi m. , netzer h. , oliva e. , lutz d. , sturm e. , 2007 , a&a , 468 , 979 malkan m. a. , gorjian v. , tam , r. , 1998 , apjs , 117 , 25 marconi a. , hunt l. k. , 2003 , apj , 589 , l21 mathur s. , 2000 , mnras , 314 , l17 matsuoka k. , et al . , 2009 , a&a , 503 , 721 mclure r. j. , dunlop , j. s. , 2004 , mnras , 352 , 1390 mineshige s. , kawaguchi t. , takeuchi m. , hayashida k. , 2000 , pasj , 52 , 499 minezaki t. , yoshii y. , kobayashi y. , enya k. , suganuma m. , tomita h. , aoki t. , peterson , b .
a. , 2004 , apj , 600 , l35 miyoshi m. , moran j. , herrnstein j. , greenhill l. , nakai n. , diamond p. , inoue m. , 1995 , nature , 373 , 127 mor r. , netzer h. , elitzur m. , 2009 , apj , 705 , 298 mushotzky r. f. , winter l. m. , mcintosh d. h. , tueller j. , 2008 , apjl , 684 , l65 nagao t. , maiolino r. , marconi a. , 2006a , a&a , 447 , 863 nagao t. , maiolino r. , marconi a. , 2006b , a&a , 447 , 157 nayakshin s. , kallman t. r. , 2001 , apj , 546 , 406 nenkova m. , sirocky m. m. , nikutta r. , ivezi . , elitzur m. , 2008 , apj , 685 , 160 netzer h. , laor a. , 1993 , apj , 404 , l51 nenkova m. , ivezi . , elitzur m. , 2002 , apj , 570 , l9 netzer h. , shemmer o. , maiolino r. , oliva e. , croom s. , corbett e. , di fabrizio l. , 2004 , apj , 614 , 558 noguchi k. , terashima y. , ishino y. , hashimoto y. , koss m. , ueda y. , awaki , h. , 2010 , apj , 711 , 144 ohsuga k. , umemura m. , 1999 , apjl , 521 , l13 ohsuga k. , umemura m. , 2001 , apj , 559 , 157 ohsuga k. , et al . , 2005 , apj , 628 , 368 ohsuga k. , 2007 , apj , 659 , 205 ohsuga k. , mineshige s. , 2007 , apj , 670 , 1283 ohsuga k. , mineshige s. , mori m. , kato y. , 2009 , pasj , 61 , l7 osterbrock d. e. , pogge r. w. , 1985 , apj , 297 , 166 otani c. , kii t. , miya k. , 1996 in rntgenstrahlung from the universe ( mpe report 263 ) , ed .
h. u. zimmermann , j. e. trmper , h. yorke ( garching , mpe press ) , 491 pentericci l. , fan x. , rix h .- w . , et al . , 2002 ,
aj , 123 , 2151 perez - olea d. e. , colina l. , 1996 , apj , 468 , 191 pier e. a. , voit g. m. , 1995 , apj , 450 , 628 pier e. a. , krolik j. h. , 1992 , apjl , 399 , l23 pogge r. w. , 2000 , new astron .
, 44 , 381 pounds k. a. , done , c. , osborne , j. , 1995 , mnras 277 , l5 reynolds c. s. , begelman m. c. , 1997 , apj , 488 , 109 rieke g. h. , 1978 , apj , 226 , 550 rodrguez - ardila a. , pastoriza m. g. , donzelli c. j. , 2000 , apjs , 126 , 63 rodrguez - ardila a. , mazzalay x. , 2006 , mnras , 367 , l57 ryan c. j. , de robertis m. m. , virani s. , laor a. , dawson p. c. , 2007 , apj , 654 , 799 sani e. , lutz d. , risaliti g. , netzer h. , gallo l. c. , trakhtenbrot b. , sturm e. , boller t. , 2010 , mnras , 403 , 1246 schartmann m. , meisenheimer k. , klahr h. , camenzind m. , wolf s. , henning , t. , 2009 , mnras , 393 , 759 schmitt h. r. , storchi - bergmann t. , cid fernandes r. , 1999 , mnras , 303 , 173 schweitzer m. , et al . , 2008 , apj , 679 , 101 shakura n. i. , syunyaev r. a. , 1973 , a&a , 24 , 337 shapiro s. l. , 2005 , apj , 620 , 59 shemmer o. , et al . , 2006 , apj , 644 , 86 shen y. , greene j. e. , strauss m. a. , richards g. t. , schneider d. p. , 2008
, apj , 680 , 169 silk j. , rees m. j. , 1998 , a&a , 331 , l1 storchi - bergmann t. , raimann d. , bica e. l. d. , fraquelli h. a. , 2000 , apj , 544 , 747 suganuma m. , et al .
, 2006 , apj , 639 , 46 thompson t. a. , quataert e. , murray n. , 2005 , apj , 630 , 167 ueda y. , akiyama m. , ohta k. , miyaji t. , 2003 , apj , 598 , 886 volonteri m. , rees m. j. , 2005 , apj , 633 , 624 wada k. , norman c. a. , 2002 , apjl , 566 , l21 wada k. , tomisaka k. , 2005 , apj , 619 , 93 wada k. , papadopoulos p. p. , spaans m. , 2009 , apj , 702 , 63 watabe y. , umemura m. , 2005 , apj , 618 , 649 watarai k .- y , fukue j. , 1999 , pasj , 51 , 725 watarai k .-
y , fukue j. , takeuchi m. , mineshige s. , 2000 , pasj , 52 , 133 watarai k .- y . , ohsuga k. , takahashi r. , fukue , j. , 2005 , pasj , 57 , 513 willott c. j. , et al .
, 2010 , aj , 140 , 546 willott c. j. , et al . , 2007 , aj , 134 , 2435 yamada t. , et al . , 2009 , apj , 699 , 1354 | we propose a new method to explore the candidate super - eddington active galactic nuclei ( agns ) . we examine the properties of infrared ( ir ) emission from the inner edge of the dusty torus in agns , which are powered by super- or sub - eddington accretion flows around black holes , by considering the dependence of the polar angle on the radiation flux of accretion flows @xcite .
we find that for super - eddington agns , of which the mass accretion rate is more than @xmath0 times larger than the eddington rate , the ratio of the agn ir luminosity and the disc bolometric luminosity is less than @xmath1 , unless the half opening angle of the torus ( @xmath2 ) is small ( @xmath3 ) .
this is due to the self - occultation effect , whereby the self - absorption at the outer region of the super - eddington flow dilutes the illumination of the torus .
such a small luminosity ratio is not observed in sub - eddington agns , whose mass accretion rate is comparable to or no more than 10 times larger than the eddington mass accretion rate , except for extremely thin tori ( @xmath4 ) .
we also consider the properties of the near - ir ( nir ) emission radiated from hot dust @xmath5 1000 k. we find that super - eddington agns have a ratio of the nir luminosity to the bolometric luminosity , @xmath6 , at least one order of magnitude smaller than for sub - eddington agns for a wide range of half opening angle ( @xmath7 ) , for various types of dusty torus model .
thus , a relatively low @xmath8 is a property that allows identification of candidate super - eddington agns .
lastly , we discuss the possibility that nir - faint quasars at redshift @xmath9 discovered by a recent deep sdss survey may be young quasars whose black holes grow via super - eddington accretion .
[ firstpage ] accretion : accretion disc black hole physics galaxies : active |
the discretized approach of 2d quantum gravity , implemented by the matrix model , exhibits , in a continuum limit the same behavior as the continuous approach , given by the liouville field theory .
thus it is certain that the two theories describe the same physics .
however , we feel that the mutual relationship is not yet fully understood . while the relation between dt and matrix model is transparent , its relation to the liouville field theory is not so evident .
intuitively , we expect that the simplicial manifolds generated by dt will tend to continuous manifolds in the large number limit of simplexes . however , such surfaces are known to be fractal@xcite , hence the way the lattice regularization simulates the continuous theory is not at all clear .
it is therefore indispensable to make sure that an universal continuum limit exists . in order to check this numerically
, we propose to measure how the network conducts currents , and compare it with the corresponding continuous medium .
an invariance under local scale transformation plays an essential role in the algorithm , and because of this property we can extract information about the complex structure which is independent of the conformal mode parametrised by moduli . despite many theoretical developments the precise nature of the transition at @xmath0 is still not fully understood .
serious numerical efforts have failed so far to convincingly establish the transition , at least by measuring the string susceptibility .
it should be much easier to observe this transition in local scale invariant quantities , rather than in the original metric which is comparatively complicated reflecting the fractal nature of the surface . here
, we show such a method by which we can define the complex structure of a dt surface and separate the conformal mode .
the basic idea of our method originates from the analog model @xcite of the dual amplitude in analogy with the electrical circuit@xcite .
let us first consider a two dimensional conducting medium with conductivity tensor density @xmath2 .
ohm s law reads @xmath3 where @xmath4 is an electric field and @xmath5 is a current density .
the joule heat @xmath6 can be estimated by a potential distribution @xmath7 on the surface ; @xmath8 . requiring @xmath9
, it leads to an equation for @xmath7 : @xmath10 , which is just the equation of continuity . for a homogeneous isotropic medium
the conductivity tensor density is given by @xmath11 with @xmath12 being the resistivity constant .
therefore , ohm s law([eq : ohm_law ] ) is invariant under the local scale transformation @xmath13 . in the following we first give an algorithm to measure the resistivity from voltage distribution . in the case of spherical topology
we can regard the surface essentially as an infinite flat sheet .
then the potential at a point specified by the complex coordinate , @xmath14 , with a source of current @xmath15 placed at @xmath16 , and a sink of the current at @xmath17 is given by @xmath18 in order to avoid the ambiguity arising from const .
in eq.([eq : v(z ) ] ) we measure the potential drop between coordinates @xmath19 and @xmath20 .
we define the potential drop between points @xmath21 and @xmath22 in the presence of a source and a sink of current at points @xmath23 and @xmath24 by @xmath25| , \label{eq : potential}\end{aligned}\ ] ] where @xmath26={z_i - z_k \over z_i - z_l } { z_j - z_l \over z_j - z_k}$ ] is known as the anharmonic ratio .
it has an important property : _ i.e. _ @xmath27 invariance .
then it allows us to fix three coordinates among the @xmath28 to any desired values without changing the potential drop eq.([eq : potential ] ) .
for example , we can fix three coordinates as @xmath29 , @xmath30 , and @xmath31 .
then the potential drop eq.([eq : potential ] ) for @xmath32 is written as @xmath33 @xmath34 all the other possible permutations of @xmath35 only give the potential drops obtained by linear combinations of eq.([eq : potd1 ] ) and eq.([eq : potd2 ] ) . since unknown variables are @xmath36 , we need the potential drop at an additional point @xmath37 .
then we get three more relations in addition to eq.([eq : potd1 ] ) and eq.([eq : potd2 ] ) which altogether are sufficient to specify five unknowns : @xmath38 .
the three extra equations are given by @xmath39 @xmath40 @xmath41 now we apply this to a random surface generated by dt .
since the local scale invariance is not fully realized on this surface , the resistivity may have certain fluctuations .
in fact , the value varies from one sample to another , and depending on the position of electrodes for the measurements .
such ambiguities should , however , vanish in the continuum limit , if it exists at all .
the dual graph of a surface consisting of @xmath42 triangles is regarded as a trivalent network , where we fix the resistance of the link connecting two neighbouring vertices to be @xmath43 .
what we want to do is to examine whether such a network behaves like a continuous medium in the limit of @xmath44 . for the measurement
we pick five vertices in the dual graph and perform the method as explained above .
the practical method we employ for the determination of potential drops is as follows ; we pick two vertices at @xmath45 and @xmath46 for the source and the sink of current with unit intensity . by writing the potential of the vertex
@xmath47 as @xmath48 , current conservation reads @xmath49 where @xmath50 and @xmath51 are the three neighboring vertices of @xmath47 .
we solved this set of equations by the _ successive overrelaxation _ method .
our method can be reconsidered from the string - theoretical point of view . whether string theory can be described by local fields or not is one of the fundamental questions .
the authors of @xcite showed that a sum of planar fishnet feynman diagrams with @xmath42 external particles is indeed approximated by a @xmath42 particle veneziano amplitude .
the basic assumption of this model is that the network corresponding to a fine planar fishnet feynman diagram can be regarded as a homogeneous isotropic conducting sheet with a constant resistivity .
once this is accepted it is straightforward to derive dual amplitudes by evaluating the heat generated by the sheet .
the purpose of our article is to justify this basic assumption numerically . in order to clarify the correspondence to the network of resistors ,
we write the amplitude of fishnet feynman diagrams as @xmath52 where @xmath53 , @xmath24 and ( @xmath54 ) indicate a link and its two ends(vertices ) of the network respectively , and @xmath55 here we have introduced a feynman parameter @xmath56 resistance ) , ( @xmath57 voltage at i ) , ( @xmath58 current through l ) and ( @xmath59 heat generation ) . ] for each propagator @xmath24 in the momentum integration , @xmath60 the basic conjecture is that the fishnet diagrams behaves as a continuum homogeneous conducting sheet after taking the average @xmath61
we fix the topology to be two sphere(@xmath62 ) and the number of triangles to be @xmath63 , @xmath64 and @xmath65 ( prohibiting tadpole and self - energy diagrams ) . for matter fields coupled to gravity we put @xmath66 scalar fields with @xmath67 and @xmath68 and/or ising spins on each triangle .
we pick - up five vertices randomly as five electrodes to obtain the resistivity constant as indicated in sec . 2 , and repeat this procedure @xmath69 times for each configuration , for about @xmath70 independent configurations as ensembles .
let us first take a look at the results for pure gravity ( fig.[fig : pure_resist ] ) .
we find that _ peaks get narrower as the size grows_. the value @xmath71 should be compared to @xmath72 of the flat network(@xmath73 a honeycomb lattice ) .
an increase in @xmath12 is understood to be a reflection of the fractal nature of the surface .
as the surfaces get larger , finite size effects due to the discreteness diminish and eventually the peak will grow indefinitely .
this is what we have expected as the continuum limit of a network of resistors .
we find similar tendency in the one scalar case(@xmath74)(fig.[fig : b1_resist ] ) .
also , in fig.[fig : random_resist ] , we show data to justify the assumptions of the analog model for the case of pure gravity .
each resistance is distributed randomly with weight @xmath75 .
these distributions indicate the same behavior with previous cases @xmath76 . in fig.[fig : b2_resist],[fig : b3_resist ] we plot the distributions of @xmath12 for the @xmath77 cases . here
no sharpening of peaks with increasing sizes is seen unlike the previous cases , which suggests that the surface does not approach a smooth continuum limit .
this is considered as the reflection of the tachyonic mode instability in the continuous theory .
the same tendency is also supported in the case of 3 critical ising spins .
next , we discuss how to extract information on the conformal mode .
once we have found the value of resistivity , we can assign a complex coordinate to each vertex of the dual graph from eq.([eq : potd1 ] ) and eq.([eq : potd2 ] ) . in fig.6
we plot the distribution of coordinates for all the vertices of a surface for the case of pure gravity with @xmath78 triangles .
the point density around a point @xmath14 should be proportional to @xmath79 , because each vertex is supposed to carry the same space - time volume . in the conformal gauge which we have been employing
, we have @xmath80 , where @xmath81 is the conformal mode .
this argument may be over - simplified in the sense that we did not pay attention to the quantum fluctuation of @xmath81 and the renormalization of the composite operator @xmath82 .
what we can say safely is that the point density at @xmath14 after taking an ensemble average should be equal to the four - point function @xmath83 in the liouville theory , where @xmath84 is a point on the complex plane and @xmath85 is the renormalization factor for the conformal mode . in a recent paper@xcite ,
an estimate for this four - point function has been given .
this should be compared with fig.6 .
we are grateful to n.d.hari dass and n.ishibashi for useful discussions and comments .
one of the authors(n.t . ) is supported by research fellowships of the japan society for the promotion of science for young scientists .
99 n.tsuda and t.yukawa , phys.lett .
b305 ( 1993 ) 223 ; + h.kawai , n.kawamoto , t.mogami , and y.watabiki , phys.lett . b306 ( 1993 ) 19 .
b.sakita and m.a.virasoro , phys.rev.lett . 24
( 1970 ) 1146 ; + h.b.nielsen and p.olesen , phys.lett .
b32 ( 1970 ) 203 .
see for example j.d.bjorken and s.d.drell , _ mcgraw hill , new york , 1965_. a.b.zamolodchikov and al.b.zamolodchikov hep - th/9506136 | the complex structure of a surface generated by the two - dimensional dynamical triangulation(dt ) is determined by measuring the resistivity of the surface .
it is found that surfaces coupled to matter fields have well - defined complex structures for cases when the matter central charges are less than or equal to one , while they become unstable beyond @xmath0 . a natural conjecture that fine planar random network of resistors behave as a continuous sheet of constant resistivity
is justified numerically for @xmath1 . |
in this paper we investigate the transverse energy per charged particle , @xmath0 , for beam energies ranging from about 1 agev up to 200 agev . in this energy range , @xmath0 at first increases rapidly from sis @xcite to ags @xcite , then saturates to a value of about 800 mev at sps @xcite energies and remains constant up to the highest available rhic energies @xcite .
the present analysis of @xmath0 uses the hadron resonance gas model ( thermal model ) which describes the final state in relativistic heavy - ion collisions as composed of hadrons , including heavy hadronic resonances as being in thermal and chemical equilibrium .
it has been known for many years @xcite that the chemical freeze - out can be described by the condition @xmath2 1 gev , where @xmath3 and @xmath4 are , respectively the total energy and particle number of the primordial hadronic resonances before they decay into stable hadrons .
this quantity can not be determined directly from experiment unless the final state multiplicity is low and hadronic resonances can be identified , which is not the case in relativistic heavy - ion collisions .
our analysis therefore starts by relating the number of charged particles seen in the detector to the number of primordial hadronic resonances and the transverse energy to the energy @xmath3 of primordial hadrons . in this paper
all thermal model calculations were performed using the thermus package @xcite .
the transverse energy , @xmath5 , is defined as the energy deposited transverse to the beam direction in a given interval of pseudo - rapidity @xmath6 .
the transverse energy has two components , the hadronic one , @xmath7 , and the electromagnetic one , @xmath8 , coming from the electromagnetic particles ( photons , electrons and positrons ) .
electromagnetic calorimeters are used to measure @xmath8 whereas hadronic calorimeters or the time projection chamber ( for particle identification and momentum information ) are used to measure @xmath7 .
the energy of a particle is defined as being the kinetic energy for nucleons , for anti - nucleons as the total energy plus the rest mass and for all other particles as the total energy @xcite . in the experiments the transverse energy and the charged particle multiplicity
are measured in a similar way so that most of the systematic uncertainties cancel out in the ratio .
experiments have reported a constant value of the ratio @xmath9 gev from sps to rhic @xcite , with the ratio being almost independent of centrality of the collision for all measurements at different energies . in all cases
the value of @xmath0 has been taken for the most central collisions , at the end of this paper we consider the centraly dependence of @xmath0 .
when this ratio is observed for the full range of center of mass energies , it shows two regions @xcite . in the first region from lowest @xmath10 to sps energy , there is a steep increase of the @xmath0 ratio with @xmath10 . in this regime ,
the increase of @xmath10 causes an increase in the @xmath11 of the produced particles . in the second region , sps to higher energies , the @xmath0 ratio is very weakly dependent on @xmath10 .
the energy pumped into the system by the increase of @xmath10 is converted mainly into particle production .
this observation is quite remarkable and requires the help of models for a better understanding of the underlying physics . to estimate @xmath0 in the thermal model we relate the number of charged particles , @xmath12 , to the number , @xmath4 , of primordial hadrons . to estimate the charged particle multiplicity at different center of mass energies from the thermal model , we proceed as follows .
first we study the variation of the ratio of the total particle multiplicity in the final state , @xmath13 , and that in the primordial i.e. @xmath14 with @xmath10 .
this ratio starts from one , since there are only very few resonances produced at low beam energy and becomes almost independent of energy after sps energy .
the value of @xmath14 in the region where it is independent of @xmath10 is around 1.7 .
the excitation function of @xmath14 is shown in fig .
[ nch](a ) .
secondly , we have studied the variation of the ratio of charge particle multiplicity and the particle multiplicity in the final state ( @xmath15 ) with @xmath10 .
this is shown in fig . [ nch](b ) .
the @xmath15 ratio starts around 0.4 at lower @xmath10 and shows an energy independence at sps and higher energies . at lower sis energy , the baryon dominance at
mid - rapidity makes @xmath16 which has a value of 0.45 for au - au collisons ( left ) and @xmath15 ( right ) with @xmath10 in the thermal model with @xmath17 = 1.08 gev as freeze - out criterium .
, width=528 ] as the next step we connect the transverse energy @xmath5 to the the energy of the primordial hadrons @xmath3 .
we start by relating the two quantities for a static fireball . in this case one
has @xmath18 where @xmath19 is the statistical distribution factor , e.g. for a boltzmann distribution it is given by @xmath20 .
it is straightforward to re - write this expression as @xmath21 thus , for a static fireball , the transverse energy is related to the total energy by a simple factor of @xmath22 . in the hadronic resonance gas model
there is a sum over all hadrons ; furthermore , taking into account the experimental configuration which leads to adding the mass of the nucleon for anti - nucleons and subtracting the same for nucleons one has @xmath23 .\end{aligned}\ ] ] the above equation relates the transverse energy measured from the data and that estimated from the thermal model . in the limit of large beam energies one has @xmath24 which is close to the value measured at rhic .
the measured @xmath5 will be affected by the radial flow and by the difference between chemical freeze - out and kinetic freeze - out temperatures ; these effects will lead to corrections which tend to largely cancel each other .
a detailed comparison of this , in the framework of a single freeze - out temperature model and limited to rhic energies , has been made in ref .
@xcite . in fig .
[ etncht ] we plot lines of constant @xmath0 in the @xmath25-diagram . for low values of @xmath0 ,
these lines are almost indepent of @xmath26 , this is mainly a consequence of subtracting @xmath27 in the definition of @xmath5 , thus taking away much of the influence of nucleons . only towards larger values of @xmath26
there is a notable dependence on this variable . to compare with the chemical freeze - out condition ,
we show also the chemical freeze - out curve in the same plane ( fig .
[ etncht ] ) . at higher energies , when @xmath26 nearly goes to zero
, the transverse energy production is mainly due to the meson content in the matter .
the intersection points of lines of constant @xmath0 and the freeze - out line give the values of @xmath0 at the chemical freeze - out .
hence at freeze - out , given the values of @xmath0 from the experimental measurements we can determine @xmath28 and @xmath26 of the system . from thermal model without flow are shown in the @xmath25 plane .
the chemical freeze - out condition of @xmath29 gev is also shown.,width=384 ] in fig .
[ etnchtmub ] , we plot the ratio @xmath0 as a function of the temperature @xmath28 and as a function of @xmath26 .
it can be seen that the relation between the @xmath0 and @xmath28 is linear to a good approximation , similarly for the relation with @xmath26 . with @xmath28 ( left ) and
the variation of @xmath0 with @xmath26 ( right),width=528 ] for the most central collisions , the variation of @xmath0 with center of mass energy is shown in fig .
[ etnchcm ] .
the data have been taken from ref .
@xcite , and are compared with the corresponding calculation from the thermal model with chemical freeze - out .
we have checked explicitly that other freeze - out criteria discussed in the literature give almost identical results for the behavior of @xmath0 as a function of @xmath10 ; this is the case for the fixed baryon plus anti - baryon density condition @xcite and also for fixed normalised entropy density condition , @xmath30 = 7 @xcite . with @xmath10 and the thermal model using @xmath29 gev as the freeze - out condition.,width=528 ]
it has been observed from sps to rhic @xcite , that the ratio @xmath0 , is almost independent of the centrality of the collisions which is represented by the number of participant nucleons . to understand the variation of @xmath0 with collision centrality
, we have estimated @xmath0 for 130 gev au+au collisions at rhic , for different centrality classes @xcite .
this is compared with the corresponding data in fig .
[ etnchnp ] .
the effect of flow is not taken into account in the model calculations for the centrality behavior .
the model agrees well with the experimental data for the centrality behavior .
again , we have checked explicitly that other freeze - out criteria lead to similar results @xcite .
with @xmath31 for 130 gev au+au collisions at rhic with the corresponding thermal model estimate.,width=528 ]
in conclusion , we have discussed the connection between @xmath0 and the ratio of primordial energy to primordial particle multiplicity , @xmath17 , from the thermal model .
this model , when combined with chemical freeze - out criteria explains the data over all available measurements for the @xmath10 behavior of @xmath0 .
it has to be noted that variables like @xmath0 , the chemical freeze - out temperature @xmath32 , @xmath33 and @xmath15 discussed in this paper , show saturation starting at sps and continuing to higher center of mass energies .
this observation along with the centrality independence of @xmath0 is not inconsistent with the simultaneity of chemical and kinetic freeze - out at higher energies @xcite .
three of us ( jc , rs , dks ) would like to acknowledge the financial support of the south africa - india science and technology agreement .
90 w. reisdorf _ et al_. , fopi collaboration , ( 1997 ) 493 .
_ et al_. , fopi collaboration , ( 1997 ) 215 .
_ et al_. , fopi collaboration , ( 2002 ) 034901 .
_ et al_. , e802 collaboration , ( 1999 ) 2173 .
+ l. ahle _
et al_. , e802 collaboration , ( 1998 ) 3523 .
ogilvie for e866 and e917 collaborations , ( 1998 ) 57c .
+ y. akiba for e802 collaboration , ( 1996 ) 139 .
+ b.b . back
_ et al_. , e917 collaboration , ( 2001 ) 1970 .
_ et al_. , e802 collaboration , ( 1998 ) 466(r ) . + l. ahle
_ et al_. , e802 collaboration , ( 2000 ) 1 .
_ et al_. , e802 collaboration , ( 2000 ) 53 .
_ et al_. , e814/877 collaboration , ( 1993 ) 2996 .
m.m . aggarwal _
et al_. , wa98 collaboration , ( 2001 ) 651 .
_ et al_. , na49 collaboration , ( 1995 ) 3814 .
j. bachler for the na49 collaboration , ( 1999 ) 45 .
+ m. van leeuwen for the na49 collaboration , ( 2003 ) 161 .
+ na49 collaboration , s.v .
et al_. , ( 2002 ) 054902 .
j. adams _ et al_. , star collaboration , ( 2004 ) 054907 . k. adcox
_ et al_. , phenix collaboration , ( 2001 ) 052301 .
et al_. , phenix collaboration , ( 2005 ) 034908 .
j. cleymans and k. redlich , ( 1998 ) 5284 ; + phys .
c * 60 * ( 1999 ) 054908 .
j. cleymans , h. oeschler , k. redlich , s. wheaton , ( 2006 ) 034905 . s. wheaton and j. cleymans , hep - ph/0407174 ; + journal of physics g31 ( 2005 ) s1069 .
d. prorok , 75 ( 2007 014903 . t. akesson _
et al_. , helios collaboration , ( 1988 ) 383 .
p. braun - munzinger and j. stachel , j. phys .
g : nucl . part .
* 28 * ( 2002 ) 1971 .
j. cleymans , m. stankiewicz , p. steinberg and s. wheaton , nucl - th/0506027 .
a. tawfik , _ j. phys .
g : nucl . part .
phys . _ * 31 * s1105 ( 2005 ) .
a. tawfik , hep - ph/0507252 and hep - ph/050824 .
j. cleymans , b. kmpfer , m. kaneta , s. wheaton , n. xu , ( 2005 ) 054901 . w. broniowski and w. florkowski , ( 2001 ) 272302 ; + ( 2002 ) 064905 , nucl - th/0204025 . | in relativistic nucleus - nucleus collisions the transverse energy per charged particle , @xmath0 , increases rapidly with beam energy and remains approximately constant at about 800 mev for beam energies from sps to rhic .
it is shown that the hadron resonance gas model describes the energy dependence , as well as the lack of centrality dependence , qualitatively .
the values of @xmath0 are related to the chemical freeze - out criterium @xmath1 gev valid for primordial hadrons . , , , , |
the most general problem setting of the wave turbulence theory can be regarded in the form of a nonlinear partial differential equation @xmath2 where @xmath3 and @xmath4 denote linear and nonlinear part of the equation correspondingly , the linear part has the standard wave solutions of the form @xmath5},\ ] ] where the amplitude @xmath6 may depend on space variables but not on time , and a small parameter @xmath7 shows that only resonant wave interactions are taken into account
. the dispersion function @xmath8 can be easily found by substitution of @xmath9 into the linear part of the initial pde , @xmath10 while @xmath11 and resonance conditions have the form @xmath12 for @xmath13 interacting waves with wave - vectors @xmath14 . for most physical applications
it is enough to regard @xmath15 or @xmath16 , and the most common form of dispersion function is @xmath17 ( for instance , capillary , gravitational and surface water waves , planetary waves in the ocean , drift waves in tokamak plasma , etc . )
+ the model of laminated wave turbulence@xcite describes two co - existing layers of turbulence - continuous and discrete - which are presented by real and integer solutions of sys.([open ] ) correspondingly .
the continuous layer is described by classical statistical methods@xcite while for the discrete layer new algorithms have to be developed .
it was shown in@xcite that an arbitrary integer lattice @xmath18 , each node of the lattice denoting a wave - vector @xmath19 , can be divided into some clusters ( classes ) and there are two types of solutions of sys.([open ] ) : those belonging to the same class and those belonging to different classes .
mathematically , a class is described as a set of wave - vectors for which the values of the dispersion function have the same irrationality .
for instance , if the dispersion function has the form @xmath20 , then a class is described as follows : @xmath21 where @xmath22 is a natural number and @xmath23 is a square - free integer .
physically , it means that waves are interacting over the scales , that is , each two interacting waves generate a wave with a wavelength different from the wave lengths of the two initial waves .
interactions between the waves of different classes do not generate new wavelengths but new phases . + in our preceding paper@xcite we presented a generic algorithm for computing all integer solutions of sys.([open ] ) within one class .
four - wave interactions among 2-dimensional gravitational water waves were taken as the main example , in this case sys.([open ] ) takes form:[prosetdef2 ] ( m_1 ^ 2+n_1 ^ 2)^1/4 + ( m_2 ^ 2+n_2 ^ 2)^1/4=(m_3 ^ 2+n_3 ^ 2)^1/4+(m_4 ^ 2+n_4 ^ 2)^1/4 + m_1+m_2=m_3+m_4 + n_1+n_2=n_3+n_4 + and classes are defined as @xmath24 , where @xmath23 , called class index , are all natural numbers containing every prime factor in degree smaller @xmath25 and @xmath26 , called weight , all natural numbers .
it can be proven that if all 4 wave - vectors constructing a solution of sys.([prosetdef2 ] ) * do not * belong to the same class , then the only possible situation is following : all the vectors belong to two different classes @xmath27 and the first equation of sys.([prosetdef2 ] ) can be rewritten then as[th2eq3 ] _ 1 + _
2=_1+_2 with some @xmath28 and @xmath29 being class indexes . in the present paper
we deal with this two - class case .
as in the previous paper @xcite , we are going to find all solutions of eq.([prosetdef2 ] ) in some finite domain @xmath30 , i.e. @xmath31 for some @xmath32 .
the first case has been studied for @xmath33 , where @xmath34 classes have been encountered .
the straightforward approach , not making use of classes , consumes , as for the first case , at least @xmath35 operations and is out of question ( see @xcite , sec.3.2.1 for discussion of this point ) .
+ straightforward application of classes also does not bring much .
the eq.([th2eq3 ] ) is now trivial - but classes are interlocked through linear conditions .
even if for each pair of classes we could detect interlocking and find solutions , if any , in @xmath36 operations ( which is probably the case , though we did not prove it ) , the overall computational complexity is at least @xmath37 - i.e. not much less than @xmath38 . for @xmath33
this implies @xmath39 pairwise class matches which is outside any reasonable computational complexity limits . + the trouble with this approach - as , for that matter , with virtually any algorithm consuming much more computation time than the volume of its input and output data implies - is , that we perform a lot of intermediary calculations , later discarded .
we develop an algorithm performing every calculation with a given item of input data just once ( or a constant number of times ) .
first of all we notice that eq.([th2eq3 ] ) can be rewritten as [ interl1 ] ( m_1l^2+n_1l^2)^1/4 = ( m_1r^2+n_1r^2)^1/4 = _ 1 + ( m_2l^2+n_2l^2)^1/4 = ( m_2r^2+n_2r^2)^1/4 = _ 2 + m_1l - m_1r = -m_2l + m_2r + n_1l - n_1r = -n_2l + n_2r + where @xmath29 are two different class indexes and @xmath40 - the corresponding weights . [
[ definition . ] ] definition .
+ + + + + + + + + + + for any two decompositions of a number @xmath41 into a sum of two squares ( see ( [ interl1 ] ) ) the value @xmath42 is called @xmath43-_deficiency _ , @xmath44 is called @xmath13-_deficiency _ and @xmath45 - _ deficiency point_. + we immediately see that for two interacting waves their deficiencies must be equal : @xmath46 . for a given weight @xmath26 , every two decompositions of @xmath47 into a sum of two squares yield , in general , four deficiency points with @xmath48 .
consider unsigned decompositions @xmath49 .
assuming @xmath50 the four points are @xmath51 and four ( symmetrical ) points in each of the other three quadrants of the @xmath52 plane .
[ [ definition.-1 ] ] definition .
+ + + + + + + + + + + the set of all deficiency points of a class for a given weight , @xmath53 , is called its _ @xmath26-deficiency set_. the set of all deficiency points of a class , @xmath54 , is called its _ deficiency set_. + the objects defined above
play the main role in our algorithm , so we compute as an illustrative example the for the number 50 .
the number @xmath55 has three decompositions into sum of two squares , namely , @xmath56 , and nonnegative deficiency points of decomposition pairs are @xmath57 they constitute a subset of the deficiency set @xmath58 , namely , the 12 points with @xmath59 . in each of three other quadrants of the @xmath60 plane there
lie 36 more points of this set , symmetrical to the ones shown with respect to the coordinate axes .
+ the crucial idea behind the algorithm of this paper is very simple and follows immediately from the exposition above : + * sys.([interl1 ] ) has a solution with vectors belonging to the two different classes @xmath61 if and only if their deficiency sets have a non - void intersection , @xmath62 , i.e. some elements belong to both classes . *
calculation of relevant class indexes @xmath23 by a sieve - like procedure , admissible weights @xmath26 and decomposition of @xmath47 into sum of two squares have all been treated in full detail in @xcite .
one new feature we introduced here is , that immediately after generating the array of class bases @xmath23 we purge away those which , whatever the admissible weight @xmath26 , do not have a decomposition into a a sum of two squares @xmath63 with both @xmath64 . for the problem considered in @xcite
this would be superfluous because virtually all these classes have been anyhow filtered away according to another criterium ( @xmath65 ) which does not apply here . in this way
we exclude 100562 classes from the 384145 which the sieving procedure returns . + evidently for any deficiency point @xmath66 inequalities
@xmath67 hold . and
if deficiency sets of two classes have a non - void intersection , they also have an intersection over points with non - negative @xmath68 .
so we start with declaring a two - dimensional array @xmath69 of type byte which serves storing and processing deficiency sets of the classes .
the array is initialized with all zeroes . in the first pass for every class @xmath23 in the main domain @xmath30 we generate its deficiency set @xmath70 .
notice that after generating deficiency set of the class for each weight @xmath26 and uniting them we must check for doubles and eventually get rid of them .
next , for every deficiency point @xmath71 of the class we increment the value of the corresponding element of the array by 1 , except elements with value 255 whose values are not changed . in the second pass
we generate deficiency sets once more and for every point of the deficiency set of a class check the values of the corresponding point of @xmath72 .
if all these values are equal to @xmath73 , no waves of the class participate in resonant interactions and the class is discarded from further considerations . + for the problem considered this pass excludes just a few ( @xmath74 ) classes , so the time gain is very modest .
however , we include this step into the presentation for two reasons .
first of all , it _ had _ to be done as no possibility of reducing the number of classes considered as much as possible and as soon as possible ( before the most time - consuming steps ) may be neglected .
second , though giving not much gain for solution of the problem at hand , this elimination techniques may play a major role in further applications of our algorithm . in the third pass
we generate a more detailed deficiency set for each class , i.e. for all classes not discarded in the previous pass : for every deficiency point @xmath66 we store @xmath75 .
we do * not * discard duplicates as we did in the previous two passes .
then we revisit the corresponding points of @xmath72 and to each point whose value is larger than @xmath73 link the structure @xmath76 described above . in the fourth pass we go through the array @xmath72 once more and store every point with value greater than one in an array @xmath77 .
we also relink structures linked to deficiency points to corresponding points of the new array .
the four passes above leave us with an array of points @xmath66 and to each of these points a list of structures @xmath78 is linked ( no less than two different @xmath79 ) . in general , a linked list is here most appropriate . every combination of two structures linked to the same point and having different @xmath79 yields a solution of sys.([prosetdef2 ] ) . from every solution found ,
we obtain four solutions changing signs of @xmath80 in the general case , i.e. for @xmath80 nonzero .
+ notice that theoretically we could skip pass 4 and extract solutions directly from the array @xmath72
. however , this is not reasonable for implementation reasons , and pass 4 is not very time - consuming . implementing the algorithm described above
, we took a few language - specific shortcuts that will be briefly described here .
+ passes 1 and 2 have been implemented one - to - one as described above .
however , manipulating linked lists in vba involves considerable overhead and for the problem considered in this paper we do not need the complete functionality of linked lists , i.e. inserting into / deleting from intermediate positions of the list .
our main data structure for pass 3 - 5 is a simple two - dimensional array arsolhalves(@xmath81 ) and in a single line of this array we store : * the class base @xmath23 ; * the coordinates of deficiency point @xmath82 ; * the coordinates of two wave vectors belonging to this deficiency point ; * the index in the array arsolhalves of the next line belonging to the same deficiency point . which is demonstrated in fig.[f : arrlista ] below . here , @xmath83 is the number of @xmath84-vectors linked to all deficiency points to which vectors belonging to two or more classes belong ( @xmath85 for @xmath33 ) .
we generate the deficiency set of each class and fill all members of this line of the array except the last one in the process , deficiency point by deficiency point .
the last member is filled later and in the following way .
+ for this pass we also declare an auxiliary array ardeficiencesprev(1 .. @xmath86 , 1 ..
@xmath86 ) initialized with zeroes .
having added a new line @xmath87 to arsolhalves , we look up the value @xmath88of ardeficiencesprev(@xmath82 ) . if it is zero ( this deficiency point being visited the first time ) we just assign this point the value of the index of the new line in the array arsolhalves .
otherwise we first assign arsolhalves(@xmath88 , 7 ) the value of the current line s index in arsolhalves , then write this number to ardeficiencesprev(@xmath82 ) ( see fig.[f : arrlistb ] ) .
+ a numerical example for this procedure is given in table 1 . in this way
, the array index of the next `` list '' member is stored with the previous one , except evidently the last one , where the corresponding field stays zero . [ cols="^,^,^,^,^,^,^,^,^",options="header " , ] + table 1 .
a few lines of the table containing solution halves for the deficiency point @xmath89 ( beginning and end of the sequence ) .
+ consider computational complexity of these steps .
for a single class index @xmath23 and weight @xmath26 , generating deficiency points in the first step consumes @xmath90 operations because every number @xmath91 has no more than @xmath92 decompositions into two squares which we combine pairwise to find deficiency points .
decompositions themselves can be found in @xmath93 time@xcite .
there are @xmath94 admissible weights to class index @xmath23 , so the overall complexity for a class can be estimated from above as @xmath95 . merging deficiency points into @xmath54 can be done in @xmath96 time for number of points x , i.e. no more than @xmath97 + taking a rough upper estimate for the number of classes @xmath98 , we obtain an estimate @xmath99 . incrementing the points of @xmath72
is linear on the point number of the set @xmath54 and need not be considered for computational complexity separately .
the same complexity estimate holds for the second pass .
notice that , having enough memory , or using partial data loading / unloading similar to that used in @xcite , we could preserve deficiency sets calculated on the first pass and not recalculate them here .
however , this would not significally improve the overall computational complexity of the algorithm .
+ we can not give an _ a priori _ estimate for the number of classes discarded at the second pass , so we ignore it and hold the initial rough upper estimate @xmath98 for the number of classes in our further considerations . in the third pass , to every point @xmath100 ( no more than @xmath101 of them )
we link the values @xmath102 for which this point has been struck .
this , as well as linking to the points of @xmath72 is , clearly , linear on the number of points and does not raise the computational complexity .
complexity of the fourth pass can be estimated as follows .
suppose the worst case , i.e. no classes are discarded at step 2 and every deficiency point is a solution point , i.e. for every @xmath103 no less than two classes have deficiency points with the same @xmath82 .
then we must make no more than @xmath104 entries into the new array @xmath105 .
we must relink no more than the mean of @xmath106 structures per point , which gives an upper estimate of @xmath107 time for the pass .
however , remember that the estimate for the deficiency point number has been made on the assumption that all @xmath108 generate distinct deficiency points . in simple words , for every point linked to @xmath109 structures we obtain @xmath110 less solution points . now
elementary consideration allow us to improve the estimate to @xmath104 time . we did not manage to obtain a reasonable estimate for the computational complexity of the fifth step .
for the worst case of all structures grouped at a single point , the estimate is @xmath111 - but this is not realistic .
if the number of solution points is @xmath104 and the number of linked deficiencies is bounded by some number @xmath112 , then we can make an estimate @xmath113 .
this , however , is also not quite the case as our numerical simulations show ) .
however , this last step deals with solution extraction and extracts them in linear time per solution .
_ any _ algorithm solving the problem has to extract solutions , so we can be sure that our step 5 is optimal - even without any estimate of its computational complexity . summing up , we obtain the overall upper estimate of computational complexity @xmath99 reached at steps @xmath73 and @xmath114 plus the time needed for solution extraction .
our algorithm has been implemented in the vba programming language ; computation time ( without disk output of solutions found ) on a low - end pc ( 800 mhz pentium iii , 512 mb ram ) is about 10 minutes .
some overall numerical data is given in the two figures below .
the number of solutions for the 2-class - case depending on the partial domain is shown in the fig.[f : solsqsir ] .
both curves are almost ideal cubic lines .
very probably they _ are _ cubic lines asymptotically - the question is presently under study .
+ partial domains chosen in fig.[f : solsqsir ] are of two types : squares @xmath115 , just for simplicity of computations , and circles @xmath116 , more reasonable choice from physical point of view ( in each circle all the wave lengths are @xmath117 ) .
the curves in the fig.[f : solsqsir ] are very close to each other in the domain @xmath118 though number of integer nodes in a corresponding square is @xmath119 and in a circle with radius @xmath30 there are only @xmath120 integer nodes .
this indicates a very interesting physical phenomenon : most part of the solutions is constructed with the wave vectors parallel and close to either axe @xmath91 or axe @xmath121 .
+ on the other hand , the number of solutions in rings @xmath122 ( corresponds to the wavelengths between @xmath123 and @xmath30 ) grows nearly perfectly linearly .
of course the number of solutions in a circle is _ not _ equal to the sum of solutions in its rings : a solution lies in some ring if and only if all its four vectors lie in that same ring .
that is , studying solutions in the rings only , one excludes automatically a lot of solutions containing vectors with substantially different wave lengths simultaneously , for example , with wave vectors from the rings @xmath123 and @xmath124 .
this `` cut '' sets of solutions can be of use for interpreting of the results of laboratory experiments performed for investigation of waves with some given wave lengths ( or frequencies ) only .
another important characteristic of the structure of the solution set is multiplicity of a vector which describes how many times a given vector is a part of some solution .
the multiplicity histogram is shown in fig.[f : vecmul222 ] .
on the axis @xmath91 the multiplicity of a vector is shown and on the axis @xmath121 the number of vectors with a given multiplicity .
the histogram of multiplicities is presented in fig .
[ f : vecmul222 ] , it has been cut off - multiplicities go very high , indeed the vector ( 1000,1000 ) takes part in 11075 solutions .
+ similar histograms computed for different 1-class - cases show that most part of the vectors , @xmath125 for different types of waves , take part in one solution , e.g. they have multiplicity 1 .
this means that triads or quartets are , so to say , the `` primary elements '' of a wave system and we can explain its most important energetic characteristics in terms of these primary elements .
the number of vectors with larger multiplicities decreases exponentially when multiplicity is growing .
the very interesting fact in the 2-class - case is the existence of some initial interval of small multiplicities , from 1 to 10 , with very small number of corresponding vectors .
for instance , there exist only 7 vectors with multiplicity 2 . beginning with multiplicity 11 ,
the histogram is similar to that in the 1-class - case . +
this form of the histogram is quite unexpected and demonstrates once more the specifics of the 2-class - case compared to the 1-class - case . as one can see from the multiplicity diagram in fig .
[ f : vecmul222 ] , the major part in 2-class - case is played by much larger groups of waves with the number of elements being of order 40 : each solution consists of 4 vectors , groups contain at least one vector with multiplicity 11 though some of them can take part in the same solution .
this sort of primary elements can be a manifestation of a very interesting physical phenomenon which should be investigated later : triads and quartets as primary elements demonstrate periodic behavior and therefore the whole wave system can be regarded as a quasi - periodic one . on the other hand ,
larger groups of waves may have chaotic behavior and , being primary elements , define quite different way of energy transfer through the whole wave spectrum . | the model of laminated wave turbulence puts forth a novel computational problem - construction of fast algorithms for finding exact solutions of diophantine equations in integers of order @xmath0 and more .
the equations to be solved in integers are resonant conditions for nonlinearly interacting waves and their form is defined by the wave dispersion .
it is established that for the most common dispersion as an arbitrary function of a wave - vector length two different generic algorithms are necessary : ( 1 ) one - class - case algorithm for waves interacting through scales , and ( 2 ) two - class - case algorithm for waves interacting through phases . in our previous paper
we described the one - class - case generic algorithm and in our present paper we present the two - class - case generic algorithm .
pacs numbers : 47.10.-g , 47.27.de , 47.27.t , 02.60.pn = msbm10 @xmath1 ) |
the accretion disc in black hole binary systems ( bhb ) is unstable at the point where hydrogen goes from being predominantly neutral to ionised .
this gives rise to dramatic , transient outbursts in which the reservoir of material built up in the outer disc during quiescence can accrete down onto the black hole .
the peak luminosity of the outburst is set by the mass in the quiescent disc and the time taken for this to accrete .
both these depend on the size of the disc taking part in the outburst , which is set by the binary separation .
hence there is a link between orbital period and peak luminosity , with short period systems showing smaller peak outburst luminosities ( @xcite ; @xcite ; @xcite ; @xcite hereafter dgk07 ) .
the peak luminosity determines the number of spectral states which are seen during the outburst .
systems with peak luminosity below @xmath1 remain in the low / hard state ( lhs ) , where the energy output peaks at @xmath2 100 kev .
brighter systems instead show a distinct transition to a much softer thermal dominated state ( tds , also termed a high / soft state ) which peaks at @xmath2 1 kev , while the brightest systems can also show a very high state ( vhs , alternatively steep power law state ) .
these dramatic changes in the spectrum are correlated with equally dramatic changes in the rapid variability properties ( see e.g. @xcite ) and the jet @xcite , implying distinct changes in the nature and geometry of the accretion flow .
these can be plausibly explained if there is a hot , optically thin , geometrically thick solution to the accretion flow equations at low luminosities ( @xcite ; @xcite).this can be put together with a cool , optically thick , geometrically thin disc @xcite in the truncated disc / hot flow model where the disc progressively replaces more of the inner hot flow as the mass accretion rate increases .
the distinct hard / soft spectral transition then marks the point at which the cool disc extends down to the last stable orbit , replacing all the hot flow ( @xcite ; dgk07 ) .
such models make a clear prediction that the inner radius of the cool disc should recede as the lhs drops in luminosity .
this can in principle be tested using the colour temperature and luminosity of the cool disc component to estimate its radius ( e.g. @xcite ; @xcite ) . however , such observations are complicated by the fact that the disc is at low temperature at low luminosities , @xmath3 kev for a @xmath4 black hole at @xmath5 , even if the disc extends down to @xmath6 ( where @xmath7 ) .
such low energy x - rays can not be seen with the 3 kev bandpass limit of _ rxte _ , the satellite which has accumulated the most bhb data to date ( @xcite ; @xcite ) .
ccd observations can extend the bandpass down to lower energies , but these are still not straightforward to interpret as the disc is not the dominant spectral component in this state .
instead , the hard x - ray emission dominates the spectrum , so irradiation can change the disc temperature ( @xcite , hereafter gdp08 ) , while the disc luminosity can be underestimated due to photons which are compton upscattered into the hard x - ray spectrum @xcite .
this component is still not easy to study even with ccd detectors since low energy x - rays are absorbed by interstellar gas .
most bhb are in the galactic plane so have gas columns of @xmath8 @xmath9 which effectively block emission below 1 kev . however , there are a few black holes which have intrinsically much lower columns which have been observed with ccd detectors , namely xte j1118 + 480 : @xmath10 @xmath9 , xte j1817 - 330 : @xmath11 @xmath9 and swift1753.5 - 0127 : @xmath12 @xmath9 @xcite .
xte j1118 + 480 has only snapshot spectra available ( @xcite ; @xcite ; @xcite ; 2003 ) , but both swift1753.5 - 0127 and xte j1817 - 330 were well sampled throughout their outbursts by the _
swift _ satellite .
however , xte j1817 - 330 is a long period system , so spends most of its time in the disc dominated state , with only two , rather faint , lhs spectra at the end of its outburst . thus the short period lhs outburst of swift1753.5 - 0127 is the best candidate to study the disc evolution in the lhs .
here we analyse these data with sophisticated models to describe the behaviour of the x - ray irradiated disc ( gdp08 ; @xcite , hereafter gdp09 ) to constrain its inner radius as the outburst declines .
= 8.5 cm = 8.5 cm [ cols="^,^,^,^,^,^,^,^,^,^,>",options="header " , ] we use publicly available _ swift _ data from swift j1753.5 - 0127 taken during the period from july 2005 to july 2007 .
we extracted both x - ray telescope ( xrt ) and uvot data .
all of our xrt data was in windowed timing mode , and we extracted source counts using a circle with radius of 20 pixels .
background was taken from an off source region using a circle of the same size . however , the source was very bright during the peak of the outburst , and some of the datasets were piled up as the count rate was above 100 counts @xmath13 .
we determined the size of the central piled up region by excluding progressively larger radii regions until the spectra stopped softening at high energies .
this gave an exclusion region of 3 pixels for brightest observations , and 2 pixels for more moderately piled up data .
the data was grouped to 20 counts @xmath13 , and fit between 0.5 - 10 kev . the evolution of the observed ( i.e. not corrected for interstellar absorption ) flux in the x - ray ( upper ) and uv ( lower ) bands is plotted in fig .
we derive the flux by integrating the absorbed model fit over 0.5 - 10 kev for the x - rays . for uv data ,
we find energy band for each filter and define the effective bandpass to be between the energies determined by fwhm ( @xmath14 to @xmath15 kev for uvw2 filter , @xmath16 to @xmath17 kev for uvw1 filter , and @xmath18 to @xmath19 kev for b filter ) .
the observations fall into 3 clearly distinct time segments .
the first covers the outburst rise and decline , after which the flux remains fairly constant . hence in this paper
we concentrate on this first data group , in order to track the evolution of the disc during the flux decline .
the _ rxte _ satellite also followed the 2005 outburst .
we use the standard extraction techniques with the bright source background to derive pca spectra from all layers of detector 2 , adding 1 percent systematic error and use these data from 3 - 16 kev .
we also extract standard 1 power spectra over the full 2 - 60 kev bandpass for intervals of 256 s , and fit these with multiple lorentzian components in order to determine the qpo frequency .
these are given in table [ bigtable ] .
we extract hexte data from cluster 0 and use these over the energy range 25 - 200 kev .
we select the subset of _ swift _ and _ rxte _ data which are taken within one day of each other .
this gives a total of 27 observations which have quasi - simultaneous coverage of the optical / uv / x - ray / hard x - ray spectrum .
table [ bigtable ] details these composite spectra .
= 8.5 cm we first concentrate on the x - ray data alone , and show the different components present in the spectrum . we fit a power law to the continuum dominated bands at 24 kev and 710 kev , and then show a ratio of the data to this model over the entire bandpass ( assuming absorption fixed at the galactic column of @xmath20 @xmath9 ) . fig .
[ residual ] shows this at different points along the outburst ( marked by arrows in fig .
1 ) , from the peak to the end of the decline . close to the peak
, the continuum clearly has a rollover at high energies , so is most likely due to thermal comptonisation .
there is also clearly an additional soft x - ray component , which decreases in strength as the outburst declines .
however , there are also more subtle features seen in the high signal - to - noise pca data , with a small excess around the iron line , followed by a slight dip and subsequent rise to 20 kev .
this is characteristic of reflection thus we use a model including a disc to produce the soft x - ray component , thermal comptonisation of these disc seed photons to produce the hard continuum , and reflection of the hard continuum from the surface of the disc .
we use the thcomp model to describe the comptonisation , developed and tested by @xcite , extended to include its reflected continuum and self - consistent line emission by @xcite .
we assume that the reflecting material is neutral , with solar abundances and inclined at @xmath21 .
we fix the galactic absorption column at @xmath22 @xmath9 corresponding to an @xmath23 @xcite .
the results for each spectrum are detailed in table 1 .
we assess the significance of the detected disc emission by removing the disc from the model , and fixing the seed photon temperature instead at 0.1 ev .
this results in a loss of 2 degrees of freedom , so the disc is only significant at greater than 99 per cent confidence for @xmath24 .
thus all the spectra apart from the final 3 ( 26 , 30 and 31 ) strongly require the presence of an additional soft component .
we similarly assess the significance of reflection ( driven mostly by the presence of the iron line ) and find that it is required in all the datasets apart from the final 3 . fig .
[ trend ] shows the derived disc temperature and radius ( see table [ bigtable ] ) . the disc radius _ decreases _ as the outburst progresses ( see also @xcite ) , in apparent conflict with the truncated disc model where the inner disc should progressively recede during the lhs decline . yet the amount of reflection decreases , as expected if the disc is receding .
additionally , the low frequency qpo seen in the variability power spectrum drops dramatically , again as expected if the truncation radius sets the qpo frequency and is increasing ( dkg07 ) .
thus while the behaviour of the reflected emission and qpo fit the truncated disc models , the behaviour of the soft x - ray component does not .
we show these x - ray model fits extrapolated down to the optical / uv data and corrected for the effects of interstellar absorption .
[ simple]a shows this for the outburst peak ( 10 ) .
this clearly shows that extrapolating the hard x - ray power law spectrum down in energy _
overpredicts _ the optical / uv data .
this implies that the hard x - ray spectrum must break at uv / soft x - ray energies , supporting our modeling of it as due to comptonisation of seed photons from the accretion disc .
however , extrapolating the soft x - ray disc emission _ underpredicts _
this emission , making it clear that intrinsic gravitational energy release in the disc is insufficient to produce the observed optical / uv data .
in general , the optical can include additional contributions from reprocessed emission from hard x - ray illumination of the outer disc , as well as contributions from the jet and the companion star(e.g .
@xcite ) . however , swift j1753.5 - 0127 is a short period , low mass x - ray binary , so the companion star must faint .
instead we try to quantify the jet contribution from observed radio flux of @xmath25 jy at 1.7 ghz ( @xcite .
we show the extrapolation of this up through our optical / uv / x - ray bandpass assuming a flat spectrum ( black line on fig . [ simple]a ) .
this is a factor @xmath26 below the observed optical flux , but this is the maximum possible jet contribution as its spectrum should break where it becomes optically thin .
it seems most likely that this break is at ir frequencies ( @xcite ; @xcite ) so the jet is also probably negligible in the optical / uv for these data .
thus x - ray reprocessing from hard x - ray illumination of the outer disc seems the most likely origin for the optical / uv flux @xcite .
= 8.5 cm there are additional radio observations from midway down the outburst decline @xcite .
[ simple]b shows this radio emission , extrapolated as above , onto spectrum 20 , the dataset closest in time to the radio observations .
again , the optical / uv points are not fit by either the hard x - ray power law extrapolated down ( though this time the mismatch is not so large ) , or by the radio emission extrapolated up , or by the disc inferred from the soft x - ray component , as also shown in @xcite .
[ simple]c shows a spectrum from the end of the outburst ( 30 ) , where there are no radio observations .
the disc emission fit to the soft x - ray component is now very weak and cool , and its extrapolation down to the optical lies dramatically below the observed optical / uv emission .
however , the optical / uv emission does lie remarkably close to the extrapolated hard x - ray flux in these low - luminosity data .
we quantify this by calculating the factor by which the disc component underpredicts the observed flux in the uvw1 filter ( uv excess : gdp09 ) in each spectrum .
we also calculate a uv deficit , which is the factor by which the comptonisation model _ overpredicts _ the flux in the uvw1 filter if its seed photons put at 0.1 ev rather than fixed to the disc temperature .
such low energy seed photons may be produced by cyclo - synchrotron emission by the same thermal electrons which produce the compton scattered emission interacting with the tangled magnetic field in the hot flow itself ( @xcite ; @xcite ; @xcite ) .
[ uvexcess ] shows how these evolutions during the outburst .
this confirms the results seen in the three individual spectra discussed above .
the uv excess ( red triangles ) increases strongly with time , as the soft x - ray disc component makes less and less contribution to the uv flux .
the disc is not significantly detected in the final three spectra , but the trend in uv excess is already clearly apparent .
conversely , the uv deficit ( blue squares ) decreases , and is roughly consistent with the observed uv flux from spectrum 21 onwards .
we first look at the possibility that irradiation of the disc is responsible for all the issues highlighted in the previous section .
irradiation of the inner disc changes its derived temperature and radius compared to that derived from fitting purely gravitational energy release models of the disc ( gdp08 ) .
similarly , irradiation of the outer disc can also increase the optical / uv emission from that expected from an unilluminated disc ( @xcite;gdp09 ) we use a slightly modified version of the diskir model of gdp08 so that the disc temperature and normalisation are set by the physical variables of the mass , distance , mass accretion rate and inner and outer radii .
this has the advantage that the outer radius is set in terms of physical units , rather than relative to the ( possibly changing ) inner radius as in diskir .
we parameterise the inner and outer radii in terms of @xmath7 , and leave the inner disc radius , @xmath27 as a free parameter in the fits .
we fix the mass and distance at the best ( though poorly constrained ) estimates of @xmath28 and 5.4 kpc , respectively @xcite .
this model uses the diskbb parameterisation of disc temperature with radius i.e. @xmath29 @xcite , so does not include a stress - free inner boundary condition or colour temperature correction .
if the disc is truncated then the lack of stress - free inner boundary condition is probably more appropriate , but the colour temperature correction , @xmath30 , is still an issue . the derived value of the inner radius is then underestimated by a factor @xmath31 for @xmath32 @xcite . however , the main aim of our spectral fitting is to track _ changes _ in the value of the inner disc radius , as its absolute value depends on the poorly known system parameters .
the key aspect of the truncated disc model which we are testing is the prediction that the truncation radius increases as the spectrum hardens during the decline .
thus we are focussing on the relative values of the inner disc radius during the decline .
the outer radius should remain constant , as should the absorption ( but see @xcite ) so we constrain these by simultaneously fitting three datasets from various points in the outburst with the modified diskir model , together with x - ray absorption parameterised by wabs and uv reddening by redden . assuming standard gas to dust ratios
allows us to tie @xmath33 ( gdp08 ) .
this gives a best fit at @xmath34 and @xmath35 , which are fixed to these values in all fits hereafter .
we fit this model to all the _ swift - rxte _ datasets .
guided by previous low / hard state data ( @xcite ; gdp08 ) , we fix the radius of the irradiated portion of the inner disc at @xmath36 . fig .
[ bigplot ] shows the evolution of the source parameters in this model .
the top panel gives the bolometric flux , which varies by an order of magnitude .
this is dominated by the hard x - ray component , whose spectral photon index ( second panel ) decreases from 1.8 to 1.6 during the first section of the decline , and then stabilizes at this value hereafter .
the source did not make a transition to the high / soft or even intermediate state , but remained in the hard state throughout the outburst .
= 8.5 cm = 8.5 cm the disc spectrum forms a component which connects the soft x - ray rise to the optical emission in this model .
part of this is powered by gravitational energy release , and the inferred mass accretion rate through the disc decreases by more than 2 orders of magnitude during the decline ( third panel ) .
thus the ratio of the hard x - ray flux to this inferred intrinsic disc flux increases by an order of magnitude ( fourth panel ) .
the intrinsic gravitational energy of the disc is augmented by thermalisation of the irradiating flux on the inner disc .
the fraction of the hard x - ray spectrum which illuminates the _ inner _ disc and is thermalised basically decreases during the decline ( fifth panel ) , as expected if the inner edge of the disc is receding .
similarly , the amount of hard x - ray reflection also declines ( sixth panel ) .
conversely , the fraction of hard x - rays which thermalise in the _ outer _ disc increases by almost an order of magnitude ( seventh panel ) .
the final panel shows the disc inner radius .
this is plainly consistent with a more or less constant value , at @xmath37 , ( corresponding to @xmath38 after a colour temperature correction ) although the error bars become large at later stages of the decline .
[ irradiated]a - c show the corresponding fits to the data of figs .
[ simple]a - c for this model .
we can use these derived parameters to explore how the mass accretion rate through the disc compares to that required to power the hard x - rays .
we calculate the mass accretion rate of the corona by : @xmath39 where @xmath40 is the relative efficiency of converting mass to radiation in a coronal flow which extends down to @xmath41 versus that in a disc extending down to the same radius .
however , the disc itself only extends down to a radius @xmath42 so its luminosity from gravitational energy release alone is @xmath43 . thus @xmath44 the spectral fitting parameters give @xmath45 , @xmath46 and @xmath27 .
close to the peak of the outburst , even a radiatively inefficient flow should have @xmath47 . here
we also expect @xmath48 , which requires @xmath41 to be around 3 times smaller than @xmath42 at this point .
we can then calculate @xmath49 assuming @xmath40 and @xmath41 remain constant .
this is plotted against @xmath46 in fig .
the red line is @xmath50 , showing clearly that these models favour a larger mass accretion rate through the corona than through the disc on the decline .
this conclusion is strengthened if the flow is increasingly radiatively inefficient at lower mass accretion rates , as the observed x - ray flux would then require an even larger coronal mass accretion rate to power the same amount of hard x - ray emission .
this conclusion is driven by the _ observational _
requirement for a soft x - ray component in the late state decline data .
this is weak , but significantly present in all but the last three datasets ( assuming that the absorption column remains constant : @xcite ) .
it combines a low luminosity with a fairly high ( soft x - ray ) temperature , which leads to the derived small radius and low mass accretion rate .
this low mass accretion rate is then much smaller than that required to power the observed hard x - ray emission .
we illustrate this by re - coding the irradiated disc model to force @xmath50 for @xmath51 .
this forces the disc to have a higher mass accretion rate , so it must truncate at a larger radius so as not to overproduce the observed soft x - ray component .
irradiation of the inner disc should then be negligible , so we fix @xmath52 for physical consistency , and we focus first on the x - ray data alone so we also fix @xmath53 .
we fit this to spectrum 24 ( the lowest luminosity data for which the disc is significantly detected in the soft x - ray flux ) , and first focus on the x - ray data alone .
the mass accretion rate through the disc is much higher ( @xmath54 g / s compared to @xmath55 in the diskir fits ) and the disc inner radius increases to @xmath56 from @xmath57 ( i.e. @xmath58 and @xmath59 after a colour temperature correction ) .
this disc contributes to the spectrum only at the softest energies of the _ swift _ xrt , and its sharp rise is rather different to the more gradual curvature seen in the soft x - ray emission .
thus this gives a significantly worse fit ( @xmath60 versus @xmath61 , where the two extra free parameters are @xmath45 and @xmath62 ) .
we can only recover the same quality of fit with @xmath63 by including an additional soft component .
this allows the disc to recede back even further , to @xmath64 ( i.e. @xmath65 ) , so that the truncated disc makes no contribution to the soft x - ray flux .
[ 24 ] shows this fit including now the optical / uv data with @xmath66 ( similar to that derived from the original fits , see fig .
[ bigplot ] ) .
irradiation can make such a hot , weak component if the irradiated disc area is _
small_. we fixed the radius of the reprocessing region at @xmath67 for the fits in fig .
[ bigplot ] , but as the disc recedes then the changing geometry means that this should also drop .
we allow this to be a free parameter with @xmath68 and can recover as good a fit as before ( @xmath69 ) with @xmath70 per cent of the bolometric flux being reprocessed in a region with @xmath71
. however , illumination of such a tiny area of the disc surface seems unreasonable from illumination by a central source .
the soft x - ray component seen at the peak of the hard state during the outburst of swift j1753.5 - 0127 is clearly from the disc .
while the system parameters are rather poorly known , the radii derived close to the peak are @xmath72 , corresponding to @xmath73 after a colour temperature correction .
this is consistent with the disc being recessed back from the last stable orbit , as required for the truncated disc / hot inner flow interpretation of the hard state .
there are two independent lines of support for this number .
the first comes from the fact that the hard x - ray spectrum is softer close to the peak , clearly consistent with the disc providing an increasing source of seed photons to compton cool the hot x - ray plasma .
since most of the gravitational potential energy to power this corona is concentrated within @xmath74 then it seems most likely that the disc extends down to similarly small radii .
this means that there can also be irradiation of the inner disc by the hard x - ray emission , leading to reflection and thermalisation of the incident flux .
we see evidence for both these processes in the spectrum , with the extra flux from irradiation increasing the inferred disc radius from the value of @xmath75 inferred from the simple ( unirradiated ) disc fits to the x - ray spectrum , to @xmath59 ( i , e , from @xmath76 to @xmath77 after colour temperature correction ) .
the second is from low frequency qpo which is at its maximum of @xmath78 hz in these data .
this implies an outer radius of @xmath79 assuming that the qpo is produced by lense - thirring ( vertical ) precession of the hot flow around a @xmath80 black hole @xcite . the mass accretion rate required to power the disc emission is comparable with that required to power the coronal hard x - ray emission assuming that the corona extends down to a radius which is @xmath81 times smaller than that of the disc i.e. @xmath82 for the parameters assumed here , i.e. @xmath83 with a colour temperature correction .
we stress again that these system parameters are poorly known , thought this appears quite reasonable .
much smaller radii could be potentially feasible for a hot inner flow , irrespective of the spin of the black hole , as numerical simulations of the mri turbulence show that the large scale height magnetic fields can extract energy from the infalling material beyond the last stable orbit ( e.g. @xcite ) .
the outburst peak spectra then form a template for comparison with the later stages of the decline , where the bolometric flux is lower by a factor 10 .
the truncated disc / hot inner flow model makes a clear prediction that these should have a larger inner disc radius , decreasing the importance of illumination on the inner edge of the disc , and decreasing the amount of associated reflection .
thus illumination of the surface of the disc is not expected to be important in distorting the derived inner disc radii in these spectra .
similarly , as there is little ( or no ) overlap in radii between the hot inner flow and disc , so the disc emission can not be strongly suppressed by comptonisation as can be the case close to the transition ( @xcite ; @xcite ) .
thus any disc component seen in these spectra should give a fairly unbiased view of the inner radius of the flow .
yet associating the observed soft x - ray component with this disc gives derived radii which are as small , if not smaller than , the radii seen at the outburst peak , in clear conflict with the truncated disc models ( see figs [ simple ] and [ bigplot ] )
. however , these radii are themselves in more subtle conflict with the observations .
the small disc has very small luminosity compared to the coronal emission .
this requires that the coronal mass accretion rate must be large compared to that through the disc if the corona is powered by matter accreting through it .
thus the corona can not be predominantly fed by material evaporating from the inner edge of the disc , but instead requires a completely separate coronal flow which incorporates most of the incoming mass accretion from the companion star .
yet it seems quite unlikely that the incoming cool roche lobe overflow stream would be able to form such a coronal flow at large radii .
instead it seems more feasible that the corona is _ not _ powered by its own mass accretion supply , as in the truncated disc models , but is instead powered by mass accreted through the disc , but whose energy is released in the corona ( e.g. @xcite ) .
however , this model itself runs into difficulties since the simplest idea would be for magnetic buoyancy to transport the energy vertically .
the energy is then released in a corona above the disc , so the corona is co - spatial with it and illuminates it .
this gives rise to a reprocessed luminosity @xmath84 ( where @xmath85 is the albedo , and the factor @xmath86 assumes the corona emits isotropically ) adds to the intrinsic disc luminosity @xcite .
for hard spectra , such as those observed here , the reflection albedo @xmath87 , hence the disc flux should be at least @xmath88 .
yet we observe a disc flux of @xmath89 towards the end of the outburst .
thus the only way to circumvent the reprocessing limits whilst having all the mass accrete via the disc is if the energy is advected radially as well as vertically .
then it can be released in a more centrally concentrated region , on size scales smaller than the ( small ! ) disc inner radius .
the alternative of having the hard x - rays be strongly beamed away from the disc , is ruled out by the very similar dim low / hard state spectra seen in xte j1118 + 480 ( @xcite ; 2003 ; @xcite ) , which has high inclination ( @xmath90 ) so that we must see a similar hard x - ray flux to that of the disc .
a potentially less arbitrary solution than using magnetic fields to transport the energy in a dissipationless fashion is if the observed soft x - ray component in the late stages of the decline is not from the disc .
the spectra clearly show that the observed soft x - ray emission can not produce the optical / uv emission without a large change in the reprocessed fraction in the outer disc from that seen during the outburst peak .
yet the outburst remained in the low / hard state , so there is no expected change in source geometry , so the fraction of the total flux which illuminates the outer disc should not change dramatically .
similarly , there is no large change in spectral shape which might produce such a large difference in thermalisation fraction .
yet the observed soft x - ray component is a factor 10 further below the uv emission than at the outburst peak ( fig .
[ uvexcess ] ) . this could be used to argue that the real disc is truncated , as in fig .
[ irradiated ] , so is larger and cooler , so that the same amount of reprocessing can produce the optical / uv .
however , it is clear that the optical / uv emission does change character during the decline .
the rapid variability of the optical flux changes dramatically during the decline , switching to an anti - correlation of the optical and x - ray emission @xcite .
clearly this shows that the optical is no longer made predominantly by reprocessing of the hard x - ray flux , so it gives only an upper limit to the irradiated disc flux .
thus the optical emission does not trace the outer disc , so can not constrain our inner disc models .
instead , we use the similarly dim low / hard state spectra from xte j1118 + 480 to argue for a non - disc origin for the soft x - ray component .
this object has an absorption column which is an order of magnitude lower than that to swift j1753.5 - 0127 .
this gives a correspondingly more sensitive view of the uv and soft x - ray emission , where it is clear that there is a large , cool disc seen in the uv and euv bandpass ( @xcite ; @xcite ) which is completely inconsistent in luminosity and temperature with the much weaker , higher temperature disc emission seen in soft x - rays ( @xcite ; 2003 ; @xcite ) .
there are plainly _ two _ components in this source , one which is consistent with a truncated disc , making no impact on the soft x - ray emission , _ and _ another component which produces weak soft x - ray flux .
this is very similar to the spectrum of 24 as shown in fig .
[ 24 ] .
thus for the late stage decline , our modeling with an irradiated disc is not supported by the data .
the cross - correlation of the rapid variability clearly shows that the optical / uv can not be made by reprocessing in the outer disc , while xte j1118 + 480 clearly shows that the soft x - rays form a separate component to the observed uv / euv emission from the ( truncated ! ) disc .
what then produces the observed soft x - ray component , and what produces the observed optical / uv emission ?
swift j1753.5 - 0127 was observed by _ xmm - newton _ in march 2006 i.e. after the end of the first intensive swift campaign covering the outburst decline .
these data are analysed by @xcite , and have much better signal - to - noise than those of the late decline _ swift _ spectra , but are very similar in shape and intensity .
the higher statistics in these data mean that the soft component is _ unambiguously _ detected , at very high significance .
the same spectra model as used in table [ bigtable ] ( diskbb+thcomp with absorption fixed at @xmath91 @xmath9)gives @xmath92 kev and norm of @xmath93 .
the timing analysis of @xcite showed that this soft x - ray component variability strongly correlates with the hard x - ray variability on short timescales i.e. is driven by reprocessing .
however , on longer timescales , there is additional variability in the soft x - ray flux , implying that there are intrinsic fluctuations in the disc emission as well . the result that much of the short term variability correlates with the hard x - rays rules out a completely separate soft x - ray component such as the jet , while the additional soft x - ray variability shows that the component is truly separate as opposed to being produced from continuum curvature of the x - ray emission or reflection ( suggested by @xcite as a potential origin for the soft x - rays ) .
these variability constraints mean that there are only two potential origins for the soft x - ray component which arises from a small area with relatively high temperature .
firstly , this could be emission from the surface of a residual inner disc , forming a small ring from the last stable orbit to some ( small ) outer radius .
secondly , this could be the inner face of a truncated disc .
we sketch these two possibilities in fig .
[ sketch ] , and outline them below .
evaporation of the disc by thermal conduction is a plausible mechanism to form a truncated disc/ hot inner flow geometry ( e.g. @xcite ) .
counterintuitively , this is not most efficient at the smallest radii , as these also have the highest coronal densities , so have the highest condensation rates .
evaporation first erodes a _ gap _ in the disc , leaving a residual inner disc .
this gap expands radially , eventually giving a fully truncated disc if the mass accretion rate is low enough@xcite .
however , close to the transition , the residual disc can remain .
this geometry allows the outer disc to carry all the mass accretion rate .
this disc evaporates into a corona , truncating it at some large radius into a coronal flow which carries all the mass accretion rate .
however , as the flow accretes to smaller radii , the increasing density gives an increasing condensation rate and some small fraction of the material can condense out of the hot flow .
this forms a cool ring at the innermost radii , with a mass accretion rate which is only a small fraction of the total mass accretion rate , with the rest of the material accreting via the coronal flow@xcite .
reprocessing is still an issue , but if the ring is small , extending over a very narrow range of radii , then it does not subtend a large solid angle to the coronal x - ray emission , so could potentially produce the large observed ratio of coronal to ( inner ) disc luminosity .
some reprocessing is _ required _ in order to produce the short term variability , but there could also be intrinsic variability in the extent of this residual disc .
the temperature of the irradiated region is required to be substantially hotter than that of the disc itself , requiring a very small reprocessing area .
this seems very unlikely for a central source illuminating the top / bottom surface of thin disc .
however , the disc has some ( small ) half thickness , @xmath94 , so its inner rim forms a distinct , small area @xmath95 and subtends a solid angle @xmath96 to the central source .
the small reprocessing area found in section 4.2 above corresponds to @xmath97 , predicting @xmath98 , where @xmath85 is the the reflection albedo .
this is a factor 5 smaller than @xmath62 derived from the data , but potentially feasible given the large uncertainties both on the parameters and on the modeling .
the irradiation origin then gives directly the rapid variability , while turbulence caused by clumping instabilities on this edge could give the required longer term additional variability .
as noted by @xcite in gx339 - 4 , the x - ray emission in the dim low / hard state can extrapolate back quite accurately to fit the optical spectrum ( see also @xcite ; @xcite ) . one way to produce
this is via a single synchrotron component from the innermost post - shock region of the jet .
this is self absorbed in the ir , strongly suppressing the emission at lower frequencies .
as the jet stretches out , each part of it produces a synchrotron component which peaks at a lower frequency than those closer to the center .
all of these make the flat spectrum seen in the radio ( e.g. @xcite ; @xcite ) .
however , the non - thermal synchrotron makes the optical and x - ray emission from a single scattering in a single region , so it is hard to see how this can make the weak anti - correlation between the optical and x - ray flux with lead of a few hundred milliseconds which seems typical of this state ( @xcite ; @xcite ; @xcite ; @xcite ) .
conversely , the cross - correlation signal _ can _ be explained if the optical emission is from the jet , while the x - ray emission is from the corona @xcite .
however , the close match of the optical and x - ray spectra is then very unexpected if these are really from different components .
instead , if the spectrum is formed from thermal comptonisation from self - produced cyclo - synchrotron photons in the hot flow then there can be complex time variability properties imprinted via propagating fluctuations through an inhomogeneous flow ( @xcite ; @xcite ) .
whether these can indeed explain the anti - correlation between optical and x - ray by such spectral pivoting @xcite remains to be seen .
the combined _ swift _ and _ rxte _ observations from the black hole transient swift j1753.5 - 0127 give one of the best datasets to probe the evolution of the inner edge of the accretion disc in the low / hard state .
these instruments cover the optical / uv and soft / hard x - ray bandpasses , giving a detailed picture of the spectral evolution during the low / hard state outburst .
we fit these with a sophisticated irradiated disc model and find that this gives a self - consistent picture around the outburst peak .
weak irradiation increases the inferred radius of the inner disc by a factor @xmath99 .
photons from the disc are the seeds for compton upscattering to produce the hard x - ray emission , and this hard x - ray emission weakly illuminates the outer disc to produce the observed optical / uv by reprocessing , as confirmed by the optical / x - ray cross - correlation @xcite . however , we find clear evidence that the model breaks down as the source flux declines .
the optical spectra require increasingly unlikely levels of reprocessing to explain the observed emission .
a change in origin of the optical emission is confirmed by the dramatic change in optical / x - ray cross - correlation signal @xcite . while the cross - correlation can be explained in models where the optical
is produced by the jet and the x - rays in a corona @xcite , this does not explain the excellent match between the optical and x - ray spectra .
instead it seems more likely that there is a single component connecting the optical and x - ray spectra .
the complex cross - correlation then remains an issue especially for a single synchrotron component from the jet , but it may potentially be explained by thermal comptonisation of ir cyclo - synchrotron emission in an inhomogeneous hot flow .
more fundamentally for the focus of this paper , the soft x - ray emission during the decline may not be thermal emission from the disc either .
if it is , its radius does not change markedly from that seen at the outburst peak , in clear conflict with the predictions of the truncated disc model .
but it also implies that the mass accreting through the disc is much less than the mass accretion rate required to power the corona .
yet it seems most likely that the mass does accrete through the outer disc , in which case the observed weak soft x - ray disc emission is a problem .
either the mass accretes through the disc but most of the energy is transported vertically _ and _ radially ( to get around the reprocessing limits ) by magnetic fields to power a small , central hard x - ray corona , or the soft x - rays are from an additional component , with the truncated disc peaking in the uv .
the lower absorption to xte j1118 + 480 allows us to distinguish between these possibilities . here
, we can see a cool component peaking in the uv which is clearly distinct from the soft x - ray emission .
the uv component fits well to a cool disc , truncated at large radii , so the soft x - ray component can not be the same material .
this may still be associated with the inner disc , perhaps from irradiation of its inner rim , or via a residual inner disc in the discontinuous disc geometry predicted by evaporation models ( e.g. @xcite ) .
however , it is also possible that the soft x - rays are instead produced in a completely different way , such as ionised reflection from grazing incidence angle illumination of the outer disc .
whatever their origin , it is clear from xte j1118 + 480 that the weak soft x - ray component seen in the dim low / hard state does not trace the inner edge of the disc , so can not be used to constrain the truncated disc models .
cyc and cd would like to thank kim page for help in extracting the _ swift _ xrt data .
frontera f. , zdziarski a. a. , amati l. , mikolajewska j. , belloni t. , del sordo s. , haardt f. , kuulkers e. , masetti n. , orlandini m. , palazzi e. , parmar a. n. , remillard r. a. , santangelo a. , stella l. , 2001 , apj , 561 , 1006 mcclintock j. e. , haswell c. a. , garcia m. r. , drake j. j. , hynes r. i. , marshall h. l. , muno m. p. , chaty s. , garnavich p. m. , groot p. j. , lewin w. h. g. , mauche c. w. , miller j. m. , pooley g. g. , shrader c. r. , vrtilek s. d. , 2001 , apj . , 555 , 477 | we test the truncated disc models using multiwavelength ( optical / uv / x - ray ) data from the 2005 hard state outburst of the black hole swift j1753.5 - 0127 .
this system is both fairly bright and has fairly low interstellar absorption , so gives one of the best datasets to study the weak , cool disc emission in this state .
we fit these data using models of an x - ray illuminated disc to constrain the inner disc radius throughout the outburst .
close to the peak , the observed soft x - ray component is consistent with being produced by the inner disc , with its intrinsic emission enhanced in temperature and luminosity by reprocessing of hard x - ray illumination in an overlap region between the disc and corona .
this disc emission provides the seed photons for compton scattering to produce the hard x - ray spectrum , and these hard x - rays also illuminate the outer disc , producing the optical emission by reprocessing . however , the situation is very different as the outburst declines .
the optical is probably cyclo - synchrotron radiation , self - generated by the flow , rather than tracing the outer disc .
similarly , limits from reprocessing make it unlikely that the soft x - rays are directly tracing the inner disc radius .
instead they appear to be from a new component .
this is seen more clearly in a similarly dim low / hard state spectrum from xte j1118 + 480 , where the 10@xmath0 lower interstellar absorption allows a correspondingly better view of the uv / euv emission .
the very small emitting area implied by the relatively high temperature soft x - ray component is completely inconsistent with the much larger , cooler , uv component which is well fit by a truncated disc .
we speculate on the origin of this component , but its existence as a clearly separate spectral component from the truncated disc in xte j1118 + 480 shows that it does not simply trace the inner disc radius , so can not constrain the truncated disc models .
= -0.5 cm [ firstpage ] x - rays : binaries accretion , accretion discs |
from the mathematical point of view there is a very interesting canonically defined neighborhood @xmath22 of the 0-section of the cotangent bundle @xmath23 of the associated symmetric space @xmath24 . in the terminology of @xcite , to which we refer for background
, it is called the _ universal domain _ associated to @xmath2 . its importance in representation theory , where it is most often called the _ crown of the symmetric space _ , has been underlined in numerous works . in the case at hand , where @xmath2 is of hermitian type ,
this is @xmath2-equivariantly and holomorphically isomorphic to the product @xmath25 of the associated hermitian bounded domain and its complex conjugate @xmath26 ( @xcite ) .
it should be noted that @xmath27 and @xmath28 are biholomorphically equivalent as complex manifolds , but are not @xmath2-equivariantly biholomorphic . in order to emphasize the concrete nature of this situation
let us provide a sketch of some basic information about the hermitian symmetric space @xmath27 ( see , e.g. , @xcite for details ) .
a symmetric space is simply a ( finite - dimensional ) riemannian manifold @xmath29 with the additional property that at every @xmath30 there exists an isometry @xmath31 which is a symmetry at @xmath32 in the sense that @xmath33 and @xmath34 .
one shows that the connected component of the identity of the isometry group is a lie group @xmath2 which acts transitively on @xmath35 and that the isotropy subgroup at a neutral point under consideration is a compact subgroup @xmath17 .
based on curvature conditions and canonical decompositions of lie groups , symmetric spaces break up into products of symmetric spaces of three irreducible types .
one such type is @xmath24 where @xmath2 is a simple ( real ) noncompact lie group and @xmath17 is a ( unique up to conjugation ) maximal compact subgroup . in some cases these symmetric spaces @xmath35 have the structure of a complex manifold such that the group @xmath2 acts as a group of holomorphic transformations
. this situation can be group - theorectically characterized by the condition that @xmath17 has a positive - dimensional center which in fact turns out to be @xmath36 . in this case
the metric can be chosen to be hermitian and therefore one refers to @xmath35 as a hermitian symmetric space and @xmath2 being of hermitian type .
note that the maximal compact subgroup @xmath20 of @xmath37 satifies this condition .
noncompact hermitian symmetric spaces of the type discussed above are naturally realized as distinguished open @xmath2-orbits in compact hermitian symmetric spaces @xmath38 which are referred to as the compact duals of the noncompact hermitian symmetric spaces .
a simple example is the case where @xmath39 acts on the space @xmath40 of lines in @xmath41 . here
an associated symmetric space is the open @xmath2-orbit of negative lines . in general
there are always exactly two @xmath2-invariant complex structures on a hermitian symmetric space . in this case
the `` other '' structure is space of positive @xmath21-planes in @xmath42 . using a precise description of the action of the complex lie group @xmath43 on @xmath38 ,
one observes that the noncompact hermitian symmetric space is contained as a bounded domain in a canonically determined ( dense ) open subset of @xmath38 which is biholomorphically equivalent to @xmath44 where @xmath45 .
conversely , every ( irreducible ) bounded domain @xmath46 in @xmath44 which is symmetric with respect to holomorphic symmetries arises in this way . in our case
we may view the hermitian symmetric spaces @xmath27 and @xmath26 as open @xmath2-orbits in the space @xmath38 of lagrangian subspaces of @xmath47 , where @xmath48 is the standard ( holomorphic ) symplectic structure . as a complex manifold @xmath49
can also be @xmath2-equivariantly realized as a @xmath2-invariant open neighborhood of the ( totally real ) orbit @xmath50 of the neutral point in the affine symmetric space @xmath51 . as mentioned above
, it is @xmath2-equivariantly identifiable with a neighborhood of the 0-section in @xmath23 .
these identifications allow us to consider natural @xmath2-invariant symplectic structures on @xmath5 and the resulting moment maps @xmath52 .
the most obvious of these are the restriction to @xmath5 of the canonical symplectic structure on @xmath23 and structures of the type @xmath53 , where @xmath54 is a strictly plurisubharmonic @xmath2-exhaustion of @xmath5 defined , e.g. , by representation theory .
one can show that with respect to any of these structures the generic @xmath2-orbits are coisotropic , i.e. , that the @xmath55-fibers are contained in the @xmath2-orbits .
thus an associated moment map @xmath56 transports a setting of geometric and representation importance to its image @xmath57 in an optimally controllable way .
our goal is to construct bosonic ensembles by equipping such an image with @xmath56-direct image measures .
however , for the symplectic structures mentioned above , @xmath57 is never contained in @xmath3 .
the fact that this happens , which was initially surprising to us , should be further investigated . here , however , we move to a slightly different complex geometric situation by implementing the @xmath2-equivariant complex conjugation @xmath58 and thereby redefining @xmath59 .
thus we still have the advantage of the symplectic geometry and representation theory of the original setup .
in addition , in this new situation the moment map @xmath56 associated to any @xmath2-invariant khlerian structure on @xmath5 satisfies @xmath60 .
furthermore , @xmath56 is generically an open map and consequently its image is not pathological .
the discussion of the @xmath56-images , both before and after complex conjugation , is contained in @xmath61 . in the following section we compute @xmath56 for a particularly beautiful symplectic structure on the product @xmath62 , where @xmath63 is the unit disk in the complex plane .
here we use the identification of @xmath2 with @xmath64 .
this is certainly a toy model .
however , the bounded domain @xmath27 contains a ( flat ) r - dimensional complex polydisk @xmath65 defined by a choice of a maximal ( abelian ) subalgebra @xmath66 in the @xmath67-part of a cartan decomposition @xmath68 such that @xmath69 . therefore certain results can be immediately carried over to @xmath27 by arguing one variable at a time in @xmath65 . on the other hand ,
our goal of understanding asymptotic information concerning spectral invariants is of course not touched here .
nevertheless we do indicate how to compute the distribution function of the eigenvalue @xmath70 with @xmath71 , and we compute this explicitly in the concrete case of the direct image of the uniform measure .
define the mixed signature hermitian form @xmath73 on @xmath74 by @xmath75 and denote by @xmath76 its group of complex linear isometries . as usual @xmath77 .
the group has exactly three orbits on the projective space @xmath78 : @xmath79 $ ] , @xmath80 $ ] and the real projective space @xmath81 $ ] .
note that the involution @xmath82 , @xmath83\mapsto [ \bar{z}_2:\bar { z}_1]$ ] defines a @xmath2-equivariant , antiholomorphic isomorphism @xmath84 . for moment - map considerations it is convenient to define the symmetric bilinear form @xmath85 by @xmath86 which is invariant under the adjoint representation of @xmath2 .
using the duality defined by @xmath87 we will regard the moment map as having image in @xmath88 .
we will use the following @xmath87-orthogonal basis of @xmath88 : @xmath89 if a moment map with values in @xmath90 is given by @xmath91 , then the corresponding map with values in @xmath88 is given by @xmath92 . here
we make two observations which make a case for considering the @xmath93 instead of @xmath25 as the appropriate phase space . [ negative ] the image of a moment map @xmath94 which is defined either by the restricted coadjoint structure or by a khler form @xmath95 of a @xmath2-invariant potential is contained in the complement of the cone @xmath3 of elliptic elements . for a complementary statement for @xmath62
we denote by @xmath96 the augmentation of either the positive or negative cone of elliptic elements by adding @xmath97 .
[ positive ] the image of a moment map @xmath98 which is defined by a khler form @xmath95 of a @xmath2-invariant potential is contained in an augmented cone @xmath99 with the diagonal being mapped to @xmath100 and the complement of the diagonal being mapped to the component of @xmath3 which is contained in @xmath96 .
it should be noted that these observations can be translated to the higher dimensional setting by using strongly orthogonal roots and the polydisk slice mentioned above . turning to the proofs of these remarks
, we first note that in the case where the khler form is defined by an invariant potential , an associated moment map is given by @xmath101 here @xmath102 is the complex structure and @xmath103 denotes the vector field on the manifold which is associated to @xmath104 .
recall that since @xmath2 is semisimple , the moment map is the unique equivariant map to @xmath90 with the hamiltonian property @xmath105 .
of course @xmath106 is tangent to the level sets of the invariant function @xmath54 and the moment map is measuring how @xmath54 is growing along @xmath107 .
in this situation it is therefore important to determine the complex tangent spaces @xmath108 of the @xmath54-level sets .
assuming @xmath109 , this cauchy - riemann tangent space is the 1-codimensional maximal complex subspace of the ( real ) tangent space @xmath110 .
_ proof of proposition [ negative]_. let @xmath111 be the base point so that @xmath112 is the unique symmetric space orbit in @xmath113 . in the standard
homogeneous coordinates this is defined by @xmath114 .
choose @xmath115 to be spanned by the matrix @xmath116 above .
let @xmath117 be the image of the half - open interval @xmath118 by the map @xmath119 .
then @xmath117 is an exact slice for the @xmath2-action on @xmath120 ( see @xcite for this and other background used here . ) .
this means that every @xmath2-orbit intersects @xmath117 in exactly one point .
the orbits @xmath121 for @xmath122 are real hypersurfaces .
the complex tangent spaces @xmath123 are calculated in a general setting in @xcite ( see p. 103 ) . applying this to the case at hand ,
we observe that these spaces are spanned by @xmath124 and @xmath125 where @xmath126 spans the @xmath127 root - space defined by @xmath116 and @xmath128 is the cartan involution . in concrete terms @xmath129 and @xmath130 .
as we have seen above , it follows that @xmath131 .
it follows that @xmath132 is contained in the @xmath116-axis where all isotropy groups are noncompact .
in particular , the image of @xmath56 is contained in the complement of @xmath117 .
it remains to prove the same result for the restriction of the canonical symplectic structure on @xmath23 .
recall that in this case @xmath56 is defined by the liouville form @xmath133 ; in particular , @xmath134 . however , if @xmath135 is the canonical projection , then @xmath136 and consequently @xmath137 for all @xmath138 .
thus again we observe that @xmath139 has empty intersection with @xmath3 .
_ proof of proposition [ positive]_. here we use the slice @xmath117 which is defined as the image of the half - open interval @xmath140 by the mapping @xmath141 in the affine coordinates @xmath142 corresponding to @xmath143 , [ 1:w])$ ] in @xmath144 .
the base point for these considerations is @xmath145 .
just as above we compute the complex tangent space @xmath123 of an orbit @xmath146 for @xmath147 . for this
observe that in this case @xmath32 is in the open @xmath148-orbit in @xmath38 .
therefore the complex tangent space @xmath149 can be represented as the quotient space @xmath150 , where @xmath151 is the ( complex ) lie algebra of the @xmath148-isotropy group at @xmath32 .
we note that this is conjugate to the subgroup of diagonal matrices in @xmath152 .
a direct calculation ( see @xcite p. 74 ) shows that @xmath151 consists of those matrices of the form @xmath153 it is then immediate that @xmath116 and @xmath154 are linearly dependent modulo @xmath151 when regarded as elements of @xmath155 .
thus the real subspace of @xmath149 which is generated by @xmath156 and @xmath157 is a complex line which by dimension arguments _ is _ the complex tangent space @xmath123 . finally , by exactly the same argument as in the proof of the previous theorem it follows that @xmath158 for all @xmath159 .
since @xmath160 , it follows that @xmath161 .
since @xmath100 is necessarily a critical point of @xmath54 , it follows that @xmath162 .
standard computations of @xmath163 show that @xmath164 for @xmath147 and the desired result follows .
actually we proved more than was stated in the above proposition . for future reference we state this here .
* zusatz*. _ the image @xmath165 is contained in the @xmath166-axis @xmath167 . _
here we let @xmath168 be the distance function defined by the poincar metric on @xmath63 .
we temporarily remove the diagonal from @xmath5 where @xmath169 is no longer smooth , show that @xmath169 is strictly plurisubharmonic so that @xmath170 is a @xmath2-invariant khler form and compute the restriction to the slice @xmath117 of the resulting moment map .
as would be expected this map can be ( continuously ) extended to @xmath5 with the diagonal being mapped to @xmath100 .
it is convenient to change coordinates on @xmath5 , letting @xmath172 and @xmath173 where @xmath142 are the affine coordinates which are used above .
thus the complex disk @xmath174 contains the slice @xmath117 as a radius .
in particular @xmath175 is transversal to the @xmath2-orbit of any point @xmath147 . note that @xmath175 is invariant by the diagonal @xmath176-action defined by the element @xmath104 .
it therefore follows that a function @xmath177 on @xmath175 is strictly plurisubharmonic if and only if @xmath178 is strictly convex .
let us begin with this step which of course requires an explicit computation with @xmath179 .
if @xmath180 is the schwarz distance function , then @xmath181 since @xmath182 is strictly convex , in order to prove the following it is enough to show that @xmath178 is strictly convex where @xmath177 is the restriction of @xmath183 to @xmath175 .
the restriction of @xmath179 to @xmath175 is strongly subharmonic . using the obvious coordinates , @xmath184 now we turn to the `` other '' direction , namely the coordinate @xmath185 . for this at each point of @xmath175
we consider the complex curve @xmath186 defined by @xmath187 . near @xmath188
the pullback @xmath189 is nonnegative with @xmath190 and @xmath191 otherwise .
furthermore , @xmath192 as above it is enough to prove the positivity for @xmath179 replaced by @xmath183 .
this is equivalent to @xmath193 for @xmath194 sufficiently small and nonzero .
this is done by elementary manipulations ( see @xcite , p.66 - 7 ) .
one can also prove ( [ positivity ] ) by making several estimates ( also see @xcite , p. 67 ) , but this can be proved in a more conceptual way : since the curve @xmath186 touches the @xmath179 level set from outside , it is immediate that @xmath195 .
if this derivative vanished at @xmath32 , then the levi - form restricted to the complex tangent bundle of the orbit @xmath196 would vanish identically .
it would then follow that @xmath196 would be foliated by complex curves .
this is for numerous reasons impossible , e.g. , continuously moving the leaf through @xmath32 away from @xmath32 so that it no longer intersects the curve @xmath186 yields a contradiction to hurwitz s theorem .
the poincar distance function is plurisubharmonic on @xmath5 and is strictly plurisubharmonic outside of the diagonal .
the above two lemmas show that @xmath197 at every point of @xmath198 .
the fact that the curves @xmath186 are tangent to the @xmath179-level sets at every point of @xmath198 implies that @xmath199 vanishes along @xmath175 and therefore the mixed derivatives @xmath200 vanish identically there as well . in other words
the full complex hessian is diagonalized along @xmath198 with positive entries along the diagonal .
since @xmath179 is only continuous along the diagonal of @xmath5 , to show that it is plurisubharmonic there , we must prove that if @xmath201 is holomorphic with @xmath202 , then @xmath203 is subharmonic .
if the image of @xmath204 is contained in the diagonal , then this is just the identically zero function .
otherwise we may assume that @xmath100 is the only point of the diagonal in the image and the result follows from the plurisubharmonicity of @xmath179 outside the diagonal and that , since @xmath205 , the mean value property is fulfilled at @xmath100 . to keep the notation straight
let @xmath206 denote the slice @xmath117 in @xmath5 .
the sign of the moment map chosen so that its image is contained in the augmented cone @xmath96 , where @xmath3 is the positive cone defined as @xmath207 where @xmath208 . by the above _ zusatz _ we know that @xmath209 . since @xmath210 and @xmath211 are slices for the respective @xmath2-actions , for our purposes here it is enough to have an explicit description of the _ restricted moment map _ @xmath212 .
this amounts to a description of the function @xmath213 along @xmath206 .
if @xmath175 is parameterized by @xmath214 and the ordered basis @xmath215 is used for @xmath88 , then @xmath216 in this parameterization the restriction of the poincar distance funtion to @xmath175 is given by @xmath217 by definition @xmath218 and the desired result follows by a direct computation of the derivative .
the moment map @xmath219 is surjective .
the diagonal in @xmath5 is the @xmath55-preimage of the vertex and the restriction of @xmath56 to the complement of the diagonal is a proper @xmath176-bundle .
furthermore , @xmath56 has the coisotropic property that @xmath220 for all @xmath221 .
let us conclude this note by explicitly computing the spectral density for the direct image measure @xmath4 on @xmath222 defined by the uniform measure @xmath126 on @xmath5 .
it appears that other natural densities , e.g. , those related to gaussian type distributions will only be numerically computable . closed form descriptions of the behaviour of these at the vertex of the cone should be possible .
note here that the matrices on the slice @xmath211 are of the form @xmath223 with @xmath224 .
the relevant distribution is therefore @xmath225 .
we will sketch the explicit computation of @xmath226 here , referring to @xcite for details .
the mapping @xmath227 is given by @xmath228 .
let @xmath229 be the corresponding inverse value .
thus @xmath230 is the euclidean volume of the region @xmath231 .
we compute this via fiber integration using the projection @xmath232 on the first coordinate . for @xmath233 let @xmath234 be the euclidean area of the fiber @xmath235 so that up to a constant @xmath236 let us begin by computing @xmath237 . since the projection @xmath242 is @xmath2-equivariant , @xmath238 is a region in the @xmath240-fiber which is bounded by an orbit of the @xmath2-isotropy group at @xmath100 in the @xmath240-image space . since this isotropy group is acting on the fiber by rotations ,
this is a disk . now
the region @xmath243 in @xmath175 has the point @xmath244 on its boundary .
thus if we apply the tranformation @xmath245 with @xmath246 then @xmath247 is mapped to the point @xmath248 which is on the boundary of the disk in question and the desired result follows .
since we are only dealing with the uniform measure on @xmath5 , it follows that @xmath249 now @xmath250 is the image of the disk @xmath251 by the transformation @xmath252 . since @xmath252 preserves the poincar metric , this is a hyperbolic disk , i.e. , a disk with respect to @xmath179 .
but hyperbolic disks are euclidean disks .
thus , @xmath253 where @xmath254 is the euclidean radius of @xmath250 which we now compute . by rotational invariance
it is enough to compute @xmath255 for @xmath256 on the positive real axis of @xmath63 .
thus the transformation @xmath257 stabilizes the interval @xmath258 of real points of @xmath63 .
since @xmath258 is orthogonal to the boundary of @xmath251 and @xmath257 is conformal , it follows that @xmath258 is orthogonal to the boundary of @xmath259 .
consequently , the length of @xmath260 is the diameter of @xmath259 .
this leads to the following fact . introducing polar coordinates and using rotational symmetry
, one shows that @xmath263 for @xmath264 as above .
this integral can actually be computed in closed form : @xmath265 using the simple dilation @xmath266 to clean up the numbers , one computes that @xmath267 thus one has a rather simple closed form description of the distribution function @xmath226 and the associated density @xmath177 . in the variable
@xmath266 the eigenvalue distribution function @xmath226 and its density @xmath177 associated to the direct image of the uniform measure by the poincar moment map @xmath219 are given by @xmath268 and @xmath269 near @xmath100 one has the power series representation @xmath270 in particular @xmath271 near @xmath100 .
for @xmath272 the expression in the above proposition implies that @xmath273 .
the expected value @xmath274 can actually be explicitly computed as @xmath275 . however , as one observes from the estimate of @xmath177 near @xmath276 the second moment @xmath277 is infinite .
this phenomena does occur in nature , but here it is probably due to the fact that the original uniform measure on the phase space @xmath5 is physically unrealistic .
gaussian densities , e.g. , of the type @xmath278 would perhaps be of greater interest , but the resulting integrals are more complicated .
xxx d. burns , s. halverscheid and r. hind , the geometry of grauert tubes and complexification of symmetric spaces , duke . j. math * 118 * ( 2003 ) , 465491 .
g. fels , a. huckleberry and j. a. wolf : cycles spaces of flag domains : a complex geometric viewpoint , progress in mathematics , volume 245 , springer / birkhuser boston , 2005 s. helgason : differential geometry , lie groups , and symmetric spaces , pure and applied mathematics * 80 * , academic press , 1978 .
t. lck , h. j. sommmers and m. r. zirnbauer , energy correlations for a random matrix model of disordered bosons , j. math . physics , * 47 * , 103304 ( 2006 ) k. schaffert , konstruktion @xmath279-ensembles mittels poincar - metrik , diplomarbeit der ruhr - universitt bochum , november 2010 | random matrix models of disordered bosons consist of matrices in the lie algebra @xmath0 . assuming dynamical stability , their eigenvalues are required to be purely imaginary . here a method is proposed for constructing ensembles @xmath1 of @xmath2-invariant sets @xmath3 of such matrices with probability measures @xmath4 .
these arise as moment map direct images from phase spaces @xmath5 which play an important role in complex geometry and representation theory . in the toy - model case of @xmath6 , where @xmath5 is the complex bidisk and @xmath4 is the direct image of the uniform measure , an explicit description of the spectral measure is given .
the goal of this article is to point out a method for constructing ensemble probability spaces in a bosonic setting .
we do this in the context of the model in @xcite where due to the requirement of dynamical stability the generators @xmath7 of the time evolution are symplectic matrices of elliptic type ,
i.e. , their eigenvalues come in pairs @xmath8 with @xmath9 .
we denote the set of these elements of the symplectic lie algebra @xmath10 by @xmath3 . to put this in perspective
it is perhaps useful to recall that in the case of fermions the basic model spaces consist of hermitian operators @xmath11 such that @xmath12 is , for example , an element of one of the lie algebras @xmath13 , @xmath14 or @xmath15 of the classical compact groups .
for example , in the case of @xmath13 the basic model space becomes the vector space of all hermitian matrices . in that case
the model probability distribution can be chosen to be gaussian , up to constants the density being @xmath16 .
since one is usually most interested in associated eigenvalue distributions , it is entirely appropriate that the model densities are invariant with respect to the action by conjugation of the compact group @xmath17 at hand , i.e. , the adjoint representation of @xmath17 . in the case of @xmath13
this is just the action of the special unitary group @xmath18 . reducing the symmetry by this action ,
one obtains associated classical weyl - group invariant distributions for the eigenvalues .
one then computes various associated distributions , e.g. , eigenvalue spacing distributions and their limits .
for example , in the case of the hermitian matrices the famous gue - density appears as a limiting distribution in this way . at the very outset
, one observes two major differences between the setting of disordered bosons and that of fermions .
first , the necessity of dynamic stability imposes a strong condition on the operators being considered . secondly , unlike the fermionic case where the model probability densities can be chosen to be invariant with repect to the full compact group of symmetries , in the bosonic case the group @xmath19 is noncompact and @xmath2-invariant densities are not available . in @xcite an interesting class of bosonic emsembles @xmath1
is introduced and the program which we sketched above for fermions is carried out . here , as in @xcite ,
ensemble densities which are invariant with respect to the ( unique up to conjugation ) maximal compact subgroup @xmath17 of @xmath2 are proposed .
the usual choice is @xmath20 .
our hope here is that @xmath17-invariant constructions which are closely related to the geometry and represention theory of the symplectic group @xmath2 will also produce bosonic ensembles of quantum mechanical interest .
we begin by explaining this approach for arbitrary @xmath21 .
calculations are carried out in what might seem to be a toy model , i.e. , the case of @xmath6 . due to the fact that we are dealing with a group of hermitian type , where the poincar disk plays a fundamental role in understanding the associated symmetric space , this is an important special case in our future work . |
microwave radiation - induced zero - resistance states@xcite are an interesting phenomenon in two dimensional electron systems ( 2des ) because , for example , coincidence of microwave - induced zero - resistance states with quantum hall zero - resistance states leads to the extinction of the associated quantum hall plateaus.@xcite the microwave radiation - induced magneto - resistance oscillations that lead into the zero - resistance states include characteristic traits such as periodicity in @xmath0,@xcite a 1/4-cycle shift,@xcite , distinctive sensitivity to temperature and microwave power,@xcite along with other specific features.@xcite .
the linear - polarization - sensitivity of these oscillations has been a topic of intensive recent experimental study @xcite ; these studies have shown that the amplitude of microwave radiation - induced magnetoresistance oscillations changes periodically with the linear microwave polarization angle .
looking at the linear polarization characteristics in greater detail , at fixed temperature and polarization angle the amplitude of microwave radiation - induced magnetoresistance oscillations increases with the microwave power .
it follows , approximately,@xcite @xmath1 , where @xmath2 is the amplitude of microwave radiation - induced magnetoresistance oscillations , @xmath3 and @xmath4 are constants and @xmath5 is microwave power . at fixed temperature and microwave power , amplitude of microwave radiation - induced magnetoresistance oscillations
changes sinusoidally with the linear polarization angle .
the experimental results have shown that the longitudinal resistance @xmath6 vs linear polarization angle @xmath7 follows a cosine square function@xcite , i.e. , @xmath8 ( @xmath2 and @xmath9 are constants , @xmath10 is the phase shift , which depend on microwave frequency@xcite ) , at low microwave power .
note tha this angular dependence can also be rewritten as @xmath11 $ ] .
deviation from this functional form was noted for higher microwave intensities . from the theoretical point of view@xcite , there are many approaches to understand the physics of the microwave radiation - induced magnetoresistance oscillations .
these include the radiation - assisted indirect inter - landau - level scattering by phonons and impurities ( the displacement model)@xcite , the periodic motion of the electron orbit centers under irradiation ( the radiation driven electron orbit model)@xcite , and a radiation - induced steady state non - equilibrium distribution ( the inelastic model)@xcite . among these approaches ,
the radiation driven electron orbit model intensively considered temperature@xcite , microwave power@xcite and microwave polarization direction@xcite as factors that could change the amplitude of microwave radiation - induced magnetoresistance oscillations . here , we report an experimental study of microwave power and linear polarization angle dependence of the radiation - induced magnetoresistance oscillations and compare the results with the predictions of the radiation driven electron orbit model .
the comparison provides new understanding of the experimental results at high microwave powers .
the radiation driven electron orbit model@xcite was developed to explain the observed diagonal resistance , @xmath6 , of an irradiated 2des at low magnetic field , @xmath12 , with @xmath12 in the z - direction , perpendicular to the 2des . here
, electrons behave as 2d quantum oscillator in the @xmath13 plane that contains the 2des .
the system is also subjected to a dc electric field in the x - direction ( @xmath14 ) , the transport direction , and microwave radiation that is linearly polarized at different angles ( @xmath7 ) with respect to the transport direction ( @xmath15-direction ) .
the radiation electric field is given by @xmath16 where @xmath17 , @xmath18 are the amplitudes of the mw field and @xmath19 the frequency .
thus , @xmath7 is given by @xmath20 .
the corresponding electronic hamiltonian can be exactly solved@xcite obtaining a solution for the total wave function , @xmath21\ ] ] where @xmath22 are fock - darwin states , @xmath23 is the center of the orbit for the electron motion , and @xmath24 ( for the x - coordinate ) and @xmath25 ( for the y - coordinate ) are the solutions for a @xmath26 driven 2d harmonic oscillator ( classical uniform circular motion ) .
the expressions for an arbitrary angle @xmath7 are given by @xmath27\frac{ee_{0}\cos wt}{m^{*}\sqrt{(w_{c}^{2}-w^{2})^{2}+\gamma^{4 } } } = a_{x}\cos wt \nonumber\\ \\
b(t)&=&\left[\frac{w\sqrt{w^{2}\cos^{2 } \theta + w_{c}^{2}\sin^{2 } \theta}+(w_{c}^{2}-w^{2})\cos\theta}{ww_{c}}\right]\frac{ee_{0}\sin wt}{m^{*}\sqrt{(w_{c}^{2}-w^{2})^{2}+\gamma^{4 } } } = a_{y}\sin wt\end{aligned}\ ] ] where @xmath28 is the electron charge , @xmath29 is a damping factor for the electronic interaction with acoustic phonons , @xmath30 the cyclotron frequency and @xmath31 is the total amplitude of the radiation electric field .
these two latter equations give us the equation of an ellipse , @xmath32 .
then , the first finding of this theoretical model is that according to the expressions for @xmath24 and @xmath25 , the center of the electron orbit performs a classical elliptical trajectory in the @xmath13 plane driven by radiation ( see inset of fig .
4 ) . this is reflected in the x and y directions as harmonic oscillatory motions with the same frequency as radiation . in this elliptical motion electrons in their orbits interact with the lattice ions being damped and emitting acoustic phonons ; in the @xmath24 and @xmath25 expressions , @xmath29 represents this damping .
the above expressions for @xmath24 and @xmath25 are obtained for a infinite 2des .
but if we are dealing with finite samples the expressions can be slightly different because the edges can play an important role
. then the key issue of symmetry / asymmetry of the sample has to be considered in regards of the polarization sensitivity . in the experiments the samples were rectangular - shaped or hall bars
( asymmetric samples ) and the @xmath6 measurements were obtained at each of the longest sides of the sample between two lateral contacts . according to this experimental set up
, we observe that along the classical elliptical trajectories the driven motion in the @xmath15 direction presents fewer restrictions since the top and bottom sample edges are far from the @xmath6 measurement points ( side contacts ) .
thus , there is no @xmath33 constraints due to the existence of edges .
however for the driven motion in the @xmath34 direction , the restrictions are very important due to the presence of the edges from the very first moment .
the lateral edges impede or make more difficult the mw - driven classical motion of the electron orbits in the @xmath34 direction .
the effect is as if the @xmath18 component of the microwave electric field were much less efficient in coupling and driving the electrons than the @xmath17 component .
this situation has to be reflected in the @xmath24 and @xmath25 expressions .
thus , we have phenomenologically introduced an asymmetry factor @xmath35 to deal with this important scenario affecting the obtained @xmath6 . since for a more intense radiation electric field
@xmath31 , the motion in @xmath34 is increasingly hindered we have introduced @xmath36 where @xmath37 is a constant that tends to @xmath38 for symmetric samples and then @xmath24 reads : @xmath39\frac{ee_{0 } } { m^{*}\sqrt{(w_{c}^{2}-w^{2})^{2}+\gamma^{4 } } } \cos wt = a^{*}_{x } \cos wt\ ] ] this @xmath40 behavior has a deep impact on the charged impurity scattering and in turn in the conductivity .
thus , first we calculate the impurity scattering rate @xmath41 between two driven landau states @xmath42 , and @xmath43@xcite . in order to calculate the electron drift velocity ,
next we find the average effective distance advanced by the electron in every scattering jump@xcite : @xmath44 , where @xmath45 is the flight time that is strictly the time it takes the electron to go from the initial orbit to the final one .
this time is part of the scattering time , @xmath46 , that is normally defined as the average time between scattering events and equal to the inverse of the scattering rate .
@xmath47 is the average advanced distance without radiation .
finally the longitudinal conductivity @xmath48 is given by : @xmath49 being @xmath50 the energy .
to obtain @xmath6 we use the relation @xmath51 , where @xmath52 and @xmath53 .
therefore , @xmath54\frac{ee_{0 } } { m^{*}\sqrt{(w_{c}^{2}-w^{2})^{2}+\gamma^{4 } } } \sin w\tau\ ] ]
experiments were carried out on high mobility gaas / algaas hetero - structure hall bar samples .
the samples were placed on a long cylindrical waveguide sample holder and loaded into a variable temperature insert ( vti ) inside the bore of a superconducting solenoid magnet .
the high mobility condition was achieved in the 2des by brief illumination with a red light - emitting diode at low temperature . a microwave launcher at the top of the sample holder excited microwaves within the cylindrical waveguide .
the angle between the long axis of hall bar sample and the antenna in the microwave launcher is defined as the linear polarization angle .
this linear polarization angle could be changed by rotating the microwave launcher outside the cryostat .
low frequency lock - in techniques were utilized to measure the diagonal and off - diagonal response of the sample . at 1.5 k ,
the longitudinal resistance @xmath6 vs magnetic field @xmath12 , exhibits strong microwave radiation - induced magnetoresistance oscillations , see fig .
1 , for frequency @xmath55 = 45.2 ghz , source power , @xmath5 = 4 mw , and vanishing linear polarization , i.e. @xmath56 . the figure shows that maxima and minima up to the fourth order are observable below 0.15 t. the first maxima and minima are designated as @xmath57 and @xmath58 . in figs . 2 and 3
, we examine the linear polarization angle dependence and the microwave power dependence at these extrema .
figure 2 exhibits @xmath6 vs linear polarization angle @xmath7 at different microwave powers at @xmath57 and @xmath58 .
the common features of the data at different microwave powers are : a ) they exhibit an oscillatory lineshape and b ) the peaks and valleys at all powers occur at the same angle . at low microwave power ,
the oscillating curve could be represented by a simple sinusoidal function . however , as microwave power increases , the amplitude of the oscillatory curves increases and , at the same time , deviations from the sinusoidal profile become observable and more prominent .
for instance , at @xmath5= 4 mw , the maxima are relatively rounded and minima are relatively sharp for @xmath57 and , in contrast , the maxima are sharp and minima are rounded for @xmath58 . at the other oscillatory extrema in the @xmath6 vs @xmath12 trace at lower @xmath12 , see fig . 1 ,
such deviations from the simple sinusoidal behavior were more prominent at the same power .
figure 3 shows @xmath6 vs. @xmath5 at different linear polarization angle @xmath7 for @xmath57 and @xmath58 . for @xmath57 , see fig .
3(a ) , @xmath6 increases non - linearly as the microwave power increases . on the other hand ,
3(b ) , @xmath6 decreases non - linearly with increasing @xmath5 at @xmath58 .
3(a ) and ( b ) , all traces start at the same resistance value at @xmath5=0.01 mw since the ( essentially ) dark resistance is invariant under polarization angle rotation .
in figure 4 , we present calculated results of irradiated @xmath6 versus @xmath12 for a microwave frequency of @xmath59 ghz and power of @xmath60 .
as in the experimental curve of fig .
2 , we obtain , at a temperature of @xmath61 , clear @xmath6 oscillations which turn out to be qualitatively and quantitatively similar to experiment . for the exhibited curve ,
the polarization angle is zero .
the most prominent peak and valley are labelled as @xmath57 and @xmath58 respectively . in the inset of this figure we present a schematic diagram showing the radiation - driven classical trajectories of the guiding center of the electron orbit .
figure 5 exhibits calculated results of irradiated @xmath6 versus linear polarization angle @xmath7 for different microwave powers for peak @xmath57 and valley @xmath58 in panels a ) and b ) respectively .
the microwave frequency is 45.2 ghz and @xmath62 is 1.5 k. the microwave power ranges from 0.01 to 6.0 mw . as in the experimental results of figure 2 , we observe that the @xmath6 curves evolve from a clear sinusoidal profile at low microwave powers to a different profile where , for instance , in upper panel ( @xmath57 ) the peaks broaden and the valleys get sharpened .
similar trend is observable in the lower panel ( @xmath58 ) .
the explanation for this peculiar behavior can be obtained from equation [ 5 ] and the square root between brackets .
when the microwave power ( electric field amplitude @xmath31 ) increases , the factor @xmath63 gets smaller and smaller . as a result , the @xmath6 curve begins to lose its simple sinusoidal profile . at high powers
the latter factor is so small that the @xmath64 factor is predominant and the square root tends to the absolute value of @xmath64 : @xmath65 \rightarrow \sqrt{\cos^{2}\theta } = abs[\cos \theta ] \ ] ] the profile observed in experiment and calculation for high powers is very similar to the one of the @xmath66 $ ] .
this evolution can be clearly observed in the panel ( c ) of figure 5 . in figure 6 , we present calculated @xmath6 under radiation versus the microwave power for different polarization angles for peak @xmath57 , upper panel , and for valley @xmath58 , lower panel .
for increasing angles from @xmath67 to @xmath68 the behavior of @xmath57 and @xmath58 are similar in the sense that the corresponding intensity of both becomes smaller and smaller . in other words , for increasing angles the height of the peak decreases and the depth of the valley decreases too .
the theoretical explanation comes from the square root term as before . for increasing angles the cosine term tends to zero becoming predominant the sine term
. however the latter gets smaller for increasing power due to the asymmetry factor .
all curves share one important feature , the non - linearity of @xmath6 versus @xmath5 suggesting a sublinear relation .
this behavior can be straightforward explained with our model in terms of : @xmath69 and then the @xmath6 profile with @xmath5 follows a square root dependence as experiments show . versus linear polarization angle @xmath7 for the magnetic fields corresponding to the peak @xmath57 ( panel ( a ) ) and valley @xmath58 ( panel ( b ) ) in panel ( c ) we present simulation of the curves evolution with mathematical functions when the @xmath70 term decreases .
the microwave photo - excitation at 45.2 ghz , the microwave power ranges from from 0.01 to 6.0 mw and @xmath62=1.5 k. , width=302 ]
we have presented together experimental- and theoretically - calculated- results concerning the microwave power and linear polarization angle dependence of the microwave irradiated oscillatory magnetoresistance in the gaas / algaas two - dimensional electron system .
experimental results show that , as the microwave power increases , the @xmath6 vs. @xmath7 traces , see fig .
2 , gradually lose the simple sinusoidal profile , as the profile begins to resemble the absolute value of the cosine function .
we presented the theoretical insight to explain this evolution using the radiation - driven electron orbit model , which suggests a profile following @xmath66 $ ] .
intuitively , one can motivate the change in the profile of the @xmath6 vs. @xmath7 curves with increasing power by noting that increasing the power , i.e. , driving the 2des more strongly , most likely increases the harmonic content in the @xmath6 vs. @xmath7 lineshape , which leads to deviations from the simple sinusoidal function proposed for low power in ref .
@xcite . in the low power limit , this theoretical prediction of the radiation driven electron orbit model matched the experimental suggestion of @xmath71 $ ] @xcite .
one might tie together the low power @xmath6 vs. @xmath7 lineshape ( @xmath72 ) with the high power @xmath6 vs. @xmath7 lineshape ( @xmath66 $ ] ) by examining the fourier expansion over the interval @xmath73 $ ] of @xmath66 = ( 2/\pi ) + ( 4/\pi ) \sum_{m=1}^{\infty } ( ( ( -1)^{m}/(1 - 4m^{2}))cos(2m\theta)$ ] . if one keeps the lowest order ( @xmath74 ) term , which is the only one likely to be observable at low power , then @xmath66 \rightarrow ( 2/\pi)(1 + ( 2/3)cos(2\theta ) ) \sim cos^2(\theta)$ ] .
magnetotransport measurements at georgia state university are supported by the u.s .
department of energy , office of basic energy sciences , material sciences and engineering division under de - sc0001762 .
additional support is provided by the aro under w911nf-07 - 01 - 015 .
j.i . is supported by the mineco ( spain ) under grant mat2014 - 58241-p and itn grant 234970 ( eu ) .
grupo de matematicas aplicadas a la materia condensada , ( uc3 m ) , unidad asociada al csic
. 0 r. g. mani , j. h. smet , k. von klitzing , v. narayanamurti , w. b. johnson , and v. umansky , nature * 420 * , 646 ( 2002 ) .
m. a. zudov , r. r. du , l. n. pfeiffer , and k. w. west , phys .
lett . * 90 * , 046807 ( 2003 ) .
r. g. mani , w. b. johnson , v. umansky , v. narayanamurti , and k. ploog , phys .
b * 79 * , 205320 ( 2009 ) .
r. g. mani , j. h. smet , k. von klitzing , v. narayanamurti , w. b. johnson , and v. umansky , phys .
* 92 * , 146801 ( 2004 ) ; phys .
b * 69 * , 193304 ( 2004 ) .
r. g. mani , c. gerl , s. schmult , w. wegscheider , and v. umansky , phys .
b * 81 * , 125320 ( 2010 ) .
j. iarrea , r. g. mani , and w. wegscheider , phys .
b * 82 * , 205321 ( 2010 ) .
a. e. kovalev , s. a. zvyagin , c. r. bowers , j. l. reno , and j. a. simmons , solid state commun .
* 130 * , 379 ( 2004 ) .
r. g. mani , v. narayanamurti , k. von klitzing , j. h. smet , w. b. johnson , and v. umansky , phys .
b * 70 * , 155310 ( 2004 ) ; phys . rev .
b * 69 * , 161306 ( 2004 ) .
b. simovic , c. ellenberger , k. ensslin , h. p. tranitz , and w. wegscheider , phys .
b * 71 * , 233303 ( 2005 ) .
r. g. mani , phys .
b * 72 * , 075327 ( 2005 ) ; appl . phys
85 , 4962 ( 2004 ) ; appl .
. lett . * 91 * , 132103 ( 2007 ) ; physica e * 25 * , 189 , ( 2004 ) ; physica e * 40 * , 1178 ( 2008 )
. s. wiedmann , g. m. gusev , o. e. raichev , t. e. lamas , a. k. bakarov , and j. c. portal , phys .
b * 78 * , 121301 ( 2008 ) .
a. n. ramanayaka , r. g. mani , and w. wegscheider , phys .
b * 83 * , 165303 ( 2011 ) . r. g. mani , a. n. ramanayaka , and w. wegscheider , phys . rev .
b * 84 * , 085308 ( 2011 ) .
o. m. fedorych , m. potemski , s. a. studenikin , j. a. gupta , z. r. wasilewski , i. a. dmitriev , phys .
b * 81 * , 201302 ( 2010 ) .
s. wiedmann , g. m. gusev , o. e. raichev , a. k. bakarov , and j. c. portal , phys .
rev . b**84 * * , 165303 ( 2011 ) .
dai , k. stone , i. knez , c. zhang , r. r. du , c. l. yang , l. n. pfeiffer , and k. w. west , phys .
b * 84 * , 241303 ( 2011 ) .
a. n. ramanayaka , r. g. mani , j. iarrea , and w. wegscheider , phys .
b * 85 * , 205315 ( 2012 ) .
r. g. mani , j. hankinson , c. berger , and w. a. de heer , nat .
commun . * 3 * , 996 ( 2012 ) .
t. ye , r. g. mani , and w. wegscheider , appl .
lett . * 102 * , 242113 ( 2013 ) ; ibid . *
103 * , 192106 ( 2013 ) .
j. iarrea , europhys . lett . * 108 * , 2 ( 2014 ) .
r. g. mani , a. kriisa , and w. wegscheider , sci .
rep . * 3 * , 2747 ( 2013 ) .
r. g. mani and a. kriisa , sci .
rep . * 3 * , 3478 ( 2013 ) . t. ye , h.c .
liu , w. wegscheider and r. g. mani , phys .
b * 89 * , 155307 ( 2014 ) .
a. n. ramanayaka , t. ye , h.c .
liu , w. wegscheider and r. g. mani , physica b * 453 * , 43 ( 2014 ) .
r. g. mani , a. n. ramanayaka , t. ye , m. s. heimbeck , h. o. everitt and w. wegscheider , phys . rev .
b * 87 * , 245308 ( 2013 ) . t. ye , w. wegscheider and r. g. mani , appl .
. lett . * 105 * , 191609 ( 2014 ) .
h. c. liu , t. ye , w. wegscheider and r. g. mani , j. appl .
phys . * 117 * , 064306 ( 2015 ) .
t. ye , h.c .
liu , z. wang , w. wegscheider and r. g. mani , sci .
rep . * 5 * , 14880 ( 2015 ) .
z. d. kvon et al .
, jetp lett .
97 , 41 ( 2013 ) .
a. d. chepelianskii , j. laidet , i. farrer , d. a. ritchie , k. kono , and h. bouchiat , phys .
b 90 , 045301 ( 2014 ) .
s. chakraborty , a. t. hatke , l. w. engel , j. d. watson , and m. j. manfra , phys .
b. * 90 * , 195437 ( 2014 ) .
a. d. levin , z. s. momtaz , g. m. gusev , o. e. raichev , and a. k. bakarov , phys .
lett . * 115 * , 206801 ( 2015 ) .
a. d. chepelianskii , m. watanabe , k. nasyedkin , k. kono , and d. konstantinov , nat .
* 6 * , 7210 ( 2015 ) .
x. l. lei and s. y. liu , phys .
b * 86 * , 205303 ( 2012 )
. j. iarrea , j. appl . phys . * 113 * , 183717 ( 2013 ) .
j. iarrea and g. platero , nanotechnology * 21 * , 315401 ( 2010 ) .
jesus inarrea and gloria platero , j. phys .
* 27 * 415801 ( 2015 ) jesus inarrea and gloria platero , phys .
b * 76 * 073311 ( 2007 ) a. c. durst , s. sachdev , n. read , and s. m. girvin , phys
lett . * 91 * , 086803 ( 2003 ) .
a. v. andreev , i. l. aleiner , and a. j. millis , phys .
* 91 * , 056803 ( 2003 ) .
v. ryzhii and r. suris , j. phys .
matt . * 15 * , 6855 ( 2003 ) .
a. a. koulakov and m. e. raikh , phys .
b * 68 * , 115324 ( 2003 ) .
x. l. lei and s. y. liu , phys .
lett . * 91 * , 226805 ( 2003 ) .
j. iarrea and g. platero , phys .
* 94 * , 016806 ( 2005 ) ; phys .
b * 76 * , 073311 ( 2007 ) .
j. iarrea and g. platero , phys .
b * 72 * 193414 ( 2005 ) i. a. dmitriev , m. g. vavilov , i. l. aleiner , a. d. mirlin , and d. g. polyakov , phys . rev .
b * 71 * , 115316 ( 2005 )
. x. l. lei and s. y. liu , phys .
b * 72 * , 075345 ( 2005 ) .
a. d. chepelianskii , a. s. pikovsky , and d. l. shepelyansky , eur .
j. b * 60 * , 225 ( 2007 ) .
j. iarrea , appl .
* 92 * , 192113 ( 2008 ) .
a. d. chepelianskii and d. l. shepelyansky , phys .
b * 80 * , 241308 ( 2009 ) .
j. iarrea , appl .
. lett . * 99 * , 232115 ( 2011 ) .
j. iarrea , appl .
* 100 * , 242103 ( 2012 ) .
j. iarrea , appl .
89 * , 052109 ( 2006 ) .
a. kunold and m. torres , physica b * 425 * , 78 ( 2013 ) .
o. v. zhirov , a. d. chepelianskii , and d. l. shepelyansky , phys .
b * 88 * , 035410 ( 2013 ) . x. l. lei and s. y. liu , j. appl
. phys . * 115 * , 233711 ( 2014 ) .
a. yar and k. sabeeh , j. phys.:cond .
* 27 * , 435007 ( 2015 ) . v. g. ibarra - sierra , j. c. sandoval - santana , j. l. cardoso , and a. kunold , annal
. phys . * 362 * , 83 ( 2015 ) .
o. e. raichev , phys .
b * 91 * , 235307 ( 2015 ) .
figure 1 ) longitudinal resistance @xmath6 versus magnetic field @xmath12 with microwave photo - excitation at 45.2 ghz , 4 mw and @xmath62=1.5 k. the polarization angle , @xmath7 , is zero .
the labels , @xmath57 and @xmath58 at the top abscissa mark the magnetic fields of the first peak and valley of the oscillatory magneto - resistance .
figure 2 ) longitudinal resistance @xmath6 versus linear polarization angle @xmath7 at the magnetic field corresponding to ( a ) @xmath57 and ( b ) @xmath58 .
the microwave frequency is 42.5 ghz .
different colored symbols represent different source microwave powers from 0 to 4 mw .
figure 3 ) figure exhibits the longitudinal resistance @xmath6 versus microwave power @xmath5 at the magnetic field corresponding to ( a ) @xmath57 and ( b ) @xmath58 .
the microwave frequency is @xmath55 = 42.5 ghz .
different color symbols represent different linear polarization angles , @xmath7 , between @xmath75 and @xmath76 . figure 4 ) calculated irradiated magnetoresistance @xmath6 versus magnetic field @xmath12 with microwave frequency of 45.2 ghz , microwave power @xmath77 mw and @xmath62=1.5 k. the polarization angle , @xmath7 , is zero .
symbols @xmath57 and @xmath58 correspond to the peak and valley , respectively , as indicated .
figure 5 ) calculated irradiated magnetoresistance @xmath6 versus linear polarization angle @xmath7 for the magnetic fields corresponding to the peak @xmath57 ( panel ( a ) ) and valley @xmath58 ( panel ( b ) ) in panel ( c ) we present simulation of the curves evolution with mathematical functions when the @xmath70 term decreases .
the microwave photo - excitation at 45.2 ghz , the microwave power ranges from from 0.01 to 6.0 mw and @xmath62=1.5 k. figure 6 ) calculated irradiated magnetoresistance @xmath6 versus microwave power .
the microwave frequency is 45.2 ghz and @xmath62=1.5 k. the polarization angle , @xmath7 , ranges from @xmath67 to @xmath68 , as indicated . | we present an experimental study of the microwave power and the linear polarization angle dependence of the microwave induced magnetoresistance oscillations in the high - mobility gaas / algaas two - dimensional electron system .
experimental results show sinusoidal dependence of the oscillatory magnetoresistance extrema as a function of the polarization angle . yet , as the microwave power increases , the angular dependence includes additional harmonic content , and begins to resemble the absolute value of the cosine function .
we present a theory to explain such peculiar behavior . |
generating spin current is one of the fundamental issues in spintronics @xcite . when spin - up and spin - down electrons travel in opposite directions , the net charge current is @xmath2 while the spin current is @xmath3 , where @xmath4 ( @xmath5 ) is the spin - up ( spin - down ) electron current .
a device which can drive a spin current into external circuits is called spin battery ( sb ) @xcite . thus far , various sbs have been proposed , e.g. , the earlier dipolar and unipolar sbs summarized in reference @xcite , and the following three - terminal devices consisting of metallic / ferromagnetic poles @xcite or even involving superconducting pole @xcite . among the existing schemes , in most of the multipolar sbs , both charge current and spin current coexist except for particular parameter regimes .
however , the charge current in a unipolar sb must be zero in the steady state since only one pole exists @xcite . from this perspective
, the unipolar sb is the superior candidate for generating a pure spin current ( psc ) with @xmath6 and @xmath7 .
a built - in spin flip mechanism is usually necessary in a unipolar sb , namely , it draws in electrons with one spin orientation from the pole , then flips the spin inside the sb , followed by pushing out the electron with opposite spin orientation to the pole @xcite .
if this kind of sb is connected to an external circuit which is not closed , it drives a psc in that circuit .
typical built - in spin flip mechanism in a unipolar sb is a ferromagnetic resonance or a rotating external magnetic field @xcite , which involves somewhat complicated time - varying external fields . .
( b ) eigenenergies of a cnt qd as a function of the magnitude of @xmath8 with @xmath9 and the cnt parameters @xmath10 , @xmath11 , and @xmath12 . here
, the values of @xmath13 , @xmath14 , and @xmath15 are extracted from reference @xcite .
the circle marks the exact crossing point @xmath16 .
( c)-(e ) illustrate the mechanism of generating pscs .
electron temperature @xmath17 in ( c ) and ( d ) is zero while in ( e ) is finite . ] in this work , we predict a unipolar sb made up of a quantum dot ( qd ) formed in a suspended carbon nanotube ( cnt ) exposed to an external magnetic field , see figure [ fig1](a ) . as we shall see below , instead of employing a time - varying external field
, a natural spin flip source is available in this sb setup .
it is known that in a cnt qd , the twofold spin and twofold orbital symmetries give rise to a fourfold degenerate energy level @xmath18 . in the presence of the spin - orbit coupling @xmath13 , intervalley scattering @xmath14 , and applying an external longitudinal magnetic field @xmath8 which couples to the electronic orbital ( @xmath15 ) and spin ( @xmath19 ) magnetic moments , the degenerate energy level @xmath18 splits into four branches @xcite @xmath20 and @xmath21 , as shown in figure [ fig1](b ) .
experimentally , by the cooperation of a gate voltage ( controlling @xmath18 ) and a longitudinal magnetic field , the four - level structure can be finely tuned @xcite . in particular , close to the exact crossing point @xmath16 ( marked by the circle in figure [ fig1](b ) ) , one has two levels of opposite spin and the same orbital and their energy separation is smaller than the energy distance from other levels .
recently , a phonon - mediated spin flip mechanism established within this two - level subspace has been addressed @xcite .
it is a consequence of the spin - vibration interaction ( svi ) resulting from the interplay between the intrinsic spin - orbit coupling and the vibrational modes of the suspended cnt .
the coupling constant of svi reads @xmath22 @xcite , with @xmath23 the profile function , and @xmath24 with @xmath25 and @xmath26 being the nanotube s mass and the @xmath27th eigenfrequency of the vibration , respectively .
one can estimate @xmath28mhz for the first odd mode in a typical cnt @xcite .
recent experiment suggests an even stronger @xmath29 due to the large @xmath13 measured @xcite .
moreover , the vibrational modes of cnts usually couple to an equilibrium thermal phonon bath representing the environment induced by thermal nuclear motions @xcite . it has been found that , at unequal electron ( @xmath17 ) and thermal bath ( @xmath30 ) temperatures , the heat current between the bath and the electrons can be converted into an electron current @xcite .
we appreciate that this nontrivial thermoelectric effect could provide the driving force needed in the present sb setup .
it is worth mentioning that the nonequilibrium between @xmath17 and @xmath30 in cnts has already been observed experimentally @xcite and discussed theoretically @xcite . in particular , the bath temperature is maintained provided that the svi strength @xmath29 is much weaker ( which is true in a typical cnt @xcite ) than the coupling of cnt vibrations to the thermal phonon bath @xcite .
alternatively , an artificial thermal phonon bath held at a temperature @xmath30 may be realized simply by an electronically insulating hard substrate touching the quantum dot @xcite , which is spatially well separated from the electronic pole .
hence , a stable temperature difference can be achieved by heating or cooling the electronic pole solely .
first of all , we illustrate why the setup sketched in figure [ fig1](a ) can work as a sb . for a large cnt qd without coulomb interaction ,
the physics is summarized in figures [ fig1](c)-[fig1](e ) . as shown in figure
[ fig1](c ) , at @xmath31 the fermi distribution in the external pole changes abruptly from unity to zero at the fermi level @xmath32 . by adjusting the gate voltage together with the applied magnetic field ,
the qd energy level can split into two sublevels such that @xmath33 . in this case
, the lower sublevel is occupied by a spin - up electron while the upper one is empty . for nonzero @xmath34 where thermal phonons are available , due to the svi
, the spin - up electron at lower sublevel can reverse its spin and transit into an excited state with energy @xmath35 by absorbing a phonon .
provided that @xmath36 , this excited spin - down qd electron can easily tunnel out to the external pole . during this combined process ,
the spin - up electrons continuously flow out of the pole into the qd while the spin - down qd electrons persist in injecting to the pole , which successfully establishes a positive psc ( @xmath37 ) . when the resonance condition @xmath38 ( i.e. , @xmath39 ) is satisfied , the magnitude of psc reaches its maximum
. similarly , a negative psc ( @xmath40 ) can be achieved under the opposite sublevel configuration @xmath41 [ see figure [ fig1](d ) ] . in figure [ fig1](e ) ,
when finite @xmath17 is considered the fermi distribution is smeared around @xmath32 and thereby a few holes ( electrons ) are available below ( above ) @xmath32 . as a result ,
the spin - up electron occupied at the lower sublevel can tunnel to the pole , while a spin - down pole electron can inject to the upper sublevel in qd and then transits immediately to the lower sublevel by emitting a phonon .
this is an opposite process to the one depicted in figure [ fig1](c ) . with this respect
, the positive psc will be suppressed at finite @xmath17 .
nevertheless , this also provides a mechanism to reverse the psc when the process in figure [ fig1](e ) prevails over the one in figure [ fig1](c ) .
the inclusion of a strong coulomb interaction in the cnt qd will not disturb the substantial physical scenario , provided that only the coulomb blockade effect @xcite survives whereas all the high - order cotunneling processes @xcite are suppressed at weak tunnel coupling or at high enough temperatures .
physically , a finite coulomb repulsion @xmath42 will induce two more sublevels with higher energies @xmath43 and @xmath44 . nevertheless , one can always adjust the upper or lower two sublvels to the vicinity of @xmath32 to act as the two relevant sublevels depicted in the illustrations , while the other two sublevels deviating largely from the fermi level do not take part in the psc generating .
this is demonstrated in section [ interacting ] by the master equation calculations incorporating the electron - electron correlations at the coulomb blockade level .
the scenarios proposed above indicate that the magnitude and the direction of psc are dependent on the strength of svi , the sublevel configuration in qd , and the electron temperature @xmath17 , ( or relatively , the bath temperature @xmath30 ) .
moreover , as we discuss below , the tunneling rate between the qd and the pole , as well as the intradot coulomb interaction , can also influence the magnitude of psc . in
what follows , we identify the physics mentioned above by solving the model hamiltonian presented in section [ model ] . in section [ noninteracting ] , the nonequilibrium green s function theory is employed to study the psc generated in a qd without coulomb interaction . in section [ interacting ] , we use the master equation method to study the effect of coulomb interaction on the psc generation .
finally , a conclusion is given in section [ summary ] .
the sb setup we consider [ figure [ fig1](a ) ] can be described by the hamiltonian with a general form @xmath45 , where @xmath46 here , @xmath47 is the single - particle energy of an electron with momentum @xmath48 in the noninteracting external pole .
@xmath49 represents a single vibrational mode of frequency @xmath50 , which describes the vibration of the suspended cnt .
@xmath51 describes the effective two - level cnt qd influenced by the svi @xcite , as mentioned above .
@xmath52 denotes the spin - dependent sublevels , @xmath42 is the coulomb repulsion energy in the qd , and @xmath29 measures the strength of svi .
@xmath53 stands for the tunneling coupling between the qd and the pole , with @xmath54 being the tunneling matrix element .
an electron and/or hole transferring between the qd and the pole is described by an effective tunneling rate @xmath55 , where @xmath56 is the pole density of states .
we first consider a large qd in which the intradot coulomb interaction could be neglected . using the standard keldysh nonequilibrium green s function theory @xcite , the steady spin - dependent electron current , @xmath57\rangle$ ] with @xmath58 , flowing through pole into the qd can be expressed as @xmath59,\label{eq1}\ ] ] where @xmath60 and @xmath61 represent the full lesser and greater green s functions ( gfs ) of the localized qd electron .
@xmath62 and @xmath63 $ ] are the lesser and greater self - energies , respectively , contributed from tunnel coupling to the pole . @xmath64 + 1\}^{-1}$ ] is the fermi distribution of pole electron ( we set @xmath65 ) . here
the spin accumulation ( spin - dependent @xmath32 ) in the pole is neglected assuming that the size of the pole is sufficiently large and the spin - relaxation time is sufficiently short @xcite .
it means that the generated electron current injects to the external circuit promptly . to solve the relevant lesser and greater qd gfs
, one has to make some approximations since the study on phonon - mediated inelastic transports is far from trivial , even though the qd itself is noninteracting .
various treatments on dealing with electron - phonon interaction from weak to strong coupling regime as well as from equilibrium to nonequilibrium have been established @xcite thus far . in this work ,
we focus on the weak coupling regime where @xmath66 , which is true in a typical suspended cnt device @xcite . in this case
, a generalization of the self - consistent born approximation @xcite , devised for treating the holstein - type electron - phonon interaction , is straightforward .
the main idea lies in that assuming the interactions of the qd with pole and phonons are independent of each other , i.e. , the total self - energies are obtained in the form @xmath67 , where @xmath68 with @xmath69 denotes the retarded ( advanced ) component .
therefore , the lesser ( greater ) qd gf can be obtained by following keldysh equation @xcite @xmath70g_\sigma^{a}(\omega ) , \label{keldysh}\ ] ] where @xmath71 \label{se}\ ] ] is the lesser ( greater ) self - energy obtained by considering the hartree and fock self - energy diagrams @xcite , regarding the svi as the perturbation term . here
, @xmath72 represents the opposite spin of @xmath73 .
@xmath74 - 1\}^{-1}$ ] is the average number of phonons in the equilibrium thermal bath to which the vibrational mode is coupled .
we note that the hartree self - energy vanishes since there is no spontaneous spin - flip term in the present hamiltonian @xmath75 . substituting equations ( [ keldysh ] ) and ( [ se ] ) into the current formula equation ( [ eq1 ] )
, one immediately obtains @xmath76\nonumber\\ & & -\sigma_0^>(\omega)[n_{ph}g_{\bar\sigma}^{<}(\omega-\varepsilon_{ph})+(1+n_{ph})g_{\bar\sigma}^{<}(\omega+\varepsilon_{ph})]\}g^a_\sigma(\omega),\label{c1}\end{aligned}\ ] ] within the framework of self - consistent born approximation , all the full gfs and self - energies have to be solved in an iterative manner , however , for weak svi as we consider , to the lowest - order of the coupling strength @xmath29 , one can replace the full gfs appear in equation ( [ c1 ] ) by their bare counterparts @xmath77 and @xmath78 . collecting these terms
we arrive at a compact current formula @xmath79 where @xmath80 and @xmath81 f ( \omega_{+})-n_{ph}f ( \omega_{- } ) [ f ( \omega_{+})-1]$ ] , with @xmath82 . here , @xmath83 is the spin - resolved density of dot states in the absence of svi .
the charge current conservation law @xmath84 is readily checked by replacing @xmath73 by @xmath72 in equation ( [ eq6 ] ) . thus the generated spin current is indeed a psc with @xmath85 .
it is not difficult to understand that this driven psc is a consequence of the nontrivial thermoelectric effect with respect to the fermi ( pole ) and bose ( thermal bath ) reservoirs .
some analytically insights about equation ( [ eq6 ] ) are summarized as follows : \i ) the magnitude of @xmath86 is proportional to @xmath87 .
\ii ) @xmath86 reverses its direction but keeps the magnitude when we exchange the positions of spin - up and spin - down sublevels .
\iii ) the function @xmath88 has exactly the same sign with @xmath89 , irrespective of the specific parameters , and it becomes zero when @xmath90 . particularly , @xmath91 converge to a constant @xmath92 in the limit @xmath93 .
on the other hand , in the limit @xmath94 , the fermi distribution reduces to the heaviside function such that the psc becomes @xmath95 and @xmath96 at different electron temperatures as ( a ) @xmath31 ( @xmath97 ) , ( b ) @xmath98 ( @xmath99 ) , and ( c ) @xmath100 ( @xmath101 ) .
the two dashed lines in ( a ) indicate the specific sublevel configurations @xmath102 ( black dashed line ) and @xmath103 ( white dashed line ) .
( d ) the evolution of psc against the temperature difference @xmath104 at three sublevel configurations indicated by the black dots in ( a ) .
the parameters used are @xmath105 , @xmath106 , and @xmath107 . ] on the tunneling rate under two specific sublevel configurations as ( a ) @xmath102 and ( b ) @xmath103 , as indicated by the black and white dashed lines , respectively , in figure [ fig2](a ) .
the parameters used are @xmath105 and @xmath108 . ] in this section , we present our numerical results based on equations ( [ eq6 ] ) and ( [ eq7 ] ) . in all the calculations , we take @xmath109 as the energy unit .
we first display the dependence of psc on the sublevel configuration at three temperatures as @xmath31 ( case i ) , @xmath110 ( case ii ) , and @xmath111 ( case iii ) . in case i , two psc islands like
baguettes develop with explicit edges [ figure [ fig2](a ) ] .
more precisely , a considerable psc is generated when the sublevel configuration satisfies the resonance condition @xmath112 together with that @xmath113 and @xmath96 have opposite signs . a positive
psc requires that @xmath114 while a negative psc needs @xmath115 , as indicated clearly in figures [ fig1](c ) and [ fig1](d ) . in case
ii , the visible areas of the psc islands are enlarged but their edges become obscure [ figure [ fig2](b ) ] , which is attributed to the blurred fermi distribution of the electron in the pole . as for case iii ,
apart from the variations of the current magnitudes and the island areas , the pscs on the two islands reverse sign simultaneously [ figure [ fig2](c ) ] , as a result of the process in figure [ fig1](e ) dominating the one in figure [ fig1](c ) .
we show in section [ interacting ] that , when a strong intradot coulomb repulsion is considered , there will be two more psc islands in the color maps , which just correspond to the pscs established between the higher two sublevels @xmath43 and @xmath44 when they are adjusted to the vicinity of @xmath32 , as we discussed in section [ intr ] .
figures [ fig2](a)-[fig2](c ) indicate that tuning the sublevel configuration along the black dashed line ( @xmath102 ) in figure [ fig2](a ) is the optimal path for a delicate controlling on the psc . in figure [ fig2](d ) , we present the detailed evolution of the pscs against the temperature difference @xmath116 at three sublevel configurations @xmath117 , @xmath118 , and @xmath119 , as indicated by the black dots aligned on the white dashed line ( @xmath103 ) in figure [ fig2](a ) .
it is observed that the pscs all collapse to zero at @xmath90 and converge to a saturation , which are direct results of the third analytical insight mentioned above .
note that the emergence of a common saturation arises from the identical zeeman splitting of the three selected configurations .
furthermore , it is evident that the dependence of pscs on the temperature difference is much sensitive at @xmath120 , for the reason that a dramatic change of the fermi distribution near @xmath32 occurs only at low @xmath17 .
the pscs at @xmath117 , @xmath118 show up qualitatively different behaviors from the ones at @xmath119 .
this is attributed to the subtle competitions between the two factors , i.e. , sublevel configuration in qd and the electron temperature , as depicted in figures [ fig1](c)-[fig1](e ) . as a function of the temperature @xmath0 at three sublevel configurations
@xmath117 , @xmath118 , and @xmath119 as indicated in figure [ fig2](a ) . here @xmath107 .
( b ) effect of the tunneling rate @xmath121 on the kinetic coefficient @xmath122 at the sublevel configuration a. ] now we discuss the dependence of psc on the tunneling rate . in figure [ fig3](a ) , the evolution of psc at @xmath31 against the specific sublevel configurations indicated by the black dashed lines in figure [ fig2](a ) is traced for various @xmath121 .
as we can see , a strong positive ( negative ) psc resonant peak is formed at @xmath38 ( @xmath123 ) with its full width at half maximum being @xmath121 .
the broadenings of resonances implicate a tolerance allowed by the resonance condition equation ( [ rc ] ) since each of the sublevels is effectively broadened by @xmath121 . as @xmath121
is increased , the absolute maximums of psc are reduced , which can be traced back to the suppressed local density of states @xmath124 involved in the function @xmath125 . physically , the reduction of psc is due to that the broadenings of sublevels diminish effectively the zeeman splitting that is important in the resonant spin flip mechanism . on the other hand , for a vanishingly small @xmath121 the local density of states @xmath124 will reduce to the dirac-@xmath126 function such that the integral in equation ( [ eq6 ] ) is divergent
this must be incorrect since no electron current and thus no psc can be set up without a tunneling coupling .
we note that the @xmath127 limit should not be taken into account here , since the perturbation theory we performed is on the parameter @xmath29 , which is considered as the smallest energy scale , instead of @xmath121 .
however , thorough insights on the role played by @xmath121 is beyond the scope of present work . in figure [ fig3](b ) , similar dependence on the tunneling rate is shown for the pscs along another sublevel configurations indicated by the white dashed line in figure [ fig2](a ) .
we would like to mention the thermoelectric psc in the linear response regime .
to this end , one can expand the current formula equation ( [ eq6 ] ) to the first order in @xmath116 at fixed @xmath0 ( @xmath128 ) and using the relation @xmath85 to obtain @xmath129 , with the kinetic coefficient ( see [ app ] ) @xmath130,\label{eq12}\ ] ] where @xmath131 represents the thermal phonon number @xmath132 at @xmath133 . in figure [ fig4](a ) ,
the dependence of kinetic coefficient on the temperature is plotted at three sublevel configurations @xmath117 , @xmath118 , and @xmath119 indicated in figure [ fig2](a ) .
one sees that the kinetic coefficient is non - monotonic in the temperature : it increases quickly to the maximum at @xmath0 about @xmath134 ( i.e. , @xmath135 ) , and then decreases more and more slowly as @xmath0 increases . particularly , it vanishes at @xmath136 , since then no thermal phonons is available .
in addition , comparing the maximums of curve @xmath117 , @xmath118 , and @xmath119 , the global optimal kinetic coefficient is predicted to be achieved at the sublevel configurations @xmath137 . in figure
[ fig4](b ) , it is shown that the kinetic coefficient is suppressed as @xmath121 increases , which is due to the same reason for the suppression of psc at large tunneling rate , as mentioned above .
now we turn to validate our discussion in section [ intr ] that the inclusion of a strong coulomb interaction in the cnt qd will not disturb the substantial physical scenario , provided that only the coulomb blockade effect survives at weak tunnel coupling or at high enough temperatures . for this purpose
, we would like to calculate the psc as a function of the sublevels in the presence of a nonzero coulomb interaction .
however , for an interacting qd , the generalized self - consistent born approximation we employed in section [ noninteracting ] is impracticable since the wick s theorem can only be applied to a quadratic unperturbed hamiltonian @xcite .
therefore , we resort to the simple but useful master equation method @xcite to incorporate the electron - electron correlations , which is reliable for @xmath138 . to proceed , we start with dividing the total hamiltonian into two parts as @xmath139 , where @xmath140 and @xmath141 . in the @xmath142 term , the pole , the vibrational mode , and the qd are independent of each other and thus they are in respective thermal equilibrium states .
it is the @xmath143 term which couples the qd to the pole and to the vibrational mode that makes the electronic transport between the qd and the pole possible .
we assume that both the tunnel coupling and the svi are so weak that the @xmath143 term can be treated as a perturbation within the framework of the master equation method . here
, we will restrict the calculations to lowest nonvanishing order of @xmath143 , which has been shown to describe the coulomb blockade quite accurately in an interacting qd connecting two poles but without the svi @xcite .
the steady - state occupation probabilities @xmath144 , @xmath145 , for the qd states are determined by the master equations @xmath146 together with the normalization condition @xmath147 .
the rates for tunneling - induced transition between the states are obtained from the generalized fermi s golden rule @xcite as @xmath148 , @xmath149 , @xmath150 $ ] , and @xmath151 $ ] .
along the same line , one can derive the svi - induced spin - flip rates as @xmath152\nonumber\\ & & \hspace{-1cm}+2\pi\hbar^{-1}\sum_{n=0}^{\infty}\rho_{n}\vert \langle n\vert
\langle \bar\sigma \vert a h_{int}\vert \sigma\rangle \vert n\rangle \vert ^{2}\delta [ \varepsilon_{d\sigma}-(\varepsilon_{d\bar{\sigma}}+\varepsilon_{ph } ) ] \nonumber\\ & & \hspace{-1.4cm}=2\pi\hbar^{-1}\lambda^{2}\sum_{n=0}^{\infty}\rho_{n}[n^{2}\delta(\varepsilon_{d\sigma}+\varepsilon_{ph } -\varepsilon_{d\bar{\sigma}})+(1+n)^{2}\delta ( \varepsilon_{d\sigma}-\varepsilon_{d\bar{\sigma}}-\varepsilon_{ph } ) ] \nonumber\\ & & \hspace{-1.4cm}=2\pi\hbar^{-1}\lambda^{2}(1 + 2n_{ph})[n_{ph}\delta(\varepsilon_{d\sigma}-\varepsilon_{d\bar{\sigma}}+\varepsilon_{ph})+(n_{ph}+1)\delta(\varepsilon_{d\sigma}-\varepsilon_{d\bar{\sigma}}-\varepsilon_{ph } ) ] , \end{aligned}\ ] ] where @xmath153 denotes the fock state occupied by @xmath27 phonons and @xmath154 is its weight factor with @xmath155 being the partition function . for the convenient of practical calculations , we would like to replace the @xmath156 function by a lorentztian function @xmath157 with a small width @xmath158 . from the solution to the master equations , the spin - dependent electron current flowing through pole into the qd is obtained as @xmath159 by the numerical calculations , we confirm that the charge current conservation law @xmath160 ( therefore , @xmath85 ) and that the charge current vanishes at @xmath161 are both respected within the master equation method . and @xmath96 in the case @xmath162 ( @xmath163 , @xmath108 ) with different coulomb interaction ( a ) @xmath164 , ( b ) @xmath165 , ( c ) @xmath166 , and ( d ) @xmath167 .
other parameters are @xmath105 , @xmath168 , and @xmath169 . ] in figure [ figapp1 ] , we show how the coulomb interaction affect the dependence of pscs on the sublevels in the case @xmath162 . in the absence of the coulomb interaction [ figure [ figapp1](a ) ] , there exists two psc islands which is consistent with figures [ fig2](a ) and [ fig2](b ) . for a weak but nonzero @xmath42 [ figure [ figapp1](b ) ] ,
the psc islands are stretched and the absolute pscs are suppressed at the same time . once @xmath42 becomes larger than the phonon energy [ figures [ figapp1](c ) and [ figapp1](d ) ] , each psc island is divided into two isolate islands .
the emergent islands just correspond to the pscs established between the two sublevels @xmath43 and @xmath44 when they are adjusted to the vicinity of @xmath32 , as we discussed in section [ intr ] .
we further explain this scenario in detail , for example , under the sublevel configuration @xmath170 .
when the sublevels are off resonance ( i.e. , @xmath171 ) , the qd is almost occupied by one spin - up electron and the coulomb repulsion prohibits the pole electrons from entering the qd .
therefore no psc flows in the pole . on the other hand ,
when the sublevels are on resonance ( i.e. , @xmath38 ) , the trapped spin - up qd electron has the possibility to reverse its spin and transit to the spin - down sublevel by absorbing one phonon , which then allows a spin - up pole electron with energy @xmath43 to enter the qd .
subsequently , the injected qd electron can absorb one phonon and transit to the sublevel @xmath44 .
finally , it tunnels easily to the pole since @xmath172 .
theses combined tunneling processes are responsible for the emergent positive psc island .
one can also notice that the absolute maximum of the psc in figure [ figapp1](d ) is roughly reduced by half in comparison to the one in figure [ figapp1](a ) , which could be explained as follows .
we first focus on the two psc islands at the top right corner in figure [ figapp1](d ) .
as illustrated in figure [ fig1](c ) for @xmath164 , when a spin - up electron at the lower sublevel transits to the upper sublevel by absorbing a phonon , another spin - up pole electron can tunnel into the lower sublevel immediately .
however , for a strong coulomb repulsion , the latter process is blockaded due to the presence of the spin - down qd electron at the upper sublevel .
only if that qd electron tunneling out to the pole will allow a spin - up pole electron to enter the qd .
mathematically , the reduction of psc is attributed to the fact that the second term in equation ( [ recurrent ] ) which contributes equal as the first term to the total current in the @xmath164 case vanishes for a strong @xmath42 , while the last two terms in equation ( [ recurrent ] ) are always vanishingly small for the parameters used in figure [ figapp1 ] . as for the two emergent psc islands ,
the relatively smaller current amplitude is due to the small empty state occupation probability @xmath173 .
figure [ figapp2 ] shows the dependence of pscs on the sublevel configuration in the case @xmath174 with finite @xmath42 .
in contrast to figure [ figapp1 ] , the pscs reverse direction as the electron temperature varies from @xmath162 to @xmath174 , which is in agreement with the @xmath164 case as shown in figure [ fig2 ] . in figure [ figapp2](a ) ,
the two emergent psc islands induced by the coulomb interaction are merged with the original islands due to the smeared fermi distribution around @xmath32 at high @xmath17 .
however , they manifest themselves at a larger @xmath42 , as shown in figure [ figapp2](b ) . and @xmath96 in the case @xmath174 ( @xmath175 , @xmath108 ) with different coulomb interaction ( a ) @xmath167 and ( b ) @xmath176 .
other parameters are @xmath105 , @xmath168 , and @xmath169 . ]
in conclusion , we have found that a qd formed in a suspended cnt exposed to an external magnetic field could act as a thermoelectric unipolar sb which generates psc . in this setup ,
the spin flip source is natural due to the interplay between the intrinsic spin - orbit coupling and the vibrational modes of the suspended cnt , rather than the previous ones induced by the somewhat complicated time - varying external fields @xcite .
moreover , the driving force of this sb setup is a consequence of the nontrivial thermoelectric effect with respect to the fermi ( pole ) and bose ( thermal bath ) reservoirs .
the magnitude and the direction of the generated psc are dependent on such four factors as i ) the strength of svi , ii ) the sublevel configuration in qd , iii ) the electron ( @xmath17 ) and bath ( @xmath34 ) temperatures , and iv ) the tunnelling rate between the qd and the pole .
in particular , for finite temperature difference between the pole and the thermal bath , a joint adjustment on the sublevel configuration and the tunneling rate suffices the delicate controlling on the psc . on the experimental aspect , the sublevels in a cnt qd
is finely tunable nowadays by the interplay of a gate voltage and an external applied magnetic field @xcite , and the tunneling rate can also be conveniently regulated by a gated tunneling barrier .
in addition , in the linear response regime , it is found that the kinetic coefficient is non - monotonic in the temperature @xmath0 and it reaches its maximum when @xmath177 is about one phonon energy .
we have also demonstrated that the existence of a strong intradot coulomb interaction is irrelevant for our sb , provided that high - order cotunneling processes are suppressed .
obviously , the sb setup we addressed in this work explicitly indicates a potential application of the versatile cnts .
we hope that our results could be helpful for obtaining controllable psc , which plays a significant role in spintronics .
this work is supported by nsfc ( grants no.11325417 and no.11674139 ) and pcsirt ( grant no.irt1251 ) .
in the spin - dependent current formula equation ( [ eq6 ] ) , while @xmath125 is independent of the temperatures @xmath88 is actually a function of @xmath17 and @xmath30 as @xmath178 [ f(\omega_{-},t_e)-1 ] f(\omega_{+},t_e)-n_{ph}(t_{b})f(\omega_{-},t_e)[f(\omega_{+},t_e)-1].\label{eqa1}\ ] ] in the linear response regime , we keep the taylor expansion of @xmath88 to the first order in the temperature difference @xmath179 , with @xmath128 denotes the fixed electron temperature , as @xmath180 notice the identity @xmath181 [ f(\omega_{-},t)-1]f(\omega_{+},t)=n_{ph}(t)f(\omega_{-},t)[f(\omega_{+},t)-1],\label{eqa3}\ ] ] the first term in equation ( [ eqa2 ] ) vanishes exactly .
the second term in equation ( [ eqa2 ] ) reads @xmath182f(\omega_{+},t)-f(\omega_{-},t)[f(\omega_{+},t)-1]\}\delta t\nonumber\\ & & \hspace{-0.5cm}=\frac{\varepsilon_{ph}n_{ph}(t)}{k_{b}t^{2}}\{[1+n_{ph}(t)][f(\omega_{-},t)-1]f(\omega_{+},t)\nonumber\\ & & \hspace{2.5cm}-[1+n_{ph}(t)]f(\omega_{-},t)[f(\omega_{+},t)-1]\}\delta t,\label{eqa4}\end{aligned}\ ] ] applying equation ( [ eqa3 ] ) to the first term in the brace of equation ( [ eqa4 ] ) one immediately obtains @xmath183\delta t.\label{eqa5}\ ] ] collecting equations ( [ eqa2 ] ) and ( [ eqa5 ] ) into equation ( [ eq6 ] ) and using the relation @xmath85 one finally obtains the equation ( [ eq12 ] ) in the main text .
10 wolf s a , awschalom d d , buhrman r a , daughton j m , von molnr s , roukes m l , chtchelkanova a y and treger d m 2001 * 294 * 1488 uti i , fabian j and sarma s das 2004 * 76 * 323 sun q - f , guo h and wang j 2003 * 90 * 258301 long w , sun q - f , guo h and wang j 2003 * 83 * 1397 brataas a , tserkovnyak y , bauer g e w and halperin b i 2002 * 66 * 060404(r ) wang d - k , sun q - f and guo h 2004 * 69 * 205312 pareek t p 2004 * 92 * 076601 chen z , wang b , xing d y and wang j 2004 * 85 * 2553 wang j , chan k s and xing d y 2005 * 72 * 115311 l h - f and guo y. 2007 * 91 * 092128 nazarov y v 2007 * 9 * 352 futterer d , governale m and knig j 2010 * 91 * 47004 wysokiski k i 2012 * 24 * 335303 zhang p , xue q - k and xie x c 2003 * 91 * 196602 wang b , wang j and guo h 2003 * 67 * 092408 rudner m s and rashba e i 2010 * 81 * 125426 jarillo - herrero p , sapmaz s , dekker c , kouwenhoven l p and van der zant h s j 2004 * 429 * 389 kuemmeth f , ilani s ,
ralph d c and mceuen p l 2008 * 452 * 448 fang t - f , zuo w and luo h - g 2008 * 101 * 246805 churchill h o h , kuemmeth f , harlow j w , bestwick a j , rashba e i , flensberg k , stwertka c h , taychatanapat t , watson s k and marcus c m 2009 * 102 * 166802 ohm c , stampfer c , splettstoesser j and wegewijs m r 2012 * 100 * 143103 plyi a , struck p r , rudner m , flensberg k and burkard g 2012 * 108 * 206811 stadler p , belzig w and rastelli g 2014 * 113 * 047201 stadler p , belzig w and rastelli g 2015 * 91 * 085432 steele g , pei f , laird e , jol j , meerwaldt h and kouwenhoven l 2013 * 4 * 1573 galperin m , ratner m a and nitzan a 2007 * 19 * 103201 zimbovskaya n a and pederson m r 2011 * 509 * 1 entin - wohlman o , imry y and aharony a 2010 * 82 * 115314 lazzeri m , piscanec s , mauri f , ferrari a c and robertson j 2005 * 95 * 236802 oron - carl m and krupke r 2008 * 100 * 127401 berciaud s , han m y , mak k f , brus l e , kim p and heinz t f 2010 * 104 * 227401 zippilli s , morigi g and bachtold a 2009 * 102 * 096804 dubi y and di ventra m 2011 * 83 * 131 fang t - f , sun q - f and luo h - g 2011 * 84 * 155417 meir y , wingreen n s and lee p a 1991 * 66 * 3048 de franceschi s , sasaki s , elzerman j m , van der wiel w g , tarucha s and kouwenhoven l p 2001 * 86 * 878 hewson a c , 1993 ( cambridge : cambridge university press ) haug h and jauho a - p 2008 2nd edn ( berlin : springer ) wirkowicz r , wierzbicki m and barna j 2009 * 80 * 195409 viljas j k , cuevas j c , pauly f and hfner m 2005 * 72 * 245415 frederiksen t , paulsson m , brandbyge m and jauho a - p 2007 * 75 * 205413 mahan g d 2000 ( new york : kluwer academic / plenum publishers ) bruus h and flensberg k 2004 ( new york : oxford university press ) averin d v , korotkov a n and likharev k k 1991 * 44 * 6199 bonet e , deshmukh m m and ralph d c 2002 * 65 * 045317 muralidharan b , ghosh a w and datta s 2006 * 73 * 155410 | a quantum dot formed in a suspended carbon nanotube exposed to an external magnetic field is predicted to act as a thermoelectric unipolar spin battery which generates pure spin current .
the built - in spin flip mechanism is a consequence of the spin - vibration interaction resulting from the interplay between the intrinsic spin - orbit coupling and the vibrational modes of the suspended carbon nanotube . on the other hand , utilizing thermoelectric effect , the temperature difference between the electron and the thermal bath to which the vibrational modes are coupled provides the driving force .
we find that both magnitude and direction of the generated pure spin current are dependent on the strength of spin - vibration interaction , the sublevel configuration in dot , the temperatures of electron and thermal bath , and the tunneling rate between the dot and the pole .
moreover , in the linear response regime , the kinetic coefficient is non - monotonic in the temperature @xmath0 and it reaches its maximum when @xmath1 is about one phonon energy .
the existence of a strong intradot coulomb interaction is irrelevant for our spin battery , provided that high - order cotunneling processes are suppressed . |
there has been mounting evidence that most of the matter in the universe is not luminous but dark .
current observations such as the cosmic microwave background ( cmb ) suggest that the dark matter ( dm ) makes up about 25 % of the universe @xcite .
the dark matter is usually assumed to be a collisionless cold component ( non - relativistic at the time of freeze - out ) , called as the cold dark matter ( cdm ) .
weakly - interacting massive particles ( wimps ) , such as the lightest neutralino in the supersymmetric standard model ( susy sm ) @xcite , and the lightest kaluza - klein particle in the universal extra dimension ( ued ) @xcite , are the popular cdm candidates .
the predicted thermal relic abundances and the large - scale structure ( @xmath6 mpc ) are in good agreement with the observed values . however , the recent high - resolution @xmath7-body simulations on the cdm - based structure formation revealed various discrepancies on smaller scales ( @xmath8 mpc ) .
the first one is so - called the `` missing satellite problem '' @xcite : the @xmath7-body simulations of the cdm particles predict significantly more virialized dark objects with mass @xmath9 ( or subhalos ) in the galaxy - sized halos with @xmath10 than those observed around the milky way .
the other one is called the `` cusp problem '' @xcite : the cdm - based models also predict a cuspy profile for mass density distributions for the cdm halos @xcite although the measurements of the rotation curves imply the presence of cores in the centers of the halo .
although these discrepancies may be circumvented by some baryonic processes @xcite , it may be worthwhile to consider the other kinds of dm particles with different clustering properties .
for instance , the superweakly interacting massive particles ( superwimps ) @xcite or the warm dark matter ( wdm ) @xcite particles can have large velocity dispersion at the epoch of radiation - matter equality . if the dm consists of the superwimps or the wdm particles , the number of less massive @xmath11 subhalos is significantly reduced , and the cusp formation is also suppressed because the primordial fluctuations at the small scales ( @xmath8 mpc ) are damped . in this note , we consider the possibilities of probing such subhalos with @xmath9 via strong lensing in which the image separations are on submilliarcsecond scales , called as `` submillilensing '' .
recently it has been pointed out that the future submillilensing observations of multiply - imaged qso - galaxy lens systems can directly probe the mass scale of subhalos in the parent galaxy halo @xcite .
the submillilensing observation might resolve whether above problems originate from various baryonic contrivances or the nature of the dm particles . in section 2 , we begin with the discussion of submillilensing and consider the possibility of detecting the small - mass subhalos via submillilensing in the next decade . in section 3 ,
we study its implications to the superwimp and the wdm scenarios .
section 4 is devoted to conclusions and discussion .
in what follows , we discuss the possibilities of direct detection of subhalos with a mass of @xmath12 via substructure lensing which is defined as lensing by @xmath13 subhalos that perturb a `` simple '' strong lensing by @xmath14 parent galaxy halo . to date , about @xmath15 quadruply - imaged gravitational lenses with flux ratio `` anomalies '' have been detected @xcite . here
`` anomaly '' refers to an observed image flux ratio that does not agree with the ratio predicted by standard macrolens models with a smooth gravitational potential . from the radio and the mid - infrared observation ,
some of those lens systems are found to be consistent with the model in which the macrolens with a smooth potential is perturbed by subhalos @xcite .
unfortunately , the mass scale of the subhalos has not been yet determined well .
first , we theoretically estimate the mass range of the subhalos that can perturb the fluxes of the multiple images produced by the parent galaxy halo with a mass of @xmath16 . in what follows , for simplicity , we assume that the low - mass subhalos are described by tidally - cut singular isothermal spheres ( siss ) with a mass function @xmath17 , where @xmath18 . at a distance @xmath19 from the center ,
an sis with a one - dimensional velocity dispersion @xmath20 has a density profile @xmath21 @xcite .
if we further assume that the parent galaxy halo is also described by an sis with a one - dimensional velocity dispersion @xmath22 , then the tidal radius @xmath23 of the sis subhalo with @xmath20 at a distance @xmath19 from the center of the parent halo is approximately given by @xmath24 , which yields the mass of the tidally - cut sis as @xmath25 .
the effects of deviation from the assumption made here will be discussed later .
the lensing cross section by a subhalo is proportional to the square of the einstein angular radius @xmath26 , and the radius for the sis is proportional to @xmath27 @xcite . using the equations for the mass function and the tidal radius , the substructure lensing cross section per logarithmic mass interval is given by @xmath28 where we have assumed that the distance @xmath19 of the subhalo to the halo center is approximately equivalent to the einstein radius of the parent halo @xmath29 .
this approximation is verified because the substructure lensing cross section depends weakly on the distance to the center @xmath19 as @xmath30 .
the contribution from the distant halos in the line of sight can boost the cross section by a factor @xcite .
thus , the contribution of massive subhalos is significant in the substructure lensing as long as the mass function satisfies @xmath31 . assuming that the mass function with @xmath32 , observed for the massive scales @xmath33 @xcite , is applicable to the mass scale @xmath34 , the lensing probability by an intermediate - mass subhalo @xmath2 is reduced by @xmath35 in comparison with the probability by a massive subhalo @xmath36 .
if the subhalos have a different mass profile such as the nfw profile @xcite , the contribution of less massive subhalos can be further reduced . because the ratio between the einstein radius of the perturber and the tidal radius at a fixed distance @xmath19 decreases as the subhalo mass decreases , the lensing cross section for a less massive subhalo is susceptible to the inner less cuspy profile .
thus , we expect that the contribution from @xmath37 subhalos is negligible in alternating the amplitude of the flux of multiple images .
next , we explore the possibility of measuring the mass scale of these subhalos from astrometric shifts of the multiply lensed images . for siss , the order of the magnitude of the astrometric shifts is typically given by the size of the angular einstein radius @xmath26 @xcite .
for an sis with one - dimensional velocity dispersion @xmath20 at @xmath38gpc , @xmath26 is approximately given by @xmath39 thus , observation with angular resolution of submilliarcsecond scales ( @xmath40mas ) , which will be achieved in the next generation satellite vlbi mission such as the vsop2 @xcite , can reveal subhalos with one - dimensional velocity dispersion @xmath41 km s@xmath42 .
it corresponds to @xmath43 at the distance equal to the einstein radius @xmath44 of the macrolens , assuming that the velocity dispersion of the parent halo is @xmath45 . from the astrometric shifts of the multiple extended images perturbed locally by a subhalo with respect to an unperturbed macrolensed image , we can break the degeneracy between the subhalo mass and the distance in the line of sight to the images if resolved at scale of an einstein radius of the perturber @xcite .
this is of great importance because otherwise we can not determine whether the flux ratio anomaly is caused by more massive intergalactic halos in the line of sight or by less massive subhalos within the macrolens halo .
even if the density profile of the perturber is shallower than an sis , we can make a distinction between models with different density profiles from astrometric shifts of the surface brightness profile within the source @xcite
. a direct detection of less massive @xmath46 subhalos within the parent halo will give a stringent constraint on the superwimp and the wdm scenarios , which will be discussed in the next section .
there exist many well motivated models for superwimps and wdm from particle physics .
the natural candidates for the superwimps are gravitino and axino , that are superpartners of graviton and axion , respectively @xcite .
the right - handed sneutrino is also the candidate when neutrino masses are dirac - type @xcite .
others are kaluza - klein graviton and axion states in the ued @xcite . as for the wdm ,
a light gravitino and sterile neutrinos have been discussed as such candidates .
thus , it is interesting to study the feasibility of probing or constraining those models in the near future experiments . in the following ,
we refer `` superwimps '' as the particles whose interactions are weaker than the weak interaction , such as the gravitino which couples to other particles only gravitationally .
superwimps are also assumed to be produced by the decays of heavier particles whose interactions are weak ( _ e.g. _ wimps ) . in the cdm - based structure formation models , the structure of the universe forms in a hierarchical manner .
protohalos , which are the first virialized objects , appear first after the mass density fluctuation becomes nonlinear , and larger objects form successively via their merger . in the ordinary wimp models ,
the comoving damping scale of the power spectrum is typically @xmath47pc , depending on the hubble radius at the kinetic decoupling temperature @xcite , and the protohalo mass is @xmath48 @xcite . in the superwimp or wdm scenarios , the primordial fluctuations on even larger scales can be damped because they have large velocity dispersion at the decoupling .
therefore , the protohalos become massive , and the total amount of subhalo mass inside the parent halo can be significantly reduced .
when the comoving damping scale of the dm power spectrum is @xmath49 , the protohalo mass is roughly given by @xmath50 here , the present matter density @xmath51 and the hubble constant @xmath52 are assumed @xcite . in the previous section
, we showed that the sensitivity of direct detection of subhalos inside the parent halo with submillilensing may reach intermediate mass @xmath53 scales .
thus , when @xmath54 kpc , the future submillilensing experiments may directly detect such protohalos .
first , we discuss the damping scale in the superwimp scenario and compare it with the sensitivities of future submillilensing . in the scenario , some wimps freeze out from the thermal equilibrium as the usual wimp dm models , and superwimps are non - thermally produced from the wimp decay , since the superwimps interact superweakly with the thermal bath .
the scenario retains the property of the ordinary wimps that the observed relic density is naturally achieved .
furthermore , they are produced by decay of some long - lived massive particle @xmath55 .
since they can have large velocities at the production epoch , the free - streaming damps the small - scale inhomogeneities .
the superwimp scenario is a possible explanation for the small - scale structure @xcite .
the comoving damping scale is typically given by the free - streaming length of the superwimp , @xmath56 , at the matter - radiation equality @xmath57 .
when the long - lived particle has lifetime @xmath58 and the superwimp is produced with three - momentum normalized by its mass @xmath59 , the comoving free - streaming scale @xmath56 is given by @xmath60\end{aligned}\ ] ] where @xmath61 and @xmath62 is the scale factor as a function of @xmath63 . when the superwimps are produced from decay of electrically - charged particlessec )
lead to overproduction of @xmath64li in the big bang nucleosynthesis due to the catalytic enhancement of the production . on the other hand
, it is argued in ref .
@xcite that there are a lot of ambiguities in the derivation . ] , the small - scale power may be further suppressed @xcite .
the charged particles are coupled to the photon - baryon fluid which oscillates on subhorizon scale .
if the scale in question enters the horizon before the decay , the density fluctuation of such dms can not grow but oscillates , which gives a rise to the suppression on small scales .
the damping scale is typically given by the hubble radius at the decay time , @xmath65 , @xmath66 in fig .
1 we show the damping scales @xmath49 of the charged and neutral @xmath55 cases as functions of @xmath58 and @xmath67 .
when @xmath55 is neutral , the damping scale is determined by the free streaming scale @xmath56 .
longer lifetime @xmath58 implies a larger free streaming scale , since the velocity dispersion at the radiation - matter equality time , which is proportional to @xmath68 , becomes larger .
when @xmath55 is charged , the damping scale @xmath49 corresponds to the larger one of @xmath56 and @xmath69 . for @xmath70 , @xmath49
is given by @xmath69 , since @xmath56 is @xmath71 .
the hatched region with damping scales larger than @xmath72 mpc is constrained from lyman alpha clouds @xcite .
the future submillilensing experiments may cover regions above the bold lines , which correspond to @xmath73 kpc .
when the @xmath55 lifetime is longer than @xmath72 sec and the superwimps are produced with relativistic momentum @xmath74 , the damping scale is larger than @xmath72 kpc , which may be constrained by future observation if the subhalos with @xmath75 were discovered .
when @xmath55 is charged , the region with @xmath76 sec may be also covered even in the small @xmath67 cases .
= 0.6 = 0.6 following ref .
@xcite , we indicated the parameter region which is suitable to solve the discrepancies in small - scale structure in fig . 1 .
the region with @xmath77 mpc is suitable to solve the `` missing satellite problem '' , while the gray region with @xmath78 is favored to solve the `` cusp problem '' , which requires a large dm velocity dispersion @xcite . as we can see in fig . 1 ,
the model parameters corresponding to these regions can be well constrained by the future submillilensing observation . for the illustrative purpose , we consider a superwimp gravitino model , in which the gravitinos are produced by slepton or sneutrino decay .
sleptons or sneutrinos are assumed to be the lightest susy particles in the susy sm .
the slepton and sneutrino lifetimes are given by @xcite @xmath79 where @xmath80 gev , @xmath81 is the gravitino mass , and @xmath82 is the mass of the parent wimp @xmath55 . in fig . 2 , @xmath56 ( solid lines ) is shown as a function of @xmath81 and the mass difference between the gravitino and the parent particle ( @xmath83 ) . dashed lines indicate the damping scale for which @xmath84 .
the gray region which corresponds to explain the `` cusp problem '' @xcite may be excluded when the future submillilensing experiments find subhalos with mass smaller than @xmath85 .
it is known that the lifetime and the mass of the long - lived particles whose decay produces the superwimps , are constrained by the big bang nucleosynthesis ( bbn ) and by the cmb planckian spectrum , depending on the decay channels .
when the hadronic modes are dominant in the decay , the bbn constrains the lifetime to be shorter than @xmath72 sec @xcite .
even if the hadronic shower is suppressed , the electromagnetic energy injection from the long - lived particle decay to the thermal bath is also constrained from the bbn while the constraint is weaker .
in addition to it , the cmb planckian spectrum also gives a constraint when the lifetime is longer than @xmath86 sec @xcite .
the future submillilensing experiments are complementary to those constraints , since the damping scale @xmath49 is independent of the decay channels . for example , when the superwimp gravitinos are produced from sneutrino decay , the energy injection to thermal bath is very tiny so that the constraints from the bbn and the cmb planckian spectrum are very weak .
even in such a case , the submillilensing will still constrain the model . = 0.6 next , we discuss the damping scale in the wdm models . among many wdm candidates discussed so far ,
the light gravitino with mass of order from @xmath87 up to @xmath88 , which is likely to be the lightest susy particle in gauge - mediated models of supersymmetry breaking , is one of the well - motivated models from the viewpoint of particle physics @xcite .
such gravitinos are in thermal equilibrium at early times but decouples when the degrees of freedom @xmath89 is @xmath90 where @xmath91 is the decoupling temperature . because the decoupling temperature of such species is higher than that of ( active ) massive neutrinos which play roles of hot dark matter , their velocity dispersion is not so large compared to that of massive neutrinos but non - negligible at the time of structure formation .
thus , they can act as the wdm .
the comoving damping scale for the free - streaming for such gravitinos or any other thermal relic can be written as @xmath92 where @xmath93 is the mean momentum over the temperature .
for thermally decoupled species , this factor gives almost unity , _ i.e. _ , @xmath94 .
the requirement from lyman alpha clouds , @xmath95 , implies @xmath96 . on the other hand ,
the energy density of wdm can be written as @xmath97 .
assuming @xmath98 , the mass of wdm should be @xmath99 kev even for @xmath100 .
we need to introduce more extra degree of freedom around @xmath91 as @xmath101 . when the constraint on @xmath56 is improved to be @xmath8kpc
, the lower bound for the mass can reach @xmath102 mev for @xmath100 .
thus , the wdm scenario may face further difficulties if future submillilensing experiments would find small - mass subhalos .
another well - motivated candidate for wdm is the sterile neutrinos @xcite . because they directly couple to the active neutrinos alone , they can be produced via neutrino oscillation .
although the evaluation of their energy density requires a numerical integration of the boltzmann equation , some useful fitting formulae are available .
the present energy density of sterile neutrinos can be written as @xcite @xmath103 where @xmath104 is the mixing angle between the active neutrinos and sterile neutrinos and @xmath105 is the mass of sterile neutrinos .
the temperature at the time when the production is most efficient is @xcite @xmath106 the sterile neutrinos can damp the small - scale inhomogeneities via the free streaming in the same manner as the thermally decoupled wdm particles do .
the free - streaming scale for a sterile neutrino wdm can also be obtained using eq .
( [ eq : damping_wdm ] ) with different values for @xmath93 from that of the thermally decoupled ones .
because the sterile neutrinos are not in the thermal bath at early times , their distribution function deviates from that of a thermal one and the above factor can be @xmath107 for the standard production mechanism @xcite . although there are some differences between the thermally decoupled wdm and the sterile neutrino wdm , they give the same predictions for the damping of matter power spectrum by identifying their masses as @xcite @xmath108 where @xmath109 is the mass of thermally decoupled relics such as a light gravitino , which is denoted as @xmath110 in eq .
( [ eq : damping_wdm ] ) . because the shape of matter power spectrum is determined by the ratio @xmath111 and the density parameter @xmath112 .
accordingly , the constraint on the sterile neutrino mass bound would be different from that on the thermally decoupled wdm mass boundthe properties of the sterile neutrinos can also be constrained from the measurement of the x - ray flux , since they can contribute to the x - ray flux due to radiative decay @xcite .
see also refs .
@xcite .. using the above formula and assuming the damping scale as @xmath113 kpc which can be reached by future submillilensing experiments , we can expect that the mass of sterile neutrinos can be constrained to be @xmath114 mev , which will be in conflict even with the current constraint @xmath115 kev @xcite .
thus , we can obtain much more insight on the robustness of these scenarios from the future submillilensing experiments , as well as for the superwimp scenarios . here
some comments on the mixed dark matter scenario are in order .
it is possible that , for example , gravitinos are produced not only from the decay of the next - lightest supersymmetric particles ( nlsp ) but also from thermal plasma . in this case
, the present dm is composed of cdm and superwimp .
in such mixed dm scenarios , the damping of the small - scale structure is less significant in comparison with the models in which the superwimps make up all of the dm .
to what extent the amplitude at small scales can be reduced depends on the ratio of the energy density of cdm and superwimp dm .
detailed discussion on this issue is beyond the scope of this letter .
some discussions on the matter power spectrum in such mixed models can be found in refs .
@xcite and @xcite . @xcite analyzed models in which the superwimps are produced from the charged particle decay while @xcite considered models in which the superwimps are produced from the neutral particle decay .
we have shown that future observation of multiply - imaged qso - galaxy lens systems with a high angular resolution @xmath40 mas will prove the small - scale clustering properties of the dm halos down to @xmath116 .
the presence of @xmath53 subhalos implies the comoving damping scale of the primordial fluctuations @xmath117 kpc assuming that the observed subhalos retain the original mass during the merger process .
the superwimp and the warm dm scenarios that predict a larger damping scale @xmath118 in comparison with @xmath47 pc in the ordinary wimp scenarios would be strongly constrained if the presence of such subhalos were proved . in the superwimp scenario ,
the superwimp dm is produced from the long - lived particle @xmath55 decay , and free - streaming of the superwimps damps the small - scale inhomogeneities . when the lifetime of @xmath55 is longer than @xmath72sec , the damping scale is larger than @xmath72 kpc unless the superwimp and @xmath55 masses are degenerate and the superwimps are non - relativistic at the production epoch .
in addition , when @xmath55 is a charged particle , it is coupled to the oscillating photon - baryon fluid before the decay .
therefore , the small - scale inhomogeneities inside the sound horizon can not grow .
the damping scale becomes larger than @xmath72 kpc@xmath119 in the case of charged @xmath55 .
one of the natural superwimp candidates is the gravitino produced from the slepton or sneutrino . in this case , the damping scale is typically larger than @xmath120 kpc .
the future submillilensing experiments , which cover the dm subhalos with mass @xmath121 , are important tests for probing such superwimp scenarios . for the wdm scenarios , such as the light gravitino and the sterile neutrinos
, the wdm mass would be further constrained .
the survivability of the protohalos during the merger process is under debate now . in the ordinary wimp scenarios ,
it is claimed that most of the earth - mass protohalos are stable against the tidal stripping @xcite .
it is also discussed whether those halos are disrupted by interaction with stars @xcite . on the one hand , subhalos that cross the galactic disk nearly perpendicularly or that fall off to the center of the parent galaxy are strongly disrupted by the tidal force , leading to a significant decrease in the total mass . on the other hand , subhalos that orbit on the plane nearly parallel to the disk or that reside in the low density region survive more - or - less intact . for simplicity , we have assumed that the observed mass scale of the subhalo is equivalent to the lowerbound on the protohalo mass scale . in practice , however , the observable mass using gravitational lensing is limited to the one within a certain radius centered at the line of sight . as a result , there remains a certain ambiguity in estimating the mass scale of the observed subhalo .
observation of astrometric shifts of lensed qso images with a substructure in the surface brightness may help to reconstruct the subhalo mass density profile , thereby reducing the ambiguity @xcite .
it has been argued that the superwimp and wdm scenarios can resolve the small - scale @xmath122 discrepancies , such as the `` missing satellite problem '' and the `` cusp problem '' if the damping scale is as large as @xmath123 mpc .
therefore , future submillilensing experiments will shed a new light on these small - scale structure problems once the above ambiguities are removed .
we would like to thank masashi chiba for useful comments .
this work is supported in part by the grant - in - aid for science research , ministry of education , science and culture , japan ( no .
18034002 and 15540255 for jh ) .
b. moore , nature * 370 * ( 1994 ) 629 ; r. a. flores and j. r. primack , astrophys .
j. * 427 * ( 1994 ) l1 ; j. j. binney and n. w. evans , mon . not .
* 327 * ( 2001 ) l27 ; a. r. zentner and j. s. bullock , phys .
d * 66 * ( 2002 ) 043003 ; j. d. simon , a. d. bolatto , a. leroy , l. blitz and e. l. gates , astrophys .
j. * 621 * ( 2005 ) 757 .
j. f. navarro , c. s. frenk and s. d. m. white , astrophys . j. * 462 * ( 1996 ) 563 ; astrophys .
j. * 490 * ( 1997 ) 493 ; b. moore , s. ghigna , f. governato , g. lake , t. quinn , j. stadel and p. tozzi , astrophys .
j. * 524 * ( 1999 ) l19 . c. r. watson , j. f. beacom , h. yuksel and t. p. walker , astro - ph/0605424 ; a. boyarsky , a. neronov , o. ruchayskiy and m. shaposhnikov , mon . not .
soc . * 370 * ( 2006 ) 213 ; a. boyarsky , a. neronov , a. neronov , o. ruchayskiy and m. shaposhnikov , astro - ph/0603368 ; a. boyarsky , a. neronov , o. ruchayskiy , m. shaposhnikov and i. tkachev , astro - ph/0603660 . | we propose to observe qso - galaxy strong lens systems to give a new constraint on the damping scale of the initial fluctuations .
we find that the future observation of submilliarc scale astrometric shifts of the multiple lensed images of qsos would find @xmath0 subhalos inside the macrolens halo .
the superweakly interacting massive particles ( superwimps ) produced from a wimp decay and the warm dark matter ( wdm ) particles that predict a comoving damping scale larger than @xmath1 can be constrained if @xmath2 subhalos are detected .
hep - ph/0608126 + august 2006 + ku - tp005 + .5 in * constraining superwimpy and warm subhalos + with future submillilensing * .45 in junji hisano@xmath3 , kaiki taro inoue@xmath4 , and tomo takahashi@xmath5 .45 in _ @xmath3 institute for cosmic ray research , + university of tokyo , kashiwa 277 - 8582 , japan + @xmath4 department of science and engineering , kinki university , + higashi - osaka , 577 - 8502 , japan + @xmath5 department of physics , saga university , saga , 840 - 8502 , japan _
.4 in |
one of the key questions in cosmology today relates to the still - unsolved problem of what is causing the observed cosmic acceleration .
primarily indicated by hubble curves constructed from luminosity distance measurements of type ia supernovae @xcite , this ( possibly apparent ) acceleration is just one aspect of the struggle for a consistent picture of the universe ; a picture that would also require an explanation of the gap between the observed clustering matter content of @xmath2 ( e.g. , * ? ? ?
* ) and the value of @xmath3 as indicated by cosmic microwave background measurements of spatial flatness @xcite , as well as a solution of the `` age problem / crisis '' for matter - only ( i.e. , decelerating ) cosmologies in which the universe appears to be younger than some of its oldest constituents ( e.g. , * ? ? ?
* ) , along with explanations of other important issues .
the standard approach to solving these problems is to introduce some form of `` dark energy '' the simplest case being the cosmological constant , @xmath1 which can fill the gap via @xmath4 , which possesses negative pressure in order to achieve cosmic acceleration ( e.g. , * ? ? ?
* ) , and which ( for non-@xmath1 cases ) recruits some form of internal nonadiabatic pressure ( e.g. , * ? ? ?
* ) in order to avoid clustering as matter does .
thus the introduction of dark energy ( often using spatially - flat @xmath1cdm models ) has led to a broadly - consistent `` cosmic concordance '' an empirical outline which has seemed so far to succeed fairly well ( e.g. , * ? ? ?
* ) at developing into a consistent cosmological picture .
there are serious aesthetic problems with dark energy , however , as is well known ; the most obvious being the problematical introduction of a completely unknown substance as the dominant component of the universe . beyond that , a static ( i.e. , cosmological constant ) form of dark energy suffers from two different fine - tuning problems : one being the issue that @xmath5 is @xmath6@xmath7 orders of magnitude smaller than what would be expected from the planck scale @xcite ; and the other being a coincidence problem " ( e.g. , * ? ? ?
* ) , questioning why observers _ today _ happen to live so near the onset of @xmath1-domination , given @xmath8 .
moving to a dynamically - evolving dark energy ( dde ) , however , invites other problems , since the _ self - attractive _ nature of negative - pressure substances ( i.e. , @xmath9 ) means that a dde may cluster spatially @xcite , possibly ruining it as a `` smoothly - distributed '' cosmic ingredient .
this could potentially be solved through the ad - hoc addition of some form of nonadiabatic support pressure for the dde @xcite , but this is a possibility which we have argued against elsewhere @xcite on thermodynamically - based cosmological grounds .
besides dark energy , various other methods have been used to attempt to explain the observed acceleration , such as employing modified gravity ( e.g. , * ? ? ?
* ) , or assuming the existence of an underdense void centered not too far from our cosmic location ( e.g. , * ? ? ?
but to avoid the substantial ( and perhaps needless ) complications which arise when assuming departures from general relativity , as well as the non - copernican ` specialness ' implied by a local void , we will instead use the feedback from cosmological structure formation itself as a _
trigger for the onset of acceleration a trigger that automatically activates at just the right time for observers to see it , due to the fact that all such observers will have been created by that very same structure formation which generates the observed cosmic acceleration .
this approach , known generally as `` backreaction '' , was used by this author in @xcite ( henceforth bbi ) to find several clustering models which managed to precisely reproduce the apparent acceleration seen in hubble curves of type ia supernova standard candles , while simultaneously driving a number of important cosmological parameters to within a close proximity of their concordance values including the age of the universe , the matter density required for spatial flatness , the present - day deceleration parameter , and the angular scale of the cosmic microwave background .
the ability of our models to achieve these goals , despite the generally pessimistic view of backreaction typically held by researchers currently ( e.g. , * ? ? ?
* ) , was due to our adoption of an explicitly causal variety of backreaction , which admits the possibility of substantial backreaction from newtonian - strength perturbations .
the standard formalism used for computing backreaction effects , developed through the extensive work of buchert and collaborators ( e.g. , * ? ? ? * ; * ? ? ? * ) , is non - causal in the sense that it drops all ` gravitomagnetic ' ( i.e. , velocity - dependent ) effects , thus rendering it unable to account for metric perturbation information flowing ( at the speed of null rays ) from structures forming in one part of the universe , to observers in another .
in similar fashion , typical studies of cosmic structure formation are also non - causal in that they use the poisson equation without time derivatives of the perturbation potential , thus computing metric perturbations from _ local _ matter inhomogeneities only , disregarding all gravitational information coming in from elsewhere in space .
the result is a mistaken ( but widespread ) notion that the entire newtonian backreaction @xmath10 can be expressed mathematically as a total divergence , thus ultimately rendering it negligible .
but by restoring causality with a `` causal updating '' integral that incorporates perturbations to an observer s metric coming from inhomogeneities all the way out to the edge of their observational horizon , we find ( bbi ) that the sum of such ` innumerable ' newtonian - strength perturbations which increase in number as @xmath11 within a spherical shell at distance @xmath12 from the observer , more than compensating for their @xmath13 weakening with distance adds up to a total backreaction effect that is not only non - negligible ( regardless of the smallness of @xmath14 for most matter flows ) , but is in fact a dominant cosmological effect that is fully capable of reproducing the observed cosmic acceleration in a fully ` concordant ' manner . despite these successes , a major problem with our model in bbi is its utter simplicity : it is clearly a toy model , with the results presented there serving primarily as ` proof - of - principle ' tests , rather than as precision cosmological predictions .
though the simplifications of the model are many , one in particular is serious in its consequences , while fortunately being not too difficult to fix : specifically , this is the dropping of what we have termed `` recursive nonlinearities '' .
unrelated to _ gravitational _ nonlinearities , or to the nonlinear regime of density perturbations in structure formation , recursive nonlinearities embody the fact that the integrated propagation time of a null ray carrying perturbation information to an observer from a distant virializing structure would itself be affected by all of the other perturbation information that has already come in to cross that ray s path from everywhere else , during all times prior to arrival . in other words ,
causal updating is itself slowed by the metric perturbation information carried by causal updating , creating an operationally nonlinear problem .
this issue was necessarily neglected in bbi , as that work was devoted to introducing our ` zeroth - order ' approach to causal backreaction .
but here we fix this problem , incorporating recursive nonlinearities into a new , ` first - order ' version of our phenomenological model
. we will find that this alteration significantly changes our results , causing a profound weakening of the backreaction effects generated by a given level of clustering , as well as significantly damping the long - term effects of information from old perturbations coming in from extreme distances . in order to retain causal backreaction as a viable model for generating the observed cosmic acceleration presuming here that this should indeed be done
it will be necessary to re - interpret the meaning of our ( inherently empirical ) ` clumping evolution functions ' to now consider the effects of hierarchical clustering on a variety of cosmic scales . doing this
, we will show that a successful alternative concordance can once again be achieved , with the right amount ( and temporal behavior ) of acceleration , and with good cosmological parameters .
this paper will be organized as follows : in section [ secrecnonlinform ] , we will re - introduce our original causal backreaction formalism , and then describe the changes implemented in order to incorporate recursive nonlinearities into the model . in section [ secrecnonlinresults ] , we will explore the results of the new formalism , and discuss the implications of the model parameters that are now needed to achieve good data fits .
furthermore , we will discuss how the damping effects due to recursive nonlinearities would alter the key factors that determine the ` ultimate ' fate of the universe , as was discussed for our original formalism in bbi , given an acceleration driven by causal backreaction rather than by some form of dark energy .
finally , in section [ secsummconclude ] , we conclude with a summary of these ideas and results , highlighting the role of causal backreaction as a fundamental component of cosmological analysis and modeling .
we recall here that the basic premise of the formalism developed in bbi is to phenomenologically represent the physical processes of structure formation complex even at the level of newtonian - strength gravitational perturbations in a simple and convenient way .
the physics at work within most clustering masses should be as follows : collapsing overdensities stabilize themselves and halt their collapse by concentrating their local vorticity ( or equivalently , by creating a large local velocity dispersion ) ; this concentrated vorticity or velocity dispersion leads to real , extra volume expansion in accordance with the raychaudhuri equation @xcite , representable ( in the final state ) at great distances by the tail of a newtonian potential perturbation to the background friedman robertson - walker ( frw ) metric ; and this newtonian tail propagates causally outward into space by inducing inward mass flows towards the virialized object from farther and farther distances as time passes .
the total perturbation at time @xmath15 for any location in space which will be independent of position , assuming similar structure formation rates everywhere will then be the combined effects of innumerable newtonian tails of this type , coming in towards the observer from the virializing masses ( in all directions ) which by that observation time have entered within the observer s cosmological `` clustering horizon '' .
as is well known , the expansion evolution ( i.e. , the friedmann equation ) for some spherical volume @xmath16 can be derived using nonrelativistic newtonian equations , in fact , for a matter - dominated universe @xcite without reference to anything outside of that sphere .
in contrast , the newtonian - level backreaction terms which we utilize here are due to perturbation information coming in from structures located predominantly outside of @xmath16 ( since one must go to cosmological distances for the effects to add up significantly ) . for any condensed structure ( at great distance ) which provides a gravitational pull upon the mass in @xmath16 ,
its main perturbative effect is simply to impose an extra ( newtonian ) perturbation potential upon @xmath16 as an addition to its original cosmological metric .
our phenomenological approach , therefore , is one in which we model the inhomogeneity - perturbed evolution of @xmath16 with a metric that contains the individually - newtonian contributions to the perturbation potential within @xmath16 ( `` @xmath17 '' ) from all clumped , virialized structures outside of @xmath16 that have been causally ` seen ' within @xmath16 by time @xmath15 , superposed _ on top of _ the background friedmann expansion of @xmath16 .
we will reiterate the mathematical essentials of this formalism below , in section [ subsecoldformalism ] ; but first we must recount the various approximations and simplifications which have gone into our analysis , to consider their importance and the feasibility of eliminating them in order to develop a greater degree of physical realism in these causal backreaction models .
first of all , though our backreaction - induced metric perturbations will indeed be time - dependent ( due to the causal flow of inhomogeneity information ) , they will be entirely spatially-_independent_. as noted above , we do not seek to achieve an observed acceleration through the mechanism of a local void ; but going even further , our model does not explicitly include any spatial variations whatsoever . rather , the system being modeled is what we term a `` smoothly - inhomogeneous '' universe , in which all perturbation information blends together evenly in a way that is essentially independent of cosmic position .
now , this simplification is one made out of practical necessity , not physical realism .
the smoothly - inhomogeneous approximation relies upon an assumption of randomly - distributed clustering which is certainly not true , as large clusters are not independent of each other , but preferentially clump near one another and are mutually correlated and this becomes ever less true during the ongoing cosmic evolution , as the universe grows more inhomogeneous with time .
furthermore , this simplification relies upon the assumption that the region of space responsible for the dominant contributions to causal backreaction within volume @xmath16 will be large enough to contain a cosmologically - representative sample of both clusters and voids ; but as we will see below in section [ secrecnonlinresults ] , adding in recursive nonlinearities ( to correct another simplification , as described below ) greatly reduces the size of the cosmological region affecting @xmath16 from what it was in our original toy model , potentially calling this assumption into question .
a proper accounting of causal backreaction in a realistically inhomogeneous universe would require the implementation of a fully spatially - detailed , three - dimensional cosmic structure simulation program perhaps along the lines of @xcite , for example but with newtonian - level backreaction effects from causal updating now added in .
the development of such a 3d simulation model is far beyond the scope of this paper ( and beyond the efforts of any individual researcher ) , but would be a useful mission to be undertaken by the cosmological community at large .
a second simplification is our use of newtonian potential terms
i.e. , the long - distance approximation of the schwarzschild metric ( e.g. , * ? ? ?
* ) to represent the tail of each individual perturbation felt from far away , rather than using a long - distance approximation of the kerr metric for spinning masses @xcite , despite the crucial role of some form of vorticity in stabilizing most structures against singular collapse .
in this case , however , the approximation is a good one .
a relatively small amount of vorticity can suffice for the self - stabilization of a clumped mass , if it is applied perpetually ; and the specific angular momentum ( i.e. , @xmath18 $ ] ) will be small for any mass not on the verge of being an extremal black hole .
furthermore , it is easy to show ( e.g. , * ? ? ?
* ) that the highest - order deviations from the newtonian expression in the diagonal metric components will go like @xmath19^{2}$ ] , thus being entirely negligible at the huge distances relevant for causal backreaction .
( and the leading - order _ off - diagonal _ kerr perturbation terms , though actually proportional to @xmath20 , will effectively cancel out due to angle - averaging in the smoothly - inhomogeneous approximation , as discussed in bbi . )
thus it appears quite safe to ignore any kerr - specific perturbation effects for physically reasonable situations .
third , our formalism neglects the purely observational effects of localized inhomogeneities , such as lensing along beam paths ( e.g. , * ? ? ?
* ) for rays from standard candles , and similar perturbative effects upon the apparent luminosity distance relationship for rays passing through inhomogeneous regions , which in some models represents the primary ` backreaction ' effect resulting from structure formation ( e.g. , * ? ? ?
even if such effects by themselves are too small to generate an observed cosmic acceleration , they will still alter the output parameters estimated while using any cosmological model ( including ours ) , and thus should be kept track of ; and in case our causal backreaction method also falls short of providing the full result of an apparent acceleration all by itself , it might successfully be combined with these other observational effects upon the light rays to produce an apparent acceleration once everything is added together ( this point to be discussed again in section [ subunityclumpweak ] ) .
combining these purely observational effects with those from causal backreaction is therefore an important task , and likely quite a feasible one ; though not one addressed yet in this current paper .
the next , most theoretically treacherous approximation is our neglect of nonlinear gravitational effects , a simplification made implicitly by our method of linearly adding together the individual metric perturbations contributed by different self - stabilized mass ` clumps ' in order to produce the total , summed , newtonian - strength perturbation potential .
( note that this is not the full `` newtonian '' approximation usually employed , since while we do assume weak - gravity , we do not completely assume ` slow - motion ' in the sense of dropping all time derivatives of the perturbation potential as that would neglect the causal flow of perturbation information . )
thus our formalism explicitly neglects the nonlinear , purely general - relativistic effects that most other researchers primarily focus upon when studying `` backreaction '' .
this approximation becomes increasingly bad as the magnitude of the summed perturbation potential approaches unity ; but since ( as will be seen below ) this potential typically does not grow to values in excess of @xmath6@xmath21 or so as @xmath22 for most of our best - fitting simulation runs , the approximation is probably good enough for our simulations to provide fairly accurate estimations of the cosmic evolution up to now , and of our measurable cosmological parameters .
( and to the extent that it is not good enough , a significant contribution due to nonlinear gravitational terms would likely only help produce the desired acceleration even more easily . )
thus it is probably not necessary for us to include higher - order gravitational terms in our formalism , in order to achieve a sufficiently reliable understanding of the currently - observable universe for our present purpose of pointing the way towards an alternative concordance ; a fortunate situation , since our model is fundamentally designed around a linearized - gravity approach , and it may be challenging to find any convenient way of modifying it to include nonlinear gravitational effects . on the other hand , given the ever - increasing strength of gravitational nonlinearities in the cosmos over time , a fully general - relativistic model of causal backreaction ( computed using a 3d simulation of realistically - distributed inhomogeneities ) would almost certainly be necessary for accurately predicting the long - term future evolution of the universe .
lastly , there is the approximation regarding what we have referred to as `` recursive nonlinearities '' .
as will be seen from the metric given below in section [ subsecoldformalism ] , one of the effects of causal backreaction is real extra volume creation .
but since causal backreaction depends upon the propagation of inhomogeneity information through space , the extra volume produced by old information from perturbations will slow down the propagation of all future inhomogeneity information ( as well as carrying all perturbing masses farther away from all observation points ) , thus feeding back upon the causal backreaction process in such a way as to strongly dampen it . of all of the simplifications and approximations
discussed so far in this subsection , the neglect of these recursive nonlinearities most likely has the strongest impact upon the quantitative predictions emerging from our causal backreaction models .
fortunately , however , fixing this problem by adding these recursive nonlinearities into our formalism is one of the simpler improvements in physical realism for us to make ; and hence , this paper focuses upon achieving this fix , and then calculating and interpreting the results produced by this ` second - generation ' causal backreaction formalism .
here we recall the technical details of our original formalism developed in bbi , to set the stage for its further development to follow . to obtain the newtonian approximation of a single ` clumped ' ( i.e. , virialized , self - stabilized ) object of mass @xmath23 , embedded at the origin ( @xmath24 ) in an expanding , spatially - flat , matter - dominated ( md ) universe , one may linearize the mcvittie solution @xcite , as can be seen from the perturbed frw expression given in @xcite .
the resulting newtonianly - perturbed frw cosmology is given by the metric : @xmath25 d t^{2 } + [ a_{\mathrm{md}}(t)]^{2 } [ 1 - ( 2/c^{2 } ) \phi ( t ) ] d r^{2 } + [ a_{\mathrm{md}}(t)]^{2 } r^{2 } [ d { \theta}^{2 } + \sin^{2}{\theta } d { \phi}^{2 } ] ~ , \label{newtpertsingleclump}\ ] ] where @xmath26 \}$ ] , and @xmath27 is the unperturbed md scale factor evolution function .
for our model of a smoothly - inhomogeneous universe , we assume a random distribution of clustered masses , being the same essentially everywhere and in every direction . so turning the above expression around
, we consider the situation for an observer at the origin , whose metric is affected by a collection of ( roughly identical ) discrete masses for now confined to a spherical shell at coordinate distance @xmath28 , and with _ total _ mass @xmath23 that are distributed randomly in direction .
the different directions of the various clumps does not matter for the observer s @xmath29 metric component ; but it does matter for the spatial metric components , due to the fact that a spatial displacement from the origin would pick up ( direction - dependent ) factors of @xmath30 in @xmath31 from the different angles of the motion with respect to the individual clumped masses , since only the radial projection of a given translation ( with respect to a particular clump ) will ` feel ' the perturbation potential from that clump in its contribution ( all within @xmath32 ) to the interval @xmath33 . averaging over direction in three dimensions ,
the spatial metric terms therefore pick up a factor of @xmath34 for the total effect when summing over all of the discrete masses ; and the total , ( gravitationally-)linearly summed and angle - averaged metric for this observer at the origin can then be written in ` isotropized ' fashion , as : @xmath35 \ }
~ dt^{2 } ~ + ~ [ a_{\mathrm{md}}(t)]^{2 } \ { 1 + ( 1/3 ) [ r_{\mathrm{sch}}(t ) / r^{\prime } ] \ } ~ \vert d \vec{r } \vert ^{2 } ~ , \label{eqnangavgbhmatdomweakflat}\ ] ] where @xmath36 \}$ ] , and @xmath37 .
( note that this factor of @xmath38 is not ` fundamental ' , but is merely the result of our approximating the linearized sum of many individual ` newtonian ' solutions , which effectively spreads out the total spatial perturbation among all three spatial metric terms , rather than confining it solely to @xmath32 , as is usual .
thus the general relativistic expectation of equal temporal and spatial potentials
i.e. , @xmath39 is not really violated here , and no actual new physics or modified gravity is implied by it . ) we regard the term multiplying @xmath40^{2 } \vert d \vec{r } \vert ^{2}$ ] as a ` true ' increase in spatial volume , and the @xmath29 term as an ` observational ' term , slowing down the perceptions of observers at any given @xmath15 . thus , even if the spatial term by itself is not sufficient to generate a real volumetric acceleration , the spatial and temporal perturbations coupled together may indeed be enough to create an
apparent acceleration " capable of explaining all relevant cosmological observations . the expression in equation [ eqnangavgbhmatdomweakflat ] represents the perturbations to the metric due to masses at some specific coordinate distance @xmath28 and thus from a specific look - back time @xmath41 as seen from some particular observational point .
the total metric at that spacetime point must be computed via an integration over all possible distances , out to the distance ( and thus look - back time ) at which the universe had been essentially unclustered .
finally , a light ray reaching us from its source ( e.g. , a type ia supernova being used as a standard candle ) travels to us in a path composed of a collection of such points , where the metric at each point must be calculated via its own integration out to its individual `` inhomogeneity horizon '' ; and only by calculating the metric at every point in the pathway from the supernova to our final location here at @xmath42 can we figure out the total distance that the light ray has been able to travel through the increasingly perturbed metric , given its emission at some specific redshift @xmath43 .
finite look - back times imply that when one is feeling the effects of clumps at cosmological distances , what one is really sampling is the clustering as it was at an earlier , retarded time .
thus the cosmological evolution at all times is dependent upon the entire history of the development of clustering . in bbi
, we defined a heuristic `` clumping evolution function '' , @xmath44 , intended to simply represent the fraction of the cosmic matter that has completed its relaxation to a self - stabilized , steady state by cosmic time @xmath15 , as opposed to that mass still freely expanding ( or collapsing ) within a still dynamically - evolving patch of space . the function @xmath44 was therefore defined over the range from @xmath45 ( perfectly smooth matter ) , to @xmath46 ( everything clumped ) .
this is an extremely simplified way of representing the virialization of cosmic structures , and we will in fact be forced to re - evaluate the meaning ( and the range ) of the function @xmath44 later on in this paper .
but for now , we just recall that the specific functions used for @xmath44 in our numerical simulations are not rigorously derived from first principles , but rather are adopted for simplicity , with their functional forms and input parameters being motivated by physically reasonable structure formation behaviors , in conjunction with certain quantitative cosmological measurements . now , to determine how hubble curves obtained from standard candle measurements are calculated in our formalism , consider a light ray emitted by a supernova at cosmic coordinate time @xmath47 , which then propagates from the supernova at @xmath48 , to us at @xmath24 , @xmath49 .
we refer here to the geometry depicted in figure [ figsnraytraceints ] . for each point @xmath50 of the trajectory
, the metric at that point will be perturbed away from the background frw form by all of the virialized clumps that have entered its causal horizon .
consider a sphere of ( coordinate ) radius @xmath51 , centered around point @xmath52 , with coordinates @xmath53 ( where @xmath54 is the retarded time ) , defined such that the information about the state of the clumping of matter on that sphere at time @xmath55 will arrive via causal updating , traveling at the speed of null rays to point @xmath52 at the precise time @xmath15 . to compute the fully - perturbed metric at @xmath52
, we must integrate over the clumping effects of all such radii @xmath51 , from @xmath56 out to @xmath57 , the farthest distance from @xmath52 out from which clumping information can have causally arrived since the clustering of matter had originally begun in cosmic history . for the remaining calculations in this subsection , note that we will be using our old formulation from bbi , _ without _ recursive nonlinearities (
those changes will be presented later , in subsection [ subsecnewformalism ] ) . in particular , this affects the speed of the propagation of inhomogeneity information through coordinate space ( as well as some other issues ) , which for now will be calculated with respect to the _ unperturbed _ frw backreaction metric , rather than referencing the perturbed metric itself in an explicitly self - consistent manner . for an frw metric with @xmath58 ,
the coordinate distance traveled by a null ray in the cosmic time span from @xmath59 to @xmath60 will be @xmath61 $ ] .
defining @xmath62 , and with @xmath63 , @xmath64 , we thus have : @xmath65 $ ] .
we then turn this into a prescription for computing @xmath55 as a function of @xmath15 ( relative to the present time , @xmath66 ) and @xmath51 , as follows : @xmath67^{3 } ~ . \label{eqntret}\ ] ] similarly , we can determine @xmath57 , given some initial time @xmath68 at which structure formation can be reasonably said to have started ( i.e. , @xmath69 ) : @xmath70 ~ . \label{eqnalphamax}\ ] ] now , how the metric at @xmath52 is affected by a spherical shell of material at coordinate radius @xmath51 depends upon the state of clumping there at the appropriate retarded time : @xmath71 $ ] . the total effect
is then computed by integrating all shells from @xmath56 out to @xmath72 ; but in order to compute the metric perturbation from each shell quantitatively , it is first necessary to relate this clumping function to an actual physical density of material . as discussed above , for now consider the @xmath44 function as representing the dimensionless ratio of matter which can appropriately be defined as ` clumped ' at a given time , expressed as a fraction of the total physical density . assuming a flat md cosmology as the initially unperturbed state , the total physical density at all times
will merely be an evolved version of the unperturbed frw critical closure density from early ( pre - perturbation ) times .
recalling equation [ eqnangavgbhmatdomweakflat ] , we have the perturbation term @xmath73 = \ { ( 2 g m / c^{2 } ) / [ r^{\prime } ~ a_{\mathrm{md}}(t ) ] \}$ ] , with @xmath74^{1/3}$ ] .
the value of @xmath23 to use here is given by the clumped matter density at coordinate distance @xmath51 times the infinitesimal volume element of the shell .
the clumped matter density at time @xmath15 , as implied above ( and still as considered with respect to the unperturbed background metric ) , will equal @xmath75 $ ] ; and the volume element in the integrand representing the effects of that shell will be @xmath76^{2 } [ a_{\mathrm{md}}(t ) ~ d \alpha]$ ] .
collecting these terms ( and letting @xmath77 ) , the integrand will thus be equal to : @xmath78 _ { r^{\prime } = \alpha \rightarrow ( \alpha + d \alpha ) } & = & \ { ( 2 g / c^{2 } ) ~ d m ~ / ~ [ a_{\mathrm{md}}(t ) ~ \alpha ] \ } \\ & = & \ { ( 2 g / c^{2 } ) ~ [ a_{\mathrm{md}}(t ) ~ \alpha]^{-1 } ~ [ \psi ( t^{\prime } ) ~ \rho _ { \mathrm{crit}}(t ) ] ~ [ 4 \pi r_{\mathrm{phys}}^{2 } d r_{\mathrm{phys } } ] \ } \\ & = & \ { ( 8 \pi g / c^{2 } ) ~ \psi ( t^{\prime } ) ~ [ a_{\mathrm{md}}(t ) ~ \alpha]^{-1 } ~ [ \rho _ { \mathrm{crit}}(t ) ~ [ a_{\mathrm{md}}(t)]^{3 } ] ~ { [ \alpha^{2 } d \alpha } ] \ } \\ & = & \ { ( 8 \pi g / c^{2 } ) ~ \psi ( t^{\prime } ) ~ a_{\mathrm{md}}(t)^{-1 } ~ [ \rho _ { \mathrm{crit}}(t_{0 } ) ~ a_{0}^{3 } ] ~ { [ \alpha d \alpha } ] \ } \\ & = & \ { ( 8 \pi g / c^{2 } ) ~ \psi ( t^{\prime } ) ~ [ ( t_{0 } / t)^{2/3 } ~ ( 3 c t_{0})^{-1 } ] ~ \ { [ 3 h_{0}^{2 } / ( 8 \pi g ) ] ~ ( 3 c t_{0})^{3 } \ } ~ { [ \alpha d \alpha } ] \ } \\ & = & \{12 ~ \psi ( t^{\prime } ) ~ [ ( t_{0 } / t)^{2/3 } ] ~ { [ \alpha d \alpha } ] \ } ~ , \end{aligned}\ ] ] [ eqniintegrandprelim ] where for simplification we have used @xmath79 and the fact that @xmath80 $ ] is constant , both true for a matter - dominated universe .
note also that only @xmath81 is evaluated at the retarded time , @xmath82 .
the _ strength _ of the metric perturbation ( at time @xmath15 for point @xmath52 ) from a point - like newtonian perturbation embedded in the expansion actually depends upon its instantaneous physical distance from @xmath52 at @xmath15 , as is obvious from @xmath83 \ } \equiv [ - g m / r_{\mathrm{phys}}(t)]$ ] in equation [ newtpertsingleclump ] .
the only relativistic " piece of propagating information which is causally delayed is the state of clumping , @xmath71 $ ] , that has just then arrived from coordinate distance @xmath51 to observer @xmath52 at @xmath84 .
the total integrated metric perturbation function due to clumping , @xmath85 , as experienced by a null ray passing through point @xmath52 at @xmath84 , is thus calculated via the causal updating integral : @xmath86 ~ [ ( t_{0 } / t)^{2/3 } ] \ } ~ \alpha ~ d \alpha ~ , \label{eqnitotintegration}\ ] ] with @xmath85 implicitly being a function of @xmath68 ( with @xmath87 for @xmath88 ) , as well as of @xmath66 . inserting this result back into the formalism of equation [ eqnangavgbhmatdomweakflat ] ,
the final clumping - perturbed metric that we will use for all of our subsequent cosmological calculations becomes : @xmath89 ~ dt^{2 } ~ + ~ \ { [ a_{\mathrm{md}}(t)]^{2 } ~ [ 1 + ( 1/3 ) i(t ) ] \ } ~ \vert d \vec{r } \vert ^{2 } ~ . \label{eqnfinalbhpertmetric}\ ] ] given this metric , we can define relationships between all of the unperturbed frw ( `` bare '' ) cosmological parameters , and their corresponding observable ( `` dressed '' ) parameters .
we must first consider cosmological redshifts , where we have : @xmath90 and : @xmath91 - 1 ~ , \label{eqndefnzobs}\ ] ] where quantities without superscripts like @xmath92 , etc .
, will generically refer to unperturbed frw parameters ; and where values of observational quantities in our backreaction - perturbed formalism will be expressly indicated ( e.g. , @xmath93 ) .
now , in order to calculate observed luminosity distances , we must compute the coordinate distance @xmath12 of a supernova going off at coordinate time @xmath15 , for which a light ray would be arriving here ( at @xmath24 ) precisely at @xmath66 . for a null ray and pure inward radial motion we have : @xmath94 [ eqnroftintegration ] this coordinate distance function can then be converted into an expression for the observed luminosity distance by adapting the expression from the homogeneous frw case , for which @xmath95 $ ] , as follows : @xmath96 ~ r^\mathrm{frw}_{\mathrm{sn}}(t ) ~ [ 1 + z^\mathrm{obs}(t ) ] \\ & = & \frac{1 + ( i_{0}/3)}{\sqrt{1 + [ i(t ) / 3 ] } } ~ \frac{c ~ t_{0}^{4/3}}{t^{2/3 } } ~ \int^{t_{0}}_{t } \ { ( t^{\prime})^{-2/3 } ~ \sqrt{\frac{1 - i(t^{\prime})}{1 + [ i(t^{\prime } ) / 3 ] } } \ } ~ d t^{\prime } \\ & = & \frac{1 + ( i_{0}/3)}{\sqrt{1 + [ i(t_{r } ) / 3 ] } } ~ \frac{c ~ t_{0}}{t_{r}^{2/3 } } ~ \int^{1}_{t_{r } } \ { ( t^{\prime}_{r})^{-2/3 } ~ \sqrt{\frac{1 - i(t^{\prime}_{r})}{1 + [ i(t^{\prime}_{r } ) / 3 ] } } \ } ~ d t^{\prime}_{r } ~ , \end{aligned}\ ] ] [ eqndlumdefn ] where @xmath97 , and @xmath98 , @xmath99 are dimensionless time ratios ( e.g. , @xmath100 ) , with no change to the essential form of @xmath85 ( i.e. , @xmath101 ) . all cosmologically - relevant curves , fits , and parameters can now be calculated from the metric ( equation [ eqnfinalbhpertmetric ] ) , and from this luminosity distance function ( equation [ eqndlumdefn ] ) and its derivatives , as investigated in bbi .
important modeled quantities include : the residual distance modulus function ( with respect to an empty coasting universe ) , @xmath102 , and the quality and probability of its fit , @xmath103 and @xmath104 , to the type ia supernova ( snia ) data ; the observed ( versus unperturbed ) hubble constant , @xmath105 ( versus @xmath106 ) ; the physically - measurable age of the universe , @xmath93 ; the `` true '' value of the total cosmic matter density , @xmath107 , which determines the primordial spatial curvature ( i.e. , @xmath108 for flatness in the pre - perturbed epoch ) , which our model must relate to some measured value of the density , @xmath109 , for normalization ( here and in bbi we use @xmath110 ) ; the observable values ( defined for @xmath111 ) of the deceleration parameter @xmath112 , the effective ( total ) cosmic equation of state @xmath113 , and the jerk parameter @xmath0 ; and finally , for comparison with complementary data sets from much earlier cosmic epochs , we compute the acoustic scale of the cosmic microwave background ( cmb ) acoustic peaks , @xmath114 . as a technical note , we recall that all evolving quantities in our numerical simulation program are calculated as discrete arrays in @xmath15 ( and in @xmath115 ) , with a tested pixelization that is fine enough for great accuracy in all parameters .
given that the discrete version of @xmath116 must be differentiated ( and evaluated specifically for @xmath111 ) to obtain cosmological parameters , we do so by using the definition of the derivative for each pixel , as follows : @xmath117 _ { \ { i \ } } = \frac{d}{d ~ z^\mathrm{obs } } [ d^{(n-1 ) \prime}_{\mathrm{l } , \mathrm{pert } } ] _ { \ { i \ } } \equiv \frac{d^{(n-1 ) \prime}_{\mathrm{l } , \mathrm{pert}}(t _ { \ { i+1 \ } } ) - d^{(n-1 ) \prime}_{\mathrm{l } , \mathrm{pert}}(t _ { \ { i \ } } ) } { z^\mathrm{obs}(t _ { \ { i+1 \ } } ) - z^\mathrm{obs}(t _ { \ { i \ } } ) } ~ , \label{eqnpixeldiffdef}\ ] ] and then obtain the @xmath111 limit from the last , latest - in - time pixel : @xmath117 _ { ( z \rightarrow 0 , ~ t \rightarrow t_{0 } ) } \equiv [ d^{n \prime}_{\mathrm{l } , \mathrm{pert } } ] _ { \ { n_\mathrm{pix } \ } } ~ . \label{eqnfirsteval}\ ] ] the cosmological results which we obtain from this procedure appear to be robust , with only minor difficulties , as will be discussed below ; and while the greatest discretization errors occur for @xmath0 , which requires three differentiations of @xmath118 , virtually all of the results which we quote here for @xmath0 should be well within @xmath119 in terms of numerical accuracy . in order to conduct specific calculations with our formalism , we must design a set of physically reasonable clumping evolution functions , @xmath44 , to serve as convenient proxies for the combined effects of the linear density evolution of early - stage clustering , the nonlinear regime and virialization for very dense clumps , and the initial development ( in many cases triggered by collisions ) of entirely new clumps with substantial mass .
guided by general cosmological considerations and simplicity , in bbi we chose three different classes of time - dependent behaviors to examine : @xmath120 , which is proportional to the evolving contrast of a density variation , @xmath121 , in the linear regime @xcite ; @xmath122 , a generally sensible choice depending simply upon the amount of time available for clumping ; and @xmath123 , an ` accelerating ' clumping model which we initially chose as a test case to see whether that would possibly help in creating an observed acceleration .
this last class of models will take on more significance in this paper , however , since our results below with recursive nonlinearities will force us to regard @xmath44 as not merely a simple percentage of clumped versus unclumped matter , but as a quantity reflecting the details of virialization brought to completion ( i.e. , extremely nonlinear density perturbations ) on multiple cosmic scales ; and appropriately , the density contrast evolution for inhomogeneities in the nonlinear regime goes as @xmath124 with @xmath125 @xcite that is , @xmath126 with @xmath127 for a matter - dominated universe .
quantitatively , we defined our three different classes of clumping evolution models as follows : @xmath128 [ eqnclumpmodels ] here @xmath68 ( or equivalently , @xmath129 $ ] , as per equation [ eqndefnzfrw ] ) represents the effective beginning of clumping , such that @xmath69 for all models ; and @xmath130 represents the current state of clumping today .
each of these models has two physically meaningful parameters to vary ( besides the universally - optimizable parameter for all models , @xmath105 ) : @xmath131 , for which we selected values by estimating the fractions of dark and baryonic matter that may likely be clumped by @xmath66 ; and @xmath132 , for which we selected values by roughly linking the beginning of clumping with the process of cosmic reionization . in bbi
, we evaluated @xmath133 different models , using @xmath134 and @xmath135 for each of @xmath136 , @xmath137 , and @xmath138 .
we found that these models produced curves which behaved very much like @xmath1cdm , with the @xmath139 runs looking like flat @xmath1cdm with @xmath140 , the @xmath141 runs looking like @xmath142 , and the @xmath143 runs looking like @xmath144 .
as was expected , due to the increase in the number of perturbing clumps with distance ( as @xmath11 ) , the total perturbative effects were typically dominated by the largest distances ( and thus the earliest look - back times ) out to which one could still see significant inhomogeneities ; hence clumping evolution functions with a more rapid onset of clustering at early times would produce a more powerful causal backreaction effect .
thus the @xmath143 models were by far the strongest , while the @xmath139 models were all too weak to reproduce the observed acceleration , with the @xmath136 models falling intermediate between the two ; and in all cases , the choice of an earlier @xmath132 led to a stronger overall backreaction .
all told , roughly a dozen of these models resulted in good fits to the standard candle snia data , reproducing the observed cosmic acceleration essentially as well as the best - fit @xmath1cdm dark energy model ; and about half of those dozen models produced especially good cosmological parameters , as well .
the overall picture was that of a successful attempt at achieving an alternative cosmic concordance without dark energy .
additionally , there emerged the bonus of a testable prediction which could distinguish causal backreaction from a cosmological constant : a result of @xmath145 for all of our models which produced good fits , in contrast to the mandatory value of @xmath146 for all flat @xmath1cdm models , regardless of the value of @xmath147 . with those results in hand ,
the next concern was to determine how the effects of recursive nonlinearities might modify these outcomes , once incorporated into our numerical models .
our initial expectation was that the overall hubble curves of each modeled cosmology would likely not change significantly , except perhaps at the end , as @xmath111 ; a result that would not alter the snia fitting results much , but which could have a measurable effect upon the cosmological parameters evaluated at @xmath66 particularly upon our ` paradigm falsifiability ' parameter , @xmath0 .
the actual outcome , however , is that the proper treatment of recursive nonlinearities ends up altering the situation to a great degree , requiring us to re - evaluate ( see section [ secrecnonlinresults ] ) the ultimate implications of causal backreaction ; although it does still remain true that an alternative concordance can be successfully achieved with it .
first , however , in the following subsection we give the precise details explaining how recursive nonlinearities for causal backreaction are technically implemented in our numerical simulations .
though in practice necessitating a complete re - write of our numerical simulation program , the essential features of recursive nonlinearities require just two fundamental changes .
the first change involves the rate at which clustering information can propagate through the inhomogeneity - perturbed universe , as depicted earlier in figure [ figsnraytraceints ] .
the result is that @xmath82 and @xmath148 are altered from their simple functional forms ( as per equations [ eqntret],[eqnalphamax ] ) for a matter - dominated frw universe , and now depend upon the metric perturbation potential function , @xmath85 .
but as we see from equation [ eqnitotintegration ] , @xmath85 is itself calculated via an integral depending upon @xmath82 and @xmath148 .
hence @xmath85 is now defined in terms of @xmath85 , in a fundamentally recursive fashion .
second , the extra spatial volume created by nonzero @xmath85 will have a dilution effect upon the strength of the perturbation induced by any ( now farther - away ) clumped mass upon the metric at our observation point .
examining equation [ eqniintegrandprelim ] ( specifically [ eqniintegrandprelim]a ) , we recall that the perturbation term for each spherical shell of clumped matter is given by @xmath73 _ { r^{\prime } = \alpha \rightarrow ( \alpha + d \alpha ) } = \ { ( 2 g / c^{2 } ) ~ d m ~ / ~ [ a_{\mathrm{md}}(t ) ~ \alpha ] \}$ ] .
now , the differential mass element in the shell , @xmath149 , does not actually change ( and therefore requires no correction factor ) , since the dilution of its mass density is precisely offset by its expanded volume .
but what _ does _ change is the effective distance of those perturbing clumps from the observation point .
recalling that the strength of a newtonian perturbation embedded in an expanding universe depends simply upon its instantaneous physical distance , we see that the denominator @xmath150 $ ] must now be replaced by the term : @xmath151 = [ a_{\mathrm{md}}(t ) ~ \sqrt{1 + ( 1/3 ) i(t ) } ~ \alpha]$ ] , which in the end just puts a factor of @xmath152 into the denominator of the integral for @xmath85 .
therefore , the modified formula ( replacing equation [ eqnitotintegration ] ) for calculating the metric perturbation function @xmath85 with the incorporation of recursive nonlinearities ( `` rnl '' ) , is given as : @xmath153 ~ [ ( t_{0 } / t)^{2/3 } ] \ } ~ \alpha}{\sqrt{1 + ( i/3 ) } } ~ d \alpha ~ , \label{eqnitotrnlintegration1}\ ] ] where the term `` @xmath154 '' inside the integral on the right - hand side represents the actual function , @xmath155 , itself .
but since the denominator term is in fact independent of the integration variable @xmath51 , we can remove it from the integral and bring it to the left - hand side of the equation
. the metric perturbation function can therefore be ( numerically ) solved for any given @xmath15 as the solution of : @xmath156 } = \int^{\alpha _ { \mathrm{max } } ( t , t_\mathrm{init } , i)}_{0 } \{12 ~ \psi [ t_{\mathrm{ret } } ( t , \alpha , i ) ] ~ [ ( t_{0 } / t)^{2/3 } ] \ } ~ \alpha ~ d \alpha ~ .
\label{eqnitotrnlintegration2}\ ] ] to evaluate the expressions @xmath157 and @xmath158 in this above integral , we utilize ( analogously with the discussion preceding equations [ eqntret],[eqnalphamax ] ) : @xmath159 } } \frac{d t^{\prime}}{(t^{\prime})^{2/3 } } ~ , \label{eqnalpharnlintegration2}\ ] ] with @xmath160 and @xmath161 .
this expression for @xmath162 could in theory be inverted to produce @xmath163 ; and in addition , we have @xmath164 . in practice , we perform these recursively - defined ` integrals ' by utilizing discrete arrays in cosmic coordinate time with some large number of pixels covering the range from @xmath165 to @xmath166 , for which the later pixels are calculated in terms of the earlier pixels . beginning with the first pixel at @xmath165
, we thus have the recipe ( with @xmath167 pixels , and @xmath168 $ ] ) : @xmath169 [ eqnrnldiscretegridloopa ] and then , for @xmath170 : @xmath171 } } \frac{\delta t}{[t_{(i-1)}]^{2/3 } } \ } ~ , \\ & x^{\mathrm{rnl}}_{i } & = \frac{12}{t_{i}^{2/3 } } ~ \sum_{k = \{1,(i-2)\ } } \ { \psi[t_{(i - k ) } ] [ \alpha _ { \mathrm{max},i } - \alpha _ { \mathrm{max},(i - k ) } ] [ \alpha _ { \mathrm{max},(i+1-k ) } - \alpha _ { \mathrm{max},(i - k ) } ] \ } ~ , \\ & i^{\mathrm{rnl}}_{i } & \sqrt{1 + [ i^{\mathrm{rnl}}_{i } /3 ] } = x^{\mathrm{rnl}}_{i } ~. \end{aligned}\ ] ] [ eqnrnldiscretegridloopb ] the result of this iterative loop is the discrete array @xmath172 , which serves as @xmath155 for all cosmological calculations , as described above . as a test , we have verified that removing the effects of recursive nonlinearities ( essentially adjusting to unity all terms inside the radical signs in equations [ eqnrnldiscretegridloopb]a , c ) succeeds in reproducing the results of our old formalism to a great degree of accuracy ; and for our new model simulations with the recursive nonlinearities included , we have checked to make sure that the full suite of results with this new formalism appears quite sensible and consistent in all cases .
we have also tested the pixelization for precision of the output results , finding that runs with @xmath6@xmath173 pixels are sufficient for quick model parameter optimization searches , producing all cosmological results to within a couple percent of their ` true ' values ( i.e. , the values to which the parameters asymptote for much larger pixelizations ) ; and that going to @xmath6@xmath174 pixels yields cosmological output results that are stable to within a small fraction of a percent .
( though going all the way to @xmath6@xmath175 pixels actually leads to discretization problems due to round - off error , which causes problems for higher derivatives of the luminosity distance function , leading to instability in the value of @xmath0 ; and so we avoid pixelizations this large . ) our high - precision output cosmological parameters
i.e. , all numerical results other than those from optimization searches which we will quote in section [ secrecnonlinresults ] , below , have all been obtained from runs with @xmath6@xmath174 pixels ( not counting the further addition of a smaller number ( @xmath6@xmath173 ) of extra pixels used for integrating the hubble curves out to the computationally simpler region past @xmath132 , before the simulated onset of clumping ) .
as mentioned previously , the incorporation of recursive nonlinearities ( henceforth rnl ) does not merely affect the late - time behavior as @xmath111 , but in fact exerts a profound damping influence upon the entire process of causal backreaction , weakening the overall effect for a given set of model input parameters .
as an example , we consider the impact of rnl upon one of the best - fitting models from bbi , @xmath136 with @xmath176 and @xmath177 .
as depicted in figure [ figstrcl1weakplot ] , we see that adjusting the simulations to include rnl greatly reduces the apparent acceleration effect no longer quite even achieving an ` acceleration ' , in fact , since @xmath178 . for ease of comparison to the results from bbi , in this paper we will still primarily use the snia data from the scp union1 supernova compilation @xcite for conducting fits of our new models with rnl .
as would be expected from figure [ figstrcl1weakplot ] , the new fit to these snia data ( not shown here ) , even after re - optimization with respect to @xmath105 , is much worse in this modified model , with its ` chi - squared ' value for @xmath179 snia minus @xmath180 model parameters yielding @xmath181 degrees of freedom increasing from @xmath182 without rnl , to @xmath183 with it ( where for comparison , @xmath184 for best - fit flat @xmath1cdm ) . visually speaking , though this causal backreaction model with rnl
does at least manage to clearly separate itself from the strongly decelerating behavior of matter - only flat scdm , it no longer matches best - fit @xmath1cdm but now falls significantly short of it , lying somewhere in the middle between scdm and @xmath1cdm ; and correspondingly , the output cosmological parameters though still being substantially modified from those for the unperturbed case without causal backreaction are now far less reflective of those required for a successful concordance , as would indeed be expected from a cosmological model producing too weak of an ` acceleration ' effect . within the set of @xmath133 model input parameter choices from bbi ( all with @xmath185 ) , the ` best ' models
now , with rnl included , are @xmath143 with @xmath186 , which fits the snia most successfully with @xmath187 ; and @xmath139 , also with @xmath186 , which despite yielding a worse overall fit ( @xmath188 ) because of too little clumping at early times , manages to produce the largest late - time backreaction effect , achieving an actual ` acceleration ' with @xmath189 .
these two models are plotted in figure [ figsqrmdcl1opth0weakplot ] , shown along with matter - only flat scdm and @xmath1cdm , where all models are now shown in the figure as optimized with respect to @xmath105 for the union1 snia data set .
one of the most important changes in causal backreaction due to rnl , is the slowdown that it causes for inhomogeneity information propagating in from great cosmological distances .
this strongly damps the backreaction effects due to perturbations at the outer edges of an observer s causal inhomogeneity horizon , which despite being the most distant ones are nevertheless very important for backreaction because they include the largest spherical shells of inhomogeneous matter , containing the greatest incremental mass ( per shell ) of perturbations overall . besides simply weakening the total backreaction effect , rnl acts specifically to inhibit the contributions from older perturbations generated during relatively early clumping epochs , due to a reduction of the coordinate distance ( and thus of the volume of clumped matter now enclosed by the sphere ) out to some given propagative look - back time ; an inhibition that is particularly strong if the perturbation potential @xmath85 has grown too close to unity very early on .
one important potential result is that this may significantly lessen the ongoing effects of causal backreaction into the far future due in the pre - rnl formalism to eternally - expanding observational horizons an issue related to the ultimate possible fates of the universe , to be discussed further in section [ subnewfuture ] , below .
but our immediate concern here is to note that rnl makes later ( i.e. , more recently generated ) clumping far more effective than earlier clumping in terms of generating a late - time acceleration .
hence in our runs here with rnl , we obtain the strongest apparent acceleration effects with the @xmath139 models , and the weakest with the @xmath143 models ; also , choosing smaller ( i.e. , later ) values of @xmath132 always makes @xmath112 more ` accelerative ' ( smaller or more negative ) , as well .
this is all very different from the results in bbi , where the @xmath139 functions were by far the weakest models at producing an apparent acceleration ; and where the @xmath143 models were so strong that small or mid - range values of @xmath132 were necessary to obtain good fits , since those models with the beginning of clumping occurring significantly _ earlier _ than @xmath190 led to an actual overkill of acceleration .
this behavior of causal backreaction with rnl suggests that we may be able to get a stronger accelerative effect by using later and later ( i.e. , smaller ) values of @xmath132 ; and indeed , going from @xmath191 to @xmath192 for the @xmath139 model with @xmath177 does strengthen the observed acceleration from @xmath193 to @xmath194 .
but this type of procedure does not save these models as a mechanism for explaining the apparent acceleration , due to several obvious problems .
first of all , the total amount of generated acceleration is still far too small , and thus the fits to the snia data remain extremely poor ( and in fact , @xmath195 gets _ worse _ for the @xmath139 runs with later @xmath132 ) . besides such problems with
the snia fits , the output cosmological parameters are also unacceptable for such runs , with quantities like @xmath93 and @xmath107 growing far too small with decreasing @xmath132 , thus spoiling any efforts at achieving an alternative cosmic concordance .
lastly , such input parameters do not even make good physical sense can we seriously estimate the effective onset of cosmic clustering as beginning at @xmath196 , or even later ?
clearly , we can not achieve our goals of reproducing the observed cosmic evolution with causal backreaction in the presence of rnl while stuck within the confines of models with @xmath197 .
these difficulties lead us to the question of what may be wrong with our formalism ; a question to which there are several possible answers .
first of all , it is certainly possible that causal backreaction is not the mechanism fundamentally responsible for the observed cosmic acceleration .
even so , given that these effects are ( as we argue ) real and significant the cosmic expansion is clearly different than it would be without them , as we have seen above in figures [ figstrcl1weakplot ] and [ figsqrmdcl1opth0weakplot ] , witnessing the clear separation between our models and unperturbed scdm even a cosmology actually dominated by some form of dark energy would be strongly modified by causal backreaction , leading to misconceptions about the evolution of the universe ( and about the nature of that dark energy , itself ) were these effects not taken into account .
alternatively , causal backreaction may be part of a ` combination ' solution for explaining the apparent acceleration in conjunction , perhaps , with observational effects ( recall section [ subsecoldapprox ] ) such as lensing or other inhomogeneity - induced perturbations to the luminosity distance function @xmath198 such that the _ total _ result from all perturbative effects , combined , may be sufficient to remove the need for dark energy .
another possibility is that our fundamental physical approach towards causal backreaction is far too oversimplified to properly model the cosmic evolution , just as we have argued was true for the standard ( non - causal ) method traditionally used for estimating backreaction .
there are many possible flaws with our approach for example , perhaps the linearized mcvittie solution ( from which we calculated our perturbative metric in section [ subsecoldformalism ] ) is inappropriate , given that the total mass of the perturbation in that metric is fixed as constant in time @xcite , in contrast to the tendency of overdensities in the real universe to keep increasing in mass due to incoming mass flows .
( though our growing @xmath199 functions are of course intended as an alternative way of accounting for this . ) on the other hand , it is quite possible that the complex processes of cosmic structure formation are too dynamically rich to be modeled accurately by a newtonianly - perturbed metric containing just a single perturbation potential function .
the very natures of our @xmath199 functions are obviously oversimplified not accounting at all , for example , for the differences between the simple linear evolution of matter overdensities , and the far more complex process of nonlinear density perturbation evolution the latter including subsidiary effects such as virialization and stabilization ( the detailed time - dependence of which may be quite important , though our formalism completely neglects it ) , as well as `` gastrophysics '' feedback ( e.g. , energy injection from star formation and from the supernovae , themselves ) resulting in baryon shock heating , and so on . considering such complications , it is possible that no simple functional form of _ any _ kind for @xmath199 and for that matter , no simple numbers for @xmath132 or @xmath131 may suffice for a realistic description of the physics
. perhaps nothing would be accurate short of a fully - detailed cosmic simulation program , calculating a realistically self - consistent representation of the intertwined behaviors of cosmic structure formation and astrophysical feedback processes as they evolve together over time . a different
, possibly significant flaw may be our complete neglect of _ gravitational _ nonlinearities ; after all , values for the ` newtonian ' perturbation potential in the range of @xmath200 are certainly not negligible , and may imply higher - order gravitational terms from general relativity that would contribute significant backreaction effects for any reasonable model of the evolution of cosmic structure . leaving out such nonlinear gravitational terms may imply that our calculations here are not serving as an ` estimate ' of the overall causal backreaction effect , but rather as a _ lower limit _ to it .
yet another possibility is that the greatly simplifying assumption of a smoothly - inhomogeneous universe is incompatible with a proper treatment of recursive nonlinearities ; after all , inhomogeneities massive on a truly cosmological scale , such as galaxy clusters , are not smoothly spread in angle , but are concentrated at particular locations on the sky . as such , rnl effects would slow down the propagation of inhomogeneity information more strongly along some directions than along others ; and so perturbative effects from new clumps , located in new directions , may be able to ` slip in ' with less delay than would inhomogeneity information coming in from the same directions on the sky as pre - existing clusters that have already imposed their causal backreaction effects upon us . ( a kind of _ cosmic _ `` channeling '' effect , so to speak . ) but quantitatively significant or not , it is likely that such issues can not be properly evaluated without the development of a fully-3d simulation program for causal backreaction , as discussed previously in section [ subsecoldapprox ] . on the other hand , the main difficulty here may really be much simpler : the basic physical assumptions and methods of our formalism might actually be capable of modeling causal backreaction effects well enough , with the problem simply being inadequate choices for our ( heuristically - adopted ) clumping evolution models and model input parameters . given the above results with rnl , it is clear that simply tweaking the detailed time - dependence of the @xmath44 functions would not be enough to obtain good data fits ; but despite such difficulties , there is actually one very simple thing that we _ can _ do , which may extend the applicability of our causal backreaction simulation formalism , while also dealing with some very real physical effects in an empirical yet useful way .
from the results in this section , it is clear that we can not achieve a new cosmic concordance with what one might call `` clumping - saturated '' models that are ultimately limited in strength to values of @xmath131 bounded above by unity .
but what about models with @xmath201 ? such models would clearly lead to much stronger backreaction effects ( with no pre - determined upper bound , obviously ) , and could perhaps include models that explain all cosmological observations , re - establishing an alternative concordance for causal backreaction even in spite of the damping effects of rnl .
given our basic definition of @xmath131 as `` the fraction of cosmic matter in the clumped , self - stabilized state '' , one may ask whether it is ever physically appropriate to use values of @xmath131 in excess of unity . in the upcoming section
, we will argue that the answer is _ yes _
, that it is entirely appropriate and in fact , both useful and necessary for a valid phenomenological description of _ hierarchical clustering _ in the universe . the convenient but ad - hoc representation of the state of cosmic clumping at some given time @xmath15 with just a single number
@xmath44 is clearly an extreme simplification , as discussed above .
even aside from our neglect of the detailed time dependence of virialization processes by having this function simply represent the clumped matter fraction over time , there is also the fact that large - scale structure is not a featureless blob : the ` clumped ' state is not a simple binary state , ` on ' or ` off ' .
most notably , clustering in the universe is not structure - free , but instead exhibits a roughly discrete form of hierarchical clustering , with identifiable structures existing on a few well - defined scales for example , galaxies existing as virialized groupings of stars , galactic clusters as virialized groupings of galaxies , and so on .
this recognizable substructure will obviously affect the resulting causal backreaction that is induced , and so the question becomes how to best ( and most simply ) estimate the effects of such substructure .
one conclusion which we can immediately draw is that clustering with multiple levels of substructure will exert _ more _ of a total backreaction effect than would a simpler form of structure without them . from elementary physical considerations ,
it is clear that the total gravitational potential energy contained within a bound collection of individually - condensed clumps will be larger in magnitude ( i.e. , more negative ) than that within a bound volume of featureless dust . by the virial theorem , the presence of a greater amount of ( negative ) potential energy requires the counterbalance of more kinetic energy , which implies more backreaction - causing vorticity and velocity dispersion . thus the existence of more ( stable ) substructure implies that more induced causal backreaction was generated during the process of stabilizing it . but one must ask , how much more ?
each distinct level of structure actually involves the _ same _ amount of mass , when considering it over a fixed cosmic volume
whether one considers the individual virializations of the tens or hundreds of galaxy clusters in a single supercluster , or of the tens or hundreds of _ thousands _ of individual galaxies in that same supercluster , or of the trillions of stellar - mass bodies in that supercluster and in virializing the same amount of mass , it seems reasonable to expect the same amount of volume - generating velocity dispersion to occur .
thus we claim that an effective phenomenological way of estimating the total backreaction effect that has been generated by a stabilized volume of mass with multiple levels of internal substructure on different length scales , is by _ counting the effective number of levels of structure _ that have attained an essentially complete virialization .
astrophysically realistic structure formation therefore provides a natural way in which @xmath44 should be expected to not only reach but to eventually exceed unity , and then to even keep growing ( without any fundamental theoretical limit ) as long as there are new levels of structure to be brought into a self - stabilized state . to avoid giving ourselves an infinite amount of leeway in searching for an alternative concordance , however , it is important to place bounds upon the ` reasonable ' values of @xmath202 that one should expect to find when using causal backreaction to explain the apparent acceleration .
the essence of the question is simple : `` how many levels ? '' in particular , we need only consider those typical length scales which became virialized during or not too long before the general cosmic time of the observed acceleration ( i.e. , @xmath203 few ) , since ( as will be shown in section [ subearlysatclumping ] ) rnl sharply limits the ongoing influence of causal backreaction effects from relatively early , ` outdated ' virializations ( _ especially _ when they are strong ) .
it seems safe to include at least two scales in our @xmath131 estimate : that of individual galaxies , and that of galaxy clusters , implying a @xmath131 of at least @xmath204 . going below these classes of structures in length scale ,
there are of course a multitude of individual star clusters and solar systems ( supported by the vorticity in their orbital motions ) that have not too long ago undergone ( or are still undergoing ) relaxation ; and even below that in scale , there are continual generations of newly - forming solid objects supported by internal body forces ( which going beyond the perfect - fluid approximation , also contribute positive terms to the raychaudhuri equation ) . below this
, of course , there is infinite regress ; though at some point it becomes absurd to consider ever - smaller structures , and significant quantities of new self - stabilization is not happening at very small scales in any case .
it is difficult to precisely estimate the total effective causal backreaction feedback that is induced by structures at sub - galaxy scales ; but for our purposes , let us consider it a ` significant fraction ' of @xmath205 .
at much higher scales , there are groupings of galactic clusters , leading to superclusters , and the largest known structures in general .
superclusters are too large to be gravitationally bound yet ( and certainly were not bound at high redshifts , back around the onset of the apparent acceleration ) , but are experiencing relaxation processes that are underway during the current epoch . once again lacking a precise estimate , we hypothesize that the contributions by stabilized structure formation to backreaction from scales larger than that of typical galaxy clusters , in terms of increasing the effective value of @xmath131 , may amount to a small but non - negligible fraction of @xmath205 .
all told , reasonable values of the present - day clumping evolution parameter for causal backreaction , when one considers hierarchical clustering , may be within the range @xmath206 ; or really pushing it , perhaps within @xmath207 . allowing ourselves such an enhanced phase space for @xmath131 ,
the question then becomes whether or not we can re - establish causal backreaction , now with rnl included , as a successful ( and ` concordant ' ) explanation of the observed acceleration
. for each of the @xmath143 , @xmath141 , and @xmath139 clumping evolution functions , different values of @xmath132 can be explored , where the highest reasonable value ( say , @xmath208 ) is set to represent a time shortly before the onset of cosmic reionization ; and where the smallest reasonable value ( say , @xmath196 ) would represent the time leading up to serious and intense virialization , signaled by such events as the supernovae themselves ( noting that the highest - redshift snia in compilations like those of the scp union collaboration tend to be at @xmath209 or so ) .
then for each value of @xmath132 that we adopt , a search is done over @xmath131 to find which value achieves a best - fit by minimizing @xmath195 .
finding a value of @xmath131 which minimizes @xmath195 for any given @xmath132 can always be done and often with good results , where the @xmath195 value is almost as low ( and sometimes even lower ) than that for the best - fit @xmath1cdm model .
but there is more to the true task of ` optimization ' than this , since besides achieving a good fit to the snia data , one must also match the observed values of a variety of complementary cosmological parameters in order to produce a proper alternative concordance . achieving all of these goals
simultaneously is more challenging ; and so to conveniently measure our ability in this regard , we define what can roughly be considered an `` average percent deviation from concordance '' parameter , as follows : @xmath210 where all of the numerical values used above to represent the ` concordance ' values are reference numbers obtained from the union1-best - fit flat @xmath1cdm model ( i.e. , @xmath211 , @xmath212 , with the approximation of zero cosmological radiation density ) , and all quantities are expressed here in typical units ( e.g. , gyr for @xmath93 ) .
an optimization search over the range of reasonable input parameters for our three cluster evolution functions , using quick simulation runs ( i.e. , 1000 pixels for @xmath213 ) , produces a number of models that provide both good snia fits and good output cosmological parameters .
a summary of the results of this search here showing only those runs which possess the optimal @xmath214 value for minimizing @xmath195 in each particular @xmath132 case is given in table [ tablernlt0satoptimruns ] .
we see that the best ( minimized ) value of @xmath195 obtainable is in fact somewhat different for different choices of the input parameter @xmath132 ; but varying even more dramatically with @xmath132 is the average deviation from concordance for its @xmath195-minimizing run .
we have highlighted ( in bold ) five of the cases in table [ tablernlt0satoptimruns ] : @xmath143 with @xmath215 and @xmath216 ; @xmath141 with @xmath217 and @xmath218 ; and @xmath139 with @xmath219 .
we will refer to these cases as ` best runs ' , chosen not by the global minimization of @xmath195 , but by selecting @xmath132 settings for which the @xmath195-minimizing @xmath214 run produces cosmological parameters with very small deviations from concordance , in addition to having values of @xmath195 that are acceptably close to the smallest ones found for the relevant @xmath44 function .
we will consider the astrophysical implications of the particular @xmath220 input values required to obtain these ` best ' runs , later on , below ; but first , we consider here the principal output results of interest . these five best runs (
now re - done with 10000 pixels each , for higher precision ) are all good fits to the snia data , as can be seen from their @xmath195 values , which are nearly as small as that for the best - fit @xmath1cdm model ( and are even smaller than that from @xmath1cdm for _ all _ of the @xmath139 runs in the table ) .
the high quality of these fits can also be inferred from figure [ figclgt1strongplots ] , where we plot three of these runs one each for @xmath143 , @xmath141 , and @xmath139 against @xmath1cdm and the matter - only flat scdm model , showing our runs to be very good at mimicking best - fit @xmath1cdm ( particularly within the redshift region @xmath221 , containing most of the snia ) , thus demonstrating their ability to successfully reproduce the cosmic ` acceleration ' as it is actually observed .
the derived output parameters for these five best @xmath201 runs are presented in table [ tablernlt0satbestruns ] . here
we see how their @xmath195 values , nearly as good or ( for @xmath139 ) better than that for best - fit @xmath1cdm , makes them comparable to @xmath1cdm in terms of the fit probability , @xmath104 ( noting that @xmath1cdm has fewer optimizable model parameters , and thus more degrees of freedom , giving it a higher @xmath104 value for the same @xmath195 ) .
we also see that their present - day ` newtonian ' perturbation values , @xmath222 , do not go much higher than @xmath223 ; this places moderate limits upon the effects of gravitational nonlinearities which are _ not _ explicitly dealt with in our formalism implying at least some reasonable level of accuracy for our models regarding this approximation . considering the observational cosmological parameters , there is an astrophysically - interesting modification of @xmath224 ( relative to @xmath225 ) of @xmath6@xmath226 ; while the best - fit observable hubble constant , @xmath105 , is in all cases extremely close ( well within @xmath227 ) of the @xmath1cdm concordance value of @xmath228 .
the cosmic ages of these models are all at least equal to the `` concordance '' value of @xmath229 gyr ( with the @xmath139 run being _ exactly _ equal to it ) ; and since cosmologies with higher ( if still reasonable ) ages would also solve the cosmic age problem just as well , all of these runs ( no more than @xmath6@xmath230 gyr older ) therefore appear to have quite acceptable values of @xmath93 .
the goal of achieving spatial flatness without dark energy is also accomplished quite successfully , with @xmath231 .
this result is even more impressive given the fact that to produce these numbers , one must first normalize the calculations to some pre - specified value of the observed matter density , @xmath109 ; and thus any uncertainties in that observed density will translate themselves into error bars on the @xmath107 values output from the simulations and with a factor _
greater _ than unity , since @xmath232 . for all of the results presented in this paper ,
we have used @xmath233 .
if one assumes that this value is more likely to be too high than too low , then models here with results of @xmath234 would be somewhat better than those with @xmath235 ; but in any case , all of the models in table [ tablernlt0satbestruns ] are clearly successful at effectively achieving flatness . in terms of providing a sufficient ( apparent ) acceleration ,
these models are again successful , with @xmath236 , effectively bracketing the concordance value of @xmath237 .
we also note that since the precise cosmological parameters obtained from snia fits are highly model - dependent ( e.g. , * ? ? ?
* ) , it is therefore not necessary for alternative - cosmology models to _ exactly _ reproduce the @xmath238 result from concordance @xmath1cdm , but merely to generate an ` accelerative ' value of @xmath238 strong enough to provide a sufficiently good fit to the supernova data , as has been done here .
one final quantity of great importance to be reproduced by an alternative concordance is the cmb acoustic scale , @xmath114 a cosmological parameter that weighs the cosmic evolution over vast distances in space and time .
we see here that for the concordance value of @xmath239 ( as computed using a toy @xmath1cdm model with no radiation ) , excellent matches are achieved by @xmath143 with @xmath216 , and @xmath141 with @xmath218 ( which is why we chose these @xmath143 and @xmath141 models to show in figure [ figclgt1strongplots ] over the other highlighted ones in table [ tablernlt0satoptimruns ] , despite their slightly worse @xmath195 values ) ; and sufficiently good matches are obtained by the remaining three models in table [ tablernlt0satbestruns ] .
although a discrepancy of @xmath240 with respect to the value expected from @xmath1cdm concordance is certainly something that would be noticeable from observations , we will see later on ( in section [ subearlysatclumping ] ) that one can do even better than this with more refined models ( for the @xmath139 case , particularly ) .
the last remaining output parameter quoted in table [ tablernlt0satbestruns ] is the `` jerk '' or `` jolt '' parameter , @xmath241 . as discussed at length in bbi , rather than serving as a well - observed cosmological parameter to be matched by theories aiming for an alternative concordance , the ( still poorly - measured ) value of @xmath241
can be viewed as a _ prediction _ of one s theory , useful for ultimately distinguishing such theories from cosmological constant dark energy ( since flat @xmath1cdm is always constrained to @xmath242 , for _ any _ value of @xmath243 ) .
we recall that the old version of our causal backreaction simulation program , without rnl , obtained a set of best - fit models which all possessed very high values of the jerk parameter
specifically , @xmath244 for the half - dozen best runs providing a clear and distinct way of observationally distinguishing our causal backreaction paradigm from a cosmological constant , or from any other slowly - evolving form of dark energy reasonably close to @xmath1cdm .
but here the results in table [ tablernlt0satbestruns ] show that when rnl is properly incorporated into our formalism , the situation becomes far less clear : the best - fitting @xmath143 and @xmath141 models now prefer _ low _ values of the jerk parameter ( @xmath245 and @xmath246 , respectively ) ; and the best - fit @xmath139 model , though still predicting a value higher than unity , now only goes so high as @xmath247
. a further analysis below in section [ subearlysatclumping ] will demonstrate that even this result may be subject to substantial change ; and thus the search for a reliable observational method for ` falsifying ' our causal backreaction paradigm ( that is , materially distinguishing it from @xmath1cdm ) will likely be more difficult than what was hoped for in bbi , where the possible impact of recursive nonlinearities was still only dealt with in terms of qualitative caveats
. with this analysis of our ` output ' results done , we now look at the _ input _ parameters which were necessary to produce these good cosmological fits .
first of all , considering @xmath132 , we see from table [ tablernlt0satoptimruns ] that to obtain good cosmological output parameters from the @xmath143 and @xmath141 models , we have had to impose an _ extremely late _
value for the time representing the ` beginning ' of structure formation specifically , @xmath248 .
this is a direct result of the effects of rnl , which for an already - perturbed universe will slow down the continued propagation of old perturbations coming in from great cosmological distances , thus rendering the causal backreaction contributions of early - developing inhomogeneities to the late - time ` acceleration ' much weaker than they were for the equivalent models simulated in bbi .
this is a serious problem for the @xmath141 and ( even more so ) for the @xmath143 functions , which are especially designed to provide more early clustering , at the expense of late - time clustering .
we must ask , in fact , whether input values like these for @xmath132 even make astrophysical sense , given the fact that @xmath249 actually marks the epoch during which the large - scale collapse of material and the resulting galactic feedback ( from star formation , etc . )
begins acting strongly to shock - heat cosmic baryons , sending a significant portion of the baryons into the superheated igm ( e.g. , * ? ? ?
* ) , thus helping to _ slow down _ further structure formation .
this problem goes away , however , for the @xmath139 runs , which have enough late - time clumping to remove the need for choosing a late ( i.e. , small ) value of @xmath132 to get good output cosmological parameters .
the runs only get better , in fact , as one goes to higher @xmath132 , and we stop at @xmath176 not because optimization tells us to ( though increasingly early initialization times eventually brings diminishing returns ) , but because it is not physically reasonable to set the start of clustering too long before the onset of reionization . besides this success of providing an alternative concordance with reasonable @xmath132 values , some reflection shows that the @xmath139 runs are the most sensible in every regard . as we now consider the @xmath199 function to represent the feedback effects of completed clustering on several different , dynamically - defined scales in hierarchical structure formation rather than as a simple percentage of ` clumped ' versus ` unclumped ' matter
the causal backreaction process becomes more naturally focused upon structure formation in the highly nonlinear regime of density perturbations . and as we recall from above , density fluctuations in the nonlinear regime evolve as @xmath126 with @xmath250 for a matter - dominated universe , meaning that @xmath199 functions evolving as @xmath251 ( or functions growing with even larger powers of @xmath15 ) are therefore the most relevant models for causal backreaction , rather than those with smaller powers of @xmath15 like the @xmath141 or @xmath143 models ( the latter type evolving as @xmath121 in the _ linear _ regime ) , which had been the dominant ones for generating backreaction in the old simulations without rnl . considering the above results once more in this light
, we see that @xmath139 has indeed performed better than @xmath141 or @xmath143 , for example looking much more like best - fit @xmath1cdm in figure [ figclgt1strongplots ] , thus providing better snia fits with significantly lower @xmath195 values ( recall tables [ tablernlt0satoptimruns],[tablernlt0satbestruns ] ) .
furthermore , the output cosmological parameters from the @xmath252-optimized @xmath139 runs ( with sufficiently large @xmath132 values ) were all extremely good , with just the possible exceptions of a slightly low total density ( e.g. , @xmath253 for @xmath176 ) , and a slightly errant cmb angular scale ( @xmath254 ) ; but even those discrepancies are quite small given the very simplified nature of these clumping evolution models .
( and those discrepancies can be made even smaller with some further development of the models , as will be seen shortly , in section [ subearlysatclumping ] ) .
the only real problem with these @xmath139 models , and quite an obvious one , is the very high degree of clumping needed in order to generate these results that is , @xmath255 .
this is right around the extreme upper edge of the ` reasonable ' range of values for @xmath131 , which we previously argued should likely lie within @xmath207 ( and more conservatively , within @xmath206 ) .
such @xmath214 values for @xmath139 ( which are greater than those for @xmath141 or @xmath143 ) are large enough as to potentially strain the credibility of causal backreaction as a cosmological solution . as we will see below , however
, this concern can be alleviated simply by adding in one important and physically reasonable model parameter as part of these clumping evolution functions : specifically , a parameter representing information about the possible _ end _ of clumping , as could be triggered in the real universe by galactic feedback .
as our clumping evolution models have so far been defined , each takes on its ` ultimate ' value , @xmath131 , at @xmath66 ; and this may make physical sense in a universe where the effective clumpiness is still growing , so that the value @xmath256 is not particularly special , but just represents a ` snapshot ' of @xmath44 taken at our particular observation time .
but if there is instead some halt to the growth of clumping due either to the limiting value of @xmath257 for our pre - rnl simulations , or due to various feedback effects in these new models ( which no longer possess a pre - determined upper bound on @xmath44 ) then from a copernican view , there is nothing necessarily special about the current cosmic epoch , and so there is no reason to assume that the asymptotic value of @xmath44 ( if there is one ) should occur specifically at the present time .
having already re - evaluated our formalism by interpreting the implementation of @xmath201 as representing multiple levels of structure on different cosmic scales , it has become possible for @xmath44 to continue increasing without any hard theoretical limit for @xmath258 , perhaps considerably beyond the effective degree of clumping that exists today . on the other hand ,
if there _ does _ happen to be some naturally - occurring , effective limit of clumping , then there is no reason to believe that the processes in charge of cosmic clustering would wait until @xmath22 to reach it . as noted above , @xcite
show that the nonlinear collapse of material and feedback from star formation have likely acted to shock - heat cosmic baryons to millions of degrees , thus inhibiting clumping the baryonic part of it , at the very least by keeping and/or sending a significant portion of the baryons into the superheated igm ; their figure 1 shows simulation results depicting the rapid increase of very hot baryons in the igm beginning strongly as of @xmath249 .
additionally , the backreaction effect itself may be helping to put the brakes on structure formation at late times ; particularly given the fact that it is generated most strongly within the vicinity of the virializing masses themselves , in high - density regions . in that case ,
signs of a cosmologically recent slow - down in the growth of clustering ( e.g. , * ? ? ?
* ) findings which are viewed as strong evidence in favor of dark energy may instead be indicating the effects of causal backreaction .
( conceivably , feedback from backreaction may also be a contributor to other poorly understood cosmological effects such as galaxy downsizing @xcite , the cuspy cdm halo problem , and the possible dearth of dwarf satellite galaxies ( e.g. , * ? ? ? * ) . )
what all of this implies is that our clumping evolution functions , as they have been defined in equations [ eqnclumpmodels]a - c , would likely not remain meaningful all the way to @xmath22 ; and thus it is useful to consider how the functional form of our models , @xmath44 , might be altered to account in some manner for this effect , to represent the possible late - time softening of the growth of clustering due to astrophysical and/or backreaction - related feedback processes .
the most straightforward way to model this behavior is by implementing an early `` saturation time '' , @xmath259 ( with @xmath260 $ ] ) , representing the time _ before _
the present epoch at which point the growth of clustering became saturated to a final , fixed value , @xmath261 .
this new clumping evolution function then fixes @xmath44 to remain equal to @xmath131 all the way from @xmath262 to @xmath66 .
recalling from section [ subsuperunityclumpstrong ] that the @xmath143 and @xmath141 models only obtain their best fits and their most ` concordant ' cosmological parameters for very late @xmath68 values ( i.e. , @xmath248 ) , it then makes little sense to outfit those models with an even later @xmath263 ; we therefore make this modification only for the @xmath139 model , which was optimized for concordance with much earlier @xmath68 values . the resulting ,
retrofitted @xmath139 function for early saturation is thus defined as : @xmath264 this is obviously a very elementary way of accounting for how astrophysical feedback acts to slow down the growth of clustering ( with the resultant lessening of its causal backreaction ) ; but even such an oversimplified model will provide us with valuable cosmological lessons . as a first run to explore the implications of early saturation ,
we consider again our earlier , cosmologically successful run with @xmath139 ( functionally identical to @xmath265 with @xmath266 ) , which had used @xmath176 and its corresponding @xmath267 ; these results are now compared to those from a new @xmath265 run still using @xmath176 and ( the now non - optimal ) @xmath268 , but with @xmath269 now imposed a conservatively early value suggested by the onset of substantial hot baryon injection into the igm .
the resulting residual hubble diagrams for the two cases are compared in figure [ figearlyvslatesat ] , where the powerful impact of moving from @xmath266 to @xmath269 upon the time - dependence of the observed backreaction can clearly be seen .
not unexpected is the strongly - amplified apparent acceleration effect at early times ( @xmath270 ) , due to the greatly compressed timescale over which clustering grows from @xmath45 to @xmath271 .
but what may be surprising is how dramatically the effects of these early - developing inhomogeneities become virtually irrelevant at late times , with the effective ` acceleration ' almost completely fading out by @xmath272 , and the residual hubble diagram relapsing to a nearly perfect scdm ( i.e. , decelerating ) cosmology not long after @xmath273 .
this is in sharp contrast to what would be expected from the results explored in bbi , where early - developing clumping provided the strongest ongoing causal backreaction effects due to the ever - growing `` inhomogeneity horizons '' of observers .
the fact that this is no longer the case is clearly due to rnl , which damps such ongoing backreaction effects by slowing down the causal propagation of inhomogeneity information ( as well as by diluting the perturbation effects of already - seen inhomogeneities via the extra volumetric expansion ) , thus greatly restricting the cosmological horizon out from which developing perturbations can affect the observer .
as can be inferred from equations [ eqnroftintegration],[eqnalpharnlintegration2 ] above , @xmath274 for inhomogeneity information propagating towards an observer at null - ray speed ; and so models with very strong early clumping particularly those with a very early @xmath275 drive up the value of @xmath155 so close to unity , so early on , that the propagation of inhomogeneity information is practically frozen to a halt at later times .
the cessation of new inhomogeneity information reaching the observer leads to a corresponding freeze in the continued evolution of @xmath155 ; and as is obvious from our metric given by equation [ eqnfinalbhpertmetric ] , a constant @xmath276 value can simply be transformed away via redefinitions of @xmath15 and @xmath277 , thus making a cosmology with static @xmath278 look exactly like a decelerating , matter - dominated scdm universe .
( interestingly , a considerably stronger late - time acceleration effect can be generated by using a _ smaller _ value of @xmath131 for these very early @xmath275 runs , which lessens this ` freezing ' effect .
the best possible snia fit for this @xmath269 case is therefore achieved with the relatively low value of @xmath279 ; though even that remains a very poor fit in absolute terms . )
the importance of such behavior is that it limits the degree to which causal backreaction with rnl can be ` self - powered ' via the ever - expanding reach of observational horizons , in defiance of a static final @xmath131 value ; a result which increases its dependence upon being ` driven ' by a continually - growing @xmath44 function . as figure [ figearlyvslatesat ]
has shown , an early saturation of clumping can lead to a virtually complete shutdown of apparent acceleration not much later .
( though noting again the caveat that nonlinear gravitational effects are not accounted for in any of our models and that they may indeed be important here , since all of the very - early saturation models with @xmath269 which we tested do have excessively large @xmath280 values , all significantly exceeding @xmath281 , with some nearly approaching unity . )
these considerations will have important implications for the eventual fate of the universe , as will be discussed below in section [ subnewfuture ] ; but first , we consider their ramifications regarding the _ past _ cosmic ` acceleration ' that has already been observed via the supernova data sets . since a value of @xmath269 effectively terminates the impact of causal backreaction far too prematurely for it to account for the apparent acceleration seen in the snia data from around @xmath273 to now , we consider runs with @xmath275 values that are similar to or smaller ( i.e. , later ) than this epoch of acceleration . specifically , we choose values of @xmath282 for study
i.e. , @xmath283 representing times by which the mass fraction of cosmic baryons possessing temperatures of @xmath284k has increased to reach @xmath6@xmath285 ( * ? ? ?
* figure 1b ) .
these values of @xmath275 are also well - placed to further illuminate the results of @xcite , which demonstrated the significant alteration in clustering behavior ( due either to dark energy or to causal backreaction ) between one cluster sample with @xmath286 ( and particularly its most distant subsample with @xmath287 ) , versus their more nearby cluster sample at @xmath288 .
re - optimizing @xmath131 for each of these chosen @xmath275 values ( with @xmath176 fixed for all of these runs ) , the resulting ` best runs ' with early saturation are those with the parameters : @xmath289 , @xmath290 , and @xmath291 ; we compare this to the best - fitting @xmath266 run , which had @xmath267 .
residual hubble curves for these four runs are plotted in figure [ figvaryingearlysatzval ] , where we see that while these early - saturation runs are no longer effectively indistinguishable from best - fit @xmath1cdm ( as was the @xmath266 run ) , they are nevertheless very close to it particularly within the redshift range containing most of the snia thus providing quite good fits to these data . the complete fit quality results and cosmological output parameters for these runs are presented in table [ tablernlearlysatruns ] .
first , we note that there are limits to the largest value of @xmath275 which may practically be used , since we see that as @xmath275 gets as high as @xmath292 , the fit probability @xmath104 gets progressively worse ( i.e. , @xmath195 begins to grow unacceptably large ) , the cosmological parameters become increasingly ` discordant ' ( particularly @xmath107 and @xmath114 ) , and there are even concerns that the formalism itself begins to break down due to gravitationally nonlinear effects , given the disturbingly large value of @xmath280 . if we restrict ourselves to @xmath293 , however , then at a cost of only a small increase in @xmath195 , the situation gets considerably better with the use of nonzero @xmath275 in several crucial ways . considering the @xmath294 case in particular , moving away from @xmath266
has improved the match to the cmb data and the verification of spatial flatness ( better @xmath114 and @xmath107 values , respectively ) , while remaining essentially as good in terms of the other cosmological parameters , with just a tiny degradation in the fit probability ( by @xmath295 ) .
but most importantly , we see that all of this been achieved with a _ much _ lower @xmath131 value
table [ tablernlearlysatruns ] showing how @xmath214 decreases with increasing @xmath275 , in general dropping all the way from @xmath267 for @xmath266 , to @xmath296 for @xmath294 .
this is now well within the @xmath297 range of ` reasonable ' values for hierarchical clustering , as was specified earlier in section [ subsuperunityclumpstrong ] .
( in addition , we have also conducted fits to a more recent supernova data set , the scp union2 snia compilation @xcite ; and when we minimize @xmath195 for the @xmath294 case with respect to the union2 data , rather than the union1 data , we still get similar results : an excellent fit with @xmath298 ( compared to @xmath299 for @xmath1cdm ) and @xmath300 ; only a slightly higher clumping strength value of @xmath301 ; and similar cosmological parameters that are also quite acceptable .
only the unperturbed matter density is slightly high with @xmath302 ; but as discussed above , this may be due to the adopted value of @xmath233 being slightly too large . ) in consequence , it is justified to say that we have in every sense re - obtained a successful alternative concordance with causal backreaction in the presence of recursive nonlinearities , having generated the proper amount ( and temporal behavior ) of apparent acceleration as evidenced by a fit to the snia data that is very nearly as good as that achievable with best - fit @xmath1cdm ( even _ without _ performing a rigorous @xmath195-minimization search over our model parameter space ) as well as having produced output cosmological parameters that are more than acceptably consistent with a variety of complementary cosmic measurements . the price to be paid to achieve this goal , due to
the incorporation of recursive nonlinearities , is the necessity of permitting @xmath131 values in excess of unity which though going against the original interpretation of this ( heuristic ) model parameter , as defined in bbi , does seem justifiable in a realistic cosmology exhibiting hierarchical structure formation on a variety of scales . on the other hand , a new benefit of these results is the shifted focus from @xmath143 models to @xmath139 models , equivalent to a shift in emphasis from linearized matter density perturbations to nonlinear density perturbations ; a new emphasis which in fact makes much more sense for a causal backreaction paradigm depending upon vorticity- and velocity - dispersion - generated virialization as the fundamental origin of structure - induced perturbations to the observed cosmological metric .
one last ( yet very important ) consideration still remains regarding early saturation , though , relating to its impact upon the very late - time behavior of the cosmic evolution .
one way to approach this , is to consider _ why _ the optimized value of @xmath131 drops so precipitously with increasing @xmath275 , so conveniently solving our problem by reducing @xmath214 to believable values . to understand why this happens ,
consider the clumping evolution functions , themselves , for the early saturation models that we have studied ; plots of these @xmath44 functions are shown in figure [ figpsitplotsvszsat ] . from these plotted curves ,
we see that despite the vast differences in @xmath214 for the @xmath266 case versus those with @xmath303 , as long as one does not use too large a value of @xmath275 ( sticking to , say , @xmath293 ) , then the degree of clumping @xmath44 at _ mid - range _ values of @xmath43 ( e.g. , @xmath304 ) actually remains fairly similar from run to run .
the distinctively huge increase from @xmath305 to @xmath306 for the @xmath266 case does not actually happen until very late times , @xmath307 .
but why should completely different behaviors of @xmath44 at late times have such a small effect upon the fits to the snia data ?
the answer is twofold : first , due to a lack of sensitivity of the data to such changes ; and second , due to the nature of causal backreaction , itself .
for the first issue , consider that most of the snia included in these supernova compilations for measuring the cosmic acceleration are located at fairly high redshift ; it is therefore unsurprising that the evolution of @xmath44 _ after _ the epoch of these supernovae should have little effect upon the acceleration measured by those high-@xmath43 snia , regardless of whether @xmath44 experiences a continued increase or a saturation .
but , even if lower-@xmath43 snia were also included in the study , and could be trusted to accurately map out the hubble flow despite their local peculiar motions , there should still be little effect other than a late - time offset , perhaps registered as a small change in @xmath105 . as noted by @xcite , simple cosmic evolution functions that are relatively insensitive to time
variations tend to measure an averaged cosmological equation of state around a ` pivot ' redshift of about @xmath308 .
and so it is understandable why the detailed behavior of @xmath44 after this crucial epoch should end up having little effect upon the precise amount of apparent acceleration measured . for the second issue , we recall from equation [ eqniintegrandprelim ] that the amount of causal backreaction due to a spherical shell at coordinate radius @xmath12 will be proportional to @xmath309 , and so very late clumping which corresponds to relatively small look - back times @xmath55 , and thus small distances @xmath12 will simply not involve a sufficiently large amount of inhomogeneous mass to generate significant causal backreaction .
thus the detailed dynamics of inhomogeneities at higher @xmath43 ( and larger @xmath12 ) assuming distances not so far away as to get damped by rnl , or having look - back times so far back relative to @xmath68 as to have very small clustering , @xmath310 will be more important in determining the detailed effects of causal backreaction than would the very late ( @xmath111 ) behavior of clumping .
the overall result is that there is a parameter degeneracy for causal backreaction models , where one can reduce @xmath131 in tandem with an increase in @xmath275 , without much change in the quality of the fit to the actual snia data .
this degeneracy has a beneficial aspect , in that it has allowed us to successfully create an alternative concordance with causal backreaction and to do so in a flat , matter - only universe without any form of dark energy while using astrophysically realistic values of @xmath131 .
but it has a negative aspect , as well , in that it lessens our ability to predict precise ranges for the late - time cosmological parameters that would be output by the various ( successfully concordant ) models in this formalism .
measurable cosmological parameters such as @xmath105 , @xmath238 , and @xmath241 are defined in theory , at least in terms of the @xmath311 behavior of taylor expansions of the luminosity distance function , @xmath312 .
hence , while a variety of @xmath44 functions may all provide good fits to the key snia located at mid - to - high redshifts , these functions may all have quite different @xmath313 behaviors , and thus very significant differences in their cosmological output parameters . in particular , consider again the first three of the four runs in table [ tablernlearlysatruns ] .
as one goes from @xmath266 to @xmath314 , we see that @xmath105 decreases below the @xmath1cdm concordance value of @xmath315 ; but yet , the difference is small enough to not be a serious concern . for @xmath238 ,
obtained from one more differentiation of @xmath316 , the difference is larger , going from @xmath317 to @xmath318 ; and yet , the precise numerical value of the cosmic ` equation of state ' @xmath238 is of no great concern to us here since it is not our task to pin down the physics of some hypothesized form of dark energy , but merely to reproduce the cosmological observations
so long as a good fit with a sufficient amount of apparent acceleration can be produced , as has indeed been accomplished .
the biggest change , however , occurs for @xmath241 obtained via yet another differentiation of @xmath316 which drops all the way from @xmath247 for @xmath266 , down through the @xmath1cdm value of @xmath319 and far past it , even going negative for @xmath275 values that still provide decent snia fits .
a small part of this change may be due to the highly simplified nature of the @xmath265 functions that we use ; in particular , since we obtain a hubble curve through two integrations of @xmath44 that is , @xmath320 , and then @xmath321 ( cf . equations [ eqnitotintegration]-[eqndlumdefn ] ) the @xmath241 parameter ( obtained from three differentiations of @xmath316 ) essentially contains one differentiation of @xmath44 .
as is obvious from figure [ figpsitplotsvszsat ] , our simple @xmath265 functions are not differentiable at @xmath275 , meaning that @xmath322 ( and thus @xmath323 ) will also be non - differentiable there , and so @xmath324 will pick up an actual discontinuity there .
still , the discontinuity is not large enough to account for the huge overall differences in @xmath241 for the different @xmath275 values ; and most of the effect of the jump in @xmath324 at @xmath275 is almost certainly real , due to the real change in the cosmic evolution at that point , and would likely still take place to a very similar degree ( just more spread out in time ) for some more complicated @xmath265 functions designed to apply smoothing at @xmath275 . in any case
, the set of runs in table [ tablernlearlysatruns ] does fairly clearly establish the trend that increasing @xmath275 away from zero leads to a steady decrease in the final value of @xmath241 .
what we can conclude with some certainty , unfortunately , is that we have lost any reasonable degree of predictability for this @xmath241 parameter in the causal backreaction paradigm , due the weak dependence of the backreaction upon the very - late - time behavior of @xmath44 , and the resulting degeneracy in @xmath325 .
this is significant , because of the different possible ways suggested in bbi to distinguish between causal backreaction and cosmological constant @xmath1cdm ( or any similar form of dark energy ) , in order to provide our paradigm with a falsifiable test , the clearest signature by far was the search for @xmath326 .
for all intents and purposes , the reliability of this signature now appears to be gone , and some more intricate means will obviously be needed to distinguish causal backreaction from even the simplest , @xmath1-like version of dark energy ( though @xmath327 would still rule out @xmath1cdm in favor of _ some _ alternative model ) .
thus while the use of early saturation has greatly enhanced the prospects for achieving an alternative cosmic concordance with realistic values of the clumping parameter @xmath131 , the incorporation of this additional degree of freedom in the models has made it significantly more challenging to definitively replace the paradigm of dark energy with that of causal backreaction , without resorting to aesthetic arguments of a subjective nature .
if appropriate observational tests should eventually be able to demonstrate causal backreaction as superior to dark energy as the driver of the cosmic evolution , then one final question of great importance would of course be : what is the ultimate fate of the universe ?
this topic was discussed in detail in bbi , in which it was found that the long - term cosmic fate depends upon which way the balance tips between the forces working to power the cosmic ` acceleration ' , versus the opposing factors acting to restrain it . for the former ,
the only real influence working to keep the apparent acceleration going ( and perhaps ultimately promote it in strength to a real volumetric acceleration ) was the ever - expanding causal horizon , growing in time , out from which an observer can ` see ' substantial inhomogeneities
i.e. , the farthest distance from the observer out to which @xmath328 , @xmath329 are still true . for opposition to the acceleration , one source of restraint is the frw expansion of the universe itself , which exerts a natural damping effect upon causal backreaction by diluting the density of inhomogeneities , simply by pulling them farther away from one another ( and from any given cosmological observer ) over time . also acting to limit a possible long - term acceleration was the inevitable saturation presumed for @xmath44 , which in our original interpretation of this measure of clumping , as described in bbi would be limited by an upper bound of @xmath330 .
considering all of these factors , analytical approximations were derived for the ( pre - rnl ) metric perturbation function @xmath85 as @xmath331 ; and it was found that @xmath85 always asymptotes to a constant numerical value for @xmath44 functions evolving as a simple power of @xmath15 , @xmath332 .
this asymptotic value of @xmath85 is larger for smaller @xmath333 , with it being equal to unity ( implying a complete breakdown of the newtonianly - perturbed metric ) for @xmath334 ( i.e. , @xmath335 ) .
thus a fully general - relativistic acceleration at late times due to causal backreaction did appear to be realistically possible , according to the formalism of bbi though of course that conclusion was made pending the still - undetermined effects of recursive nonlinearities , and other complications . in this paper
, however , we will have to revise our expectations , since the incorporation of rnl ( recursive nonlinearities ) has made the prospects of an ` eternal ' , self - powered acceleration seem far less likely
. we can conclude this despite the fact that @xmath85 is no longer necessarily amenable to a simple analytical analysis , even for @xmath331 .
one point in favor of such a conclusion is that rnl has forced us to change to @xmath332 functions with larger exponents @xmath333 in order to re - establish a concordance
i.e. , moving from @xmath143 and @xmath141 before , to @xmath139 now and as our previous results have shown , larger exponents in @xmath44 lead to smaller late - time values of @xmath85 ( even if we no longer know its exact asymptotic behavior ) , thus implying a less general - relativistic , less ` accelerative ' long - term evolution . but
an even more important factor is what we learned about the impact of the ultimate ` saturation ' of the clumping evolution function @xmath44 at some final , fixed numerical value , as was depicted in figure [ figearlyvslatesat ] .
this result ( and other simulation runs that we have done along these lines ) clearly demonstrate that rnl , which acts to stall the expansion of an observer s `` inhomogeneity horizon '' particularly so in cosmologies with strong early - time backreaction will largely ` shut off ' all apparent acceleration effects not long after the ongoing clustering and virialization have ceased , once @xmath44 has settled down to some mostly - constant final value .
this behavior certainly works against the possibility of an ` eternally ' ( or even long - term ) ` accelerating ' universe .
the one enhancement introduced in this paper which might possibly help lead to a long - term acceleration , is the now fundamentally unbounded nature of @xmath44 due to hierarchical clustering ( assuming that an effective end to clustering is not imposed explicitly by early saturation , as we did in fact choose to impose for our models in section [ subearlysatclumping ] ) .
but while this change in interpretation allowing @xmath336 is indeed based upon real physics , is it strong enough to keep the ` acceleration ' going continually , deep into the future ?
the answer to this question is naturally uncertain : on ( relatively ) small scales , clustering has never ceased , as star clusters and small galaxies continue to merge ( with new virialization ) into large galaxies ; galaxies continue to merge into galactic clusters ; and so on .
yet , one would realistically expect to find steadily diminishing returns on such ` small ' scales .
any real hope for a continuing acceleration would seem to rest upon the future clustering behavior on extraordinarily _ large _ scales a realm of structure formation with no real upper limit , as superclusters eventually manage to self - virialize internally , then begin themselves to merge into even larger structures , and so on , ad infinitum .
higher and higher scales of clustering take exponentially longer to complete than the levels below them , though , and it is not clear how large the effective ` inhomogeneity horizon ' for causal backreaction can ever practically become especially given the fact that the cosmic acceleration itself tends to impede the expansion of observational horizons . a truly long - term cosmic acceleration ( apparent or real ) into the far future
would therefore seem very difficult to accomplish using causal backreaction with recursive nonlinearities ; though such conclusions can not be considered definitive without more detailed cosmological simulations , employing a significantly more sophisticated treatment of the effects of vorticity and virialization than we have used in these calculations so far , given our toy models for @xmath44 , and therefore @xmath85 .
there are still two remaining wildcards , however , which were discussed in bbi and must be mentioned again here .
the first one is the possible advent of truly general - relativistic perturbations , causing the breakdown of our newtonianly - perturbed metric approximation ( equation [ eqnangavgbhmatdomweakflat ] ) , and the failure of our treatment of the individual perturbations due to the multitude of locally - clumped masses as being linearly summable .
the possibility of such a complication still exists , as is made obvious ( for example ) by the very large @xmath280 values that we typically find for early - saturation runs with large @xmath275 settings ( e.g. , the @xmath337 case from table [ tablernlearlysatruns ] ) . as noted above in section [ subearlysatclumping ]
, the fact that the propagation of inhomogeneity information obeys the proportionality @xmath274 , means that the flow of such information will be choked off whenever @xmath276 approaches unity , thus freezing the expansion of the relevant inhomogeneity horizons , thereby locking @xmath155 nearly in place at whatever actual value that it has managed to grow to by that time .
( this is in fact the biggest reason why the @xmath269 model from figure [ figearlyvslatesat ] experienced such an abrupt shut - off of its apparent acceleration so quickly after the growth of its clumping evolution function @xmath44 had ceased . )
what we can not know from our formalism , however , is how to physically interpret the effects of a metric perturbation function @xmath155 that is hovering around a nearly fixed value , but where that value is always slowly asymptoting towards unity . does it look almost exactly like a decelerating scdm universe , since a constant @xmath155 can simply be transformed away through coordinate redefinitions ? or would it instead be a universe perpetually appearing to be right on the verge of undergoing a runaway acceleration , just never quite doing it , basically ` riding the edge ' forever ? ( perhaps succeeding , at least , at producing a hesitating , stop - and - go accelerating behavior . ) or on the other hand , does the proximity of @xmath155 to unity manage to override all other considerations , and lead to an actual , volumetric , runaway acceleration ?
the real answers to these questions can not be determined without at least some nonlinear gravitational terms being added into the formalism , if not actually requiring a fully general - relativistic treatment .
but while the quantitative calculation of such effects to determine their physical implications is beyond the scope of the causal backreaction formalism that has been presented here ( or in bbi ) , it is clear that the sum of innumerably many newtonian - level perturbations is indeed capable of driving the total cosmological metric perturbation right up to the breaking point , where a ` real ' cosmic acceleration ( by any definition ) may very possibly take over , and for an indeterminate period of time .
last of all considerations , though , is the one that can least safely be neglected : and that is the eventual breakdown of our smoothly - inhomogeneous approximation , as increasing clumpiness leads to fundamentally non - negligible anisotropies across much of the observable universe itself ( e.g. , the dark flow of * ? ? ? * ) ; a scenario which in bbi was termed the `` big mess '' .
this final , virtually certain breakdown of the cosmological principle is the ultimate game - changer , and questions about the possibility of a ` permanent ' cosmic acceleration due to causal backreaction then become just as hard to define as they are to answer , as the fundamental frw basis of cosmological analysis finally breaks down entirely .
in this paper , we have revisited the causal backreaction paradigm introduced in @xcite , for which the apparent cosmic acceleration is generated not by any form of dark energy , but by the causal flow of information coming in towards a typical cosmological observer from a multitude of newtonian - strength perturbations , each one due to a locally clumped , virializing system .
self - stabilized by vorticity and/or velocity dispersion , such perturbations are capable of generating positive volume expansion despite their individually - newtonian natures . noting that previous ` no - go ' arguments against newtonian - level backreaction are based upon non - causal backreaction frameworks , we see that the sum total of these small but innumerable perturbations adds up to an overall effect that is strong enough to explain the apparent acceleration as detected by type ia supernovae , as well as permitting the formulation of an alternative cosmic concordance for a matter - only , spatially - flat universe .
our purpose here has been to develop and test a second - generation version of this causal backreaction formalism , filling in one of the most important gaps of the original ` toy model ' by including what we have termed `` recursive nonlinearities '' specifically referring to the process by which old metric perturbation information tends to slow down the causal propagation of all future inhomogeneity information , therefore reducing the effective cosmological range of causal backreaction effects , and thus damping the strength of their overall impact upon the cosmic evolution and upon important cosmological observations
. utilizing the new simulation program introduced here , which now incorporates recursive nonlinearities into causal backreaction , we find profound differences in the resulting cosmological model calculations .
for a given magnitude of self - stabilized clustering assumed for large - scale structure , denoted by dimensionless model input parameter @xmath131 , the overall power of causal backreaction is now considerably weaker , in addition to fading out relatively rapidly after the growth of clustering ceases .
this is unlike the results of the original model , in which causal backreaction effects would continue to grow regardless of any late - time saturation of clustering , due to the causally - expanding `` inhomogeneity horizon '' seen by an observer which continually brings more ` old ' inhomogeneities into view from ever - greater cosmic distances .
after discussion of some of the possible reasons for which causal backreaction may now appear to fall short of its cosmological goals either due to issues regarding the fundamental mechanism itself , or due to our highly simplified treatment of it we then considered a very straightforward way in which the paradigm may be fully recovered as a cosmological replacement for dark energy : all that is needed is the adoption of @xmath131 values greater than unity . though representing an ad - hoc modification of the original formalism
, the change makes astrophysical sense in a number of ways . rather than viewing the clumping evolution function @xmath44 as simply representing the fraction of cosmic matter in the ` clumped ' versus ` unclumped ' state at any given time ,
@xmath44 can now be recognized ( more realistically ) as representing the total backreaction effect of hierarchical structure formation in the universe , where clustering and virialization take place simultaneously on a number of different cosmic length scales from stellar clusters , to individual galaxies , to galaxy clusters , etc .
model input parameter @xmath338 is now interpreted as the effective number of ` levels ' of completed clustering that exists ( at current time @xmath66 ) in the large - scale structure when one sums over the ( partial or total ) clustering of matter on all relevant cosmic scales .
given this enlarged parameter space with @xmath201 now permitted , we once again find a selection of model cosmologies that succeed ( despite the damping effects of recursive nonlinearities ) at reproducing the observed cosmic acceleration , while also re - establishing an alternative cosmic concordance by producing output parameters that match the observables derived from several of the most important cosmological data sets .
furthermore , astrophysical considerations regarding the necessary _ input _ parameters for these apparently successful models specifically , the need to assume a sufficiently early beginning of clustering
results in a preference by the new formalism for @xmath44 models that reflect the late , nonlinear phase of structure formation .
this is an improvement over the old formalism without recursive nonlinearities which had preferred models that embody the early phase of clustering , with linearized matter fluctuations since it is this final , nonlinear stage of clustering during which virialization occurs via the generation of vorticity and velocity dispersion , and hence represents the more astrophysically reasonable source of substantial causal backreaction . noting that the only problem still remaining for this new concordance is the somewhat excessively large amount of clustering required to achieve it that is , @xmath339 , rather than what we consider to be more reasonable values like @xmath206 we then determined that this problem could be successfully fixed ( i.e. , a good concordance generated with @xmath340 ) by introducing `` early saturation '' , in which the clumping evolution function @xmath44 reaches its ultimate value of @xmath131 somewhat in the past ( @xmath341 ) , and then changes little thereafter .
this is a highly reasonable adjustment to the formalism , since in the real universe `` gastrophysics '' feedback exists which creates superheated baryons , sending large amounts of material back into the intergalactic medium , thereby slowing down the continued clustering of matter at late times ; not to mention the likely slowdown of clustering due to the feedback effects of the backreaction , itself .
the only major drawback of this new feature is the addition of an extra model input parameter
the epoch of saturation , @xmath275 which results in a degeneracy within @xmath325-space , providing a range of models that all fit the type ia supernova data well , yet lead to significant differences for certain output cosmological parameters . the greatest variation in the output results due to
this degeneracy occurs for the observable jerk parameter , @xmath0 , hence implying a loss of predictability for @xmath0 by our causal backreaction formalism .
this is a significant loss , given the previous findings from @xcite ( without recursive nonlinearities ) which had indicated that @xmath342 was the most distinctive signature of causal backreaction , thus serving as the clearest way for distinguishing it from cosmological constant @xmath1cdm ( or from anything close to it ) , since flat @xmath1cdm always requires @xmath242 .
it thus becomes more difficult to find a falsifiable test of the causal backreaction paradigm , a test that is needed to definitively distinguish it from dark energy in order to eventually rule out one cosmological approach in favor of the other .
finally , concerning the ` ultimate ' fate of the universe , we note that the incorporation of recursive nonlinearities tends to shut down any strong apparent acceleration effects fairly quickly once the ongoing clustering ( i.e. , the continued growth of @xmath44 ) finally stops .
even more dramatic is the way in which the metric perturbation function , @xmath155 , becomes essentially locked in place when approaching too close to unity , making it an even greater obstacle in terms of preventing the acceleration ( apparent or otherwise ) from completely taking over the cosmic evolution .
this makes the scenario of a perpetual , ` eternal ' acceleration seem even less likely than it already did in @xcite ; though the now - unbounded nature of @xmath44 could potentially provide some aid in producing a long - term acceleration , as long as virialized structure can continue to form on ever - larger cosmic scales , without any fundamental upper limit to the sizes of coherent structures .
furthermore , the question of the ultimate cosmic fate is once again complicated by the possible backreaction contributions of gravitationally nonlinear terms , and the ( unavoidable ) eventual breakdown of the approximation of the universe as `` smoothly - inhomogeneous '' both complications representing scenarios which our toy - model formalism is not presently designed to account for . in summary
, we conclude that our causal backreaction formalism remains successful at generating an alternative cosmic concordance for a matter - only universe , without requiring any form of dark energy ; though the necessary incorporation of recursive nonlinearities into these models implies that a significantly stronger amount of such backreaction than before is now needed , acting throughout the crucial ` acceleration epoch ' of @xmath343 or so , in order to provide a degree of observed acceleration sufficient to match the cosmological standard candle observations .
bochner , b. 2011 , preprint ( arxiv:1109.4686v3 ) ; a briefer overview version is available as : 2011 , preprint ( arxiv:1109.5155v3 ) buchert , t. , & ehlers , j. 1997 , 320 , 1 ; preprint ( arxiv : astro - ph/9510056v3 ) with @xmath344 , where the upper solid line represents the old version of the simulated cosmology without recursive nonlinearities ( rnl ) , and the lower solid line represents the new version with rnl .
shown along with them for comparison ( broken lines ) are the flat scdm and ( union1-best - fit , @xmath211 ) concordance @xmath345 cosmologies . ]
runs ( as described in the text ) , selected from the new simulations with rnl but with the choice of model input parameters restricted to those used in bbi ; specifically , these @xmath143 and @xmath139 curves both have @xmath186 .
also plotted here are the union1-best - fit flat scdm and concordance @xmath345 ( @xmath211 ) cosmologies .
shown along with these curves are the scp union1 snia data , here binned and averaged for visual clarity ( bin size @xmath346 = 0.01 $ ] ) .
each theoretical model is individually optimized in @xmath105 to minimize its @xmath195 with respect to the union1 snia data set ; and for simplicity , instead of moving the snia data up or down for each different optimized @xmath105 value , the optimization is depicted here by plotting the residual hubble diagram of the snia data versus a coasting universe of a single , fixed hubble constant ( @xmath347 ) , and then displacing each theoretical curve vertically , relative to the snia data , as appropriate for each fit . ]
( as described in the text ) , using model input parameters ( solid lines increasing from lowest to highest ) : @xmath139 with @xmath348 ( curve almost indistinguishable from @xmath1cdm ) ; @xmath141 with @xmath349 ; and @xmath143 with @xmath350 . shown along with them for comparison ( broken lines )
are the flat scdm and ( union1-best - fit , @xmath211 ) concordance @xmath345 cosmologies . ] with @xmath348 , where the solid line peaking at @xmath272 ( the curve almost indistinguishable from @xmath1cdm ) represents a simulated cosmology with rnl but restricted to @xmath266 ( thus rendering it equivalent to @xmath139 ) ; and where the solid line peaking at @xmath351 represents the shift to @xmath269 . shown along with them for comparison ( broken lines )
are the flat scdm and ( union1-best - fit , @xmath211 ) concordance @xmath345 cosmologies . ]
models with fixed @xmath176 and varying @xmath275 , with @xmath131 re - optimized for each chosen @xmath275 value . from lowest to highest ,
the solid lines represent the runs with model input parameters : @xmath352 ( curve almost indistinguishable from @xmath1cdm ) ; @xmath289 ; @xmath290 ; and @xmath291 .
also plotted here ( broken lines ) are the union1-best - fit flat scdm and concordance @xmath345 ( @xmath211 ) cosmologies . shown along with these curves are the binned and averaged scp union1 snia data ( bin size @xmath346 = 0.01 $ ] ) .
each theoretical curve is displaced vertically , relative to the snia data , to depict its individualized @xmath105-optimization . ] , plotted versus @xmath353 $ ] . in order of the positions of the functions ` corners ' from left to right ( increasing @xmath275 ) , the lines depict @xmath265 functions with the parameters : @xmath352 ; @xmath289 ; @xmath290 ; and @xmath291 . ]
ccccccccccccc @xmath132 & @xmath131 & @xmath195 & @xmath104 & @xmath280 & @xmath224 & @xmath105 & @xmath106 & @xmath93 & @xmath107 & @xmath238 & @xmath241 & @xmath114 + + 1.5 & 2.8 & 314.7 & 0.324 & 0.53 & 1.16 & 69.54 & 41.96 & 13.95 & 0.948 & -0.639 & 0.59 & 297.8 + 2 & 2.9 & 315.1 & 0.319 & 0.62 & 1.17 & 69.58 & 38.62 & 14.44 & 1.162 & -0.639 & 0.51 & 288.7 + + 2 & 3.2 & 313.6 & 0.340 & 0.53 & 1.16 & 69.75 & 42.10 & 13.88 & 0.946 & -0.675 & 0.89 & 297.3 + 3 & 3.3 & 314.0 & 0.334 & 0.62 & 1.16 & 69.68 & 38.72 & 14.30 & 1.159 & -0.667 & 0.83 & 287.2 + + 25 & 4.1 & 311.8 & 0.367 & 0.53 & 1.14 & 70.07 & 42.32 & 13.64 & 0.943 & -0.751 & 1.73 & 294.5 + + & & 311.9 & 0.380 & & 1.0 & 69.96 & 69.96 & 13.64 & 0.287 & -0.713 & 1.0 & 285.4 + + & & 608.2 & 3.4e-22 & & 1.0 & 61.35 & 61.35 & 10.62 & 1.0 & 0.0 & 1.0 & 287.3 + ccccccccccccc @xmath275 & @xmath214 & @xmath195 & @xmath104 & @xmath280 & @xmath224 & @xmath105 & @xmath106 & @xmath93 & @xmath107 & @xmath238 & @xmath241 & @xmath114 + + 0 & 4.1 & 311.8 & 0.351 & 0.53 & 1.14 & 70.07 & 42.32 & 13.64 & 0.943 & -0.751 & 1.73 & 294.5 + 0.25 & 2.6 & 313.5 & 0.326 & 0.58 & 1.15 & 69.60 & 40.24 & 14.00 & 1.054 & -0.620 & 0.15 & 289.7 + 0.5 & 2.3 & 316.6 & 0.284 & 0.68 & 1.15 & 69.40 & 36.32 & 14.65 & 1.338 & -0.585 & -0.14 & 279.8 + 1 & 2.2 & 320.2 & 0.238 & 0.80 & 1.14 & 68.77 & 29.54 & 15.75 & 2.086 & -0.488 & -0.94 & 259.9 + + & & 311.9 & 0.380 & & 1.0 & 69.96 & 69.96 & 13.64 & 0.287 & -0.713 & 1.0 & 285.4 + | we revisit the causal backreaction paradigm , in which the need for dark energy is eliminated via the generation of an apparent cosmic acceleration from the causal flow of inhomogeneity information coming in towards each observer from distant structure - forming regions .
a second - generation version of this formalism is developed , now incorporating the effects of `` recursive nonlinearities '' : the process by which metric perturbations already established by some given time will subsequently act to slow down all future flows of inhomogeneity information . in this new formulation ,
the long - range effects of causal backreaction are damped , substantially weakening its impact for simulated models that were previously best - fit cosmologies . despite this result
, we find that causal backreaction can be recovered as a replacement for dark energy through the adoption of larger values for the dimensionless ` strength ' of the clustering evolution functions being modeled a change justified by the hierarchical nature of clustering and virialization in the universe , occurring as it does on multiple cosmic length scales simultaneously . with this and
with the addition of one extra model parameter used to represent the slowdown of clustering due to astrophysical feedback processes , an alternative cosmic concordance can once again be achieved for a matter - only universe in which the apparent acceleration is generated entirely by causal backreaction effects .
the only significant drawback is a new degeneracy which broadens our predicted range for the observed jerk parameter @xmath0 , thus removing what had appeared to be a clear signature for distinguishing causal backreaction from cosmological constant @xmath1cdm . considering the long - term fate of the universe , we find that incorporating recursive nonlinearities appears to make the possibility of an ` eternal ' acceleration due to causal backreaction far less likely ; though this conclusion does not take into account potential influences due to gravitational nonlinearities or the large - scale breakdown of cosmological isotropy , effects not easily modeled within this formalism . |
17 natexlab#1#1bibnamefont # 1#1bibfnamefont # 1#1citenamefont # 1#1url # 1`#1`urlprefix[2]#2 [ 2][]#2 , , , * * , ( ) .
, , , , * * , ( ) . , * * , ( ) . , , , , , , , * * , ( ) . , , , * * , ( ) . , , , , , , , , * * , ( ) . , , , , * * , ( ) . , , , , , * * , ( ) . , , , , * * , ( ) . , , , * * , ( ) . , , , , * * , ( ) . , * * , ( ) . , , , _
( , , ) . , , , , , , * * , ( ) . , _
( , , ) . ,
* * , ( ) . , , , * * , ( ) . , , , , , * * , ( ) .
* fig . 1*.
( color online ) ( a ) the surface band structure for the ( 10 ) surface of a square phc slab of dielectric rods in air , the radius of rods is @xmath55 ( @xmath5 is the lattice constant ) , the height of the rods is @xmath56 , and the dielectric constant of the rods is @xmath57 .
the radius of the rods at the surface is @xmath58 .
( b ) the surface band structure for the @xmath59 surface of a triangular phc slab of air holes extending through a high - index ( @xmath2 ) finite - height dielectric slab , the radius of holes is @xmath45 , and the high - index slab is of thickness @xmath30 .
the distance between the centers of first right holes and the right boundary of the dielectric is @xmath60 . *
2*. ( color online ) ( a ) top view of the microcavity composed by a 2d square phc slab .
( b ) top view of the microcavity composed by a 2d triangular phc slab .
the length of the cavities is denoted by @xmath10 . *
3*. ( color online ) ( a ) field cross sections for the surface resonant mode in the square phc slab cavity , showing the @xmath16 component of the electric field .
( b ) and ( c ) field cross sections for the two surface resonant modes [ @xmath61 and @xmath62 in the triangular phc slab cavity , showing the @xmath16 component of the magnetic field .
solid lines are outlines of the rods / holes / slabs . *
4*. ( color online ) the quality factor ( q ) is plotted as a function of the length ( l ) for the microcavity in ( a ) a square phc slab and ( b ) a triangular phc slab .
corresponding structures are illustrated in fig .
( color online ) ( a ) the surface band structure for the ( 10 ) surface of a square phc slab of dielectric rods in air , the radius of rods is @xmath55 ( @xmath5 is the lattice constant ) , the height of the rods is @xmath56 , and the dielectric constant of the rods is @xmath57 .
the radius of the rods at the surface is @xmath58 .
( b ) the surface band structure for the @xmath63 surface of a triangular phc slab of air holes extending through a high - index ( @xmath2 ) finite - height dielectric slab , the radius of holes is @xmath45 , and the high - index slab is of thickness @xmath30 .
the distance between the centers of first right holes and the right boundary of the dielectric is @xmath64 .
, width=8 ] ( color online ) ( a ) top view of the microcavity composed by a 2d square phc slab .
( b ) top view of the microcavity composed by a 2d triangular phc slab .
the length of the cavities is denoted by @xmath10.,width=6 ] ( color online ) ( a ) field cross sections for the surface resonant mode in the square phc slab cavity , showing the @xmath16 component of the electric field .
( b ) and ( c ) field cross sections for the two surface resonant modes [ @xmath61 and @xmath62 in the triangular phc slab cavity , showing the @xmath16 component of the magnetic field .
solid lines are outlines of the rods / holes / slabs.,width=5 ] ( color online ) the quality factor ( q ) is plotted as a function of the length ( l ) for the microcavity in ( a ) a square phc slab and ( b ) a triangular phc slab .
corresponding structures are illustrated in fig .
2 . , width=8 ] | optical microcavities based on zero - group - velocity surface modes in photonic crystal slabs are studied .
it is shown that high quality factors can be easily obtained for such microcavities in photonic crystal slabs . with increasing of the cavity length , the quality factor is gradually enhanced and the resonant frequency converges to that of the zero - group - velocity surface mode in the photonic crystal . the number of the resonant modes with high quality factors is mainly determined by the number of surface modes with zero - group velocity .
optical microcavities have attracted much attention due to their potential applications in many various fields , such as ultrasmall optical filters , high efficiency light emission diodes , low threshold lasers , nonlinear optics and quantum information precessing @xcite .
all applications require microcavities to confine light strongly and densely , that is , microcavities should have both a high - quality factor @xmath0 and small modal volume @xmath1 .
total internal reflection and/or bragg reflection are generally used for light confinement . for the smaller cavity
, @xmath0 is greatly reduced .
one of the best approaches to resolve the problem is the extension of the bragg reflection in a two - dimensional ( 2d ) or a three - dimensional ( 3d ) photonic crystal ( phc ) @xcite .
it has been also shown that extremely small - mode - volume high-@xmath0 microcavities can be realized by 2d phc slabs @xcite , owing to strong optical confinement for both the in - plane and vertical direction .
surface waves are propagating electromagnetic waves , which are bound to the interface between materials and free space .
for dielectric interfaces , they usually do not exist on dielectric materials .
however , due to the existence of photonic band gaps , it has been shown that dielectric phcs may support surface waves for some cases , e.g. , a truncated or deformed structure at the interface @xcite . in the present letter , we study phc microcavities utilizing surface modes with zero - group velocity .
though the concepts apply for all phcs , for simplicity , only surface resonant modes in 2d phc slabs are considered in this letter .
these cavities are open cavities in the sense that one of the in - plane boundaries is exposed to exterior .
although the similar structure has been realized by yang et .
al .
experimentally @xcite , the dependence of the quality factor and resonant frequency to the cavity length has not been analyzed yet .
let us first consider a 2d square phc slab with dielectric rods in air .
the permittivity of the rods is @xmath2 , the height of the rods is @xmath3 , and the radius of the rods is @xmath4 , where @xmath5 is the lattice constant .
surface defects at the interface are introduced by reducing the radius of rods ( @xmath6 ) , the top view of which is illustrated in the inset of fig .
[ fig1 ] ( a ) .
the dispersion relation for the reduced - radii rod - slab transverse magnetic ( tm ) surface modes is calculated using the 3d finite - difference time - domain ( fdtd ) method @xcite and is shown in fig .
[ fig1 ] ( a ) , where the shadow regions are the projected band structure for tm modes .
it can be seen from fig .
[ fig1 ] ( a ) that such a surface structure only supports one surface mode .
group velocity governed by @xmath7 for the surface mode is zero for the wave vector @xmath8 ( mode a ) .
suppose an optical microcavity that is composed by the 2d square phc slab , as shown in fig .
[ fig2 ] ( a ) .
the grey rods , acting as reflecting mirrors in the @xmath9 direction , are the same as those interior rods .
the length of the cavity is denoted by @xmath10 .
surface modes can go through the central surface but are terminated by the reflecting mirror .
only the mode for obeying the fabry - perot condition ( the round - trip accumulated phase @xmath11 is a multiple of @xmath12 , where @xmath13 is the phase shift associated with reflection from the boundary ) will stay in the cavity for a long time , i.e. , become a surface resonant mode .
for such a microcavity , with no absorption by the material , the quality factor @xmath0 is mainly determined by the reflection loss at the interface between the surface defect rods and the mirror rods .
quality factors and resonant frequencies of surface resonant modes are analyzed by the 3d fdtd method @xcite with a boundary treatment of perfectly matched layers @xcite .
we excite the surface mode with a gaussian pulse and then monitor the radiative decay of the field . the frequencies @xmath14 and quality factors @xmath0 of the resonant modes are calculated using a combination of fdtd techniques and pad approximation with baker s algorithm @xcite . in fig .
[ fig3 ] ( a ) , we show the electric field cross sections of a high-@xmath0 surface resonant mode for the square phc slab cavity with the length of @xmath15 . figure [ fig3 ] ( a ) depicts horizontal and vertical cross sections of the electric - field @xmath16 component at the center plane of third axis ( @xmath17 ) and at the plane center of the surface defects , where the dotted line represents the position of the surface defect rods with @xmath6 .
the mode has an angular frequency @xmath18 and quality factor @xmath19 , in which @xmath20 and @xmath21 ( @xmath22 ) .
it can be clearly seen from fig .
[ fig3 ] ( a ) that the resonant mode is really governed by wave vector @xmath8 , which is consistent with what we analyzed above . from fig .
[ fig3 ] ( a ) , one can see that the electric field is almost concentrated in the centers of the defect rods .
we believe that the reflection phase shift @xmath13 is close to @xmath23 , which is roughly independent of the cavity length @xmath10 .
therefore , the cavity modes can be related to surface modes which satisfy @xmath24 , ( @xmath25 and @xmath26 is an integer ) .
however , our results also show that the quality factor of the resonant mode corresponding to @xmath8 , i.e. , zero - group - velocity surface mode , is order of magnitude larger than those of cavity modes with other @xmath27 vectors .
@xmath0 values for such a cavity mode with different cavity length are shown in fig .
[ fig4 ] ( a ) .
it can be seen from fig .
[ fig4 ] ( a ) that @xmath0 gradually increases as the cavity length becomes larger .
our results also show , with increasing of the cavity length , that the angular frequency of the surface resonant mode converges to @xmath28 , corresponding to that with the zero - group - velocity surface mode in the phc slab .
let us next consider a triangular phc slab with air holes extending through a high - index ( @xmath2 ) finite - height dielectric slab , the top view of which is shown in the inset of fig .
[ fig1 ] ( b ) .
the holes have a radius of @xmath29 , while the high - index slab is of thickness @xmath30 .
the distance between right boundary and the centers of the first right holes is @xmath31 .
the dispersion relation for such a triangular phc slab transverse electric ( te ) surface modes is shown in fig .
[ fig1 ] ( b ) , where the shadow regions are the projected band structure for te modes .
one can see from fig .
[ fig1 ] ( b ) that there exist three surface mode curves in the triangular phc slab . for simplicity
, we only consider the angular frequency between @xmath32 and @xmath33 , where there is one surface mode in a large region .
two zero - group - velocity surface modes ( mode a with @xmath34 and mode b with @xmath35 ) can be found in our considered frequency region .
now , imagine an optical cavity composed by such a 2d triangular phc slab , as shown in fig .
[ fig2 ] ( b ) .
the reflecting mirrors in the @xmath9 direction are introduced by enlarging the radius of surface holes to @xmath36 , together with the dielectric background . using the method mentioned above , we do find two high - quality - factor surface resonant modes with the cavity length of @xmath37 .
one of the modes ( corresponding to mode a ) has a resonant angular frequency @xmath38 , @xmath39 and another ( corresponding to mode b ) is @xmath40 , @xmath41 .
figures [ fig3 ] ( b ) and [ fig3 ] ( c ) show the magnetic - field cross sections for the two resonant modes with the frequency of @xmath42 and @xmath43 , respectively .
the upper and bottom figure represent the horizontal and vertical cross sections of the magnetic field at the center plane of third axis ( @xmath17 ) and at the vertical @xmath44 plane ( the distance between surface boundary and the plane is @xmath45 ) , respectively .
our results show that , when the cavity length increases , the frequencies of the resonant modes converge to those two surface modes with zero - group velocity .
however , @xmath0 values fluctuate as the cavity length @xmath10 increases , as shown in fig .
[ fig4](b ) , where the diamond line and the asterisk line represent the results for modes a and b , respectively . as described in ref .
@xcite , @xmath0 should be enhanced when @xmath11 is a multiple of @xmath12 . due to the complicated reflection interfaces , as seen in fig .
[ fig2](b ) , the reflection phase shift @xmath13 is not close to @xmath23 for the both modes , different from the case of the square lattice .
furthermore , for the cavity length with numbers of lattice constant , @xmath46 is not anymore close to a multiple of @xmath12 for the mode b with @xmath47 .
peaks of @xmath0 appear when @xmath48 is close to a multiple of @xmath12 .
the positions of peaks for the mode a are different with those for the mode b , mainly arisen from the different wave vectors for the two resonant modes .
@xmath49 reaches at @xmath50 when @xmath51 and @xmath52 is @xmath53 when @xmath54 .
if the cavity length is further increased , @xmath0 values will again fluctuate but in average increasing . in this letter , we study optical microcavities based on surface modes in phcs .
high quality factors can be obtained for microcavities based on zero - group - velocity surface modes .
we also study the influence of the microcavity length to the quality factor and resonant frequency .
it has been shown that , with increasing of the cavity length , the quality factor is gradually enhanced and resonant angular frequency converges to @xmath28 , corresponding to that for the zero - group - velocity surface mode in phcs . the number of the resonant modes with high quality factors is mainly determined by the number of surface modes with zero - group velocity .
it is well known that surface plasmons exist on a metal - dielectric interface .
we believe similar ideas can be applied to a metallic structure to obtain surface resonant modes .
this work was supported by the swedish foundation for strategic research ( ssf ) on ingvar program , the ssf strategic research center in photonics , and the swedish research council ( vr ) under project no .
2003 - 5501 . |
9 siddiqi i _ et al .
_ 2004 _ phys . rev .
lett . _ * 84 * 207002 | for the model of a linearly driven quantum anharmonic oscillator , the role of damping is investigated .
we compare the position of the stable points in phase space obtained from a classical analysis to the result of a quantum mechanical analysis .
the solution of the full master equation shows that the stable points behave qualitatively similar to the classical solution but with small modifications .
both the quantum effects and additional effects of temperature can be described by renormalizing the damping . in recent years
, driven anharmonic oscillators have generated a large amount of interest because of their use as qubit readout devices ( i.e. , the josephson bifurcation amplifier ) @xcite .
all these devices operate close to or in the quantum regime .
there are two possible stable states in phase space , which can be distinguished by their respective amplitude and phase .
many studies have been performed on the transition between the stable states @xcite and further effects like multiphoton resonances @xcite and dynamical tunneling @xcite make driven anharmonic oscillators an ideal model to study thermo - dynamics and quantum effects in a non - equilibrium system .
most bifurcation amplifiers are operated in the limit of strong damping to speed up the classical read - out process of the amplifier . in this paper
, we define an equivalent to the classical stable state derived from a quantum master equation and discuss its position in phase space as a function of damping . in phase space .
the extrema correspond to the two stable states , i.e. , the high amplitude state ( * _ m _ * ) and the low amplitude state ( * _ m _ * ) .
the saddle point corresponds to an unstable state ( * _ s _ * ) .
the three kinds of classical orbits are also depicted by black closed curves .
( b ) a cut through the quasienergy potential @xmath0 for @xmath1 .
the quantum levels nearest to maximum ( * _ m _ * ) , minimum ( * _ m _ * ) and saddle point ( * _ s _ * ) are shown in blue , red and green short lines respectively . shown in long black line
is the quantum level on the outer torus , which is close to and can be generate with level * _ m_*. ( c ) squared amplitude of the wave functions corresponding to quantum levels shown in fig.1(b ) .
we use the same colors and labels to distinguish them .
the two subfigures correspond to different parameters , i.e. , wave function on the outer torus is degenerate with the low amplitude state ( top subfigure ) and non - degenerate ( bottom subfigure ) .
we used @xmath2 and @xmath3 for this plot ( degeneracy can be created by only minor changes of @xmath4 ) . ]
the driven anharmonic oscillator is described by @xmath5 where @xmath6 stands for the external driving force . in the rotating frame and after a rotating wave approximation ( rwa )
the hamiltonian reads @xmath7 here , we have introduced the detuning between the oscillator s natural frequency and the driving frequency @xmath8 , @xmath9 .
we then introduce the position operator @xmath10 and momentum operator @xmath11 in the rotating frame @xmath12 where @xmath13 is the effective planck constant .
finally , the scaled quasienergy hamiltonian is defined as @xmath14 where @xmath15 here , we have neglected some constant terms .
now all the properties of this system are dependent on one single quantity @xmath16 . the quantum properties are contained in the commutation relation @xmath17=i\lambda$ ] .
the transition to the classical regime is equivalent to the limit @xmath18 .
the dissipative environment is taken into account using the lindblad master equation @xmath19 + \eta \ { ( 1+\bar{n}){\cal d}[a]{\rho } + \bar{n}{\cal d}[a^{\dagger}]{\rho } \ } , \end{aligned}\ ] ] where the lindblad operator is defined through @xmath20{\rho}\equiv a{\rho}a^{\dagger}-\frac{1}{2}\{a^{\dagger}a,{\rho } \ }
$ ] , @xmath21 is the bose - einstein distribution , and @xmath22 is the dimensionless damping strength @xcite .
let s first discuss the classical dynamics of the driven duffing oscillator in the weak - damping limit . for @xmath23 ,
the hamilton function exhibits three extrema in phase space , as indicated in fig.1(a ) , i.e. , a maximum @xmath24 , a minimum @xmath25 and a saddle point @xmath26 . without friction ,
there are three kinds of possible periodic orbits .
they correspond to different isoenergetic sections of the hamilton function .
the orbits with an ear - shape around the maximum @xmath24 form one group .
those circling the minimum @xmath25 form another group .
the third group consists of the orbits on the outer torus .
one can calculate the explicit expressions of these orbits based on the classical equation of motion : @xmath27 in the vicinities of the extrema @xmath24 and @xmath25 the system is equivalent to an underdamped harmonic oscillator .
therefore , if damping is included the orbits nearby will shrink towards the extrema . as a result , they are stable points , corresponding to the oscillations with _ high _ and _ low _ amplitudes respectively .
they are separated by a phase space barrier associated with the unstable saddle point @xmath26 . in the quantum limit
, the classical orbits will be quantized into a series of discrete energy levels .
they can be calculated by diagonalizing the scaled hamiltonian ( [ hq ] ) , which results in a set of eigenvalues and eigenstates , i.e. , @xmath28 . in fig.1(b ) we show a cut through the quasienergy potential @xmath0 for @xmath1 .
the classically stable states are the extrema of the potential .
the quantum levels corresponding to the three extrema are shown by colored short lines and labelled by * _ m_*,*_m _ * and * _ s _ * respectively .
the black long line indicates one quantum level on the outer torus , which is close to and can be resonant with the level * _ m_*.
fig.1(c ) shows the quantum analogues ( i.e. , squared amplitude of the wave functions , @xmath29 ) of the classical orbits corresponding to the quantum levels shown in fig.1(b ) .
we use the same colors and labels in each figures . in principle , the small amplitude state is coupled via tunnelling to states on the outer torus .
if a state on the outer and inner torus are degenerate , the eigenstates are mixed states and this results in a large change of the degenerate wave functions ( see the red and black curves in fig.1(c ) ) .
however , one should note that in most of the relevant parameter regime the dynamics of the system is dominated by intra - well transitions @xcite . and
@xmath30 respectively .
( b ) the difference between the effective damping @xmath31 and real quantum simulation damping @xmath22 at zero temperature ( red solid circles ) and @xmath32 ( black solid squares ) respectively .
( c ) the relationship between effective damping and temperature at fixed damping @xmath33 .
other parameters are @xmath2 and @xmath34 . ] as damping increases , the positions of the stable points in phase space will shift .
we can find the classical solution for the stable points from the classical equation of motion including damping:@xmath35 . for finite damping
the system is classically bistable for @xmath36 , with @xmath37 @xcite . in fig.2(a ) , we plot the possible positions of the low amplitude state ( red solid line ) , high amplitude state ( blue solid line ) and the unstable state ( green dashed line ) in phase space . in order to describe the damping effects in the quantum regime
, we turn to the master equation ( [ me ] ) .
we can also define a stable state using our master equation . at exactly zero temperature , it is possible to get an analytical expression for the stationary distribution in the so called p - representation @xcite , which is one of several quasi - probability distributions @xcite . with the help of the p - representation , the exact analytical solution for the stationary density matrix for @xmath38 can be obtained in the photon number representation ( see eq.(15 ) in ref.@xcite ) .
we can diagonalize it and use its eigenstates as our basis .
a diagonal representation will not only give us a clear physical picture of the dynamics of the system but also simplifies any numerical simulation . to understand how our solution compares to the solution of the classical equation , we use the following procedure : ( 1 ) diagonalize the exact solution ; ( 2 ) identify the eigenstates which correspond to the two stable states and calculate the expectation value of position and momentum ( or amplitude and phase ) .
an example for the results of such a calculation are the black empty circles in fig.2(a ) ; ( 3 ) compare the expectation values of the wave function with those expected from the classical equation of motion . since the states are located on the same curve as the classical state we can assign each state a value @xmath31 that corresponds to the damping of the equivalent classical stable point .
due to the finite effective planck constant @xmath39 used in this work , our results display some quantum effects .
one is the disappearance of the small - amplitude state already for weaker damping than classically predicted .
the reason is that near the bifurcation point , the low amplitude well is so shallow that the zero point energy exceeds the potential barrier . as a result
, there is no quantum level in the vicinity of the minimum * m*. the situation is similar for the large - amplitude state if @xmath4 is small .
another quantum signature is the renormalization of damping , @xmath40 .
we show a comparison of @xmath22 , as chosen for the master equation and the effective damping @xmath31 of the equivalent classical stable point in fig.2(b ) .
as @xmath39 becomes smaller we find @xmath41 @xcite .
the effect of temperature can be included in an exact numerical solution of the master equation .
then we again follow the steps outlined above and find that the eigenstates that diagonalize the matrix have average momenta and coordinates that fall on the curve of the classically stable states ( see yellow and red solid circles in fig.2(a ) ) .
we find that the temperature also introduces a weak renormalization of the damping @xmath22 .
we plot the difference , @xmath42 , as function of @xmath22 in fig.2(b ) at a finite temperature ( black solid squares in fig.2(b ) ) .
we also show the renormalized damping as a function of @xmath43 for fixed @xmath22 in fig.2(c ) .
as one can see , there is a small quantum correction for @xmath44 , and temperature has , as expected , no effect as long as @xmath45 . as temperature increases we get an additional small shift along the curve of the classically stable state .
generally we find that the states that diagonalize the density matrix at @xmath44 still remain an efficient diagonal basis up to @xmath46 . |
let @xmath7 be vectors in @xmath8 , and denote @xmath9 by @xmath10 .
a _ simultaneous integer relation _ ( sir ) for @xmath11 is a vector @xmath2 such that @xmath12 , i.e. @xmath3 for @xmath13 . for short ,
we also call @xmath14 an sir for @xmath10 .
when @xmath15 , we say that @xmath14 is an integer relation for @xmath16 .
the problem of detecting integer relations for a rational or real vector is quite old .
historical surveys can be found in @xcite . among these integer relation detecting algorithms , the hjls algorithm @xcite and the pslq algorithm
@xcite have been used frequently . in the present paper , using the technique to construct the hyperplane matrix in hjls and a generalized method of the matrix reduction from pslq we propose an algorithm sird , which can be used to detect an sir for @xmath17 real vectors .
the cost of our algorithm is at most @xmath5 exact arithmetic operations for detecting an sir for @xmath10 , where @xmath6 represents the least euclidean norm of sirs for @xmath10 .
furthermore , our detecting algorithm sird either always finds an sir for @xmath10 if one exists or proves that there are no sirs for @xmath10 of norm less than a given size .
experimental results show that sird is practical . in application
, we successfully apply sird to find the minimal polynomial of an algebraic number @xmath18 with degree and height at most @xmath19 and @xmath20 respectively from its an approximation @xmath21 satisfying @xmath22 , and propose the corresponding algorithm mpf , where the minimal polynomial of an algebraic number @xmath23 is the unique primitive polynomial @xmath24 $ ] of least degree such that @xmath25 .
in fact , for @xmath26 from @xmath27 to @xmath19 we run sird with @xmath28 , @xmath29 as its input and then an exact sir for @xmath30 has been detected .
we provide a sufficient controlling on @xmath31 and prove that such an @xmath31 is sufficient to enable an exact sir for @xmath32 and @xmath33 to be also an sir for @xmath34 and @xmath35 , where @xmath31 depends only on @xmath19 and @xmath20 , as in ( [ eq : error - controlling ] ) .
it implies the correctness of mpf and is better than already existing results in @xcite . in @xcite , j. hastad , b.
just , j. c. lagarias , and c. p. schnorr not only presented the hjls algorithm and the first rigorous proof of a ` polynomial time ' bound for a relation finding algorithm but also proposed a simultaneous relations algorithm ( see @xcite ) , whereas hjls is numerically unstable .
the unstable examples can be found in @xcite . in their draft @xcite , c. rssner and c. p. schnorr studied the case of @xmath36 by using a modified hjls algorithm .
but for the moment , @xcite is still in a preliminary state with some open problems .
the pslq algorithm , together with related lattice reduction schemes such as lll @xcite , was named one of ten `` algorithms of the twentieth century '' by the publication _ computing in science and engineering _ ( see @xcite ) , and is now extensively used in experimental mathematics , with applications such as identification of multiple zeta constants , a new formula for @xmath37 , finding algebraic relations and so on ( see @xcite ) .
moreover , pslq is numerically stable and can be easily generalized to complex number field and hamiltonian quaternion number field ( see @xcite ) , but it is not suitable to detect an sir for several real vectors .
+ the sird algorithm in this paper is to detect an sir for @xmath17 real vectors and can be applied to detect an integer relation in @xmath38 for a complex vector or a hamilton quaternion number vector .
a significant body of experimental data shows that sird is practical and better than the hjls simultaneous relations algorithm .
in fact , the mpf algorithm in this paper is a positive answer to the following interesting question : suppose we are given an approximation to an algebraic number @xmath23 , and two bounds on the degree and the size of the coefficients of its minimal polynomial respectively .
is it possible to infer the minimal polynomial ?
the question was raised , independently , by manuel blum in theoretical cryptography ( see @xcite ) and the last author of this paper in automated reasoning ( see @xcite ) .
the first complete answer to this question , kll algorithm , was presented by r. kannan , a.k .
lenstra and l. lovsz in @xcite by using the celebrated lattice reduction algorithm lll @xcite . in the computer algebra system _
maple _ , the built - in function ` polynomialtools :- minimalpolynomial ( ) ` is a function to find a polynomial of degree @xmath19 ( or less ) with small integer coefficients which has the given approximation @xmath39 of an algebraic number as one of its roots and is based on kll algorithm .
the correctness of the polynomial returned by the built - in function depends on the accuracy of the approximation ( see _ maple _ s help ) . from another aspect
, the minimal polynomial of an algebraic number @xmath23 with exact degree @xmath19 can be found by detecting an integer relation for the vector @xmath40 .
besides hjls , b. just also presented an algorithm to detect integer relations for a given vector consists of algebraic numbers in @xcite .
we can apply just s algorithm or hjls to the vector @xmath41 for finding the minimal polynomial of @xmath23
. however , both just s algorithm and hjls are not numerically stable , as mentioned previously .
all these algorithms are based on lll .
two authors of this paper presented a method to reconstruct a rational number from its an approximation by using continued fraction in @xcite .
it may be viewed as an answer to a special case of the question . based on pslq
, one can find algebraic relations , such as @xcite , whereas these articles did not involve the minimal polynomial finding .
the authors of this paper also presented an algorithm in @xcite for finding the minimal polynomial of a real algebraic number from its an approximation .
however , these pslq based algorithms can not deal with complex algebraic numbers since pslq only outputs a relation in gaussian integer ring for a complex vector .
fortunately , our simultaneous integer relation detection algorithm sird in present paper can be used to overcome these pitfalls . applying sird to one or two real vectors , we present another affirmative answer , the mpf algorithm , to the question above .
we show that mpf is a more efficient minimal polynomial finding algorithm comparing with the algorithms in @xcite and provide a sufficient condition on the error controlling , from which we can claim that the polynomial returned by mpf is the exact minimal polynomial of the algebraic number that we only know an approximate value and two bounds on its degree and height .
although a similar even better complexity can be obtained by kll , mpf has its own meaning since it is a new method without using lll reduction .
* road - map . * in section [ sec : the simultaneous integer relation detecting algorithm ] and [ sec : the sird algorithm ] we first give some preliminaries , and then present the sird algorithm and analyze it .
we report on some experimental results about the performance of sird in section [ sec : performance results ] , apply sird to find the minimal polynomial of an algebraic number from its an approximation and propose the mpf algorithm in section [ sec : finding - minimal - polynomial ] , in which we also analyze mpf and present the result of error controlling .
we conclude this paper with section [ sec : conclude ] .
* notations . * throughout this paper , @xmath42 , @xmath43 , and @xmath44 are the sets of integers , real numbers , and complex numbers respectively .
the real and imaginary parts of @xmath45 will be denoted re@xmath46 and im@xmath46 respectively . for @xmath47 , @xmath48 .
all vectors in this paper are column vectors , and will be denoted in bold . if @xmath49 , then @xmath50 represents its euclidean norm , i.e. @xmath51 , where @xmath52 is the inner product of two vectors .
we denote @xmath53 identity matrix by @xmath54 . given a matrix @xmath55 , we denote its transpose by @xmath56 , its trace by tr@xmath57 , its determinant by @xmath58 , and its frobenius norm by @xmath59 , i.e. @xmath60 .
we say that a matrix @xmath61 is lower trapezoidal if @xmath62 for @xmath63 .
@xmath64 is the group of @xmath53 unimodular matrix with entries in @xmath42 .
the height of a vector is defined by the maximum of all the absolute values of its entries .
for a polynomial @xmath65 , we denote by @xmath66 its degree with respect to @xmath67 , @xmath68 its one norm , @xmath69 its euclidean length , and height@xmath70 its height .
in what follows we always suppose that @xmath1 are linearly independent vectors in @xmath8 , where @xmath71 . obviously , we have @xmath72 .
we denote by @xmath10 the matrix @xmath73 , and suppose that @xmath74 satisfies @xmath75 unless otherwise specified . for @xmath76 not satisfying ( [ eq
: suppose - on - x ] ) , exchanging some rows of @xmath10 produces @xmath77 , where @xmath78 is an appropriate matrix in @xmath64 .
and then we detect an sir for @xmath79 .
if @xmath14 is an sir for @xmath79 , then @xmath80 is an sir for @xmath10 .
[ def : hyperplane - matrix ] let @xmath81 .
a hyperplane matrix with respect to @xmath10 is any matrix @xmath82 such that @xmath83 and the columns of @xmath84 @xmath85 @xmath86 .
now we introduce a method to construct a hyperplane matrix for @xmath10 .
let @xmath87 form a standard basis of @xmath88 , i.e. the @xmath26-th entry of @xmath89 is @xmath27 and others are @xmath90 . by performing the process of standard gram - schmidt orthogonalization
to @xmath91 in turn we have @xmath92@xmath93where @xmath94 [ lem : hyperplane - matrix - construction ] let @xmath95 , @xmath96 , @xmath97 and @xmath98 be as above .
. there exist @xmath17 elements in @xmath99 denoted by @xmath100 such that @xmath101 .
2 . @xmath102 .
part @xmath27 easily follows from the process of standard gram - schmidt orthogonalization .
we next prove @xmath103 when ( [ eq : suppose - on - x ] ) holds .
set @xmath104 taking each side as a column vector and observing the last @xmath17 components of two sides , we have @xmath105 . and
since @xmath106 are linearly independent , we have @xmath107 .
thus the @xmath19 vectors @xmath108 are linearly independent .
this implies that @xmath102 .
[ def : h_x ] for @xmath109 satisfying ( [ eq : suppose - on - x ] ) , define @xmath110 to be the @xmath111 matrix @xmath112 .
[ lem : property - of - hx ] let @xmath109 and @xmath110 be as above . then 1 .
2 . @xmath114 .
@xmath115 is an orthogonal matrix .
\4 . @xmath116**0 * * , i.e. @xmath110 is a hyperplane matrix of @xmath10 . \5 .
@xmath110 is a lower trapezoidal matrix and every diagonal element of @xmath110 is nonzero .
since every two columns of @xmath110 are orthogonal , part @xmath27 follows . and part @xmath117 follows from part @xmath27
. let @xmath118 . obviously , @xmath119 is an orthogonal matrix .
from part 3 and standard gram - schmidt orthogonalization we have @xmath120 and @xmath121 respectively , where @xmath122 is an appropriate @xmath123 invertible matrix .
thus @xmath124 and hence that part 4 follows .
we now prove part 5 .
denote the @xmath125-th element of @xmath126 by @xmath127 .
the diagonal elements of @xmath110 are @xmath128 for @xmath129 . before normalizing @xmath126 we have @xmath130 , and at the same time , @xmath131 thus all the diagonal elements of @xmath110 are nonzero .
now we only need to show that @xmath110 is lower trapezoidal . from standard gram - schmidt orthogonalization , we can check that @xmath132 holds for @xmath133 .
this completes the proof .
so far , we have had a method to produce a hyperplane matrix @xmath134 for @xmath74 .
the basic idea is from hjls ( see @xcite ) .
the same strategy was also used in pslq , however , in which partial sum was adopted instead of gram - schmidt orthogonalization . [
lem : property - of - px ] for @xmath135 define @xmath136 . then 1 .
2 . @xmath138 .
3 . @xmath139 .
4 . @xmath140 .
@xmath141 for any @xmath142 .
particularly , @xmath143 for any sir @xmath14 for @xmath10 .
the proof of the first part is easy .
let @xmath144 .
from lemma [ lem : property - of - hx ] we have @xmath145 .
thus part @xmath117 follows .
part @xmath146 and part @xmath147 follow from @xmath148 and @xmath149 tr@xmath150 tr@xmath151 tr@xmath152 respectively
. since @xmath153 , we have @xmath154 for @xmath155 .
and the process of standard gram - schmidt orthogonalization implies @xmath156 .
thus we have @xmath157 from part 2 . from lemma [ lem :
property - of - hx ] and lemma [ lem : property - of - px ] we can easily generalize the theorem @xmath27 in @xcite to the case of @xmath74 .
[ thm : lower - bound - of - simultaneous - relations ] let @xmath74 and @xmath110 be as above .
suppose that for any matrix @xmath158 there exists an orthogonal matrix @xmath159 such that @xmath160 is lower trapezoidal and all of the diagonal elements of @xmath161 satisfy @xmath162 .
then for any sir @xmath14 of @xmath10 we have @xmath163 as this theorem easily follows from the proof of theorem 1 of @xcite with little modifications , the detail has been omitted here .
the lower bound given in ( [ eq : lower - bound - of - simultaneous - relation ] ) when @xmath164 is consistent with a similar lower bound in @xcite .
moreover , if a method to reduce the norm of @xmath134 by multiplication by some unimodular @xmath165 on the left has been developed , then it will produce an increasing lower bound on @xmath6 , where @xmath6 is the least euclidean norm of sirs for @xmath10 .
in fact this theorem suggests a strategy to detect an sir for @xmath10 .
we now study how to reduce the hyperplane matrix @xmath134 .
first we recall ( modified ) hermite reduction in @xcite . a lower trapezoidal matrix @xmath166 with @xmath162 . a reducing matrix @xmath167 of @xmath20 . the @xmath53 matrix @xmath167 .
[ algo : modified - herimite - reduction ] if algorithm [ algo : modified - herimite - reduction ] output @xmath167 for an @xmath168 matrix @xmath20 , we say that @xmath169 is the modified hermite reduction of @xmath20 and that @xmath167 is the reducing matrix of @xmath20 .
this reduction develops the left multiplying modified hermite reducing matrix @xmath167 .
hermite reduction is also presented in @xcite , and is equivalent to modified hermite reduction for a lower triangular matrix @xmath20 with @xmath162 ( see ( * ? ? ? * lemma 3 ) ) .
both the two equivalent reductions have the following properties : 1 .
the reducing matrix @xmath170 .
2 . for all @xmath171 , the ( modified )
hermite reduced matrix @xmath172 satisfies @xmath173 satisfy the two properties above , we need the following generalized hermite reduction . a lower trapezoidal matrix @xmath174 with @xmath162 . a reducing matrix @xmath167 of @xmath20 .
[ algostep : ghr - hermite - reduction - start ] [ algostep : generalized-2-hermite - reduction - start ] [ algostep : generalized-2-hermite - reduction - end ] the @xmath53 matrix @xmath167 .
[ algo : generalized - herimite - reduction ] if algorithm [ algo : generalized - herimite - reduction ] output @xmath167 for an @xmath111 matrix @xmath20 , we call @xmath169 the generalized hermite reduction of @xmath20 and @xmath167 the reducing matrix of @xmath20 .
obviously , generalized hermite reduction is equivalent to modified hermite reduction when @xmath164 . in addition
, we can easily check that generalized hermite reduction remains the two properties mentioned above .
[ rem : generalized - hermite - reduction ] there are two main differences between ( modified ) hermite reduction and generalized hermite reduction .
firstly , the last @xmath175 rows of @xmath20 will also be reduced by the first @xmath176 rows of @xmath20 in generalized hermite reduction , while ( modified ) hermite reduction can not do so .
secondly , generalized hermite reduction exchanges the @xmath177-th row and the @xmath178-th row of @xmath167 if @xmath179 , @xmath180 and @xmath181 ( from step [ algostep : generalized-2-hermite - reduction - start ] to step [ algostep : generalized-2-hermite - reduction - end ] ) .
this implies that if @xmath182 after generalized hermite reduction then @xmath183 .
this property plays an important role in the proof of lemma [ lem : hn , n-2=0 ] .
using the hyperplane matrix constructing method and generalize hermite reduction in the previous section we can get a simultaneous integer relation detecting algorithm sird . @xmath184 satisfying ( [ eq : suppose - on - x ] ) either output an sir for @xmath10 or give a lower bound on @xmath6 .
compute the hyperplane matrix @xmath110 , set @xmath185 , @xmath186 .
call algorithm [ algo : generalized - herimite - reduction ] to reduce @xmath110 producing the reducing matrix @xmath187 . set @xmath188 .
[ algostep : exchange ] _ exchange .
_ let @xmath189 .
choose an integer @xmath39 such that @xmath190 for @xmath191 , where @xmath192 .
define the permutation matrix @xmath193 to be the identity matrix with the @xmath39 and @xmath194 rows exchanged .
update @xmath195 .
[ algostep : corner ] _ corner .
_ let @xmath196 let @xmath197 . if @xmath198 , then let the submatrix of @xmath122 consisting of the @xmath39-th and @xmath199-th rows of columns @xmath39 and @xmath200 be @xmath201 . update @xmath202 .
[ algostep : reduction ] _ reduction .
_ call algorithm [ algo : generalized - herimite - reduction ] to reduce @xmath110 producing @xmath167 .
update @xmath203 .
compute @xmath204 .
then there exists no sir whose euclidean norm is less than @xmath205 .
[ algostep : bound - of - possible - relations ] [ algo : simultaneous - relation ] let @xmath206 be the result after @xmath125 iterations of sird .
why do we set the parameter @xmath207 at step [ algostep : exchange ] ?
suppose the @xmath39 chosen in step [ algostep : exchange ] is not @xmath208 .
in this case we let @xmath209 be as in ( [ eq : alph - beta - lambda - delta ] ) .
then @xmath210 is the submatrix of @xmath211 consisting of the @xmath39 and @xmath200 rows of columns @xmath39 and @xmath200 , where @xmath198 .
after step [ algostep : exchange ] has been performed @xmath212 may not be zero , which makes that @xmath20 is not lower trapezoidal .
after step [ algostep : corner ] the result is @xmath213 since @xmath39 is chosen such that @xmath214 is as large as possible , and @xmath198 we have @xmath215 , hence @xmath216 . from the property of generalized hermite reduction
we have that @xmath217 , which then gives @xmath218 thus @xmath219 is reduced as long as @xmath220 , i.e. @xmath207 . as was pointed out by borwein ( see bor2002 ) , although this increases @xmath221 , this is not a significant problem . at each step
we force the larger diagonal elements of @xmath20 toward @xmath222 , where their size can be reduced by at least a factor of @xmath117 when @xmath223 . as a matter of fact
, the parameter @xmath224 can be freely chosen in the open interval @xmath225 .
[ lem : hn , n-2=0 ] if @xmath226 for some @xmath227 and no smaller @xmath125 , then @xmath228 and an sir for @xmath10 must appear as a column of the matrix @xmath229 . by the hypothesis on @xmath125
we know that all diagonal elements of @xmath230 are not zero . now
, suppose the @xmath39 chosen in step [ algostep : exchange ] is not @xmath176 .
since generalized hermite reduction does not introduce any new zeros on the diagonal , and from the analysis of step [ algostep : exchange ] and step [ algostep : corner ] above , we have that no diagonal element of @xmath206 is zero .
this contradicts the hypothesis on @xmath125 and our assumption that @xmath198 was false .
thus we have @xmath231 after the @xmath232-th iteration has been completed .
next we show that there must be an sir for @xmath10 appeared as a column of the matrix @xmath229 .
we have @xmath233 from lemma [ lem : property - of - hx ] and hence that @xmath234 , where @xmath122 is an appropriate orthogonal @xmath235 matrix .
let @xmath236 , where @xmath237 .
then @xmath238 we know @xmath239 and @xmath240 from @xmath241 . from remark [ rem : generalized - hermite - reduction ] and
@xmath242 we have @xmath243 which implies the last equality . since @xmath240
, it follows that @xmath244 .
thus the @xmath245-th column of @xmath229 is an sir for @xmath10 .
from theorem [ thm : lower - bound - of - simultaneous - relations ] and lemma [ lem : hn , n-2=0 ] , the correctness of sird has been proved .
moreover , we have [ thm : upper - bound - for - sirs ] let @xmath6 be the least euclidean norm of any sir for @xmath10 . let @xmath14 be an sir detected by sird .
then @xmath246 for all @xmath247 .
assume @xmath248 with @xmath249 and @xmath250 at the @xmath125-th iteration of sird .
then from theorem [ thm : lower - bound - of - simultaneous - relations ] and the exchange rule of sird we have @xmath251 at this time , @xmath252 holds from the same strategy in the proof of lemma 10 in @xcite .
[ def : pi - function ] for the @xmath125-th iteration in sird , define @xmath253 the routine of analyzing the number of iterations in @xcite can be carried over here with redefining the @xmath254 function as above .
so we state the following lemma directly without proof .
[ lem : pi - function ] for @xmath255 we have 1 .
@xmath256}\geq\pi(k)\geq 1 $ ] , where @xmath6 is the least norm of sirs for @xmath10 .
2 . @xmath257 . from this lemma
, it follows that the @xmath254 function is increasing with respect to @xmath125 and has an upper bound for a fixed @xmath258 .
thus we have [ thm : the - number - of - iterations ] if @xmath109 has sirs , then the number of iterations such that sird finds an sir for @xmath10 will be no more than @xmath259 \frac{\log(\gamma^{n - t}\lambda(x))}{\frac{1}{2 } \log\left(\frac{4\gamma^2}{\gamma^2 + 4}\right)}.\ ] ] from definition [ def : pi - function ] we can infer @xmath260 . and by lemma [ lem : pi - function ] we know that @xmath261 solving @xmath125 from this inequality gives the conclusion , as was to be shown .
[ thm : time - complexity - of - simultaneous - relation ] if @xmath109 has sirs , then there exists a @xmath224 such that sird will find an sir for @xmath10 in polynomial time @xmath262
. let @xmath263 .
then sird will construct an sir for @xmath10 in no more than @xmath264 iterations .
sird takes @xmath265 exact arithmetic operations per iteration , and hence that @xmath266 exact arithmetic operations is enough to produce an sir for @xmath10 .
since @xmath267 , the proof is complete . [
rem : simultaneous - relation - algorithm ] from this corollary , we can claim that our detecting algorithm always return an sir for @xmath10 if one exists .
additionally , sird will produce lower bound on the euclidean norm of any possible sirs for @xmath10 ( theorem [ thm : lower - bound - of - simultaneous - relations ] ) .
thus sird can be used to prove that there are no sirs for @xmath10 of norm less than a given size .
pslq may be viewed as a particular case of sird when @xmath268 . similarly with pslq
, sird can be easily generalized to complex field with @xmath269 such that the outputs are in gaussian integer ring and all conclusions mentioned above hold with corresponding modifications .
moreover , sird can also be applied to detect an integer relation in @xmath38 for a given complex vector .
for example , suppose @xmath270 in @xmath271 with vector components @xmath272 where @xmath273 .
then sird can give an sir @xmath14 for @xmath274 , and hence that @xmath275 is an integer relation for @xmath276 , but pslq only can give a gaussian integer relation in @xmath277^n$ ] .
this is one of the biggest differences between sird and pslq .
furthermore , the matrix reducing method in sird is generalized hermite reduction , which avoids lll - type reduction .
this is a difference not only between sird and hjls , but also between sird and pslq because that ( modified ) hermite reduction is not suitable to detect sirs any more . and just the generalized hermite reduction guarantees the correctness of sird .
in theory , the costs of sird and the hjls simultaneous relations algorithm ( see @xcite ) are the same as in corollary [ thm : time - complexity - of - simultaneous - relation ] in the worst case , whereas in practice sird usually needs fewer iterations . for @xmath278 and @xmath279 , hjls outputs @xmath280 after 5 iterations while sird outputs @xmath281 after only 2 iterations .
.comparison of performance results for hjls and sird [ cols="^,^,^,^,^,^,^,^",options="header " , ] table [ tab : comparison of mpfs ] gives a comparison of the digits and complexity of 4 different minimal polynomial finding algorithms in the worst case .
since the algorithm in @xcite can only find the minimal polynomial of a real algebraic number , we do nt compare the complexity with it .
it seems that a lower complexity can be achieved by using some new type lll algorithms , such as l@xmath282 @xcite and h - lll @xcite , but when we apply these new algorithms to find the minimal polynomial we have to choose @xmath31 as in a similar formula with ( [ eq : error - controlling ] ) .
thus multiple precision arithmetic is inevitable .
_ example [ exmp : x^2 - 4x+7 ] ( con . ) .
_ for @xmath283 , its minimal polynomial @xmath284 .
set @xmath285 and @xmath286 .
computing the error tolerance as in equation ( [ eq : error - controlling ] ) gives @xmath287 .
corollary [ cor : epsilon - control ] implies that @xmath288 correct decimal digits are sufficient to guarantee the output is correct .
this example also illustrates that @xmath31 in ( [ eq : error - controlling ] ) is only a sufficient condition on error controlling , but not a necessary one .
the number of iterations and the cost of sird algorithm are related to the parameter @xmath224 . for @xmath289 and @xmath290 ,
if we choose @xmath291 then sird outputs @xmath292 after 12 iterations , however , if we choose @xmath293 , sird outputs @xmath294 after only 6 iterations . in future work
we expect to find the best choice for @xmath224 .
additionally , how to choose the digits such that sird under floating - point arithmetic finds an exact sir is also in our interests .
finally , we see that the mpf algorithm can be used to factor @xmath295 in @xmath296 $ ] like this : solve an approximation root with accuracy satisfying equation ( [ eq : error - controlling ] ) , and call mpf for finding its minimal polynomial which corresponds an irreducible factor of @xmath295 , and then repeat the two steps until @xmath295 has been factored completely .
it is symbolic - numeric and different from traditional algorithms based on hensel lifting . .
shanghai scientific and technological education publishing house , 1996 ( in chinese ) . obtaining exact value by approximate computations .
_ science in china series a : mathematics 50 _ , 9 ( 2007 ) , 13611368 . | let @xmath0 . a simultaneous integer relation ( sir ) for @xmath1 is a vector @xmath2 such that @xmath3 for @xmath4 . in this paper , we propose an algorithm sird to detect an sir for real vectors , which constructs an sir within @xmath5 arithmetic operations , where @xmath6 is the least euclidean norm of sirs for @xmath7 .
one can easily generalize sird to complex number field .
experimental results show that sird is practical and better than another detecting algorithm in the literature . in its application
, we present a new algorithm for finding the minimal polynomial of an arbitrary complex algebraic number from its an approximation , which is not based on lll . we also provide a sufficient condition on the precision of the approximate value , which depends only on the height and the degree of the algebraic number . |
the @xmath8 meson production is an important part of the study program of different nucleus - nucleus collision experiments in the whole range of collision energies : at the ags @xcite , the sps @xcite , rhic @xcite , and even at deeply subthreshold energies at the sis @xcite . as the longest - living vector meson , the @xmath8
is considered to be a good probe of the collision dynamics : it would decay mainly outside the fireball and the daughter hadrons would be weakly affected by rescattering .
the dominant hadronic decay is @xmath9 , hence , the production of @xmath8 s interplays with the production of strange mesons in nucleus - nucleus collisions and can provide a complementary information on the strangeness dynamics . as other neutral vector mesons , the @xmath8 contributes to dilepton production via decays @xmath10 and @xmath11 .
the comparison of @xmath8 meson yields in the hadronic and electromagnetic channels @xcite can reveal interesting information about the later stage of the collision @xcite . according to the su(3 ) quark model the dominant component constituting the @xmath8 meson is a spin - one bound state of @xmath12 and @xmath13 quarks .
hence , the hadronic interactions of phi mesons are subject to okubo - zweig - iizuka rule , which tells that the interactions of a pure @xmath14 state with non - strange hadrons are suppressed . in practice
, the strict implementation of this rule would imply vanishing of @xmath15 and @xmath16 couplings .
indeed , the ozi - forbidden reactions are typically orders of magnitude smaller than the ozi - allowed , e.g. , the ratio of @xmath17 to @xmath8 meson production cross section is about @xcite ~75 .
[ ozisupp ] this value is , however , three times smaller than the one which follows from the su(3 ) symmetry and the experimentally known deviation from the ideal mixing between singlet and octet vector mesons @xcite .
the phenomenological models are able to reproduce the experimental cross section of the @xmath5 reaction by taking into account the ozi - violating @xmath18 coupling adjusted to reproduce the @xmath19 decay width @xcite .
the role of baryon resonances in this process was studied in ref .
@xcite .
since the ozi suppression weakens the @xmath8 production only by the ordinary hadronic matter and is lifted in the quark - gluon medium , the strong , of the order indicated in ( [ ozisupp ] ) , enhancement of the @xmath8 yield was proposed in @xcite as a signal of the quark - gluon plasma formation .
an enhancement of the @xmath8 yield was indeed observed in experiments albeit to a lesser degree : the ratio of the @xmath8 yield in a nucleus - nucleus collision normalized by the number of participating nucleon pairs to the @xmath8 yield in a proton - proton collision lies between 3 and 4 for ags / sps energies @xcite . within the hadrochemical model @xcite the factor 3 of enhancement in @xmath8 production could be explained only by assuming a decrease of hadron masses in hot and dense hadronic matter . in @xcite the main contribution to the @xmath8 yields
are given by the ozi - allowed process with _ strangeness coalescence _ @xmath20 and @xmath21 .
the string - hadronic transport model urqmd @xcite , in which the @xmath8 is produced mainly via @xmath22 reactions and no in - medium effects are included , systematically underestimates the @xmath8 production at sps energies , cf .
figure 9 in ref .
the experiments suggest some similarity in production mechanism of @xmath8 mesons and kaons . in au+au collisions at 11.7@xmath23 gev / c @xcite and in pb+pb collisions at 158@xmath23
gev / c @xcite it was observed that the @xmath8 yield grows with the centrality more strongly than linearly in a similar way as the @xmath24 and @xmath25 yields do .
in contrast , the @xmath26 production scales linearly with a number of participating nucleons @xcite . at the sis the @xmath8 production
was studied at the beam energies about 2 gev per nucleon , which are less than the threshold energy of @xmath8 production in the nucleon - nucleon collision 2.6 gev .
the fopi collaboration measured @xmath8s in ni+ni reaction at 1.93 @xmath23gev @xcite . the reported ratio of @xmath8 to @xmath25 multiplicites is @xmath27 or @xmath28 depending on the temperature of the emitting source ( 130 mev and 70 mev , respectively ) which has to be assumed for an extrapolation to a full solid angle .
these results are considerably larger ( at least by factor 4 ) than the very preliminary estimates announced in @xcite
. such a large @xmath8 abundance can not be described by the transport model @xcite where @xmath8s are produced in reactions @xmath29 and @xmath30 ( @xmath31 ) with the dominant contribution from pion - nucleon reactions .
the calculations with the iqmd code @xcite underestimate the @xmath8 production data @xcite , too .
even the inclusion of other resonance channels @xmath32 and @xmath33 studied in ref .
@xcite would not fully accommodate the results @xcite on @xmath8 multiplicites .
note that the strangeness coalescence process could not contribute much to the @xmath8 yield at these energies since kaons have a long mean free path and most likely leave the fireball right after its creation without any further interaction .
the large value of @xmath34 reported in @xcite seems to be confirmed by further analyses of the data @xcite .
the preliminary results by hades experiments with ar+kcl collision at 1.756 @xmath23gev give also a comparable value for @xmath35 @xcite . in this paper
we would like to address some puzzles of the @xmath8 production in heavy - ion collisions at various energies and propose a new type of the @xmath8 production mechanism_the catalytic @xmath8 production _ by strange particles , e.g. the reactions yy , |k ny , y= , .
[ cathprod ] in contrast to the strangeness coalescence reaction , here the strangeness does not hide inside the @xmath8 but stays in the system and the presence of @xmath36 mesons is unnecessary .
the efficiency of these reactions should be compared with the process @xmath5 , which is found to be dominating in @xcite .
the reactions ( [ cathprod ] ) are ozi allowed , so we win in cross sections compared to @xmath5 a factor of the order of 75 , cf .
( [ ozisupp ] ) .
we lose , however , in the smaller concentration of hyperons and anti - kaons compared to nucleons and pions , respectively .
lccccc @xmath37 & 2 gev/@xmath23 & 4 gev/@xmath23 & 6 gev/@xmath23 & 8 gev/@xmath23 & 10 gev/@xmath23 + @xmath38 & @xmath39 & @xmath40 & @xmath41 & @xmath42 & @xmath43 + @xmath44 & @xmath45 & @xmath46 & @xmath47 & @xmath48 & @xmath49 + @xmath50 & & @xmath51 & @xmath41 & @xmath52 & @xmath53 + @xmath54 & 23 mb & 3.5 mb & 2 mb & 1.5 mb & 1 mb + @xmath55 & & 5.2 mb & 2 mb & 1.3 mb & 0.9 mb + there is no experimental information about the cross sections for the reactions ( [ cathprod ] ) , therefore we will have to rely on some modeling .
let us first estimate how large the cross sections for these reactions should be so that their contributions become comparable to @xmath56 ?
this can be deduced from the following relations ( y ) ( nn ) , + ( |k ny ) ( nn ) .
here @xmath57 number of nucleons participating in the collision , @xmath58 , @xmath59 , @xmath60 , and @xmath61 are the multiplicities of the corresponding particles produced in the collisions . using mid - rapidity data for collisions at ags energies , cf .
@xcite , listed in table [ tab : mult ] and the experimental value @xmath62 mb we get an estimate for the lower bound of the @xmath63 and @xmath64 cross sections .
we see that a few mb is enough to make the cathalitic @xmath8 meson production comparable with the conventional mechanisms at ags energies . for sps energies
these bounds could be even lower , since there are more kaons and hyperons .
although encouraging these rough estimates are not convincing yet .
indeed , the important is not only how much strangeness is produced in the collision but also when it is produced and at what temperature .
the catalytic reactions ( [ cathprod ] ) will be operative if there is enough strangeness produced and the medium is still hot enough . to study various scenarios quantitatively , in section 2
we estimate the cross section of the reactions ( [ cathprod ] ) within a simple model . in section 3
we set up a hadrochemical model for strangeness production .
the @xmath8 production is discussed in section 4 .
we consider the lagrangian of lightest strange particles , i.e. hyperons and kaons , interacting with pions and nucleons and include the @xmath8 meson as a gauge boson of the local u(1 ) symmetry associated with the strangeness conservation & = & | ( i+-m _ ) + |_a ( i+-m _ ) ^a + & + & 12 ( -i ) _ k^ ( + i ) ^ k-12 m_k^2k^ k + & + & c_(|__5 ^a + |^a__5)^_a + c _ i_abc ( |^a__5 ^b ) ^^c + & + & c_kn ( |n _ _ 5 _ a^a ( ^+i g _ ^ ) k+ ( ^-i g _ ^ ) k^|_a^a _ _ 5
n ) + & + & c_kn ( |n _ _ 5 ( ^+i g _ ^ ) k+ ( ^-i g _ ^ ) k^| _ _ 5 n ) , [ lag ] where @xmath67 and @xmath68 denote the isospin triplets of @xmath69 baryons and pions ; kaon and nucleon isospin doublets are @xmath70 and @xmath71 .
the coupling @xmath72 is fixed from the @xmath73 decay , @xmath74 .
the couplings @xmath75 , @xmath76 , @xmath77 and @xmath78 we take from the analysis of meson baryon interactions @xcite .
reaction , width=340 ] in figure [ fig : diag ] the lowest - order diagrams contributing to the @xmath65 process are shown .
the corresponding amplitude reads t&= & g_c_|yy i|u_|y(|p;|s)^ s_|y(|p+|q ) ( q ) _ 5 u_y(p;s)^*_(|q ; ) + & + & g_c_|yyi|u_|y(|p;|s ) ( q)_5s_y(p-|q ) ^u_y(p;s)^*_(|q ; ) + & = & g_c_|yy|u_|y(|p;|s ) j^_|yy(q ) u_y(p;s)^*_(|q ; ) , + & & s_y(p)=,s_|y(p)=. [ ampl ] here we use the following notations : @xmath79 and @xmath80 are the 4-momenta of a hyperon and a pion , respectively , in the initial state , @xmath81 and @xmath82 are the 4-momenta of a hyperon and a phi meson in the final state ; @xmath83 and @xmath84 are masses of the initial and final hyperons ; @xmath85 denotes the dirac bispinor of a hyperon @xmath86 with the momentum @xmath79 and the spin projection @xmath12 ; @xmath87 is the wave function of an outgoing @xmath8 meson with the polarization @xmath88 . for the moment we do not specify in ( [ ampl ] ) the isospin coefficient for the particular reaction . taking into account the on - shell relations for hyperons
@xmath89 we can reduce the current @xmath90 introduced in ( [ ampl ] ) to the on - shell equivalent form j_|yy^(q)=i(m+|m)\ { ^ s_|y(|p+|q ) _ 5 + _ 5s_y(p-|q)^}. the current has the convenient properties @xmath91^\dag\gamma_0 = j_{y\bar{y}}^\mu(-q)$ ] and @xmath92 .
the cross section is determined by the amplitude squared , summed over the spin and polarization of the final hyperon ( @xmath93 ) and phi meson ( @xmath88 ) , and averaged over the spin of the initial hyperon ( @xmath12 ) _
|yy&=&12_s,|s , tt^ + & = & - g^2_c^2_|yy12tr\ { ( |p+|m ) j_,|yy(q)(p+ m ) j^_y|y(-q ) } + & = & g^2_c^2_|yy(m+|m)^2 , where f_|yy(|p,|q , p ) & = & 12tr\ { ( |p+|m)^ s_|y(|p+|q ) _ 5(p+
m ) ^ s_y(p-|q ) _ 5 } + & + & 12tr\ { ( |p+|m ) _ 5s_y(p-|q)^ ( p+ m ) _
5s_|y(|p+|q ) ^ } + & = & , + g_|yy(|p,|q , p ) & = & 12 tr\ { ( |p+|m)^ s_|y(|p+|q ) _
5s_|y(|p+|q ) ^ } + & = & .
using the relations @xmath94 , @xmath95 , @xmath96 , and @xmath97 we can express @xmath98 as a function of two invariants @xmath12 and @xmath99 .
then the total cross section can be evaluated as ( cf .
@xcite ) _ |yy(s)= _ -1 ^ 1 d x _ |yy(s , u(s , x ) ) [ crosssec ] where u(s , x)=12 ( m^2_+m_^2+|m^2+m^2-s)+ - 2p_y(s)p_|y(s ) x , + p_y^2(s)=((s - m_^2-m^2)^2 - 4 m_^2 m^2 ) , p_|y^2(s)=((s - m_^2-|m^2)^2 - 4 m_^2 in calculations of the particle production in heavy - ion collisions one usually uses isospin - averaged cross sections which can be defined as the sum of cross sections for active ( allowed ) isospin channels divided by the total number of isospin combinations of colliding particles . in @xmath100 channel the following three reactions are possible @xmath101 . since all these processes have the same cross section with the isospin coefficient equal to one ,
the isospin - averaged cross section is _ ( s)=_(s ) .
[ sigls ] in the @xmath102 channel we have six possible reactions @xmath103 , @xmath104 , and @xmath105 .
the entrance channel , @xmath106 , counts 9 possible isospin combinations .
the isospin averaged cross section is then _
( s)=23_^+^-^0(s ) = 23_(s ) .
[ sigss ] the @xmath107 channels is represented by three reactions @xmath108 and @xmath109 out of nine available for the @xmath106 entrance channel . _ ( s)=13_^+^-(s ) = 13_(s ) .
[ sigsl ] meson production in @xmath110 ( left panel ) and in @xmath66 ( right panel ) reactions . on the left panel two lowest lines depict the parameterisations of the @xmath111 reaction from ref .
@xcite ( solid line ) and ref .
@xcite ( dashed line ) in comparison with the experimental data from @xcite .
both the paremeterizations and the data are scaled up by factor 5.,title="fig:",width=264 ] meson production in @xmath110 ( left panel ) and in @xmath66 ( right panel ) reactions . on the left panel two lowest lines depict the parameterisations of the @xmath111 reaction from ref .
@xcite ( solid line ) and ref .
@xcite ( dashed line ) in comparison with the experimental data from @xcite .
both the paremeterizations and the data are scaled up by factor 5.,title="fig:",width=264 ] the isospin - averaged cross sections ( [ sigls],[sigss],[sigsl ] ) are shown in figure [ fig : cross ] as functions of the collision energy above the threshold @xmath112 .
for comparison , we depict the cross section for the @xmath113 reaction in the parameterisation @xcite ( solid line ) and @xcite ( dashed lines ) scaled up by factor 5 .
we see that the cross sections of the phi production on hyperons are in average about 50 times larger than the production on nucleons ( at @xmath114 mev ) reaction is @xmath115 . ] .
the diagrams contributing to @xmath116 processes in the lowest order are depicted in figure [ fig : diagkn ] .
reaction , width=453 ] the reaction amplitude is it_|yn&= & g _ c_kn|y
|u_|y(|p;|s ) ( ^-(q-|q ) d_k(q-|q)(2q-|q)^)_5 u_n(p;s)^*_(|q ; ) + & -&g _ c_kn|y |u_|y(|p;|s ) ^s_|y(|p+|q)(q)_5 u_n(p;s ) ^*_(|q ; ) + & = & g _ c_kn|y
|u_|y(|p;|s ) j_|y n(|q , q ) u_n(p;s ) ^*_(|q ; ) , + & & d_k(q)=(q^2-m_k^2)^-1 .
[ amplkn ] as before , we do not write explicitly the isospin factors for each reaction .
they will be restored later . using the on - shell conditions we rewrite the current @xmath117 in the on - shell equivalent form j_|y n^(|q , q)=(|m+ m ) ( d_k(|q- q ) ( 2 q-|q)^+^ s_|y(|p+|q ) ) _ 5 . the squared and spin summed amplitude is then _
|yn&=&12_s,|s , t_|ynt_|yn^=12tr\ { ( |p+|m)_5(p+ m_n)_5 } d_k^2(q-|q ) ( 2 q-|q)^2 + & + & 12tr\ { ( |p+|m)_5(p+ m_n ) _ 5 s_y(p-|q ) _ } d_k(q-|q ) ( 2 q-|q)^ + & + & 12tr\ { ( |p+|m )
_ s_y(p-|q ) _ 5(p+ m_n ) _ 5 } d_k(q-|q ) ( 2 q-|q)^ + & + & 12 tr\ { ( |p+|m)^ s_|y(|p+|q ) _ 5(p+ m_n ) _
5s_|y(|p+|q ) ^ } + & = & g_|y n(|p,|q , p ) - 2 + & - & 4 the function @xmath118 follows from @xmath119 after the replacement @xmath120 . using the kinematic relations @xmath121 , @xmath122 , @xmath123 , @xmath124 , @xmath125 , and @xmath126 we can express @xmath127 as a function of invariants @xmath12 and @xmath128 .
the cross section of @xmath129 is given by _
-1 ^ 1 d x _ |yn(s , t(s , x ) ) , [ sigkn ] + t(s , x)=m_^2+m_k^2 - 2 + 2p_k n(s)p_|y(s ) x , + p_kn^2(s)=((s - m_k^2-m_n^2)^2 - 4 m_k^2 m_n^2),p_|y^2(s)=((s - m_^2-|m^2)^2 - 4 m_^2 |m^2 ) .
the isospin averaged cross section for the @xmath130 reaction is equal to _ kn(s)=13_n(s ) [ signl ] where we use that @xmath131 .
evaluating the isospin averaged cross section for the @xmath132 reaction we take into account that the reactions with charged @xmath69 hyperon in the final state have additional isospin factor 2 , then _
kn(s)=_n(s ) . [ signs ]
the cross sections ( [ signl ] ) and ( [ signs ] ) are shown on the right panel in figure [ fig : cross ] . in comparison with the @xmath133
cross section the @xmath66 cross sections are about 60 times larger at @xmath134 mev .
in order to estimate the efficiency of reactions with the @xmath8 production on hyperons and kaons , we have to study , first , the dynamics of the strangeness production in the course of a nuclear collision .
this question was addressed within hadro - chemical models in @xcite . in our consideration
we make the following assumptions : ( a ) the fireball matter is baryon - dominated ; ( b ) strangeness can be considered perturbatively , i.e , the number of produced strange particles , say kaons , is much smaller than the baryon number . in the baryon - dominated matter
the particles carrying strange quarks , @xmath25 , @xmath135 , @xmath136 , @xmath69 and heavier hyperons have short mean free path .
hence , we can assume that they are confined inside the fireball and remain in thermal equilibrium with other species till the fireball breaks up .
oppositely , the particles with anti - quark , @xmath137 and @xmath138 mesons , have larger mean free paths and can leave the fireball at some earlier stage of the collision . based on this observation the strangeness separation scenario
was formulated in refs .
it can be suitable for nucleus - nucleus collisions at ags energies .
the analysis in @xcite shows that it can be also partially valid at sps energies .
however , at what stage of the collision kaons ( @xmath24 and @xmath138 ) leave the fireball is not important for our further calculations if the concentrations of strange particles remain small .
since in such a case the reactions with strangeness annihilation , @xmath139 and @xmath140 , can always be neglected and the number of strange particles accumulated in the system is equal to the number of produced kaons .
the time dependence of the fireball temperature and baryon density will be parameterized in the form of a scaling solution of the hydrodynamic equations @xcite t(t)=,_b(t ) & = & , [ tr - evol ] where @xmath141 and @xmath142 are the initial ( maximal ) temperature and density of the fireball , and @xmath143 is the typical time scale of the fireball expansion .
the evolution of the kaon density is described by the differential equation -_k(t ) & = & ( t),_k(0)=0 .
[ kprod ] the second term on the left - hand side takes into account the reduction of the kaon density due to the expansion of the system ( dilution ) .
we exploit here the relation for the volume change @xmath144 . the kaon production rate on the right - hand side is determined by the processes with @xmath145 , @xmath146 , @xmath147 , @xmath148 and @xmath149 in the initial states & = & _ n^k x _ _ n + _ ^kx _ _ + _ nn^kx _ n^2 + _ ^k|k _ ^2 + _ n^kx _ n _ .
[ rate ] as argued above we do not include on the right hand side in ( [ kprod ] ) the rates of kaon annihilation processes .
they are clearly absent if @xmath36 mesons leave fireball immediately after their creation .
if @xmath36 mesons stay in the system the annihilation reactions are still very suppressed as being proportional to @xmath150 .
the transport coefficient @xmath151 is the cross section @xmath152 of the binary reaction @xmath153 averaged with the particle relative velocity @xmath154 over the momentum distributions of colliding particles .
following ref .
@xcite we write _
ab^kx(t ) [ kappadif ] = , where @xmath155 and @xmath156 are the macdonald s functions of the first and second orders , and @xmath157 is the reaction threshold . among the @xmath145 induced reactions we include @xmath158 , @xmath159 and @xmath160 processes : @xmath161 . for the isospin
averaging we use the relations @xcite ( nk ) & = & 12 ( ^- pk^0 ) , + ( nk ) & = & 12 ( ( ^- p^- k^+)+ ( ^- p^0 k^0)+ ( ^+ p^+ k^+ ) ) , + ( n n k |k ) & = & 3 ( ^- pn k^0 k^- ) . the parameterizations for the elementary @xmath162 cross sections are taken from refs .
@xcite and for the @xmath163 cross section from ref . @xcite . in the @xmath146 channel
we include only the process with @xmath164 and @xmath106 in the final state : @xmath165 .
the isospin averaged cross section of the @xmath166 reactions we derive at hand of the relations ( 21 - 31 ) in ref .
@xcite ( k ) & = & ( 5 ( ^-^++^0 k^+ ) + 6 ( ^0 ^ 0 ^ - k^+ ) + & & + 21 ( ^+^0 ^ 0 k^+ ) + 8 ( ^+^-^- k^+ ) ) + ( k ) & = & 13 ( ^-^++k ) with the elementary cross section given in ref .
the transport coefficient of kaon production in the @xmath167 channel has three contributions @xmath168 and the parameterisations of the elementary @xmath169 and @xmath170 cross sections are taken from ref .
the cross section for the @xmath171 reaction which determines @xmath172 we take from ref .
@xcite too . the cross sections for reactions @xmath173 which contribute to @xmath174 are taken from ref .
@xcite in our model consideration we neglect contributions from other process with kaon production which have higher thresholds and , therefore , are less relevant in the range of temperatures we consider below @xmath175 .
the interaction among pions , nucleons and deltas is strong enough to sustain them in thermal and chemical equilibrium during the whole fireball evolution .
hence , their densities are given by _
( t ) & = & _ 0^ p p^2 ( -/t(t ) ) , + _
n(t ) & = & _ 0^ d p p^2 ( _ b(t)/t(t)-/t(t ) ) , + _ ( t ) & = & _ 0^ d p p^2 ( _ b(t)/t(t)-/t(t ) ) .
we disregard the finite width of @xmath176s and treat them as stable particles with the mass @xmath177 mev . neglecting the contributions of hyperons , heavier resonances and anti - particles we can find the baryon chemical potential from the following relation _
b(t)=-t(t ) .
[ mub ] in order to study different fireball evolution scenarios we consider three sets of parameters for the maximal temperature and density & i. & t_m=110 mev,_m=4 _ 0,=0.2 , + & ii.&t_m=130 mev,_m=5 _ 0 , = 0.3 , + & iii . &
t_m=150 mev,_m=6 _ 0,=0.5 , [ cases ] where @xmath178@xmath179 is the nuclear saturation density .
the dependence of the temperature and density on time is exemplified on the left panel in figure [ fig : t - r ] . on the right panel in figure
[ fig : t - r ] the rates of kaon productions in various processes are shown .
the most efficient ones are processes with @xmath145 and @xmath149 in the incoming channel .
when all ingredients entering the rate ( [ rate ] ) are specified the density of produced kaon follows after integrating ( [ kprod ] ) _
k(t)=_b(t ) _ 0^t dt = _ b(t)t_0 _ 0^t / t_0 dx .
[ kdens ] the produced kaons escape from the fireball leaving behind a medium with some net strangeness accumulated in form of @xmath180 , @xmath136 and @xmath69 .
( @xmath181 and @xmath182 can be neglected since their abundances are of the higher orders in terms of the small parameter @xmath183 . )
the occupation of particles caring the strange quark is according to their statistical weights _
i(t ) & = & _ k(t ) , i=|k , , , + n_|k(t ) & = & _ 0^ ( -/t(t ) ) , + n_(t ) & = & _ 0^ ( _ b(t)/t(t)-/t(t ) ) , + n_(t ) & = & 3_0^ ( _ b(t)/t(t)-/t(t ) ) .
[ strange - dist ] we assume , thereby , that the reactions redistributing the strange quark among @xmath180 and hyperons are swift
. the evolution of strangeness in the course of a collision is illustrated in figure [ fig : str ] where we depict the ratios of strange particles , @xmath36 , @xmath184 , @xmath136 , and @xmath69 , to the baryon density _
i(t)=_i(t)/_b(t ) . note that in view of eqs .
( [ tr - evol ] ) and ( [ strange - dist ] ) the kaon density ( [ kdens ] ) and the densities of other strange particles are linearly proportional to the parameter @xmath143 for the fixed @xmath185 . presenting the results in figure [ fig : str ]
we take into account this scaling , i.e. , the lines are to be scaled by factor @xmath186 .
we observe that strangeness is dominantly presented in form of @xmath69 hyperons and that the number of produced strange particles doubles roughly with an increase of the temperature by every 20 mev from case i to case iii in ( [ cases ] ) .
in this section we discuss how the catalytic reactions can influence various aspects of @xmath8 production in heavy - ion collisions .
first we consider whether the catalytic reaction can be an efficient source of @xmath8 mesons compared to a conventional reactions .
the rate equation for the @xmath8 meson production has the form similar to ( [ kprod ] ) -_(t ) = r^_n(t)+_|y , y= , r^_y|y(t)+ r_|k n^(t)+ , [ phirates ] where on the right hand side we explicitly write down the @xmath8 production rates from the conventional @xmath5 reactions @xmath188 and the catalytic reactions on hyperons @xmath189 and on anti - kaons @xmath190 the ellipses stand for further @xmath8 production reactions like those discussed in @xcite and the @xmath8 absorption processes , e.g. , @xmath191 and @xmath192 , etc .
the rates of various process are shown figure [ fig : frates ] .
the @xmath8 production in @xmath145 collisions starts of course at the very beginning and gradually falls off as the fireball cools down and expands .
the rates of catalytic reactions increase , the more strange particles are produced , reach the maximum at ( 0.30.4)@xmath193 and drop off later . for set i with the maximal temperature @xmath194 mev the net rate of catalytic reactions yields to that of @xmath5 . for set ii with @xmath195 mev the rates become comparable for times @xmath196 . for @xmath197
mev in set iii the catalytic production of @xmath8 exceed the rate of the @xmath198 reaction already at @xmath199 . note that the numerical values for the rates in figure [ fig : frates ] correspond to the fireball expansion time @xmath200 fm .
if the collision lasts longer then the curves for catalytic reactions have to be scaled up by a factor @xmath201 , since the number of the strange particles is proportional to the expansion time .
this will make the catalytic reaction efficient even for the smaller temperature in set i. the above estimates show that the catalytic processes can be potentially important for the @xmath8 production in heavy - ion collisions .
it is interesting now to investigate how they can affect other @xmath8 production patterns , and whether their dominance can lead to any contradiction with experimental data .
we discuss now the centrality dependence of the @xmath8 production . changing centrality
one changes the volume of the system .
we will use the mean number of projectile participants , @xmath202 , as the measure for initial volume of the fireball created in the collision , @xmath203 .
the @xmath202 can be directly related to the energy deposited in the zero - degree calorimeter , cf .
if there is only one changing parameter with the unit of length as in the case of a symmetrical collision at a fixed collision energy , @xmath204 , the scaling properties of hydrodynamics imply that the collision time is of the order @xmath205
the multiplicities of particle species which reach the full chemical equilibrium , e.g. pions and @xmath176s , are proportional to the volume of the fireball and scale , therefore , as @xmath206 and @xmath207 .
the multiplicity of kaons ( @xmath24 ) , which have a long mean free path and leave the system right after being produced , scales as @xmath208 . according to ( [ strange - dist ] )
other strange particles should follow the @xmath24 scaling , @xmath209 .
the number of produced @xmath8 mesons can be estimated as n_&~ & _ n t_0+_y t_0 + _ |k n t_0 + & ~ & a_convn_pp^4/3 + a_cat n_pp^5/3 the term @xmath210 is due to the conventional production reactions like @xmath198 whereas the term @xmath211 one corresponds to the catalytic reactions which contribute with an additional factor @xmath143 . for the ratios experimentally observed in ref .
@xcite we find ~a ( ) ^1/3+b ( ) ^2/3 , ~a + b ( ) ^1/3 , [ phiratios ] where @xmath212 and @xmath213 parameterize the relative strength of conventional and catalytic processes , and @xmath23 is the number of nucleons in the colliding nuclei . in figure
[ fig : central ] we compare these parameterizations with the only available data of the @xmath8 yield centrality dependence for au+au collisions at 11.7 @xmath23gev @xcite .
first , we adjust parameters @xmath214 and @xmath215 separately and obtain dashed and solid curves , respectively .
because of large error bars both fits with @xmath216 or @xmath217 equally well go through data points .
so the data on centrality dependence can not rule out even the dominance of the catalytic reactions .
the optimal fits are reached when both parameters are activated , dashed - dotted lines in figure [ fig : central ] . comparing the values of parameters @xmath214 and @xmath215 from these fits
we can conclude that the catalytic reaction contribution can be about 30%-40% for @xmath218 .
the systematics of @xmath8 rapidity distributions in pb+pb collisions at the sps is reported in ref .
the distributions can be fitted with a sum of two gaussian functions placed symmetrically around mid - rapidity f(y , a,)= = , [ 2gauss ] here @xmath219 is the rapidity in c.m system .
the width of the distribution is characterized by the root mean square @xmath220 .
the left panel of figure [ fig : rap ] depicts the @xmath221s versus the beam rapidity for @xmath8 mesons , constructed from the parameters @xmath222 and @xmath223 given in table v in @xcite , together with those for @xmath224 , @xmath24 and @xmath25 @xcite .
@xcite pointed out that the width of the @xmath8 meson distribution does not fit into the systematics observed for the other mesons but increases much faster with the collision energy .
it was emphasized in @xcite that the steep rise of the @xmath8 distribution width can not be explained by the hadronic process @xmath225 .
the rapidity distribution of @xmath8 from the latter process can be estimated as the product of @xmath24 and @xmath25 rapidity distributions .
the width of the product was calculated in @xcite under the assumption of the single - gaussian sources for kaons with @xmath226 and @xmath227 as _ k^+k^-=. [ gauss - rms ] we show this result on the right panel of figure [ fig : rap ] by the dash - dot - dot line .
the resulting width is much smaller than that observed in the experiment .
question is if any other @xmath8 production mechanism can produce the steep rise of the @xmath221 with the collision energy .
we note that the rapidity distribution of hyperons increases much faster than those for mesons as the hyperons are dragged with nucleons to forward and backward rapidities ( an effect of partial transparency ) .
the @xmath221s of @xmath136 reported in ref .
@xcite are shown on the left panel in figure [ fig : rap ] by solid circles .
the distribution width for 158 @xmath23gev was not determined experimentally , therefore we calculated it from the rapidity distribution obtained theoretically in @xcite within a hydrodanamic model which successfully reproduces the particle production .
we assume that the rapidity distributions of particles do not change after some initial stage when nuclei are passing throuhgh each other .
this implies the absence or weakness of acceleration and diffusion processes .
the collision kinematics is restricted mainly to the exchange of transverse momenta .
then the rapidity distribution of @xmath8s produced in the reaction @xmath228 is roughly proportional to the product of rapidity distributions of colliding particle species 1 and 2 .
if the distributions are given by the two gaussian distributions ( [ 2gauss ] ) the rms of the resulting @xmath8 distribution is given by ^2_12= + = + + .
[ rms12 ] this expression differs from ( [ gauss - rms ] ) applied in @xcite . using the parameters of @xmath24 and @xmath25 from @xcite and eq .
( [ rms12 ] ) we obtain rmss for the @xmath229 reaction shown in figure [ fig : rap ] ( right panel ) by dotted line .
they are somewhat larger than those obtained in @xcite but still much smaller than empirical ones .
in contrast , the width of @xmath8 rapidity distributions from the reactions involving @xmath136 particles , @xmath230 and @xmath231 ( solid and dashed lines in figure [ fig : rap ] ) rises much faster and is comparable with the experimental results for beam energies between 20 @xmath23gev and 80 @xmath23gev . in collisions at 158 @xmath23gev
the @xmath8 distribution is much broader than our estimates .
perhaps , some new non - hadronic mechanism of the @xmath8 production becomes operative at this energy .
in this paper we study a new mechanism of @xmath8 meson production in nucleus - nucleus collisions the catalytic reactions on strange particles , e.g. , @xmath110 and @xmath232 .
these reactions are ozi - allowed and their cross section can be by an order of magnitude larger than the cross section of conventional ozi - suppressed @xmath8 production reactions @xmath5 and @xmath233 , etc , considered so far .
these reactions require only one strange particle in the entrance channel and therefore are less suppressed than the strangeness coalescence reactions , @xmath234 and @xmath235 in collisions where strangeness is statistically suppressed . using a hadronic lagrangian ( [ lag ] ) we estimate the @xmath110 and @xmath236 cross sections to be roughly of the order of 1 mb . in order to estimate the efficiency of the new reactions we calculate the evolution of the strangeness content of the fireball within a hadrochemical model .
the main assumption we follow is that kaons ( @xmath24 and @xmath138 ) can leave the fireball freely after being created whereas the strange particles ( hyperons , @xmath25 and @xmath135 ) remain in the fireball in thermal and chemical equilibrium .
thus the net strangeness accumulated in the fireball is growing in the course of the collision .
we compare the rates of @xmath8 production in the catalytic reactions and in the @xmath5 reaction .
the former can be competitive in collisions with the maximal temperature above 110 mev and the collision time @xmath237 fm .
the efficiency of catalytic reaction increases if some strangeness is presented in the fireball right in the beginning after the first most violent nucleon - nucleon collisions and if the fireball lifetime is longer .
we discuss how the catalytic reaction could affect the centrality dependence of the @xmath8 yield .
since the catalytic rates depend on the concentration of strange particles accumulated during some time , the resulting number of @xmath8s produced in such a reaction would grow @xmath238 , where @xmath143 is the collision time , in contrast to the number of @xmath8s from conventional reactions @xmath56 growing @xmath239 . in the scaling regime of hydrodynamics the collision time can be related to the typical length scale of the system determined by its volume @xmath240 , where @xmath241 is the centrality criterion the mean number of projectile participants .
taking two types of processes into account we can parameterize the experimental centrality dependence of the @xmath242 and @xmath243 ratios measured in au+au collision at 11.7 @xmath23gev as follows 10 ^ 3 & = & 7 ( ) ^1/3 + 3 ( ) ^2/3 , + 10 ^ 2 & = & 1.7 + 1.4 ( ) ^1/3 , a=197 . here
the first terms are due to the conventional reactions involving non - strange particles and the second terms are due to the catalytic @xmath8 production on strange particles .
relative strength of two terms shows that the contribution of the catalytic reactions can be up to 30%40% at the ags energies for @xmath218 . analyzing the @xmath8 rapidity distributions at sps energies we find that the strong rise of the distribution width with the collision energies between 20 @xmath23gev and 80 @xmath23gev can be explained by the @xmath8 production in @xmath164 and @xmath244 collisions .
the present analysis of the catalytic mechanisms of the @xmath8 production is only exploratory and should merely serve as an invitation for further investigations .
particularly , a more reliable calculation of the cross sections for the @xmath245 and @xmath236 reactions and fireball evolution is mandatory .
we gratefully acknowledge the support by vega under nr .
99 akiba y _ et al _
( e802 collaboration ) 1996 _ phys .
lett . _ * 76 * 2021 - 4 back b b _ et al _ ( e917 collaboration ) 2004 _ phys .
_ c * 69 * 054901 ( 10pp ) afanasiev s v _
( na49 collaboration ) 2000 _ phys .
_ b * 491 * 59 - 66 alessandro b _
et al _ ( na50 collaboration ) 2003 _ phys .
_ b * 555 * 147 - 55 adamova d _
et al _ ( ceres collaboration ) 2006 _ phys .
lett . _ * 96 * 152301 ( 4pp ) de falco a _
( na60 collaboration ) 2006 _ nucl .
a * 774 * 719c-22c alt c _ et al _ ( na49 collaboration ) 2008 _ phys .
_ c * 78 * 044907 ( 15pp ) adams j _ et al _ ( star collaboration ) 2005 _ phys .
_ b * 612 * 181 - 9 adler s s _ et al _ ( phenix collaboration ) 2005 _ phys
_ c * 72 * 014903 ( 23pp ) mangiarotti a _ et al _ ( fopi collaboration ) 2003 _ nucl .
phys . _ a * 714 * 89 - 123 friese v for na49 collaboraion 1997 _ j. phys .
g : nucl . part .
* 23 * 1837 ( 4pp ) jouan d for na50 collaboraion 2008 _ j. phys .
g : nucl . part
. phys . _ * 35 * 104163 ( 4pp ) floris m for na60 collaboraion 2008 _ j. phys .
g : nucl . part .
phys . _ * 35 * 104054 ( 4pp ) johnson s c , jacak b v and drees a 2001 e_ur . phys .
c * 18 * 645 - 9 filip p and kolomeitsev e e 2001 _ phys .
_ c * 64 * 054905 ( 8pp ) kolomeitsev e e and filip p 2002 _ j. phys .
_ * 28 * 1697 - 1705 bleicher m 2003 _ nucl .
phys . _ a * 715 * 85c-94c sibirtsev a , haidenbauer j and meiner u - g 2006 _ eur .
a * 27 * 263 - 8 tsushima k , huang s w and faessler a 1994 _ phys .
_ b * 337 * 245 - 53 tsushima k , huang s w and faessler a 1997 _ austral .
j. phys . _
* 50 * 35 ( reprint nucl - th/9602005 ) cassing w , bratkovskaya e l , mosel u , teis s and sibirtsev a 1997 _ nucl .
phys . _ a * 614 * 415 - 32 | the phi meson production on hyperons , @xmath0 and anti - kaons @xmath1 is argued to be a new efficient source of phi mesons in a nucleus - nucleus collision .
these reactions are not suppressed according to okubo - zweig - izuka rule in contrast to the processes with non - strange particles in the entrance channels , @xmath2 and @xmath3 with @xmath4 .
a rough estimate of the cross sections within a simple hadronic model shows that the cross sections of @xmath0 and @xmath1 reactions can exceed that of the @xmath5 reaction by factors @xmath6 and @xmath7 , respectively . in the hadrochemical model for nucleus - nucleus collisions at sis and lower ags energies
we calculate the evolution of strange particle populations and phi meson production rate due to the new processes .
it is found that the catalytic reactions can be operative if the maximal temperature in nucleus - nucleus collisions is larger than 130 mev and the collision time is larger than 10 fm .
a possible influence of the catalytic reactions on the centrality dependence of the @xmath8 yield at ags energies and the @xmath8 rapidity distributions at sps energies is discussed . |
the mass of the @xmath0 meson has been a controversial issue in recent years . before 2000 ,
three different experiments @xcite yielded comparable masses for the @xmath0 meson .
the particle data group ( pdg ) then used these results to calculate a weighted mean mass @xmath4mev @xcite . in 2002 , the na48 collaboration published @xcite a very precise result , @xmath5 mev , which deviated significantly from the world average adopted by the pdg . including the na48 measurement in the average , the pdg in 2004 @xcite obtained the value @xmath6mev , almost 0.5mev higher than the value reported previously .
this created the motivation to repeat the previous mainz @xcite measurement at mami , especially after another precise measurement by the gem collaboration at the cosy facility @xcite gave the result @xmath7mev , in agreement with the old measurements of the @xmath0 mass . since the previous @xmath0 mass measurement at mainz @xcite , the mami electron accelerator @xcite
was significantly improved to provide a more precisely known electron beam energy and much higher beam stability .
in addition , on - line monitoring of the electron and photon beam positions in the experimental hall was introduced .
furthermore , the old taps setup with its limited angular coverage was replaced with the large acceptance crystal ball @xcite photon spectrometer , which allowed an improvement in the detection efficiency for the two most prominent neutral decays of the @xmath0 meson , @xmath8 ( br=39.31% ) and @xmath9 ( br=32.56% ) .
together with the high energy resolution of the tagged photon beam offered by the recently developed tagger focal - plane microscope detector @xcite , these improvements provided the more accurate determination of the @xmath0 mass presented in this paper .
this work describes the measurement of the @xmath0 photoproduction threshold @xmath10 @xcite from data measured in 2004 - 2005 with the crystal ball / taps detector system and the tagger focal - plane microscope detector . from kinematics of the reaction @xmath1 the @xmath0 meson mass @xmath11
was calculated using @xmath12 where @xmath13 is the proton mass and @xmath14 is the speed of light in vacuum .
in mainz the real photon beam was produced by bremsstrahlung of the 883mev electrons from mami - b @xcite on a 100@xmath15 m thick diamond radiator .
the absolute electron energy , @xmath16 , of the incident beam was precisely determined @xcite in the third race - track microtron of mami - b with a total uncertainty of about @xmath17kev .
the photon energies were determined using the glasgow photon tagging spectrometer ( tagger ) @xcite , which provided a tagged photon flux of roughly @xmath18s@xmath19mev@xmath20 at a beam current of about 35na .
for the first time the tagger focal - plane microscope detector @xcite was used to improve the tagged photon energy resolution .
the microscope detector was placed in front of the main focal - plane spectrometer ( see fig.[fig : tagger ] ) , so that it covered the region around the @xmath0 production threshold ( @xmath21mev ) from @xmath22mev to @xmath23mev at an electron beam energy @xmath24mev .
made of 96 scintillator strips overlapping to one third , it provided 191 tagging channels with a higher energy resolution of about 0.29mev per channel compared to approximately 1.8mev available from the main focal - plane detector . into a separate experimental area , where they can induce reactions in the experimental target . making a coincidence between the reaction products in the experiment and the electron in the tagger
, one can determine the energy of the photon @xmath25 as the difference between the main beam energy @xmath26 and the electron energy @xmath27 measured with the tagger @xmath28 .
the 4.76 cm long liquid hydrogen target was located at the center of the crystal ball photon spectrometer .
the crystal ball @xcite , covering polar angles between @xmath29 and @xmath30 , consisted of 672 nai(tl ) crystals ( see fig.[fig : cb ] ) and had two openings for the beam in forward and backward directions .
each nai(tl ) crystal had the form of a truncated 41 cm long pyramid and was equipped with an individual photomultiplier . in order to distinguish between neutral and charged particles detected by the crystal ball , the system was equipped with a particle identification detector ( pid ) @xcite .
pid was a cylindrical detector , consisting of 24 2 mm thick plastic scintillator strips , arranged parallel to the photon beam axis .
the forward wall detector , taps @xcite , had 510 baf@xmath31 hexagonally shaped crystals , each equipped with a 5 mm thick plastic scintillator for identifying charged particles .
a single baf@xmath31 crystal was 25 cm long and had an inner diameter of 5.9 cm .
the taps detector , intended for detecting particles in the forward direction ( @xmath32 ) , was located at a distance of 173 cm from the crystal ball center , making it possible to use the time - of - flight analysis for the particle identification .
the experimental trigger comprised two levels . in the first level , the total energy deposited in the crystal ball was checked .
if the sum of all photomultiplier analog signals exceeded a threshold corresponding to about 390mev , the event was accepted .
the second - level trigger included a condition on the crystal ball sector multiplicity .
the 672 crystals of the spectrometer were grouped into 45 sectors of up to 16 crystals each . if at least one of the 16 signals exceeded a threshold of 20 to 40mev , depending on the relative energy calibration of the photomultiplier signals , the sector contributed to the multiplicity .
all events with multiplicity m@xmath333 and every third event with m@xmath332 were recorded for further analysis .
the latter condition was especially important for detection of the @xmath34 decay .
special care was taken of the energy calibration of the tagger microscope with electrons of different known energies from mami .
the mami accelerator can produce beam energies from 180mev upwards in steps of 15mev , and it is possible to make accurate measurements of the beam energy .
originally designed to produce electrons of maximum energy 855mev , mami can also produce a beam of energy 883mev by slightly increasing the magnetic field of the bending magnets and slightly raising the energy gain per circulation . [
cols="<,<,>",options="header " , ] the average of the three values listed in the first row of table [ tab : res ] resulted in the @xmath0 mass @xmath35 where the first uncertainty is due only to statistics , and the second originates from the uncertainty in the photon beam energy calibration .
the result of this work supports the three most precise measurements by the na48 @xcite , kloe @xcite and cleo @xcite collaborations and disagrees by about @xmath3 with the smaller value obtained by the gem @xcite collaboration .
the @xmath0 mass determined in this work is plotted in fig.[fig : prev_eta_mass ] with the other measurements in the order of the year of publication .
the disagreement with the previous measurement @xcite at mami is most probably due to the lack of online beam monitors that resulted in underestimation of the systematic uncertainty for the photon tagger energy calibration .
this article describes the determination of the @xmath0 mass by measuring the threshold of the reaction @xmath1 at the mami accelerator .
the high resolution tagger microscope was used for the first time to get the total cross section of the reaction .
the @xmath0 mesons were selected by identifying the decay products of the two most prominent neutral decays , @xmath8 and @xmath9 , in the crystal ball detector .
the three precise @xmath0 mass measurements by the na48 , kloe and cleo collaborations were confirmed , though the result presented in this article disagrees with the gem collaboration measurement .
the uncertainty for the new @xmath0 mass measurement has been improved in comparison to the previous mainz experiment by a factor of @xmath36@xmath37 .
the disagreement with the previous mami measurement is most probably due to the underestimated systematic uncertainty of the old result .
+ + _ acknowledgments : _ the authors wish to thank the accelerator group of mami for the precise and very stable beam conditions .
this work was supported by the deutsche forschungsgemeinschaft ( sfb 443 , sfb / tr 16 ) , the dfg - rfbr ( grant no .
05 - 02 - 04014 ) , european community - research infrastructure activity under the fp6 structuring the european research area programme " ( hadronphysics , contract no .
rii3-ct-2004 - 506078 ) , the nserc ( canada ) , schweizerischer nationalfonds , u.k .
epsrc , primenet network , and the u.s .
doe and u.s .
we thank the undergraduate students of mount allison university and the george washington university for their assistance . | a new precise determination of the @xmath0 meson mass is presented .
it is based on a measurement of the threshold for the @xmath1 reaction using the tagger focal - plane microscope detector at the mami - b facility in mainz .
the tagger microscope has a higher energy resolution than the standard tagging spectrometer and , hence , allowed an improvement in the accuracy compared to the previous @xmath0 mass measurement at mami - b .
the result @xmath2 mev agrees very well with the precise values of the na48 , kloe and cleo collaborations and deviates by about @xmath3 from the smaller , but also very precise value obtained by the gem collaboration at cosy . |
this paper intends to shed some light on the following question in the theory of finite dimensional completely integrable hamiltonian systems : _ suppose that @xmath15 form an integrable system on a @xmath16-dimensional symplectic manifold @xmath17 .
what is the relation between the following conditions ? _ * _ the flow of at least one of the functions @xmath18 , @xmath19 , is periodic ; _ * _ existence of a set @xmath20 of non - degenerate focus - focus singularities of the joint map @xmath21 ; _ * _ existence of a set @xmath22 of non - degenerate hyperbolic singularities of @xmath23 ; _ * _ existence of set @xmath24 of degenerate singularities of @xmath23 .
_ several results are known which give partial answers to this question .
for instance , under general assumptions on @xmath25 and @xmath3 , if condition ( 1 ) holds for all @xmath26 , then @xmath27 in fact , there is a complete theory for these systems - called _ toric integrable systems _ - when @xmath3 is compact , by atiyah , guillemin - sternberg , and delzant @xcite .
the results of this paper imply that there are integrable systems for which ( 1 ) , ( 2 ) , ( 3 ) , and ( 4 ) hold simultaneously with @xmath28 , @xmath29 , and @xmath30 ( this is the content of theorems [ gen : thm ] , [ mt2 ] , and [ thm : main ] ) .
moreover , it seems quite plausible that under some general condition , ( 1 ) , ( 2 ) and ( 3 ) with @xmath28 and @xmath29 will imply that @xmath30 in ( 4 ) ( see question [ hc ] ) . because several of the technical tools we use are exclusive to dimension four , from now on we assume that @xmath31 .
let @xmath32 be a connected symplectic @xmath1-manifold , that is , @xmath3 is a smooth connected @xmath1-manifold , and @xmath33 is a non - degenerate closed @xmath34-form on @xmath3 , i.e. a _
symplectic form_. a smooth function @xmath35 induces a vector field @xmath36 on @xmath3 by means of hamilton s equation : @xmath37 the vector field @xmath36 is called the _ hamiltonian vector field _ induced by @xmath38 .
given any two smooth functions @xmath39 we may define their _ poisson bracket _
@xmath40 an _ _ integrable system _
_ is a triple @xmath41 where @xmath0 is a connected symplectic @xmath1-manifold and @xmath42 are smooth functions for which @xmath43 on @xmath3 , and such that the differentials @xmath44 , @xmath45 are linearly independent almost everywhere on @xmath3 .
near each point in @xmath3 there are coordinates @xmath46 in which the symplectic form @xmath33 is given by @xmath47 , and the equation @xmath43 may be written as a partial differential equation @xmath48 which is equivalent to @xmath49 ( respectively @xmath50 ) being constant along the flow lines of @xmath51 ( respectively @xmath52 ) .
the following result gives a method to attach hyperbolic singularities to an integrable system , by modifying locally the system near its focus - focus singularities .
focus - focus singularities come endowed with a hamiltonian circle action near the focus - focus singular fiber , it is the hamiltonian action generated by the @xmath49-component of the system in local normal form ( as explained in section [ els ] ) .
the image of the joint map @xmath53 is called by physicists _ the classical spectrum _ of @xmath5 ( in analogy with the semiclassical spectrum of quantum mechanics ) . throughout the paper
we assume that each focus - focus singularity is in a singular fiber in which it is the only singularity ( a system satisfying this property is often called simple " ) .
topologically , this means that the singular fiber containing a focus - focus singularity is a torus pinched precisely once ( as opposed to a multipinched torus ) .
[ gen : thm ] let @xmath54 be an integrable system on a connected symplectic @xmath1-manifold @xmath0 with only non - degenerate singularities and for which @xmath4 is a proper map .
assume that @xmath5 does not have singularities with hyperbolic blocks and that @xmath6 are all the focus - focus singularities of @xmath5 , each of which is simple .
for each subset @xmath7 , @xmath55 , there is an integrable system @xmath56 such that its classical spectrum @xmath57 contains @xmath11 loops consisting of three piecewise smooth curves of singular values .
one smooth piece corresponds to non - degenerate transversally hyperbolic singularities of @xmath58 and two pieces correspond to non - degenerate transversally elliptic singularities .
moreover the focus - focus singularities of @xmath58 are precisely @xmath13 , @xmath14 , and each @xmath59 , is non - degenerate .
the proof of theorem [ gen : thm ] uses eliasson s linearization theorem for non - degenerate singularities , and properties of the hamiltonian hopf bifurcation .
the properness of @xmath49 in theorem [ gen : thm ] means that the preimage under @xmath49 of a compact set is compact in @xmath3 .
_ structure of the paper_. in section [ sing : sec ] we review the different types of singularities an integrable system can have , state eliasson s linearization theorem , and recall the notion of degenerate singularity . in section [ hyp ]
we introduce hyperbolic semitoric systems , and prove a version of theorem [ gen : thm ] for semitoric systems that takes into account the additional symmetry given by the semitoric property . in section [ els ]
we explain the existence of a global hamiltonian circle action semiglobally near a singularity of focus - focus type , using eliasson s linearization theorem . in section [ mt : sec ]
we state the general theorem of the paper , theorem [ thm : main ] , of which theorem [ gen : thm ] is a consequence . in section
[ sec : hopf ] we briefly recall the basics about hopf bifurcations which we need for the proof of theorem [ thm : main ] . the remaining of the paper is devoted to the proof of theorem [ thm : main ] . _
acknowledgements_. we would like to thank joachim worthington for his contribution in the analysis of the example in section [ sec : examples ] . the first author is partially supported by arc grant dp110102001 .
the second author is partially supported by nsf grants dms-1055897 and dms-1518420 .
let @xmath60 be an integrable system and let @xmath61 .
we say that @xmath62 is a _ regular point _ if @xmath63 has rank @xmath34 .
a point @xmath64 is called a _ regular value _ if every point in @xmath65 is regular ; in this case , @xmath65 is called a _
regular fiber_. we say that @xmath62 is a
_ critical point _ ( or a _ singularity _ ) , if the rank of @xmath66 is @xmath67 or @xmath68 .
the fiber @xmath65 is a _ singular fiber _ if it contains one or more critical points let @xmath69 be a connected component of a regular fiber @xmath65 and assume that the hamiltonian vector fields @xmath70 are complete on @xmath65 .
then it follows from the definition of integrable system that @xmath69 is diffeomorphic to @xmath71 , @xmath72 , or @xmath73 . in the particular case that @xmath74 is compact , then @xmath70 are complete , and therefore @xmath69 is diffeomorphic to @xmath75 ; this is always true for instance if at least one of @xmath76 is a proper map . throughout this paper
we use the notion of _ non - degeneracy _ of a singularity .
suppose that @xmath32 is a connected symplectic four - manifold .
let @xmath77 be an integrable system on @xmath32 and @xmath78 a critical point of @xmath5 .
if @xmath79 , then @xmath62 is called _ non - degenerate _ if the hessians @xmath80 , @xmath81 , span a cartan subalgebra of the symplectic lie algebra of quadratic forms on the tangent space @xmath82 to @xmath3 at @xmath62 .
it follows from the work of williamson @xcite that such a cartan subalgebra has a basic building block of three types : two uni - dimensional ones ( the elliptic block @xmath83 and the real hyperbolic block : @xmath84 ) and a two - dimensional block called focus - focus : @xmath85 if @xmath86 one may assume that @xmath87 .
let @xmath88 be an embedded local @xmath34-dimensional symplectic submanifold through the point @xmath62 , such that @xmath89 and @xmath90 is transversal to the hamiltonian vector field @xmath91 defined by @xmath92 .
the critical point @xmath62 of @xmath5 is transversally _ non - degenerate _ if @xmath93 is a non - degenerate symmetric bilinear form on @xmath94 .
the notion of non - degeneracy is independent of the choice we make of @xmath95 .
the existence of @xmath95 follows from the flow box theorem ( see , eg .
* theorem 5.2.19 ) , also called the darboux - carathodory theorem ( * ? ? ? * theorem 4.1 ) ) . for the notion of non - degeneracy of a critical point in arbitrary dimension see @xcite , (
* section 3 ) . near a regular or singular value .
the transversally - hyperbolic , elliptic - hyperbolic , and hyperbolic - hyperbolic cases are not possible if @xmath5 is semitoric ( definition [ d ] ) , but they are possible if @xmath5 is hyperbolic semitoric ( definition [ def : andsish]).,width=340 ] non - degenerate critical points may be characterized ( @xcite ) using the williamson normal form @xcite .
eliasson s theorem stated below describes the local normal form around non - degenerate singularities of an integrable system ; the result holds in any dimension , but we state it here in dimension four since it is the case we are concerned with in the present paper .
the analytic version of the theorem is due to vey @xcite .
[ eliasson ] the non - degenerate critical points of an integrable system @xmath96 are linearizable .
that is , if @xmath97 is a nondegenerate critical point of the integrable system @xmath98 then there exist symplectic coordinates @xmath99 near @xmath62 , in which @xmath62 is represented as @xmath100 , such that @xmath101 , for all @xmath102 , where we have the following possibilities for each @xmath103 , @xmath104 , each of which is defined on a neighborhood of @xmath105 : * elliptic component : @xmath106 or @xmath107 . *
hyperbolic component : @xmath108 or @xmath109 .
* focus - focus component : @xmath110 and @xmath111 ( note that this component appear in `` pairs '' ) . *
non - singular component : @xmath112 or @xmath113 .
moreover if @xmath62 does not have any hyperbolic block , then the system of commuting equations @xmath101 , for all indices @xmath102 , may be replaced by the single equation @xmath114 where @xmath115 and @xmath116 is a diffeomorphism from a small neighborhood of the origin in @xmath117 into another such neighborhood , such that @xmath118 see figure [ singularities_in_image ] for a local description of the singularities appearing in theorem [ eliasson ] .
the analytic case of theorem [ eliasson ] was proved by rmann in @xcite when @xmath31 , and then in any dimension by vey @xcite .
a simple way to check eliasson non - degeneracy is as follows : [ lem : xxx ] let @xmath119 be an integrable system .
a critical point @xmath62 of @xmath5 of rank @xmath67 is non - degenerate if the hessians @xmath120 and @xmath121 are linearly independent , and there is a linear combination @xmath122 for which there are no multiple eigenvalues . here
@xmath123 is the symplectic matrix corresponding to the symplectic form @xmath33 . in particular
if @xmath124 has no multiple eigenvalues then @xmath62 is non - degenerate .
this criterion is based on the fact ( see for instance @xcite ) that a commutative subalgebra of the symplectic algebra is a cartan subalgebra if and only if it is two - dimensional and if it contains an elements whose eigenvalues are different .
theorem [ gen : thm ] may be applied in particular to enlarge the category of _ semitoric systems _ , these are systems which have an additional circular hamiltonian symmetry coming from a global hamiltonian action of the circle @xmath125 .
many integrable systems from classical mechanics ( see eg . the book by holm @xcite , or the article @xcite ) , exhibit symmetries of this nature including the lagrange top , the two - body problem , and the spherical pendulum . recall that an action of the circle @xmath125 on a symplectic manifold @xmath0 by symplectomorphisms is _ hamiltonian _ if there exists a smooth map @xmath4 , called the _ momentum map _ such that @xmath126 where @xmath127 is the vector field ( or infinitesimal generator ) of the @xmath125-action .
the category of semitoric systems includes many examples from the physics literature such as integrable systems of jaynes - cummings type , and the jaynes - cummings system . in this section we state a theorem which allows us to construct a semitoric system with hyperbolic singularities , from a semitoric system without hyperbolic singularities ; below we give the precise definitions of these notions .
the famous _ jaynes - cummings system _ @xcite is given on phase space @xmath128 by @xmath129 where @xmath130 is the unit sphere in @xmath131 with coordinates @xmath132 , and @xmath73 is equipped with coordinates @xmath133 .
we endow @xmath134 with the product symplectic structure @xmath135 where @xmath136 is the standard symplectic form on @xmath130 and @xmath137 is the standard symplectic form on @xmath73 . the @xmath49 component is the momentum map of the hamiltonian @xmath125-action that simultaneously rotates about the vertical axes of @xmath130 and the origin in @xmath138 . under the flow of @xmath49 , the points @xmath132 and @xmath139 move along the flows of @xmath140 and @xmath141 , respectively , with same angular velocity , so @xmath142 is constant and commutes with @xmath49 . the completely integrable system @xmath143 given by ( [ jc ] ) has been extensively studied , and recently attracted a lot of interest in both the physics and mathematics communities , see for instance @xcite . in (
* corollary 2.2 ) it was proved that @xmath144 is an example of a so called _ semitoric system _ , which we define in general next .
we start with the notion of a semitoric system ( ( * ? ? ?
* definition 2.1 ) ) .
[ d ] a _ semitoric system _ ( or _ semitoric integrable system _ ) is an integrable system @xmath60 such that : * @xmath49 is the momentum map for a hamiltonian @xmath125-action ; * @xmath49 is proper ; * @xmath5 has only non - degenerate singularities , without hyperbolic blocks .
if @xmath5 is a semitoric system in the sense of definition [ d ] , we have the following possibilities for the map @xmath145 in theorem [ eliasson ] , depending on the rank of the singularity : * if @xmath62 is a singularity of @xmath5 of rank zero , then the building blocks are * * @xmath146 and @xmath147 . * * @xmath148 and @xmath149 ; * if @xmath62 is a singularity of @xmath5 of rank one , then * * @xmath146 and @xmath150 .
semitoric systems as in definition [ d ] are classified in @xcite in terms of five symplectic invariants .
this classification excluded systems with hyperbolic singularities , which are however prominent in the physics literature . by eliminating conditions ( ii ) and ( iii ) in definition [ d ] one would expand significantly the class of integrable systems which are semitoric " in spirit , that is , those for which the component @xmath49 still generates a hamiltonian @xmath125-action .
for the purpose of this paper , we keep item ( ii ) due technical reasons is proper as a map into @xmath138 , as in @xcite , but this still requires substantial work . ] , and show that there are many interesting examples of systems which satisfy the assumptions ( i ) and ( ii ) but not ( iii ) ; precisely , we will be concerned with _
hyperbolic semitoric systems_. [ def : andsish ] a _ hyperbolic semitoric system _ is an integrable system @xmath151 for which : * @xmath49 is the momentum map for a hamiltonian @xmath125-action ; * @xmath49 is proper ; * the set of hyperbolic singularities of @xmath5 is non - empty ; * the set of degenerate singularities @xmath5 is isolated .
our next goal is to construct hyperbolic semitoric systems from semitoric systems .
the main theorem of this paper ( theorem [ thm : main ] ) implies that any semitoric system in the more strict sense of definition [ d ] may be suitably modified to create a new semitoric " system in the more relaxed sense of definition [ def : andsish ] .
that is , we have the following result . [ mt2 ]
given a semitoric system @xmath152 with focus - focus singularities @xmath153 and an integer @xmath154 , there exists a smooth hamiltonian function @xmath155 such that @xmath156 is a hyperbolic semitoric system with non - degenerate focus - focus singularities at @xmath157 .
the new aspect of this theorem is that it says that the global hamiltonian @xmath125-action of the original semitoric system may be preserved , in addition to the other information about the old and newly created singularities given in theorem [ gen : thm ] .
the proof in section [ xx ] is based on the knowledge of eliasson s normal form near a focus - focus singular point . here
we quickly review the ingredients we use .
[ mm ] a smooth map @xmath158 is a _ momentum map _ on @xmath159 if @xmath160 is surjective almost everywhere in @xmath159 and @xmath161 .
[ ll ] a _ singular liouville foliation _ @xmath162 is a union of connected subsets of @xmath3 , called the _ leaves _ of the foliation , which are pairwise disjoint and such that there is a momentum map @xmath163 in the sense of definition [ mm ] such that the leaves of @xmath162 coincide with the connected components of the fibers @xmath65 for @xmath164 varying in an open subset of @xmath138 .
let @xmath165 be an integrable system on a connected symplectic @xmath1-manifold @xmath0 ( semitoric or not ) .
let @xmath162 be the singular liouville foliation of @xmath3 associated to @xmath5 , as in definition [ ll ] .
the leaves of this foliation are the connected components of the fibers @xmath65 .
let @xmath62 be a critical point of focus - focus type .
for simplicity suppose that @xmath166 and that the fiber @xmath167 contains no critical point other than @xmath62 .
it is well known that @xmath168 is a pinched torus " ( that is , an immersion of the sphere @xmath130 with a transversal double point ) .
all the fibers of @xmath5 in a neighborhood of the fiber @xmath168 are @xmath34-tori . by theorem
[ eliasson ] there exist symplectic coordinates @xmath169 in a neighborhood @xmath170 about the focus - focus point @xmath62 in which if @xmath171 then @xmath172 is a momentum map for @xmath162 ; here the critical point @xmath62 corresponds to @xmath173 . near @xmath62 ,
the hamiltonian flow of @xmath174 is periodic , and assuming v to be invariant with respect to this flow the associated @xmath125-action is free in @xmath175 .
that is , _ we always have existence of a hamiltonian action of the circle @xmath125 that commutes with the flow of the system semiglobally near a focus - focus singularity _ ( i.e. in a neighborhood of the singular fiber that contains the focus - focus singular point @xmath62 ) .
this is a property of focus - focus singularities which is essential for theorem [ mt2 ] and our upcoming theorem [ thm : main ] .
fix a point @xmath176 .
let @xmath177 denote a small 2-dimensional manifold transversal to @xmath162 at @xmath178 . since @xmath162 in a neighborhood of @xmath177 is regular for @xmath5 and @xmath145 simultaneously
, there exists a diffeomorphism @xmath179 from a neighborhood @xmath159 of @xmath180 into a neighborhood of @xmath181 in @xmath73 such that @xmath182 .
thus there exists a smooth momentum map @xmath183 ( in the sense of definition [ mm ] ) for @xmath162 , defined on a neighborhood @xmath184 of @xmath168 , which agrees with @xmath172 on @xmath170 .
write @xmath185 and @xmath186 ( notice that @xmath187 ) .
it follows from equations ( [ equ : cartan1 ] ) and ( [ equ : cartan2 ] ) that near @xmath62 the orbits corresponding to the hamiltonian @xmath188 are periodic . on the other hand ,
the vector field @xmath189 is hyperbolic and it has a local stable manifold in the @xmath190-plane transversal to its local unstable manifold in the @xmath191-plane .
moreover , the vector field @xmath189 is radial in the sense that the flows approaching the origin do not spiral on the local ( un)stable manifolds ( for further details see ( * ? ? ?
* section 3 ) and ( * ? ? ?
* section 5.2.3 ) ) .
for an integrable deformation of the spin - oscillator with @xmath192 , illustrating theorem [ thm : main ] .
see section [ jc : sec ] and section [ sec : examples].,width=302 ] note that , by definition , a semitoric system can never be hyperbolic semitoric , or conversely .
the main theorem of this paper is the following result about attaching hyperbolic singularities to an integrable system with non - degenerate focus - focus singularities @xmath6 , by modifying the system locally near a certain subset @xmath193 of the set of focus - focus singularities . by relabeling
if needed we may assume that @xmath194 for some @xmath195 .
[ thm : main ] let @xmath196 be an integrable hamiltonian system where @xmath197 is a proper map .
assume that among the singularities of @xmath5 are @xmath198 simple non - degenerate focus - focus singularities @xmath199 , where @xmath200 .
then for any @xmath154 there exists an integrable system @xmath201 with the following properties : 1 .
@xmath202 , @xmath203 , and @xmath204 ; 2 .
@xmath12 has non - degenerate singularities at @xmath205 of focus - focus type ; 3 .
@xmath12 has non - degenerate singularities at @xmath206 of elliptic - elliptic type ; 4 . there
exist @xmath207 closed piecewise smooth curves with corners in the classical spectrum @xmath10 , each of which consists of three smooth curves @xmath208 \to \widetilde{f}(m),\,\,\,\,\ , j=1,2,3,\,\,\,\,\,\ , k+1
{ \leqslant}i { \leqslant}m_f\ ] ] such that for each fixed @xmath209 the images of the interior @xmath210 under each of @xmath211 and @xmath212 consists of non - degenerate transversally elliptic singular values , and under @xmath213 it consists of non - degenerate transversally hyperbolic singular values . the two endpoints where the elliptic and hyperbolic families meet non transversally are degenerate singularities .
the third endpoint is where the two elliptic families meet transversally at the non - degenerate elliptic - elliptic singularity @xmath13 ( see figure [ fig : triangle ] ) .
5 . if @xmath214 is semitoric ( see definition [ d ] ) then @xmath215 is hyperbolic semitoric ( see definition [ def : andsish ] ) .
we will prove theorem [ thm : main ] in section [ xx ] .
[ hc ] _ are there are hyperbolic semitoric systems @xmath196 for which the set of degenerate singularities is empty ? _
the proof of theorem [ thm : main ] in section [ xx ] gives evidence that the answer to the question is probably going to be _ no _ , at least under some fairly general conditions .
the answer is no in all the physical examples we are aware of .
that is , it is quite possible that the existence of hyperbolic singularities satisfying ( a),(b ) and ( c.1 ) in definition [ def : andsish ] forces the existence of at least some degenerate singularities in item ( c.2 ) , at the points where the various non - degenerate families of singularities under consideration meet with each other .
if this were not the case , there would be loops of singularities , which we can not rule out at the time being ( it would be crucial to understand whether the existence of the global @xmath125-action may not be compatible with the existence of such a loop ) .
another way of formulating the theorem is to say that for a non - degenerate focus - focus equilibrium point in an integrable system it is possible to locally modify the hamiltonian so that a hamiltonian hopf bifurcation is induced .
the main point is that this modification can be achieved without destroying the integrability of the system .
here we are going to present a brief review of some background on the hopf bifurcation .
the original references are @xcite .
a hamiltonian hopf bifurcation in the linear approximation occurs when pure imaginary eigenvalues of an equilibrium collide and move into the complex plane . at the collision
point the linearization is not diagonalizable . at the bifurcation point
the hamiltonian may be put into sokolskii s normal form @xmath216 up to flat terms where @xmath217 contains cubic and higher order terms . here
the abbreviations @xmath218 are used . adding the bifurcation parameter @xmath219 gives the unfolding @xmath220 this is the hamiltonian hopf bifurcation normal form , where @xmath221 , @xmath222 , @xmath223 , @xmath224 .
there are two different cases , depending on the sign of @xmath225 .
the cases @xmath226 ( respectively @xmath227 ) are called the subcritical ( respectively supercritical ) hamiltonian hopf bifurcation .
when the elliptic - elliptic equilibrium point looses stability in the subcritical case there is a family of stable periodic orbits near the unstable focus - focus equilibrium point . in the supercritical case
there is no such family .
up to flat terms the hamiltonian hopf bifurcation normal form @xmath228 is a family of integrable system @xmath229 to find the image of @xmath230 and its critical values introduce symplectic polar coordinates @xmath231 so that @xmath232 in these coordinates @xmath233 @xmath234 and @xmath235 hence define a reduced hamiltonian with @xmath236 as a parameter and canonical variables @xmath237 as @xmath238 solving this equation for @xmath239 gives the reduced action @xmath240 and the discriminant of the polynomial @xmath241 contains the set of critical values of @xmath242 . for more details on this derivation using singular reduction instead of polar coordinates see @xcite .
since @xmath243 only when the double root is non - negative does the corresponding part of the discriminant surface of @xmath244 belong to the set of critical values of @xmath230 .
we are going to deform the hamiltonian @xmath50 to a new hamiltonian @xmath245 in a neighborhood of each focus - focus singularity .
the deformation vanishes outside a sufficiently small neighborhood .
we will show that we can choose a deformation which turns a focus - focus point into an elliptic - elliptic point with the properties described in ( 4 ) .
this can be done independently for each focus - focus point , and hence we can create @xmath246 elliptic - elliptic points .
in the construction we use eliasson s coordinates near each focus - focus point , and we preserve the @xmath125 action which always exist near a focus - focus point , see section [ els ] . if we are in the semitoric setting , then one of eliasson s integrals _ is _ the global @xmath125 action @xmath49 , and hence the global hamiltonian @xmath125 action is preserved , establishing ( 5 ) . to establish ( 3 ) and ( 4 ) for a single equilibrium point there are three steps .
[ [ step-1 ] ] _ step 1 _ + + + + + + + + ( a new integrable system ) let @xmath247 be the eliasson coordinates given by theorem [ eliasson ] near @xmath62 of the original semitoric system and let the hamiltonian of the original system in these coordinates be @xmath50 .
then there exists a smooth function @xmath248 such that @xmath249 with @xmath250 where @xmath251 and @xmath252 now @xmath253 all have vanishing poisson bracket with @xmath254 .
thus @xmath255 and @xmath256 is an integrable system .
[ [ step-2 ] ] _ step 2 _ + + + + + + + + ( type of equilibrium ) the quadratic part of the original semitoric integrable system in eliasson coordinates near the focus - focus singularity is @xmath257 with given real parameters @xmath258 and @xmath259 . .
the discriminant of @xmath260 is marked in bold ( red ) lines , delineating the regions hyperbolic - hyperbolic ( hh ) with 4 real roots , elliptic - elliptic ( ee ) with 4 pure imaginary roots , elliptic - hyperbolic ( eh ) with 2 real and 2 pure imaginary roots , and focus - focus ( ff ) with a complex quadruplet . for eigenvalues coming from @xmath261 the grey region of hyperbolic - hyperbolic and elliptic - hyperbolic is not accessible.,width=226 ] the quadratic part of the modified integrable system is @xmath262 with real parameters @xmath263 .
we can choose these parameters by choosing the function @xmath264 from step 1 .
we are now going to show that by choosing these parameters we can make @xmath261 have a non - degenerate elliptic - elliptic equilibrium point at the origin .
this does not seem to help in creating hyperbolic singularities , but we will see in step 3 that by adding appropriate higher order terms in @xmath264 we can create hyperbolic - transverse singularities . except for cases with multiple roots
the type of equilibrium of @xmath261 is determined by the characteristic polynomial @xmath265 where @xmath266 and @xmath267 the discriminant of @xmath260 is @xmath268 the line @xmath269 and the parabola @xmath270 divide the plane into four regions , see figure [ fig : ffee ] .
when @xmath271 and @xmath272 there are four distinct pure imaginary eigenvalues @xmath273 now we verify that by a choice of @xmath274 we can satisfy these conditions .
note that @xmath275 and by choosing @xmath276 this can be made positive .
now @xmath276 also implies @xmath272 .
finally we can achieve @xmath277 by increasing @xmath278 if necessary .
since the eigenvalues of the linearization of the hamiltonian vectorfield of @xmath279 are distinct the singularity at the origin is non - degenerate , according to lemma [ lem : xxx ] .
the construction may be done independently at each focus - focus point , and thus we established part ( 3 ) of the theorem . because of the way @xmath280 and @xmath281 depend on @xmath274 it is not possible to have @xmath282 nor is it possible to have @xmath283 and @xmath284 .
thus the idea to directly choose the parameters @xmath274 so that @xmath261 has a singularity of hyperbolic - hyperbolic type or elliptic - hyperbolic type does not work .
it is easy to see that when @xmath270 and @xmath272 and in addition @xmath285 the eigenvalues are pure imaginary , the matrix @xmath286 is not semi - simple , and the singularity is degenerate .
it is possible to choose a smooth family @xmath287 with parameter @xmath288 $ ] such that @xmath289 , @xmath290 has a non - degenerate focus - focus singularity for @xmath291 $ ] , @xmath290 has a non - degenerate elliptic - elliptic singularity for @xmath292 $ ] , and only @xmath293 is degenerate .
[ [ step-3 ] ] _ step 3 _ + + + + + + + + ( higher order terms ) we are now going to determine higher order terms in @xmath245 that will generate a family of hyperbolic - transverse singularities .
this will be done by transforming a special case of the general hopf normal form @xmath228 into @xmath245 .
we are going to show that the image of @xmath294 has the form schematically shown in figure [ fig : triangle ] as described precisely in the statement part ( 4 ) of our main theorem .
after the hopf bifurcation .
elliptic - elliptic is the non - degenerate elliptic equilibrium point , e denotes the two transversally elliptic non - degenerate families of singularities ( stable isolated periodic orbits ) attached to the equilibrium , h denotes the transversally hyperbolic non - degenerate families of singularities ( unstable isolated periodic orbits ) .
nt denotes that there are n tori in the preimage of the corresponding regular values .
the cusps where the e and h families meet are degenerate singularities ( saddle - centre bifurcation of periodic orbits ) .
, width=226 ] to obtain the hopf normal form we need to introduce a new set of local coordinates related to eliasson s coordinates by a linear symplectic transformation .
we now choose a parameter @xmath295 such that @xmath296 and set @xmath297 the transformation is @xmath298 where @xmath299 the transformation @xmath300 has the property that it maps the quadratic part of @xmath228 into @xmath261 .
specifically it transforms the function @xmath301 as defined in section [ sec : hopf ] as follows : @xmath302 e.g. the first identity for @xmath303 follows from @xmath304 we now specialize the hopf normal form @xmath228 discussed in section [ sec : hopf ] to @xmath305 because that is sufficient for our purpose . transforming the hopf normal form ( given in variables with hat ) and with sufficiently small @xmath219 @xmath306 into variables without hat gives @xmath307 thus the quadratic part of @xmath245 has parameters in the notation of step 2 given by @xmath308 and @xmath309 now we choose the function @xmath264 by setting @xmath310 so that @xmath311 note that @xmath264 is not only producing the desired @xmath245 , but it also annihilates any unwanted higher order terms which may be present in @xmath50 .
recall that @xmath312 and @xmath313 , and therefore the parameters in step 2 are @xmath314 and they put @xmath245 in the elliptic - elliptic region as long as @xmath219 is positive .
this may be seen by computing the eigenvalues of the linearized vector field of @xmath228 at the origin : * when @xmath315 the equilibrium is of type focus - focus with eigenvalues @xmath316 * when @xmath317 it is of type elliptic - elliptic with eigenvalues @xmath318 hence for sufficiently small and positive @xmath219 the equilibrium is of elliptic - elliptic type and it is non - degenerate because the eigenvalues are distinct . for the subcritical hopf bifurcation with @xmath226 .
critical values of @xmath230 are shown in red .
in addition , the vanishing of the discriminant of @xmath244 is shown in blue .
the parameter values are @xmath319 , @xmath320 , and @xmath321 or @xmath322.,width=340 ] for the supercritical hopf bifurcation with @xmath227 .
critical values of @xmath230 are shown in red .
in addition , the vanishing of the discriminant of @xmath244 is shown in blue .
the image is to the left for @xmath323 and to the right for @xmath324 .
the parameter values are @xmath319 , @xmath320 , and @xmath321 or @xmath322.,width=340 ] we are now going to establish the properties of the image of @xmath294 claimed in ( 4 ) of the main theorem .
see figure [ fig : hopf1 ] and [ fig : hopf2 ] for an illustration of the various cases .
first of all notice that @xmath294 and @xmath230 are related by a linear symplectic transformation , so we may as well study the image of @xmath230 . as described in section [ sec
: hopf ] the critical values of @xmath325 are contained in the discriminant surface of the cubic polynomial @xmath326 the discriminant of this polynomial has a rational parametrization that can be found by the ansatz @xmath327 and solving for the double root @xmath328 , @xmath49 and @xmath329 .
the result is @xmath330 where the double roots of @xmath244 occur at @xmath331 assume that @xmath317 .
three piecewise smooth curves of critical values are found for @xmath332 $ ] , for @xmath333 $ ] , and @xmath334 $ ] . both derivatives @xmath335
vanishes at @xmath336 only .
we call these corresponding critical values the cusp
. the other segment endpoints @xmath337 map to the same critical value @xmath338 .
we now establish the type and the ( non-)degeneracy of the critical values on the three curves .
we need to compute the determinant of the hessian of the reduced hamiltonian @xmath339 from section [ sec : hopf ] at the critical points .
the determinant of the hessian at an arbitrary point is @xmath340 and evaluating this at @xmath341 , @xmath342 , and @xmath343 gives @xmath344 which is non - zero unless @xmath336 , corresponding to the cusps .
moreover for @xmath345 the determinant is negative , so this segment is transversally - hyperbolic .
similarly , the other two segments @xmath346 and @xmath347 lead to a positive determinant and are hence transversally - elliptic . finally at the cusps
the determinant of the hessian vanishes and hence these are degenerate critical values .
the three curves are graphs over @xmath49 because @xmath348 on the interior of the three segments , and when @xmath349 then also @xmath350 creating the two cusps .
when the two curves of transversally elliptic values intersect at the origin they have slopes @xmath351 and hence the intersection is transversal for @xmath317 .
not all points in the discriminant surface of @xmath352 are actually critical values of @xmath242 . to be in the image of @xmath230 we need @xmath353 .
the double root @xmath354 is non - negative for @xmath355 $ ] only when @xmath226 .
hence we are now choosing @xmath356 such that @xmath226 .
accordingly we have one of the two right situations shown in figure [ fig : hopf1 ] . a particular point on the transversally - hyperbolic curve
is obtained from @xmath357 where @xmath358 hence the sign of @xmath356 determines whether this critical value is above or below the equilibrium point at the origin .
the cusps corresponding to @xmath336 are created by a saddle - centre bifurcation of an elliptic and a hyperbolic periodic orbit .
this is one of the typical degenerate singularities described in @xcite . for a detailed analysis of the universal features of the dynamics near this bifurcation using an integrable model
see @xcite . as in step 2
, we can think of the function @xmath264 as a one - parameter family that smoothly deforms the original hamiltonian into the new hamiltonian @xmath245 . from this point of view
it makes sense to consider @xmath315 as well , and the corresponding sets of critical values of @xmath230 are shown on the left column of figure [ fig : hopf1 ] and [ fig : hopf2 ] .
the main result of this paper is stated in the form of several theorems , but in fact it is more a method than a result .
the following theorem has a similar proof .
[ thm : noff ] each focus - focus singular point in a semitoric integrable system @xmath359 may be continuously deformed via a path of semitoric systems into an elliptic - elliptic point on the boundary of the image of the new integrable system @xmath360 constructed in section [ xx ] , either at the top boundary ( maximal @xmath245 for fixed @xmath49 ) or at the bottom boundary ( minimal @xmath245 for fixed @xmath49 ) .
this may be done simultaneously for any subset of the set of focus - focus singular points of the system .
the proof is as above , the only difference is that now we choose @xmath264 with @xmath227 .
the corresponding critical values are shown in the two right cases in figure [ fig : hopf2 ] .
the sign of @xmath356 determines which one of the two cases occurs . under
what conditions the resulting semitoric system is in fact toric is an open question .
a deformation of the spherical pendulum that models floppy triatomic molecules ( eg .
hcn ) has been studied in @xcite . strictly speaking this is not an example for out theorem since the momentum map @xmath49 of the spherical pendulum is not proper .
but if @xmath49 is not proper the same idea still works as long as @xmath230 is proper , which can be easily checked in this particular example .
in fact , our theorem can be generalised to the situation where @xmath49 is not proper by adding higher order terms to @xmath264 that make @xmath245 proper . here
we give a brief description of a deformation of the semitoric coupled spin - oscillator @xcite ( jaynes - cummings system ) that makes it pass through a subcritical hopf bifurcation with @xmath227 .
a family for which the focus - focus point is driven into the boundary in a supercritical hopf bifurcation with @xmath226 has been presented in @xcite .
the image of the momentum map of the jaynes - cummings is shown in figure [ fig : jc ] in particular examples it may be simpler to avoid the eliasson normal form , since it may not be easy to find .
instead we start with @xmath362 in the original variables @xmath363 where @xmath364 and @xmath365 then we ask for a function @xmath264 that commutes with @xmath49 , and define @xmath366 in an example it will also typically be nicer to work with a global function @xmath264 , instead of one that is only defined in a neighborhood of the equilibrium under consideration . for the spin - oscillator any function of the form @xmath367 works .
we chose @xmath368 , and the characteristic polynomial of the linearization at the equilibrium point @xmath369 is given by @xmath370 and undergoes a hopf bifurcation when @xmath371 . now set @xmath372 in the @xmath373 plane
we obtain the horizontal line @xmath374 for @xmath375 the singularity is in the elliptic - elliptic region .
the non - linear analysis of this example is somewhat involved and will be presented in detail in a forthcoming paper . here
we simply present the image of @xmath230 for @xmath376 , see figure [ fig : spinoschopf ] .
abraham , r. and marsden , j.e . : _ foundation of mechanics _ , second edition , revised and enlarged .
with the assistance of tudor ratiu and richard cushman .
benjamin / cummings publishing co. , inc .
, advanced book program , reading , mass .
, 1978 .
holger r. dullin + school of mathematics and statistics + university of sydney + sydney , nsw 2006 , australia + _ e - mail _ : ` [email protected] ` + + lvaro pelayo + university of california , san diego + mathematics department + 9500 gilman dr # 0112 + la jolla , ca 92093 - 0112 , usa .
+ _ e - mail _ : ` [email protected] ` | let @xmath0 be a connected symplectic @xmath1-manifold and let @xmath2 be a completely integrable system on @xmath3 with only non - degenerate singularities and for which @xmath4 is a proper map .
assume that @xmath5 does not have singularities with hyperbolic blocks and that @xmath6 are the focus - focus singularities of @xmath5 .
for each subset @xmath7 we will show how to modify @xmath5 locally around any @xmath8 , in order to create a new integrable system @xmath9 such that its classical spectrum @xmath10 contains @xmath11 smooth curves of singular values corresponding to non - degenerate transversally hyperbolic singularities of @xmath12 .
moreover the focus - focus singularities of @xmath12 are precisely @xmath13 , @xmath14 , and each of these @xmath13 is non - degenerate . the proof is based on eliasson s linearization theorem for non - degenerate singularities , and properties of the hamiltonian hopf bifurcation . |
the @xmath0 , a @xmath42 state in the charmonium family , was found in the inclusive photon spectra from @xmath4 and @xmath43 @xcite decays , as well as in hadronic decays @xcite .
a number of decay modes of @xmath0 were then measured @xcite .
more recent measurements of hadronic decays of @xmath44 are listed in ref .
@xcite . according to ref .
@xcite , the @xmath44 is expected to have numerous decay modes into hadronic final states .
although a number of decay modes of @xmath44 have been measured by different experimental collaborations , the number of measured @xmath0 decay channels are few .
this means that many decay modes of @xmath44 are unknown .
the 58 million , @xmath45 @xcite , @xmath4 events taken at besii provide a chance to observe new decays . in this analysis , @xmath44
decaying into @xmath46 and @xmath47 are studied using @xmath48 .
the upgraded beijing spectrometer detector , located at the beijing electron - positron collider ( bepc ) , is a large solid - angle magnetic spectrometer which is described in detail in ref .
the momentum of the charged particle is determined by a 40-layer cylindrical main drift chamber ( mdc ) which has a momentum resolution of @xmath49/p=@xmath50 ( @xmath51 in gev / c ) .
particle identification is accomplished by specific ionization ( @xmath52 ) measurements in the drift chamber and time - of - flight ( tof ) information in a barrel - like array of 48 scintillation counters .
the @xmath52 resolution is @xmath53 ; the tof resolution for bhabha events is @xmath54 ps . radially outside of the time - of - flight counters is a 12-radiation - length barrel shower counter ( bsc ) comprised of gas tubes interleaved with lead sheets .
the bsc measures the energy and direction of photons with resolutions of @xmath55 ( @xmath56 in gev ) , @xmath57 mrad , and @xmath58 cm .
the iron flux return of the magnet is instrumentd with three double layers of counters that are used to identify muons . a geant3 based monte carlo package ( simbes ) with detailed consideration of the detector performance
is used to obtain the detection efficiency .
the consistency between data and monte carlo has been carefully checked in many high purity physics channels , and the agreement is reasonable @xcite .
these events are observed in the topology @xmath60 .
events with six good charged tracks and at least one isolated photon are selected .
the selection criteria for good charged tracks and isolated photons are described in detail in ref .
. each charged track must be well fitted to a helix , originating from the interaction region of r@xmath61 @xmath62 2 cm and @xmath63 20 cm , and have a polar angle @xmath64 in the range @xmath65 @xmath62 0.8 . here
r@xmath61 is the distance from the beam axis , and @xmath66 is along the beam axis .
isolated photons are those that have energy deposited in the bsc greater than 60 mev , the angle between the direction at the first layer of the bsc and the developing direction of the cluster less than 30@xmath67 , and the angle between photons and any charged tracks larger than @xmath68 .
two of the charged tracks should be identified as kaons by combined tof and de / dx information . a four - constraint ( 4c )
kinematic fit is performed under the hypothesis of @xmath69 , and the @xmath70 is required to be less than 10 . to reject background from @xmath71
, @xmath72 is required to be less than @xmath73 .
background events from @xmath74 are eliminated by requiring @xmath75 @xmath76 and @xmath77 mev / c , where @xmath78 is the missing momentum of charged tracks . for selected events .
the histogram with error bars is from data , the shaded part is the background estimated from @xmath79 monte carlo simulation , and the curve represents the fitting results described in the text.,scaledwidth=50.0% ] after the above selection , the @xmath80 invariant mass , @xmath81 , distribution is shown in fig .
[ metackk ] .
a peak at the @xmath44 mass is observed .
the shaded histogram is the background estimated from 58 million @xmath82 monte - carlo events generated with the lund - charm generator @xcite ; no prominent signal in the @xmath44 mass region is seen . also , 100,000 events for the two possible background channels @xmath83 and @xmath84 are simulated . after final selection ,
no events remain in the @xmath44 mass region . a breit - wigner folded with a gaussian to take into account the mass resolution of 12.3 mev / c@xmath85 at the @xmath0 and a polynomial background
are used in the fit .
the fit gives @xmath86 @xmath0 events with a statistical significance of 4.0 @xmath87 , where the mass and width of @xmath0 are fixed to the pdg values @xcite . using this sample , we search for the decay mode @xmath88 . to select @xmath89 events , we require that the invariant masses of @xmath90 and @xmath91 must statisfy @xmath92 gev/@xmath93 .
after the @xmath94 and @xmath95 selection , the @xmath1 invariant mass is shown in fig .
[ metac2kst ] .
a small peak at the @xmath44 mass is observed .
the background events corresponding to the shaded histogram in fig .
[ metac2kst ] are estimated from @xmath94 and @xmath95 sidebands ( 0.1 gev / c@xmath85 @xmath96 gev/@xmath97 and 0.1 gev / c@xmath85 @xmath98 gev/@xmath97 ) , and there is no evident @xmath44 signal .
@xmath99 events are obtained by fitting the mass spectrum with a breit - wigner folded with a gaussian to account for the @xmath44 mass resolution plus a second polynomial background .
the corresponding mass and width of the @xmath0 are fixed to pdg values @xcite .
since the significance of the peak is only 3@xmath87 , we also give the upper limit for @xmath88 . with the bayes method ,
the fit of this distribution yields 65 events at the 90% confidence level . for @xmath88 candidate events .
the histogram with error bars is for data , the shaded part is the background estimated from @xmath94(@xmath95 ) sidebands , and the curve is the fitting results described in the text.,scaledwidth=50.0% ] the @xmath100 sample can also be used to search for @xmath101 . for selecting a @xmath102 signal ,
the @xmath103 mass , @xmath104 , is required to be in the region @xmath105 gev / c@xmath85 .
after this selection , no clear @xmath0 signal is found in the distribution of @xmath106 , as shown in fig .
[ phi4pi ] . using bayes method , a fit to @xmath107 with a breit - wigner folded with a gaussian and a polynomial background gives 13.5 @xmath0 events at the 90% confidence level . from monte - carlo simulation , in which the angle ( @xmath64 ) between the direction of the @xmath108 and @xmath44 in the laboratory frame is generated according to a @xmath109 distribution and uniform phase - space
is used for @xmath44 decaying into @xmath1 and @xmath110 , the detection efficiencies of @xmath111 , @xmath112 , and @xmath113 are determined as @xmath114 , @xmath115 , and @xmath116 , respectively
. therefore , the branching fractions obtained are @xmath117 , @xmath118 + @xmath119 , and @xmath120 . for @xmath101 candidate events.,scaledwidth=50.0% ]
these events are observed in the topology @xmath122 .
events with six good charged tracks and at least one isolated photon are selected . no particle identification is required . to suppress background ,
a 4c kinematic fit is performed under the hypothesis @xmath123 , and the @xmath124 is required to be less than 10 . to reject background from @xmath125 and @xmath126 , we require @xmath127 to be less than @xmath128 and @xmath129 . for selected events .
the histogram with error bars is data , the shaded part is the background estimated from monte carlo simulation , and the curve is the fitting result described in the text.,scaledwidth=50.0% ] figure [ etac1 ] shows the @xmath47 invariant mass spectrum after the above selection .
a clear @xmath44 peak is observed .
the shaded histogram in fig .
[ etac1 ] corresponds to background estimated from 58 million @xmath82 monte - carlo events generated using the lund - charm generator @xcite , where no @xmath44 signal is evident .
a fit of the @xmath130 distribution , which is shown as the solid curve in fig .
[ etac1 ] , using a breit - wigner convoluted with a gaussian to represent the signal and a polynomial for the background , yields @xmath131 @xmath44 events with a statistical significance of @xmath132 . in the fit , the mass and width of @xmath0
are again fixed to pdg values @xcite .
the detection efficiency for @xmath133 is determined to be @xmath134 , by monte - carlo simulation with the distribution of @xmath64 , the angle between the directions of @xmath108 and @xmath44 in the laboratory frame , being generated with a @xmath109 and with @xmath44 decaying into @xmath2 being generated with a uniform phase - space distribution .
the branching ratio is then found to be @xmath135 .
the systematic errors mainly come from the following sources : ( 1 ) mdc tracking efficiency + this has been measured with clean channels like @xmath136 and @xmath137 .
it is found that the monte carlo simulation agrees with data within 1 - 2% for each charged track .
therefore , 12% is conservatively taken as the systematic error in the tracking efficiencies for the 6-prong final states analyzed here .
( 2 ) photon detection efficiency + this has been studied using different methods with @xmath138 events @xcite . the difference between data and monte carlo simulation is less than 2% for each photon , and 2% is taken as the systematic error for the photon efficiency in this analysis .
( 3 ) particle identification ( pid ) + this has been studied with @xmath139 .
the efficiency of the pid from data is consistent with that from monte carlo simulation .
the average difference is less than 2% . for @xmath100 decay , 4%
is taken as the systematic error from pid .
( 4 ) kinematic fit + the kinematic fit is useful to reduce background . using the same method for estimating the systematic error as in ref .
@xcite , the decay mode @xmath140 is also analyzed .
the efficiency difference of the kinematic fit for data and monte carlo is 7.7% .
since the decay of @xmath140 is similar to the two channels analyzed in this paper , 7.7% is also taken here as the systematic error of the kinematic fit .
( 5 ) @xmath0 parameters + although the @xmath44 signal is clear , the number of events is not large enough to determine the breit - wigner parameters and the background shape well .
the variation of the fit solution due to changes of the @xmath0 mass and width corresponding to the uncertainties in the pdg , as well as changes in the fitting mass region used , is taken as a systematic error and listed in table [ syserr ] .
( 6 ) background + for @xmath141 , the biggest background comes from @xmath142 .
when the invariant mass of @xmath143 is required to be within the @xmath144 mass region ( @xmath145 gev / c@xmath15 ) , five events remain in the @xmath44 mass region .
if all of them are regarded as signal from @xmath142 , the background from this decay mode is about 5.1% , and this is taken as the systematic error associated with background for this channel .
no events remain for @xmath88 and the upper limit is 2.3 events at 90% confidence level . then the uncertainty caused by @xmath146 is 5.1% . for the @xmath147 , monte carlo simulation is used to estimate the background from @xmath146 . using the branching fraction for @xmath148 , obtained from @xmath149 @xcite , monte carlo simulation indicates that 33 background events contribute to the @xmath44 signal .
compared to the 416 signal events from fitting the mass spectrum , the background fraction is 7.9% which is taken as the background systematic error for this channel . ( 7 ) number of @xmath4 events + the number of @xmath4 events is @xmath150 , determined from @xmath4 inclusive four - prong events .
the uncertainty is taken as a systematic error in the branching ratio measurement .
table [ syserr ] lists the systematic errors from all sources , and the total systematic error is the sum of them added in quadrature . the bes collaboration thanks the staff of bepc and the computing center for their hard efforts .
this work is supported in part by the national natural science foundation of china under contracts nos .
19991480 , 10225524 , 10225525 , the chinese academy of sciences under contract no .
kj 95t-03 , the 100 talents program of cas under contract nos .
u-11 , u-24 , u-25 , and the knowledge innovation project of cas under contract nos .
u-602 , u-34 ( ihep ) ; and by the national natural science foundation of china under contract no.10175060 ( ustc ) , and no .
10225522 ( tsinghua university ) ; and by the u. s. department of energy under contract n0 .
de - fg02 - 04er41291 .
p. partridge et al .
( 1980 ) 1150 .
t. himel et al.m , phys .
( 1980 ) 1146 .
r. m. baltrusaitis et al .
rev.d33 ( 1986 ) 629 ; d. bisello et al.,phys .
lett . 179b ( 1986 ) 294
; d. bisello et al . ,
phys . lett .
192b ( 1987 ) 294 ; j. z. bai et al .
d62 ( 2000 ) 072001 .
bai et al . , phys .
b555 ( 2003 ) 174 ; j.z .
bai et al . , phys .
b578 ( 2004)16 ; h .- c huang et al .
, hep - ex/0305068 ; f. fang et al . ,
( 2003 ) 071801 ; b. aubert et al .
d70 ( 2004 ) 011101 .
c. quigg and j. l. rosner , phys .
d16 ( 1977 ) 1497 .
s. s. fang et al . , high energy phys .
27 ( 2003 ) 277(in chinese ) . | the decays of @xmath0 to @xmath1 and @xmath2 are observed for the first time using a sample of @xmath3 @xmath4 events collected by the besii detector .
the product branching fractions are determined to be @xmath5 @xmath6 , @xmath7 , and @xmath8 .
the upper limit for @xmath9 is also obtained as @xmath10 at the 90% confidence level . _
pacs : _
13.25.gv , 14.40.gx , 13.40.hq m. ablikim@xmath11 , j. z. bai@xmath11 , y. ban@xmath12 , j. g. bian@xmath11 , x. cai@xmath11 , j. f. chang@xmath11 , h. f. chen@xmath13 , h. s. chen@xmath11 , h. x. chen@xmath11 , j. c. chen@xmath11 , jin chen@xmath11 , jun chen@xmath14 , m. l. chen@xmath11 , y. b. chen@xmath11 , s. p. chi@xmath15 , y. p. chu@xmath11 , x. z. cui@xmath11 , h. l. dai@xmath11 , y. s. dai@xmath16 , z. y. deng@xmath11 , l. y. dong@xmath11@xmath17 , q. f. dong@xmath18 , s. x. du@xmath11 , z. z. du@xmath11 , j. fang@xmath11 , s. s. fang@xmath15 , c. d. fu@xmath11 , h. y. fu@xmath11 , c. s. gao@xmath11 , y. n. gao@xmath18 , m. y. gong@xmath11 , w. x. gong@xmath11 , s. d. gu@xmath11 , y. n. guo@xmath11 , y. q. guo@xmath11 , z. j. guo@xmath19 , f. a. harris@xmath19 , k. l. he@xmath11 , m. he@xmath20 , x. he@xmath11 , y. k. heng@xmath11 , h. m. hu@xmath11 , t. hu@xmath11 , g. s. huang@xmath11@xmath21 , x. p. huang@xmath11 , x. t. huang@xmath20 , x. b. ji@xmath11 , c. h. jiang@xmath11 , x. s. jiang@xmath11 , d. p. jin@xmath11 , s. jin@xmath11 , y. jin@xmath11 , yi jin@xmath11 , y. f. lai@xmath11 , f. li@xmath11 , g. li@xmath15 , h. h. li@xmath11 , j. li@xmath11 , j. c. li@xmath11 , q. j. li@xmath11 , r. y. li@xmath11 , s. m. li@xmath11 , w. d. li@xmath11 , w. g. li@xmath11 , x. l. li@xmath22 , x. q. li@xmath23 , y. l. li@xmath24 , y. f. liang@xmath25 , h. b. liao@xmath26 , c. x. liu@xmath11 , f. liu@xmath26 , fang liu@xmath13 , h. h. liu@xmath11 , h. m. liu@xmath11 , j. liu@xmath12 , j. b. liu@xmath11 , j. p. liu@xmath27 , r. g. liu@xmath11 , z. a. liu@xmath11 , z. x. liu@xmath11 , f. lu@xmath11 , g. r. lu@xmath28 , h. j. lu@xmath13 , j. g. lu@xmath11 , c. l. luo@xmath29 , l. x. luo@xmath24 , x. l. luo@xmath11 , f. c. ma@xmath22 , h. l. ma@xmath11 , j. m. ma@xmath11 , l. l. ma@xmath11 , q. m. ma@xmath11 , x. b. ma@xmath28 , x. y. ma@xmath11 , z. p. mao@xmath11 , x. h. mo@xmath11 , j. nie@xmath11 , z. d. nie@xmath11 , s. l. olsen@xmath19 , h. p. peng@xmath13 , n. d. qi@xmath11 , c. d. qian@xmath30 , h. qin@xmath29 , j. f. qiu@xmath11 , z. y. ren@xmath11 , g. rong@xmath11 , l. y. shan@xmath11 , l. shang@xmath11 , d. l. shen@xmath11 , x. y. shen@xmath11 , h. y. sheng@xmath11 , f. shi@xmath11 , x. shi@xmath12@xmath31 , h. s. sun@xmath11 , j. f. sun@xmath11 , s. s. sun@xmath11 , y. z. sun@xmath11 , z. j. sun@xmath11 , x. tang@xmath11 , n. tao@xmath13 , y. r. tian@xmath18 , g. l. tong@xmath11 , g. s. varner@xmath19 , d. y. wang@xmath11 , j. z. wang@xmath11 , k. wang@xmath13 , l. wang@xmath11 , l. s. wang@xmath11 , m. wang@xmath11 , p. wang@xmath11 , p. l. wang@xmath11 , s. z. wang@xmath11 , w. f. wang@xmath11@xmath32 y. f. wang@xmath11 , z. wang@xmath11 , z. y. wang@xmath11 , zhe wang@xmath11 , zheng wang@xmath15 , c. l. wei@xmath11 , d. h. wei@xmath11 , n. wu@xmath11 , y. m. wu@xmath11 , x. m. xia@xmath11 , x. x. xie@xmath11 , b. xin@xmath22@xmath21 , g. f. xu@xmath11 , h. xu@xmath11 , s. t. xue@xmath11 , m. l. yan@xmath13 , f. yang@xmath23 , h. x. yang@xmath11 , j. yang@xmath13 , y. x. yang@xmath33 , m. ye@xmath11 , m. h. ye@xmath15 , y. x. ye@xmath13 , l. h. yi@xmath14 , z. y. yi@xmath11 , c. s. yu@xmath11 , g. w. yu@xmath11 , c. z. yuan@xmath11 , j. m. yuan@xmath11 , y. yuan@xmath11 , s. l. zang@xmath11 , y. zeng@xmath14 , yu zeng@xmath11 , b. x. zhang@xmath11 , b. y. zhang@xmath11 , c. c. zhang@xmath11 , d. h. zhang@xmath11 , h. y. zhang@xmath11 , j. zhang@xmath11 , j. w. zhang@xmath11 , j. y. zhang@xmath11 , q. j. zhang@xmath11 , s. q. zhang@xmath11 , x. m. zhang@xmath11 , x. y. zhang@xmath20 , y. y. zhang@xmath11 , yiyun zhang@xmath25 , z. p. zhang@xmath13 , z. q. zhang@xmath28 , d. x. zhao@xmath11 , j. b. zhao@xmath11 , j. w. zhao@xmath11 , m. g. zhao@xmath23 , p. p. zhao@xmath11 , w. r. zhao@xmath11 , x. j. zhao@xmath11 , y. b. zhao@xmath11 , z. g. zhao@xmath11@xmath34 , h. q. zheng@xmath12 , j. p. zheng@xmath11 , l. s. zheng@xmath11 , z. p. zheng@xmath11 , x. c. zhong@xmath11 , b. q. zhou@xmath11 , g. m. zhou@xmath11 , l. zhou@xmath11 , n. f. zhou@xmath11 , k. j. zhu@xmath11 , q. m. zhu@xmath11 , y. c. zhu@xmath11 , y. s. zhu@xmath11 , yingchun zhu@xmath11@xmath35 , z. a. zhu@xmath11 , b. a. zhuang@xmath11 , x. a. zhuang@xmath11 , b. s. zou@xmath11 + ( bes collaboration ) + @xmath11_institute of high energy physics , beijing 100049 , people s republic of china _
+ @xmath15_china center for advanced science and technology , beijing 100080 , people s republic of china _
+ @xmath33_guangxi normal university , guilin 541004 , people s republic of china
_ + @xmath24 _ guangxi university , nanning 530004 , people s republic of china
_ + @xmath28 _ henan normal university , xinxiang 453002 , people s republic of china _
+ @xmath26_huazhong normal university , wuhan 430079 , people s republic of china
_ + @xmath14 _ hunan university , changsha 410082 , people s republic of china _ + @xmath22_liaoning university , shenyang 110036 , people s republic of china _ + @xmath29_nanjing normal university , nanjing 210097 , people s republic of china
_ + @xmath23 _ nankai university , tianjin 300071 , people s republic of china _ + @xmath12 _ peking university , beijing 100871 , people s republic of china
_ + @xmath20 _ shandong university , jinan 250100 , people s republic of china _ + @xmath30 _ shanghai jiaotong university , shanghai 200030 , people s republic of china
_ + @xmath25 _ sichuan university , chengdu 610064 , people s republic of china
_ + @xmath18 _ tsinghua university , beijing 100084 , people s republic of china _ + @xmath19 _ university of hawaii , honolulu , hawaii 96822 , usa _ + @xmath13 _ university of science and technology of china , hefei 230026 , people s republic of china
_ + @xmath27 _ wuhan university , wuhan 430072 , people s republic of china _
+ @xmath16 _ zhejiang university , hangzhou 310028 , people s republic of china _
+ @xmath36 current address : iowa state university , ames , iowa 50011 - 3160 , usa .
+ @xmath37 current address : purdue university , west lafayette , indiana 47907 , usa .
+ @xmath38 current address : cornell university , ithaca , new york 14853 , usa .
+ @xmath39 current address : laboratoire de lacclratear linaire , f-91898 orsay , france .
+ @xmath40 current address : university of michigan , ann arbor , michigan 48109 , usa .
+ @xmath41 current address : desy , d-22607 , hamburg , germany .
+ 0.3 cm |
delayed feedback was found to be a highly efficient tool for controlling the coherence of noisy oscillators @xcite .
even weak feedback with sufficiently long delay time can diminish or enhance the phase diffusion constant , quantitative measure of coherence , by one order of magnitude and even more .
the theory of the delay feedback control of the phase diffusion was developed in @xcite .
however , for long delay times and vanishing noise , the multistability of the mean frequency can occur . in the presence of multistability , noise results in intermittent switchings between states with different mean frequencies and , thus , the phase diffusion is contributed not only by fluctuations around the mean linear growth of the phase but also by the alternation of ` local ' mean growth rates . for weak noise which is a physically relevant situation
it turns - out to be possible to find natural variable in terms of which the switching between two stable phase growth rates becomes a perfect telegraph process . in @xcite , frequency multistability and noise - induced switchings were studied for extremely long delay times , although without consideration of coherence and phase diffusion . in this work , on the basis of the ` telegraphness ' property
, we construct the analytical theory of the effect of delayed feedback on the phase diffusion in the presence of multistability . in agreement with the results of numerical simulation
we analytically derive that the phase diffusion constant has giant peaks close to the points where the residence times in two states are equal .
the dynamics of a limit - cycle oscillator subject to weak action can be described within the framework of the phase reduction , where the system state in determined solely by the oscillation phase @xcite .
the phase reduction can be used as well for the case of weak noise , including a @xmath0-correlated one @xcite . for nearly harmonic oscillators
subject to white gaussian noise and linear delayed feedback the phase equation reads @xcite @xmath1+\varepsilon\xi(t)\ , , \label{eq-01}\ ] ] where @xmath2 is the natural frequency of the oscillator , @xmath3 and @xmath4 are the strength and delay time of the feedback , respectively , @xmath5 is the noise strength , @xmath6 is the normalised @xmath0-correlated gaussian noise : @xmath7 recently , a strong impact of weak anharmonicity for recursive delay feedback has been reported for the effects under our consideration @xcite ; however , for a ` simple ' delay the harmonic approximation and model reduction remain well justified for many physical , chemical and biological systems ( _ e.g. _ , see @xcite ) .
noise disturbs the linear growth of the phase , resulting in its diffusion according to the law @xmath8 ^ 2\rangle = dt$ ] for @xmath9 . in the control - free case , @xmath10 . the diffusion constant @xmath11 measures the coherence of oscillations ; the smaller @xmath11 the better coherence . within the framework of the linear - in - noise approximation
, delayed feedback has been found to change the mean oscillation frequency @xmath12 and the diffusion constant @xmath11 as follows @xcite ; @xmath13
let us first discuss the system dynamics for vanishing noise . in the solution of equation is plotted for the noise - free case with the solid line .
one can notice that for large @xmath4 the frequency @xmath12 is not always unique , which has already been noticed in @xcite .
the linear stability analysis for this system reveals that the linear growth solution @xmath14 is unstable when @xmath15 . in @xcite
, the stability analysis can be found for the monotonous perturbations , the analysis for the oscillatory perturbations ( to be published elsewhere ) does not change this condition .
if one moves along the curve in from the origin , this condition will be fulfilled on the segments of the line going from the minimal values of @xmath12 towards the maximal ones between points with vertical inclination .
these segments , say @xmath16 , are plotted with gray and separate the branches of stable solutions which are plotted with blue , @xmath17 , and red , @xmath18 , in the domains of multistability .
equation system is derived within the framework of the linear - in - noise approximation , which considers only small fluctuations around a state with constant phase growth rate @xmath12 ; it does not handle hopping between several states with @xmath19 .
equation has to be considered as a ` local ' phase diffusion at the state with corresponding @xmath12 .
the system dynamics is dominantly controlled by the delay term , which makes the events of transition between two stable states of the noise - free system non - local in time ; for instance , @xmath20 can be kicked - out by noise from @xmath17 , reach @xmath18 and then again return to @xmath17 .
thus , the noise - driven dynamics of the system in the presence of multistability is a non - trivial problem requiring subtle consideration . for longer delay time @xmath4
there are domains with more than two locally stable regimes . as a first step , in this work we restrict our consideration to the case of multistability between two stable regimes ; for @xmath21 and @xmath22 ( )
this is the case as long as @xmath23 .
let us rewrite equation in the form @xmath24 where @xmath25 is the phase growth per delay time .
this integral form of our dynamic system suggests that @xmath26 can be a suitable ` natural ' variable to trace the switchings between two states .
indeed , in one can see that in terms of @xmath27 the system state is a ` telegraph ' signal ; transitions between two states occur without immediate reversals
each reverse transition occurs as an event independent of the previous transition the residence time distributions for two states are exponential .
this ` telegraphness ' can be used to develop analytical theory of the dynamics of a system with multistability .
the diffusion constant of an asymmetric telegraph process altering between two states with growth rates @xmath17 and @xmath18 and mean residence times @xmath28 and @xmath29 , respectively , can be calculated ; @xmath30 the net phase diffusion is contributed by ` telegraph ' diffusion and local diffusions at states @xmath31 ; @xmath32 where @xmath33 ( equation ) .
singular behaviour is observed for the local phase diffusion at the edges of the multistability domains ( ) . in the local phase diffusions @xmath34 are plotted with the solid line .
the line is black for the domains with no multistability and coloured otherwise ; the same colours mark corresponding regimes in .
one can notice that at the edges of the multistability domains for the regime experiencing the tangential bifurcation at this edge the local phase diffusion tends to infinity .
this divergence of local diffusion rises the question how the net diffusion behaves on the edges of the multistability domains . to answer this question
, one can perform a normal form analysis of the system near the tangential bifurcation in the presence of weak noise and derive scaling properties for quantifiers of interest .
let us consider a point of the tangential bifurcation where regimes @xmath16 and @xmath18 appear ( annihilate ) , @xmath35 , and describe the noise - induced behaviour of the system for @xmath36 .
the normal form of the system dynamics near the bifurcation point reads @xmath37 with @xmath38 and @xmath39 , where @xmath40 is a positive number of order of magnitude of @xmath41 . for this normal form the mean first passage time from the potential
well can be calculated for weak noise ; @xmath42^{-\frac{1}{2 } } \exp\left(\frac{4}{3\varepsilon^2}[k_2(\tau-\tau_{2\ast})]^\frac{3}{2}\right ) .
\label{eq-06}\ ] ] since @xmath43 , the telegraph diffusion becomes @xmath44 , which tends to zero for @xmath45 .
subtle analysis of equations yields the scaling law for the local diffusion @xmath46 ; it will be convenient to introduce constant @xmath47 as follows @xmath48 .
the behaviour of @xmath49 is regular around @xmath35 .
hence , equation takes the form @xmath50^{3/2}}\ , .
\label{eq-07}\ ] ] one can see that for @xmath51 , where the weak noise approximation and our derivations are valid , the last term in rapidly decreases as @xmath4 tends to @xmath52 .
thus , both the telegraph diffusion and the contribution of the diverging local diffusion into the net diffusion vanish at the edge of the multistability domains , and the behaviour of the net diffusion is regular , as it can be also seen from the results of numerical simulation ( see ) . within a multistability domain
the mean residence time @xmath28 decreases as @xmath4 increases and tends to zero at the right edge of the multistability domain , while @xmath29 increases starting from zero at the left edge of the domain . at certain value
@xmath53 these times become equal , @xmath54 .
let us consider the vicinity of this point .
for small deviations @xmath55 , one can write @xmath56 \equiv t_\ast\exp[\mp(q_{1,2}/\varepsilon^2)\delta\tau]$ ] .
notice , since the mean residence times depend on @xmath4 exponentially strong with small number @xmath57 in the denominator of the argument of the exponential , we linearise not @xmath58 but @xmath59 next to @xmath53 .
coefficients @xmath60 are positive number of order of magnitude of @xmath41 . hence , @xmath61 the exponential in the nominator is a small correction to the cube of hyperbolic cosine in the denominator .
one has to expect a strong peak near @xmath62 with height somewhat above @xmath63 .
notice , for the weak noise , the latter expression is exponentially large , as @xmath64 is exponentially large .
moreover , for vanishing noise the width of the peak @xmath65 , while its height is exponentially large in @xmath66 , _ i.e. _ , the integral of this peak over @xmath4 diverges exponentially fast . in one
can see these peaks in the results of numerical simulation .
noteworthy , these peaks have a well pronounced triangular shape in the linear log scale , meaning equation to represent the behaviour of the phase diffusion remarkably well for the major part of each multistability domain , not only at its centre . in one
can see how the noise strength influences peaks in multistability domains .
the numerical simulation data in all the figures are calculated with time series of length @xmath67 .
in this paper we have developed the theory of the effect of delayed feedback on coherence of noisy phase oscillators in the presence of frequency multistability induced by this time delay .
the coherence has been quantified by the phase diffusion constant @xmath11 .
the process of alternation between two states has been demonstrated to be well representable as an asymmetric markovian ` telegraph ' process .
employment of this ` telegraphness ' property allows constructing the framework for analytical study on the problem .
the behaviour of the phase diffusion constant has been revealed to be smooth on the edges of the multistability domains .
giant peaks in the dependence of the phase diffusion constant on delay time have been discovered near points where the mean residence times in two states are equal ( see equation and figures [ fig3 ] and [ fig4 ] ) . remarkably , their ` integral strength ' increases for vanishing noise , as their width @xmath65 while height is exponentially large in @xmath66 .
for the longer delay time the peaks become taller . | for self - sustained oscillators subject to noise the coherence , understood as a constancy of the instantaneous oscillation frequency , is one of the primary characteristics .
the delayed feedback has been previously revealed to be an efficient tool for controlling coherence of noise - driven self - sustained oscillator .
the effect of the delayed feedback control on coherence is stronger for a longer delay time .
meanwhile , the instantaneous frequency of a noise - free oscillator can exhibit multistability for long delay time .
the impact of the delay - feedback - induced multistability on the oscillation coherence , measured by the phase diffusion constant , of a noisy oscillator is studied in this work both numerically and analytically . |
the class of _ @xmath1-nearly planar graphs _
contains all graphs embeddable in the plane with at most @xmath1 crossings per edge ( see figure [ fig : this2one ] for an example of a @xmath3-nearly planar graph ) .
this graph family was introduced for the first time by pach and tth @xcite under the name _
@xmath1-quasi planar graphs _ ( we adopt the term _
@xmath1-nearly planar _ as the term _ @xmath1-quasi planar _ has been extensively used to denote embedded graphs without @xmath1 pairwise crossing edges , see @xcite ) .
the class of @xmath1-nearly planar graphs does not have the good closedness properties of topologically - defined graph classes .
in particular , they are not closed under topological minors or contractions see corollary [ notclosed ] however they are closed under subgraphs .
we stress that the study of nearly planar graphs marks a direction that is `` orthogonal '' to the one of ( topological ) minor free graphs as every graph is a topological minor of some @xmath4-nearly planar graph : take any drawing of a graph in the plane and for each edge @xmath5 , introduce a vertex in each segment delimited by two consecutive crossings along @xmath5 .
several combinatorial and algorithmic results on planar graphs have been extended to @xmath1-nearly planar graphs .
for instance , according pach and tth @xcite , @xmath1-nearly planar graphs are sparse : each @xmath6-vertex @xmath1-nearly planar graph has at most @xmath7 edges .
moreover bodlaender and grigoriev proved that several problems in graphs admit eptas ( efficient polynomial time approximation schemes ) when restricted to @xmath1-nearly planar graphs @xcite .
consider a distance metric between the vertices of graphs .
we define the _ @xmath2-power _ of a graph @xmath0 under this metric as the graph with the same vertex set as @xmath0 and where two vertices @xmath8 and @xmath9 are adjacent if and only if their distance is upper bounded by @xmath2 .
our aim is to define such a metric that `` captures '' the concept of @xmath1-nearly planar graphs .
this metric is defined on embedded graphs and is called _ wall - by - wall _ distance ( for the definition see section [ sec : classes ] ) . to explain how wall - by - wall expresses @xmath1-nearly planar graphs , we introduce the concept of _ @xmath2-flat graphs _ that are subgraphs of the @xmath2-powers of plane graphs under the new metric .
our main result is a proof , that the classes of @xmath2-flat graphs and @xmath1-nearly planar graphs are _ parametrically equivalent _ , i.e. , there are functions @xmath10 and @xmath11 such that every @xmath1-nearly planar graph is @xmath12-flat and every @xmath2-flat graph is @xmath13-nearly planar .
this enables us to represent every nearly planar graph @xmath0 by a suitably - defined plane graph @xmath14 whose structure `` geometrically represents '' @xmath0 .
the paper is organized as follows : in section [ defprel ] , we give some basic definitions and results . in section [ sec : classes ] we introduce the notion of @xmath2-flat graphs and we prove several properties of them .
the proof of our main result is presented in section [ parma ] .
all graphs in this paper are undirected , without loops , and may have multiple edges . if a graph has no multiple edges or loops we call it _
simple_. given a graph @xmath0 , we denote by @xmath15 its vertex set and by @xmath16 its edge set .
we say that @xmath14 is a _ subgraph of @xmath0 via an injection @xmath17 _ if for every edge @xmath18 , @xmath19 is an edge of @xmath0 and we denote this fact by @xmath20 .
we also say that @xmath14 is a subgraph of @xmath0 if @xmath20 via such an injection @xmath21 and we write @xmath22 . for any set of vertices @xmath23 , we denote by @xmath24 $ ] the subgraph of @xmath0 induced by the vertices from @xmath25 .
moreover , we denote by @xmath26 the neighbors of the vertices in @xmath25 that do not belong to @xmath25 .
we also set @xmath27=s\cup n_{g}(s)$ ] and @xmath28 .
given two graphs @xmath14 and @xmath0 , we write @xmath29 and call @xmath14 a _ minor _ of @xmath0 , if @xmath14 can be obtained from a subgraph of @xmath0 by edge contractions ( the _ contraction _ of an edge @xmath30 in a graph @xmath0 is the operation of replacing @xmath8 and @xmath9 by a new vertex @xmath31 that is made adjacent with all the neighbors of @xmath8 and @xmath9 in @xmath0 that are different from @xmath8 and @xmath9 ) .
moreover , we say that @xmath14 is a _ contraction _ of @xmath0 if @xmath14 can be obtained from @xmath0 by contracting edges .
we say that @xmath14 is a _ topological minor _ of @xmath0 if a subdivision of @xmath14 is a subgraph of @xmath0 ( the _ subdivision _ of an edge @xmath30 is its replacement by a path of length 2 whose endpoints are @xmath8 and @xmath9 ) .
to simplify notations , we use `` @xmath32 '' instead of saying that `` @xmath33 for some recursive function @xmath34 '' . * plane graphs*. throughout this paper , we use the term _ graph embedding _ to denote an embedding of a graph in the sphere where if two edges cross , they cross to a point that is not a vertex of the graph .
we use the term _ plane graph _ for an embedding of a graph without crossings .
a graph is _ planar _ if it has some planar embedding .
given a plane graph @xmath0 , we denote by @xmath35 the set of faces of @xmath0 , i.e. , the connected components of @xmath36 , that are open sets . * radial and medial graph*. given a plane graph @xmath0 with at least one edge , we define its _ radial graph _ @xmath37 as the plane graph whose vertex set is @xmath38 ( @xmath39 is the dual graph of @xmath0 ) and whose edges are defined as follows : let @xmath40 be the connected components of @xmath41 isomorphic to open discs .
observe that for any @xmath42 , the component @xmath43 is an open set whose boundary is incident to one vertex @xmath44 and to one vertex @xmath45 .
the edge set of @xmath37 is the set @xmath46 where edge @xmath47 has multiplicity @xmath4 if both , @xmath48 and @xmath49 , have degree at least @xmath3 in @xmath0 and @xmath39 , respectively , otherwise , the edge multiplicity of @xmath47 is @xmath3 .
notice that @xmath50 is a bipartite graph , whose parts are the vertex and face sets of @xmath0 , respectively
. given a plane graph @xmath0 , we define its _ medial graph _ @xmath51 as the dual of its radial graph , i.e. , @xmath52 . clearly , the vertices of @xmath0 correspond to faces of @xmath51 .
* plane transformation*. we describe a procedure we will be using in order to transform any simple embedding of a ( not necessarily planar ) graph @xmath0 on the sphere , to a plane graph @xmath53 . *
modify the embedding of @xmath0 so that no more than two edges are crossing at the same point . *
subdivide the edges of @xmath0 and draw the resulting graphs such that each edge is crossed only once . in the resulting graph , for every crossed pair of edges @xmath30 and @xmath54 * add the cycle @xmath55 ( notice that this cycle can be drawn without creating further crossings ) , and * remove every such pair of crossing edges @xmath5 and @xmath56 .
notice that , by construction , @xmath57 is indeed a plane graph . for an illustration of the construction of @xmath57 , see figure [ fig : gadget ] .
we will be referring to this procedure as _ plane transformation _ of a graph .
let @xmath0 be a plane graph . for any vertex @xmath58 ,
the edges incident to @xmath59 appear in the drawing in the natural cyclic order which we denote by @xmath60 ( such a cyclic order is well defined as @xmath0 is loopless ) .
two edges are called _ attached _ if they share a vertex , say @xmath58 , and appear consecutively in the cyclic order @xmath60 or its inverse . * wall by wall distance*. let @xmath61 be two edges in @xmath0 . an @xmath62-_wall - by - wall walk _ is a sequence of distinct edges @xmath63 such that @xmath64 and @xmath65 and every two consecutive edges in @xmath63 are attached .
intuitively , such a walk can be explained in the following way .
let the plane graph represent a building map where edges represent walls , each having a door .
then , a _ wall - by - wall - walk _ , denoted also as a * wbw*-walk , is a walk from wall @xmath56 to wall @xmath66 continuously touching walls and possibly going through doors . now , let @xmath8 and @xmath9 be two vertices of the plane graph @xmath0 .
the * wbw*-walk between @xmath8 and @xmath9 , also called _ @xmath67-*wbw*-walk _ , is an @xmath68-*wbw*-walk such that the edges @xmath69 and @xmath70 are incident to the vertices @xmath8 and @xmath9 , respectively .
the _ wall - by - wall distance between @xmath8 and @xmath9 _ , denoted by * wbw * , is the number of edges in the shortest @xmath67-*wbw*-walk , if such a walk exists , otherwise it is infinite ( in the case that no path in @xmath0 connects @xmath8 and @xmath9 ) .
it is easy to verify that if @xmath71 for some integer @xmath72 , then there exists a _ wall - by - wall path _ of length @xmath72 , namely a * wbw*-walk whose edges form an acyclic subgraph of @xmath0 .
again , we use the notations _ * wbw*-path _ and _ @xmath67-*wbw*-path_.
notice that the wall - by - wall distance is lower bounded by the standard distance in graphs .
however , the difference between the two distances may be arbitrarily large , [ e.g. the distance of two non adjacent vertices of a star ] .
the following observation gives an easy way to express the wall by wall distance in terms of the common distance .
[ obs : tdhisplo ] let @xmath0 be a plane graph and let @xmath51 be its medial .
let also @xmath8 and @xmath9 be two vertices in @xmath0 and let @xmath73 and @xmath74 be the faces of @xmath51 corresponding to @xmath8 and @xmath9 respectively .
then the wall - by - wall distance between @xmath8 and @xmath9 in @xmath0 is one more than the minimum distance in @xmath51 between a vertex in the boundary of @xmath73 and a vertex in the boundary in @xmath74 . * @xmath2-flat graphs*. we consider taking powers of a plane graph with respect to the new metric .
let @xmath14 be a simple plane graph and let @xmath2 be a positive integer . for @xmath75 , we define graph @xmath76 as a graph with vertex set @xmath77 and edge set @xmath78 , where @xmath79 contains all pairs @xmath80 that do not belong to @xmath81 and @xmath82 .
we say that a graph @xmath0 is _ @xmath2-flat _ if it is a subgraph of @xmath83 for some planar graph @xmath14 .
notice that , according to the above definition , @xmath4-flat graphs are exactly the planar graphs .
* @xmath1-nearly planar graphs*. we use the term _
@xmath1-nearly planar graph embedding _ for an embedding with at most @xmath1 crossings per edge .
a graph @xmath0 is _ @xmath1-nearly planar _ if it has a @xmath1-nearly planar graph embedding . for simplicity , we use the same notation for both @xmath1-nearly planar graphs and their embeddings . in contrary to the case of planar graphs , even the recognition of @xmath4-nearly planar graphs is a np - hard problem @xcite . the following theorem states our main result , which links the two aforementioned graph classes .
[ equiv ] the classes of @xmath2-flat graphs and @xmath1-nearly planar graphs are parametrically equivalent .
from @xcite , each @xmath6-vertex @xmath1-nearly planar graph has at most @xmath7 edges , i.e. is sparse .
an immediate consequence of theorem [ equiv ] is that @xmath2-flat graphs are also sparse .
[ smcor ] every @xmath2-flat graph on @xmath6 vertices has @xmath84 edges . by definition , @xmath2-flat graphs are closed under subgraphs ( i.e. a subgraph of a @xmath2-flat graph is also @xmath2-flat ) .
we prove that this is not the case for topological minors and contractions .
[ notclosed ] for every @xmath85 , @xmath2-flat graphs are not closed neither under taking topological minors nor under taking contractions .
the same holds also for @xmath1-nearly planar graphs for @xmath86 .
we present the proof for the case of @xmath2-flat graphs .
the proof for @xmath1-nearly planar graphs is very similar and uses the same gadgets .
let @xmath87 be a complete graph that is not @xmath2-flat , for some integer @xmath2 ( such a value for @xmath6 exists because of corollary [ smcor ] ) .
we will show that a @xmath3-flat graph @xmath88 contains the clique @xmath87 as a topological minor / contraction .
for the case of the topological minor relation , we construct @xmath88 as follows .
let @xmath89 be the plane transformation of an embedding of @xmath87 .
let also @xmath88 be the graph obtained after the third step of the plane transformation of @xmath87 to @xmath89 , namely before eliminating the pair of crossed edges . by construction , the graph @xmath88 contains @xmath90 as a topological minor .
we now claim that @xmath88 is @xmath3-flat . to see this ,
consider the plane graph @xmath89 .
for any pair of deleted crossed edges @xmath67 and @xmath91 , @xmath89 has a face bounded by the cycle @xmath55 .
this implies that * wbw*@xmath92 and * wbw*@xmath93 .
thus @xmath94 contains all deleted edges .
this implies that @xmath95 and therefore @xmath88 is @xmath3-flat . for the case of the contraction relation ,
we construct @xmath88 as follows .
let @xmath89 be a @xmath96-grid and let @xmath97 .
certainly , @xmath88 is a @xmath3-flat graph .
moreover , it is easy to verify that @xmath89 can be contracted to the complete bipartite @xmath98 and this in turn to a clique of @xmath6 vertices
. therefore the same holds for @xmath99 . by corollary [ smcor ]
, there is no @xmath2 such that all graphs are @xmath2-flat .
we define the following two functions from graphs to integers @xmath100 according to theorem [ equiv ] , the two functions @xmath101 and @xmath102 are parametrically equivalent , i.e. , the one is bounded by a function of the other .
this means that the _ wall - by - wall _ metric permits us to treat @xmath1-nearly planar graph embeddings as subgraphs of powers of plane graphs .
this section is dedicated to the proof of theorem [ equiv ] . [
cross : flat ] let @xmath0 be a simple @xmath1-nearly plane graph on @xmath6 vertices , then , there exists a graph @xmath14 on @xmath103 vertices such that @xmath104 via an injection @xmath105 , where every vertex in @xmath14 is within * wbw*-distance at most @xmath1 from some vertex of @xmath106 .
let @xmath14 be the plane transformation of @xmath0 .
consider an edge @xmath107 crossed @xmath108 times .
notice , for such an edge , it is sufficient to subdivide it with @xmath109 additional vertices . then , by construction , in @xmath14 there is a @xmath67-*wbw*-path of length @xmath110 ( for an illustration see figure [ fig : k - cross ] ) .
thus , @xmath111 .
moreover , each vertex on this path has * wbw*-distance at most @xmath1 from either @xmath8 or @xmath9 . and
since all vertices of @xmath14 not in @xmath106 are created as a subdivision of an edge of @xmath0 , every vertex in @xmath14 is within * wbw*-distance at most @xmath1 from some vertex of @xmath106 .
[ triangulate ] let @xmath14 be a simple and connected plane graph .
then , there exists a simple plane graph @xmath112 and an injection @xmath113 such that 1 .
@xmath112 is @xmath114-connected and triangulated , 2 .
@xmath115 , 3 . for every @xmath116 it holds that @xmath117 , 4 .
every vertex in @xmath112 is within * wbw*-distance at most @xmath118 from some vertex of @xmath119 , 5 . for every non - negative integer @xmath120 @xmath121 ( where @xmath122 ) .
for every vertex @xmath59 of @xmath14 , where @xmath123 denotes the cyclic order of its edges , we do the following .
if its degree @xmath124 is not less than two , we add in the plane @xmath125 vertices and join them to @xmath59 , so that between every two consecutive edges in @xmath123 now lies a new edge . otherwise , if @xmath126 , we add two vertices and join them to @xmath59 and if @xmath127 ( i.e. @xmath0 contains only one vertex ) , we add three vertices and join them to @xmath59 .
next , for every face @xmath128 in the given embedding of @xmath14 , we add edges between the new vertices that lie now inside @xmath128 , which form a cycle that bounds an empty open disc in the plane
. note that in the procedure described , all added edges can be drawn so that they do nt cross each other or already existing edges .
thus , the resulting graph is again plane . we denote it as @xmath129 .
notice that we have effectively surrounded each edge of @xmath14 with two more edges in @xmath129 , in a way that each pair of vertices in @xmath129 , is connected by at least three disjoint paths .
hence , @xmath129 is @xmath114-connected .
consider now a particular triangulation of @xmath129 , we denote it by @xmath130 , which is the union of @xmath129 and its radial @xmath131 , i.e. the graph with vertex set @xmath132 and edge set @xmath133 . as @xmath14 is a simple graph , all faces of @xmath134 are open disks with at least 3 vertices in their boundary .
this means that the graph @xmath130 is also 3-connected and therefore it satisfies the first two properties of the lemma .
let @xmath135 be arbitrary two vertices in @xmath14 .
consider a @xmath67-*wbw*-path @xmath136 in @xmath14 of length @xmath137 .
we construct the corresponding * wbw*-path @xmath138 in @xmath130 where , for every two sequential edges @xmath139 of @xmath136 , we follow the steps below . the edges @xmath139 are attached in @xmath14 .
let @xmath140 be their common endpoint
. by construction , between them lies an edge of @xmath129 , say @xmath141 .
add this edge to the path @xmath138 .
since @xmath129 is @xmath114-connected , @xmath142 and @xmath143 are incident on two different faces of @xmath129 .
add the two edges of @xmath131 that have as one endpoint the vertex @xmath140 and as the other the vertices of @xmath131 in the faces of @xmath129 incident to @xmath142 and @xmath143 .
observe , the resulting sequence of edges is a @xmath67-*wbw*-path in @xmath130 . since we inserted three edges between every two sequential edges of @xmath136 and the length of @xmath136 is @xmath144 , the length of the @xmath67-*wbw*-path @xmath138 in @xmath130 is @xmath145 and property 3 holds .
note , that this also immediately implies property 5 . to see that property 4 holds ,
recall that if a vertex @xmath59 of @xmath130 is not in @xmath119 , then either @xmath146 so @xmath147 for a vertex @xmath148 , or @xmath149 so @xmath150 for a vertex @xmath151
. in the worst case , a * wbw*-path of length @xmath118 in @xmath130 connects @xmath59 with a vertex in @xmath119 .
[ flat : cross ] let @xmath14 be an @xmath6-vertex 3-connected triangulated planar graph and @xmath2 a positive integer .
then , there exists an embedding of @xmath152 with at most @xmath153 crossings per edge .
given a plane graph @xmath14 and an integer @xmath75 , the edges of @xmath81 are called _ old _ edges of @xmath152 while the edges in @xmath154 are called _ new _ edges of @xmath152 . consider two arbitrary vertices @xmath155 with @xmath156 , i.e. there exists a @xmath67-*wbw*-path @xmath157 .
we have to draw a new edge @xmath5 in @xmath158 joining @xmath8 and @xmath9 .
let @xmath159 and @xmath160 be incident to @xmath8 and @xmath9 , respectively . since sequential edges in the * wbw*-path are incident to a face , for any @xmath161 , edges @xmath162 and @xmath163 characterize uniquely the shared face @xmath164 .
we start drawing @xmath5 in @xmath165 crossing @xmath166 .
then , @xmath5 goes through @xmath167 crossing @xmath168 and so on till it reaches @xmath9 through @xmath169 .
so , in total @xmath5 passes through @xmath170 faces . observe that , for every new edge crossing an old edge @xmath5 and entering face @xmath128 , there are three possibilities : either crossing one of the other two old edges incident to @xmath128 or to end in the vertex of the triangle lying opposite to @xmath5 .
since in a triangulated graph all vertices within * wbw*-distance of at most 2 are already joined , this yields that from any vertex @xmath59 of @xmath158 we can reach with new edges at most @xmath171 other vertices , per each face incident to @xmath59 . calculating carefully , we derive that any old edge can be crossed by at most @xmath172 new edges .
now , let us count how many times the new edges can be crossed .
first we count that the total number of new edges passing through a face of @xmath130 is at most @xmath173 .
exaggeratively , assume that every new edge is crossed within the face by all other entering edges , i.e. a new edge is crossed inside the face at most @xmath174 times .
any new edge goes through at most @xmath175 faces and crosses additionally at most @xmath176 old edges .
therefore , any new edge is crossed at most @xmath177 times . combining the previous lemmata
we prove the following corollary , which immediately implies theorem [ equiv ] .
[ lambda : xi ] for every graph @xmath0 , @xmath178 .
let @xmath179 be a plane graph and @xmath2 an integer such that @xmath180 .
w.l.o.g we assume that @xmath179 is connected ( otherwise , we work with each of its connected components separately ) .
let @xmath14 be the graph obtained from @xmath179 after subdividing once each of it edges .
observe that @xmath14 is a simple and connected planar graph where @xmath181 .
then , we may apply lemma [ triangulate ] and obtain that @xmath0 is a subgraph of @xmath182 , for a planar 3-connected triangulated graph @xmath130 . by lemma [
flat : cross ] , we can draw this graph on the plane with at most @xmath183 crossings per edge .
we derive that @xmath184 .
consider now an embedding of @xmath0 with at most @xmath185 crossings per edge . by lemma [ cross : flat
] , there is a plane graph @xmath186 such that @xmath187 , i.e. , @xmath188 .
in this note we introduced the concept of @xmath2-flat graphs as an alternative for @xmath1-nearly planar graphs .
we believe that this class has independent graph theoretic interest .
the first question is whether the exponential bound of corollary [ lambda : xi ] can become polynomial .
also , it is an open issue whether this new concept can be useful to extend known algorithmic results and techniques on @xmath1-nearly planar graphs . | a graph @xmath0 is _ @xmath1-nearly planar _ if it can be embedded in the sphere so that each of its edges is crossed at most @xmath1 times .
the family of @xmath1-nearly planar graphs is widely extending the notion of planarity .
we introduce an alternative parameterized graph family extending the notion of planarity , the @xmath2_-flat graphs _ , this time defined as powers of plane graphs in regard to a novel notion of distance , the _ wall - by - wall distance_. we show that the two parameterized graph classes are parametrically equivalent . *
keywords * : planarity , crossing number , graph powers |
the compact and time varying bright regions located in the form of umbral dots , light bridges and penumbral features harbor signatures of underlying physical processes below the visible layers @xcite .
last two decades of ground based high spatial resolution observations have revealed that umbra is inhomogeneous divided by light bridges and filled with bright points and elongated structures that can not be primarily produced by granular convection processes @xcite .
the penumbra is more dynamic with stratified temperature , velocity and magnetic field structure and consists of elongated channels with intermittent evershed flow @xcite .
extremely bright grains and regions of hot upflows are also observed at the middle of penumbra @xcite .
the features are integral part of the magnetic structure , which could be aggregation of many separate flux tubes pressed together to form a flux bundle or monolithic large flux tube @xcite . theoretically , there are three mechanisms that can produce bright points with some of the observed properties in the sunspot
. one of them is primarily based on the properties of fluids with high electrical conductivity trapped in magnetic field described by @xcite that were also observed in laboratory systems @xcite . in case of sunspots
, the plasma is trapped between the flux tubes in a cluster .
@xcite proposed that the sunspots are dynamical clustering of many separate flux tubes pressed together as single flux bundle at the visible surface keeping their separate identity beneath . while the magnetic pressure at the center of the flux tube axis is smaller by about 0.5 to 0.75 compared to the boundary near the photospheric layers , the field is more concentrated below .
the space between the flux tubes below the photosphere are filled with the field free plasma which become unstable and subjected to oscillation appearing as umbral dots @xcite .
such oscillatory plasma can also penetrate the surface with velocities of about 270 m s@xmath0 @xcite .
the observations of formation and decay process of the sunspots appears to be consistent with these theories of penetrative convection @xcite .
the second category of physical processes are developed for monolithic sunspots as a result of non - linear magnetic convection leading to oscillatory plasma columns @xcite .
the simulations show that the monolithic sunspots models can also explain , the formations of umbral dots with magnetic fields of 3.8 kilo gauss , central upward flowing plasma of upto 4 km s@xmath0 , surrounded by downward motion @xcite .
the velocity field at the top of plasma columns in these processes would display vertical velocities up to 1 to 2 km s@xmath0 .
the third type of theoretical model consider magnetohydrodynamic consequences of thin vertical gas column embaded in a sunspot umbra @xcite .
geometrically , this is a converging gas column with a week field than that of the ambient medium . in this case , both up and down flowing plasma exists depending on the temperature and density at the top of the column . while the appearance of near circular structure of the compact bright features are consistent with the theoretical models , other properties such as mass motion and magnetic field configuration needs to be examined for a better understanding .
several observations show not all the umbral dots are circular and the associated velocities vary .
the penumbra consists of elongated horizontal flux tubes through which the intermittent evershed flow occurs .
some of the theoretical models describe the penumbra as a result of the returning flux tubes anchored by the inter - granular lane through turbulent pumping @xcite .
other models interpret penumbral fine structures as due to the convection of field - free radially aligned gaps just below the visible surface @xcite .
these different competing models can now be examined in detail with the availability of high spatial resolution observations . with the high resolution spectropolarimetry observations and modeling it has been possible to look for finer descripencies between the theory and obserrvations .
for example the tip and polis observations show that the penumbral models with flux tubes embedded in non - magnetic plasma needs to account for the flow @xcite .
the high resolution images of sunspots obtained with adoptive optics system at dunn solar telescope show elongated structures associated with umbral dots towards the penumbral boundary that can not be entirely due to magnetoconvections @xcite .
the imaging and spectropolrimetry observations with the solar optical telescope ( sot ) on hinode provides best possible high spatial resolution observations to study these structures without seeing interruptions .
these observations have already contributed to our understanding of the sunspot bright points @xcite .
the similarity of penumbral grains and peripheral umbral dots have been found by sobotka and puschmann ( 2009 ) . on the other hand
, there is a difference in the evolution of these two structures @xcite . in this paper
, we use the broad band images and spectropolarimetry observations of sunspots to study the thermal and magnetic properties of isolated bright points in and around sunspots .
we find that there is continuity in the properties of bright features from the umbral dots that exists deep inside the umbra to the bright points in the vicinity of sunspots .
we discuss these results with contemporary models through a simple cartoon .
we have used the images of the sunspots with focal plane package of solar optical telescope ( sot ) on the hinode spacecraft @xcite .
the sunspot images through broadband filter imager ( bfi ) with spatial resolution resulting from 1 @xmath1 1 and 4 @xmath1 4 summing were used for this study .
after initial processing of the images for removing instrumental affects with standard procedures provided by hinode team and co - alignment , we perform flux calibration . in order to convert the data number into the units of solar irradiance , we compared the data numbers of solar flux obtained by the charge coupled device on ground and space before and after the launch of hinode to check the consistency of detector performance .
we measured average data numbers of area covering 100 square pixels in relatively non - plage locations of quiet sun away from the sunspots that may represent the average solar flux measured on ground @xcite .
the ground based measurements were performed by observing the disk center and averaging the entire field of view .
while the ground measurements may contain some magnetically active locations , we have excluded them from space measurements .
figure 1 presents the ground and space data number comparison for 29 april 2007 .
we find the general trend of data numbers of the detector remains same both on ground and on space with somewhat larger scatter through the filters with shorter wavelength .
possible inclusion of pixels with magnetic elements in the ground based data could result differences especially in caii and g - band flux . after correcting for the reduced filter transmission for the observing day , we compare the quiet sun values with a standard zero airmass solar spectrum to obtain the calibration factor to convert the data numbers into solar flux units @xcite .
the flux calibrated images were used to obtain the intensity of bright points in the sunspots .
the sunspots that were used for this study are given in table 1 .
the bright features representing umbral dots , peripheral umbral dots , penumbral grains and bright points in the vicinity of sunspots were identified from the ratio images obtained by dividing g - band and caii h frames as shown in figure 2 .
the contrast of the bright points in the inner portion of sunspots becomes much higher leading to their easy detectability .
the higher contrast in the ratio images are due to the featureless umbra in caii h images which represent the solar atmospheric height of about 150 km above the g - band height formation @xcite .
most of the sunspot flux tubes , that result the moving magnetic features , do not reach at lower atmospheric heights below about 300 km leading to their non - appearance at lower chromosphere @xcite . in this case , the caii h images act as high pass filter enhancing the features that do not reach the heights of about 150 km .
identification of bright features by comparing g - band and caii h images preclude the points that might be effected due to the presence of brighter locations in close proximity .
the photometry of the bright points were performed following method similar to the crowded - field photometry known as daophot ( stetson , 1987 ) .
the main steps are identification and flux measurement by fitting a gaussian surface function to the bright point .
initially , the coordinate of the bright points were visually determined from the ratio images .
the coordinates match very well with the locations of well defined , isolated bright points found by adopting the automatic procedures that compares the adjacent pixels to look for enhancement of local intensities ( riethmuller et al , 1008 , watanabe et al , 2009 ) . the additional constraint for selecting the bright point for our analysis is successful fitting of gaussian surface leading to a convergence for photometry flux extraction .
the exact center of the bright points were determined by using a derivative search at which the local derivative in the neighborhood of the manually found positions of brightness goes to zero .
a 20 @xmath1 20 pixel image around the center was extracted for flux measurements .
the image was fitted with a two dimensional elliptical gaussian function given in equation ( 1 ) , around the centroid of the rectilinearly gridded bright point image .
@xmath2 ( 1 ) where @xmath3 .
the x , y are the center and a , b are the axis of the ellipse along the line joining to the centroid of the large portion of the sunspot umbra .
the flux was calculated by summing the intensity of the pixels within half width of the gaussian surface .
average size of the bright point was computed by taking average of both axis .
figure 3 show typical bright points and corresponding gaussian surface .
figure 4 shows the histogram of the bright point radius .
the extracted flux over the gaussian function half - width from the co - aligned images through the filters represents the brightness of the different features in the sunspot that are shown in figures 5 .
figure 6 show the continuum flux superposed with the planck function for different temperatures .
the g - band images over a period of 5 hours were co - aligned using the adjacent image shifts to determine their light curve .
the integrated flux was measured following the same procedure .
since the extracted image for gaussian surface fitting is larger with 20 pixels , the inward migration of umbral dots , which is at most about three pixels , does not contribute to the flux measurement .
figure 7 show the typical light curves of various locations .
the spectropolarimeter ( sp ) observations of the sunspot close to the imaging observations , obtained on 28 april 2007 beginning at 09:00 ut , are used to determine the magnetic and velocity properties of the bright features . by using the inversion of radiative transfer equation for all components of polarized spectrum
, we obtained the magnetic and velocity over the sunspot from the sp data .
the inversion code uses milne - eddington atmosphere for producing the spectral lines .
the magnetic and velocity field maps and the degree of polarization in the spectral lines in the sunspot are shown in figure 8 .
the reference velocity ( zero velocity ) was obtained by averaging the values of the pixels in the umbra that did not contain bright points .
we notice the enhanced magnetic and velocity compared to their surrounding .
we obtained the magnetic field and velocity of several common identified points of images by following a similar feature of flux extraction by two dimensional gaussian function fitting .
the compact bright points in the continuum image of 6301.5 were close to the coordinates of bright points in the red continuum image .
figure 9 show the filling factor of the bright points that are common in sp and imaging observations plotted along with the g - band bright point intensities in the same scale .
other properties such as magnetic field , doppler velocity , field inclination and degree of polarization of the bright points derived from the sp observations are shown in figure 10 .
there are two classes of theoretical sunspot structures that can be used to build contemporary models to explain the formation of umbral dots .
one of them is developed with the hypothesis that the subsurface magnetic field of a sunspot contains several separate flux tubes filled with magnetic field free fluid between the gaps . through penetrative
convection the field free gas oscillates or occasionally penetrates through the surface appearing as bright points @xcite .
the other models umbral dots as the result of rising plumes triggered by the oscillatory magnetoconvections @xcite . in both these models the up flowing plasma at the bright dots has higher density with reduced blackbody temperature compared to the quiet sun . in case of penetrative convection
the fluid is magnetic field free , where as in the case of magnetoconvection of monolithic sunspots , the umbral dot field is expected to be mostly in the vertical direction .
the plasma columns with lower field compared to the surrounding embedded in a flux concentration will also produce the bright points @xcite . in the following
, we shall discuss our results in context of these theoretical models .
the ratio of sunspot image obtained through g - band filter and caii h filter is shown in figure 2 that shows bright points in the umbra , penumbral boundary and in the penumbra . @xcite
have shown that for the sot / bfi g - band filter that includes a broad upper photospheric contribution samples about 74 km , where as caii h filter samples higher region at the height of 247 km of solar atmosphere .
these heights are derived using the quiet sun atmospheric models @xcite . in sunspots ,
the atmospheric model will change significantly , the contribution in caii h would be at a higher atmospheric level compared to that of g - band .
the sunspot atmospheric models show that the density and pressure has a higher gradient in sunspots compared to the quiet sun @xcite .
the contribution height between the two lines in the umbra might be smaller than 173 km , which is less than the height of the magnetic features obtained earlier @xcite . in the penumbra , however , the caii formation height could be higher than in the umbra @xcite .
the ratio images show peripheral umbral dots are mostly located at the foot points of the origin of penumbral fibrils and much brighter than the umbral dots .
these bright points are also seen to make inward migration with time .
however , with in the time gap of images used for the ratio images , which is about 3 seconds , the motion would result the displacement within a pixel of the detector without affecting the contrast of these points .
these points may be considered as the penumbral foot points as they are located mostly at the origin of penumbra .
we notice most umbral dots are not completely circular but has a slightly asymmetry directed towards the penumbra as shown in figure 3a .
the ratio of elliptical axis of gaussian surface is about 1.1 .
some of the umbral dots , specially located towards the umbral boundary , display elongated filamentary structures pointed towards the penumbra as shown at the location with arrow mark in contrast enhanced image in figure 2 . such elongated features are reported earlier @xcite .
similar to the radial filamentary structures discovered by @xcite , many umbral dots appear to be linked with magnetic structures , contrary to the expectation of theoretical models that consider them only due to the rising plasma columns in sunspots .
the dots are not uniformly distributed in the umbra and are more concentrated at that part of the umbra that decayed sooner .
we have examined at least 16 sunspot ratio images . in all these sunspots
the umbral dots located at the foot points of the penumbra towards the disk center appear brighter compared to their limb side counter part .
such effect would be expected if the bright points represent inclined magnetic structures in which the limb side has lower depth compared to the ones towards the center of the sun .
earlier observations have also showed the similarity between the umbral dots and peripheral umbral dots and non - uniform distribution within the umbra @xcite .
these imaging observations suggest that the bright points in the umbra and its periphery may not be only due to the protrusion of non magnetic material from below the visible surface , but are actually parts of rapidly diverging inclined flux tubes similar to the ones that make penumbra .
the ratio images show that the penumbra is divided into two parts at midway between the umbra and quiet sun .
the inner penumbra has higher ratio compared to the outer penumbra .
this is roughly the location where the inward and outward migration of penumbral grains are observed @xcite .
this division of brightness of penumbra show the temperature stratification with height is different in two parts of the sunspot .
the temperature distribution leads to the photospheric layer of inner sunspot hotter , whereas in the outer part the lower chromosphere is hotter contributing more flux through the caii h filters .
the penumbral grains are brighter and elongated with the ratio of axis of 1.4 and few with higher values as shown in figure 3b .
most theoretical models describe penumbra as horizontal flux tubes embedded in field free plasma with gaps between them similar to the parker s model of subsurface magnetic structure of sunspots @xcite . in these models
the bright features are considered as the hotter field free material surrounding the flux tubes carrying the cooler plasma .
some times they are also considered as the view of photosphere through the gaps of penumbral flux tubes as first reported by @xcite . while , it is not possible to distinguish between the photospheric view and in - situ hot plasma around the flux tube , we notice that the bright features are elongated and aligned at the middle portion of the penumbral features indicating that they are related to the geometry of the flux tubes .
there are other mechanisms such as localized reconnections that can heat the magnetized plasma due to the magnetic reconnection processes as primarily observed in caii h images @xcite .
such contribution we would observe as dark features in outer penumbra in the ratio images . however , if the reconnection heats the plasma at lower heights in the flux tubes the features would be elongated .
the bright points observed in the penumbra are long lived unlike the reconnection events that has shorter life time .
the size distribution derived from the ratio images of the bright points is bimodal peaking below 0.2 arc second and near 0.5 arc sec as shown in figure 4 .
the smaller bright points are umbral dots similar to the features observed by @xcite .
the larger ones are peripheral umbral dots or penumbral foot points .
the sample with large size than 0.5 arc second are the penumbral grains .
the photometry measurement of the bright point flux through blue ( 4505 ) , green ( 5550 ) and red ( 6684 ) continuum filters along with the planck functions for the temperature range of 4600 to 6600 k are shown in figure 6 .
the figure shows that there is a gradual increase of temperature from cool umbral dots to the hotter g - band bright point in the vicinity of the sunspots with in range of 2000 k. earlier spectroscopic observations showed that the sunspot umbra mainly consists of a two component atmosphere @xcite .
the cold component has temperature of @xmath4 4000 k , where as the hot component temperature is @xmath4 5700 k. the hot component occupies about 10% of the area and contribute to the 50% of the radiation .
our observations show that while these temperatures are consistent with the hot component umbra derived using the spectroscopic diagnostic , we find that the hotter component actually is due to the umbral dots located at the boundary of penumbral fibrils @xcite . while the umbral dots are cooler , the penumbral grains and outside bright points are hotter .
consistent with the blackbody character of the bright points , the continuum flux through green and red filter change linearly with respect to the blue continuum flux as shown in figure 5 .
we also notice that the g - band brightness of the points surrounding the sunspots represented by the blue plus sign are brighter when located near the solar limb compared to the disk , a trend similar to the quiet sun g - band bright points . as shown in figure 5 , the flux measured through the filter in molecular bands ( cn and g - band )
does not show the trend for black body character of the bright points in which case the corresponding values would have been similar to continuum flux trend . instead of a linear relationship between the blue ( @xmath5 ) and g - band ( @xmath6 ) flux , we find the non linear dependence represented by the dashed line following the equation 2 .
@xmath7 ( 2 ) where a and b are constants .
the theoretical modeling of g - band bright points are focused to explain the measurements carried out the quiet sun @xcite . in quiet sun , the g - band bright points are mostly associated with the magnetic concentrations that has elevated temperature leading to the dissociation of primarily ch molecules . at these wavebands the opacity remains unchanged compared to the neighboring continuum opacity that is due to h@xmath0 . as a result
the deep photospheric layers of hotter flux tubes are exposed through the ch - wavelength window @xcite .
the flux tube geometry plays an important role , where the evacuated magnetic elements for maintaining the pressure balance with the surrounding exposes the hotter layers of photosphere that is enhanced for inclined flux tubes leading to higher brightness near the solar limb compared to their disk counterpart @xcite . here
also we notice the enhanced g - band brightness of circum - sunspot bright points compared to their disk counterpart ( figure 5a ) .
however , the bright points in the sunspot do not show such disk - to - limb variation trend of the brightness .
this behavior may be expected from already inclined flux tubes that would show the g - band brightness with an upper limit . in that case ,
our observation imply that unlike their quiet sun counterpart , the sunspot bright points would represent inclined flux tubes , which is consistent with the magnetic field inclination values obtained from sp observations shown in figure 10 , .
the g - band spectral synthesis with model solar atmosphere show that the g - band intensity does not vary in a linear fashion with respect to the continuum intensity and depends on the magnetic properties of the location @xcite .
we observe a similar trend for cn band as shown in figure 5b . in caii h filter
, we also observe a similar trend with more scatter shown in figure 5c . in case of caii
h the enhanced ionization in hot environment could result a similar situation as molecular dissociation alter the opacity of the region .
the non linear dependence of g - band flux as well as flux measured through cn and caii h filters could be due to the magnetic properties of the bright locations in the sunspots as discussed later .
the bright point light curves in figure 7 show that the large contrast enhancements observed in umbral dots , peripheral umbral dots and umbral grains compared to the circum - sunspot bright points .
the life time of bright points are in the range of 15 minutes to about one hour .
this is in agreement with several earlier measurements @xcite .
the higher contrast in the umbral dots compared to the g - band bright points could be due to the intense heating of the cold gas in side the spot compared to the ones outside . inside the sunspot , the material is generally cold , isolated by the magnetic field from the hot plasma in the solar convection zone .
the physical mechanism that can heat the material does so for a limited time exposing the hot material underneath .
the g - band bright points outside the sunspot are generally embedded in a hot environment .
the external heat input does not have a dramatic effect as in case of sunspot material .
the total magnetic flux , doppler velocity and the degree of polarization map derived from the sp observations are shown in figure 8 .
while the magnetic field map show that the bright points corresponding to higher magnetic field compared to the surroundings , the polarization map show enhanced polarization in agreement with earlier results @xcite .
both these observations indicate the magnetic nature of the bright points .
while the magnetic field are enhanced at these location of bright points that appear as blobs in the magnetograms , the field at the center of these blobs appear to be some what lower than the surroundings .
the line of sight velocity map show continuous channels of enhanced velocity elongated towards the penumbra starting at some of the bright points .
the quantitative magnetic properties of the bright points are presented in figure 9 and 10 .
the relationship between the magnetic filling factor and the g - band flux of bright points derived from the nearly co - temporal observations of bfi and sp observations are shown in figure 9 in the same scale .
we measure the flux at the bright point in g - band and their corresponding red continuum using the bfi images .
similarly , the flux of the bright points at the continuum wavelength of 6301.5 and their filling factor from inverted sp data are extracted .
these are the two different data sets that are used to determine the continuum flux of sunspot bright points in red wavelength , g - band intensity and the magnetic filling factors .
although the spatial coordinates of the bright points in bfi and sp data are close , we are not sure if they are the same objects due to the temporal gap between two observations .
therefore , we plot the g - band flux and magnetic filling factor as a function of red continuum and 6301.5 wavelength respectively to examine the general trend .
we notice a remarkably tight relationship between the magnetic filling factor with g - band bright point intensity where the magnetic features with smaller filling factor are brighter .
although the sot on hinode has very small scattered light contribution , it would effect most for the umbral dots resulting their lower value indicated by the arrow in figure 9 .
the reduced filling factors are expected for the flux tube that `` flare out '' more @xcite leading to the exposure of larger cross - section of photosphere devoid of ch molecules .
figure 10a show that the brighter points are of lower magnetic flux compared to the darker counterpart , which are mostly umbral dots similar to the earlier results @xcite .
however , our the doppler velocity measurements also show smaller values for the brighter spots compared to their darker counterpart in the sunspot ( figure 10b ) .
the brighter points situated in outer penumbra and outside the sunspot also display large downward velocity values as represented by six points shown in a circle in figure 10b .
these locations could harbor the evershed flow material draining along the flux tube .
although , we shall discuss the velocity properties of the bright sunspot features in a later paper , here we confine the comparison with the velocity measurements having high spatial resolution observations .
several velocity measurements in different wavelength show diverse properties of umbral dots and penumbral grains .
while some penumbral grains showed upflows , the umbral dots displayed large down flows @xcite .
the relative red shift velocities are observed at these points indicates the presence of down flowing plasma @xcite .
the proper motion studies show the bright features moving towards the umbra @xcite . the ground based high resolution observations using adoptive optics by @xcite showed the the bright points displaying @xmath8 300 m s@xmath0 , with brighter ones near zero velocity .
however , the hinode observations also showed relative blue shifts indicating a upward flow plasma with respect to the surrounding at these locations @xcite .
the observations with la palma stokes polarimeter show very small upflows with velocity of about 100 m s@xmath0 .
hinode sp data inversion also find downflow velocities at higher photosphere @xcite .
the velocity measurements , being relative to the surrounding , they are effected by the choice of reference velocity region .
some of the discrepancy of the velocity measurements could arise due to the selection of the reference region .
our velocity measurements are derived from the stokes profile inversion that gives doppler velocity measured from the line shift .
the velocity of the umbral region devoid of bright features was taken as the reference , with respect to which all except two penumbral bright points show down - flow velocities .
figure 10c show a wide range of the field inclination for the brighter points with smaller range for umbral dots .
the horizontal field are found at these locations in earlier measurements @xcite .
combined with the discovery of association of elongated features , the flatter field of these locations indicate the presence of inclined flux tubes at the sunspot bright points .
the degree of polarization is higher at these location of bright points with decreasing trend for brighter points as shown in figure 10d .
as the polarized signal is essentially due to the magnetic field , the bright features are likely the site of magnetic flux tubes instead of field free plasma .
the observational properties of the sunspot bright points , presented in section 3 , show their magnetic structure . while several properties of the bright points agree well with earlier results ,
our g - band flux measurements of four sunspots , indicate that these locations are showing properties of flaring flux tubes " as described by @xcite .
we also notice that careful examination of the umbral bright points in the sunspots show association of elongated filamentary structure extending towards the penumbra similar to the features reported earlier @xcite .
following the theory for the umbral dots , originally proposed by @xcite , we interpret the results with a model as illustrated figure 11 . the dark vertical lines with divergence at the top in the umbral dot in figure 11 a represent the flaring magnetic field " above the visible solar surface shown by the dashed line .
the diverging magnetic flux tubes have gas pressure decreasing in downward direction as the magnetic field increases leading to higher magnetic pressure .
the inter flux tube space is filled with field free hotter plasma , compared to that inside tube , penetrating through the gaps in the subsurface clustered flux tube bundle from the quiet sun .
as the plasma column is trapped between the magnetic field , they would be subjected to magnetic instabilities in a similar fashion as described by @xcite .
the hotter oscillating plasma will heat the adjacent flux tubes through heat exchange @xcite .
the heated plasma would eventually flow downwards following the pressure gradient in the flux tubes .
however , if these oscillating plasma column were manifesting as the umbral dots , than we would have observed the g - band intensity consistent with the blackbody curve in which the brightness increase depends linearly with blue band intensity .
but , since we observe that the g - band intensity is nonlinearly dependent on blue band intensity in a manner that would be expected from the flux tubes as described by @xcite , we suggest that the g - band bright points correspond to the diverging flux tube locations adjacent to the oscillating magnetic field free plasma column . since these are the same points that appear brighter in the continuum wave bands , the umbral dots represent the photospheric interaction of the diverging flux tubes that show the magnetic properties described in the previous section .
the penumbral foot points are more inclined and located at the edge that is in contact with the field free convective zone material .
the flux tube plasma at these locations will receive more heat flux from the adjacent field free convective zone plasma leading to the heating of a large cross - section of hot wall in the line of sight .
this effect would result their higher brightness appearance compared to the bright points inside the umbra .
the sunspot has a gradually lower elevation from out side to the center resulting the wilson depression .
the same effect would enable viewing a larger cross - section of hot wall of the flux tube leading to the brighter appearance of sun center side penumbral foot points . in this model
the penumbral grains are due to the viewing of a larger cross section of inclined flux tubes embedded in field free plasma @xcite .
the localized heating could arise either by in - situ heating by magnetic reconnection @xcite or other plasma processes @xcite .
the g - band bright points in the vicinity of the sunspot are similar to that described in @xcite .
the analysis of high spatial resolution imaging and spectropolarimetry observations of sunspots show the following properties of their bright points .
( a ) the bright points are non - circular and associated with elongated filamentary structures .
( b ) the g - band intensity deviates from the expected blackbody temperature.(c ) the locations of bright points display magnetic properties .
these properties can be best understood if the subsurface structure of sunspots are porous with intersprsed convective zone plasma .
understanding the stability and decay of the sunspots is a major problem of modern solar physics .
the properties of bright points in and around the sunspot is a crucial part of the sunspot structure playing important role in this processes .
our study shows that the bright points are mostly magnetic concentration compared to their surroundings .
the properties are compatible with the description in which the bright points represent the locations inside the diverging flux tubes instead of the inter - flux tube gaps as proposed by @xcite . if we consider them as purely due to magneto - convection the nonlinear relationship between the blue and g - band brightness can not be explained . in this case
the upwelled regions of umbra would still show blackbody flux distribution .
also , the brightness asymmetry observed in sun center side compared to limb side may not be consistent . in conclusion
, the location of bright points at the photospheric intersection of the diverging umbral flux tubes seems to be the natural model that accommodates most of the observed features .
one of the authors dpc is grateful the institute of space and astronautical sciences for the hospitality during this work .
we would like to thank dr .
eugene parker for stimulating discussions .
a part of support comes from nsf under grant atm 05 - 48260 and nasa grant nnx 08aq32 g .
the data was obtained using hinode solar optical telescope .
hinode is a japanese mission developed and launched by isas / jaxa , with naoj as domestic partner and nasa and stfc ( uk ) as international partners .
it is operated by these agencies in co - operation with esa and nsc ( norway ) .
beck , c. 2008 , , 480 , 825 beck , c. , bellot rubio , l. r. , schlichenmaier , r. & stterlin , p. 2007
, , 472 , 607 beckers , j. m. & schrter , e. , h. 1969 , solar phys . , 10 , 384 berger , t. e. , lfdahl , m. g. , shine , r. s. , title , a. m. 1998 , , 495 , 973 carlsson , m. , hansteen , v. h. , de pontieu , b. , mcintosh , s. , tarbell , t. , shine , d. , tsuneta , s. , katsukawa , y. , ichimoto , k. , suematsu , y. , shimizu , t. & nagata , s. 2007 , pub .
japan , 59 , s663 chandrasekhar , s. 1961 , hydrodynamic and hydromagnetic stability ( oxford : clarendon press ) choudhuri , a. r. 1981 , , 302 , 809 choudhary , d. p. , balasubramaniam , k. s 2007 , 664 , 1228 degenhardt , d. & lites , b. w. 1993 , , 404 , 383 degenhardt , d. & lites , b. w. 1993 , , 416 , 875 del toro iniesta , j. c. 2001 , magnetic field across the hertzsprung - russel diagram , solanki , s. k & wickramasinghe , asp conf .
248 , 35 denker , c. , tritschler & a. , deng , n. 2008 , astron .
nachr , 7 , 1 deng , n. , choudhary , d. p. , tritschler , a. , denker , c. , liu , c. , wang , h. , 669,1372 ding , m. d. & fang , c. 1989 , , 225 , 204 fontenla , j. m. , avreet , e. h. , & loeser , r. 1990 , , 355 , 700 garcia la de rosa , j. i. 1987 , solar phys , 112 , 49 hamedivafa , h & sobotka , m. 2004 , 428 , 215 hartkorn , k. 2003 , current theoretical models and high resolution solar observations , asp conf .
proc , pevtsov , a. a. , & uitenbroek , h. , 286 , 281 hienmann , t. , nordlund , , scharmer , g. b. & spruit , h. c. , 2007 , , 669 , 1390 jhan , k. 1981 , in physics of sunspots , j. h. thomas , n. weiss ( eds ) , cambridge university press , 139 kneer , f. 1973 , solar phys . , 28 , 361 knobloch , e. , weiss , n. o. 1984 mon .
, 207 , 203 katsukawa , y. , berger , t. e. , ichimoto , k. , lites , b. w. , nagata , s. , shimizu , t. , shine , r. a. , suematsu , y. , tarbell , t. d. , title , a. m. & tsuneta , s. 2007 , science , 318 , 1594 kitai , r. 1986 , solar phys , 104 , 287 kosugi , t. , matsuzaki , k. , sakao , t. , and 15 coauthors 2008 , solar phys . , 243 , 3 kusoffsky , u. & lundstedt , h. 1986 , , 160 , 51 livingston , w. 1991 , nature , 350 , 45 lites , b. w. , bida , t. a. , johannesson , a. , scharmer , g. b. 1991 , , 373 , 683 malcolm w. e. 1992 , solar phys . , 137 , 215 makita , m. 1963 , pub .
japan , 15 , 145 moore , r. l. 1981 , 249 , 390 moore , r. l. 1981 , the physics of sunspots , l. e. cram and j. h. thomas , sacramento peak observatory , sunspot , new mexico , 259 nakawaga , y. 1957 , proceedings of the royal society of london , series a , mathematical and physical sciences , 240 , 108 obridko , v. n. & staude , j. 1988 , , 189 , 232 parker , e. n. 1979 , , 230 , 905 parker , e. n. 1979 , , 234 , 333 proctor , m. r. e. & wise , n. o. 1982 , rep .
phys . , 44
, 1317 rempel , m. , schssler , cameron , r. h. & knlker , m. 2009 , science , 325 , 171 riethmller , t. l. , solanki , s. k. & lagg , a. 2008 , , 678 , l157 riethmller , t. l. , solanki , s. k. , zakharov , v. & gandorfer , a. 2008 , , 492 , 233 rimmele , t. , r. 2004 , , 604 , 906 rimmele , t. , r. 2008 , , 678 , 684 rimmele , t. , r. 1994 , , 290 , 972 roupee van der voort , l. h. m. 2002 , , 389 , 1020 roupee van der voort , l. h. m. 2003 , , 397 , 757 ruiz cobo , b. & bellot rubio , l. r. 2008 , , 488 , 749 rutten , r. j. , kiselman , d. , va der voort , l. r. 2001 , advanced solar polarimetry - theory , observation and instrumentation , m. sigwarth , astron .
conf . series .
236 , 445 snches almeida , j. , asensio ramos , a. , trujillo buneo j. & cernicharo , j. 2001 , , 555 , 978 scharmer , g. b. & spruit , h. c. 2006 , , 460 , 605 schlichmaier , r. & solanki , s. k. 2003 , , 411 , 257 schmidt h. u. 1991 , geophysics astrophys .
fluid dyn . , 62 , 249 schssler & vger , a. 2006 , , 641 , l73 shimizu , t. , kubo , m. , tarbell , t. d. , berger , t. e. , suematsu , y. , ichimito , k. , katsukawa , y. , elmore , d. & lites , b. 2007 , new solar physics with solar - b mission , asp conference ser . , 369 , 51 sobotka , m. , bonet , j. a. & vzquez , m. 1993 , , 415 , 832 sobotka , m. , & jurk , j. 2009 , , 694 , 1080 sobotka , m. & puschmann , k. g. 2009 , , 504 , 575 socas - navarro , h. , martnez pillet , v. , sobotka , m & vzquez , m. 2004 , , 614 , 448 spruit , h. c. 1981 , in physics of sunspots , l. e. cram , j. h. thomas ( eds ) , national solar obs .
, sunspot , nm , 98 spruit , h. c. & scharmer , g. b. 2006 , , 447 , 343 staude , j. 1981 , , 100 , 284 stetson , p. 1987
paci . , 99 , 191 steiner , o. , hauschildt , p. h. & bruls , j. 2001 , , 372 , l13 steiner , o. , bruls , j. & hauschildt , p. h. 2001 , advanced solar polarimetry - theory , observation and instrumentation , m. sigwarth , astron .
conf . series .
, 236 , 453 thomas , j. h. , weiss , n. o. , tobias , s. m. , brummell , n. h. 2002 , nature , 420 , 390 tsuneta , s. , ichimoto , k. , katsukawa , y. and 22 coauthors 1973 , solar phys .
, 249 , 167 watanabe , h. , kitai , r. & ichimoto , k. 2009 , , 702 , 1048 watanabe , h. , kitai , r. , ichimoto , k. & katsukawa , y. 2009 , publ .
japan , 61 , 193 westendrop plaza , c. , del toro iniesta , ruiz cobo , b. , martniz pillet , v. , lites , b. w. , skumanich , a. 2001 , , 547 , 1130 westendrop plaza , c. , del toro iniesta , ruiz cobo , b. , martniz pillet , v. 2001 , , 547 , 1148 wiehr , e. & degenhardt , d. 1993 , , 278 , 584 wehrli , c. 1985 extraterrestrial solar spectrum " , publication number 615 , physikalisch - metrorologisches observatorium + worle radiation center ( pmo / wrc ) davos dorf , switzerland ccrrrrrrrrcrl 2006 november 14 & 10940 & s10e35 & @xmath9/@xmath10 & 500 + 2007 february 28 & 10944 & s05w07 & @xmath11 & 640 + 2007 february 02 & 10923 & s06w03 & @xmath10 & 100 + 2007 april 28 & 10953 & s04w20 & @xmath11 & 290 + | we used the flux calibrated images through the broad band filter imager and stokes polarimeter data obtained with the solar optical telescope onboard the hinode spacecraft to study the properties of bright points in and around the sunspots .
the well isolated bright points were selected and classified as umbral dot , peripheral umbral dot , penumbral grains and g - band bright point depending on their location .
most of the bright points are smaller than about 150 km .
the larger points are mostly associated with the penumbral features .
the bright points are not uniformly distributed over the umbra but preferentially located around the penumbral boundary and in the fast decaying parts of umbra .
the color temperature of the bright points , derived using the continuum irradiance , are in the range of 4600 k to 6600 k with cooler ones located in the umbra .
the temperature increases as a function of distance from the center to outside . the g - band , cn - band and caii h flux of the bright points as a function of their blue band brightness increase continuously in a nonlinear fashion unlike their red and green counterpart .
this is consistent with a model in which the localized heating of the flux tube deplete the molecular concentration resulting the reduced opacity which leads to the exposition of deeper and hotter layers .
the scatter in caii h irradiance is higher compared to the g - band and cn - band irradiance .
the light curve of the bright points show that the enhanced brightness at these locations last for about 15 to 60 minutes with the least contrast for the points out side the spot .
the umbral dots near the penumbral boundary are associated with elongated filamentary structures .
the spectropolarimeter observations show that the g - band brightness closely follows their magnetic filling factor . as the filling factor decreases the brightness increases .
generally , the umbral dots have higher magnetic field and larger doppler velocity compared to their counterpart outside the spot .
the flux tubes " in the umbra are less inclined compared to the ones associated with the penumbra .
we discuss these observation in the framework of sunspot models that include the creation of umbral dots by the plasma trapped in between the magnetic flux tubes .
the upward intruding plasma through the nonmagnetic columns between the fluxtubes of sunspot produce the bright points and heat the matter inside of adjacent tubes .
the heated plasma flows in the direction of reduced gas pressure . similar localized heating of penumbra leads to the origin of penumbral grains .
the g - band bright points are mostly the vertical flux tubes embedded in photospheric material .
the varying degree of g - dand , cn - band and caii h brightness of these points are due to the combined effect of different degree of heating and exposure of hot flux tube walls for viewing deeper photospheric layers . |
several candidates for a constructive definition of superstring theory have recently attracted attention . here
we focus on the ikkt model ( or iib matrix model ) @xcite , which is supposed to correspond to iib superstring theory in the large @xmath0 limit . that correspondence is supported by some analytical arguments .
this matrix model is formally obtained from ordinary super ym gauge theory in the zero volume limit ( one point ) . here
we study directly the large @xmath0 dynamics of large @xmath0 reduced matrix models .
some results were obtained before for the `` bosonic case '' ( where the fermions are dropped by hand ) , but we now want to address mainly the susy case .
in particular we simulate the 4d counterpart of the ikkt model @xcite we denote it as 4d iib matrix model which corresponds again to the dimensional reduction of 4d super ym gauge theory .
this model has also been studied analytically @xcite and numerically in the framework of dynamical triangularization @xcite . here
we report on direct monte carlo simulations using the hybrid r algorithm @xcite .
conclusive results can be obtained because in the 4d version the fermion determinant is real positive in contrast to the 10d ikkt model , where simulations would be plagued by a sign problem .
we got away with a computational effort of @xmath1 in the susy case , and of @xmath2 in the bosonic case .
the 4d iib matrix model is given by @xmath3^{2 } \nonumber \\ s_{f } & = & -\frac{1}{g^{2 } } \ , { \rm tr } ( \bar \psi_{\alpha } \gamma^{\mu}_{\alpha \beta } [ a_{\mu},\psi_{\beta}])\end{aligned}\ ] ] where @xmath4 ( @xmath5 ) are complex , traceless @xmath6 matrices , and the @xmath7 ( only ) are hermitean .
we use @xmath8 .
in addition to susy and so(4 ) invariance , this model has a su(@xmath0 ) symmetry , which is inherited from gauge invariance .
the first question about this model is if it is well - defined as it stands .
since the integration domain of @xmath9 is non - compact , divergences are conceivable .
however , our results , as well as results on a number of special cases @xcite confirm consistently that this model _ is _ well - defined at any @xmath0 ; there is no need to impose a ir cutoff .
this implies that the only parameter @xmath10 is simply a _
scale parameter _ , rather than a coupling constant .
it can be absorbed by introducing dimensionless quantities .
the challenge is , however , to tune @xmath10 as a function of @xmath0 so that the correlators are finite at @xmath11 , see sections 4 and 5 .
in the iib matrix model , the space coordinates arise dynamically from the eigenvalues of the matrices @xmath7 @xcite . in general the latter
can not be diagonalized simultaneously , which implies that we deal with a non - classical space .
we measure its uncertainty by @xmath12 \ , \ ] ] and the `` maximizing '' matrix @xmath13 is also used for introducing the coordinates of @xmath0 points , @xmath14 what we are really interested in is their pairwise separation @xmath15 , and we show the distribution @xmath16 in fig . 1 .
we observe @xmath17 at short distances ( @xmath18 ) , hence a uv cutoff is generated _
dynamically_. we also see that increasing @xmath0 favors larger values of @xmath19 . to quantify this effect we measure
the `` extent of space '' diverges , since @xmath20 at large @xmath19 ( see second ref . in @xcite ) . ]
@xmath21 fig .
2 shows @xmath22 and @xmath23 as functions of @xmath0 ( at @xmath24 ) . the inclusion of fermions enhances @xmath22 and suppresses @xmath23 , keeping their product approximately constant .
the lines show that both quantities follow the same power law , @xmath25 , in susy and in the bosonic case . in susy
this behavior is consistent with the branched polymer picture : there one would relate the number of points as @xmath26 , where @xmath27 is some minimal bond , which corresponds to the above uv cutoff .
the hausdorff dimension @xmath28 then reveals consistency with our result .
we define the polyakov loop @xmath29 and the wilson loop @xmath30 which is conjectured to correspond to the string creation operator as @xmath31 of course the choice of the components of @xmath7 is irrelevant , and the parameter @xmath32 can be considered as a `` momentum '' .
now @xmath33 has to be tuned so that @xmath34 remain finite as @xmath35 .
this is achieved by @xmath36 which leads to a beautiful large @xmath0 scaling ; fig .
3 shows the invariance of @xmath37 for @xmath38 in susy .
also the bosonic case scales accurately @xcite . the historic 2d eguchi - kawai model @xcite had a @xmath39 symmetry , which implied @xmath40 , a property which was crucial for the proof of the eguchi - kawai equivalence to gauge theory .
as we see , this property is not fulfilled here , but @xmath41 falls off rapidly , towards a regime where the assumption of this proof holds approximately .
we proceed to a more explicit test of eguchi - kawai equivalence by checking the area law for @xmath42 .
4 shows that the area law seems to hold in a finite range of scale .
remarkably , the behavior is very similar @xcite in the bosonic case .
there we further investigated the behavior at much larger @xmath0 @xcite , and we observed that the power law regime does neither shrink to zero as it was generally expected nor extend to infinity a scenario which seems possible from fig .
4 . at least in the bosonic case
its range remains finite at large @xmath0 .
recently the same behavior was observed in a study of the 10d bosonic case @xcite .
we now consider connected multipoint functions @xmath43 , @xmath44 being a polyakov or a wilson loop .
we wonder if it is possible to renormalize all of those multipoint functions simply by inserting @xmath45 , so that a single factor @xmath46 renders all functions @xmath47 ( simultaneously ) finite at large @xmath0 .
it turns out that such a universal renormalization factor seems to exist in susy .
we have to set again @xmath48 , and then @xmath49 provides large @xmath0 scaling , as we observed for a set of 2 , 3 and 4-point functions .
two examples are shown in fig .
our observation can be summarized by the susy rule @xmath50 this implies that large @xmath0 factorization holds , @xmath51 , as in gauge theory , although coupling expansions are not applicable here . for the bosonic case
, a @xmath52 expansion @xcite suggests large @xmath0 factorization to hold as well , but it also predicts @xmath53 .
this is confirmed numerically @xcite : in particular the 3-point functions shown in fig .
5 now requires @xmath54 .
therefore no universal renormalization factor @xmath46 exists in the bosonic case , which is an important qualitative difference from the susy case .
we also performed simulations in @xmath55 @xcite , where we simplified the ikkt model as follows : \(1 ) we use a 1-loop approximation , which is expected to capture the low energy dynamics .
this amounts to an effective action , keeping track of off - diagonal elements only to the quadratic order , in the spirit of ref .
@xcite .
\(2 ) we omit the phase of the pfaffian by hand , in order to avoid the sign problem .
thus a monte carlo study becomes feasible .
the validity of ( 1 ) is supported by our results for @xmath22 , but ( 2 ) is certainly a drastic step . still one could hope to observe basic properties of the ikkt model at least qualitatively .
these simplifications allow for a simulation effort of only @xmath2 .
our main interest here is if the eigenvalue distribution of the @xmath7 indicates a spontaneous symmetry breaking ( ssb ) of so(10 ) invariance , as it was suggested in the formulation of the ikkt model @xcite . to this end
, we consider the moment of inertia @xmath56 ( @xmath57 ) , and we measure its 10 eigenvalues
. a gap in this spectrum would indicate the ssb of lorentz symmetry .
however , this can not be observed , even though we raised @xmath0 up to 512 . on the contrary ,
we observe that the eigenvalue distribution becomes more and more isotropic as @xmath0 increases , see fig . 6 , and the same is true in @xmath58 @xcite .
we conclude that if ssb of lorentz symmetry occurs in the ikkt model , then it must be driven by the imaginary part of the action .
we first simulated the 4d iib matrix model , both , susy and bosonic .
we varied @xmath0 up to 48 , which turned out to be sufficient to study the large @xmath0 dynamics .
we confirmed that the model is well - defined as it stands , hence @xmath10 is a pure scale parameter .
the space coordinates arise from eigenvalues of the bosonic matrices @xmath7 .
the extent of space follows a power law . in susy
this agrees with the branched polymer picture .
fermions leave the power unchanged but reduce the space uncertainty though it remains finite at large @xmath0 .
the large @xmath0 scaling of polyakov and wilson loops and their correlators requires @xmath48 in susy and in the bosonic case , but the wave function renormalization is qualitatively different : only in susy a universal renormalization exists .
the area law for wilson loops holds in a finite range of scale for the susy and the bosonic case .
the latter comes as a surprise , and we checked up to rather large @xmath0 that this range remains indeed finite .
hence eguchi - kawai equivalence to ordinary gauge theory @xcite may hold in some regime . | we simulate a supersymmetric matrix model obtained from dimensional reduction of 4d su(@xmath0 ) super yang - mills theory ( a 4d counter part of the ikkt model or iib matrix model ) .
the eigenvalue distribution determines the space structure .
the measurement of wilson loop correlators reveals a universal large @xmath0 scaling .
eguchi - kawai equivalence may hold in a finite range of scale , which is also true for the bosonic case .
we finally report on simulations of a low energy approximation of the 10d ikkt model , where we omit the phase of the pfaffian and look for evidence for a spontaneous lorentz symmetry breaking . |
in previous microlensing models the rates and optical depth per star are usually estimated ( kiraga and paczyski , 1994 han and gould , 1995 , mollerach and roulet , 1996 ) .
below some limiting magnitude these stars can not be resolved individually and form a dense stellar background on all the surface of the images .
i will call these star `` background stars '' .
these background stars can be lensed , and this issue has already been investigated in the case of the lmc ( bouquet , 1993 ) , and for the galaxy ( nemiroff , 1994 ) . however
the stellar fields of the current microlensing experiments macho ( alcock _ et al .
_ 1996 , alcock _ et al .
_ 1993 ) , ogle ( udalski _ et al .
_ 1994 ) , duo ( alard 1995a , alard _ et al . _
1995b ) , eros ( aubourg _ et al .
are extremly crowded , so tha inside the resolution radius of a monitored stars , there are plenty of these unresolved background stars .
thus if one of these background stars is lensed , we will see a blended event .
the amplification , duration and impact parameter will be biased by the blending .
we may now wonder if their contribution to the total number of events observed is significant , compared to those expected from the resolved stars .
as the blending of a background star with a resolved star will be high we expect a large reduction of the effective einsten radius ( di stefano and esin , 1995 )
. consequently the observed number of events will be a competition between the decrease of the einsten radius , and the increasing number of stars at larger magnitudes .
also the effective duration of the event will be much shorter than it is for to events on resolved stars , and consequently it may produce a tail of short events .
these short events could be confused with low mass lens events , and especially lensing by brown dwarf , so that a detailed study of these events is particulary important .
another problem will be to recognise these blended events .
the blending of light produces a modified light curve which might be distinguished from the unblended light curve .
however to make an unambiguous separation a good time sampling is required , and a very accurate photometry is required . in the case of these short events , we expect that it will be difficult to achieve a sampling dense enough .
the study of lensing of unresolved stars is very sensitive to the shape of the adopted luminosity function . in previous works ( kiraga and pazcyski , roulet and mollerach , han and gould ) ,
the number of stars at a given luminosity @xmath0 is expressed as : @xmath1 , with usually @xmath2 .
a more realistic luminosity function can be derived from the holtzman ( holtzman _ et al .
_ 1993 ) observations of a field in baade window with the hst .
the final holtzman _ et al .
_ luminosity function can be approximated by three continuous straight segments with different slopes .
the first change in the slope occur at @xmath3 . with the extinction and distance to the galactic center adopted by holtzman _
, it correponds to a value of @xmath4 . .
the holtzman _
_ study agrees well with terndrup ( terndrup 1989 ) investigations of the galactic bulge which found a value of @xmath5 for the bulge turn - off at @xmath6 . in the holtzman luminosity function
the turn - off is situated around @xmath7 also .
the second change in the slope occur around @xmath8 and seeems to continue until the holtzman _
limit situated at @xmath9 . a comparison with the luminosity function of the globular clusters reviewed by mould ( mould 1996 ) indicates that the shape of their luminosity function is almost the same in this range of magnitude .
the mould diagram allows also to a better determination of the last segment of the luminosity function .
this straight segment seems to hold until @xmath10 where a possible turnover in the luminosity function is observed .
however @xmath10 is about 7 magnitudes fainter than the limiting magnitude of the current microlensing surveys , and such faint unresolved stars do not contribute to the optical depth .
the duration of the events on stars blended with a star 7 magnitudes brighter would be so short that it is not observable .
consequently the exact behavior of the luminosity function in this region is not important for our study , and we will make the reasonable choice to cut the luminosity function at @xmath10 .
looking now at the bright side of the luminosity function , we find that at magnitudes brighter than the turnoff , the slope is very steep indicating that the number of stars drops rapidly . in this region
the bulge stars are essentially sub giants .
the density rises again in the clump giant region , however these stars do not makes more than 5 percent of the stars in the current surveys .
consequently they have not a very significant influence on the total luminosity function .
these stars should be treated separately ( alard 1996 ) , and in the present study , we will ignore them .
this will put the bright end of the luminosity function at @xmath11 in the holtzman _
the adopted value for the distance to the galactic center in this study is 8.5 kpc .
the final adopted luminosity function is shown in figure 1 .
it is clear that , in an investigation of the lensing of unresolved stars , consistency requires that we calculate the rates and optical depth also with the same luminosity function .
this involves some slight changes from previous analysis .
therefore i will first investigate the case of resolved stars considering a more general form for the luminosity function .
i first calculate the change in the optical depth . the numbers of source stars at distance @xmath12 , and with absolute luminosity l can be written as : @xmath13 where k is a constant .
+ suppose that the experiment is able to find a total of @xmath14 stars up to a limit in apparent luminosity @xmath15 .
this leads to the following expressions : @xmath16 with : @xmath17 and : @xmath18 , where @xmath15 is the apparent limiting luminosity .
+ @xmath19 . where @xmath20 is defined as the saturation limit of the detector .
it can be expressed as : @xmath21 , where @xmath22 is the detector dynamic range , a reasonable value is about 6 magnitudes in crowded fields , either for ccd or photographic plates . the optical depth associated with these stars can be calculated as ( paczyski 1991 , kiraga and paczyski 1994 , hereafter refered as kp ) : @xmath23 with : @xmath24 and @xmath25 + to get the total optical depth we have now to integrate eq ( 1 ) over the source distances @xmath12 and the absolute luminosity @xmath0 of these stars .
+ @xmath26 with : @xmath27 and : @xmath28 here the factor @xmath29 is put in this expression to allow direct comparison with the standard optical depth formulae ( kiraga and paczyski 1994 ) : @xmath30 we see that the kiraga and paczyski expression is simply equivalant to taking @xmath31 at all distances .
a slight improvement is to take @xmath32 beyond a given distance ( roulet and mollerach , 1995 ) .
the plot in figure 2 allows a direct comparison between @xmath33 , and the approximation @xmath31 .
figure 2 illustrates the calculations both for duo and for the ogle experiment , which is about one magnitude deeper .
let s see now the changes in the total optical depth introduced by the @xmath33 function .
the optical depths are computed using the cobe bar model for the bulge ( dwek _ et al .
_ , .. ) , and the bahcall model for the disk ( bahcall and soneira , 1984 ) the ratio of the modified optical depth to the standard one is shown in table 1 .
the change is already important for ogle , and is more significant again for duo .
it is easy to understand that the lower optical depth of the new formulae is essentially due to the fact that due to the limiting magnitude , the experiments are more sensitive to sources in the near end of the bar , for which the density of lenses on the line of sight is smaller .
llll optical depth ratio & bulge lenses & disk lenses & total + ogle & 0.85 & 0.89 & 0.86 + duo & 0.75 & 0.8 & 0.77 + the formulae given for the lensing rates by kiraga and paczyski is : @xmath34 where : @xmath35 @xmath36}^{1/2 } \ dd_d \
dv_i\ ] ] @xmath37 is the relative velocity between lens and source , and the @xmath38 are the four components of this velocity . + now , it is straightforward to introduce the @xmath33 function in this expression , exactly as was done for the optical depth .
+ + this leads to : @xmath39 the result of the calculation of the @xmath40 and @xmath41 rates is shown in figure 3 for duo , and in figure 4 for ogle .
for all the figures the velocity dispersion given by han and gould ( han and gould 1995 ) was adopted .
the mass function is a salpeter mass function with a lower cut off at @xmath42 and an upper cut off at @xmath43 .
the einsten radius is defined as the distance for which the amplification of a microlensing event is a factor of 1.34 . in the case of an unresolved background star of luminosity @xmath15 , blended with a resolved star of magnitude @xmath44 ,
the background star amplification required to make a total amplification in the combined image of 1.34 is : @xmath45 hereafter , i will refer to @xmath46 as the blending factor : @xmath47 this gives the following expression for the blended einsten radius : @xmath48 with : @xmath49^{1/2}\ ] ] let us consider unresolved background stars with absolute luminosity l at a distance @xmath12 , whose number is dn(@xmath12,l ) , and let us suppose also that they are blended with stars of apparent luminosity @xmath50 .
the optical depth per resolved star associated with these stars can be calculated using equation ( 2 ) by replacing the einsten radius by the blended einsten radius . @xmath51
the expression for @xmath52 contains the factor @xmath53 , which takes into account that only those unresolved objects which are lensed close to a resolved star will be seen .
the value of @xmath53 will be close to the ratio of the area of the seeing disk , to the mean space occupied by a star on the image . + with : + @xmath54 ^ 2 $ ] , and @xmath55 .
+ + we have now to take into account that the luminosity @xmath50 of the resolved star is not constant , but comes from the luminosity function .
it was shown by zhao ( zhao 1995 ) using the ogle data that the number of resolved stars in baadeswindow can be well described by a power law , with an exponent close to @xmath56 . to be precise
we have a fraction of stars near luminosity @xmath50 : @xmath57 with : @xmath58 and @xmath59 + consequently , @xmath60 to get the total optical depth associated with the unresolved stars , we have now to integrate over the luminosity range of these stars , from the absolute luminosity ( @xmath61 ) to the faint end of their luminosity function ( @xmath62 ) .
in addition , we have also to integrate over all the possible distances for the sources , and all apparent luminosities for the blends .
+ + this leads to : @xmath63 @xmath64 + if now we uses eqs , ( 1 ) , ( 2 ) , ( 3 ) , we can express @xmath65 as : + + @xmath66 @xmath67 this expression can be integrated over the variables @xmath0 and @xmath50 , consquently : @xmath68 with : @xmath69 @xmath70 note that the dependance in distance is hidden in the variable @xmath71 , and also in the boundary @xmath61 .
+ + figure 5 illustrates the new @xmath72 function for the unresolved stars in the case of the duo and ogle experiments . note that these functiond reach their maxima at larger distances than the previous @xmath33 functions , it means that a large contribution will come from unresolved sources on the far side of the bar .
it is now possible to quantify the contribution of the unresolved sources to the total optical depth . in table 2
the ratio of the optical depth of unresolved sources to the the optical depth of resolved sources is indicated for duo and ogle .
the calculation of this ratio require an estimation of the r constant .
a crude estimate of r is the ratio of the surface covered by the resolution radius to the mean area occupied by a star . for the duo experiment
the resolution radius is close to 3 pixels , and the mean area occupied by a star is about 60 pixels ( alard and guibert 1996 ) .
this gives @xmath73 .
for ogle the resolution radius is probably close to the seeing value ( this is due to smaller pixels ) .
consequently with a mean seeing of @xmath74 in las campanas , and a pixel of @xmath75 we can estimate the resolution radius as 2.7 pixels .
the ogle experiment follows @xmath76 stars on 14 images of @xmath77 pixels each .
this gives again a value for r , of @xmath73 .
llll @xmath78 & bulge lenses & disk lenses & total + ogle & 0.58 & 0.55 & 0.57 + duo & 1.93 & 1.81 & 1.89 + the rates per resolved star for unresolved events can be calculated in the same way as for the optical depth .
+ we have : @xmath79 with : @xmath80 @xmath70 it is clear that the change of the einsten radius by the factor @xmath81 changes also the duration by the same factor .
consequently in the differential rates calculation , each duration @xmath82 will have to be scaled by the factor @xmath81 .
however , we see in equation 4 that for a given einsten radius there is a distribution of the scaling factors @xmath81 hidden in the double integral @xmath83 . this distribution can be calculated for each distance @xmath12 , and for the realisation of the calculations , the scaling factor of the einsten radius will be choosen by a monte - carlo method using this distribution .
some examples of such distributions for different distances are shown in figure 6 , in the case of the duo experiment .
+ + the result of the calculation of the @xmath40 and @xmath41 rates is shown in figure 7 for duo , and in figure 8 for ogle .
it is evident that the duo experiment is dominated by unresolved stars .
this result is in very good agreement with the data ( alard and guibert 1996 ) . even for ogle which has the deepest photometry ,
the bias is still very significant .
it is rather straightforward to relate these events to the excess in the rates and optical depth observed towards the bulge .
this excess was not been explained , even using a bar in the galactic model .
we find here a natural explaination .
the light curve of a blended event is affected by the additional light coming from the companion , so that the light curve is modified .
this blended light curve has more extended wings than an unblended event , so that it might be possible to differentiate it from the unblended one using a chi - square test . +
another important issue is the possibilty that the amplified star and its companion may have different colors .
for instance if the companion is more blue than the amplified star , we expect more blending in the blue , and thus a larger reduction of the amplitude in this color than in the red .
let us now investigate the problem of color variation during the event in more detail .
in the current microlensening experiments the sources are concentated in the range @xmath84 .
an examination of terndrup ( terndrup , 1988 ) color magnitude diagram in baadeswindow indicates that the mean @xmath85 color is about 1.3 to 1.4 . for fainter stars , where expect to find the unresolved sources , the holtzman _
_ color magnitude diagram shows that the color changes again very slightly . the mean color a close to 1.4 at @xmath86 .
we see that in the apparent magnitude range of interest we expect a diferential color variation of about 0.1 magnitude at most between the unresolved source and the blending companion .
we have to reach @xmath87 to expect more significant color variation between sources and blend ( about 0.2 magnitudes ) .
unfortunately the photographic technique does not perform very well in the red band , thus an accurate investigation of the color changes during the event is not possible .
it means that the duo survey will not be able to detect the slight color changes expected .
the ogle survey has a good coverage only in the i band , the sampling in v is very sparse , and does not allow color analysis during the event . in the macho case ,
photometric data are available in two large bandpasses , which is transformed to a @xmath88 color index .
the analysis of the lmc events by macho shows that they can identify color variations due to blending for some of the events .
however there is a small residual noise on the color variations of all the events of about 0.02 to 0.03 magnitudes , associated with the photometric errors . at a 3 @xmath89 confidence level
we may require a color variation of about @xmath90 magnitude to firmly established the chromaticity of the event .
converted to @xmath85 it requires a color changes of about 0.15 magnitudes at least to recognise a blended event with a good confidence level .
+ we consider now with a such accuracy how many of the bulge microlensing events which involves unresolved sourves could be identified as such .
+ + let us make the following simple model : + + and we assume that the difference of color can be described by a gaussian distribution , shifted by a systematic value .
thus the number density @xmath91 of stars with color a difference in color c will be expressed as : @xmath92 the holtzman _
_ color magnitude diagram indicates that for the gaussian we can assume @xmath93 . in the range @xmath94 to @xmath86 , a conservative value for the shift is @xmath95 .
the color shift is then given by : @xmath96 i and v stand where the subscripts for the two photometric colors . with : @xmath97 and : @xmath98 + a and c are defined by the following expressions : + + @xmath99 the number of events with color blending signature identified will be simply the number of events with a color shift @xmath100 .
it can be easily calculated if we assume that the unblended impact parameter @xmath101 has a uniform distribution . in this case
it is sufficient to integrate the number of events with @xmath100 over the @xmath101 distribution with a given color difference @xmath102 .
the calculation is completed by suming over the color difference distribution defined in eq ( 5).the maximum value of the impact parameter is set by the blending factor @xmath103 using the formulae : @xmath104^{1/2}\ ] ] with : @xmath45 the calculation of the percentage of events with color shifts is performed for several values of the blending factor @xmath46 .
an illustration of these calculations is given in figure 9 , where the color shift is expressed as a function of the impact parameter for a few @xmath46 values , and color diffrence of 0.2 magnitude .
looking at this diagram we can already guess that the number of events with detectable color shitfts will be small .
the final values of the events with color shifts identified is given in table 3 .
the values are extremly small . a shift of 0.15 mags in the color difference distribution , even for high values of the blending factor .
the values for a shift of 0.3 magnitudes are also given , to illustrate the case of the very faint unresolved stars .
however these stars will be extremly blended , and consequently will give events will very short duration which will be almost all removed by the low efficiency of the experiments in this range .
consequently the color shift method will give poor results in the case of the galactic bulge .
in the case of the magellanic clouds , the color changes rapidly with the magnitude close to the limit of the microlensing experiments , which explains why so many events with color shifts are found by macho ( alcock , _ et al .
_ , 1996)towards the lmc .
lllll blending factor & 2 & 5 & 10 & 50 + @xmath105 &
0.0025 & 0.0219 & 0.0364 & 0.0485 + @xmath106 & 0.1677 & 0.3079 & 0.3559 & 0.3906 + i will investigate another possibility to identify a blended microlensing event based on the shape of the light curve in this section .
the blending of light modifies the shape of the light curve , and consequently we may use a statistical test to see if a light curves differs significantly from an unblended one .
however we have to realise that the likely difference due to blending is rather small , and we may expect that with the current available photometry it will be difficult to identify a blended light curve at a significant level of confidence .
it is possible to adress this problem in more general terms using monte - carlo simulation of microlensing events . assuming a noise distribution
, it is easy to simulate microlensing events with different blending factors .
these blended events can be analysed by fitting an unblended curve to the simulated data set .
we expect a systematic difference in the chi - square of the fir due to the different shape of the blended light curve , especially in the wings .
however the problem is to know how significant is this difference compared to the normal chi - square variations for unblended events due to noise .
this problem can be easily addressed if we are able to build the chi - square distribution for a series of blended light curves .
such chi - square distributions are illustrated in fig .
10 , the line shows the limit within which 95 percent of the chi - square distribution for unblended events is contained . beyond this limit
we have a 95 percent confidence level that the chi - square indicates a light curve which is systematically different from the unblended model .
it is now sufficient to sum the fraction of the blended distribution beyond this limit to get the fraction of blended events @xmath107 which can be identified with a good confidence level .
the ability to recognise a blended event will of course depend on the amplitude of the event , for a given blending factor @xmath103 and a given duration @xmath108 .
the amplitude is related to the impact parameter @xmath101 . consequently to get the expected fraction @xmath107 , at @xmath103 and @xmath108 it is sufficient to sum on the impact parameter @xmath101 , in the range 0 to @xmath109 ( @xmath109 is defined in eq 6 ) .
the result of the corresponding calculation is illustrated in fig 11 , as a function of the blending factor , for a series of durations .
the blended events are well identified only for small impact parameters . beyond @xmath110 ,
the efficiency drops significantly , which makes the total efficiency quite low .
these tabulated expressions of @xmath107 as a function of @xmath103 and @xmath108 are now introduced directly in the calculation of the unresolved event rates . then new rates corrected for the number of blended events which could be recognised are computed .
fig 12 shows a comparison between this rate and the uncorrected rate in a case resembling to the ogle experiment .
typical errors distributions are extrpolated from the duo experiment ( alard and guibert 1996 ) , and are divided by a scaling factor of 2 to take into account the better quality of the ogle photometry .
the resulting errors distribution has a sigma about 0.07 magnitude for most of the points , and reaches 0.1 or slightly more in 10 percent of the cases ; it seems to be an acceptable description of the ogle photometry on stars close to the limiting magnitude , which represents most of the sample .
we see that the difference is small , and considering that ogle has the best photometric accuracy , it is certainly worse again for the others experiments .
+ + the conclusion is again , as for the color test , that given the mean quality of the photometry available in the microlensing experiments only , a slight percentage of blended events should be recognised on the basis of the light curve shape .
only the the events with good photometry and small impact parameter should be identified as blended .
however for some of the monitored stars the ogle experiment is able to perform photometry much better again than for the majority of the stars .
this gives an interesting opportunity to look for a blending signature in the light curve .
the photometry achieved on the ogle 5 event is of an interesting quality , the errors bars are as good as a few percent on many points . however
this event is not well fitted with the standard unblended model ( see figure 13 ) .
the discrepancy is especially large in the wings of the event , so that it is natural to try to fit this event with blending .
14 shows the dramatic improvement of the chi - square when a blended light curve is fitted .
the best fit is obtained for a blending factor of 2.45 ( fig 15 . ) .
thus makes the lensed source 1.34 magnitude fainter , and places it below the ogle detection limit .
consequently it is very likely that ogle 5 is an example of lensing on an unresolved star .
the better chi - square per degree of freedom is different from unity , but we have to recall that an event on an unresolved star is seen only by blending on a resolved star . for the photometry , the position of the resolved star is used , but it might be significantly different from the position of the true magnified star .
the achieved errors in a psf fitting routine are rather sensitive to the quality of the positioning , thus we expect somewhat larger errors in the case of an unresolved event .
the small discrepancy in the chi - square per degree of freedom is thus perfectly consistent with the scenario of an unresolved event .
in this paper i show that the rates of microlensing events and optical depth to microlensing due to the unresolved stars can be modelled .
however i emphasize that the exact rates and optical depth are closely related to the number of unresolved stars which will contribute per resolved star . this number is well constrained in crowded fields , essentially because the surface occupied by a star on the image is set by the crowding limit . on the other hand , this crowding limit is closely related to the resolution radius , thus the surface occupied by a star is just a function of this resolution radius .
a value of 0.5 was found for @xmath53 ( the ratio of the surface covered by the resolution radius to the mean surface occupied by a star ) , both for duo and ogle .
this simply means that probably there is a linear relation between the resolution radius and the mean radius per star .
+ however in the crowded fields , if a new star appear at more than a resolution radius , it does not mean necessarily that it will be resolved .
there are so many stars that it might well be confused with another one .
this issue should be clarified later using monte - carlo simulation of crowded fields .
this effect will probably lead to a slight increase of @xmath53 .
another problem is that if the amplified star is situated at some distance from the resolved star the photometry which assumes fixed positions ( except for duo ) is certainly affected .
this possibilty was already discussed in the case of ogle 5 and will cause a slight drop in the efficiency , which will increase as the unresolved star will be more distant form the resolved one .
this issue could be also clarified later using monte - carlo simulation .
but globally , this effect will more or less cancell out with the confusion effect exposed just before .
thus to conclude , the value @xmath111 might be an acceptable approximation .
the modeling of the unresolved star microlensing rates demonstrates that they induce an important bias for the short events .
i show that given the current photometric errors , most of these events would be hard to distinguish from the unblended events .
thus this is an important and annoying bias .
if this bias is not taken into account , it will certainly influence the shape of the lens mass function estimated from the data .
for instance , if a salpeter mass function is used , we expect that the lower cut - off will be shifted towards the brown dwarfs .
this idea is illustrated in fig .
16 where the total ogle rates ( resolved+unresolved ) are compared to the rates for salpeter mass functions with different lower cut - offs ( i recall that initially i used a salpeter mass function with a lower cut - off of @xmath112 to compute the ogle and duo rates ) . a better agreement with the total rates is found if the different trial models are slightly scaled .
i calculated this scaling factor , so that the trial models give the same total rates in the range 0 to 80 days as the total rate ( see fig 16 . ) .
the case of the duo experiment will be treated in another paper ( alard and guibert 1996 ) .
while the unresolved star bias has to be taken into account , as demonstrated in the previous section , the modelling includes some uncertainties .
however for the data set already assembled by the microlensing experiments , it is certainly an acceptable solution to the unresolved star bias . for the future ,
a very interesting possibilty is offered by the planet ( sackett , 1995 ) collaboration , who will provide a dense and accurate sampling of the events .
it will allow a much better control of the bias caused by the unresolved , using both color shift but importantly the shape of the light curve , as demonstrated for ogle 5 .
+ to conclude , i will also emphasize that the contribution of the unresolved stars drops if the magnitude limit is increased .
this is well illustrated by the comparison ogle vs. duo .
consequently a first solution is to try to reach stars as faint as possible to minimise the bias .
it is certainly a possible orientation for the ogleii experiment , which will achieve improved resolution better again with a new telescope . on the other hand for macho and erosii experiments , which use bigger pixels , and cover larger fields , an interesting solution is certainly to monitor the bright clump giants .
these stars are much brighter ( about 3 magnitudes ) than most of the bulge stars , and their density per image is also much lower .
thus , the unresolved stars per giant will be low and they will be so blended that the effective einsten radius will be very small .
this idea leads to the conclusion that the unresolved star bias will be very small on clump giants .
some other advantages of the clump giants include the good photometry expected , and also the possibility to have a sample with a very high completeness ( gould 1995 ) .
alard and guibert , c. 1996 , in preparation alard , c. 1996 , in preparation alard , c. , 1995 , iau symposium 173 , melbourne , july 9 - 14 alard , c. , mao , s. , and guibert , j. 1995 , a&a , 300 , l17 alcock c. et al
. 1996 , apj v.461 , p.84 alcock , c. et al .
1993 , nature , 365 , 621 aubourg , e. et al .
1993 , nature , 365 , 623 bahcall , j.n .
, and soneira , r.m . , 1984 ,
apjs v. 55 , 67 bouquet , a. , 1993 a&a , v. 280 , 1 di stefano r. and esin a. , 1995 , apjl , v. 448 , 1 dwek , e. , et al . , 1995 ,
apj , 445 , 716 gould , livermore workshop 1995 .
han , c. , and gould a. , 1995 , apj , 447 , 53 holtzman , j. et al . ,
aj.,v . 106 , 1826 kiraga and paczyski 1994 , apjl , v. 430 , l 101 mollerach , s. and roulet , e. 1996 , apjl v.458 , p.l9 mould , j. , 1996 , pasp , v. 108 , 35 nemiroff , r. , 1994 , apj . , v. 435 , 682 sackett , p. , 1995 , baas , 18711704a terndrup , d. m. , 1988 , aj , v. 96 , 884 udalski , a. et al .
1994 , acta astron .
, 44 , 165 zhao , h. , spergel , d.,n . , and rich , r.m . , 1995 , apj , v. 440 l13 | this paper presents an analysis of lensing of unresolved background stars .
previous calculations of the lensing rates and optical depths considered only resolved stars . however , if a faint unresolved star lens is close enough to a resolved star , the event will be seen by the microlensing experiments and attributted to the bighter star .
the blending biases the duration , making the contribution of the unresolved stars very significant for short events .
it is especially annoying , because this contribution is confused with lensing by brown dwarfs .
the exact rates of these blended events are extremly sensitive to the limiting magnitude achieved in the microlensing search .
appropriate calculations of the optical depth and rates are provided here , and illustrated in the case of the duo and ogle experiments .
the additional contribution of unresolved stars is very significant .
it probably explains the high optical depth and rates observed towards the galactic bulge .
the blended unresolved event can be identified using either the color shift or the light curve shape .
however , neither of these two methods is apropriate to identify a large number of blended events towards the bulge . in some case of good
photometry and small impact parameter an indentification is possible .
this is illustrated by the case of the ogle 5 event , which clearly appears as a case of lensing of an unresolved star .
the recent results obtained by the planet collaboration indicate that a high resolution and dense sampling of the light curve is possible , and will probably provide a very interesting possibility to correct the blending bias , as demonstrated for ogle 5 .
this possiblity , is certainly better than a statistical estimation of the lensing rates , which are always prone to some uncertainty .
but , at this time , the analysis of microlensing events found in the various microlensing experiments requires the uses of modelisations of the contribution of unresolved stars . |
in this appendix , we present the tree and one - loop compton amplitudes for virtual photon scattering on a transversely polarized nucleon in the unfactorized form . to simplify the expressions ,
we assume @xmath140 so that the amplitudes are purely real . with @xmath14 ,
we write the compton tensor as @xmath141 in qcd ,
we can write , @xmath142 - ( x_b\rightarrow - x_b ) \ , \end{aligned}\ ] ] where the @xmath143 s are unrenormalized parton distributions as defined in the text , @xmath112 s are perturbation series in @xmath61 and have infrared poles . at tree level , the result is well known , @xmath144 at the one - loop level , the following amplitudes are associated with distributions with a covariant derivative , @xmath145 \nonumber \\ & & -2\left({2y\over 1-x } - { y\over x } + 3\right ) -\left({3y\over 1-x } - { 2y\over x^2 } \right . \nonumber \\ & & + \left . { 12\over x } + { 4\over y - x } - { y\over x } -2-{2y\over y - x}\right)\log(1-x ) \nonumber \\ & & -{2y\over y - x}\left(1 -{2x\over 1-y } -{4\over y - x}\right ) \log(1-y+x ) \nonumber \\ & & + \left({2\over x } + { y\over x } - { y\over x^2 } + { 2y\over 1-x } \right ) \log^2(1-x ) \nonumber \\ & & \left . - { y\over y - x}\left({2\over y - x } + { x\over 1-y}\right ) \log^2(1-y+x ) \right\ } \nonumber\\ & & + { \alpha_{sb}\over 2\pi}{c_a\over 2 } \left({\bar\mu^2\over q^2}\right)^{\epsilon/2 } \left\ { { 2\over \epsilon } \left [ - { 4 \over y - x}\log ( 1-x ) \right.\right.\nonumber \\ & & \left . - { 2y\over y - x } \left({2\over y - x } + { x\over 1-y}\right)\log ( 1-y+x)\right ] \nonumber \\ & & -2\left({1\over 1-y } + { 4\over y - x } \right)\log(1-x ) \nonumber \\ & & + { 2y\over y - x}\left(1 -{2x\over 1-y}- { 4\over y - x}\right ) \log(1-y+x ) \nonumber \\ & & + { 2\over y - x}\log^2(1-x ) \nonumber \\ & & \left .
+ { y\over y - x } \left({2\over y - x } + { x\over 1-y}\right ) \log^2(1-y+x ) \right\}\ , \end{aligned}\ ] ] @xmath146 \nonumber \\ & & -2\left({2y\over 1-x } - { y\over x}\right ) -\left({3y\over 1-x } - { 2y\over x^2 } - { y\over x}+2\right)\log(1-x ) \nonumber \\ & & -{2y\over y - x}\left(1 + { 2x\over 1-y}\right ) \log(1-y+x ) \nonumber \\ & & + \left({y\over x } - { y\over x^2 } + { 2y\over 1-x } \right ) \log^2(1-x ) \nonumber \\ & & \left . + { xy\over ( y - x)(1-y ) } \log^2(1-y+x ) \right\ } \nonumber \\ & & + { \alpha_{sb}\over 2\pi}{c_a\over 2 } \left({\bar\mu^2\over q^2}\right)^{\epsilon/2 } \left\ { { 2\over \epsilon } { 2xy\over ( y - x)(1-y)}\log ( 1-y+x)\right .
\nonumber \\ & & + { 2\over 1-y } \log(1-x ) \nonumber \\ & & + { 2y\over y - x}\left(1 + { 2x\over 1-y}\right ) \log(1-y+x ) \nonumber \\ & & \left .
- { xy\over ( y - x)(1-y)}\log^2(1-y+x ) \right\}\ , \end{aligned}\ ] ] r. mertig and w. l. van neervan , z. phys .
c * 60 * , 489 ( 1993 ) ; + g. altarelli , b. lampe , p. nason and g. ridolfi , phys .
b * 334 * , 187 ( 1994 ) ; + j. kodaira , s. matsuda , t. uematsu , and k. sasaki , phys .
b * 345 * , 527 ( 1995 ) ; + p. mathews , v. ravindran , and k. sridhar , hep - ph/9607385 ; + a. gabieli , g. ridolfi phys , lett .
b * 417 * , 369 ( 1998 ) .
p. g. ratcliffe , nucl .
b * 264 * , 493 ( 1989 ) ; + i. i. balitsky and v. m. braun , nucl .
b * 311 * , 541 ( 1989 ) ; + x. ji and c. chou , phys . rev .
d * 42 * , 3637 ( 1990 ) ; + b. geyer , d. mller and d. robaschik , hep - ph/9611452 ; + j. kodaira , y. yasui , k. tanaka and t. uematsu , phys .
b387 , 855 ( 1996 ) . | we consider the one - loop factorization of the simplest twist - three process : inclusive deep - inelastic scattering of longitudinally - polarized leptons on a transversely - polarized nucleon target . by studying the compton amplitudes for certain quark and gluon states at one loop , we find the coefficient functions for the non - singlet twist - three distributions in the factorization formula of @xmath0 .
the result marks the first step towards a next - to - leading order ( nlo ) formalism for this transverse - spin - dependent structure function of the nucleon .
= 0.5 cm deep - inelastic scattering ( dis ) of leptons on the nucleon is a time - honored example of the success of perturbative quantum chromodynamics ( pqcd)@xcite . the factorization formulae for the leading structure functions @xmath1 and @xmath2 , augmented by the dokshitzer - gribov - lipatov - altarelli - parisi ( dglap ) evolution equations for parton distributions @xcite , can describe the available dis data collected over the last 30 years exceedingly well .
although the same formalism is believed to work for the so - called higher - twist structure functions @xcite , e.g. @xmath3 and @xmath4 , which contribute to physical observables down by powers of the hard momentum @xmath5 , there are few detailed studies of them in the literature beyond the tree level .
the qcd radiative corrections to @xmath0 need be investigated as accurate data have recently been taken @xcite and more data will be available in the future @xcite . in this paper , we report a one - loop study of inclusive deep - inelastic scattering of longitudinally - polarized leptons ( e.g. electrons ) on a transversely - polarized nucleon target @xcite .
the subject was first investigated in the context of single parton scattering in ref .
@xcite , and studies along the same line have continued in the literature @xcite .
however , due to the subtlety of the twist - three process @xcite , those results are sensitive to the treatment of quark masses and are incomplete in the context of qcd fatorization .
indeed , even at tree level one must go beyond the single quark process to derive the correct @xmath3 expression in terms of the parton distributions @xcite .
when loop corrections are included , one needs a general strategy to systematically calculate their contribution to higher twist processes . for a transversely polarized nucleon of four - momentum @xmath6 and polarization vector @xmath7 , the hadron tensor @xmath8ps\rangle$ ]
can be expressed as @xmath9 where @xmath10 is the photon four - momentum , @xmath11 , @xmath12 , @xmath13 , and @xmath14 .
@xmath15 is the electromagnetic current of the quarks .
thus it is the combination @xmath16 that naturally appears in the @xmath17-suppressed transverse polarization asymmetry . in this paper
, we concentrate only on the non - singlet part of @xmath18 ; the singlet case is left to a separate publication .
we seek a factorization formula for @xmath19 to one - loop order in terms of the perturbative coefficient functions @xmath20 and the generalized parton distributions ( correlations ) @xmath21 with two light - cone ( or feynman ) variables @xmath22 and @xmath23 .
the starting point is one - loop forward virtual - photon compton scattering off a few on - shell " quark and gluon states .
> from the scattering amplitudes , we examine the validity of infrared factorization and extract the one - loop coefficient functions @xmath20 .
we begin by outlining a general approach to the factorization of higher - twist observables , generalizing the method of ref .
@xcite for the twist - two structure functions . for deep - inelastic scattering , it is convenient to consider first the forward compton amplitude @xmath24 the spin - dependent part of the compton amplitude ( antisymmetric in @xmath25 and @xmath26 ) defines two invariant amplitudes @xmath27 , @xmath28 according to the optical theorem , the imaginary part of the @xmath29 amplitudes ( divided by @xmath30 ) is just the nucleon spin structure functions @xmath31 . hence , factorization of @xmath32 naturally leads to factorization of the structure function @xmath18 .
the former has a straightforward feynman - dyson perturbative expansion , making it easily handled .
according to ref .
@xcite , the nucleon compton amplitude @xmath33 can be expressed as a sum of terms that are convolutions of quark - gluon compton scattering amplitudes @xmath34 and the bare quark - gluon correlation functions @xmath35 in the nucleon , as shown schematically in fig . 1
, @xmath36 implicitly involved in the convolution are integration over the intermediate quark - gluon four - momenta @xmath37 and summation over the spin and color indices .
@xmath38 , without external quark - gluon legs and self - energies , is the sum of a complete set of feynman diagrams for compton scattering .
these diagrams are calculated in unrenormalized perturbation theory in the sense that all parameters in the expansion are bare .
we use dimensional regularization to regularize both infrared and ultraviolet divergences in the diagrams .
the pair of light - cone four - vectors @xmath39 and @xmath40 define a collinear basis in which the nucleon and photon momenta are written @xmath41 and @xmath42 , respectively , where @xmath43 is an arbitrary dimensionful parameter .
intermediate quark and gluon momenta can also be expressed in this basis : @xmath44 and the full result for @xmath38 can be expanded about @xmath45 .
the leading term in this expansion can be interpreted as scattering of collinear partons with feynman momentum fractions @xmath46 .
because the partons are on - shell , @xmath38 can be viewed as the parton scattering s - matrix element . in principle
, one must multiply by parton wave function renormalization factors to get the proper s - matrix element .
however , in dimensional regularization , the absence of a physical scale at the massless poles of the quark and gluon propagators reduces these contributions to unity .
subleading terms in the expansion of @xmath38 are parton scattering s - matrix elements with insertions of certain vertices associated with powers of @xmath47 and @xmath48 .
for instance , with one power of quark momentum @xmath47 , the subleading term is calculated with one insertion of the vector vertex @xmath49 to one of the quark propagators . because s - matrix elements and their relatives are gauge invariant
, one may choose any gauge for the internal gluon propagators in @xmath38 .
we use feynman s choice in our calculation . after the collinear expansion and integration over the quark - gluon four momenta , the convolution of eq .
( [ conv ] ) involves integrating over the parton feynman variables @xmath50 .
the correlation functions @xmath35 are matrix elements of gauge - invariant nonlocal quark and gluon operators in one - to - one correspondence with the external parton states in @xmath38 .
in particular , when @xmath38 contains momentum - related vertex insertions , the quark and/or gluon fields in @xmath35 appear with partial derivatives .
the collinear expansion results in all qcd fields separated in spacetime along the light - cone direction @xmath51 . as in the leading - twist case , expanding the gluon polarization indices and summing over all contributions from longitudinally - polarized gluons generates straightline gauge - links which connect fields at separate spacetime points . to simplify the derivation of a factorization formula , we use the light - cone gauge ( @xmath52 ) , or link - free , expression for @xmath35 .
this gauge choice allows one to focus on the physical partons without worrying about the effects of the longitudinally polarized gluons .
the gauge - invariant form of the factorization is recovered simply by imposing gauge invariance on the final form of parton correlations . as we have argued above ,
@xmath34 is ultraviolet finite because the on - shell wave function renormalization is trivial in dimensional regularization .
nevertheless , due to the massless on - shell external states , @xmath34 has infrared divergences showing up as @xmath53 poles .
these divergences may be factorized in the perturbative sense @xmath54 where @xmath55 is the finite coefficient function and @xmath56 contains only the @xmath53 poles . on the other hand , the infrared - finite quantities @xmath35 contain ultraviolet divergences which also show up as @xmath53 poles .
when the infrared poles in @xmath56 cancel all the ultraviolet poles in @xmath35 , @xmath33 is said to be factorizable .
the product @xmath57 defines the renormalized parton correlation functions @xmath58 .
the final factorization formula for the compton amplitude is then @xmath59 where @xmath60 is a well - defined perturbation series in @xmath61 and @xmath58 is a finite nonperturbative distribution .
we apply the above discussion to compton scattering on a transversely polarized nucleon . as shown in fig .
2 , the non - singlet twist - three process involves two possible intermediate parton states in the light - cone gauge . first is the two - quark state with the transverse momentum flowing through @xmath62 , as shown in fig .
2a .
after taylor - expansion , we keep only the contribution with exactly one insertion of the @xmath49 vertex .
the corresponding correlation function is @xmath63 where the dirac and color indices on quark fields are open and @xmath64 is perpendicular to @xmath65 and @xmath51 .
the second intermediate parton state involves two quark lines and one gluon line , as shown in fig .
2b .
if we use the feynman rule @xmath66 for the gluon attachment to @xmath67 , the corresponding correlation function is @xmath68 we remind the reader that all fields and the couplings here are bare . to decouple the spin and color indices in @xmath38 and @xmath35 , we write explicitly @xmath69 where @xmath70 is the number of colors , @xmath71 , and @xmath72 .
in dimensional regularization , @xmath73 must be defined explicitly . here , we follow t hooft and veltman s convention . with appropriate insertions of light - cone gauge links , which can be generated by summing over intermediate states with additional longitudinally - polarized gluons , we define the following gauge - invariant parton correlations @xcite
: @xmath74 one may argue that this is not the only way to obtain gauge invariant correlations .
in particular , the combination @xmath75 is the light - cone expression for a gauge invariant distribution involving the gluon field strength rather than the covariant derivative . on the other hand ,
these contributions necessarily involve coefficients @xmath76 that vanish when @xmath77 .
eq.([vanish ] ) then implies that one can consider instead the combination of eq.([combine ] ) free of charge .
hence these are the only distributions relevant to our process .
using hermitian conjugation , it is easy to show that @xmath78 is symmetric in @xmath22 and @xmath23 and @xmath79 is antisymmetric
. we will use these symmetries to simplify our presentation of the coefficient functions .
let us consider the use of the above approach to the tree diagrams shown in fig .
3 , where we have shown @xmath80 only .
those diagrams for @xmath62 can be obtained simply by replacing the gluon interaction @xmath66 by the transverse vector vertex @xmath49 .
figure 3 corresponds to the following on - shell process : a quark and gluon with four - momenta @xmath81 and @xmath82 , respectively , scatter with a photon of four - momentum @xmath10 , producing a quark of four - momentum @xmath83 and a forward photon .
the compton amplitude can easily be calculated using the usual feynman rules . when the quark and gluon combine before or after interacting with the photons , the intermediate propagator , @xmath84 , drawn with a bar in fig .
3 , is singular and needs regularization .
adding an infinitesimal @xmath85 , one arrives at @xmath86 the first term yields zero when acting on the external quark wave function , so only the second term contributes . combining the result from @xmath62 and taking into account crossing symmetry , we find the following gauge - invariant expression for the tree - level compton amplitude @xmath87-(x_b\rightarrow -x_b)\ , \end{aligned}\ ] ] where @xmath88 is the non - singlet part of the quark charge squared and @xmath89 . the integration has support only for @xmath90 .
when @xmath91 , the above expression develops an imaginary part through @xmath92 . using the optical theorem ,
one obtains the following tree result for @xmath93 , @xmath94 > from qcd equations of motion , one can show @xmath95 where @xmath96 is defined as @xmath97 this object seems to have a simple physical interpretation .
the tree level result suggests the following all - order factorization formula for the non - singlet part of @xmath98 , @xmath99 @xmath100 can be written in terms of a perturbation series in the strong coupling @xmath61 , @xmath101 the tree result yields @xmath102 now we come to the main subject of the paper : one - loop radiative corrections for @xmath103 .
all the feynman diagrams for @xmath104 are shown in fig .
4 . the first 12 diagrams have color factor @xmath105 , the next 8 come with the factor @xmath106 , and the final 4 with @xmath107 .
since the external transverse momentum of the quark only goes through the quark propagators , the taylor - expansion of @xmath62 yields the vector vertex insertion on the quark line only .
these diagrams correspond to the first 20 diagrams in fig .
4 with the external gluon vertex replaced by the vector vertex .
an inspection of the one - loop feynman integrals reveals a simple rule : @xmath62 is just the @xmath105 part of @xmath104 and , hence , the former need not be calculated separately .
since all feynman diagrams are computed in bare perturbation theory , the result depends on the bare coupling @xmath108 .
we replace it with a renormalized coupling in the @xmath109 scheme , with a difference of higher order in @xmath110 . after a lengthy calculation involving one - loop integrals with up to five internal feynman propagators ,
we find @xmath111 - ( x_b\rightarrow - x_b)\ . \end{aligned}\ ] ] explicit expressions for the @xmath112 s in the region @xmath113 can be found in the appendix .
now we are ready to show that @xmath114 is factorizable at the one - loop level , i.e. , the infrared poles @xmath53 in @xmath115 match the ultraviolet poles in @xmath116 . to this end , we use the infrared poles in @xmath117 to generate a scale evolution equation for the parton distributions .
from the result in the appendix , we find @xmath118 \nonumber \\ & & + c_f k_2(x , y , \mu^2 ) \left[{-1\over x(x_b - x ) } + { 2\over x^2 } - 2\left({1\over x^2 } - { x_b\over x^3 } + { 2\over x(x_b - x)}\right ) \right .
\nonumber \\ & & \left .
\times \log\left ( { x_b - x\over x_b}\right ) - { 2\over ( y - x)(x_b - y)}\log \left({x_b - y+x\over x_b } \right)\right ] \nonumber \\ & & + { c_a\over 2}k_1(x , y,\mu^2 ) \left [ - { 4x_b \over xy(y - x)}\log \left({x_b - x\over x_b}\right ) \right . \nonumber \\ & & \left . - { 2x_b\over y - x } \left({2\over x(y - x ) }
+ { 1\over x_b - y}\right ) \log\left({x_b - y+x\over x_b}\right)\right ] \nonumber \\ & & \left .
+ { c_a\over 2}k_2(x , y,\mu^2 ) { 2\over ( y - x)(x_b - y)}\log\left({x_b - y+x\over x_b}\right ) \right\ } \ . \end{aligned}\ ] ] expanding the above in the large @xmath119 limit , we get the evolution equation for the moments of the parton distributions .
a detailed check shows that the equation is identical to that in ref .
@xcite obtained by studying the ultraviolet divergences present in the twist - three operators . in particular , in the large @xmath70 limit , the above equation becomes autonomous @xcite .
the final step of the calculation is to take the imaginary part of the factorized @xmath114 to get a factorized expression for the structure function @xmath120 in the physical region @xmath121 . with the definition in eq .
( 19 ) we find the following result for the coefficient functions , @xmath122 \nonumber \\ & & + \theta\left({y - x\over x_b}-1\right ) { 1\over y - x}\left[1 - { 2x\over x_b - y } - { 4x_b\over y - x } \right .
\nonumber \\ & & \left .
+ \left({2x_b\over y - x } + { x\over x_b - y } \right)\log{\left(y - x - x_b\over x_b\right)}\right ] \nonumber \\ & & + \delta(x - x_b)\log{\left(1-x_b\over x_b\right ) } \left[-{3\over 2 } + \log{\left(1-x_b\over x_b\right)}\right ] \nonumber \\ & & \left .
+ \delta(y - x_b){2x\over x_b - x } \log\left({x\over x_b}\right ) \left[1-{1\over
4}\log\left({x\over x_b}\right)\right ] \right\ } \nonumber\\ & & + { c_a\over 2}y \left\ { \theta\left({x\over x_b}-1\right){x_b\over y } \left [ { 1\over x_b - y } + { 4\over y - x}\left(1-{1\over 2 } \log\left({x_b - x\over x_b}\right)\right)\right ] \right . \nonumber \\ & & -\theta\left({y - x\over x_b}-1\right){1\over ( y - x)}\left [ 1 - \left({2x\over x_b - y } + { 4x_b\over ( y - x)}\right ) \right . \nonumber \\ & & \left .
\times \left(1-{1\over 2 } \log\left({y - x - x_b\over x_b}\right)\right)\right ] - \delta(y - x_b)\left [ \log\left({x - x_b\over x_b}\right ) \right .
\nonumber \\ & & \left.\left .
+ { 2x\over ( x_b - x ) } \left(1-{1\over 4}\log\left({x\over x_b}\right ) \right)\log\left({x\over x_b}\right)\right ] \right\ } \ , \end{aligned}\ ] ] @xmath123 \nonumber \\ & & + \theta\left({y - x\over x_b}-1\right ) { 1\over y - x}\left[1 + { 2x\over x_b - y } - { x\over x_b - y}\log{\left(y - x - x_b\over x_b\right)}\right ] \nonumber \\ & & + \delta(x - x_b)\log{\left(1-x_b\over x_b\right ) } \left[-{3\over 2 } + \log{\left(1-x_b\over x_b\right)}\right ] \nonumber \\ & & \left
. - \delta(y - x_b){2x\over x_b - x } \log\left({x\over x_b}\right ) \left[1-{1\over 4}\log\left({x\over x_b}\right)\right ] \right\ } \nonumber \\ & & + { c_a\over 2}y \left\ { -{x_b\over y(x_b - y ) } \theta\left({x\over x_b } -1\right ) \right .
\nonumber \\ & & - { 1\over y - x}\left[1 + { 2x\over x_b - y } \left ( 1- { 1\over 2}\log\left({y - x - x_b\over x_b}\right)\right)\right ] \theta\left({y - x\over x_b}-1\right ) \nonumber \\ & & \left .
+ \delta(y - x_b ) \left [ \log\left({x - x_b\over x_b}\right ) + { 2x\over x_b - x}\left(1-{1\over 4}\log \left({x\over x_b}\right ) \right)\log\left({x\over x_b}\right)\right ] \right\ } \end{aligned}\ ] ] the definition of the + functions can be found in ref .
@xcite .
the support for the parton correlations limits @xmath124 to the interval @xmath125 $ ] . to check the burkhardt - cottingham sum rule @xcite
, we integrate @xmath103 over @xmath119 . assuming the integration over @xmath22 and @xmath23 can be interchanged with that of @xmath119 , one obtains @xmath126 here the coefficient @xmath127 reduces to @xmath128 if we define @xmath73 so that the non - singlet axial current is conserved . compared with the factorization formula for @xmath2
, we have the burkhardt - cottingham sum rule at one loop @xmath129 if the order of integration can not be interchanged because of the singular behevior of the parton distributions at small @xmath22 and @xmath23 , the above sum rule may be violated .
indeed some small @xmath119 study does indicate such singular behavior @xcite .
finally , we consider the next - to - leading order correction to the non - singlet part of the @xmath130 moment of @xmath18 . in the leading order ,
it is well known : @xmath131 where @xmath132 is the second moment of the @xmath133 structure function and @xmath134 is a twist - three matrix element @xcite .
using the coefficient functions found above , we see that @xmath135 using the next - to - leading result for @xmath136@xcite , we find @xmath137 notice that the combination of @xmath98 and @xmath138 relevant to @xmath134 receives no radiative correction . to summarize
, we have extended the leading - twist factorization formalism to higher twist .
the extension mainly involves a correct identification of the intermediate parton states and arranging different calculations into gauge invariant combinations . because of
the multipartons in the initial state , higher - twist perturbative calculations are , in general , much more complicated than the leading twist cases . as an example
, we have presented the one - loop coefficient functions the twist - three process involving forward compton scattering on a transversely polarized nucleon target .
this result can be combined with a two - loop calculation of the twist - three evolution equation to yield a next - to - leading order formalism for the @xmath139 structure function .
the authors wish to acknowledge the support of the u.s .
department of energy under grant no .
de - fg02 - 93er-40762 .
x. song was supported in part by the institute of nuclear and particle physics , department of physics , university of virginia , and the commonwealth of virginia . |
solar eruptive events , including flares , filament eruptions , and coronal mass ejections ( cmes ) are due to magnetic reconnection or loss of equilibrium in the solar corona ( see , e.g. , a recent review by webb & howard 2012 ) . from time sequence magnetograph observations , irreversible and rapid changes of surface magnetic fields associated with a large number of major flares have been observed ( e.g. , wang 1992 , 2006 ; kosovichev & zharkova 2001 ; wang et al .
1994 , 2002 , 2004a , b ; liu et al . 2005 ; sudol & harvey 2005 ; petrie & sudol 2010 ; burtseva & petrie 2013 ) .
wang & liu ( 2010 ) synthesizes earlier studies and presented analysis of new events , and found a trend of increase of horizontal field at the polarity inversion line ( pil ) associated with almost all the flares studied by the authors .
the results agree with our previous finding of rapid changes of sunspot structure associated with flares ( liu et al . 2005 ; deng et al .
2005 ) , where the spot feature near the flaring pil darkens while part of peripheral penumbra decays . recently
, the helioseismic and magnetic imager ( hmi ) on board the solar dynamics observatory ( sdo ) is providing state - of - the - art observations under seeing - free condition , which help advance the study of magnetic field changes associated with flares ( e.g. , liu et al .
2012 ; wang et al .
2012a , b ; petrie 2012 ) .
all these photospheric magnetic field changes are interpreted due to either the change of field line orientation or the appearance of newly formed magnetic loops near the surface . in both cases ,
the changes are more prominent in the horizontal rather than the vertical component . from the viewpoint of the theory of the flare phenomenon , hudson et al .
( 2008 ) and fisher et al .
( 2012 ) pointed out that the coronal energy release and magnetic restructuring may cause back reaction to the solar surface and interior .
the authors found that after a flare , the photospheric magnetic field may become more horizontal at the pil .
the authors used the simple principle of energy and momentum conservation , and specifically predicted that flares can be accompanied by rapid and irreversible changes of photospheric magnetic field .
melrose ( 1997 , 2012 ) also provided an explanation for the enhancement of magnetic shear at the flaring pil , using the concept of reconnection between two current - carrying flux systems .
such a magnetic share increase at the pil is often observed ( wang et al .
, 2012a , b ) .
the work of melrose and those of hudson and fisher , may have physics linkage : there are two current systems , one moves upward as part of an eruption and the other moves toward the surface .
these might be linked to the tether - cutting reconnection model for sigmoids ( moore et al .
2001 , 2006 ) .
it is noticeable that even with the sdo / hmi observations , the resolution of data is no better than 1 . therefore , the detailed evolution of sunspot structure may not be well observed .
the high - resolution ( 0@xmath01 ) observation with the 1.6 m new solar telescope ( nst ; goode et al . 2010 ; cao et al . , 2010 ) at big bear solar observatory ( bbso ) thus provides a unique opportunity to understand the fine - scale structure change of sunspots associated with flares .
compared with most of the previous observations mainly for x - class flares , nst images also allows such a study related to weaker flares . in this letter
, we provide a detailed analysis of rapid formation of a @xmath1 sunspot associated with the c7.4 flare on 2012 july 2 from the noaa ar 11525 .
we describe the multiwavelength observations and results in section [ result ] , and summarize and discuss our major findings in section [ summary ] .
more dynamic detail can be seen in the accompanying animations in the online journal .
noaa ar 11525 produced numerous c - class and m - class flares during its disk passage from 2012 june 29 to july 12 .
figure [ f1](f ) displays a hmi vertical field covering the entire active region .
multiple pils and the possible interaction between emerging flux systems could be the primary reasons of the high flare productivity .
we concentrate our analysis on a small field - of - view ( fov ) ( 27 @xmath2 27 ) as marked by the black box in ( f ) , in order to scrutinize the flare - related changes of sunspot structure .
this fov includes a main flaring pil , and is within the larger 70 @xmath2 70 fov of nst diffraction - limited observation .
nst best covers the c7.4 flare on july 2 , which started at 18:45 ut , peaked at 18:56 ut , and ended at 19:02 ut in goes 18 flux .
the related nst observations are at three wavelengths , the h@xmath3 line center , [email protected] offband , and the tio band at 7057 which is a good proxy for the photospheric continuum .
the cadence of h@xmath3 blue - wing images is 6 s , while that of h@xmath3 line center and tio images is 15 s. all the observations were taken with the bbso s 76 element adaptive optics , which is able to correct most of the atmospheric seeing at these wavelengths .
speckle reconstruction using 100 frames ( obtained within 3 s ) is subsequently applied in order to obtain diffraction - limited image sequences ( i.e. , around 0@xmath01 at h@xmath3 and tio ) .
figures [ f1](a ) and ( b ) are the h@xmath3 line center and offband images at the flare peak time , showing the double flare ribbons and the erupting material ( pointed to by arrows ) .
in fact , this flare is most probably preceded by the eruption of a filament lying along the pil from @xmath518:33:50 onward ( see the time - lapse movie ) .
a complete study of the erupting filament will be presented elsewhere .
figures [ f1](c ) and ( d ) compare preflare and postflare tio images , with the preceding spots p1p3 and the following spot f1 labeled .
it can be distinctly seen that after the flare , there is a newly formed penumbra ( pointed to by the arrow ) that directly connects p1 and f1 ( see further discussion below ) .
contours of hmi line - of - sight ( los ) magnetic field is also over - plotted on the pre - flare tio image in figure [ f1](e ) to clearly exhibit the polarity of each spot .
bbso did not have magnetograph observations on that day . to shed light on the 3d magnetic field structure
, we resorted to the nonlinear force - free field extrapolation ( nlfff ) , using hmi vector magnetograms as the boundary condition .
this extrapolation model uses the `` weighted optimization '' method as discussed by wiegelmann ( 2004 ) .
the photospheric boundary is pre - processed to simulate the force - free condition ( wiegelmann et al .
the calculation was performed to a volume of @xmath5171 @xmath2 171 @xmath2 171 mm@xmath6 .
selected nlfff lines are superimposed on figures [ f1](e ) and ( f ) , demonstrating a possible sheared flux rope lying along the pil ( red ) and the overlying arcade field ( blue ) .
these are nicely consistent with the above observation of the erupting filament and flare ribbons . in the standard flare picture ,
filament eruption can stretch open the overlying fields , which then reconnect and produce flare ribbons at their footpoints .
we present in figure [ f2 ] the time sequence of tio images right across the flare ( panels ( a)(g ) ) and a corresponding hmi vector magnetic field map ( panel ( h ) ) .
it is evident that there exists strongly sheared fields along the pil between the positive , ridge - like p2 and negative f1 spots , and that before the flare , a sheared penumbra lies between p2 and f1 ( figure [ f2](a ) ) . with the occurrence of the flare , a significant new section of penumbra then rapidly forms and
is obviously seen to directly connect the main positive spot p1 with the negative f1 ( figures [ f2](e)(g ) ) and hence link part of the flare ribbons ( cf . figures [ f1](b ) and ( d ) ) .
we note that some penumbral fibrils may apparently join positive spots p1 and p2 .
this is most probably because that the inclined penumbral field stemming from p1 could have a dip near the region of p2 but then curves up again to reach f1 ( e.g. , figure 8 of wang , deng , & liu 2012 ) . in the sense that the main spots p1 and f1 with opposite magnetic polarity now share a common penumbra
, the flare effectively creates a new @xmath1 sunspot . in order to further demonstrate the rapidness of the flare - related penumbral formation
, we show in figures [ f3](a ) and ( b ) the temporal evolution of the mean tio intensity within the red boxed region in figure [ f2 ] along with that of the corresponding horizontal magnetic field , in comparison with the goes 18 flux ( the red line ) .
it is remarkable that the tio intensity begins to sharply drop with the onset of the flare at 18:45 ut and reaches a nearly 20% lower level around 19:30 ut . at the meantime ,
the horizontal field starts to gradually increase from @xmath518 ut but exhibits a highest increase rate ( see the cyan line in ( b ) ) simultaneously with the flare peak .
similar to our previous studies ( e.g. , liu et al .
2005 ; wang et al .
2012b ) , we believe that the rapid intensity darkening and horizontal field enhancement are resulted from the 3d magnetic field reconstruction due to the flare .
the long - term gradual strengthening of the horizontal field , however , may be related to the overall evolution of the active region , i.e. , the converging motion between p1/p2 and f1 as discussed below .
the spatial - temporal relationship of the formation of the new penumbra can be best seen in figure 4 , in which we show the time slice image along the slit as marked in figure [ f2](a ) . before the flare ,
obvious convective pattern is present in the upper region ( @xmath51019 ) with a time scale of 1020 minutes and a spatial scale of about 12 , which are typical scales of granulation . after the flare ,
such a pattern is replaced by a typical penumbral structure with alternating dark and bright fibrils .
it is also convincing that the transition from granulation to penumbrae is rapid and occurs right after the flare peak ( the over - plotted solid line ) . since long - term evolution of magnetic field
can provide hints for understanding the triggering of flares , we also examine in figure [ f3](c ) the time profiles of magnetic flux within the cyan box in figure [ f2](h ) that encompasses the main flaring pil .
we realize the difficulty in separating the short - term change in the smaller scale from the long - term change in the larger scale .
nevertheless , with the help of the accompanied magnetograph movie , there is some indication of magnetic cancellation in this region beginning from @xmath518 ut , which reduces the positive flux with the highest cancellation rate ( see the cyan line ) co - temporal with the peak of the flare .
the cancellation could be caused by the collision between p2 and the eastward motion of the emerging negative f1 field .
however , it is also possible that such a reduction of positive flux is partially due to its southward migration . the evolution of negative field can not be used to identify the flux cancellation , as f1 is part of another larger scale flux emergence .
the sharp increase of negative flux ( see the green line ) around the flare peak time is then due to the emerging flux moving into the calculation box .
all these may be resulted from the interplay between the long - term and short - term magnetic restructuring , for which we presently can not draw a definite conclusion .
it is worth emphasizing two facts in our plots .
first , a rapid change is only observed in the tio intensity observed with nst .
in contrast , all the hmi magnetic quantities have a more gradual change with an onset time even before the flare , although the sharp increase of the change rate is apparently associated with the flare .
this could be attributed to the very different image scales of the two sets of observations and the long - term evolution of the entire active region in a large scale .
second , there is another peak in the field change rate around 21:15 ut , while we can not find any goes x - ray signature around this time .
after carefully examining the nst h@xmath3 data , we noted a sub - flare peaked at 21:16 ut , which has a similar morphology as the c7.4 flare .
we speculate that the magnetic field as well as intensity changes are associated with this sub - flare .
taking advantage of the 0@xmath01 spatial resolution and 15 s cadence images of the nst , we observed in detail the rapid formation of a sunspot penumbra across the pil closely associated with the 2012 july 2 c7.4 flare . the penumbral formation is unambiguously evidenced by the transformation from patterns of typical granulation to penumbral fibrils , and is accompanied with rapid tio intensity darkening and horizontal field enhancement .
the formation of this penumbra produces a larger scale @xmath1 sunspot .
such a fine - scale study is only possible with observations at sufficiently high resolution .
based on the observational results , we propose the following interpretation
. there could exist a low - lying flux rope right above the initially sheared pil , which can be supported by sheared arcade fields . some evidence of magnetic cancellation at this pil and the filament eruption preceding the flare are consistent with the formation and eruption of the flux rope ( e.g. , green , kliem , and wallace 2010 ) . as the flux rope moves upward ,
the closing of the opened overlying fields are then pushed back toward the surface as a result of the downward momentum ( fisher et al .
this leads to the formation of the new penumbra and the @xmath1-spot configuration . several supporting evidences for this scenario are as follows .
( 1 ) sheared fields exist around the original pil before the vertical fields cancel across it .
( 2 ) nlfff extrapolation clearly shows both the flux rope and the overlying arcade fields . ( 3 ) rising of the flux rope is visible in the h@xmath3 movie , and part of the bundle is apparent in the h@xmath3 blue - wing image in figure [ f1](b ) . ( 4 ) as shown in h@xmath3 images ( figures [ f1](a ) and ( b ) ) , the flare ribbons are not located beside the compact flaring pil , but lie further apart in the main sunspots p1 and f1 that are finally connected by the newly formed penumbra .
we also note that there are four homologous flares occurred during this period ( 1722 ut ) .
each flare may have a partial contribution to the flux rope eruption but the c7.4 flare plays a major role .
the penumbral formation is also only associated with this flare .
we emphasize that only high - resolution observing sequence would allow such a detailed examination of sunspot structure change .
the size of the new penumbra is only about 5 . without these new nst observations
, we might only conclude on the feature darkening near the flaring pil as what have been described in our previous observations ( e.g. , liu et al .
the formation of new penumbra also signifies the downward push of fields due to lorentz force change , as implied by the more horizontal field lines after flares ( hudson et al .
2008 ; fisher et al .
as soon as some horizontal fields subside close to the surface , the penumbra can be formed .
the opposite process was also observed before ( wang , deng & liu 2012 ) , in which the peripheral sunspot penumbra may disappear suddenly after flares , so that convection in the granule scale appears immediately as the fields turn from horizontal to vertical state . in both cases
, penumbra may be quite a thin layer of structure , as discussed by wang , deng & liu ( 2012 ) .
although the study of larger scale flare activity is out of the scope of this letter , it is worth mentioning that the present flare involves additional brightenings in other areas in the active region .
in fact , by looking at figure [ f1](f ) , we find a magnetic field configuration favorable for circular - ribbon flares ( e.g. , pariat et al .
that is , positive field surrounds the central negative field , and the negative field also exhibits obvious motion . indeed , during this flare , euv images taken by the atmospheric imaging assembly ( aia ) show a semi - continuous circular ribbon in the outer positive field together with the more compact ribbon in the central negative field region .
the two ribbons in figure [ f1 ] as seen by nst represent part of the outer ribbon and most of the inner flare core . this circular flare picture is supported by the outward ejecting jets as clearly seen in the aia movies .
flares having a circular ribbon have rarely been reported , although it is expected in the fan - spine magnetic topology involving reconnection at a 3d coronal null point ( masson et al .
2009 ; reid et al . 2012
; wang & liu 2012 ) .
it is speculated that the eruption of flux rope may have been triggered by the outer null - point reconnection ( e.g. , jiang et al .
2013 ) , which cause homologous jets as observed by aia .
we are grateful to the referee for the valuable comments to improve the paper .
we thank the bbso staff for obtaining the outstanding nst observations , and the sdo / hmi team for the magnetic field data .
we are indebted to dr .
yurchyshyn for pre - processing and transferring the nst data .
this work is supported by nsf grants ags 1153226 and ags 1153424 , and nasa grants nnx13ag13 g and nnx13af76 g . | rapid , irreversible changes of magnetic topology and sunspot structure associated with flares have been systematically observed in recent years .
the most striking features include the increase of horizontal field at the polarity inversion line ( pil ) and the co - spatial penumbral darkening . a likely explanation of the above phenomenon is the back reaction to the coronal restructuring after eruptions : a coronal mass ejection carries the upward momentum while the downward momentum compresses the field lines near the pil .
previous studies could only use low resolution ( above 1 ) magnetograms and white - light images .
therefore , the changes are mostly observed for x - class flares .
taking advantage of the 0@xmath01 spatial resolution and 15 s temporal cadence of the new solar telescope at big bear solar observatory , we report in detail the rapid formation of sunspot penumbra at the pil associated with the c7.4 flare on 2012 july 2 .
it is unambiguously shown that the solar granulation pattern evolves to alternating dark and bright fibril structure , the typical pattern of penumbra .
interestingly , the appearance of such a penumbra creates a new @xmath1 sunspot .
the penumbral formation is also accompanied by the enhancement of horizontal field observed using vector magnetograms from the helioseismic and magnetic imager .
we explain our observations as due to the eruption of a flux rope following magnetic cancellation at the pil .
subsequently the re - closed arcade fields are pushed down towards the surface to form the new penumbra .
nlfff extrapolation clearly shows both the flux rope close to the surface and the overlying fields . |
recently , mid infrared spectroscopy has been used increasingly as a tool for studies of ultraluminous infrared galaxies ( ulirgs , @xmath5 , see sanders & mirabel ( 1996 ) for a review ) .
the presence of starbursts and active galactic nuclei ( agn ) in ulirgs ( and coexistence in some of them ) has been known for quite some time , but the question which of the two dominates the luminosity has been difficult to answer because of the large columns of obscuring dust found towards the nuclear regions of these gas- and dust - rich systems .
high sensitivity and complete coverage of the infrared spectrum by esa s infrared space observatory ( iso , kessler et al .
1996 ) have considerably advanced the use of infrared spectroscopy to penetrate this obscuring dust , and to probe for the sources of the huge luminosity of ulirgs .
fine structure line and aromatic emission feature observations with iso - sws and isophot - s of a sample of 15 bright ulirgs suggest that most ulirgs are predominantly starburst - powered @xcite .
the pah method has been extended to a larger sample @xcite , allowing to probe for evolutionary effects expected in a scenario where starburst activity gives way to a quasar - like agn buried inside the ulirg @xcite .
comparison of optical and mid - infrared spectroscopic diagnostics @xcite demonstrated surprisingly good agreement if optical liners are interpreted as starbursts .
this finding also suggests that agn in ulirgs usually make their presence known optically at least in certain directions , instead of being fully embedded by large obscuring columns of dust .
this paper presents the result of a solicited program in iso open time , which addresses two main issues only partly covered by previous work .
first , and most important , at which luminosity does the transition from ` predominantly starburst - powered ' to ` predominantly agn - powered ' occur ?
on one hand , the iso observations of genzel et al .
( 1998 ) and lutz et al .
( 1998 ) suggest that not only luminous infrared galaxies with @xmath6 , but also most ulirgs above this threshold are predominantly starburst powered . on the other hand , optical and mid - infrared observations of hyperluminous infrared galaxies @xcite support a previous consensus that agn dominate these hyperluminous systems ( but note the intriguing possibility that far - infrared peaks in some luminous agn may trace even higher star formation related luminosities , e.g. haas et al .
the luminosity at which _ local _ ulirgs ( @xmath7 ) on average switch from starburst to agn dominated is a quantity with implications for various fields ranging from the interpretation of sources found in recent submm surveys @xcite to explanation of the hard x - ray background @xcite .
simple arguments based on the free fall time of the gas mass concentrated in the inner region of ulirgs and the nucleosynthesis efficiency suggest a maximum luminosity of a starburst event approaching @xmath8 @xcite , the limit depending on gas mass , spatial scale , and initial mass function properties .
observational evidence is clearly needed . despite its size of @xmath960 sources , the ulirg sample presented by lutz et al .
( 1998 ) and rigopoulou et al .
( 1999 ) is not optimal for this task : it is selected basically as flux limited at 60@xmath1 m and dominated by low to moderate luminosity ulirgs , because of the steep ulirg luminosity function @xcite .
the preliminary analysis of lutz et al .
( 1998 ) in the mid - infrared and a recent study of veilleux et al .
( 1999 ) in the optical / nir infrared showed that the presence of agn in ulirgs increases with the luminosity of the object .
however , these studies clearly call for improved statistics above @xmath10 in order to better determine the transition to agn - dominated systems . the second issue addressed by this paper
is related to the use of pah features as an agn / starburst diagnostic .
mid - infrared spectra of most galaxies show the 6.2 , 7.7 , 8.6 , and 11.3@xmath1 m features attributed to aromatic carbonaceous material @xcite . among several popular designations for these bands
will be designated hereafter by the terms pah ( polycylic aromatic hydrocarbon ) or uib ( unidentified infrared bands , because of remaining uncertainties on the precise nature of the aromatic carrier ) .
groundbased observations of these features and a companion at 3.3@xmath1 m first demonstrated that their equivalent width is larger in starburst galaxies than in classical agns @xcite .
iso spectroscopy has further strengthened this link by demonstrating the anti - correlation between feature strength relative to the continuum and the ionization state of the gas @xcite .
spatially resolved isocam - cvf observations of nearby agn support this interpretation by showing the pah features contrast to be weak near the central agn but strong in the circumnuclear region likely dominated by star formation ( e.g. cen a , ngc 1068 : @xcite , circinus : @xcite ) . despite this solid empirical basis , special care is required in analyzing pah spectra of ulirgs for agn or starburst dominance .
all components of the spectrum continuum and emission features , agn or starburst may be unusually obscured , and the limited wavelength coverage of isophot - sl spectra ( 5.8 - 11.6@xmath1 m observed wavelength ) makes continuum definition difficult . here , the extended wavelength range of isocam - cvf ( 5 - 16.5@xmath1 m ) is beneficial in reaching the long wavelength side of the broad 9.7@xmath1 m silicate absorption feature .
our paper is organized as follows : we discuss sample selection , observations and data analysis in section 2 . section 3 presents the observational results .
section 4 establishes a new quantitative method to determine the contribution of starburst and agn activity to the mid - infrared spectra , and compares it with the method used by genzel et al .
( 1998 ) , lutz et al .
( 1998 ) and rigopoulou et al .
( 1999 ) . in section 5 ,
we discuss the properties of ulirgs as a function of luminosity , combining our sample with the results of the isophot - s sample and with isocam results on hylirgs .
finally , we conclude in section 6 .
the solicited observing program presented here was defined in the last third of the iso mission when it had become apparent that low resolution spectroscopy of ulirgs was a promising tool , and that studies of ulirgs needed to be extended towards higher luminosities .
the sample selection is strongly driven by the visibility constraints of the iso satellite during the final six months of the mission .
ulirgs were selected from the samples of fisher et al .
( 1995 ) , kim ( 1995 ) , clements et al .
( 1996 ) , and the qdot sample ( lawrence et al . 1999 ) .
six sources were chosen from these catalogues with @xmath11 that had reasonable iso visibility during the remaining mission , plus an additional ten slightly lower luminosity sources that could be accommodated within the available observing time .
table [ tab : sources ] lists basic properties of the sample galaxies . throughout this paper
, we follow the convention reported by sanders & mirabel ( 1996 ) to compute the 8 - 1000@xmath1 m luminosity @xmath12 , adopting h@xmath13=75kms@xmath14mpc@xmath14 and q@xmath13=0.5 , and iras fsc fluxes ( iras 03521 + 0028 , iras 18030 + 0705 : psc ) .
most of our sources are lacking detections in some of the iras photometric bands . in those cases , we replaced the missing fluxes by estimates based on the 60@xmath1 m fluxes and average colors of the sanders et al .
( 1988 ) ulirgs .
the observations included the source iras 03452 + 2320 taken from the sample of fisher et al .
( 1995 ) , which is not listed in table [ tab : sources ] .
we did not detect a signal at its position .
we believe this reflects an infrared cirrus related misidentification in the fisher et al .
( 1995 ) sample rather than mid - infrared faintness of a true ulirg , since crawford et al .
( 1996 ) failed to detect the same source in radio continuum .
our sample is certainly heterogeneous : it was selected from ulirg surveys of different limiting 60@xmath1 m flux , with a preference for high luminosity targets but not implementing a strict lower cutoff , and strongly driven by iso visibilities .
none of the selection criteria , however , should significantly bias our sample in mid - ir spectral properties or agn content .
in particular , no agn - related iras color criteria like the 25/60@xmath1 m flux ratio were applied @xcite . a slight tendency towards favoring agn at higher redshifts / luminosities
is induced by the 60@xmath1 m selection which is effectively at @xmath945@xmath1 m rest wavelength for the most distant targets .
we have not attempted to correct for this since detailed far - infrared spectral energy distributions are unavailable for these faint iras sources , and since the effect will be negligible for our small and still fairly local sample .
the sample is hence suited to study the agn content of ulirgs , bridging the gap between the lower luminosity ulirgs studied with isophot - s @xcite and the two hyperluminous infrared galaxies ( hylirgs ) observed by iso @xcite .
figure [ dist_lir ] shows the distribution of infrared luminosities both for our sample and the combined iso database . the observations presented here used the circular variable filter ( cvf ) mode of the isocam camera @xcite rather than the isophot - s mode with which most other low resolution ulirg spectra were obtained and which is efficient in building a large sample of relatively bright sources .
the ultimate sensitivity achieved in a long cvf integration time is better , and the need for stabilization overheads was no major concern for our moderate number of very faint sources .
in addition , the extended long wavelength coverage is crucial in targeting objects up to [email protected] , and in probing an extended rest wavelength range .
since our sources are expected to be spatially unresolved , we have chosen the most sensitive setting with 6 pixels and a total fov of about 3 .
the observed wavelength range varies from source to source , the short end being defined by 5@xmath1 m rest wavelength and the long end either by 12@xmath1 m rest wavelength or , for high redshift sources , by 16@xmath1 m observed wavelength where the sensitivity of cvf observations starts to decrease quickly . in order to increase redundancy and to enable correction of the transients of the isocam detector , each cvf scan has been done in up and down direction . to reconcile the need for detector stabilization ( 10 frames of 5sec exposure at each wavelength ) with the total number of wavelength steps and practical limits on observing time , we have executed up- and down - scans with interleaved wavelength steps .
the total observing time per source including overheads was typically 2.5 hours . each final spectrum is composed by about 120 wavelength steps and has an average resolution of @xmath15 .
reductions have been carried out with the cam interactive analysis ( cia ) software .
the dark current has been estimated with the dark model described by biviano & sauvage ( 1999 ) .
this gives better results than using the two offset observations we have taken before and after the cvf observation .
we used the multi - median resolution deglitching method @xcite to remove glitches , and then additionally deglitched manually a square of 3 by 3 pixels centered on the source . to correct the transient effect of the detector the inversion method developed at ias
@xcite was used .
the calibration of the flux was performed using the standard spectral response function of cia .
appendix [ psf_red ] presents a reduction scheme which allows to quantify the spectrum of a target , with its s / n ratio and the systematic error introduced by the reduction method ( the total uncertainty on the flux is about 5 mjy ) .
it also allows us to determine the position of the fairly faint source with sub - pixel accuracy .
the typical uncertainty of the position is 5 , including s / n and iso pointing uncertainty .
one target ( iras 00406 - 3127 ) of the set had not been acquired by isocam because of a telemetry drop . for iras f02115 + 0226 there was no source detection .
the s / n ratio observed for iras f02455 - 2220 is not high enough to reliably determine the psf correction described in the appendix .
however , the observed spectrum can be taken as a lower limit , using only the signal from the brightest pixel .
for the remaining 13 targets , the data reduction has been applied successfully .
the rest - frame spectra for the 13 sources are shown in figure [ results_spectrum ] .
the shaded confidence area around the spectrum represents the rms plus the systematic error .
table [ tab : sources ] lists the position of the sources as obtained by the psf fitting method ( see appendix ) , and the iras fluxes .
all targets were found to be point like sources at isocam resolution .
one target has been found to be significantly offset from the iras psc position : iras 18030 + 0705 .
the offset is about one arcminute .
we checked the original iras data using xscanpi and found a source with fluxes similar to those listed in the psc , but centered on the isocam position .
the origin of this position discrepancy is unknown .
the fsc fluxes were adopted for the ulirgs detected by isocam except for iras 18030 + 0705 and iras 03521 + 0028 for which we used psc fluxes .
> from previous studies , the following components are expected to be present in ulirg spectra , but not necessarily together in _ one _ spectrum : * uib emission features at 6.2 , 7.7 , 8.6 and 11.3 microns , mainly originating from star forming regions of the galaxies .
this uib emission can be also coupled with a continuum of hot dust present in the most energetic regions of the starburst . *
an agn continuum from the hot dust present in the vicinity of the agn , e.g. in a dust torus or at larger scale in the narrow line region . * strong absorption , both continuous and in features like the silicate 9.7 @xmath1 m one .
absorption may arise either in the surroundings of the agn or the larger scale ism .
the amount of absorption may be different for agn and for starburst .
the silicate feature at 9.7@xmath1 m that is usually found in absorption may also mimic emission features . in dust configurations as in some torus models
@xcite , self - absorbed silicate emission can create an emission peak near 8@xmath1 m that is superficially similar to a faint uib emission and may be observed in one hylirg @xcite .
discrimination of such a ` fake uib ' peak and real uib emission has to rely on the detailed shape of the 8@xmath1 m peak and presence of the other uib features .
table [ tab : sources_rk ] summarizes the presence of the uib bands .
about half of the targets exhibit a spectrum with at least the 6.2 , 7.7 and 11.3 @xmath1 m mid infrared bands clearly visible .
the remaining targets do not exhibits all these bands , and two none of them significantly .
it is striking that almost all the observed ulirgs show a strong feature at or near 7.7 @xmath1 m .
iras f23529 - 2919 is an exception . here , the continuum might be so strong that it can dominate the emission of the 7.7 @xmath1 m emission band . also , the shape of the 7.7@xmath1 m feature is very different in two other galaxies from the typical uib band .
iras 22192 - 3211 and iras 00275 - 2859 have a band which is broader than the one observed in a `` normal '' starburst galaxy . the signal to noise for iras 23515 - 2917 is lower
so it is more difficult to draw a definitive conclusion .
a strong continuum has been detected in the two former targets , which supports the presence of an agn in both sources .
the change of the shape of the 7.7 @xmath1 m feature can be explained in two ways .
the carriers of the uib might be processed in the agn environment , changing the shape of the spectral emission of these carriers .
such changes in the shape of the 7.7/8.6@xmath1 m complex have in fact been observed in some ultra - compact regions @xcite .
alternatively , a `` fake '' 7.7 @xmath1 m feature could arise when one observes a high optical depth source with self - absorbed silicate emission .
there is no agreement in detail , however , between the broadish 8@xmath1 m features in these ulirgs and the spectral shapes predicted by torus models @xcite .
a further problem of a ` torus ' identification of these 8@xmath1 m features is that they are not observed in the spectra of template seyferts and qsos @xcite .
another source , iras f00183 - 7111 , has a maximum near 8@xmath1 m which we do not ascribe to uibs .
its spectrum is remarkably similar to some deeply obscured objects ( dhendecourt et al .
1996 , ngc 7538 irs1 : d.cesarsky unpublished ) , showing a strong 5 - 8@xmath1 m continuum and an extremely deep silicate absorption .
even an apparent `` 6.5 @xmath1 m emission '' is indicated in this object which is very likely the result of the absorption of water ice at 6 @xmath1 m and an absorption feature at 6.85@xmath1 m which is of unclear origin @xcite .
the spectrum of iras f00183 - 7111 is thus dominated by a heavily absorbed ( agn ) continuum with a possible weak 7.7@xmath1 m emission , which might be either due to a classical uib or self - absorbed silicate emission . in total
, it appears to be the only unambiguous example in our data for a type of source considered typical in the classical sanders et al .
( 1988 ) scenario : an extremely obscured agn .
the use of the line - to - continuum ratio of the 7.7 @xmath1 m band ( hereafter l / c ) as proposed by genzel et al .
( 1998 ) and lutz et al .
( 1998 ) to distinguish in the _ mid - infrared _ between starburst - dominated and agn - dominated ulirgs is based on evidence that a starburst galaxy presents strong uib features while agn presents strong continuum with a low contrast emission features in the nuclear part of the galaxy .
this is suggested by ground based observations @xcite , iso spectroscopy @xcite , and iso spectro - imaging @xcite .
this also implies that the uib luminosity of a galaxy scales with its star formation .
recent studies tend to reinforce this idea : in spiral galaxies , the global h@xmath16 emission of the disk is correlated with the total mir emission of uib @xcite . when using these features as tracers of the sources of the total ulirg luminosity , a further assumption is that the mid - infrared actually probes the active regions and not just a surface layer .
this is not trivial since the active regions of ulirgs are small @xcite and dusty .
depending on dust configuration , a major part of the luminosity could be hidden from mid - infrared view .
genzel et al .
( 1998 ) have addressed this problem using the starburst - type mid - infrared nebular spectra of ulirgs .
for ulirgs , the ratio of the inferred starburst ionizing luminosity to the bolometric luminosity approaches that for template starbursts , indicating that more than half of the luminosity has been traced by the mid - infrared observations .
similar analyses can be made using the luminosity of the pah features .
rigopoulou et al .
( 1999 ) and genzel & cesarsky ( 2000 ) find that observed pah luminosities of ulirgs are well above those of lower luminosity sources , the ratio to bolometric luminosity being only a factor 1.4 - 2 lower than in starbursts .
this factor is plausibly explained by the higher extinction in ulirgs and perhaps the effect of the more intense radiation field on the pah emission .
we have tentatively repeated this comparison using the extinction correction derived from our fits described below , which however give only fairly uncertain extinction for individual noisy spectra .
the mean discrepancy narrows down to a factor 1.2 .
thus , the mid - infrared diagnostics are probing the major part of the luminosity . for a narrow line and emission from a single region ,
both line and continuum are obscured by the same amount .
the width of the pah feature and the strong wavelength dependence of extinction at the onset of the silicate feature make the pah l / c less ideal . in some objects
the absorption might be so strong that it can mimic a 7.7 band .
there are , however , differences in shape and the l / c will normally stay below the cutoff of 1 adopted in previous papers .
iras f00183 - 7111 is the only clear example in our sample .
another technical point that has to be verified is the determination of the continuum under the 7.7@xmath1 m feature . in the l / c diagnostic , this continuum is deduced from the linear interpolation of the continuum levels at 5.9 and 10.9@xmath1 m , i.e. the l / c ratio is affected by the flux uncertainties at these wavelengths .
this is also a systematic source of uncertainty since the 10.9@xmath1 m point is still within the silicate feature and does not probe a clean continuum .
we have explored alternative methods in order to verify the robustness of the l / c ratio with respect to the systematic error induced in the 7.7@xmath1 m continuum determination .
first , we constructed the `` narrow band '' ratio between the flux density at 7.7 and 6@xmath1 m .
strong uib emission will create a high ratio which is not reached by a pure continuum unless it is unrealistically steep over this small wavelength interval .
this diagnostic should hence identify strong uib emission easily but will face problems discriminating weak uibs from continuum slope variations .
the strength of this diagnostic is to avoid any assumption on the shape of the spectrum in the region obscured by the silicate feature .
figure [ lc_nb ] shows the correlation between the l / c and f(7.7)/f(6 ) diagnostics seen for ulirgs and template spectra observed by phot - s @xcite , sws @xcite , and cam - cvf ( this paper ) . the dispersion in this plot
is explained by the variation in continuum slope and the systematic error of the l / c discussed before .
the good correlation of the two diagnostics supports the use of the l / c diagnostic to identify systems that are clearly starburst - like and show strong uibs . by definition of the f7.7/f6 diagnostic , figure [ lc_nb ] can not test the effect of continuum placement uncertainty in the l / c method on measurements of faint residual uibs in agn - like systems . the line f(7.7@xmath1m)/f(6@xmath1m)=l
/ c+1 is also overplotted in figure [ lc_nb ] to indicate the asymptotic case where the continuum is flat .
to investigate further how the different spectral components influence the mir spectrum of a galaxy , we created another diagnostic whose basis is a more physical parameterization of the spectrum .
our aim is an empirical fitting procedure using the complete mid - infrared low resolution spectra to derive the contribution of the agn and starburst components to the mid - infrared spectrum . in a second step
, results will be extrapolated to the bolometric luminosities .
the fit is based on template spectra of dusty starbursts and agn , and applied to the whole set of ulirg and template spectra observed by phot - s @xcite , sws @xcite , and cam - cvf ( this paper ) .
since absorption plays an important role for the spectra of dust - rich ulirgs , we incorporate the effects of absorption that is higher than in the templates . in principle , explaining the mid - infrared spectra of ulirgs is a problem of three dimensional radiative transfer requiring a consistent treatment of dust , ices , and pah features .
considerable insight can be gained , however , from a simple superposition of a screen - obscured starburst and a screen - obscured agn .
arguments in favour of this approach are : ( 1 ) the mid - infrared diagnostic reaches most of the active region .
unknown more obscured sources will be minor contributors .
( 2 ) the use of an observed agn template implicitly accounts for warm dust close to the agn .
screen extinction is a reasonable first approximation for the effect of even more dust around the agn , since this dust will mostly just absorb in the mid - infrared consider here , and reemit in the far - infrared .
( 3 ) an aggregate of star forming regions is in principle better represented by a case where emission and absorption are well mixed .
mixed case extinctions are difficult to determine , however , unless there are probes measured at high s / n over a large range of optical depths , a requirement not met by our spectra .
the screen case adopted here will give similar results for optical depths around one , fall short at higher optical depth , but be conservative in the sense of not inferring excessive starburst activity in uncertain cases .
genzel et al .
( 1998 ) , rigopoulou et al .
( 1999 ) and sturm et al .
( 2000 ) showed that in the range of the 6 - 9@xmath1 m uib bands there is little variation in starburst mir spectra observed by sws and phot - s .
starburst spectra present some variations at longer wavelength , however .
( 1996 ) , noted that in some regions of ngc 4038/4039 the flux at 15 @xmath1 m can rise dramatically and dominate the mid infrared spectrum .
it can be shown that the mir spectrum of starburst galaxies is the superposition of two components : one from photodissociation regions ( pdr ) and one from regions @xcite .
the pdr component is responsible for the uib emission and the region for the hot dust continuum which becomes important at 12 @xmath1 m @xcite .
this hot dust continuum is probably due to very small grains heated to temperature as high as 200k in some regions .
this continuum is very faint at 7@xmath1 m @xcite , and is not really noticeable in the spectrum of the majority of starburst galaxies where the pdr contribution ( _ i.e. _ the uib contribution ) dominates in this part of the spectrum @xcite . for some dwarf
metal - deficient starburst galaxies like ngc 5253 and iizw40 ( rigopoulou et al . 1999 , s.madden private communication ) , uibs are nearly absent .
in such a case a strong featureless continuum will not be an agn signature .
since ulirg are dust rich and not known to be metal - deficient this effect is likely not relevant , and we still consider a strong ulirg continuum with weak uibs an agn signature .
as the ( rest)-wavelength range investigated in this paper is between 5 and 12 @xmath1 m , the contribution of the hot dust continuum from regions will not change drastically the spectrum of the starburst contribution .
but one _ must _ keep in mind that strictly speaking a significant continuum at short wavelengths ( 5 - 8 @xmath1 m ) is not necessarily a tracer of an agn , but could reveal a higher weight on regions in the ism with respect to photodissociation regions . to break the degeneracy , one has to look at the continuum at 15 @xmath1 m since the continuum of an region is steeper than the continuum of an agn @xcite . in the present model
we assume that the template spectrum of a starburst does not change in a significant way between 5 and 12 @xmath1 m ( see section 4.2 for the self consistency of this hypothesis ) .
* we have used the m82 sws spectrum as starburst template spectrum * , supposed to represent the standard line to continuum ratio of the uib seen in a dusty starburst .
nuclear agn spectra have low contrast uib features and mainly present a continuum emitted by hot dust .
obviously , when looking at the total spectrum of a galaxy hosting an agn , one naturally observes uib bands from circumnuclear star formation or from the disk of the agn host ( see for instance @xcite ) . *
the agn spectrum itself can be described by an intrinsic continuum which can be well fitted by a power law of variable index * @xcite subject to absorption by dust especially in the region of the 9.7@xmath1 m silicate absorption feature .
as discussed in the previous subsection , a continuum seen in the mid - infrared may also be due to regions . in the wavelength range considered ( 5 - 12 @xmath1 m )
, there is a degeneracy between the agn continuum and a continuum coming from an region , although an continuum is usually steeper and fainter .
our fit invokes a single ` continuum component ' including both possibilities of an agn continuum and an unusually strong continuum .
we will come back to the degeneracy between these two types of continua when discussing fit results .
because ulirgs are gas and dust rich galaxies @xcite , the extinction is expected to play a crucial role on shaping the spectrum of these galaxies , even in the mid - infrared ( see for instance arp220 , @xcite ) .
it is important to test how dust obscuration can affect both uib dominated and featureless spectra @xcite .
one can argue that mid - infrared ulirg spectra could be completely dominated by absorption , and that the 7.7@xmath1 m feature , which we identify as an uib , could result from a deep absorption of a flat continuum @xcite
. this could be the case in the strongly absorbed galaxy f00183 - 7111 .
we want to test more generally whether it is possible to produce an emission feature around 8 @xmath1 m by a power law continuum affected by an absorption law or whether a simple uib spectrum affected by absorption is preferred .
we used two absorption laws , the ` classical ' draine and lee ( 1984 ) absorption law and one composite absorption law derived from the galactic center - sgra * spectrum ( hereafter gc or gc - sgra * ) .
the gc extinction law has a higher extinction level between 3 and 7@xmath1 m @xcite .
since this law is derived from hydrogen recombination lines and does not sample the silicate feature well , we have supplemented it with points deduced from the gc continuum spectrum itself . they have been computed assuming that the intrinsic gc spectrum is a power law spectrum absorbed by a dust screen .
the extinction law and the exponent of the power law have been adjusted in order to fit the extinction deduced from the hydrogen lines and the gc spectrum itself .
this procedure is adequate since we cover only a limited wavelength range , and since the dust in front of the galactic center is close to the ideal screen case .
this absorption law has two main peculiarities : a flatter absorption between 3 and 7@xmath1 m , and a narrower silicate feature than what the draine and lee curve predicts ( see figure [ gc_ext ] ) .
these two absorption laws have been used to represent variations in dust properties as possible in extreme environments such as close to an agn .
comparison of the m82 and ngc1068 sws spectra @xcite suggests such variations at least for absorption _
features_. because of the possibility of such variations , absolute a@xmath17 values derived from the model are somewhat uncertain .
the main emphasis is to test whether the shape of the ulirg spectra can be reproduced well .
the two spectral components and obscuration have been used to construct four types of models in two classes : * the first class is based on the assumption that the starburst component and the continuum component are obscured by the same dust screen .
models of this class vary in extinction properties : the dust extinction law is either a `` pure '' draine and lee law , or a `` pure '' gc extinction law or a mixture of the two .
* the second class allows different obscuration of the two components and is clearly more realistic .
the two different contributions ( continuum and starburst ) are not necessarily located in the same physical region and therefore do not suffer the same extinction .
table [ tab : mod_descr ] provides the formal description of the different models used .
models 1a , 1b and 2 refer to the first class , while models 3a - d and 4 refer to the second class . in two cases ( 2 and 4 ) , the extinction law is a linear combination of the gc extinction law and the draine and lee one .
this combination is not constrained by observations but included in our modeling as a comparison to other models with different degrees of freedom . to validate the model , the fit quality
is quantified first .
second , the results from test fits to template starburst and agn spectra are investigated , and compared to observations of `` mixed '' sources where the agn and starburst relative contributions are well known from spatially resolved data .
figure [ disp_fit ] shows some examples of fits to ulirg spectra using model 3a .
the figure includes agn and starburst - like spectra of both high and low s / n .
fits using other models tend to be slightly worse but still acceptable .
the first conclusion is that the models presented in table [ tab : mod_descr ] can give a reasonable fit for all ulirg spectra .
this implies that the assumptions of these models ( _ e.g. _ a mir spectrum of a ulirg is a superposition of a continuum component and a starburst component extincted by a dust - rich ism ) are not in obvious contradiction with observations . to better quantify the quality of the various models we adopt the indicator @xmath18 where @xmath19 is the observed flux density
, @xmath20 is the model flux density , @xmath21 is the rms of the observed flux density .
the summation is done over _ all _ ulirg spectra , and @xmath22 is the sum of degrees of freedom of modeled spectra . with this indicator
the quality of the different models can be compared : the closer to 1 . , the better .
table [ tab : comp_stddev ] presents the mean @xmath23 both for our complete ulirg sample , and for a high quality subset with observed s / n greater than 3 .
we have obtained model fits for various starburst , agn , and mixed template sources .
a first result can be obtained from a comparison of the two classes of models for the mixed templates circinus and centaurus a. figure [ comp_visu ] shows the obvious superiority of models where the extinction towards the agn is different from that towards the starburst . fully equivalent good fits
are achieved for models 3a-3c .
figures [ verif_stb ] and [ verif_agn ] show the observed spectrum and the fitted spectrum computed with model 3a , for several starburst and agn templates .
the galactic center has been grouped with the agns since the very center shown here is also a continum source devoid of uibs @xcite .
the absence of these features in the gc - sgra * spectrum is due to the presence of an -region with an intense radiation field that probably destroys the uib carriers .
we overlay the contributions of the starburst component and the contributions of the continuum component .
as expected , the starburst component dominates totally the modelled spectrum for the starburst galaxies .
similarly , a quasi - pure continuum contribution is needed to fit the agn spectra , with extinction being negligible in many cases but significant in others ( ngc 5506 ) .
the low metallicity galaxy ngc 5253 ( not shown ) is well fitted by a strong continuum contribution and weak uibs .
the fits to the starburst templates ( fig .
[ verif_stb ] ) show at their long wavelength end a varying contribution of a faint steeply rising continuum .
most likely , this is due to variations in the strength of the 15@xmath1 m vsg ( very small grains ) continuum , a conclusion also reached by sturm et al .
( 2000 ) on the basis of the sws spectra of m82 and ngc 253 . a further test is to apply the model to objects which are well known to be composite , containing both starburst and an agn . in figure [ verif_mix ]
spectra of four such objects are presented , showing plausible decompositions into the starburst and continuum components .
a stringent test can be obtained for centaurus a and circinus which have both been observed by isocam / cvf at high spatial resolution ( mirabel et al .
1999 ; moorwood et al .
1999 and in preparation ) .
these observations allow us to separate the contribution of the nucleus and the surrounding star - forming regions , and thus a direct test of our model decomposition .
the spectrum of the centaurus a nucleus , extracted from the isocam - cvf observation , is overplotted in figure [ verif_mix ] .
the modelled continuum contribution is indeed very similar to the observed spectrum of the nucleus , confirming the validity of our approach .
we performed the same comparison for the circinus galaxy ( a.f.m .
moorwood , private communication ) and find the same good agreement between the observed nuclear continuum and the modelled continuum contribution .
at least for these two sources , the separation into two contributions ( continuum and starburst ) is reliable .
table [ tab : comp_stddev ] shows that the two classes of models do not reproduce the ulirg spectra equally well .
it is obvious that the models of the second class , where extinctions of continuum and starburst are different , are in better agreement with the observations .
this better accuracy is not explained trivially by the increasing number of free parameters .
model 2 which has the same number of free parameters as the ` good ' models 3a - c is clearly less successful .
that means , more flexibility in the shape of the extinction does not improve the fit significantly , while invoking different extinction towards the starburst and agn does .
this lends support to the plausible idea that two different extinctions are applicable to the physically distinct starburst and agn parts of a ulirg . comparing the continuum and starburst extinction derived by our fit procedure
, we find the continuum extinction to be slightly smaller on average .
this is plausible if the obscuring dust is not distributed with spherical symmetry .
we note , however , that this result should be considered tentative given the considerable uncertainty of individual a@xmath24 values .
figure [ lc_ratas ] compares the l / c ratio used in previous papers with the ratio of the integrated 5 - 10@xmath1 m fluxes contained in the continuum contribution and in the starburst contribution as deduced from the model 3a . in this plot
we also add the starburst and agn comparison samples as described by rigopoulou et al .
the two indicators are well ( anti)correlated .
this correlation shows _ a posteriori _ the consistency of the two indicators .
table [ tab : ulirg_ratas_lc ] displays the individual value of l / c and the ratio of the integrated 5 - 10@xmath1 m fluxes contained in the continuum contribution and in the starburst contribution .
genzel et al .
( 1998 ) and subsequent papers adopted a l / c threshold of 1 to discriminate between starburst and agn dominated systems .
at this threshold , the flux ratio of the continuum and starburst components , each integrated over the 5 - 10@xmath1 m range , will be larger than one , simply because the uibs are relatively narrow features .
this is reflected in figure [ lc_ratas ] .
the purpose of the threshold , however , is to identify the component dominating the bolometric luminosity and not just the 5 - 10@xmath1 m flux .
the mid - infrared contributions have to be extrapolated to the bolometric contributions . using the spectral energy distributions of m82 and of the central region of ngc1068 as starburst and agn templates ( sturm et al .
2000 for m82 , including iras points and correcting for aperture effects ) , we find the ratio @xmath25 about 2.5 times greater in the case of the agn than for starbursts .
very similar values are obtained for other agn templates ( ngc 4151 , 3c273 ) , consistent with the notion of stronger mid - ir continua in seyferts than in starbursts @xcite .
orientation related variations in mid - infrared seyfert continua @xcite caution that there will be some scatter in these bolometric corrections for individual agn . in the remainder of this paper
, we will apply the factor of 2.5 wherever total infrared ( i.e. bolometric ) quantities are discussed .
taking into account the correction factor 2.5 , figure [ lc_ratas ] confirms the l / c threshold of 1 as a sensible one for differentiating between sources dominated by the starburst component and those dominated by the agn component .
determination of the power source of the highest luminosity ulirgs is a prime goal of our project , extending the isophot - s sample which is dominated by lower luminosity sources @xcite .
figure [ lc_comp ] displays the l / c ratio for the cam - cvf ulirgs together with data from lutz et al .
( 1998 ) and the two hyperluminous sources observed by iso @xcite .
the uib l / c is used to classify sources as starburst - like if l / c @xmath261 or agn - like if l / c @xmath27 1 . from our enlarged database ,
we confirm that most ulirgs are starburst dominated in the mid - infrared but that there is a trend towards agn - like ulirgs at higher luminosities . in our sample which contains 17 sources above @xmath28 , the highest luminosity source classified as starburst is at @xmath29 ( iras 03835 - 6432 ) .
such a luminosity limit for the ` most luminous starburst ' is in agreement with the timescale argument given by heckman ( 1994 ) . adopting the bolometric luminosities of our sources and sizes like those derived by soifer et al .
( 2000 ) for the mid - ir emitting region of ulirgs , the surface brightness will exceed the limit for starbursts suggested by meurer et al .
individual star forming region are known to exceed this limit , however , and the huge gas concentration in the centers ulirgs provides sufficient fuel
. additionally , meurer et al . (
1997 ) derived their limit using , for similar sources , h@xmath16-based sizes which are larger than the mid - ir ones found by soifer et al .
( 2000 ) , thus staying below the actual surface brightness . even in our sample enriched in luminous ulirgs , statistics at the highest luminosities
is limited , thus maintaining the possibility for higher luminosity starbursts .
given the modest sample size , it is difficult to assign an accurate value for the luminosity at which half of the sources are starburst and half agn dominated . from figure [ lc_comp ] , this occurs most likely at luminosities @xmath30 to @xmath31 .
the trend from starburst dominance for most ` normal ' ulirgs to agn dominance at the highest luminosities now consistently emerges from a variety of indicators : mid - infared continuum @xcite , mid - infrared spectroscopy ( lutz et al .
1998 and this paper ) , and optical spectroscopy @xcite .
as noted by @xcite , the good agreement between optical and mid - infrared suggests that agn in ulirgs do not remain fully embedded for long but manage to break the obscuring screen at least in certain directions .
the highest luminosity ` starburst - like ' systems appear well above our cutoff l / c of 1 , but below the l / c@xmath32 typical for starbursts and lower luminosity ulirgs . this can be interpreted in different ways .
either they contain a noticeable though not dominant agn component , or the increased continuum contribution is due to an increase in the -region related continuum , as might be reasonable for sources with very high star formation rates in a small region .
the simple l / c method will not be able to break the degeneracy for the origin of ulirg mid - infrared continua . also , the parameterization chosen for our fits ( sect .
[ pres_mod ] ) describes the spectrum of a ulirg in terms of just two components : a uib - dominated starburst component and a continuum component which is either due to an agn or extra region contributions .
if the relative weight of ` pdr - like ' and ` region - like ' components varies for different starburst conditions , as seen e.g. for different regions of the antennae galaxies @xcite , assigning the continuum entirely to an agn may be misleading .
the nature of the continuum has to be investigated in more detail to quantify the fraction of energy radiated by the agn ( agn continuum ) and by star formation activity ( uib - dominated component plus region continuum ) .
it is possible to break the continuum degeneracy by using the slope of the continuum ( laurent _ et al .
the ratio of the fluxes at 6@xmath1 m and 15 @xmath1 m is different for an region and a continuum emitted by the dust ( torus ) surrounding an agn .
dust in the inner parts of regions emits a mid - infrared spectrum that is steep over our wavelength range .
this can be seen both in galactic regions @xcite and in extragalactic regions of particularly concentrated star formation , e.g. the star forming knot a in the antennae galaxies @xcite .
in contrast , the continuum emitted by hot agn dust is much flatter in the mid - infrared , as shown by the examples of cen a @xcite and ngc1068 @xcite . because of the presence of the uib emission and silicate absorption features ,
the steepness of the continuum can be determined reliably only with sufficient wavelength coverage , ideally from 6@xmath1 m or less to 15 @xmath1 m ( rest wavelength ) .
this condition is not met by the isophot - s spectra which form most of the iso database of low resolution ulirg spectra but cut at less than 11@xmath1 m rest wavelength for most ulirgs .
it is partially met for the isocam - cvf spectra presented in this paper which typically extend to rest wavelength about 12@xmath1 m .
it is instructive to discuss the case of arp 220 for which a full cvf spectrum is available @xcite .
arp 220 shows 6.2 and 7.7@xmath1 m uibs on no or weak continuum , whereas the continuum in the 11 - 15@xmath1 m region is clearly stronger than in a standard m82-type starburst spectrum .
this suggests a strong but steep continuum as in the star forming knot a of the antennae , subject to even higher extinction ( see figure [ arp220_knota_comp ] ) .
the case is less clear for our spectra which do not extend to 15@xmath1 m .
some spectra suggest a strong but steep continuum at the long wavelength end of a uib dominated spectrum ( e.g. iras 03521 + 0028 ) while others are fit by a uib dominated spectrum plus a flat continuum ( e.g. iras 18030 + 0705 ) . at the limited wavelength coverage ,
the uncertainties are considerable , also taking into account that fits are affected by some 7 - 8@xmath1 m features not showing the typical pah shape ( sect .
[ results ] ) .
our model can be used to better quantify the contribution of starburst and agn activity independent from the l / c criterion used earlier .
we measured the mid infrared fluxes emitted by starburst and continuum contribution ( not dereddened ) integrated over the wavelength range 5 - 10@xmath1 m ( similar to the filter lw2 of isocam ) .
when discussing total infrared ( bolometric ) contributions instead of 5 - 10@xmath1 m ones , the different agn and starburst seds are considered by increasing the starburst contribution by the correction factor 2.5 derived in sect .
[ res_mod ] . since we were not able to reliably break the degeneracy of agn and region continuum for our wavelength coverage and signal - to - noise ratio , we explored two extreme cases for the origin of the continuum , reality most likely being in between .
the first scenario assumes that the continuum is coming entirely from an agn .
the examples of region continua discussed above , however , clearly argue for _ some _ star formation contribution to the continuum .
it is unlikely that an almost pure continuum spectrum is produced in this way . the only star - formation powered galaxies approaching this case are low metallicity dwarfs like ngc 5253 , an unlikely match to the ulirg situation .
we hence have adopted the following second extreme scenario : all faint continua , defined as emitting less than 50% of the total 5 - 10@xmath1 m flux , are assumed to be due to star formation , while all stronger ones are due to an agn .
figure [ lir_mirfrac ] shows the median ratio of the energy emitted by the starburst component to the total energy emitted ( @xmath33/@xmath34 ) of ulirgs presented in table [ tab : ulirg_ratas_lc ] and two hylirgs @xcite , versus the total infrared luminosity .
a trend towards agn dominance at high luminosity is again clearly seen .
the cutoff between starburst - dominated galaxies and agn - dominated galaxies lies at @xmath35 .
the significance of this trend can be tested by dividing the luminosity range into the bins @xmath36 and @xmath37 and using respectively the kolmogorov - smirnov s test and the student s t - test to test if the two samples follow the same distribution or have the mean value .
the probability for such hypotheses are @xmath38(resp .
@xmath39 ) , demonstrating that the trend is statistically very significant .
we also perform the two statistical test while removing the two most luminous galaxies ( @xmath40 ) .
probabilities are then @xmath41 and @xmath42 .
to quantify the agn contribution at different luminosities , we identified either all or none of the continuum of each source as agn according to the two scenarios described previously .
we then determined the ratio @xmath43/f@xmath44 , with @xmath45 being the sum of uib - dominated starburst flux and that part of the continuum flux that is ascribed to star formation under the given scenario . for each scenario , we averaged the results for the ulirgs in each of the two luminosity bins .
table [ tab : ratas_lir ] displays the resulting average values along with their uncertainties .
the average contribution of star formation to the 5 - 10@xmath1 m flux is 65 to 84% for the low luminosity bin and 24 to 33% for the high luminosity bin , the ranges representing the two extreme scenarios . extrapolating from mid - infrared fluxes to total far - infrared ones , the average contributions of star formation are 82 to 94% for the low luminosity bin and 44 to 55% for the high luminosity bin .
these results from quantitative modeling considerably reinforce the conclusions from the earlier work based on the l / c ratio : low luminosity ulirgs are predominantly starburst powered , while agn powering dominates above @xmath46 to @xmath3 .
we have presented isocam - cvf low resolution mid - infrared spectroscopy of 16 ultraluminous infrared galaxies obtained within a program targeted at investigating the most luminous ulirgs .
the data are analyzed along with the complete iso database of low resolution spectra of ulirgs / hylirgs , totalling 76 sources .
we use the presence of the mid - infrared ` uib ' aromatic bands as a diagnostic of starburst activity , finding starburst - dominated systems up to a luminosity of @xmath47 .
other spectra show a strong agn continuum with weaker features of uncertain origin .
we have found one highly obscured agn .
the mid - infrared spectra can be modeled in terms of a superposition of a uib - dominated starburst spectrum and a continuum , both potentially strongly obscured .
this continuum contains both agn and additional region contributions .
the fits prefer a two zone model in which the extinction to starburst and continuum regions may differ .
results from these fits agree well with previous simpler diagnostics based on the line - to - continuum of the uib features .
low luminosity ulirgs are mostly starburst dominated , but the agn fraction increases with luminosity and most ulirgs above @xmath2 to @xmath3 appear agn like .
we have separated ulirgs into two luminosity bins of @xmath48 and @xmath37 .
we find that the average contributions of star formation to the infrared luminosity are 82 to 94% for the low luminosity bin and 44 to 55% for the high luminosity bin .
we are grateful to michael rowan - robinson for allowing us access to the qdot ulirgs prior to publication .
the work of qdt is supported by verbundforschung ( 50 or 9913 7 ) .
dch acknowledges support from nasa grant nag5 - 3359 to us iso observers .
dr acknowledges support of the ec tmr network `` infrared surveys '' ( under contract number erbfmrx - ct96 - 0068 ) .
yt is supported by the ministry of education , science , sports and culture of japan under grants 10044052 , and 10304013 .
sws and the iso spectrometer data center at mpe are supported by dlr ( dara ) under grants 50 qi 8610 8 and 50 qi 9402 3 .
aiming for a good cam - cvf spectrum including a reliable continuum , a main source of uncertainty is due to the point spread function ( psf ) . because the width of the psf changes with the observed wavelength , it is crucial to determine the observed total flux for each wavelength separately .
this is easily done for bright point sources because it is easy to sum the observed flux over a large box ( typically 5 by 5 pixels ) or to fit the psf profile to the observed image . since our sources are point - like at cam spatial resolution , it is not necessary to consider the complex situation for extended sources . even for point sources , however , the psf correction is difficult if they are faint .
then , only a few pixels around the brightest one have a significant s / n ( hereafter the brightest pixel will be called `` central pixel '' ) . in such a case , the s / n of the corrected spectrum obtained by summing over a box or fitting the psf profile is usually very low .
a better s / n can be obtained just using the signal of the central pixel and applying a psf correction factor that varies slowly with wavelength .
the main difficulty in determining this factor is to obtain the position of the source within the central pixel .
this could be done by fitting a model of the psf to the observed images at each wavelength .
but , because the s / n is quite low for the pixels around the central pixel , direct application of this method gives uncertain positions and a final spectrum where the s / n is low , comparable to the s / n obtained by summing over a box . in order to improve the s / n we adopted the following method : in cia , theoretical psfs are available at every half micron of the isocam - cvf wavelength range , residual variations over such ranges are small .
we divided the observed wavelength range into sub - ranges of one micron centered around the wavelength of a theoretical psf , and summed images corresponding to wavelengths in this range of one micron around the given wavelength ( i.e. about 10 images ) .
we then fitted a psf profile to the resulting image ( hereafter the summed image ) .
this method improves the s / n , and the deduced position of the source in the central pixel is known with a better confidence level .
we used this position and the psf to derive the correction factor from flux in the central pixel to total flux , which was applied for each image in the considered wavelength range .
although this method gives a better position of the source , it is still ncessary to compute the systematic error introduced by this correction .
the formal way is to compute the s / n for each pixel of the summed image , and then to use an inverse monte - carlo method to compute the s / n of the spectrum .
another equivalent way which avoids a monte - carlo simulation is to consider the images used to obtain the summed image _ as guesses of a monte - carlo simulation _ .
this is equivalent to a simulation with ten guesses because the s / n of the summed image is deduced from these images .
we then used the positions of the source obtained for all these images to correct the flux of the summed image . > from the scatter of these measurements , it is then possible to deduce the systematic error introduced by the fitting method of the psf profile .
lrrcrrrrr iras name&ra(2000 ) & dec(2000)&z & @xmath49&@xmath50&@xmath51&@xmath52&log(@xmath53 + & & & & jy & jy & jy & jy & @xmath54 + f00183 - 7111 & 0:20:35.3&-70:55:22 & 0.327 & @xmath270.06 & 0.13 & 1.20 & 1.19 & 12.77 + 00188 - 0856 & 0:21:26.6 & -8:39:22 & 0.129 & @xmath270.12 & 0.37 & 2.59 & 3.40 & 12.31 + 00275 - 2859 & 0:30:04.4&-28:42:23 & 0.279 & @xmath270.08 & 0.17 & 0.69 & 0.73 & 12.46 + 00406 - 3127 & 0:43:03.2&-31:10:50 & 0.342 & @xmath270.06 & @xmath270.09 & 0.72 & 0.99 & 12.64 + 02113 - 2937 & 2:13:33.0&-29:23:35 & 0.194 & @xmath270.08 & @xmath270.12 & 0.94 & 1.88 & 12.29 + f02115 + 0226 & 2:14:10.3 & 2:40:00 & 0.400 & @xmath270.11 & @xmath270.16 & 0.32 & 0.64 & 12.48 + f02455 - 2220 & 2:47:51.3&-22:07:38 & 0.296 & @xmath270.08 & @xmath270.10 & 0.82 & 1.27 & 12.57 + 03000 - 2729 & 3:02:11.5&-27:07:24 & 0.221 & @xmath270.07 & @xmath270.11 & 0.92 & 2.04 & 12.41 + 03538 - 6432 & 3:54:25.3&-64:23:39 & 0.310 & @xmath270.06 & 0.06 & 0.99 & 1.30 & 12.65 + 03521 + 0028 & 3:54:42.4 & 0:37 : 8&0.152 & @xmath270.25 & 0.23 & 2.64 & 3.83 & 12.46 + 04384 - 4848 & 4:39:50.8&-48:43:11&0.213 & @xmath270.04 & 0.07 & 0.99 & 1.34 & 12.32 + 17463 + 5806&17:47:04.6 & 58:05:24&0.309 & @xmath270.04 & @xmath270.04 & 0.65 & 0.95 & 12.48 + 18030 + 0705&18:05:32.5 & 7:06 : 9&0.146 & @xmath270.25 & @xmath270.25 & 0.84 & 4.40 & 12.18 + 22192 - 3211&22:22:09.6&-31:56:34&0.231 & @xmath270.12 & 0.19 & 0.89 & 1.42 & 12.43 + 23515 - 2917&23:54:06.7&-29:00:58&0.336 & @xmath270.08 & @xmath270.14 & 0.65 & 1.06 & 12.60 + f23529 - 2919&23:55:33.3&-21:03 : 7&0.430 & @xmath270.08 & @xmath270.16 & 0.33 & 0.63 & 12.55 + lccccl iras name & 6.2 @xmath55 & 7.7 @xmath55 & 11.3 @xmath55 & l / c & note + f00183 - 7111 & n & n & n & 0.88 & + 00188 - 0856 & n & y & y & 1.53 & + 00275 - 2859 & n & y & n & 0.58 & + 00406 - 3127 & - & - & - & - & not acquired + 02113 - 2937 & n & y & y & 2.53 & + f02115 + 0226 & n & n & n & - & no detection + f02455 - 2220 & n & y & n & - & ( see sect .
2.2 ) + 03000 - 2729 & n & y & y & 1.74 & + 03538 - 6432 & y & y & n & 1.70 & + 03521 + 0028 & n & y & y & 3.23 & + 04384 - 4848 & y & y & n & 2.08 & + 17463 + 5806 & n & y & n & 1.85 & + 18030 + 0705 & y & y & y & 1.88 & large position offset + 22192 - 3211 & n & y & n & 0.85 & + 23515 - 2917 & n & y & n & 1.36 & + f23529 - 2919 & n & n & n & 0.38 & + lcccc model & type of model & number of & @xmath23 & @xmath23 + & & parameters & all ulirgs & high s / n ulirgs + 1 a & @xmath56 & 4 & 2.48 & 4.79 + 1 b & @xmath57 & 4 & 2.24 & 4.30 + 2 & @xmath58 & 5 & 2.23 & 4.31 + 3 a & @xmath59 & 5 & 1.70 & 2.81 + 3 b & @xmath60 & 5 & 1.54 & 2.52 + 3 c & @xmath61 & 5 & 1.64 & 2.79 + 3 d & @xmath62 & 5 & 1.72 & 3.05 + 4 & @xmath63 & 7 & 1.42 & 2.21 + & @xmath64 & & + 00153 + 5454 & 0.0 & 3.34 & 17028 + 5817 & 0.47 & 3.58 + f00183 - 7111 & 5.31 & 0.88 & 17068 + 4027 & 0.0 & 1.40 + 00188 - 0856 & 1.48 & 1.53 & 17179 + 5444 & 1.59 & 1.38 + 00199 - 7426 & 6.26 & 1.08 & 17208 - 0014 & 0.07 & 5.54 + 00275 - 2859 & 100 . & 0.57 & 17463 + 5806 & 2.38 & 1.85 + 00397 - 1312 & 100 . & 0.22 & 18030 + 0705 & 0.60 & 1.88 + 01003 - 2238 & 100 . & 1.44 & 18443 + 7433 & 6.09 & 1.70 + 01166 - 0844 & 0.00 & 2.16 & 18470 + 3233 & 100 . & 0.68 + 01199 - 2307 & 0.00 & 0.97 & 18531 - 4616 & 1.59 & 2.59 + 01298 - 0744 & 1.79 & 1.50 & 19254 - 7245 & 3.47 & 0.89 + 01355 - 1814 & 0.00 & 0.93 & 19420 + 4556 & 0.0 & 3.69 + 01388 - 4618 & 0.00 & 3.93 & 19458 + 0944 & 0.0 & 2.44 + 01494 - 1845 & 0.00 & 1.95 & 20049 - 7210 & 0.0 & 3.77 + 01569 - 2939 & 100 . & 1.14 & 20100 - 4156 & 0.0 & 1.92 + 02113 - 2937 & 0.64 & 2.53 & 20446 - 6218 & 0.0 & 0.94 + 02364 - 4751 & 0.66 & 3.48 & 20551 - 4250 & 0.42 & 2.33 + 02411 + 0354 & 0.18 & 3.17 & 21396 + 3623 & 0.0 & 2.31 + 03000 - 2719 & 0.88 & 1.73 & 22055 + 3024 & 1.12 & 0.96 + 03158 + 4227 & 0.53 & 1.72 & 22192 - 3211 & 100 . & 0.84 + 03521 + 0028 & 0.21 & 3.23 & 22491 - 1808 & 1.04 & 2.85 + 03538 - 6432 & 1.45 & 1.70 & 23060 + 0505 & 100 . & 0.08 + 04063 - 3236 & 100 . & 2.55 & 23128 - 5919 & 0.96 & 2.86 + 04103 - 2838 & 100 .
& 1.37 & 23129 + 2548 & 0.0 & 3.55 + 04114 - 5117 & 1.03 & 0.97 & 23230 - 6926 & 0.58 & 1.50 + 04384 - 4848 & 0.93 & 2.08 & 23253 - 5415 & 0.79 & 1.58 + 06009 - 7716 & 1.64 & 3.63 & 23327 + 2913 & 0.0 & 0.97 + 06035 - 7102 & 0.92 & 1.11 & 23365 + 3604 & 0.0 & 4.47 + 06206 - 6315 & 0.33 & 3.69 & 23389 - 6139 & 0.0 & 1.33 + 06301 - 7934 & 0.48 & 1.02 & 23515 - 2917 & 2.87 & 1.36 + 06361 - 6217 & 2.0 & 0.67 & f23529 - 2119 & 15.32 & 0.38 + 09104 + 4109 & 100 . & 0.48 & arp220 & 0.0 & 4.20 + 12112 + 0305 & 0.39 & 3.13 & mrk1014 & 100 . &
0.62 + 14348 - 1447 & 0.0 & 3.55 & mrk231 & 100 . & 0.31 + 15250 + 3609 & 0.46 & 2.88 & mrk273 & 1.24 & 1.90 + 15307 + 3252 & 100 . & 0.27 & ngc6240 & 0.0 & 2.58 + 16474 + 3430 & 0.35 & 3.58 & ugc5101 & 0.33 & 2.12 + 16487 + 5447 & 0.52 & 2.19 & & & + dum abergel , a. , et al .
1996 , , 315 , l73 alexander , d.m . , efstathiou , a. , hough , j.h . ,
aitken , d.k , lutz , d. , roche , p.f .
, & sturm , e. 1999 , , 310 , 78 aussel , h. , gerin , m. , boulanger , f. , dsert , f.x . ,
casoli , f. , cutri , r.m . , & signore , m. 1998 , , 334,l73 barger , a.j . ,
cowie , l.l . , sanders , d.b . ,
fulton , e. , taniguchi , y. , sato , y. , kawara , k. , & okuda , h. 1998 , , 394 , 248 biviano , a. , & sauvage , m. 1999 , in preparation cesarsky , c.j .
, et al . 1996 , , 315 , l32 cesarsky , d. , et al .
1996 , , 315 , l309 charmandaris , v. , mirabel , i. f. , tran , d. , laurent , o. , cesarsky , c. j. , gallais , p. , sauvage , m. , & vigroux , l. 1997 in : proceedings of the xvii rencontres de moriond , `` extragalactic astronomy in the infrared '' , eds .
mamon , trinh xuan thuan , and j. tran thanh van , p.283 clavel , j. , et al .
2000 , , 357,,839 clements , d.l . , sutherland , w.j . , saunders , w. , efstathiou , g.p . ,
mcmahon , r.g .
, maddox , s. , lawrence , & a. , rowan - robinson , m. 1996 , , 279 , 459 comastri , a. , setti , g. , zamorani , g. , & hasinger , g. 1995 , , 296 , 1 condon , j.j . , huang , z .- p .
, yin , q.f . , & thuan , t.x . 1991 , , 378 , 65 coulais , a. , abergel , a. 1999 , in : the universe seen by iso , eds .
p.cox & m.f.kessler , esa sp-427 , ( noordwijk : esa ) , p. 61
crawford , t. , marr , j. , partridge , b. , & strauss , m.a .
1996 , , 460 , 225 de grijp , m.h.k . ,
miley , g.k . ,
lub , j. , & de jong , t. 1985 , , 314,240 draine , b. , & lee , h. 1984 , , 285 , 89 dudley , c.c . , & wynn - williams , c.g .. 1997 , , 488 , 720 duley , w.w . , &
williams , d.a .
1981 , , 196 , 269 fisher , k.b . ,
huchra , j.p . ,
strauss , m.a .
, davis , m. , yahil , a. , & schlegel d. 1995 , , 100 , 69 genzel , r. , et al .
1998 , , 498 , 579 genzel , r. , & cesarsky , c.j .
2000 , , in press ( astro - ph/0002182 ) haas , m. , chini , r. , stickel , m. , lemke , d. , klaas , u. , & kreysa , e. 1998 , , 503 , l109 heckman , t.m .
1994 , in : mass - transfer induced activity in galaxies , ed .
i. shlosman , ( cambridge : cambridge university press ) , p.234 dhendecourt , l. , et al .
1996 , , 315 , l365 hines , d. c. , schmidt , g.d . ,
smith , p. s. , cutri , r. m. , & low , f. j. 1995 , , 450 , l1 hines , d. c. , schmidt , g. d. , wills , b. j. , smith , p. s. , & sowinski , l. g. 1999 , , 512 , 145 hughes , d.h . , et al .
1998 , , 394 , 241 kessler , m.f . , et al .
1996 , , 315 , l27 kim d.c .
1995 , phd thesis , university of hawaii laurent , o. , mirabel , i.f . , charmandaris , v. , gallais , p. , vigroux , l. , & cesarsky c.j .
1999 , in : the universe as seen by iso , eds .
p. cox and m.f .
kessler , esa sp-427 ( noordwijk : esa ) , p. 913
lawrence , a. , et al .
1999 , , 308 , 897 lger , a. , & puget , j .- l . 1984 , , 137 , l5 lutz , d. , et al .
1996 , 315 , l269 lutz , d. , spoon , h.w.w . ,
rigopoulou , d. , moorwood , a.f.m .
, & genzel , r. 1998 , , 505 , l103 lutz , d. , veilleux , s. , & genzel , r. 1999 , , 517 , l13 lutz , d. 1999 , in : the universe as seen by iso , eds .
p. cox and m.f .
kessler , esa sp-427 ( noordwijk : esa ) , p. 623
lutz , d. , sturm , e. , moorwood , a.f.m . ,
alexander , t. , netzer , h. , & sternberg , a. 2000 , 536 , 697 meurer , g.r . , heckman , t.m .
, lehnert , m.d . ,
leitherer , c. , & lowenthal , j. 1997 , 114 , 54 mirabel , i.f . , et al .
1998 , , 341 , 667 mirabel , i.f . , et al .
1998 , , 333 , l1 moorwood , a.f.m .
1986 , , 166 , 4 moorwood , a.f.m .
1999 , in : the universe as seen by iso , eds .
p. cox and m.f .
kessler , esa sp-427 ( noordwijk : esa ) , p. 825 papoular , r. , conrad , j. , giuliano , m. , kister , j. , & mille , g. 1989 , , 217 , 204 pier , e.a . , & krolik , j.h .
1992 , , 401 , 99 rigopoulou , d. , spoon , h.w.w . ,
genzel , r. , lutz , d. , moorwood , a.f.m , & tran , q.d 1999 , , 118 , 2625 roche , p.f . ,
aitken , d.k . ,
smith , c.h . , & ward , m.j .
1991 , , 248 , 606 roussel , h. , sauvage , m. & vigroux , l. 2000 , in preparation roelfsema , p.r . , et al .
1996 , , 315 , l289 sakata , a. , wada , s. , onaka , t. , & tokunaga , a.t .
1987 , , 320 , l63 sanders , d.b . ,
soifer , b.t . , elias , j.h . ,
madore , b.f . , matthews , k. , neugebauer , g. , & scoville , n.z .
1988 , , 325 , 74 sanders , d.b . ,
soifer , b.t . , elias , j.h .
, neugebauer , g. , & matthews , k. 1988a , , 328 , l35 sanders , d.b . , & mirabel , i.f .
1996 , , 34 , 725 solomon , p.m. , downes , d. , radford , s.j.e . & , barrett , j.w .
1997 , , 478 , 144 soifer , b.t . ,
neugebauer , g. , matthews , k. , egami , e. , becklin , e.e , ressler , m. , werner , m.w . , weinberger , a.j , evans , a.s . ,
scoville , n. , surace , j.a . , & condon , j.j .
2000 , , 119 , 509 schutte , w. 1998 in solid interstellar matter : the iso revolution ( les houches workshop , february 2 - 6 ) , edp sciences .
starck , j.l . , abergel , a. & aussel , h. 1999 , , 134 , 135 sturm , e. , et al .
1996 , , 315 , l144 sturm , e. , lutz , d. , tran , d. , feuchtgruber , h. , genzel , r. , kunze , d. , moorwood , a.f.m . , & thornley , m. 2000 , , 358 , 481 taniguchi , y. , sato , y. , kawara , k. , murayama , t. , & mouri , h. 1997 , , 318 , l1 taniguchi , y.,yoshino , a. , ohyama , y. , & nishiura , s. 1999 , , 514 , 660 tran , d. 1998 , phd thesis , universit de paris xi verstraete , l. , puget , j .- l . ,
falgarone , e. , drapatz , s. , wright , c.m . , & timmermann , r. 1996 , , 315,l337 veilleux , s.,sanders , d. b. & kim , d.c . , 1999 , , 522,139 vigroux l. , et al .
1996 , , 315,l93 vigroux l. , et al . 1999 , in : the universe as seen by iso , eds .
p. cox and m.f .
kessler , esa sp-427 ( noordwijk : esa ) , p. 805 | we present low resolution mid infrared spectra of 16 ultraluminous infrared galaxies ( ulirgs ) obtained with the cvf spectroscopy mode of isocam on board the infrared space observatory iso .
our sample completes previous iso spectroscopy of ultra- and hyperluminous infrared galaxies towards higher luminosities .
the combined samples cover an infrared luminosity range of @xmath0 .
to discriminate agn and starburst activity , we use the agn - related mid - infrared continuum and the starburst - related 6.2 , 7.7 , 8.6 , 11.3@xmath1 m mid infrared emission bands attributed to aromatic carbonaceous material .
for about half of the high luminosity ulirgs studied here , strong aromatic emission bands suggest starburst dominance .
other spectra are dominated by a strong agn - related continuum with weak superposed emission features of uncertain nature .
our sample contains one unusual example , iras f00183 - 7111 , of an agn that is highly obscured even in the mid - infrared .
an improved method to quantitatively characterize the relative contribution of star formation and agn activity to the mid - infrared emission of ulirgs is presented .
the ulirg spectra are fitted by a superposition of a starburst and an agn spectrum , both of which may be obscured at different levels .
models in which starburst and agn obscuration differ are significantly more successful than models with a single extinction .
previous results based on a simpler line - to - continuum measure of aromatic emission strength are confirmed , further supporting the robustness of the aromatic emission feature as a diagnostic of ulirg power sources . as dominant source of the bolometric luminosity ,
starbursts prevail at the lower end and agns at the higher end of this range .
the transition between mostly starburst and mostly agn powered occurs at @xmath2 to @xmath3 , and individual luminous starbursts are found up to @xmath4 . |
the moduli space @xmath0 of stable curves of genus @xmath1 has been an important subject of study in algebraic geometry .
it is a smooth deligne - mumford stack ( orbifold ) .
the moduli space @xmath2 of stable maps of degree @xmath3 from genus-@xmath1 curves to the projective space @xmath4 is a natural generalization of @xmath0 , but this generalization leads us from a smooth space to a space that can contain singularities as bad as possible ( vakil s murphy s law ) . for the purposes of some programs ,
it is important to obtain the ( local ) structures of the moduli space @xmath2 . a point @xmath5 $ ] of @xmath2 is ( the isomorphism class of ) a map @xmath6 where @xmath7 is an algebraic curve of arithmetic genus @xmath1 with at worse nodal singularities such that the automorphism group of the map @xmath5 $ ] is finite . here
a nodal singularity locally is given by @xmath8 ; the point corresponding to the origin is called a node . an automorphism of @xmath5 $ ] means a morphism @xmath9 such that @xmath10 the automorphism group of the map @xmath5 $ ] is finite if and only if any genus-0 irreducible component of @xmath7 contains at least three nodes whenever it is contracted by the map and any genus-1 irreducible component of @xmath7 contains at least one node whenever it is contracted by the map .
one may add @xmath11 marked points to the domain curve @xmath7 away from the nodes and obtain the moduli space of stable maps with @xmath11 markings , denoted @xmath12 .
but , as far as singularity is concerned , @xmath12 and @xmath2 have exactly the same local singularity types simply because each marked point moves in a smooth local domain .
hence , as far as singularity types are concerned , we may only consider moduli spaces without markings .
when @xmath13 , @xmath14 is smooth ( as a stack or orbifold ) ; when @xmath15 , @xmath2 is singular ( as a stack or orbifold ) . indeed ,
if allowing arbitrary @xmath1 and @xmath3 , ravi vakil showed that @xmath2 can contain all possible singularity types over @xmath16 .
this seems to be a piece of bad news , but thinking positively , it also makes the spaces @xmath2 ultimately rich as modular singularity models to investigate . historically though
, @xmath2 was introduced for the ( sole ) purpose of defining the gromov - witten invariants .
an important standing problem in this area is to enumerate the ( virtual ) number of curves with fixed genus and degree in a smooth calabi - yau threefold in @xmath17 , or more generally , in a complete intersection @xmath18 in @xmath4 . in principle
, the curve - counting business on @xmath18 can be done on the ambient space @xmath4 , using the defining equations of @xmath18 .
the moduli space @xmath19 of stable maps of degree @xmath3 from curves of genus @xmath1 to @xmath18 is naturally a submoduli space of @xmath20 .
for example , let us assume that @xmath18 is a smooth hypersurface defined by @xmath21 for some section @xmath22 with some positive integer @xmath23 .
if we let @xmath24 be the universal family over the moduli space @xmath2 with the universal map @xmath25 and let @xmath26 then we have @xmath27 if @xmath28 were locally free , then it would have a natural euler class , and this euler class would bridge the intersection theory on @xmath19 to the intersection theory on @xmath2 . thus , resolving the non - locally free locus of the direct image sheaf @xmath28 is essential in gw theory of complete intersections in @xmath4 . here , by resolving the sheaf , we mean a diagram @xmath29 where @xmath30 is a blowup and @xmath31 such that the direct image sheaf @xmath32 is locally free .
thus , one sees that on the one hand , it is important for gw theory to resolve the sheaf @xmath28 ; on the other hand , it is important for singularity theory to resolve @xmath2 . the two problems are related .
the point of @xcite is that it is more natural to resolve the sheaves @xmath28 first , and at least when the genus is low , resolving the sheaves will also ensure a resolution of the moduli space . to see this point , observe that a stable map @xmath33\in \mpdg$ ] , as a morphism , is given by the data @xmath34 : c\lra\pn,\quad u_i\in h^0(u\sta \so_{\pn}(1));\ ] ] its deformation is determined by the combined deformation of the curve @xmath7 and the sections @xmath35 .
since the deformation of the curve is unobstructed , the irregularity of @xmath2 is closely related to the non - local - freeness of the direct image sheaf @xmath36 this alludes that desingularizations of @xmath2 _ should be _ governed by desingularizations of @xmath36 . for genus-1 ,
this is true in the simplest form , a desingularization of @xmath37 implies a desingularization of @xmath38 .
vakil and zinger first discovered a desingularization of the main component of the moduli space @xmath38 in @xcite .
their method is analytic in nature .
they found a natural sequence of blowups that resolve singularities of @xmath38 and then showed that the same blowups also resolve the sheaves @xmath39 .
as hinted in the last paragraph , from the algebro - geometric approach , it is more natural to resolve the sheaves first . in @xcite
, we first obtain local structures of the sheaf @xmath40 .
the structures of @xmath36 allow us to derive local defining equations of the moduli space @xmath38 .
having obtained these local equations , it is rather clear what loci one should blow up , how the resulting space turns out to be smooth , and why the resulting direct image sheaves become locally free .
this note , through some concrete examples , is solely devoted to reveal the structures of the sheaf @xmath41 . by working out these examples ,
we hope the student will familiarize himself / herself with the aspects of families of elliptic curves that are useful for obtaining the local structures of the moduli space of stable maps . during the summer school , i gave four lectures on tropical curves and their applications to plane enumerative geometry . when the school was over , the organizers asked every speaker to write up his lecture notes for the proceeding .
however , as there have been already several excellent expository articles on tropical curves that the students can easily find on arxiv and i did not feel that i could make any meaningful improvement , instead , i thought that an introductory note on elliptic stable maps should help students to learn this important subject .
the idea is that through some examples the students will gain the intuition about the approach to the local structures of the stable map muduli @xcite .
i hope that the material will be a more converging addition to the proceeding , and it will be more useful for graduate students as well as for researchers . most of the note should be accessible to any graduate student with some backgrounds on algebraic geometry .
_ this note recollects the toy examples that jun li and i calculated in the summer of 07 as the warm - up as well as the guide for our approach toward the general theory @xcite , but , needless to say , all the mistakes in this detailed presentation must be due to my own oversight . _
i thank jun li , from whom i have learned a lot , for the collaboration .
i also thank cms of zhejiang university and the organizers of the summer school , especially lizhen ji , for the excellent environment and for supporting the idea to include this note in the proceeding .
we begin with motivating the setups of our examples .
let @xmath42 be the universal family with the universal map @xmath43 .
as for any sheaf , the question on the structures of the direct image sheaf @xmath44 is ( tale ) local . given any point @xmath45=[u : c \lra \pn ] \in \mpd,\ ] ] we choose a small tale neighborhood @xmath46 \subset \mpd$ ] . by choosing @xmath23 general hyperplanes @xmath47 of @xmath4 , we can assume that @xmath48 is a simple divisor @xmath49 of degree @xmath50 ( @xmath51 are disjoint ) .
let @xmath52 , then @xmath53 by an tale base change , we may assume that @xmath54 where each @xmath55 is a section of @xmath56 such that @xmath57 . hence , the local structures of @xmath58 is reflected in the structure of @xmath59 .
this motivates us to consider some examples of flat families of elliptic curves , @xmath60 with a section @xmath61 , and study the associated direct image sheaf @xmath62 .
the core @xmath63 of a connected genus - one curve @xmath7 is the unique smallest ( by inclusion ) subcurve of arithmetic genus one .
( 3 , 5 ) ( 90,5 ) .5 cm figure 1 .
an elliptic curve @xmath63 attached with 3 rational tails given any connected genus - one curve @xmath7 , upon removing the core of @xmath7 , the rest of irreducible components are all rational curves ( i.e. , genus zero curves ) , we denote their union by @xmath64
. we will call each connected component of @xmath64 a tail , it is a tree of rational curves .
if there are @xmath65 such connected components , we will say the curve @xmath7 has @xmath65 ( rational ) tails .
see figure 1 for an example of elliptic curve with three rational tails . by riemann - roch
, we can check that @xmath66 at point @xmath67 where the section @xmath61 meets the core of @xmath68 but @xmath69 does not vanish otherwise . in particular , @xmath62 is locally free at point @xmath67 where the section meets the core of @xmath68 , but it is not locally free otherwise .
so we will study examples of families @xmath60 together with a section @xmath61 having a point @xmath70 such that * for @xmath71 , the fiber @xmath68 is smooth ; * the fiber @xmath72 is an elliptic curve with @xmath65 tails ; * @xmath61 misses the core of @xmath72 but meets the tails .
( 3 , 5 ) ( 90,5 ) .5 cm figure 2 .
the family @xmath73 we will begin with three toy examples in order of generality : a warm - up 1-tail case , an @xmath74-tail case ( @xmath75 ) , and a case of more general type . throughout the note , unless otherwise stated , we can work over any fixed algebraically closed base field @xmath76 .
we now begin to construct a one - parameter family of genus-1 curves whose central fiber is a smooth elliptic curve attached with a smooth rational curve .
one may consult figure 2 for picture of such a construction .
the details are in the next two paragraphs .
let @xmath77 be the blow up of @xmath78 at @xmath79 where @xmath80 a fixed point on @xmath81 and @xmath82 is a fixed point on @xmath83 , respectively .
let @xmath84 be the projection to the first factor and @xmath85 note that this provides a one - parameter smoothing of an elliptic curve with one rational tail .
we will denote the fiber @xmath86 by @xmath87 , @xmath88 .
choose a generic point @xmath89 .
let @xmath90 be the proper transform of @xmath91 and @xmath92 , respectively .
note that if we write the central fiber of @xmath93 as @xmath94 where @xmath95 is elliptic and @xmath96 is rational , then @xmath97 .
we will consider the direct image sheaf @xmath98 .
we have a short exact sequence @xmath99 and a long exact sequence @xmath100 here observe that @xmath101 is locally free because one calculates that @xmath102 and @xmath103 since @xmath104 and @xmath105 is an isomorphism , we see that @xmath106 our goal is to describe explicitly the entire sequence ( [ longsd ] ) .
the case @xmath107 is somewhat special , we isolate it below .
@xmath108 it is easy to see that this is @xmath109 toward the general case of ( [ longsd ] ) , we first consider another short exact sequence : @xmath110 noting that @xmath111 and @xmath112 we obtain a short exact sequence of locally free sheaves @xmath113 we claim [ splitdd ] the sequence ( [ dd ] ) splits and consequently @xmath114 we prove it by induction . for @xmath107 ,
it is clear that @xmath115 assume that the lemma is true for the case of @xmath11 .
then by ( [ dd ] ) , we have @xmath116 using serre duality , one calculates that @xmath117 hence the exact sequence above must be trivial .
let @xmath118 $ ] be the homogeneous coordinates of @xmath81 .
we may assume that @xmath119 $ ] .
then we have a canonical map @xmath120 , @xmath121 which induces a canonical exact sequence @xmath122 where @xmath123 is the one - dimensional skyscraper sheaf supported at @xmath80 and @xmath124 is the evaluation at @xmath80 .
we make an observation here that any map @xmath125 such that @xmath126 is a scalar multiple of @xmath120 .
this is because @xmath127 and @xmath128 has a basis @xmath129 .
in particular , any such nontrivial @xmath130 determine the same cokernel , namely , @xmath123 .
[ prop1tail ] assume @xmath131 .
up to isomorphism , we have 1 .
@xmath132 2 .
the map @xmath133 is the natural inclusion to the corresponding factors .
the map @xmath134 is given by @xmath135 .
the map @xmath136 is the evaluation at @xmath80 . in particular , all @xmath137 are isomorphic to the one - dimensional skyscraper sheaf @xmath138 supported at @xmath80 . to start , observe that for any nonnegative integer @xmath23 , a map from @xmath139 to @xmath140 is equivalent to a map from @xmath141 to @xmath142 , thus the space of all such maps is @xmath143 .
now consider the case @xmath144 first .
we have @xmath145 by the observation in the start , we can express @xmath146 as @xmath147 where @xmath148 and @xmath149 ( @xmath150 ) are two fixed sections .
since the cokernel of @xmath146 supports at @xmath80 , which follows from @xmath151 by the base change property , hence @xmath152 and @xmath153 is a non - zero constant multiple of @xmath154 .
in particular this also implies that @xmath155 is the one - dimensional skyscraper sheaf @xmath138 supported at @xmath80 , and @xmath156 is the evaluation at the point @xmath80 .
this proves the case @xmath144 . for @xmath157 ,
consider the following natural commutative diagram of long exact sequence , @xmath158 @xmath159 note that the third downward arrow is an isomorphism .
the diagram gives rise to the following one which we will use soon @xmath160 @xmath161 we take splits of all @xmath162 @xmath163 so that all the inclusions @xmath164 are given by the natural factor inclusions . by the second square of the last diagram
, we see that the @xmath134 restricted to the factor @xmath165 is nontrivial .
so , up to a non - zero scalar , we can assume that it is given by multiplication by @xmath154 .
as in the case of @xmath144 , we can express @xmath134 as @xmath166 for some fixed @xmath167 , where @xmath168 . using the base change property
, we see that @xmath169 is supported at @xmath80 , thus @xmath152 and @xmath170 for @xmath171 .
so , we can write @xmath172 with @xmath173 .
hence we obtain that @xmath174 is given by @xmath175 this clearly is isomorphic to @xmath176 .
therefore , by expressing an arbitrary element @xmath177 as @xmath178 @xmath179 we conclude that after a ( possibly ) new split of @xmath180 we can identify @xmath174 with the direct summand @xmath181 and the map @xmath134 is given by the restriction to the summand @xmath182 which , in turn , is given by multiplying by @xmath154 .
all the rest of the statements follow immediately .
@xmath183 ( 3 , 5 ) ( 90,5 ) figure 3 .
the central fiber is a smooth elliptic with @xmath23 tails we can generalize the above to the case where the central fiber is an elliptic curve with @xmath23 many tails . to do this , again on @xmath184 ,
we pick up @xmath23 points @xmath185 , and blow up @xmath184 at @xmath186 .
this way , we obtain a smooth surface @xmath187 whose projection to @xmath81 provides a family @xmath188 with the central fiber an elliptic curve with @xmath23 rational tails , elsewhere the fiber is isomorphic to @xmath83 . in the moduli space , this represents a general direction along which all the nodes are smoothed simultaneously
. see figure 3 .
let @xmath189 be the proper transform of @xmath190 , @xmath191 and @xmath192 be the proper transform of @xmath193 where @xmath194 is a general point on @xmath83 .
let @xmath195 and @xmath196 .
let @xmath197 and @xmath198 .
we consider the pushforward sheaf @xmath199 .
then all the previous results extend almost word by word to this case with @xmath11 , @xmath61 replaced by @xmath200 , @xmath201 .
this suggests that the direct image sheaf is sensitive only to the smoothing direction , but not to the number of tails .
the point is that after carefully treating the case of one tail , it is almost routine to treat the case of multiple tails
. one can consult [ rtails ] below to see how to formulate similar statements and arguments .
for example , in the inductive proof in this case , to go from @xmath11 to @xmath202 , one goes from @xmath203 to @xmath204 .
since the results of this subsection will not be used elsewhere in this work , we omit the routine details . we will now construct a family @xmath205 such that the central fiber is a smooth elliptic curve attached with a connected chain of rational curves .
( 3 , 5 ) ( 90,5 ) figure 4 .
the central fiber has two rational components consider the space @xmath206 .
let @xmath207 \in { \mathbb{p}}^1 $ ] .
pick @xmath65 distinct points @xmath208 in @xmath83 . for any @xmath209 , set @xmath210 where @xmath80 occurs in the i - th factor .
set @xmath211 , @xmath209 .
let @xmath212 be the blowup of @xmath206 along the disjoint union @xmath213 .
let @xmath189 be the proper transform of @xmath214 and @xmath192 be the proper transform of @xmath215 where @xmath89 is a point distinct from all @xmath208 .
let @xmath93 be the projection @xmath216 followed by the projection to @xmath217 .
then this provides a family of nodal elliptic curves : the fiber over @xmath218 is a smooth elliptic curve with @xmath65 many rational tails ; moving to a general direction represented by coordinates indexed by @xmath219 ( @xmath220 ) , the nodes created by blowing up @xmath221 will be smoothed .
figure 4 shows the case of @xmath222 . here , the central fiber over @xmath223 is @xmath224 where @xmath225 and @xmath226 ; the section @xmath192 passes @xmath63 ; the sections @xmath189 passes @xmath227 ( @xmath228 ) .
let @xmath229 be the exceptional divisor of @xmath187 corresponding to @xmath230 .
observe that @xmath231 .
let @xmath232 and @xmath233 .
let @xmath197 and @xmath198 .
we will study the local freeness of the direct image sheaf @xmath234 .
we have a short exact sequence @xmath235 and a long exact sequence @xmath236 as in [ 1-tail ] , @xmath237 is locally free and @xmath238 since @xmath239 and @xmath240 is an isomorphism , we see that @xmath241 the case of ( [ longsdr ] ) when @xmath107 is special , we isolate it below . @xmath108
it is easy to see that this is @xmath242 to treat the general case , similar to [ 1-tail ] of the 1-tail case , we will first prove a formula for @xmath243 . we begin with a lemma .
[ owj ] [ ow = o1 ] @xmath244 is isomorphic to @xmath245 . and both are isomorphic to @xmath246 where @xmath247 is the projection to the j - th factor .
any bundle on @xmath248 is of the form @xmath249 .
we will determine the integer @xmath250 for each of @xmath244 and @xmath245 .
first , it is easy to determine that @xmath251 by looking at the intersection numbers of of @xmath252 with the coordinate lines
. hence @xmath253 , as desired .
to show the rest , again we can argue by computing intersection numbers .
let @xmath254 be the pre - image of the i - th coordinate line by the isomorphism @xmath255 .
then we have @xmath256 when @xmath257 because @xmath254 can be moved to become a section of @xmath258 , disjoint from @xmath259 .
when @xmath260 , we have @xmath261 hence @xmath262 . but @xmath263 .
hence @xmath264 because @xmath265 .
we may also argue without calculating the intersection numbers . since @xmath259 is the proper transform of the constant section @xmath266 in @xmath267
, we have that the normal bundle @xmath268 is isomorphic to @xmath269 where the exceptional divisor in @xmath259 is the pre - image of @xmath252 in @xmath259 .
this implies that @xmath270 .
now , consider the short exact sequence : @xmath271 noting that we have @xmath272 @xmath273 and @xmath112 hence we obtain a short exact sequence of locally free sheaves @xmath274 [ lemmsplit ] the sequence ( [ ddr ] ) splits and consequently @xmath275 we proceed by induction on @xmath276 .
when @xmath107 , we have @xmath277 hence the lemma holds in this case . assume that when @xmath276 , the lemma holds .
then for the case of @xmath202 , by ( [ ddr ] ) , we have @xmath278 now applying serre s duality and lemma , we obtain @xmath279 hence the sequence is the trivial one .
our aim is to explicitly describe the sheaf @xmath280 . by ( [ bfm=0 ] ) , we may assume @xmath281 .
let @xmath282 $ ] be the homogeneous coordinates of the i - th factor @xmath81 in the product @xmath217 .
then @xmath283 note that @xmath284 .
let @xmath285 then we have a canonical map @xmath120 , @xmath286 which induces a canonical exact sequence @xmath287 where @xmath288 is the structure sheaf of @xmath289 .
this sequence is a higher dimensional generalization of ( [ beta ] ) and plays similar role in the proof .
set @xmath290 [ proprtail ] assume , we can ignore @xmath252 , @xmath259 , and the family @xmath229 from the consideration , this tail will not affect the sheaf @xmath291 , and the problem is reduced to the previous @xmath292-tail case .
this is analogous to the situation of [ 1-tailr ] . ]
that @xmath293 for all @xmath220 . then up to isomorphism
, we have a commutative diagram @xmath294 @xmath295 where the first two downward arrows are inclusions , and the last two are isomorphisms . moreover ,
we have 1 . @xmath296 .
2 . @xmath297 .
3 . @xmath298 .
the proof is totally analogous to the proof given in proposition [ prop1tail ] with main exception that @xmath299 needs not to be trivial in this higher dimensional case .
below , we will provide enough details .
\(1 ) is just a rephrase of lemma [ lemmsplit ]
. we will prove the rest all together .
the initial case is when all @xmath300 , @xmath220 .
denote @xmath301 by @xmath302 .
we have @xmath303 hence @xmath304
so , we can express @xmath305 as @xmath306 where @xmath307 , @xmath308 is a fixed scalar , and @xmath309 are fixed sections .
the same argument as in the case of @xmath310 implies that @xmath152 and @xmath311 . by changing coordinates if necessary , we can write @xmath312
. then @xmath313 if and only if @xmath314 .
that is , @xmath315 .
hence the case @xmath316 of the lemma follows .
when @xmath317 , meaning all @xmath318 , we have @xmath319 @xmath320 hence we can assume that @xmath321 .
then the map @xmath322 is given by @xmath323 as in proposition [ prop1tail ] , @xmath324 and we may write @xmath325 ( @xmath326 ) . hence @xmath327 if and only if @xmath328 we can split an arbitrary point @xmath329 as the sum of @xmath330 and @xmath331 this gives a splitting of @xmath237 as the direct sum of @xmath332 and @xmath333 . under this direct sum
, we see that @xmath334 .
the rest follows straightforwardly .
clearly , @xmath299 , independent of @xmath335 , is the sole cause for non - local freeness of the sheaf @xmath291 .
blowup is to make @xmath299 locally free , thus resolves the sheaf @xmath280 .
1 . note here that the essence of the arguments on the sheaf is independent of the genus of the family . 2 .
also , the decomposition @xmath336 seems to suggest a relation with either restricting the family to the coordinate directions or projections to the coordinate directions .
let @xmath337 be the blowup at the point @xmath338 .
we have the square @xmath339 the sheaf @xmath340 is locally free .
this can be checked directly .
( see @xcite for a general proof .
the reader is encouraged to do this example . ) to gain further intuition around deeper strata of the moduli space @xmath38 , let us look at the stratum of the type @xmath341 $ ] .
@xmath341 $ ] is the square bracket notation for the first tree of figure 5 .
( the interested reader is referred to @xcite for further explanation of this notational scheme ; he is also encouraged to construct a family for the second curve in figure 5 . ) here , any curve in the stratum consists of a smooth elliptic curve @xmath95 , a rational curve @xmath96 attached to @xmath95 , and two more rational curves @xmath342 and @xmath343 attached to @xmath96 . a line bundle over the family of
the interest is trivial on both @xmath95 and @xmath96 , and positive on @xmath342 and @xmath343 .
we will construct explicitly such a family @xmath187 over a three dimensional base @xmath344 .
we work over @xmath345 in this subsection .
let @xmath346 be the family of nodal elliptic curve constructed in
[ 1-tail ] .
that is , @xmath346 is the blowup of @xmath347 at @xmath79 .
we would like to have two disjoint sections that go through the rational curve of the central fiber .
so , we take a small analytic disc @xmath348 of the base center at @xmath80 .
we denote the family restricted to @xmath348 by @xmath349 .
so , let @xmath350 be two disjoint sections of @xmath349 that go through the rational curve of the central fiber , @xmath351 .
let @xmath352 be the proper transform of @xmath353 in @xmath346 , @xmath354 .
consider @xmath355 .
let @xmath356 and @xmath357 .
let @xmath187 be the blowup of @xmath355 along @xmath358 .
then we obtain a family of nodal elliptic curve over @xmath359 @xmath360 such that : moving along the first coordinate direction , the node @xmath361 is smoothed , but the elliptic curve comes with two solid tails ; moving along the second direction , the node @xmath362 is smoothed ; moving along the last direction , the node @xmath363 is smoothed .
( 3 , 5 ) ( 30,5 ) .5 cm figure 5 . two elliptic curves and two trees .5 cm
let @xmath192 be the proper transform of @xmath364 in @xmath187 , @xmath365 the proper transform of @xmath366 , and @xmath367 the proper transform of @xmath368 .
let @xmath369 , @xmath370 , @xmath371 .
let @xmath372 be the projection of @xmath373 to the i - th factor , @xmath374 .
a main new feature in this example is @xmath375 @xmath376 both can be checked directly by calculating the degree on each coordinate direction .
let @xmath377 . then , almost identical arguments as in the sections
[ 1-tail ] and [ rtails ] will lead to a canonical homomorphism @xmath378 hence we have , up to isomorphism , we have a commutative diagram @xmath379 @xmath380 where the first two downward arrows are inclusions , and the last two are isomorphisms . moreover , we have 1 .
2 . @xmath382 .
3 . @xmath383 .
parallel to to the proofs in [ 1-tail ] and [ rtails ] . in the next two subsections [ generalcase - r ] and [ generalstrata ] , we will give interpretations of the main results of @xcite on the direct image sheaves in the spirit of the examples that the reader has already been through .
we refer the proofs to @xcite .
let @xmath384 be a flat family of nodal genus 1 curves .
let @xmath385 be a fixed point with the fiber @xmath386 such that @xmath387 is a nodal elliptic curve , @xmath227 are smooth rational curves attached to @xmath387 at distinct smooth points .
we will first make sense of the subscheme @xmath388 that are locus of the @xmath389-th tail not smoothed ( @xmath209 ) .
we let @xmath390 be the stabilization contracts any rational components that has fewer than three nodes . ] , let @xmath391 be the exceptional divisors , let @xmath392 be the irreducible component of @xmath393 that contains @xmath394 .
we define @xmath395 the subset @xmath388 is canonically a closed subscheme whose ideal sheaf @xmath396 is generated by a regular function @xmath397 .
so we can write @xmath398 .
for every @xmath209 , let @xmath399 be the restriction of the total family to @xmath400 .
choose a section @xmath189 over @xmath401 that meets @xmath392 ( @xmath209 ) , and a generic section @xmath192 that misses all @xmath392 .
all these can be done by shrinking @xmath401 if necessary .
again we will consider sheaves of the form @xmath402 we have a short exact sequence @xmath403 and a long exact sequence @xmath404 since @xmath401 is affine , we have an isomorphism @xmath405 .
we fix such an isomorphism . as in [ rtails ] , consider the short exact sequence : @xmath406 because @xmath407 , we have @xmath408 , hence we obtain a short exact sequence of locally free sheaves @xmath409 because @xmath410
. then by arguing that @xmath411 , we obtain that the sequence splits ; then by induction , we get [ lemmsplit - rtails ] the sequence ( [ ddr - tails ] ) splits and consequently @xmath412 let @xmath413 then we have a canonical map @xmath120 , @xmath414 which induces a canonical exact sequence @xmath415 where @xmath416 .
set @xmath417 we have [ propmoduli - rtails ] assume that @xmath293 for all @xmath220 .
then up to isomorphism , we have a commutative diagram @xmath418 @xmath419 where the first two downward arrows are inclusions , and the last two are isomorphisms . moreover , we have we let @xmath423 be a flat family of nodal curves of arithmetic genus one .
we let @xmath424 be a fixed point with the fiber @xmath386 such that @xmath387 is a nodal elliptic curve , @xmath227 are connected trees of rational curves attached to @xmath387 at distinct smooth points . for each @xmath209 ,
let @xmath425 be the irreducible component of @xmath227 that meets @xmath387 .
note that as a nodal elliptic curve , @xmath387 contains either a smooth elliptic curve or a unique loop of rational curves .
this is the core @xmath63 of @xmath426 .
we let @xmath427 be an invertible sheaf on @xmath428 whose restriction to @xmath387 is trivial and whose restriction to @xmath425 are effective and with positive degree ( @xmath209 ) .
thus , @xmath387 is the largest subcurve of @xmath426 containing the core @xmath63 such that @xmath429 is trivial . as in [ generalcase - r ] , by shrinking @xmath401 if necessary , we can choose sections @xmath189 , @xmath209 , and @xmath192 such that @xmath427 is of the form @xmath430 for some @xmath431 . for each @xmath227 , @xmath209 , let @xmath432 be the list of all irreducible ( ghost ) rational curves through which @xmath227 is attached to @xmath63 . observe that some ( ghost ) rational curves @xmath433 are allowed to repeat as @xmath434 for some @xmath435 .
we have local variables @xmath436 , @xmath437 , @xmath209 , @xmath438 , each of which is the smoothing parameter of the corresponding node .
let @xmath439 and @xmath440 , @xmath441 , @xmath442 .
let @xmath443 the reason that @xmath444 occurs is because of the sequence @xmath445 and @xmath446 .
then we have a canonical map @xmath120 , @xmath447 which induces a canonical exact sequence @xmath448 where @xmath449 . set @xmath450 [ theomoduli ] assume that @xmath293 for all @xmath220 .
then up to isomorphism , we have a commutative diagram @xmath451 @xmath452 where the first two downward arrows are inclusions , and the last two are isomorphisms . moreover , we have 1 . @xmath453 .
2 . @xmath454 .
3 . @xmath298 .
let @xmath455 .
then this reduces to the case of @xmath65-tails ( proposition [ propmoduli - rtails ] ) .
see also @xcite for the general proof .
consider the flat family of nodal elliptic curves @xmath456 . in this section ,
we look at sections in @xmath457 over the central fiber and study their extensions by deformation theory .
this is useful because of the following .
set theoretically , our deformation space is the union @xmath458 the deformation space is singular at 0 if the core curve of @xmath72 is ghost due to the nonvanishing of the first cohomology .
naively , this is caused by the existence of some sections in @xmath459 that do not extend to neighboring fibers .
thus , it is instrumental to determine , at least for some examples , exactly what sections can be extended and what sections can not .
( the singularities occurring in the genus one case is relatively simple comparing to the more general case . for more general deformation of sections of invertible sheaves ,
one may consult the book @xmath460deformations of algebraic schemes@xmath461 by edoardo sernesi . )
we write the central fiber @xmath72 of @xmath93 as @xmath470 where @xmath7 is elliptic and @xmath96 is rational .
choose homogeneous coordinates @xmath471 $ ] of @xmath96 such that @xmath472 $ ] and @xmath473 $ ] .
then the space of sections of @xmath474 restricted to @xmath96 has a basis @xmath475 since @xmath474 restricted to @xmath7 is trivial , sections on @xmath7 are constant , hence these constants are uniquely determined by the values of sections on @xmath96 at the point @xmath472 $ ] .
this shows that @xmath474 restricted to @xmath72 has @xmath202 dimensional global sections spanned by @xmath476 we first observe that @xmath477 extends to @xmath187 .
indeed , the tautological inclusion @xmath478 induces a section @xmath479 that vanishes along @xmath61 of order @xmath11 .
thus this section restricts @xmath72 must be the @xmath477 mentioned .
when @xmath144 , we claim that @xmath480 is the section that does not extend to the nearby fiber . for this , note that @xmath480 does not vanish on @xmath481 $ ] .
consider @xmath482 where @xmath483 , then a section at a nearby fiber must vanish at the point @xmath484 . as @xmath154 specializes to 0 ,
the limit section on the central fiber must vanish at the point @xmath485 $ ] .
combined , since @xmath486 extends , @xmath487 extends if and only if @xmath488 .
we begin with studying the sheaf @xmath493 and its global sections @xmath494 . for this , we need to gain knowledge on @xmath495 , which will be determined by the long exact sequence of cohomology of the exact sequence @xmath496 alternatively , this is just @xmath497 where @xmath498 is the ideal sheaf of @xmath7 .
in particular , @xmath499 . by adjunction formula ,
@xmath500 where @xmath501 . as a consequence @xmath502 and @xmath503 .
since @xmath508 consists of constant sections , by lifting constant sections over @xmath7 to constant sections over @xmath509 , we obtain a canonical split @xmath510 thus , an arbitrary element of @xmath511 can be ( canonically ) expressed as @xmath512 for scalars @xmath361 and @xmath362 .
then consider the exact sequence @xmath513 first , we have @xmath514/(u_1 ^ 2).\ ] ] we already know that @xmath515 let @xmath516 be as indicated in the sequence , then sections of @xmath517 are pairs @xmath518 such that @xmath519 we can write @xmath520 and @xmath521 , by solving @xmath522 we see that @xmath523 and @xmath524 . hence we conclude @xmath525 in the previous subsection , we have already determined the space @xmath527 . to find sections of @xmath528
, we shall use induction based on the exact sequence @xmath529 taking global sections , we obtain exact sequence @xmath530 @xmath531 because @xmath532 and @xmath533 , @xmath534 to prove the claim , we use an exact sequence similar to @xmath536 then taking cohomology , we have @xmath537 @xmath538 is easily seen to be surjective . also , we have @xmath539 @xmath540 thus the claim is proved .
first observe that @xmath548 has the space of sections generated by @xmath549 next , using the sequence @xmath550 and the fact that @xmath551 where @xmath552 , @xmath553 , we see that each @xmath554 , @xmath555 , extends by zero to @xmath556 .
thus @xmath557 to proceed further , we need to determine @xmath558 for this , we look at @xmath559 note that since @xmath560 is annihilated by @xmath561 , it is naturally a sheaf of @xmath562-modules . and as such , @xmath563 ( recall that @xmath501 ) . taking the long exact sequence , we then get @xmath564 because @xmath565 , we conclude that the last term is zero .
hence we obtain @xmath566 @xmath567 where the last isomorphism holds by riemann - roch .
we claim that the evaluation homomorphism @xmath568 has image @xmath569/(u_1 ^ 2)=\so_{2c\cap c_a}$ ] . indeed , by evaluating sections of @xmath570
, we obtain an exact sequence @xmath571 because as @xmath562-modules , @xmath572 thus its first cohomology group vanishes .
this proves the claim .
now consider the sequence @xmath573 by the above , we may choose a generator @xmath574 of @xmath575 so that its image under @xmath576 is @xmath480 .
then it is easy to see that @xmath577 extends to @xmath578 if and only if @xmath579 .
so , we find that @xmath580 this is the same as @xmath581
. we let @xmath589 be the exceptional divisor ; we let @xmath590 be the analytic neighborhood @xmath591 of @xmath589 .
we let @xmath471 $ ] be the homogeneous coordinate of @xmath83 so that over @xmath592 : @xmath593 we let @xmath594 , @xmath595 and @xmath596 .
then we have since @xmath603 , @xmath604 has an obvious trivialization . also , we assume that @xmath605 , hence @xmath606 also an obvious trivialization .
we assume that @xmath61 intersects @xmath83 at @xmath607,0 ) \in v_0 $ ] .
we use the following local trivializations of @xmath474 : @xmath608 @xmath609 @xmath610 from @xmath597 to @xmath600 , the transition function is the same as the transition function for @xmath611 , hence we have @xmath612 since we use 1 to trivialize both @xmath600 and @xmath598 , the transition from @xmath600 and @xmath598 is just identity @xmath613 the two transitions patch all the local trivializations . 10 yi hu and jun li , _ genus one stable maps , local equations and vakil - zinger s desingularization . _
( 2010 ) , no .
4 , 929 - 963 .
r. vakil and a. zinger , _ a desingularization of the main component of the moduli space of genus - one stable maps into @xmath4 .
12 ( 2008 ) , no . | this informal note provides some elementary examples to motivate the local structural results of @xcite on the moduli space of genus one stable maps to projective space .
the hope is that these examples will be helpful for graduate students to learn this important subject . |
it is shown in @xcite that any tensor ( i.e. , multi - dimensional array ) can be decomposed into the product of orthogonal matrices and an _ all - orthogonal _ core tensor .
this decomposition generalizes the matrix svd and is today commonly called higher - order singular value decomposition ( hosvd ) or multilinear svd . in applications , people are usually interested in seeking a low - multilinear - rank approximation of a given tensor , such as the multilinear subspace learning @xcite and multilinear principal component analysis @xcite . unlike the matrix svd ,
_ truncated _ hosvd can give a good but not necessarily the best low - multilinear - rank approximation of the given tensor . to obtain a better approximation , people ( e.g. , @xcite ) solve the best rank-@xmath0 approximation problem @xmath1 where @xmath2 is a given tensor
, @xmath3 denotes mode-@xmath4 tensor - matrix multiplication ( see the definition in below ) , and @xmath5 with @xmath6 fixed , the optimal core tensor is given by @xmath7 . absorbing this @xmath8 into the objective
, one can write equivalently to ( see ( * ? ? ? * theorem 3.1 ) for detailed derivation ) @xmath9 one popular method for solving is the higher - order orthogonality iteration ( hooi ) ( see algorithm [ alg : hooi ] ) .
although hooi is commonly used and practically efficient ( already coded in the matlab tensor toolbox @xcite and tensorlab @xcite ) , existing works only show that the objective value of at the generated iterates increasingly converges to some value while the iterate sequence convergence is still an open question ( c.f .
@xcite ) . the iterate sequence convergence ( or equivalently the multilinear subspace convergence ) is important because without convergence , running the algorithm to different numbers of iterations may give severely different multilinear subspaces , and that will ultimately affect the results of applications . in this paper , we address this open question .
our main results are summarized in the following theorem .
[ thm : main ] let @xmath10 be the sequence generated by the hooi method .
we have : _ ( i)_. if @xmath10 has a block - nondegenerate ( see definition [ def : nondeg ] ) limit point @xmath11 , then @xmath11 is a critical point and also a block - wise maximizer of .
in addition , @xmath12 , where @xmath13 _ ( ii)_. if the starting point @xmath14 is sufficiently close to any block - nondegenerate local maximizer of , then the entire sequence @xmath15 must converge to some point @xmath16 and @xmath11 is a local maximizer of .
we make some remarks on the assumption and the convergence results .
[ rm : main ] the block - nondegeneracy assumption is also necessary because even starting from a critical point @xmath11 , the hooi method can still deviate from @xmath11 if it is not block - nondegenerate ( see remark [ rm : non - deg ] ) , that is , a degenerate critical point is not stable ( see @xcite for the perturbation analysis ) . in practice ,
the block - nondegeneracy is always observed because @xmath17 .
] , and it is implied by @xmath18 , where @xmath19 is defined by .
the assumption is similar to the one assumed by the orthogonal iteration method @xcite for computing @xmath20-dimensional dominant invariant subspace of a matrix @xmath21 .
typically , the convergence of the orthogonal iteration method requires that there is a positive gap between the @xmath20-th and @xmath22-th largest eigenvalues of @xmath21 in magnitude , because otherwise , the @xmath20-dimensional dominant invariant subspace of @xmath21 is not unique . for a block - wise maximizer @xmath11 , its block - nondegeneracy is equivalent to negative definiteness of each block hessian over the stiefel manifold @xmath23 .
the definition of our block - nondegeneracy is different from the nondegeneracy in @xcite .
a nondegenerate local maximizer in @xcite is one local maximizer that has negative definite hessian , so the nondegeneracy assumption in @xcite is strictly stronger than our block - nondegeneracy assumption . since the solution to each subproblem
( see ) of the hooi method is not unique and actually still a solution after multiplying any orthogonal matrix to its right , we can only hope to establish convergence of the projection matrix sequence @xmath15 instead of @xmath10 itself . before proceeding with discussion ,
we first review some basic concepts about tensor that we use in this paper ; see @xcite for more review .
the @xmath24-th component of an @xmath25-way tensor @xmath26 is denoted as @xmath27 . for @xmath28 ,
their inner product is defined in the same way as that for matrices , i.e. , @xmath29 the frobenius norm of @xmath26 is defined as @xmath30 a _ fiber _ of @xmath26 is a vector obtained by fixing all indices of @xmath26 except one .
the mode-@xmath4 _ matricization _ ( also called _ unfolding _ ) of @xmath26 is denoted as @xmath31 , which is a matrix with columns being the mode-@xmath4 fibers of @xmath26 in the lexicographical order .
the mode-@xmath4 product of @xmath32 with @xmath33 is written as @xmath34 which gives a tensor in @xmath35 and is defined component - wisely by @xmath36 if @xmath37 , then for any @xmath4 , @xmath38 the hooi method updates @xmath6 by maximizing the objective of alternatingly with respect to @xmath39 , one factor matrix at a time while the remaining ones are fixed . specifically , assuming the iterate to be @xmath40 at the beginning of the @xmath41-th iteration , it performs the following update sequentially from @xmath42 through @xmath25 : @xmath43 where we have used , and @xmath44 any orthonormal basis of the dominant @xmath45-dimensional left singular subspace of @xmath19 is a solution of .
the pseudocode of hooi is given in algorithm [ alg : hooi ] .
* input : * @xmath46 and @xmath0 * initialization : * choose @xmath47 with @xmath48 it is easy to implement algorithm [ alg : hooi ] by simply setting @xmath49 to the left @xmath45 leading singular vectors of @xmath19 .
this implementation is adopted in the matlab tensor toolbox @xcite and tensorlab @xcite .
however , such choice of @xmath49 causes difficulty to the convergence analysis of the hooi method . while preparing this paper
, we did not find any work that gives an iterate sequence convergence result of hooi , except for our recent paper @xcite that establishes subsequence convergence by assuming a strong condition on the entire iterate sequence .
the essential difficulty is the non - uniqueness of the solution of , and the leading singular vectors are not uniquely determined either . to tackle this difficulty
, we first analyze a greedy method , which always chooses one solution of that is closest to @xmath50 as follows : @xmath51 where @xmath52 the pseudocode of the greedy implementation is shown in algorithm [ alg : ghooi ] .
the subproblem in can be solved by the method given in remark [ rm : sol - sub ] .
although can in general have multiple solutions , we will show that near any limit point of the iterate sequence , it must have a unique solution . with the greedy implementation , we are able to establish iterate sequence convergence of the greedy hooi method ( i.e. , algorithm [ alg : ghooi ] ) , as shown in sections [ sec : subseq ] and [ sec : glb - cvg ] . through relating
( see and figure [ fig : cvg - bh ] ) the two iterate sequences generated by the original ( i.e. , algorithm [ alg : hooi ] ) and greedy hooi methods , we then establish the iterate sequence convergence of the original hooi method , as shown in section [ sec : pf - of - main ] .
* input : * @xmath2 and @xmath0 * initialization : * choose @xmath47 with @xmath48 besides the hooi method , several other methods have been developed for solving the low - multilinear - rank tensor approximation problem
. one of the earliest methods , called tuckals3 , was proposed in @xcite .
tuckals3 also sequentially updates @xmath53 through @xmath54 and then cycles the process , but different from hooi , it obtains approximate leading left singular vectors of @xmath19 by carrying out only one step of the so - called bauer - rutishauser method @xcite starting from @xmath50 .
this update is equivalent to solving a linearized version of the subproblem , and it prevents @xmath49 being far away from @xmath55 .
subsequence convergence of tuckals3 was established under the assumption that @xmath56 is positive definite for all @xmath4 and @xmath41 .
although tuckals3 has slightly lower per - iteration complexity than hooi , it does not converge as fast as hooi as demonstrated in figure [ fig : cvg - bh ] .
recently , some newton - type methods on manifolds were developed for the low - multilinear - rank tensor approximation problem such as the newton - grassmann method in @xcite and the riemannian trust region scheme in @xcite .
these methods usually exhibit superlinear convergence .
numerical experiments in @xcite demonstrate that for small - size problems , the riemannian trust region scheme can take much fewer iterations and also less time than the hooi method to reach a high - level accuracy based on the gradient information .
however , for medium - size or large - size problems , or if only medium - level accuracy is required , the hooi method is superior over the riemannian trust region scheme and also several other newton - type methods . under negative definiteness assumption on the hessian of a local maximizer ,
the newton - type methods are guaranteed to have superlinear or even quadratic local convergence ( c.f .
compared to our block - nondegeneracy assumption , their assumption is strictly stronger because as mentioned in remark [ rm : main ] , for a local maximizer , its block - nondegeneracy is equivalent to the negative definiteness of each block hessian . only with block - nondegeneracy assumption ,
it is not clear how to show the local convergence of the newton - type methods .
[ fig : cvg - bh ] [ cols="^,^ " , ] we summarize our contributions as follows .
* we propose a greedy hooi method .
for each update , we select from the best candidates one that is closest to the current iterate . with the greedy implementation ,
we show that any block - nondegenerate limit point is a critical point and also a block - wise maximizer , and if a block - nondegenerate limit point exists , then the entire iterate sequence converges to this limit point .
* through relating the iterates by the original hooi method to those by the greedy hooi method , we for the first time establish global convergence to a critical point by assuming the existence of a block - nondegenerate limit point and local convergence to a local maximizer by assuming sufficient closeness of the starting point to a block - nondegenerate local maximizer .
* as a result , we show that the iterate sequence converges to a globally optimal solution , if the starting point is sufficiently close to any block - nondegenerate globally optimal solution .
we use bold capital letters @xmath57 to denote matrices , caligraphic letters @xmath58 for ( set - valued ) mappings , and bold caligraphic letters @xmath59 for tensors .
@xmath60 denotes an identity matrix , whose size is clear from the context .
the @xmath61-th largest singular value of a matrix @xmath21 is denoted by @xmath62 .
the set of all orthonormal matrices in @xmath63 is denoted as @xmath64 . throughout the paper ,
we focus on real field , but our analysis can be directly extended to complex field .
[ def : nondeg ] a feasible solution @xmath6 of is block - nondegenerate if @xmath65 , where @xmath66 [ rm : non - deg ] in general , we are only able to claim convergence with existence of a block - nondegenerate limit point .
the original hooi method can deviate from a critical point if it is block - nondegenerate . to see this ,
suppose @xmath11 is a block - wise maximizer and thus a critical point .
assume @xmath67 .
let the original hooi method start from @xmath11 and update the first factor to @xmath68 .
then @xmath68 may not span the same subspace as that by @xmath69 because @xmath70 has more than one dominant @xmath71-dimensional left singular subspaces .
therefore , we can not guarantee the convergence of the learned multilinear subspace .
the rest of the paper is organized as follows .
section [ sec : subseq ] shows subsequence convergence of the greedy hooi . in section [ sec : glb - cvg ] , global convergence of the greedy hooi is established under the assumption of the existence of a block - nondegenerate limit point .
the convergence of the original hooi is shown in section [ sec : pf - of - main ] .
finally , section [ sec : discussion ] concludes the paper .
in this section , we show the subsequence convergence of algorithm [ alg : ghooi ] , namely , the criticality on the limit point of the iterates . if @xmath11 is a critical point of , then letting @xmath72 , we have @xmath73 to be a critical point of .
therefore , our analysis will only focus on .
the lagrangian function of is @xmath74 where @xmath75 is the lagrangian multiplier .
the kkt conditions or first - order optimality conditions of can be derived by @xmath76 , namely , [ eq : kkt0 ] @xmath77 where @xmath78 is defined in . from , we have @xmath79 .
hence , the condition in can be written to [ eq : kkt ] @xmath80 we say a point @xmath11 is a critical point of if it satisfies the conditions in and
. the following result is well known , and we will use it several times in our convergence analysis . [
lem : von - ineq ] for any matrices @xmath81 , it holds that @xmath82 the inequality holds with equality if @xmath21 and @xmath83 have the same left and right singular vectors . to show the convergence of algorithm [ alg : ghooi ] , we analyze the solution of the subproblem , which can be written in the following general form : @xmath84 where @xmath85 and @xmath86 are given , and @xmath87 given a matrix @xmath86 and positive integer @xmath88 , define @xmath89 for any @xmath90
, if @xmath91 , i.e. , they span the same subspace , we say they are equivalent . by this equivalence relation , we partition @xmath92 to a set of equivalence classes and form a quotient set denoted as @xmath93 . throughout the paper , we regard @xmath93 as the finite set of orthonormal matrices , and each of its elements is a representative of the bases that span the same subspace .
if @xmath94 , then @xmath83 has a unique dominant @xmath20-dimensional left singular subspace , and @xmath93 is a singleton .
however , if @xmath95 , then @xmath83 has multiple dominant @xmath20-dimensional left singular subspaces , and @xmath93 has more than one element .
[ prop : projh ] the problem has a unique solution if the following two conditions hold : 1 . if @xmath96 , then @xmath97 is nonsingular ; 2 . for any @xmath98 , if @xmath99 , then @xmath100 ; where @xmath101 denotes matrix nuclear norm , defined as the sum of all singular values of a matrix .
assume @xmath102 and @xmath103 are both solutions of .
note that @xmath104 in is exactly the set @xmath92 .
hence , @xmath105 and @xmath106 for @xmath107 and some @xmath108 .
note @xmath109 then by lemma [ lem : von - ineq ] and the optimality of @xmath102 on solving , we have @xmath110 hence , from items 1 and 2 , it follows that @xmath111 , and similarly @xmath112 .
let @xmath113 be the full svd of @xmath97 and @xmath114 , so @xmath115 .
then from , it holds that @xmath116 note that @xmath117 and @xmath118 .
the equality @xmath119 holds only if @xmath120 .
since @xmath121 is orthogonal , we must have @xmath122 .
hence , @xmath123 and @xmath124 . for the same reason , @xmath125 .
therefore , @xmath126 , and the solution of is unique .
it is easy to see that the two conditions in items 1 and 2 are also necessary for uniqueness of the solution of .
define @xmath127 then for any @xmath128 , has a unique solution , which we denote as @xmath129 . in this way
, @xmath130 defines a mapping on @xmath131 .
[ rm : sol - sub ] the proof of proposition [ prop : projh ] provides a way for finding a solution of .
find @xmath96 and get full svd of @xmath113 .
then @xmath132 is a solution of .
using proposition [ prop : projh ] , one can easily show the following two corollaries .
if @xmath21 is sufficiently close to one @xmath133 in @xmath92 , then the solution of is unique . if @xmath134 , then @xmath135 , i.e. , @xmath21 is a fixed point . furthermore , we can show the continuity of @xmath130 .
[ thm : ops ] the mapping @xmath130 is continuous on @xmath131 . for convenience of the description , in this proof
, we simply write @xmath136 and @xmath130 to @xmath137 and @xmath138 , respectively . for any @xmath139 , let @xmath140 . if @xmath138 is not continuous at @xmath21 , then there exists @xmath141 and a sequence @xmath142 in @xmath143 such that @xmath144 and @xmath145 , where @xmath146 . by the definition of @xmath143 , we know that there is @xmath147 such that @xmath148 for any @xmath149 . similarly , there is a sequence @xmath150 in @xmath151 such that for each @xmath41 , @xmath152 for any @xmath153 .
let @xmath154 .
there is a sufficiently large integer @xmath155 such that for all @xmath156 , it holds @xmath157 and @xmath158 .
note @xmath159 .
hence , @xmath160 , i.e. , @xmath161 . therefore , by the definition of @xmath162 , it must hold that @xmath163 .
hence , we can write @xmath164 and @xmath165 for all @xmath156 , where @xmath166 . note @xmath167 as @xmath168 .
then from the proof of proposition [ prop : projh ] , we have @xmath169 and thus @xmath170 as @xmath168 .
this contradicts to @xmath145 .
therefore , @xmath138 is continuous at @xmath21 . since @xmath21 is an arbitrary point in @xmath143 , this completes the proof .
one can also show the following result .
[ thm : conty ] assume @xmath94 and @xmath171 as @xmath168 .
if @xmath128 , then there is a sufficiently large integer @xmath155 such that @xmath172 for all @xmath156 , and @xmath173 by the assumption @xmath94 , @xmath93 is a singleton .
let @xmath98 . then from @xmath174
, it follows that @xmath175 is nonsingular . since @xmath171 as @xmath168
, there exists an integer @xmath155 , such that @xmath176 , i.e. , @xmath177 is a singleton for all @xmath156 .
let @xmath178 .
we can choose the representative satisfying @xmath179 , since @xmath171 .
therefore , taking another larger @xmath155 if necessary , we have that @xmath180 is nonsingular and thus @xmath172 for all @xmath156
. finally , using remark [ rm : sol - sub ] and @xmath179 , we have and complete the proof .
we also need the following result .
[ lem : crit - pt ] for any feasible solution @xmath11 , if @xmath181 , then @xmath11 is a critical point and also a block - wise maximizer of , where @xmath182 note that @xmath181 implies that @xmath183 is a basis of the dominant @xmath45-dimensional left singular subspace of @xmath184 . hence , @xmath185
therefore , @xmath11 is a critical point . in addition
, @xmath181 implies that @xmath183 is a solution to @xmath186 over @xmath23 for all @xmath4 .
hence , @xmath11 is a block - wise maximizer .
this completes the proof .
now we are ready to show the subsequence convergence result .
[ thm : subseq ] let @xmath10 be the sequence generated from algorithm [ alg : ghooi ] .
then any block - nondegenerate limit point of @xmath10 is a critical point and a block - wise maximizer of .
suppose that @xmath11 is one block - nondegenerate limit point and the subsequence @xmath187 converges to @xmath11 .
from the update rule in , it is easy to see @xmath188 we claim that @xmath69 is a solution of @xmath189 .
otherwise , @xmath190 .
note @xmath191 which contradicts to .
hence , @xmath192 .
note that @xmath193 as @xmath194 and @xmath195 as @xmath196 is sufficiently large . from the block - nondegeneracy of @xmath11 and theorems [ thm : ops ] and [ thm : conty ] ,
we have @xmath197 hence , taking a sufficiently large @xmath196 , we can make @xmath198 sufficiently small , and thus we can repeat the above arguments for @xmath199 to conclude @xmath200 therefore , from the definition of @xmath201 , it holds that @xmath181 , and @xmath11 is a critical point and a block - wise maximizer of from lemma [ lem : crit - pt ] .
the result in is a key step to have the subsequence convergence . in general , without the block - nondegeneracy assumption , it may not hold .
in this section , we assume the existence of one block - nondegenerate limit point and show global convergence of algorithm [ alg : ghooi ] .
the key tool we use is the so - called kurdyka - ojasiewicz ( kl ) property ( see definition [ def : kl ] below ) .
let @xmath202 and @xmath203 be the indicator function on @xmath23 for @xmath204 .
also let @xmath205 then is equivalent to @xmath206 , and @xmath11 is a critical point of _ if and only if _
@xmath207 , where @xmath208 denotes the limiting frchet subdifferential ( see @xcite for example ) .
we show the global convergence of algorithm [ alg : ghooi ] also by analyzing the solution of the subproblem .
as shown below , if there is a positive gap between @xmath209 and @xmath210 , the distance between @xmath50 and @xmath49 can be bounded by the objective difference .
[ thm : key - ineq ] given @xmath85 and @xmath86 , any solution @xmath211 of satisfies @xmath212 note @xmath213 for some @xmath98 and @xmath214 .
let @xmath215 be the full svd of @xmath83 .
also , let @xmath216 and @xmath217 then @xmath218 and @xmath219 from @xmath220 . as in the proof of proposition [ prop : projh
] , we have @xmath221 and @xmath222 where the last equality is from lemma [ lem : von - ineq ] and the optimality of @xmath211 for .
also , note that @xmath223 assume @xmath224 to be the full svd of @xmath225 . then @xmath226 let @xmath227 be the first @xmath20 largest singular values of @xmath225 .
then @xmath228 , and using lemma [ lem : von - ineq ] again , we have @xmath229 and @xmath230 hence , from and through , we have @xmath231 where the last inequality is from @xmath232 . using the fact @xmath233 $ ] , we have @xmath234 and thus from , it follows that @xmath235 plugging the above inequality into , we have the desired result . using theorem [ thm : key - ineq ] , we show the following result . [
lem : sq - bd ] let @xmath10 be the sequence generated from algorithm [ alg : ghooi ] .
assume it has a block - nondegenerate limit point @xmath11 .
then there is a constant @xmath236 such that if @xmath40 is sufficiently close to @xmath11 , we have @xmath237 it is easy to see that there exists a small positive number @xmath162 such that if @xmath238 , then @xmath239 where the strict inequality is from the block - nondegeneracy of @xmath11 .
assume @xmath40 is sufficiently close to @xmath11 such that @xmath240 from theorem [ thm : key - ineq ] , it follows that @xmath241 where @xmath242 is defined in , and we have used .
hence , @xmath243 and @xmath244 repeating the above arguments , in general , we have for all @xmath4 that @xmath245 and @xmath246 therefore , every intermediate point @xmath247 is in @xmath248 , and thus for all @xmath4 , @xmath249 let @xmath250 .
summing the above inequality from @xmath42 to @xmath25 gives the desired result .
using lemma [ lem : sq - bd ] and the kl property of @xmath251 , we show the global convergence of algorithm [ alg : ghooi ] .
[ def : kl ] a function @xmath252 satisfies the kl property at point @xmath253 if there exists @xmath254 such that @xmath255 is bounded around @xmath256 under the notational conventions : @xmath257 in other words , in a certain neighborhood @xmath258 of @xmath256 , there exists @xmath259 for some @xmath260 and @xmath254 such that the kl inequality holds @xmath261 where @xmath262 and @xmath263 .
the kl property was introduced by ojasiewicz @xcite on real analytic functions , for which the term with @xmath264 in is bounded around any critical point @xmath256 .
kurdyka extended this property to functions on the @xmath265-minimal structure in @xcite .
recently , the kl inequality was extended to nonsmooth sub - analytic functions @xcite .
the works @xcite give a lot of concrete examples that own the property .
the function @xmath251 is one of their examples and thus has the kl property .
[ thm : glb - cvg ] if @xmath11 is a block - nondegenerate limit point of the sequence @xmath10 generated from algorithm [ alg : ghooi ] , then @xmath11 is a critical point of , and @xmath266 from theorem [ thm : subseq ] , we have the criticality of @xmath11 , so we only need to show . note that @xmath267 is nondecreasing with repsect to @xmath41 and thus converges to @xmath268 .
we assume @xmath269 otherwise , if for some @xmath155 , @xmath270 , we must have @xmath271 .
since @xmath251 has the kl property , then in a neighborhood @xmath272 , there exists @xmath259 for some @xmath260 and @xmath254 such that @xmath273 if necessary , taking a smaller @xmath274 , we assume @xmath275 where @xmath162 and @xmath276 s are defined in the same way as those in the proof of lemma [ lem : sq - bd ] .
note that there is a constant @xmath277 such that @xmath278 where @xmath279 since @xmath11 is a limit point , there is a subsequence @xmath187 convergent to @xmath11 .
hence , we can choose a sufficiently large @xmath280 such that @xmath281 is sufficiently close to @xmath11 .
without loss of generality , we assume @xmath14 ( otherwise set @xmath281 as a new starting point ) is sufficiently close to @xmath11 such that @xmath282 and @xmath283 which can be guaranteed from lemma [ lem : sq - bd ] and where @xmath236 is the same as that in lemma [ lem : sq - bd ] .
assume @xmath284 for @xmath285 .
we go to show @xmath286 and thus @xmath287 by induction . for any @xmath204 , from the optimality of @xmath50 on problem @xmath288
, it holds that @xmath289 hence , @xmath290 letting @xmath291 and @xmath292 , we have @xmath293 which implies @xmath294 summing the above inequality from @xmath295 to @xmath296 and simplifying the summation gives @xmath297 and thus @xmath298 therefore , @xmath286 and @xmath287 , by induction .
hence , holds for all @xmath296 , and letting @xmath299 , we conclude that @xmath10 is a cauchy sequence and converges .
since @xmath11 is a limit point , then @xmath300 as @xmath168 .
this completes the proof .
as long as the starting point is sufficiently close to any block - nondegenerate local maximizer , algorithm [ alg : ghooi ] will yield an iterate sequence convergent to a local maximizer as summarized below .
[ thm : loc - min ] assume algorithm [ alg : ghooi ] starts from any point @xmath14 that is sufficiently close to one block - nondegenerate local maximizer @xmath301 of @xmath302
. then the sequence @xmath10 converges to a local maximizer .
first , note that if some @xmath281 is sufficiently close to @xmath301 and @xmath303 , then @xmath281 must also be a local maximizer and block - nondegenerate . in this case , @xmath271 .
hence , without loss of generality , we can assume @xmath304 .
secondly , note that in the proof of theorem [ thm : glb - cvg ] , we only use @xmath305 and the sufficient closeness of @xmath14 to @xmath11 to show @xmath10 to be a cauchy sequence .
therefore , repeating the same arguments , we can show that if @xmath14 is sufficiently close to @xmath301 , then @xmath10 is a cauchy sequence and thus converges to a block - nondegenerate point @xmath11 near @xmath301 . from theorem [ thm : subseq ]
, it follows that @xmath11 is a critical point .
we claim @xmath306 , i.e. , @xmath11 is a local maximizer .
if otherwise @xmath307 , then by the kl inequality , it holds that @xmath308 , which contradicts to @xmath207 .
hence , @xmath306 .
this completes the proof . from theorem [ thm : loc - min ] , we can easily get the following local convergence to a globally optimal solution .
[ thm : glb - opt ] assume algorithm [ alg : ghooi ] starts from any point @xmath14 that is sufficiently close to one block - nondegenerate globally optimal solution @xmath301 of
. then the sequence @xmath10 converges to a globally optimal solution .
in this section , we analyze the convergence of the original hooi method by relating its iterate sequence to that of the greedy hooi method .
because any solution to each subproblem of the original hooi method is still a solution after arbitrary rotation , we do not hope to establish convergence on the iterate sequence @xmath10 itself .
instead , we show the convergence of the projection matrix sequence @xmath15 .
first note that @xmath309 we also need the following two lemmas .
[ limit - pt ] if @xmath310 and @xmath311 is a critical point of , then @xmath11 is also a critical point .
since @xmath311 is a critical point of , it holds that @xmath312 and @xmath313 for all @xmath4 .
note that @xmath310 implies @xmath314 .
hence , for any @xmath4 , @xmath315 multiplying @xmath183 to both sides and noting @xmath316 gives @xmath317 and thus @xmath11 is a critical point .
[ lem : nchg ] let @xmath318 be the sequence generated by the original hooi method and assume it has a block - nondegenerate limit point @xmath11 . if for some @xmath155 , @xmath319 , then there is an integer @xmath320 such that @xmath321 .
because @xmath267 is nondecreasing and upper bounded , we have @xmath322 and @xmath323 , so if @xmath319 , then @xmath324 .
since @xmath11 is a limit point , there must be an integer @xmath320 such that @xmath325 is sufficiently close to @xmath11 and @xmath325 is block - nondegenerate .
hence , @xmath326 has a unique dominant @xmath71-dimensional left singular subspace .
note @xmath327 therefore , @xmath328 and @xmath329 both span the dominant @xmath71-dimensional left singular subspace of @xmath326 , and thus @xmath330 . using , we can repeat the arguments to have @xmath331 , i.e. , @xmath332 . now starting from @xmath333 and repeating the arguments
, we have the desired result . by lemma [ lem : nchg ] , without loss of generality , we assume @xmath334 in the remaining analysis . with lemmas [ limit - pt ] and [ lem : nchg ] , we are now ready to prove the main theorem .
* part ( i ) : * since @xmath11 is a limit point of @xmath335 , there is a subsequence @xmath187 convergent to @xmath11 , and there is @xmath280 such that @xmath281 is sufficiently close to @xmath11 . without loss of generality ,
we assume that @xmath14 is sufficiently close to @xmath11 because otherwise we can set @xmath281 as a new starting point and the convergence of @xmath318 is equivalent to that of @xmath336 .
let @xmath337 be the sequence generated by the greedy hooi method starting from @xmath338 .
we go to show that if @xmath14 is sufficiently close to @xmath11 , then @xmath339 repeating the same arguments in the proof of lemma [ lem : sq - bd ] , we have that if @xmath14 is sufficiently close to @xmath11 , then @xmath340 is also sufficiently close to @xmath11 .
note that when @xmath14 is sufficiently close to @xmath11 , it is block - nondegenerate and @xmath341 .
hence , @xmath342 and @xmath343 both span the dominant @xmath71-dimensional left singular subspace of @xmath344 and thus @xmath345 . since both @xmath346 and @xmath340 are sufficiently close to @xmath11 , we have @xmath347 .
note @xmath348 .
hence , @xmath349 and @xmath350 both span the dominant @xmath351-dimensional left singular subspace of @xmath352 and thus @xmath353 . repeating the above arguments , we have @xmath354 , i.e. , @xmath355 .
assume that for some integer @xmath356 , it holds @xmath357 and @xmath358 for all @xmath359 , where @xmath274 is sufficiently small and plays the same role as that in the proof of theorem [ thm : glb - cvg ] . from
, it follows that @xmath360 . through the same arguments as those in the proof of theorem [ thm : glb - cvg ] , we have @xmath361 , and thus @xmath362 by the above arguments that show @xmath355 . by induction
, we have the result in . taking another subsequence if necessary
, we can assume @xmath363 converging to @xmath311 and thus @xmath310 by .
note that the block - nondegeneracy of @xmath311 is equivalent to that of @xmath11 .
hence , @xmath311 is block - nondegenerate and is a critical point and a block - wise maximizer by theorem [ thm : subseq ] , and @xmath364 converges to @xmath311 by theorem [ thm : glb - cvg ] .
therefore , @xmath365 converges to @xmath366 . from lemma [ limit - pt ]
, we have that @xmath11 is a critical point of , and from , @xmath11 is a block - wise maximizer .
this completes the proof of part ( i ) . *
part ( ii ) : * let @xmath367 be the sequence generated by the greedy hooi method starting from @xmath338 . from theorem [ thm : loc - min ] , it follows that @xmath364 converges to a local maximizer @xmath311 of .
in addition , by similar arguments as those in the proof of part ( i ) , we can show that still holds .
hence , @xmath365 converges to @xmath368 , and this completes the proof .
we proposed a greedy hooi method and established its iterate sequence convergence by assuming existence of a block - nondegenerate limit point . through relating the iterates by the original hooi to those by the greedy hooi ,
we have shown the global convergence of the hooi method , for the first time .
in addition , if the starting point is sufficiently close to any block - nondegenerate locally optimal point , we showed that the original hooi could guarantee convergence to a locally optimal solution .
, _ proximal alternating minimization and projection methods for nonconvex problems : an approach based on the kurdyka - lojasiewicz inequality _ , mathematics of operations research , 35 ( 2010 ) , pp
. 438457 .
height 2pt depth -1.6pt width 23pt , _ perturbation theory and optimality conditions for the best multilinear rank approximation of a tensor _ , siam journal on matrix analysis and applications , 32 ( 2011 ) , pp .
14221450 . ,
_ a block coordinate descent method for regularized multiconvex optimization with applications to nonnegative tensor factorization and completion _
, siam journal on imaging sciences , 6 ( 2013 ) , pp . | the higher - order orthogonality iteration ( hooi ) has been popularly used for finding a best low - multilinear - rank approximation of a tensor .
however , its iterate sequence convergence is still an open question . in this paper
, we first analyze a greedy hooi , which updates each factor matrix by selecting from the best candidates one that is closest to the current iterate . assuming the existence of a block - nondegenerate limit point , we establish its global convergence through the so - called kurdyka - ojasiewicz ( kl ) property .
in addition , we show that if the starting point is sufficiently close to any block - nondegenerate globally optimal solution , the greedy hooi produces a sequence convergent to a globally optimal solution . relating the iterate sequence by the original hooi to that by the greedy hooi , we then show that the same convergence results hold for the original hooi and thus positively address the open question .
higher - order orthogonality iteration ( hooi ) , global convergence , kurdyka - ojasiewicz ( kl ) property , greedy algorithm , block coordinate descent |
the jpa device was fabricated on a 275 @xmath77 m thick intrinsic silicon chip with standard e - beam lithography and bilayer resist process .
the josephson junctions in the squid were made using double - angle evaporation with a 200 thick @xmath89 layer at @xmath90 , followed by a three minute oxidation in 0.7 torr of pure oxygen , and a 500 thick second layer of @xmath89 at @xmath91 .
all depositions were done in an e - beam evaporator . from room temperature resistance measurements ,
the total critical current of the squid was estimated to be 2.8 @xmath77a .
the capacitors in parallel plate geometry were designed to have an area of @xmath92 with a 300 thick @xmath93 sandwiched between two 700 thick @xmath89 layers .
the al@xmath94o@xmath95 and second @xmath89 layer were deposited with the sample stage continuously rotating and a deposition angle of @xmath96 and @xmath97 respectively to prevent shorting between the two @xmath89 layers .
the estimated capacitance was 3.4 pf .
we used a voltage biased tunnel junction as a calibrated source of noise @xcite to measure the system noise temperature of our amplification chain . the shot noise tunnel junction ( sntj )
is a standard @xmath98 tunnel junction with @xmath99 nominal resistance and is embedded in a @xmath99 coplanar wave guide transmission line .
a strong rare earth magnet is typically used to suppress superconductivity in aluminum to realize a normal tunnel junction at dilution temperatures .
the noise power coupled to a matched load by a voltage biased sntj at temperature @xmath100 and frequency @xmath101 is given by @xcite @xmath102,\ ] ] where @xmath103 is the gain of the amplification chain , @xmath104 is the boltzmann constant , @xmath105 is the bandwidth over which noise is measured , @xmath106 is the noise temperature of the amplification chain , @xmath107 is planck s constant and @xmath108 is the voltage applied across the tunnel junction . using the above equation , a measurement of noise power @xmath109 as a function of @xmath108 allows determination of system temperature @xmath100 , gain @xmath103 and most importantly the noise temperature @xmath106 . given the system gain , one can extract the input signal power at the plane of the sample allowing one to calibrate the 1-db compression point as well .
we use the high voltage limit of this equation which results in a linear relation between @xmath109 and @xmath108 given by @xmath110 we can then extract @xmath103 and @xmath106 using straight line fits . in a separate experiment , we had ensured that identical results were obtained for @xmath103 and @xmath106 when using eq . or .
this allowed us to eliminate the strong magnet since at voltages much greater than the gap , the superconducting tunnel junction behaves like a normal tunnel junction . a cryogenic microwave switch was used to route the output of the sntj or the jpa to the amplification chain as shown in fig .
[ fig : setup ] . with
the sntj switched in , we measured the system noise temperature @xmath106 and the result is shown in fig .
[ fig : tnsys](a ) .
the periodic variations are a result of impedance mismatch primarily in the circulator but the mean value of @xmath106 was consistent with our hemt noise temperature after accounting for signal attenuation between the sntj and the hemt . of the amplification chain measured using a sntj noise source without the jpa .
( b ) system noise temperature @xmath111 of the amplification chain with the jpa switched in . the error bars ( @xmath112 )
reflect the error related to the difference in signal attenuation between the signal paths from the sntj and the jpa respectively , to the cryogenic switch.,scaledwidth=100.0% ] we tried to keep the signal path from the jpa and the sntj to the cryogenic switch identical , so that the measured @xmath106 ( using sntj ) was still valid with the jpa switched in .
however , we estimated that there was a mismatch of about @xmath113 db in the transmission between the two signal paths .
this directly leads to a fractional error in the estimation of the @xmath106 of @xmath114 which is reflected by the error bars in fig . [
fig : tnsys](a ) .
other sources of uncertainty like reduced power coupling due to impedance mismatch of the sntj , calibration error in the voltage bias across the sntj , and statistical errors in fitting were insignificant in comparison . with the jpa switched in , its noise temperature @xmath115 is determined by measuring the signal to noise ratio improvement @xmath116 of a probe signal when the jpa is biased to provide 20 db maximum gain at the pump frequency : @xmath117=t_\text{n , sys}^{\rm jpa}(f)-t_\text{n , sys}(f)g(f)^{-1},\ ] ] where @xmath118 is the gain of the jpa . here
@xmath111 is the system noise temperature with the jpa switched in and is plotted in fig .
[ fig : tnsys](b ) .
the fractional error in @xmath111 and @xmath119 is the same as that in @xmath106 which is reflected in the error bars in fig .
[ fig : tnsys](b ) and fig . 3 in the main text .
here , we briefly outline why one can use the effective two - mode hamiltonian to describe the jpa , when it involves only a single physical cavity mode .
the standard approach to a driven jpa is to model it as a driven cavity with a kerr nonlinearity .
the cavity hamiltonian in this case ( in an interaction picture at the pump frequency ) has the form@xcite : @xmath120 while the equations of motion for the fourier - transformed cavity mode operators take the form : @xmath121 & = \left(-i \omega_{d } ' - \kappa_0/2 \right ) \hat{a}[\omega ] - \lambda \hat{a}^\dagger[\omega ] , \\
-i\omega\hat{a}^\dagger[\omega ] & = \left(i \omega_{d } ' - \kappa_0/2 \right ) \hat{a}[\omega ] - \lambda \hat{a}[\omega],\end{aligned}\ ] ] where we have used the following convention for the fourier - transform @xmath122 & = \int a(t)e^{i\omega t}dt , \\
a^\dagger[\omega ] & = \int a^\dagger(t)e^{i\omega t}dt.\end{aligned}\ ] ] note that @xmath123 = \left ( \hat{a}[-\omega ] \right)^\dagger$ ] . for @xmath124 ,
the operators @xmath125 $ ] and @xmath123 $ ] are commuting operators describing independent degrees of freedom ( namely photons at a frequency @xmath27 versus those at @xmath28 ) . to stress this independence
, we can relabel operators ( for @xmath126 ) as : @xmath127 & \rightarrow \hat{a}_1[\omega ] , \\
\hat{a}[-\omega ] & \rightarrow \hat{a}_2[-\omega ] .
\end{aligned}\ ] ] the above equations of motion could then be equivalently derived from the starting hamiltonian given in eq .
( 1 ) of the main text .
when we use the jpa as a phase - preserving amplifier , it is convenient to use this effective two - mode description , as then one has a direct equivalence to a standard two - mode non - degenerate parametric amplifier . while this description is convenient , we stress that identical results would be obtained if one stuck to a single mode description .
in this section , we will show how the susceptibility matrix @xmath128 can be obtained from the classical equations of motion of a driven non - linear oscillator .
[ fig : jpadrive ] depicts a jpa consisting of a squid loop with a total critical current @xmath129 shunted by a capacitor @xmath130 .
the jpa is coupled to an environment of frequency independent shunt impedance @xmath36 .
the equation of motion of the gauge - invariant phase @xmath131 across the josephson junction is then given by @xmath132 where @xmath133 is the reduced flux quantum and @xmath134 is the current applied from an external source . @xmath134 in general consists of two terms : a drive @xmath135 to bias and a signal @xmath136 to be amplified . keeping only the first non - linear term in the expansion of @xmath137 , eq . becomes @xmath138 where @xmath139 is the damping rate of the oscillator amplitude and @xmath140 is the plasma frequency of the jpa .
the steady state dynamics @xmath141 is controlled by the strong drive @xmath142 and is obtained by solving eq . while setting @xmath143 to zero .
we write this solution as @xmath144 next we perform a perturbative expansion around the steady state solution @xcite in order to find the response due to the signal @xmath143 by considering @xmath145 with the assumption @xmath146 . substituting @xmath131 into eq . and retaining terms only up to first order in @xmath147 one obtains the following differential equation @xmath148 where @xmath149 .
. represents a harmonic oscillator driven parametrically at @xmath150 and its solution can be easily obtained by going to the frequency domain . taking the fourier transform of eq . ,
applying rotating wave approximation , and moving to a frequency reference centered at @xmath6 one finds @xmath151+\lambda e^{i\theta}\delta_s^\dagger[-\omega]=e^{-i\theta}\dfrac{ia_s[\omega]}{2\omega_p},\ ] ] where @xmath152\equiv\delta_s[\omega ] , \ \delta_s^\dagger[-\omega]\equiv\delta_s[\omega-2\omega_d ] , \
a_s[\omega]\equiv a_s[\omega],$ ] with @xmath153 $ ] and @xmath154 $ ] being the fourier transforms of @xmath147 and @xmath143 respectively . here
@xmath155 is the effective parametric pump which is controlled by the drive amplitude @xmath156 .
the other symbols have the same meaning as described in the main text . defining @xmath157 = e^{-i \theta } \delta_s[\omega]$ ] , @xmath158 = e^{i \theta } \delta_s^\dagger[- \omega]$ ] and @xmath159
= i e^{-i \theta}a_s[\omega]/(2 \omega_p)$ ] , we get , @xmath160+\lambda \hat{a}_2^\dagger[-\omega]=\hat{a}_\text{in}[\omega].\ ] ] combining eq . along with its hermitian conjugate , one finds the susceptibility matrix @xmath16 $ ] that relates the intracavity fields @xmath17=(\hat{a}_1[\omega],\hat{a}_2^\dagger[-\omega])$ ] to the input fields @xmath18 $ ] as @xmath17=\chi[\omega]\vec{a}_{\rm in}[\omega]$ ] , where @xmath19)^{-1}= -i\begin{pmatrix } ( \omega-\omega_{d}')+i\kappa_0/2 & i\lambda \\ i\lambda & ( \omega+\omega_{d}')+i\kappa_0/2 \end{pmatrix}.\ ] ] this susceptibility is the same as eq .
( 2 ) in the main text in the absence of any auxiliary resonator so that @xmath47=-i \kappa_0/2 $ ] . in the most general case , when the environment is frequency
dependent , one has to use a frequency dependent self energy to compute the susceptibility matrix as shown in the main text .
the self energy is related to the environmental impedance as explained in the next section .
in this section , we show how the effects of a general input admittance seen by the jpa can be treated by using a self energy , i.e. frequency dependent terms which modify the jpa susceptibility defined in eq .
( 2 ) of the main text .
we stress that as the admittance is associated with a linear system , its treatment via a self energy is exact .
classically , one can obtain the self energy simply by solving the ( linear ) equations of motion for the admittance degrees of freedom , and then substituting these into the equations of motion for the jpa degrees of freedom ; this is shown explicitly in the next section . here , we present a derivation of the self energy based on quantum linear response .
this approach yields additional insights , and is convenient if one is also interested in understanding quantum noise properties of the amplifier .
we stress that the frequency - dependent self energy approach we describe is both a powerful and general way of approaching our system , as it is applicable to an arbitrary input admittance , including cases where several resonances contribute .
we start by working with the single - mode description of the jpa resonator , as described in the previous section , but in the lab frame .
the coupling between the jpa resonator and external environment ( comprising of the auxiliary circuit plus transmission line ) has the form ( @xmath161 ) : @xmath162 here @xmath163 is the flux operator for the jpa cavity mode , and @xmath164 is the current operator for the external environment .
as the environment is a linear circuit , it corresponds to a collection of independent bosonic modes ( see , e.g. , ref . ) ; further , @xmath164 is linear in the creation and annihilation operators for these modes .
it follows that the retarded self energy @xmath165 $ ] we want is directly determined by the retarded correlation function of @xmath164 : @xmath166 & = \frac { \omega_p l_p}{2 } g^r[\omega ] \\ & = \frac{\omega_p l_p}{2 } \left(-i \int_0^{\infty } dt \ , \langle \hat{i}_{\rm env}(t ) \hat{i}_{\rm env}(0 ) \rangle
e^{i \omega t } \right ) \\ & = \frac { \omega_p l_p}{2 } \left(-i \omega \tilde{y}_{\rm in , lab}[\omega ] \right).\end{aligned}\ ] ] in the last line , we have used the kubo formula to relate the required correlation function to the admittance of the environment @xmath167 $ ] as seen by the jpa .
the tilde here indicates that we are using the standard physicist convention for the fourier transform in writing the admittance .
the frequency dependent admittance using the standard engineering convention @xmath168 $ ] is given by @xmath168 = \tilde{y}_{\rm in , lab}[\omega]^*$ ] .
we use this in what follows . finally , we move to the interaction picture ( with respect to the drive frequency @xmath6 ) , and use our two - mode representation ( c.f .
( 1 ) of the main text ) .
this lets us identify the self - energies for each of these effective modes . for the effective signal mode self energy ( frequencies above @xmath6 ) , we have : @xmath169 & \equiv \sigma_{\rm lab}[\omega_d + \omega ] = -i \frac { \omega_p l_p}{2 } \left(\omega_d + \omega \right ) y_{\rm in , lab}^*[\omega_d + \omega ] \\ & \simeq -i \frac { 1 } { 2 c_p } y_{\rm in , lab}^ * [ \omega_d + \omega ] , \end{aligned}\ ] ] where in the last line , we are using the fact that we will be interested in signal frequencies @xmath20 satisfying @xmath170 and @xmath171 . similarly , @xmath172 & \equiv -\left(\sigma_{\rm lab}[\omega_d - \omega]\right)^ * \\ & \simeq -i \frac { 1 } { 2 c_p } y_{\rm in , lab } [ \omega_d - \omega].\end{aligned}\ ] ] again , we stress that apart from simplifying to frequencies @xmath170 , this derivation is exact , for the simple reason that linear response is exact for a linear system ( i.e. the input admittance ) .
the possible hybridization between resonances in the input admittance and the jpa is fully retained in the self energy approach ; we show this explicitly below . in the main part of the text , we focus on the simple case where the input admittance seen by the jpa corresponds to a single damped series lc resonance . the self energy corresponding to eqs .
( 9 ) of the main text is simply : @xmath173 = \sigma_2[\omega ] = \frac{1}{2 \alpha c_p } \frac{1}{\omega + i r / \alpha}.\end{aligned}\ ] ]
the simple pole in this self energy reflects the response of the series lc resonance ( whose resonance frequency is shifted to zero in our rotating frame ) . as our system in this case
is composed of two coupled harmonic oscillators ( i.e. the main jpa resonance , and the extra series lc resonance ) , we expect to find coupled normal modes .
we stress that this coupling and corresponding hybridization is fully retained in the self energy approach .
consider the case where there is no effective parametric driving , and hence @xmath174 in eq .
( 2 ) of the main text for the susceptibility matrix @xmath16 $ ] .
the susceptibility has simple poles at frequencies satisfying @xmath175 = 0 $ ] . using the form for the self energy above
, one easily finds that these poles are at @xmath176 , with : @xmath177 these simply represent the two ( damped ) normal modes of the system .
one can easily see that the effective splitting of these normal modes scales as @xmath178 , while their broadening scales as @xmath179 .
they thus become resolved in the limit of large @xmath57 ( i.e. large @xmath58 ) . as discussed in the main text after eq .
( 11 ) , one can use the normal mode basis to get an intuitive picture of behavior of the frequency - dependent gain in fig .
2 . for large @xmath58 ,
the normal modes are resolved , and the two peaks in the gain curve correspond to contributions from each normal mode .
in contrast , for the optimal value of @xmath58 , the two normal modes are not resolved , and they contribute together to yield a gain profile which is extremely flat .
qualitatively , this behavior can be understood by considering the leading frequency dependence of @xmath180 $ ] near @xmath46 .
we stress however that our full quantitative results keep the full frequency dependence of @xmath180 $ ] , and thus are not based on any assumption of an almost frequency - independent self energy .
while the normal - mode picture provides important intuition , we prefer in the main text to focus the discussion on our general self energy approach ( as opposed to explicitly focusing on a system with two resonances ) .
this lets us understand the general principle behind the flattening out of the gain , something that applies even for input admittances that are more complicated than that of a simple series lc resonance .
we also stress that even in the simple case where the input admittance is a series lc circuit , working in the basis of the dissipation - free ( i.e. @xmath181 ) system normal modes can be misleading , as the dissipation is strong in the most interesting regime , and because the dissipation is highly asymmetric ( e.g. the series lc is damped , the main jpa cavity has negligible damping ) .
as discussed in the previous section , given the frequency dependent admittance of the environment one can compute the response of the parametric amplifier .
this technique is quite powerful since it allows one to compute the response for any linear circuit which might be used to modify the environment seen by the jpa . in our experiment
, we used a combination of @xmath68 and @xmath0 transformers to introduce a positive linear slope in the imaginary part of the input impedance seen by the jpa .
the equivalent circuit model is a simple series lc circuit and the corresponding self energy is given by eqs . ( 9a ) and ( 9b ) in the main text .
we will now show that one gets identical results by considering the same system as two coupled oscillators with a frequency independent environment .
the classical equations of motion of the circuit in fig .
1(a ) of the main text can be written as @xmath182 where @xmath183 with @xmath184 being the charge on the series capacitor @xmath185 , @xmath186 while the other symbols have been defined earlier .
we follow the same procedure as described in section [ sec : classicaleqns ] and first define the steady state solutions of @xmath131 and @xmath187 with @xmath188 as @xmath189 and @xmath190 respectively .
next we perform a perturbative expansion around the steady state solution in order to find the response due to the signal @xmath143 by considering [ eq : deltaandq ] @xmath191 where @xmath189 is defined in eq . with the assumption @xmath146 and @xmath192 . substituting @xmath131 and @xmath187 into eqs . and and retaining terms only up to first order in @xmath147 one obtains the following differential equations @xmath193 as before
, we now take the fourier transform of the above equations , use the rotating wave approximation and move into a reference frequency frame centered at @xmath6 . from eq .
we can solve for @xmath194 $ ] in terms of @xmath152 $ ] , and we get @xmath195= \dfrac{1}{(-i\omega + \kappa/2)}\left ( \dfrac{i}{2\omega_d}a_s[\omega ] - \dfrac{\kappa_0}{2}\delta_s[\omega ] \right).\ ] ] substituting @xmath194 $ ] in the fourier transform of eq . , we get @xmath196+\lambda e^{i\theta}\delta_s^\dagger[-\omega]=e^{-i\theta}\dfrac{ia_s[\omega]}{2\omega_p}\left[1+\dfrac{i\omega}{(-i\omega+\kappa/2 ) } \right].\ ] ] defining @xmath157=e^{-i\theta}\delta_s[\omega ] , \
\hat{a}_2^\dagger[-\omega]=e^{i\theta}\delta_s^\dagger[-\omega]$ ] and @xmath159 $ ] as the rhs of eq .
, we get , @xmath197+\lambda \hat{a}_2^\dagger[-\omega]=\hat{a}_\text{in}[\omega].\ ] ] comparing eq . and its hermitian conjugate with eq .
( 2 ) in the main text , and substituting for @xmath139 and @xmath198 , we obtain , [ eq : deltakappa1 ] @xmath55&=\dfrac{\kappa_0 \kappa}{4}\left ( \dfrac{\omega}{\omega^2 + \kappa^2/4}\right ) = \dfrac{1}{2 \alpha c_p}\left(\dfrac{\omega}{\omega^2 + r^2/\alpha^2}\right ) , \\ { \kappa}_{1(2)}[\omega]&=\dfrac{\kappa_0 \kappa}{4}\left ( \dfrac{\kappa}{\omega^2 + \kappa^2/4}\right)=\dfrac{r}{\alpha^2 c_p}\left(\dfrac{1}{\omega^2 + r^2/\alpha^2}\right),\end{aligned}\ ] ] which are identical to eqs . ( 9a ) and ( 9b ) in the main text obtained using the self energy approach .
we would again like to emphasize that the self energy approach is more general and a lot more powerful as it allows us to compute the parametric amplifier response for any linear auxiliary circuit as long as we can calculate the frequency dependent input admittance .
the method used in this section will not be amenable for more complex impedance transforming circuits .
in what follows , we will write the input admittance as [ eq : genzin ] @xmath199 \right)^{-1 } \equiv \left ( y_{\rm in } [ \omega ] \right)^{-1 } & = z_\text{in}[\omega ] \\ & = r_\text{in}[\omega]+ix_\text{in}[\omega]+i\alpha(\omega-\omega_\text{off}),\end{aligned}\ ] ] where @xmath200 $ ] and @xmath201 $ ] are the real and imaginary parts of the environmental impedance ( without the auxiliary circuit ) which we nominally assume to be @xmath99 . however , even small impedance mismatches in various components of the measurement chain lead to a frequency dependent complex input impedance .
the second imaginary term is due to the auxiliary resonator but it now allows for the more general case where the resonant frequency of the auxiliary circuit can be shifted by @xmath202 with respect to the pump frequency of the jpa . the susceptibility matrix ( eq . ( 2 ) of main text ) can be written as @xmath203)^{-1}= -i\begin{pmatrix } ( \omega-\omega_d')-\sigma_1[\omega ] & i\lambda \\ i\lambda & ( \omega+\omega_d ' ) -\sigma_2[\omega ] \end{pmatrix},\ ] ] with @xmath47=\delta_{1(2)}[\omega]-i{\kappa}_{1(2)}[\omega]/2 $ ] .
the effective photon potentials " @xmath204 $ ] and induced damping rates @xmath205 $ ] of the jpa are directly determined by the input admittance @xmath22 = 1/z_\text{in}[\omega]$ ] as @xmath206 & = -\text{im}\left[\dfrac{y_\text{in}[\omega]}{2c_p}\right ] , \ \
\delta_2[\omega ] = \text{im}\left[\dfrac{y_\text{in}[-\omega]}{2c_p}\right ] , \label{eq : deltadef } \\ \kappa_1[\omega ] & = \ \ \text{re}\left[\dfrac{y_\text{in}[\omega]}{c_p}\right ] , \
\ \kappa_2[\omega ] = \ \text{re}\left[\dfrac{y_\text{in}[-\omega]}{c_p}\right ] .
\label{eq : kappadef } \end{aligned}\ ] ] the reflection coefficient is given by @xmath207= 1-\kappa_1[\omega ] \chi_{11}[\omega]= 1-\kappa_1[\omega]\dfrac{-i(\omega+\omega_d'-\delta_2[\omega])+\kappa_2[\omega]/2}{(-i\tilde\omega[\omega]+\kappa_-[\omega]/2)(-i\tilde\omega[\omega]+\kappa_+[\omega]/2)},\ ] ] where [ eq : defs ] @xmath31 & = \omega-\delta[\omega ] , \\ \kappa_\pm[\omega ] & = \kappa[\omega]\pm2\sqrt{(\lambda^2-\omega_d'^2)+\epsilon[\omega]},\\ \epsilon[\omega ] & = \omega'_d\delta[\omega]-\delta[\omega]^2/4,\\ \delta[\omega ] & = ( \delta_2[\omega]-\delta_1[\omega])+i({\kappa}_1[\omega]-{\kappa}_2[\omega])/2,\end{aligned}\ ] ] with [ eq : defs1 ] @xmath208 & = \dfrac{\delta_1[\omega]+\delta_2[\omega]}{2 } = \text{im}\left[\dfrac{y_\text{in}[-\omega]-y_\text{in}[\omega]}{4c_p}\right ] , \\ \kappa[\omega ] & = \dfrac{\kappa_1[\omega]+\kappa_2[\omega]}{2 } \
= \text{re}\left[\dfrac{y_\text{in}[-\omega]+y_\text{in}[\omega]}{2c_p}\right ] .\end{aligned}\ ] ] we have introduced @xmath56 $ ] and @xmath209 $ ] to capture the effects of the symmetry of @xmath22 $ ] around @xmath33 .
the term @xmath210+{\kappa}_-[\omega]/2)$ ] in eq .
( [ eq : rw ] ) predominantly determines the gain and bandwidth of the jpa and that term depends on @xmath56 $ ] and @xmath209 $ ] given in eq . .
note that @xmath56 $ ] and @xmath209 $ ] control the impedance engineering effect i.e. they will be constants if the input impedance was a frequency independent @xmath99 .
( [ eq : defs1 ] ) implies that symmetric variations in @xmath211 $ ] around @xmath33 will leave @xmath56 $ ] unchanged while anti - symmetric variations in @xmath212 $ ] leave @xmath209 $ ] unchanged . as discussed in the main text , the last term in eq .
( [ eq : genzin]b ) which is due to the auxiliary circuit leads to @xmath54 = y_\text{in}[\omega]^*$ ] when @xmath213 . for small @xmath20 , this results in
@xmath214 $ ] being anti - symmetric about @xmath20 and hence the gain profile is modified significantly .
as mentioned earlier , standing waves resulting from the finite circulator cable length and impedance mismatch at the circulator , lead to a frequency dependence in @xmath200 $ ] and @xmath201 $ ] .
use of a circulator cable is practically unavoidable because one would like to protect the jpa from the magnetic field of the circulator by keeping it inside a magnetic shield and away from the circulator . as shown in fig .
[ fig : grip](b),(c ) , this introduces additional periodic oscillations in @xmath22 $ ] where the period is set by the length of the circulator cable and magnitude by the impedance mismatch .
this results in the ripple in the gain profile as shown in the red curve of fig .
[ fig : grip](a ) . ) profile shown in ( b ) and ( c ) . the periodic variations in @xmath215
are used to model the effect of the circulator cable and a small impedance mismatch at the circulator . note that the ripple in the gain is maximized ( red curves ) when @xmath212 $ ] is symmetric and @xmath211 $ ] is anti - symmetric about @xmath33 . by tuning the symmetry of @xmath212 $ ] and @xmath211
$ ] one can minimize the ripple in the gain ( blue curves ) .
the black dashed line in ( a ) is a reference gain curve with no circulator cable and hence no periodic variations in @xmath212 $ ] .
@xmath215 is plotted with @xmath216 for clarity.,scaledwidth=50.0% ] now if one could adjust the periodic pattern in @xmath22 $ ] in such a way that @xmath211 $ ] is symmetric while @xmath212 $ ] is anti - symmetric about @xmath33 , then the ripples in the gain profile can be minimized as shown in the blue curve in fig .
[ fig : grip](a ) .
the black dashed curve is the gain profile with no circulator cable shown for reference .
this adjustment can be done either by changing the circulator cable length or by slightly adjusting the pump frequency .
however , one can not completely eliminate these ripples because of the small contributions from @xmath32 $ ] in eq .
( [ eq : defs]b ) which actually gets enhanced when one performs the adjustment explained above .
[ fig : grip](b ) and [ fig : grip](c ) show the real and imaginary parts of @xmath22 $ ] ( with @xmath216 for clarity ) when the ripple is maximized ( red curves ) and minimized ( blue curves ) . in practice , the variations in @xmath22 $ ] are not simply sinusoidal but can have an overall frequency dependence which also limits the effectiveness of this trick . since we could nt tune the circulator cable length in - situ in our experiments , we changed the pump frequency in small steps to achieve the ripple cancellation . note that this leads to @xmath217 and in general affects the gain profile .
however , we checked from our theory that provided @xmath202 is small compared to the enhanced bandwidth , it does nt affect the gain and bandwidth significantly . in order to test the effect of the circulator cable length
, we performed different experimental runs with cable sections varying in length . in fig .
[ fig : gcomp](a ) the jpa was directly mounted ( with an sma male - male barrel ) on the circulator showing 830 mhz bandwidth while in fig .
[ fig : gcomp](b ) we obtained 600 mhz bandwidth with an 18 inch cable section which is three times longer than the cable used to obtain the data in the main text . the large variation in bandwidth can be attributed to different quality factors set by different re@xmath218 $ ] in these two cases .
the effectiveness of ripple minimization also varied between the two samples . even though the data in fig .
[ fig : gcomp](a ) shows more bandwidth than the one presented in the main text , the amplifier was significantly noisier .
we believe that this is due to the quality factor being too low in this configuration , leading to the amplifier becoming unstable . in practice , we do nt know the exact value of @xmath219 and we tune @xmath6 ( around the expected @xmath219 ) and @xmath4 ( by flux ) till we get @xmath220 and the largest bandwidth while ensuring minimal ripples in the gain profile .
the ripple minimization procedure is robust and we were able to achieve this optimization for several different configurations with different @xmath22 $ ] ) as shown in fig .
[ fig : gcomp ] . ]
the gain profile in ( b ) has less ripples than that in ( a ) .
we suspect that this is due to the input impedance of the circulator not being symmetric about @xmath33 which makes the ripple minimization procedure less effective .
even though the longer cable in ( b ) adds more ripples to @xmath215 , it is probably more uniform around @xmath33 and hence better suited for ripple minimization.,scaledwidth=50.0% ]
in order to compute @xmath75 , we reproduce eqs .
( 9a , b ) from the main text below : [ eq : deltakappa - s ] @xmath55&=\dfrac{1}{2 \alpha c_p}\left(\dfrac{\omega}{\omega^2 + r^2/\alpha^2}\right ) , \\ { \kappa}_{1(2)}[\omega]&=\dfrac{r}{\alpha^2 c_p}\left(\dfrac{1}{\omega^2 + r^2/\alpha^2}\right).\end{aligned}\ ] ] here , @xmath221 is the tuning parameter which can be used to cancel the @xmath222 dependence of @xmath49 $ ] for enhancing bandwidth . at the maximum gain point with @xmath223 ,
the optimal value of @xmath224 is obtained by setting the coefficient of @xmath222 in @xmath49 $ ] to zero and we get @xmath225
the input impedance seen looking towards a load @xmath226 through a transmission line section of impedance @xmath227 and length ( in units of wavenumber ) @xmath228 for a design frequency @xmath229 is given by @xcite @xmath230=z_0 \dfrac{z_l+i z_0 \tan\left(2\pi \bar{\nu } ( 1+\omega/\omega_0)\right)}{z_0+i z_l \tan(2\pi \bar{\nu } ( 1+\omega/\omega_0))}.\ ] ] eq .
shows that any @xmath0 @xmath231 or @xmath68 @xmath232 transformer provides a positive linear slope in the imaginary component of @xmath233 close to its design frequency . for small values of @xmath20
, we can model this slope with a series lc circuit with characteristic impedance @xmath224 . in case of a @xmath0 transformer of impedance @xmath234 ,
the characteristic impedance is given by @xmath235 similarly , for a @xmath68 transformer of impedance @xmath236 , @xmath237 note that the @xmath68 section also transforms the real part of the impedance .
however the frequency dependence in the real part over the frequency range of interest is quite small and can be ignored for computing the bandwidth .
since we used a @xmath68 transformer followed by a @xmath0 transformer , the characteristic impedance of the effective auxiliary resonator can be written as @xmath238 \dfrac{\pi}{4},\ ] ] where we have substituted @xmath239 , the characteristic impedance of the measurement chain .
we designed our impedance transformer to have @xmath240 and @xmath241 resulting in @xmath242 with @xmath69 .
5ifxundefined [ 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 \doibase http://dx.doi.org/10.1063/1.3497008 [ * * , ( ) ] link:\doibase 10.1126/science.1084647 [ * * , ( ) ] link:\doibase 10.1103/physreva.83.033803 [ * * , ( ) ] link:\doibase 10.1103/physreve.49.1198 [ * * , ( ) ] @noop * * ( ) @noop _ _ , ed .
( , , ) | we present an impedance engineered josephson parametric amplifier capable of providing bandwidth beyond the traditional gain - bandwidth product .
we achieve this by introducing a positive linear slope in the imaginary component of the input impedance seen by the josephson oscillator using a @xmath0 transformer .
our theoretical model predicts an extremely flat gain profile with a bandwidth enhancement proportional to the square root of amplitude gain .
we experimentally demonstrate a nearly flat 20 db gain over a 640 mhz band , along with a mean 1-db compression point of -110 dbm and near quantum - limited noise .
the results are in good agreement with our theoretical model .
josephson parametric amplifiers ( jpas ) have become a crucial component of superconducting qubit @xcite measurement circuitry , enabling recent studies of quantum jumps @xcite , generation and detection of squeezed microwave field @xcite , quantum feedback @xcite , real - time tracking of qubit state evolution @xcite and quantum error detection @xcite .
although jpas based on josephson junctions embedded in a resonator @xcite regularly achieve 20 db power gain and quantum - limited noise , typical bandwidth is restricted to 10 - 50 mhz @xcite , making them suitable for single qubit measurements only . the rapid progress towards multi - qubit architectures @xcite for fault - tolerant quantum computing @xcite demands an amplifier with much larger bandwidth to enable simultaneous readout of multiple qubits with minimal resources .
there have been several attempts in this direction in recent years .
one such attempt used a broadband impedance transformer @xcite to lower the quality factor of a lumped - element josephson oscillator which is the main component of a parametric amplifier .
while the observed large bandwidth was qualitatively explained by a model consisting of a negative resistance @xcite coupled to a frequency dependent impedance , no clear prescription on the design principle was provided .
a different approach using josephson non - linear transmission lines @xcite was recently demonstrated @xcite with nearly 4 ghz of bandwidth .
however , this design requires fabrication of about 2000 nearly identical blocks of oscillator stages , demanding fairly sophisticated fabrication facilities .
multimode systems utilizing dissipative interactions have also been suggested theoretically as a route for enhancing bandwidth @xcite , but have not been realized experimentally . in this letter , we present a simple technique for enhancing the bandwidth of a jpa and beating the standard gain - bandwidth limit .
it involves engineering the _ imaginary _ part of the environmental impedance : in particular , we introduce a positive linear slope in the imaginary component of the impedance shunting the jpa , while keeping the real part unchanged at the pump frequency .
our design uses a combination of a @xmath1 and a @xmath2 impedance transformers which are significantly easier to fabricate than a broadband impedance transformer @xcite .
our theoretical model explains why the imaginary part of the impedance plays a crucial role in determining the amplifier bandwidth , and predicts an extremely flat gain profile with a bandwidth beyond the standard gain - bandwidth product .
we experimentally demonstrate 640 mhz of nearly flat 3 db bandwidth with 20 db peak gain at a center frequency of @xmath3 ghz .
the observed near - quantum - limited noise performance with a mean 1-db compression point of -110 dbm over the whole bandwidth makes it a promising candidate for use in multi - qubit architecture .
we model the jpa as a non - linear resonator with a plasma frequency @xmath4 coupled to a frequency dependent environment specified by an input impedance @xmath5 $ ] .
we consider the directly pumped ( at @xmath6 ) case @xcite but our approach is equally applicable to flux pumping @xcite as well .
similar to ref . , we will be interested in using the jpa in a phase - preserving mode of operation .
it is thus useful to treat the signal and idler degrees of freedom of the single physical cavity mode as two separate modes @xcite . at the mean - field level , the system is then analogous to a detuned non - degenerate parametric amplifier ( ndpa ) @xcite . in an interaction picture at the drive frequency ,
the effective hamiltonian in the rotating wave approximation is given by @xmath7 where @xmath8 is the strength of effective parametric pumping .
here , @xmath9 and @xmath10 are canonical lowering operators describing the signal and idler frequency components of the same physical cavity mode ; they can be treated as independent as long as the signal frequency is not equal to the pump frequency @xcite .
also , @xmath11 where @xmath12 is the detuning of the pump from the oscillator resonant frequency .
the effective detuning @xmath13 depends on the pumping strength @xmath8 for parametric amplifiers based on linear driving of a quartic kerr non - linearity @xcite ; in contrast , if one flux - pumps , one has @xmath14 ( and could thus easily tune to have @xmath15 ) . solving the equations of motion@xcite yields
the system susceptibility matrix @xmath16 $ ] that relates the intracavity fields @xmath17=(\hat{a}_1,\hat{a}_2^\dagger)$ ] to the input fields @xmath18 $ ] as @xmath17=\chi[\omega]\vec{a}_{\rm in}[\omega]$ ] , where @xmath19)^{-1}= -i\begin{pmatrix } ( \omega-\omega_{d}')-\sigma_{1}[\omega ] & i\lambda \\ i\lambda & ( \omega+\omega_{d}')-\sigma_{2}[\omega ] \end{pmatrix}.\ ] ] here @xmath20 is the signal frequency detuning from @xmath6 while @xmath21 $ ] is the self energy and is related to the input admittance ( @xmath22 = 1/z_\text{in}[\omega]$ ] ) seen by the jpa as @xcite [ eq : sigmayin ] @xmath23&=\delta_{1}[\omega]-i{\kappa}_{1}[\omega]/2 \ = - \frac { i } { 2 c_p } y_{\rm in}^ * [ \omega ] , \\ \sigma_{2}[\omega]&=\delta_{2}[\omega]-i{\kappa}_{2}[\omega]/2 \ = - \frac { i } { 2 c_p } y_{\rm in } [ -\omega],\end{aligned}\ ] ] where @xmath24 $ ] and @xmath25 $ ] are related to the real and imaginary parts of @xmath26 $ ] at the signal ( @xmath27 ) and idler ( @xmath28 ) frequencies respectively .
the photon number gain of the signal mode @xmath29 $ ] is given by [ eq : gain ] @xmath30&=|{r}[\omega]|^2=|1-\kappa_1\chi_{11}[\omega]|^2 , \\
\chi_{11}[\omega]&= \dfrac{-i(\omega+\omega_d'-\delta_2[\omega])+\kappa_2[\omega]/2}{(-i\tilde\omega[\omega]+\kappa_-[\omega]/2)(-i\tilde\omega[\omega]+\kappa_+[\omega]/2)},\end{aligned}\ ] ] with [ eq : omegakappa ] @xmath31 & = \omega-(\delta_1[\omega]+\delta_2[\omega])/2 , \\
\kappa_\pm[\omega ] & = ( \kappa_1[\omega]+\kappa_2[\omega])/2\pm2\sqrt{(\lambda^2-\omega_d'^2)+\epsilon[\omega]},\end{aligned}\ ] ] here @xmath32 $ ] is a small quantity @xcite which depends on the symmetry of @xmath22 $ ] about @xmath33 .
both the peak gain @xmath34 $ ] and bandwidth ( fwhm of @xmath29 $ ] ) are controlled by the slow " effective damping rate @xmath35 $ ] . for the simplest case of the jpa coupled to a frequency - independent environment with characteristic impedance @xmath36 , @xmath22 = 1/r$ ] , leading to @xmath24=\kappa_0 $ ] ,
@xmath25=0 $ ] and @xmath32=0 $ ] .
here @xmath37 is the oscillator linewidth , with @xmath38 being the characteristic impedance of the oscillator ( fig .
[ fig : paramplc](a ) ) . for a given @xmath39 ,
there is an optimal @xmath8 which maximizes @xmath34 $ ] and is given by @xcite @xmath40 where @xmath41 , while the gain and bandwidth ( in the large gain limit ) are given by @xmath42 \approx \dfrac{\mathcal{g}_\text{max}}{1 + ( \omega/\gamma_\text{bw})^2 } , \ \
2\gamma_{\text{bw } } \approx \kappa_0 \left ( \dfrac { 1 } { \mathcal{g}_\text{max } } \right ) ^{1/2}.\ ] ] eq .
gives the standard gain - bandwidth product @xcite where the bandwidth falls off as the square - root of power gain .
for a given @xmath43 , the simplest way to enhance bandwidth @xcite is to increase @xmath44 by reducing the shunt impedance @xmath36 .
however , lowering the quality factor @xmath45 requires one to pump harder and at some point the strong pump makes the amplifier unstable and noisy due to higher order non - linearities @xcite which are ignored in our linearized hamiltonian ( eq . ) .
we consider a different approach where we leave the shunting impedance at the pump frequency ( @xmath33 ) undisturbed so that the same optimal value of @xmath8 provides @xmath43 as before . instead of increasing @xmath44 , we modify the coherent cavity dynamics to make the amplifier less sensitive to the effects of detuning the signal from the pump ( i.e. non - zero @xmath20 ) ; as such , strong amplification will persist even away from @xmath46 .
this can be achieved by introducing a frequency dependence in @xmath22 $ ] with the help of an auxiliary system such that @xmath47 $ ] exactly cancels the explicit @xmath20 dependence of @xmath16 $ ] in eq . .
the result would be a modified susceptibility @xmath48 $ ] and gain @xmath49 $ ] that are frequency independent .
a simple step in this direction is to use an additional linear resonant mode as the auxiliary system . here
, we introduce a series lc resonator as shown in fig .
[ fig : paramplc](a ) .
we further make the optimal choice of having the auxiliary lc resonance frequency match
the drive frequency i.e. @xmath50 . for small @xmath20 ,
the input impedance can now be written as @xmath51=r+i\alpha \omega,\ ] ] where @xmath52 is the slope of the imaginary component and @xmath53 is the characteristic impedance of the auxiliary resonator .
this choice of @xmath5 $ ] leads to @xmath54 = y_\text{in}[\omega]^*$ ] and hence [ eq : deltakappa ] @xmath55&=\dfrac{1}{2 \alpha c_p}\left(\dfrac{\omega}{\omega^2 + r^2/\alpha^2}\right ) , \\ { \kappa}_{1(2)}[\omega]&=\dfrac{r}{\alpha^2 c_p}\left(\dfrac{1}{\omega^2 + r^2/\alpha^2}\right),\end{aligned}\ ] ] with @xmath32=0 $ ] .
for small @xmath20 , @xmath56 \sim \alpha \omega/(2r^2c_p)$ ] and can be used to cancel the intrinsic @xmath20 dependence in @xmath16 $ ] to leading order : one just tunes @xmath57 by adjusting @xmath58 .
we see that it is @xmath57 which controls the frequency - dependence of the induced potential " @xmath25 $ ] that is crucial to our scheme while @xmath59 is constant for small @xmath20 .
the modified gain can now be computed by using eqs .
, and .
the leading - order quadratic dependence of @xmath49 $ ] in our modified jpa can be canceled exactly for an optimal choice of @xmath58 , given by @xmath60 where @xmath61 is a prefactor of order unity @xcite .
the zero frequency gain is still given by @xmath43 since @xmath62 $ ] is left unmodified in our scheme ( i.e. @xmath63 = \kappa_0 , \delta_{1(2)}[0]=0 $ ] ) , while the gain @xmath49 $ ] roll - off is much slower near @xmath33 . in the large gain limit ,
the modified gain profile and corresponding bandwidth are given by @xmath64=\dfrac{\mathcal{g}_\text{max}}{1 + ( \omega /\gamma_\text{bw})^4 } , \ \
2\tilde\gamma_{\text{bw } } \approx \kappa_0 \left ( \dfrac { 1 } { \mathcal{g}_\text{max } } \right ) ^{1/4},\ ] ] with a bandwidth enhancement factor @xmath65 . for finite gain
, @xmath66 is somewhat larger than this asymptotic value .
fig .
[ fig : thgain ] shows how the frequency - dependence of the gain is modified for different choices of @xmath58 ; the flattest profile is obtained for @xmath67 . for a value of @xmath58 larger than the optimal value
, the gain profile transforms to a double peaked shape ( green curve of fig . [
fig : thgain ] ) where the two peaks correspond to the normal modes of the two coupled oscillators .
the optimal value for @xmath58 ensures that the contribution of the two modes overlap in just the right way to give a flat gain response near @xmath33 .
having established the basic theory underlying our design , we now discuss its experimental implementation .
this was done in two steps ( fig .
[ fig : paramplc](b ) ) .
first , we used a @xmath68 section ( at @xmath6 ) to transform the standard @xmath69 environment to @xmath70 re@xmath71 = 30 \ \omega$ ] .
this not only provides some conventional bandwidth enhancement but also keeps the value of @xmath72 ( for @xmath73 ) in a range which is easy to implement .
second , we used a @xmath0 section ( at @xmath6 ) to independently set the slope of im@xmath74 $ ] , and hence @xmath75 , while keeping @xmath36 unchanged .
both transformer sections were made on a single rf circuit board in the co - planar waveguide ( cpw ) geometry for a design frequency of 6 ghz .
our jpa device was made using standard e - beam lithography and double - angle evaporation to fabricate @xcite the squid with a total critical current @xmath76 @xmath77a and the capacitor @xmath78 pf .
the squid geometry allowed the tuning of @xmath79 in the range 5 - 7.25 ghz with the help of a superconducting coil .
the device was connected to the impedance transformer using wire - bonds and measured at 20 mk temperature inside a dilution refrigerator .
we used standard cryogenic microwave reflectometry setup to measure the devices @xcite .
a magnetically shielded cryogenic circulator connected to the jpa was used to separate the incident and reflected signal and both were housed inside a magnetic and radiation shielded can .
fig .
[ fig : gain ] shows the gain profile obtained using our impedance transformer design .
the best performance was obtained at @xmath80 ghz , close to the design frequency of 6 ghz , with a -3 db bandwidth of 640 mhz ( vertical dashed line ) and 20 db gain at the pump frequency .
we measured the input impedance of a nominally identical transformer at room temperature ( dashed lines of fig .
[ fig : ripple ] ( a ) ) and obtained @xmath81 and @xmath82 .
we also extracted @xmath36 from cryogenic measurements using eq . and
obtained @xmath83 , consistent with our room temperature measurements .
using @xmath83 , the predicted value of bandwidth using eqs .
( [ eq : gain]a ) and ( [ eq : gain]b ) is @xmath84 mhz for @xmath85 which is in excellent agreement with our measured bandwidth . without the auxiliary resonator the same device ( with @xmath83 ) would have had 180 mhz of bandwidth , giving us a bandwidth improvement of about a factor of 3.6 .
the total bandwidth enhancement when compared to the standard case of @xmath69 is about 5.7 .
the noise temperature of the jpa was measured by first calibrating the system noise temperature using a shot noise tunnel junction @xcite and then switching to the jpa using a cryogenic microwave switch in the same experimental run . the jpa s noise temperature ( fig .
[ fig : gain ] , red curve ) was then extracted by measuring the signal - to - noise ratio improvement obtained using the jpa @xcite .
the performance was near quantum - limited ( @xmath86 ) in the full 640 mhz band .
the 1-db compression point is shown in the bottom panel of fig .
[ fig : gain ] and has a mean value of about -110 dbm throughout the band . in practice , the ideal input impedance required ( eq . and dashed lines in fig . [
fig : ripple]a ) is extremely difficult to implement . a more realistic @xmath5 $ ]
is shown in fig .
[ fig : ripple]a ( solid lines ) where the periodic deviations are primarily due to standing waves in the cable section between the impedance transformer and the circulator ( fig .
[ fig : paramplc](b ) , henceforth called the circulator cable ) , and usually lead to ripples in the gain profile . however , we could minimize these ripples @xcite by a slight adjustment of the pump frequency . fig .
[ fig : ripple](b ) shows the variation of gain profile as pump frequency is varied while leaving other bias conditions unchanged .
the arrow corresponds to the data in fig .
[ fig : gain ] where ripples are minimized .
this ripple minimization effect can be understood using eqs .
( [ eq : omegakappa]a ) and ( [ eq : omegakappa]b ) which show us that variations in @xmath87 $ ] which are symmetric about @xmath33 do not affect the gain . by positioning the pump frequency ( @xmath6 ) at the right value with respect to @xmath22 $ ] , one can minimize the ripples while not affecting the optimal choice of @xmath88 significantly .
the two larger humps at the band edge are probably due to significant deviation of the impedance seen by the jpa from our simplified model .
we would like to emphasize that the ripple minimization procedure is robust and we were able to achieve this optimization for several different configurations @xcite with different @xmath22 $ ] .
our scheme thus benefits from having introduced an extra ( linear ) resonant mode in a judicious manner .
the practical utility of having two non - linear modes was recently discussed in ref .
, though that design was still subject to the standard gain - bandwidth limitation .
it is also interesting to compare our use of two modes to the traveling wave parametric amplifier ( twpa ) @xcite which can be thought of as consisting of a large number of coupled non - linear modes ; here , we show that just a single extra linear mode can provide significant advantages .
further , the phase matching achieved by dispersion engineering the twpa @xcite can be compared to our introduction of an imaginary impedance at the signal and idler frequencies to enhance the bandwidth . in conclusion , we have designed a simple impedance transformer which introduces a positive linear slope in the imaginary part of the impedance seen by the jpa . our theory explains the significance of the imaginary part of input impedance in enhancing the bandwidth .
we demonstrate a device with a nearly flat 20 db gain over a 3 db bandwidth of 640 mhz while maintaining near quantum - limited noise performance , in good agreement with theoretical predictions .
these characteristics along with a 1-db compression point of -110 dbm clearly indicates its capability for simultaneous multi - qubit readout @xcite .
our technique is readily adaptable to flux pumped jpas @xcite , amplifiers based on the josephson parametric converter @xcite , and flux @xcite and charge @xcite sensors based on parametric amplifiers .
it could also be useful as a source of broadband squeezed microwaves @xcite for experiments in microwave quantum optics .
this work was supported by the department of atomic energy of government of india .
rv acknowledges funding from the department of science and technology , india via the ramanujan fellowship .
we also acknowledge the tifr nanofabrication facility .
ac acknowledges support from the army research office under grant no .
w911nf-14 - 1 - 0078 .
36ifxundefined [ 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 * * , ( ) link:\doibase 10.1103/physrevlett.106.110502 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.106.220502 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevlett.109.240502 [ * * , ( ) ] link:\doibase 10.1126/science.1226897 [ * * , ( ) ] link:\doibase 10.1038/nature12539 [ * * ( ) , 10.1038/nature12539 ] link:\doibase 10.1038/nature13559 [ * *
( ) , 10.1038/nature13559 ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevb.83.134501 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.2773988 [ * * , ( ) ] @noop
* * , ( ) \doibase http://dx.doi.org/10.1063/1.4821136 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.4764940 [ * * , ( ) ] link:\doibase 10.1103/physreva.86.032324 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.4886408 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevb.87.144301 [ * * , ( ) ]
link:\doibase 10.1103/physrevlett.113.157001 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.4922348 [ * * , ( ) ] link:\doibase 10.1126/science.aaa8525 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.112.133904 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.2964182 [ * * , ( ) ] @noop link:\doibase 10.1103/revmodphys.82.1155 [ * * , ( ) ] link:\doibase 10.1103/physreva.83.033803 [ * * , ( ) ] link:\doibase 10.1103/physreve.49.1198 [ * * , ( ) ] link:\doibase 10.1103/physrevb.76.014524 [ * * , ( ) ] link:\doibase 10.1126/science.1084647 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.3497008 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevapplied.4.014018 [ * * , ( ) ] link:\doibase 10.1103/physreva.81.013814 [ * * , ( ) ] * supplementary material for broadband parametric amplification with impedance engineering : beyond the gain - bandwidth product " * + |
the terminology and notation used in this paper are standard and consistent with that used in @xcite .
therefore , in this section we only settle the notation and terminology that could vary among texts .
the vertex set @xmath23 of a graph @xmath24 is denoted by @xmath25 , its edge set by @xmath26 , its girth by @xmath27 , its adjacency matrix by @xmath28 and its diameter by @xmath29 ; when there is no place for confusion , we drop the symbol @xmath24 .
we often use the letter @xmath7 to denote the order of @xmath24 .
the identity matrix of order @xmath7 is denoted by @xmath30 , while by @xmath8 we denote the @xmath9 matrix whose entries are all 1 s . for a matrix @xmath6
the set formed by its @xmath31 distinct eigenvalues @xmath32 with respective multiplicities @xmath33 is called the _ spectrum _ of @xmath6 and is denoted by @xmath34^{m_0},\ldots,[\lambda_r]^{m_r}\}$ ] . the characteristic polynomial @xmath35 of @xmath6 is denoted by @xmath36 . for a graph @xmath24 , we often write @xmath37 rather than @xmath38 .
we denote the eigenspace of @xmath6 corresponding to the eigenvalue @xmath39 by @xmath40 .
we call a cycle of order @xmath7 an _ @xmath7-cycle _ and denote it by @xmath41 . if a graph @xmath24 is a union of @xmath42 vertex - disjoint cycles , we consider the multiset of their @xmath31 distinct lengths @xmath43 and respective multiplicities @xmath33 , and write that the _ cycle structure _ of @xmath24 is @xmath44^{m_0},[l_1]^{m_1}\ldots [ l_r]^{m_r}\}$ ] with @xmath45 and @xmath46 . the degree of a polynomial @xmath47 is denoted by @xmath48 .
as it is customary , we denote the real chebyshev polynomial of the second kind by @xmath49 @xcite .
recall that the polynomial @xmath49 , defined on @xmath50 $ ] , satisfies the following recurrence equations .
@xmath51 \end{cases}\ ] ] it is known that the moore bound , denoted by @xmath52 and defined below , represents both an upper bound on the order of a graph of maximum degree @xmath0 and diameter @xmath53 and a lower bound on the order of a graph of minimum degree @xmath0 and odd girth @xmath54 @xcite . @xmath55
non - trivial _ moore graphs _ ( graphs whose order equals the moore bound , with @xmath56 and @xmath57 ) exist only for @xmath58 ( or equivalently , for @xmath59 ) , in which case @xmath60 , 3 , 7 and possibly 57 @xcite . by virtue of the rarity of moore graphs , it is important to consider graphs which are somehow close to the ideal moore graphs .
graphs of maximum degree @xmath0 , diameter @xmath1 and order @xmath61 are called _
@xmath62-graphs _ , where the parameter @xmath3 is called _ defect_. graphs of minimum degree @xmath0 , odd girth @xmath2 and order @xmath63 are called _
@xmath64-graphs _ , where the parameter @xmath3 is called the _
excess_. graphs with defect or excess 1 were completely classified by bannai and ito @xcite ; for any degree @xmath65 , the only graphs of defect 1 are the cycles on @xmath66 vertices , while the only graphs of excess 1 are the _ cocktail party graphs _ ( the complement of @xmath67 copies of @xmath68 , with even @xmath0 ) .
however , for @xmath69 the story is quite different . for maximum degree 2 and
diameter @xmath70 the path of length @xmath71 is the only @xmath72-graph . for degree @xmath15 and diameter @xmath70
there are only 5 known graphs with defect 2 , all of which are shown in fig .
[ fig : defect2 ] . for degree 2
there is no graph with excess 2 , while for degree @xmath20 and girth 3 the complement of the cycle @xmath73 is the only graph with excess 2 . for degree @xmath15 and odd girth @xmath74
there are only 4 graphs with excess 2 known at present ( see fig .
[ fig : excess2 ] ) . for those familiar with the theory of voltage graphs ( see ( * ? ? ?
* chapter 2 ) ) , in fig .
[ fig : defect2 ] we present the @xmath75-graph , the @xmath76-graph and the @xmath77-graph as lifts of voltage graphs .
the @xmath78-graph takes voltages on the group @xmath79 , while the @xmath76-graph and the @xmath77-graph take voltages on the group @xmath80 . in all cases the undirected edges have voltage 0 and the directed edges have voltage 1 .
it is worth mentioning that we gave an alternative voltage graph construction of a graph when this construction was simpler than the selected drawing of the graph .
as principle failed for the @xmath78-graphs , we omitted their respective voltage graph representation .
-graph , ( c ) a voltage graph of the unique @xmath81-graph , ( d ) the unique @xmath81-graph , ( e ) a voltage graph of the unique @xmath82-graph , ( f ) the unique @xmath82-graph , ( g ) a voltage graph of the unique @xmath77 graph , and ( h ) the unique @xmath83-graph . ] -graphs , ( c ) the unique @xmath84-graph ( the robertson graph ) , and ( d ) the unique @xmath85-graph ( the mcgee graph ) . ]
it is not difficult to see that if @xmath86 and @xmath87 , a @xmath88-graph must be @xmath0-regular . similarly ,
if @xmath89 and @xmath87 , a @xmath64-graph must be @xmath0-regular .
henceforth we consider graphs with defect or excess 2 , and to avoid trivial cases , we only analyze graphs with degree @xmath22 and diameter @xmath90 for defect 2 , and graphs with degree @xmath22 and girth @xmath91 for excess 2 .
note that all these graphs must be regular . in a graph @xmath24 with defect 2 ,
if there are at least 2 paths of length at most @xmath29 from a vertex @xmath92 to a vertex @xmath93 , then we say that @xmath92 is a _ repeat _ of @xmath93 ( and _ vice versa _ ) . in this case we have two repeats ( not necessarily different ) for each vertex of @xmath24 .
then , we define the _ defect ( multi)graph _ of @xmath24 as the graph on @xmath25 , where two vertices are adjacent iff one is a repeat of the other .
then , the defect graph is a union of vertex - disjoint cycles of length at least 2 .
similarly , in a graph @xmath24 with excess 2 , we define the _ excess graph _ of @xmath24 as the graph on @xmath25 , where two vertices are adjacent iff they are at distance @xmath29 ( with @xmath94 ) .
therefore , the excess graph is a union of vertex - disjoint cycles of length at least 3 .
next we present the cycle structure of the defect or excess graphs of the known non - trivial graphs with defect or excess 2 .
cyclic structure of graphs of defect 2 : : for the mbius ladder on 8 vertices @xmath95 ^ 1\}$ ] , for the other @xmath96-graph @xmath97 ^ 2,[2]^1\}$ ] , for the unique @xmath81-graph @xmath98 ^ 4\}$ ] , for the unique @xmath82-graph @xmath99 ^ 2,[3]^1\}$ ] , and for the unique @xmath83-graph @xmath97 ^ 6,[2]^3\}$ ] .
cyclic structure of graphs of excess 2 : : for the only @xmath100-graphs ( depicted in fig .
[ fig : excess2 ] as ( a ) and ( b ) ) we have that ( a ) @xmath101 ^ 1,[3]^1\}$ ] and ( b ) @xmath95 ^ 1,[4]^1\}$ ] , for the unique @xmath84-graph ( the robertson graph ) @xmath97 ^ 1,[12]^1,[4]^1\}$ ] , and for the unique @xmath85-graph ( the mcgee graph ) @xmath102 ^ 6\}$ ] . for a graph @xmath24 of degree @xmath0 with adjacency matrix @xmath6 , we define the polynomials @xmath103 for @xmath104 : @xmath105 it is known that the entry @xmath106 counts the number of paths of length at most @xmath42 joining the vertices @xmath107 and @xmath108 in @xmath24 ; see @xcite .
regular graphs with defect @xmath3 and order @xmath7 satisfy the equation @xmath109 and regular graphs with excess @xmath3 and order @xmath7 satisfy the equation @xmath110 where @xmath8 is the @xmath9 matrix whose entries are all 1 s , and @xmath10 is a matrix with the row and column sums equal to @xmath3 . the matrix @xmath10 is called the _ defect _ or _ excess matrix _ accordingly . for moore graphs ,
the matrix @xmath10 is the null matrix and @xmath111 is their minimal polynomial . for graphs with defect or excess 1 ,
@xmath10 can be considered as the adjacency matrix of a matching with @xmath7 vertices @xcite . for a graph @xmath24 with defect or excess 2 ,
the matrix @xmath10 is the adjacency matrix of the defect graph ( respectively , of the excess graph ) . with a suitable labeling of @xmath24
, @xmath10 becomes a direct sum of matrices representing cycles @xmath112 of length @xmath113 ( respectively , @xmath114 ) .
@xmath115{6}\\ 1&0&0&\dots&1&0 \end{pmatrix}\ ] ] the previous point about the labelling of a graph @xmath24 is illustrated in fig .
[ fig:322graph ] , where a @xmath96-graph is labelled such that the defect matrix @xmath10 displays the aforementioned structure . -graph that produces the desired structure of the corresponding defect matrix @xmath10 . ] for graphs with defect or excess 2 , equation ( [ eq : repm ] ) has been studied for diameter @xmath58 @xcite , and equation ( [ eq : exm ] ) has been studied for girths 5 and 7 @xcite . if @xmath10 is the adjacency matrix of a cycle of order @xmath7 ( i.e. @xmath116 ) , then the solution graphs of equations ( [ eq : repm ] ) and ( [ eq : exm ] ) are called _ graphs with cyclic defect _ and _ graphs with cyclic excess _ , respectively . among all the known non - trivial graphs with defect or excess 2 , only one has cyclic defect , the mbius ladder on 8 vertices @xcite , and none has cyclic excess .
in this paper we focus on graphs with cyclic defect or excess .
basically , we deal with the following problems : [ probrep ] classify the graphs of degree @xmath20 , diameter @xmath117 and order @xmath7 such that @xmath118 .
[ probex ] classify the graphs of degree @xmath20 , odd girth @xmath119 and order @xmath7 such that @xmath120 . as problem [ probrep ]
was completely settled for @xmath58 in @xcite , from now on , we assume @xmath121 . the main result of the paper is the provision of the asymptotic upper bound of @xmath19 for the number of graphs of odd degree @xmath20 and cyclic defect or excess .
this bound turns out to be quite generous as our next results show .
there are no graphs of degree 3 or 7 , for diameter @xmath15 and cyclic defect or for odd girth @xmath17 and cyclic excess , nor any graphs of odd degree @xmath15 , girth @xmath21 or 9 and cyclic excess .
other non - existence outcomes are the non - existence of graphs of any degree @xmath15 , diameter 3 or 4 and cyclic defect ; and graphs of degree @xmath18 , girth 7 and cyclic excess . to obtain our results we rely on algebraic methods , specifically on connections between the polynomials @xmath103 and the classical chebyshev polynomials of the second kind @xcite , on eigenvalue techniques , and on elements of algebraic number theory .
the rest of this paper is structured as follows . in section [ sec : cycles ] we provide some old and new combinatorial conditions for the existence of graphs with cyclic defect .
in section [ sec : algcond ] we present several algebraic approaches to analyze graphs with cyclic defect or excess , while section [ sec : nonexres ] presents the main results of the paper .
finally , section [ sec : conclusion ] summarizes our results and gives some concluding remarks .
next we present some results about @xmath122-graphs .
we denote by @xmath123 the graph which is the union of three independent paths of length @xmath71 with common endvertices .
* lemma 2 ) let @xmath93 be a vertex of a @xmath122-graph @xmath24 .
then either : 1 .
@xmath93 is a branch vertex of a @xmath123 and every cycle of length at most @xmath66 in @xmath24 containing @xmath93 is contained in this @xmath123 ; or 2 .
@xmath93 is contained in one cycle of length @xmath124 and no other cycle of length at most @xmath66 ; or 3 .
@xmath93 is contained in exactly two cycles of length @xmath66 and no other cycle of length at most @xmath66 .
[ prop : cycles ] [ cor:2dcycles ] let @xmath24 be a @xmath122-graph with cyclic defect
. then every vertex lies in exactly 2 cycles of length @xmath66 .
[ ddivn ] the order @xmath7 of a @xmath122-graph with cyclic defect is a multiple of @xmath71 . by corollary [ cor:2dcycles ] ,
the number of @xmath66-cycles in a @xmath122-graph with cyclic defect is @xmath125 , and thus , the result follows .
[ congruences ] the allowed degrees for a @xmath122-graph with cyclic defect are restricted to some congruence classes modulo @xmath71 .
when @xmath71 is even , @xmath0 is odd .
when @xmath71 is a power of an odd prime , @xmath126 is a multiple of @xmath71 .
when @xmath127 is a power of 2 , @xmath126 is a multiple of @xmath128 . if @xmath129 , then @xmath130 . as @xmath131
, it follows that @xmath132 , which implies @xmath133 , and thus , @xmath134 .
suppose that @xmath71 is a power of a prime @xmath135 .
. then , @xmath137 and @xmath138 , which is incompatible with @xmath139 .
suppose @xmath140 . by the little fermat theorem @xcite
, we have that @xmath141 , and thus , that @xmath142
. therefore , @xmath143 .
also , since @xmath7 is a multiple of @xmath71 , it follows that @xmath144 .
it remains to see what happens when @xmath145 for @xmath146 with @xmath147 , that is , @xmath148 with @xmath149 .
as @xmath148 , we have that @xmath150 .
therefore , it follows that @xmath151 .
thus , to have @xmath152 it is necessary that @xmath153 if @xmath135 is odd and @xmath154 if @xmath155 .
this completes the proof of the corollary .
we start this section by giving some known results .
if @xmath10 is the adjacency matrix of the @xmath7-cycle then its characteristic polynomial @xmath156 satisfies the following @xmath157 where @xmath158 is a monic polynomial of degree @xmath159 if @xmath7 is even and @xmath160 if @xmath7 is odd .
recall that @xmath161 , where @xmath162 denotes the _ @xmath163-th cyclotomic polynomial _ , where @xmath164 denotes all @xmath163th primitive roots of unity , and @xmath165 denotes the _
euler s totient function _
, that is , the function giving the number of positive integers @xmath166 and relatively prime to @xmath163 . ] .
the cyclotomic polynomial @xmath162 is an integer polynomial , irreducible over the field @xmath167 $ ] of polynomials with rational coefficients , and _ self - reciprocal _ ( that is , @xmath168 ) .
a consequence of @xmath162 being irreducible over @xmath167 $ ] and self - reciprocal is that the degree of @xmath162 is even for @xmath169 . using the previous facts on cyclotomic polynomials , we obtain the following factorization of @xmath170 : @xmath171 , where @xmath172 is an integer polynomial of degree @xmath173 satisfying @xmath174 . also , @xmath172 is irreducible over @xmath167 $ ] . in particular
, we have that @xmath175 , @xmath176 , @xmath177 , @xmath178 , @xmath179 , @xmath180 , @xmath181 , @xmath182 . more concretely , @xmath183 ^ 1,[2\cos{(\frac{2\pi}{n}\times 1)}]^{2},\ldots , [ 2\cos{(\frac{2\pi}{n}\times \frac{n-2}{2})}]^{2},[-2]^1\ } & \text{if $ n$ is even } \\
\{[2]^1,[2\cos{(\frac{2\pi}{n}\times 1)}]^{2},\ldots , [ 2\cos{(\frac{2\pi}{n}\times \frac{n-1}{2})}]^{2}\ } & \text{if $ n$ is odd } \end{cases}\ ] ] it is also very well known that @xmath184 ^ 1,[0]^{n-1}\}$ ] .
considering equations [ eq : repm ] and [ eq : exm ] , we obtain that the eigenspace @xmath185 equals both the eigenspace @xmath186 and the eigenspace @xmath187 .
furthermore , for each eigenvalue @xmath39 @xmath188 of @xmath6 , we have that @xmath189 is an eigenvalue @xmath190 @xmath191 of @xmath10 . in this case , we say that the eigenvalue @xmath39 is paired with the eigenvalue @xmath190 .
therefore , for each eigenvalue @xmath190 @xmath192 of @xmath10 , the eigenspace @xmath193 contains the eigenspace of the eigenvalue of @xmath6 paired with @xmath190 .
[ prop : integerrootdef ] let @xmath6 be the adjacency matrix of a @xmath122-graph of order @xmath7 . if @xmath7 is even , then @xmath6 has a simple eigenvalue @xmath39 such that @xmath39 is an integer root of the polynomial @xmath194 .
consider equations ( [ eq : repm ] ) and ( [ eq : bspec ] ) .
if @xmath7 is even , @xmath195 is a simple eigenvalue of @xmath10 , and the eigenspace of @xmath195 is spanned by the vector @xmath196 .
let @xmath39 be the simple eigenvalue of @xmath6 which is a root of @xmath194 .
then , @xmath93 is also an eigenvector of @xmath6 , implying that @xmath39 must be integer .
let @xmath24 be a graph with cyclic defect .
if we substitute @xmath197 into @xmath198 , we obtain a polynomial @xmath199 of degree @xmath200 such that @xmath201 of its roots are eigenvalues of @xmath6 , and thus , @xmath202 for each eigenvector @xmath93 of @xmath6 orthogonal to the all-1 vector @xmath203 .
setting @xmath204 we have @xmath205 [ elemdiv1 ] for each polynomial @xmath206 , where @xmath207 and @xmath208 , the kernel of @xmath209 , denoted by @xmath210 , is formed by the direct sum of the eigenspaces associated with the roots of @xmath206 , and thus , @xmath210 is a @xmath165-dimensional space on @xmath167 $ ] .
since @xmath6 commutes with @xmath10 , we have that @xmath210 is stable under the multiplication by @xmath6 .
furthermore , as @xmath211 is null on @xmath210 , it follows that @xmath212 is null on @xmath210 and that @xmath213 is @xmath165-dimensional on @xmath167 $ ] .
consider a factor @xmath214 of @xmath215 .
the kernel of @xmath216 is stable under the multiplication by @xmath10 , since @xmath211 is null on @xmath210 .
thus , its dimension on @xmath167 $ ] is either 0 , or @xmath173 or @xmath165 .
hence , corresponding to the factor @xmath206 of the minimal polynomial of @xmath10 , the polynomial @xmath215 has either 2 factors of degree @xmath173 or one factor of degree @xmath165 . by using observation [ elemdiv1 ]
, we obtain our first simple necessary condition on the existence of graphs with cyclic defect .
[ prop : irr ] for @xmath121 and @xmath208 such that @xmath207 , if there is a @xmath122-graph with cyclic defect , then @xmath215 must be reducible over @xmath167 $ ] .
recall that @xmath217 . if @xmath215 is irreducible over @xmath167 $ ] , then all its roots must be eigenvalues of @xmath6 .
however , by observation [ elemdiv1 ] , only @xmath165 roots of @xmath215 can be eigenvalues of @xmath6 , a contradiction for @xmath121 .
note that @xmath218 iff @xmath219 , and that @xmath219 iff @xmath220 .
thus , we have the following useful corollary . [ cor:3,4,6 ]
let @xmath7 be the order of a graph with cyclic defect and diameter @xmath221 .
then , 1 .
if @xmath222 then @xmath223 must be reducible over @xmath167 $ ] .
if @xmath224 then @xmath225 must be reducible over @xmath167 $ ] .
3 . if @xmath226 then @xmath227 must be reducible over @xmath167 $ ] . knowing that @xmath175 , @xmath176 , and @xmath177 , the result follows from proposition [ prop : irr ] . for @xmath224 we can even prove a result slightly stronger than the one of corollary [ cor:3,4,6 ] . note that if @xmath224 then @xmath228 . as @xmath224
, @xmath229 is an eigenvalue of @xmath10 with multiplicity 2 .
the vectors @xmath230 and @xmath231 form a basis of @xmath232 . as @xmath6 and @xmath10 commute , @xmath233 and @xmath234 .
therefore , we have that @xmath235 for some @xmath236 define a matrix @xmath237 , called the _ restriction of @xmath6 on @xmath238 _ , as @xmath239 .
note that the characteristic polynomial @xmath240 of @xmath237 is the polynomial having as roots the two eigenvalues of @xmath6 paired with the eigenvalue 0 of @xmath10 .
let us consider @xmath241 , where @xmath242 is the all-1 vector .
all components of this sum are even .
thus , since all entries of @xmath6 are integers , @xmath243 has only even components .
consequently , @xmath244 and @xmath245 are even . as @xmath6 is symmetric , @xmath246 ( recall that if @xmath247 and @xmath248 are matrices then @xmath249 ) .
then , it follows that @xmath250 and that @xmath251 , since @xmath252 , @xmath253 and @xmath254 .
thus , @xmath255 and @xmath256 . in this way
, we have obtained the following proposition .
[ parity ] let @xmath6 be the adjacency matrix of a graph with cyclic defect . if @xmath257 then the restriction of @xmath6 on the kernel of @xmath10 has an even trace .
[ cor : trmdef ] for @xmath58 the characteristic polynomial of the restriction of @xmath6 on the kernel of @xmath10 ( i.e. @xmath259 ) must be reducible over @xmath167 $ ] .
@xmath258 the previous results on graphs with cyclic defect can be readily extended to cover graphs with cyclic excess .
therefore , we limit ourselves to give the results . [ prop : integerrootex ] let @xmath6 be the adjacency matrix of a @xmath260-graph of order @xmath7 .
if @xmath7 is even , there is a simple eigenvalue @xmath39 of @xmath6 such that @xmath39 is an integer root of the polynomial @xmath261 .
@xmath258 let @xmath24 be a graph with cyclic excess .
substituting @xmath262 into @xmath198 , we obtain a polynomial @xmath263 of degree @xmath264 such that @xmath265 for each vector @xmath93 orthogonal to the all-1 vector . setting @xmath266
we have that @xmath267 [ prop : irrbis ] for @xmath268 and @xmath208 such that @xmath207 , if there is a @xmath260-graph with cyclic excess and order @xmath7 then @xmath269 must be reducible over @xmath167 $ ] .
[ cor:3,4,6bis ] let @xmath7 be the order of a graph with cyclic excess . then , 1 .
if @xmath222 then @xmath270 must be reducible over @xmath167 $ ] .
if @xmath224 then @xmath271 must be reducible over @xmath167 $ ] .
3 . if @xmath226 then @xmath272 must be reducible over @xmath167 $ ] .
[ parityex ] let @xmath6 be the adjacency matrix of a graph with cyclic excess .
if @xmath257 then the restriction of @xmath6 on the kernel of @xmath10 has even trace .
@xmath258 [ cor : trmex ] for @xmath59 and odd @xmath0 , the characteristic polynomial of the restriction of @xmath6 on the kernel of @xmath10 ( i.e. @xmath259 ) must be reducible over @xmath167 $ ] .
@xmath258 to establish some relations between the polynomials @xmath103 and @xmath49 , we make use of their respective generating functions ( ordinary power series ) @xmath274 it is convenient to introduce @xmath275 . then , it follows that @xmath276 thus , @xmath277 equation ( [ eq : gdd_ud ] ) allows us to establish some bounds for the eigenvalues of a graph with cyclic defect ( proposition [ prop : eigbounds ] ) .
[ prop : eigbounds ] for a graph of degree @xmath20 , diameter @xmath117 and cyclic defect , if @xmath108 is real and @xmath278 , then the roots of @xmath279 are real and belong to the open interval @xmath280 .
set @xmath281 , and notice that @xmath282 , with equality only when @xmath58 and @xmath283 . from equation ( [ defu ] ) observe that @xmath284 and @xmath285 .
then , using equation ( [ eq : gdd_ud ] ) we obtain that @xmath286 and that @xmath287 we compute @xmath288 for @xmath289 . since @xmath290 , it follows that @xmath291 hence , for any @xmath292 , @xmath20 and @xmath70 , with the exception of @xmath293 , @xmath58 and @xmath283 , each of the @xmath71 open intervals @xmath294 with @xmath295 contains a root of @xmath279 ( by the intermediate value theorem ) . in the case
@xmath293 , @xmath58 and @xmath283 , the roots of @xmath296 are 0 and @xmath297 , which belong to ( @xmath298 , @xmath299 ) . by virtue of proposition
[ prop : eigbounds ] , we can assume that every eigenvalue ( @xmath300 ) of @xmath6 has the form @xmath301 with @xmath281 and @xmath302 .
in this case @xmath303 as a corollary of proposition [ prop : eigbounds ] , we obtain a very useful necessary condition on the existence of graphs with cyclic defect and even order . [
cor : evenn ] if @xmath304 then a graph with cyclic defect must have an integer eigenvalue @xmath39 such that @xmath305 and @xmath306 .
the corollary follows immediately from propositions [ prop : integerrootdef ] and [ prop : eigbounds ] . extending proposition [ prop : eigbounds ] and corollary [ cor : evenn ] to graphs with cyclic excess , we obtain the following assertions . [ prop : eigboundsex ] for a graph of degree @xmath20 , odd girth @xmath119 and cyclic excess , if @xmath108 is real and @xmath278 , then the roots of @xmath307 are real and belong to the open interval @xmath280 . @xmath258
[ cor : evennex ] if @xmath304 then a graph with cyclic excess must have an integer eigenvalue @xmath39 such that @xmath305 and @xmath308 .
here we basically prove the non - existence of graphs of degree @xmath20 , diameter 4 and cyclic defect , or graphs of degree @xmath20 , girth 9 and cyclic excess .
our proof is decomposed into three parts .
we first show that the polynomial @xmath309 has an integer root , then we find those values of @xmath0 making the existence of this root possible .
finally we show that for none of these values of @xmath0 the polynomial @xmath310 has an integer root ( contradicting propositions [ prop : integerrootdef ] and [ prop : integerrootex ] ) .
[ theo : d4 ] there is no regular graph of degree @xmath20 , diameter 4 and cyclic defect , nor any regular graph of odd degree @xmath20 , girth 9 and cyclic excess .
considering graphs of diameter 4 and cyclic defect , from corollary [ congruences ] it follows that @xmath228 and that @xmath224 , while for regular graphs of odd degree @xmath20 and cyclic excess it follows that @xmath224 .
set @xmath311 , then @xmath312 . by corollary [ cor:3,4,6 ] ( for cyclic defect ) and corollary [ cor:3,4,6bis ] ( for cyclic excess )
the polynomial @xmath313 must be reducible over @xmath167 $ ] , and thus , it must have a factor of degree at most 2 .
we first claim that for @xmath314 @xmath313 must have an integer root .
* claim 1 . * for @xmath314 @xmath313 must have an integer root .
* proof of claim 1 .
* we proceed by contradiction , assuming that there is a factorization of @xmath313 into factors of degree 2 irreducible over @xmath167 $ ] .
then , from the roots @xmath315 of @xmath313 we can obtain two sets , say @xmath316 and @xmath317 , such that @xmath318 , @xmath319 , @xmath320 and @xmath321 are all integers .
using vite s formulas we obtain that @xmath322 therefore , we can compute the coefficients of the equation @xmath323 with the 3 roots @xmath324 , @xmath325 and @xmath326 ; indeed , we have that @xmath327 thus , we have to find integer solutions for @xmath328 .
discarding the uninteresting solution @xmath329 , we may write @xmath330 as @xmath331 , where @xmath332 is rational and @xmath333 .
this equation in @xmath93 has discriminant @xmath334 , which can be a perfect square only if @xmath335 , a contradiction .
therefore , @xmath336 can not have integer roots , and the claim follows .
@xmath258 since @xmath313 must have an integer root , we search the integer pairs @xmath337 such that @xmath338 .
the discriminant @xmath339 of the equation @xmath340 in @xmath341 is a perfect square iff @xmath342 ; multiplying this equation by 5 and setting @xmath343 , we obtain @xmath344 equation ( [ eq : pell ] ) is closely related to the well - known pell equation ( namely , @xmath345 , where @xmath346 ) .
the infinitely many solutions @xmath347 of equation ( [ eq : pell ] ) are given by @xmath348 and @xmath349 , where @xmath350 and @xmath351 denote the @xmath42th lucas number and @xmath42th fibonacci number , respectively ; see @xcite .
for all integers @xmath42 the recurrence equations of the lucas and the fibonacci numbers can be defined as follows .
@xmath352 if we set @xmath353 ( the so - called _ golden ratio _ ) , then @xmath354 and @xmath355 . in order to retain integer values for @xmath356 , we have that @xmath357 , and thus , that @xmath358 .
set @xmath359 , then @xmath360 and @xmath361 .
we first rule out the existence of graphs of diameter 4 and cyclic defect
. * claim 2 .
* there is no regular graph of degree @xmath20 , diameter 4 and cyclic defect .
* proof of claim 2 .
* for the aforementioned values of @xmath362 , by proposition [ prop : integerrootdef ] , the polynomial @xmath363 must have an integer root .
our goal now is to prove that this is not the case . from the two possible values for @xmath362
take @xmath364 .
note that for any two integer values @xmath93 and @xmath92 , @xmath365 is an integer .
suppose that @xmath366 is an integer root of @xmath363 , then , for @xmath366 and @xmath367 we have that @xmath368 is an integer , which implies that @xmath369 or @xmath370 . as a result
, it follows that @xmath371 note that @xmath372 is a polynomial in @xmath373 of degree 6 .
investigating the real roots of @xmath372 for each value of @xmath374 , we see that their absolute values lie between 0.05 and 9 .
but , since @xmath359 , we have that for @xmath375 , the values of @xmath373 are at least 29 , and that for @xmath376 , the values of @xmath373 lie between 0 and 0.01 .
for @xmath377 it can be easily verified that @xmath378 , which contradicts the fact that @xmath379 .
we have excluded the values of @xmath380 because they give the trivial solution @xmath381 .
therefore , for @xmath364 , there is no integer value of @xmath366 that makes @xmath382 zero .
analogously , for @xmath383 , the absolute values of the real roots of @xmath372 for each value of @xmath374 lie between 0.12 and 19 . for @xmath377 , observe that @xmath384 .
consequently , there is no integer value of @xmath366 that makes @xmath382 zero .
@xmath258 * claim 3 .
* there is no regular graph of odd degree @xmath20 , girth 9 and cyclic excess .
* proof of claim 3 . * in this case we proceed as in claim 2 , then @xmath385 and @xmath372 is a polynomial in @xmath373 of degree 6 . for @xmath364 ,
the absolute values of the real roots of @xmath372 for each value of @xmath374 lie between 0.05 and 8 . for @xmath377 , observe that @xmath378 , a contradiction . consequently , there is no integer value of @xmath366 that makes @xmath386 zero . finally , for @xmath383
, the absolute values of the real roots of @xmath372 for each value of @xmath374 lie between 0.12 and 17 .
consequently , there is no integer value of @xmath366 that makes @xmath386 zero .
@xmath258 the theorem follows from claims 2 and 3 .
[ theo : d3 ] there is no regular graph of degree @xmath20 , diameter 3 and cyclic defect . from corollary [ congruences ]
it follows that @xmath387 and that @xmath222 . in this case
we see that @xmath297 is an eigenvalue of @xmath10 with multiplicity 2 .
thus , @xmath388 must have factors of degree at most 2 , and therefore an integer root @xmath39 congruent to 1 modulo 3 . since @xmath389
, we see that @xmath390 divides @xmath391 , and thus , @xmath390 divides 9 ( because @xmath392 ) and @xmath393 . however , from these values only @xmath394 is congruent to 1 modulo 3 . for @xmath394 and @xmath395
, we have that @xmath396 and @xmath397 . by proposition [ prop : irr ]
the polynomial@xmath398 must be reducible over @xmath167 $ ] ( see also observation [ elemdiv1 ] ) .
however , @xmath399 from where we obtain that @xmath398 is irreducible over @xmath167 $ ] .
[ theo : odddiam6]there is no regular graph of odd degree @xmath20 , diameter @xmath400 , and cyclic defect . since @xmath401 , by corollary [ ddivn ] ,
the order @xmath7 of these graphs is a multiple of @xmath71 , implying that @xmath7 is a multiple of 3 and 4 . in this case , by proposition [ prop : integerrootdef ] the polynomial @xmath194 should have an integer root @xmath39 . on the other hand
, from @xmath401 it follows that @xmath402 .
set @xmath403 where @xmath404 is a polynomial of degree @xmath405 .
thus , @xmath406 should be congruent to 0 modulo 6 . but
no integer @xmath39 satisfies @xmath407 .
[ theo : oddg5 ] there is no regular graph of odd degree @xmath20 , girth 5 and cyclic excess . in this case equation ( [ eq : exm ] )
takes the form @xmath408 . if @xmath0 is odd then @xmath409 . by proposition [ prop : integerrootex ]
, there is a simple integer eigenvalue @xmath39 of @xmath6 satisfying @xmath410 as @xmath411 , 0 is an eigenvalue of @xmath10 with multiplicity 2 .
therefore , the eigenvalues of @xmath6 paired with 0 satisfy the equation @xmath412 denote by @xmath413 and @xmath414 the roots of equation ( [ eq2 ] ) . if both are eigenvalues of the restriction of @xmath6 on @xmath238 , the trace is @xmath297 ( see corollary [ cor : trmex ] )
therefore , only one of them can be an eigenvalue , say @xmath413 , implying that @xmath413 has multiplicity 2 and is an integer .
the discriminant of equation ( [ eq2 ] ) is @xmath415 and , like the discriminant @xmath416 of equation ( [ eq1 ] ) , must be a perfect square .
the only pair of perfect squares differing by 8 is \{1,9 } , implying @xmath417 , contradicting the hypothesis @xmath57 .
[ theo : g7 ] there is no regular graph of degree @xmath418 , girth 7 , and cyclic excess .
such a graph has an order multiple of 3 .
therefore , the polynomial @xmath419 has a factor of degree 1 or 2 , and thus , an integer root @xmath39 .
since @xmath0 is an integer , @xmath390 divides @xmath420 , and thus , divides 7 ( because @xmath421 ) .
the possible values for @xmath39 are @xmath422 , @xmath297 , 0 and 3 , and the corresponding values for @xmath0 are 8 , 2 , 0 and 6 .
the orders for the interesting degrees 6 and 8 are @xmath423 and @xmath424 , respectively .
but in both cases , substituting @xmath425 in @xmath426 , we obtain an irreducible polynomial @xmath427 of degree 9 , contradicting proposition [ prop : irrbis ] .
thus , none of these graphs exists . in this section we show how to use corollaries [ cor : evenn ] and [ cor : evennex ] , and the software maple@xcite in order to prove the non - existence of graphs of small degree with cyclic defect or excess .
specifically , we analyze the existence of an integer root in the polynomials @xmath428 for @xmath429 and small degrees .
cubic graphs with cyclic defect or excess are considered in subsection [ sec : cubic ] , while the case of @xmath59 for all graphs of odd degree and cyclic excess was dealt in subsection [ subsec : g5 ] . in this subsection
we assume @xmath430 and @xmath268 .
[ theo:5def ] for @xmath431 there is no graph of degree 5 , diameter @xmath71 and cyclic defect .
furthermore , for @xmath432 , @xmath433 odd , there is no graph of degree 5 , girth @xmath433 and cyclic excess . for @xmath429
we analyze the polynomial @xmath434 , for @xmath435 and @xmath436 . for @xmath437 , we have that @xmath438 if @xmath439 . for
@xmath429 @xmath440 only takes the values @xmath441 for @xmath442 , and in these cases @xmath443 . however , as the order of such graphs is a multiple of 4 . by corollary [ cor:3,4,6 ]
, both @xmath444 and @xmath445 must be reducible over @xmath167 $ ] , but they are not . for
@xmath429 @xmath446 takes the values @xmath441 only for @xmath447 , and then @xmath448 .
but , @xmath257 and @xmath449 is irreducible over @xmath167 $ ] , contradicting corollary [ cor:3,4,6 ] .
[ theo:7def ] for any @xmath121 there is no graph of degree 7 , diameter @xmath71 and cyclic defect .
furthermore , for any @xmath268 , @xmath433 odd , there is no graph of degree 7 , girth @xmath433 and cyclic excess .
since @xmath450 , it suffices to look at @xmath451 for @xmath435 and @xmath436 .
indeed , @xmath452 for @xmath436 ; for any @xmath453 and @xmath454 , we have that @xmath455 ; and for @xmath439 and @xmath456 , it follows that @xmath457 .
[ theo:9def ] for @xmath431 there is no graph of degree 9 , diameter @xmath71 and cyclic defect .
furthermore , for @xmath432 , @xmath433 odd , there is no graph of degree 9 , girth @xmath433 and cyclic excess . since @xmath458
, it suffices to look at the values of @xmath459 for @xmath435 and @xmath460 . for @xmath461 and @xmath462 ,
the value @xmath459 is a multiple of 4 . for @xmath429 and @xmath463
, we have that @xmath459 never takes the values @xmath441 .
[ theo:11def ] for @xmath431 there is no graph of degree 11 , diameter @xmath71 and cyclic defect .
furthermore , for @xmath432 , @xmath433 odd , there is no graph of degree 11 , girth @xmath433 and cyclic excess .
since @xmath464 , it suffices to look at @xmath465 for @xmath435 and @xmath466 .
first , for @xmath467 , @xmath468 does not take the values 2 or @xmath195 .
then , for @xmath453 and @xmath469 , we have that @xmath465 is a multiple of 4 , while for @xmath439 and @xmath470 , @xmath465 is a multiple of 10 .
finally , for @xmath471 @xmath472 never takes the values @xmath441 .
[ theo:13def ] for @xmath431 there is no graph of degree 13 , diameter @xmath71 and cyclic defect .
furthermore , for @xmath432 , @xmath433 odd , there is no graph of degree 13 , girth @xmath433 and cyclic excess .
since @xmath464 , it suffices to look at the values of @xmath473 for @xmath435 and @xmath466 . for @xmath453 and @xmath474
, we have that @xmath473 is a multiple of 4 , while for @xmath439 and @xmath475 , the polynomial @xmath473 is a multiple of 6 . for @xmath476 and @xmath477 , @xmath473 does not take the value @xmath478 or @xmath195 .
finally , the polynomial @xmath479 never takes the values @xmath441 for @xmath466 .
this approach is likely to work for graphs of higher degrees and larger diameters or girths , but its application quickly becomes monotonous and uninteresting .
however , the aforementioned non - existence results of graphs of odd degree with cyclic defect or excess motivated us to unveil a deeper phenomenon , namely , the finiteness of such graphs ( see subsection [ sec : finiteness ] ) . in this section we prove the most important results of the paper , namely , the finiteness of all graphs of odd degree @xmath430 and cyclic defect or excess ( see theorem [ theo : finiteness ] ) , and the non - existence of cubic graphs with cyclic defect or excess ( see theorem [ theo:3def ] ) . the idea behind the proof of theorem [ theo : finiteness ] is the following . for any odd degree @xmath430 graphs of diameter @xmath13 and cyclic defect , or graphs of girth @xmath480 and cyclic excess have an order multiple of 4 , implying that the polynomial @xmath11 must have an algebraic integer of degree at most 2 as a root . making use of equation ( [ eq : gcos ] ) and the fact that any eigenvalue @xmath481 has the form @xmath482 ( with @xmath302 ) ,
we show that @xmath483 must be an algebraic integer of degree at most 4 .
we then note that if , for a given @xmath0 and an eigenvalue @xmath39 , equation ( [ eq : gcos ] ) has at least two values of @xmath13 , then @xmath107 must be rational . in the case of @xmath107 being rational and @xmath483 being an algebraic integer of degree at most 4
, we verify that , for @xmath430 and all the possible values of @xmath483 , the polynomial @xmath484 has no algebraic integer of degree at most 2 as a root .
this last result implies that for a given @xmath0 the number of different eigenvalues @xmath39 of equation ( [ eq : gcos ] ) represents an upper bound for the number of graphs of degree @xmath0 and cyclic defect or excess .
finally , we proceed to provide an asymptotic bound for the number of such eigenvalues , knowing that they are algebraic integers of degree 2 lying between @xmath485 and @xmath486 .
[ theo : finiteness ] there are finitely many graphs of odd degree @xmath430 and cyclic defect or excess .
furthermore , an asymptotic bound for the number of such graphs is given by @xmath487 .
for graphs of diameter @xmath1 and cyclic defect , and graphs of girth @xmath2 and cyclic excess , if its degree @xmath0 is odd then its order @xmath7 is a multiple of 4 , which implies , by corollary [ cor:3,4,6 ] ( for cyclic defect ) and corollary [ cor:3,4,6bis ] ( for cyclic excess ) , that the polynomial @xmath11 must be reducible over @xmath167 $ ] . from propositions [ prop : eigbounds ] and [ prop : eigboundsex ]
it follows that an eigenvalue @xmath39 @xmath188 of such graphs has the form @xmath488 with @xmath302 and @xmath281 , and that @xmath39 lies between @xmath489 and @xmath490 . in this case , because of equation ( [ eq : gcos ] ) the equation @xmath491 implies that @xmath492 also , by observation [ elemdiv1 ] , such an eigenvalue is an algebraic integer of degree at most 2 .
we first claim the following .
* claim 1 . * for a given eigenvalue @xmath493 , the number @xmath483 is an algebraic integer of degree at most 4 .
* proof of claim 1 . * because of equation ( [ eq : gdd_ud ] ) , we can expressed equation ( [ eq : qsin ] ) as @xmath494 then , as @xmath495 , it follows that @xmath496 from equation ( [ eq : u_k ] ) it follows that @xmath497 is a polynomial of degree @xmath498 with integer coefficients , having @xmath483 as a root . therefore , @xmath483 is an algebraic integer of degree at most @xmath498 . to see that @xmath483 is in fact an algebraic integer of degree at most @xmath499 , we need the following facts from algebraic number theory ( i )
if @xmath190 is an algebraic number of degree @xmath500 then @xmath501 is also an algebraic number of degree @xmath500 , and ( ii ) if @xmath190 and @xmath502 are algebraic numbers of degree @xmath500 and @xmath503 , respectively , then @xmath504 is an algebraic number whose degree divides @xmath505 .
as @xmath493 is an algebraic number of degree at most 2 and @xmath506 is an algebraic number of degree 2 , by the previous facts , @xmath483 is an algebraic number of degree at most 4 .
@xmath258 * claim 2 . * for a given odd degree @xmath430 and an eigenvalue @xmath39 , there is only one value of @xmath13 satisfying equation ( [ eq : qsin ] ) . *
proof of claim 2 .
* we proceed by contradiction , assuming that for a given odd degree @xmath430 and an eigenvalue @xmath39 ( @xmath507 ) , there are at least two values @xmath508 and @xmath509 for which equation ( [ eq : qsin ] ) holds .
observe that in this case @xmath510 , where @xmath511 . indeed , assuming that @xmath512 ( for otherwise @xmath513 with @xmath514 ) , equation ( [ eq : qsin ] ) is equivalent to @xmath515 ( since @xmath516 ) .
if there are two values @xmath508 and @xmath509 for which equation ( [ eq : qsin ] ) holds , then @xmath517 then , as @xmath518 is a function with period @xmath519 , we have that @xmath520 , where @xmath521 . in other words , @xmath522 is rational .
therefore , for @xmath302 we have three cases according to the degree of @xmath483 ; see @xcite . * if @xmath523 is an algebraic integer of degree 1 , then @xmath524 . *
if @xmath523 is an algebraic integer of degree 2 , then @xmath525 @xmath526 * if @xmath523 is an algebraic integer of degree 4 , then @xmath527 with @xmath528 and @xmath529 , or equivalently , @xmath530
@xmath531 we analyze each case in order , that is , for the aforementioned values of @xmath483 we look for the values of odd @xmath0 and @xmath13 satisfying equation ( [ eq : u_k ] ) , or equivalently , @xmath532 to do this task we sometimes rely on the software maple@xcite . * case ( i ) * @xmath524 . for @xmath533 and @xmath534
, we have that @xmath535 , and thus , @xmath60 ; for @xmath536 and @xmath537 , we have @xmath538 ; and for @xmath539 and @xmath534 , we have @xmath540 . therefore , there are no feasible values for @xmath0 and @xmath13 .
* case ( ii ) * @xmath541 .
the only viable value of @xmath542 is @xmath543 , which implies that @xmath283 .
this case occurs when @xmath544 and @xmath545 .
* case ( iii ) * @xmath527 with @xmath528 and @xmath529 .
in this case it can be verified that the only feasible values of @xmath107 are @xmath546 ( @xmath547 ) and @xmath548 ( @xmath549 ) . for these values of @xmath483
, we have that @xmath283 ( @xmath550 ) when @xmath551 . as a result ,
when @xmath522 is rational there is no odd degree @xmath430 satisfying equation ( [ eq : qsin ] ) , and thus , the claim follows .
@xmath258 claim 2 also tells us that , for a given odd degree @xmath0 , the number of distinct eigenvalues @xmath39 is an upper bound for the number of graphs of degree @xmath0 and cyclic defect and excess .
recall that , since characteristic polynomials have integer coefficients , if @xmath39 is an eigenvalue , so is its conjugate @xmath552 .
* claim 3 . *
let @xmath39 be an eigenvalue of a graph of odd degree and cyclic defect or excess , such that @xmath305 and @xmath553 .
then , the number of such eigenvalues lying in @xmath554 is @xmath487 .
* proof of claim 3 .
* we first state a very well known fact about the ring @xmath555 of integers of @xmath556 , where @xmath557 is a square - free integer ( see ( * ? ? ?
* theorem 1 on pp .
35 ) ) : if @xmath558 then @xmath559 , while if @xmath560 then @xmath561 with @xmath562 . by virtue of the previous fact and as @xmath305 and @xmath553
, we can assume that @xmath39 has the form either @xmath563 with @xmath564 or @xmath565 with @xmath564 , @xmath566 and @xmath567 .
in the former case it follows that @xmath568 and @xmath569 , while in the latter we have that @xmath570 and @xmath571 .
an asymptotic bound can be obtained from the remark that the number of pairs @xmath572 such that @xmath573 and that @xmath574 is at most @xmath575 .
indeed , @xmath576 the lemma follows from considering the bounds for @xmath341 and @xmath577 .
@xmath258 note that the representation of a number @xmath39 is not unique ; for instance , the numbers @xmath578 and @xmath579 represent the same @xmath39 .
however , this detail only makes our bound rougher .
a more careful counting shows that @xmath580 is indeed an upper bound for the number of algebraic integers of degree 2 lying in @xmath280 . as a way of illustration , see that for @xmath283 there are @xmath581 algebraic integers between @xmath298 and @xmath299 while for @xmath582 there are @xmath583 algebraic integers lying in @xmath584 .
the proof of theorem [ theo : finiteness ] follows immediately from claims 1 , 2 and 3 .
theorem [ theo : finiteness ] did not settle the finiteness of cubic graphs with cyclic defect or excess . in this subsection
we take care of this case .
the proof of theorem [ theo:3def ] first exploits the fact that for a cubic graph of diameter @xmath13 and cyclic defect , or a cubic graph of girth @xmath480 and cyclic excess , the polynomial @xmath585 must have an integer root @xmath500 between @xmath586 and @xmath587 . in this direction
we prove that @xmath588 , and in this case , @xmath589 and @xmath590 , thus ruling out the existence of graphs with cyclic defect . as the order @xmath7 of cubic graph with cyclic
excess is a multiple of 4 , @xmath591 has an algebraic root of degree at most 2 .
in addition , if @xmath590 then @xmath222 , so @xmath592 must also have algebraic roots of degree 1 or 2 . recall that all these algebraic integers lie on the interval @xmath593 .
the next step of the proof is to settle that @xmath594 is the only algebraic integer of degree at most 2 for which the polynomial @xmath595 , where @xmath596 .
the conditions that @xmath545 , @xmath597 and @xmath589 greatly narrow down the numbers @xmath356 that could make @xmath598 ; we then prove that the only such numbers are @xmath599 and @xmath600 .
the proof ends when we prove that , in fact under all the previous conditions , for @xmath599 @xmath600 the polynomial @xmath601 .
[ theo:3def ] for @xmath221 there is no graph of degree 3 , diameter @xmath71 and cyclic defect .
furthermore , for odd @xmath74 there is no graph of degree 3 , girth @xmath433 and cyclic excess .
suppose , for the sake of contradiction , that there is at least a cubic graph of diameter @xmath602 and cyclic defect , and at least a graph of girth @xmath89 and cyclic excess .
relying on corollaries [ cor : evenn ] and [ cor : evennex ] , next we discard the values of @xmath439 and @xmath603 for which @xmath591 is different from @xmath478 or @xmath195 .
note that @xmath604 , @xmath605 , @xmath606 are multiples of 4 if @xmath453 and that @xmath607 if @xmath56 ; this can be checked easily by induction .
for @xmath467 the polynomial @xmath608 has no factor of degree 1 or 2 to be used with the eigenvalue 0 of the matrix @xmath609 ( see corollary [ cor:3,4,6 ] ) .
therefore , there are no cubic graphs of diameter 3 ( for cyclic defect ) or girth 7 ( for cyclic excess ) . for @xmath453 we can check by induction that @xmath610 if @xmath13 is even and that @xmath611
if @xmath13 is odd . therefore , from now on we can assume @xmath589 and @xmath537 .
as a consequence , there is no cubic graph of diameter @xmath453 and cyclic defect .
we now concentrate on cubic graphs of girth @xmath612 and cyclic excess , for @xmath537 .
computing modulo 32 we see that the value @xmath589 can be attained only if @xmath613 . in this case , these graphs have an order multiple of 4 and @xmath614 , and therefore , we must add the conditions that @xmath591 and @xmath592 have algebraic roots of degree 1 or 2 .
henceforth , together with the integers @xmath615 and 2 , we analyze the set of the 38 algebraic integers of degree 2 between @xmath298 and @xmath299 .
these numbers are as follows : @xmath616 for @xmath617 , @xmath618 for @xmath619 , @xmath620 for @xmath621 and at last @xmath622 .
we now prove the following .
* claim 1 . * among all the algebraic integers of degree at most 2 in the interval @xmath593 , the polynomial @xmath591 with @xmath596 takes 0 only for @xmath594 .
* proof of claim 1 . *
the function @xmath606 in @xmath13 is null for @xmath596 .
the polynomial @xmath591 never takes the value 0 for @xmath623 , because @xmath624 , @xmath625 and @xmath626 .
these assertions can be proved by induction .
the value @xmath595 is also obtained several times for @xmath627 , but only if @xmath628 .
the function @xmath629 in @xmath13 is periodic and never null for @xmath596 ; the same happens for @xmath630 and @xmath631 .
consider the algebraic integers @xmath632 of degree 2 and odd _ norm _
@xmath633 of quadratic number @xmath356 is @xmath634 . ] in the _ algebraic extension _ is called algebraic if every element of @xmath635 is algebraic over @xmath636 . ] @xmath637 .
then , if @xmath638 does not belong to the _ principal ideal _ of the ring @xmath555 of algebraic integers that forms an additive group and has the property that if @xmath639 and @xmath640 then @xmath641 .
the _ principal ideal generated by @xmath107 _ with @xmath642 is defined as @xmath643 . ]
generated by 2 in the ring @xmath555 of algebraic integers , then the polynomial @xmath644 never enters into that ideal , and thus , never vanishes .
this is the case for the numbers @xmath645 , @xmath646 , @xmath647 , @xmath648,@xmath622 , @xmath649 , and @xmath650 . for the numbers
@xmath651 we note that for @xmath652 the number @xmath653 has the form @xmath654 with @xmath655 , and that for @xmath656 the number @xmath657 has the form @xmath658 with @xmath655 .
furthermore , @xmath659 .
therefore , we have ruled out all the numbers @xmath660 .
note that if @xmath661 then @xmath662 . finally , observe that for @xmath652 the number @xmath663 has the form @xmath664 with @xmath655 .
this leaves the numbers @xmath665 out .
this completes the proof of the claim .
we finalize the proof of the theorem by showing the following two claims .
* claim 2 . *
provided that @xmath652 with @xmath545 , @xmath597 and @xmath589 , the only numbers @xmath356 that could make @xmath598 are @xmath599 and @xmath600 .
* proof of claim 2 . * set @xmath632 , then to have simultaneously @xmath597 , @xmath589 and @xmath598 we must have that @xmath666 and that @xmath667 .
since the coefficients of @xmath591 are integers , the former condition means that @xmath668 divides 1 in the ring of integers of @xmath669 , while the latter condition implies that @xmath638 divides 1 in the aforementioned ring .
these conditions also imply that the norms of @xmath668 and @xmath638 in the algebraic extension @xmath670 must be @xmath671 or @xmath297 .
thus , only the numbers @xmath672 and @xmath673 satisfy both conditions . @xmath258 * claim 3 . * for @xmath652 with @xmath545 , the polynomials @xmath674 and @xmath675 never take the value 1 . *
proof of claim 3 .
* we first consider the value @xmath676 , and claim that for @xmath677 with @xmath678 , the values @xmath679 are never integers .
observe that @xmath680 and that @xmath681 .
as @xmath682 and @xmath683 , by induction @xmath684 . ]
, we obtain that @xmath685 , which is not the norm of an integer .
recall that the norm of integers is congruent to 0 or 1 modulo 4 .
this approach also rules out the value @xmath686 .
the values of @xmath687 are never 1 for @xmath688 .
indeed , computing modulo 4 in the ring of integers of @xmath689 of @xmath669 , we say that @xmath107 divides @xmath108 , denoted by @xmath690 , if @xmath691 , and that @xmath692 if @xmath693 . ] , we see that @xmath694 if @xmath695 is multiple of 3 , that @xmath696 if @xmath697 , and that @xmath698 if @xmath699 .
this approach also shows that @xmath700 is never 1 for @xmath688 .
this completes the proof of the claim . @xmath258
combining claims 1 , 2 and 3 the theorem follows .
an immediate corollary of theorem of [ theo:3def ] is the finiteness of cubic graphs with cyclic defect or excess ( corollary [ cor : cubicfiniteness ] ) , settling , in this way , the finiteness of all graphs of odd degree and cyclic defect or excess .
[ cor : cubicfiniteness ] for @xmath56 , apart from the mbius ladder on 8 vertices , there is no cubic graph of diameter @xmath13 and cyclic defect nor any cubic graph of girth @xmath480 and cyclic excess .
using a number of algebraic approaches , we proved the non - existence of infinitely many graphs with cyclic defect or excess , and the finiteness of graphs of odd degree and cyclic defect or excess .
while substantial progress in this direction was made through algebraic approaches , definitive solutions to problems [ probrep ] and [ probex ] still seem to be elusive , mainly due to the complexity of the theoretical problems that emerged during our investigation .
for instance , the approach which ruled out the existence of cubic graphs with cyclic defect or excess may work for higher degrees , but the complexity of the analysis also increases considerably .
the condition of having cyclic defect or excess imposes heavy constraints on the structure of graphs with defect or excess 2 , so we firmly believe that the mbius ladder on 8 vertices is the only such graph , and accordingly , conjecture it .
j. conde and j. gimbert , _ on the existence of graphs of diameter two and defect two _ , discrete mathematics * 309 * ( 2009 ) , no . 10 , 31663172 , http://dx.doi.org/10.1016/j.disc.2008.09.017[doi:10.1016/j.disc.2008.09.017 ] .
p. kovcs , _ the non - existence of certain regular graphs of girth 5 _ , journal of combinatorial theory , series b * 30 * ( 1981 ) , no . 3 , 282284 , http://dx.doi.org/10.1016/0095-8956(81)90045-9[doi:10.1016/0095-8956(81)90045-9 ] .
m. miller , m. h. nguyen , and g. pineda - villavicencio , _ on the non - existence of odd degree graphs with diameter 2 and defect 2 _ , proceedings of iwoca 2007 , the 18th international workshop on combinatorial algorithms ( lake macquarie , nsw , australia ) ( l. brankovic , y. q. lin , and w. f. smyth , eds . ) , iwoca proceedings , college publications , nov 2007 , pp .
143150 .
r. c. singleton , _ on minimal graphs of maximum even girth _ , journal of combinatorial theory * 1 * ( 1966 ) , no . 3 , 306332 , http://dx.doi.org/10.1016/s0021-9800(66)80054-6[doi:10.1016/s0021-9800(66)80054-6 ] . | the moore bound constitutes both an upper bound on the order of a graph of maximum degree @xmath0 and diameter @xmath1 and a lower bound on the order of a graph of minimum degree @xmath0 and odd girth @xmath2 .
graphs missing or exceeding the moore bound by @xmath3 are called _ graphs with defect or excess @xmath3 _ , respectively .
regular graphs with defect @xmath3 satisfy the equation @xmath4 , and regular graphs with excess @xmath3 satisfy the equation @xmath5 , where @xmath6 denotes the adjacency matrix of the graph in question , @xmath7 its order , @xmath8 the @xmath9 matrix whose entries are all 1 s , @xmath10 a matrix with the row and column sums equal to @xmath3 , and @xmath11 a polynomial with integer coefficients such that the matrix @xmath12 gives the number of paths of length at most @xmath13 joining each pair of vertices in the graph . for _ moore graphs _ ( graphs with @xmath14 ) ,
the matrix @xmath10 is the null matrix . for graphs with defect or excess 1 ,
@xmath10 can be considered as the adjacency matrix of a matching with @xmath7 vertices , while for graphs with defect or excess 2 , @xmath10 can be assumed to be the adjacency matrix of a union of vertex - disjoint cycles .
graphs with defect 1 do not exist for any degree @xmath15 and diameter @xmath16 , while graphs with excess 1 do not exist for any degree @xmath15 and odd girth @xmath17 .
however , graphs with defect or excess 2 represent a wide unexplored area . graphs with defect or excess 2 having the adjacency matrix of a cycle of order @xmath7 as the matrix @xmath10 are called _ graphs with cyclic defect or excess _ ; these graphs are the subject of our attention in this paper .
we obtain the following results about graphs with cyclic defect or excess .
we prove the non - existence of infinitely many such graphs , for example , graphs of any degree @xmath15 , diameter 3 or 4 and cyclic defect ; and graphs of degree @xmath18 , girth 7 and cyclic excess . as the highlight of the paper we provide the asymptotic upper bound of @xmath19 for the number of graphs of odd degree @xmath20 and cyclic defect or excess .
this bound is in fact quite generous , and as a way of illustration , we also show that there are no graphs of degree 3 or 7 , for diameter @xmath15 and cyclic defect or for odd girth @xmath17 and cyclic excess , nor any graphs of odd degree @xmath15 , girth @xmath21 or 9 and cyclic excess .
actually , we conjecture that , apart from the mbius ladder on 8 vertices , no non - trivial graph of any degree @xmath22 and cyclic defect or excess exists . to obtain our results we rely on algebraic methods , for instance , on the connection between the polynomial @xmath11 and the classical chebyshev polynomials , on eigenvalues techniques and on elements of algebraic number theory .
* keywords : * moore bound , moore graph , defect , excess , chebyshev polynomial of the second kind , cyclic defect , cyclic excess , pell equation . *
ams subject classification : * 05c12 , 05c35 , 05c50 , 05c75 . |
quantum mechanics is known to differ from classical mechanics due to the existence of the position momentum uncertainty relation by heisenberg @xcite . the uncertainty relation containing the correlation of the position and momentum
was found by robertson @xcite and schrdinger @xcite .
there exists the uncertainty relation of the position and momentum for mixed quantum states @xcite .
new kinds of the uncertainty relations were obtained by trifonov @xcite .
extensions of the uncertainty relations of @xcite for mixed states were found by karelin @xcite .
review of the uncertainty relations in quantum mechanics is given in @xcite . there
exist specific uncertainty relations called `` entropic uncertainty relations '' based on the notion of shannon entropy and information @xcite .
these relations , which read as inequalities for entropy associated with the position - and - momentum probability distributions , were discussed , for example , in @xcite .
recently a new formulation of quantum mechanics where the quantum states are described by tomographic - probability distributions ( instead of the wave function or density matrices ) was suggested @xcite . for a system with continuous degrees of freedom ,
such probability is the symplectic tomogram of the quantum state @xcite .
the corresponding symplectic tomographic entropy was introduced for quantum states in @xcite and in signal analysis in @xcite . in @xcite
the symplectic entropy was discussed for the bec solitons , in view of the tomogram of the solution to gross
pitaevskii equation . a general approach to quantum information including the application of different kinds of tomographic entropies was developed in @xcite .
the aim of this study is to establish a new kind of entropic uncertainty relations formulated as inequality for the entropy associated with the symplectic tomogram of the quantum state of a system with continuous degrees of freedom .
the paper is organized as follows . in section [ sect2 ] , a review of known entropic uncertainty relations for systems with continuous variables is presented while , in section [ sect3 ] ,
the symplectic - tomography approach is discussed .
entropic inequalities for symplectic entropy are studied in section [ sect4 ] and examples of new inequalities for the gaussian packets ( squeezed states ) and soliton solution of gross pitaevskii equation ( bose einstein condensates ) are given in section [ sect5 ] .
finally , conclusions are summarized in section [ sect6 ] .
in the context of information theory , entropy is related to an arbitrary probability - distribution function @xcite .
for example , given the probability distribution @xmath0 , where @xmath1 is a discrete random variable , i.e. , @xmath2 together with the normalization condition @xmath3 one has , by definition , the entropy @xmath4 in quantum mechanics , the discrete probability distributions are standard ingredients in the description of spin ( see , e.g. , tomographic probability of spin states in @xcite and related entropy for spin tomograms in @xcite ) . for continuous variables ,
the wave function @xmath5 provides the probability - distribution density @xmath6 the corresponding entropy reads ( see , e.g. , @xcite ) @xmath7 in the momentum representation , one has the wave function @xmath8 the corresponding entropy related to the momentum - probability density @xmath9 reads @xmath10 it is worthy noting that one can construct entropies @xmath11 and @xmath12 not only in quantum mechanics .
if the function @xmath5 is replaced by a signal function @xmath13 depending on time @xmath14 , the function @xmath15 is replaced by the function @xmath16 describing the signal spectrum . in this case , the entropy of the signal @xmath17 and its spectrum @xmath18 provide some information characteristics of the signal . from mathematical point of view
, there exists the correlation of entropies @xmath11 and @xmath12 @xmath19 and @xmath20 , since the function @xmath5 @xmath21 $ ] determines the fourier component @xmath15 @xmath22 $ ] .
this means that the entropies @xmath11 and @xmath12 have to obey some constrains .
these constrains are entropic uncertainty relations ( some inequalities ) . for the one - mode system ,
the inequality reads ( see @xcite , p. 28 ) @xmath23 or @xmath24 for the gaussian wave functions ( gaussian signals ) describing the states without correlations of the position and momentum , e.g. , the ground state of the harmonic oscillator @xmath25 one has @xmath26 consequently , @xmath27 the equality takes place for squeezed states with the wave function @xmath28 thus , one has @xmath29 where @xmath30 and @xmath31 read @xmath32 and @xmath33 for squeezed and correlated @xcite states , the wave functions have the gaussian form , i.e. , @xmath34 and @xmath35 here @xmath36 is the correlation coefficient of the position and momentum , i.e. , @xmath37 and for squeezed but not correlated states @xmath38 .
the sum of entropies for the squeezed and correlated states reads @xmath39 for squeezed but not correlated states , the entropy @xmath11 differs from @xmath12 . for multimode systems ( multicomponent signals ) ,
the entropy uncertainty relation reads @xmath40 where @xmath41 is the number of degrees of freedom of the system and @xmath42 the functions @xmath43 and @xmath44 are connected by the fourier transform @xmath45 for the gaussian wave function corresponding to factorized squeezed state of several modes , @xmath46
there exists an invertable map of the density operator ( matrix ) @xmath47 onto the symplectic tomogram @xcite which is the probability - density of random quadrature @xmath48 @xmath49 here @xmath47 is the density operator .
the parameters @xmath50 and @xmath51 are real parameters and operators @xmath52 and @xmath53 are quadrature operators .
the map ( [ t1 ] ) has the inverse @xcite @xmath54dx\,d\mu\,d\nu.\ ] ] for a pure state @xmath55 , the transform ( [ t1 ] ) yields @xcite @xmath56 the function @xmath57 ( the state tomogram ) is the probability density of the position @xmath48 , i.e. , @xmath58 and @xmath59 transformation ( [ t1 ] ) can be expressed in terms of the real wigner function @xcite @xmath60 where @xmath61 is the density matrix in the position representation , and for @xmath62 one has @xmath63 the density matrix reads @xmath64 in terms of the tomogram , one has @xmath65}\,dy\,d\mu.\ ] ] the tomographic - probability density has the homogeneity property following from its definition ( [ t1 ] ) and the relation for the dirac delta - function @xmath66 the homogeneity property reads @xmath67 also for the pure state , one has @xmath68 and @xmath69 where @xmath70 is the wave function in the position representation and @xmath71 is the wave function in the momentum representation . these properties are connected with the expression of symplectic tomogram in terms of the wigner function @xcite @xmath72 the inverse transform reads @xmath73\,dx\,d\mu\,d\nu.\ ] ] since for the pure state , @xmath74 and @xmath75 relations ( [ t11 ] ) and ( [ t12 ] ) are easily obtained .
tomogram ( [ t4 ] ) can be rewritten in the form @xmath76dy\right|^2.\ ] ] for @xmath77 and @xmath78 , one has the optical tomogram @xcite @xmath79\frac{dy}{\sqrt{2\pi i\sin t}}\right|^2.\ ] ] on the other hand , this tomogram formally equals to @xmath80 where the wave function reads @xmath81\psi(y)\,{dy},\ ] ] being the fractional fourier transform of the wave function @xmath82 .
this wave function corresponds to the wave function of a harmonic oscillator with @xmath83 taken at the time moment @xmath14 provided the wave function at the initial time moment @xmath84 equals to @xmath82 . for mixed state with density operator @xmath47 written in the form of the spectral decomposition , @xmath85 where @xmath86 are nonnegative eigenvalues and @xmath87 are the eigenvectors of the density operator
, the optical tomogram reads @xmath88{dy}\right|^2.\ ] ] in eqs .
( [ t22 ] ) and ( [ t24 ] ) , we use the identity of the kernel of fractional fourier transform to the green function of the schrdinger evolution equation for the harmonic oscillator @xcite .
tomogram of a mixed state takes the form of convex sum of tomograms of pure states @xmath87 , i.e. , @xmath89 where @xmath90 are given by eq .
( [ t17 ] ) with @xmath91 in view of eqs .
( [ t11 ] ) and ( [ t12 ] ) , one has for mixed state @xmath92 and @xmath93 where @xmath94 is the complex wave function in the position representation of the eigenstate @xmath87 and @xmath95 is the complex wave function in the momentum representation of this state .
thus , we pointed out that symplectic tomogram of a quantum state can be interpreted as modulus squared of the harmonic - oscillator s wave function for pure state or as convex sum of modulus squared of such functions for mixed state .
another possibility for analogous interpretation follows in view of considering the tomogram @xmath57 within the framework of fresnel - tomography approach @xcite .
in fact , formula ( [ t4 ] ) with @xmath96 can be rewritten in the form @xmath97\psi(y)\,dy\right|^2.\ ] ] the fresnel tomogram @xmath98 is related to the optical tomogram by @xmath99 on the other hand , @xmath100 can be considered as the wave function of free particle at the time moment @xmath101 if the initial value of the wave function at the time moment @xmath84 is equal to @xmath82 .
thus , the fresnel tomogram for pure state can be interpreted as modulus squared of the wave function of free particle .
the fresnel tomogram is the probability distribution satisfying the normalization condition @xmath102 for mixed state ( [ t23 ] ) , the fresnel tomogram reads @xmath103\psi_k(y)\,dy\right|^2\ ] ] and @xmath104
since the symplectic tomogram has the standard probability distribution features , one can introduce entropy associated with the tomogram of quantum state @xcite or of analytic signal @xcite .
thus one has entropy as the function of two real variables @xmath105 in view of the homogeneity and normalization conditions for tomogram ( [ t10 ] ) , ( [ t5 ] ) one has the additivity property @xmath106 for pure state @xmath107 , one obtains the entropies @xmath11 and @xmath12 , namely , @xmath108 and @xmath109 in view of inequality ( [ a10 ] ) , one has the inequality for tomographic entropies @xmath110 for multimode system , the symplectic entropy reads @xmath111 since the symplectic entropy is related to entropies @xmath112 and @xmath113 of the multimode - system state , one can use inequality ( [ a21 ] ) to obtain the entropic uncertainty relation in the form of inequality for symplectic entropies @xmath114 where @xmath115 means @xmath116 and @xmath117 means @xmath118 .
the fresnel tomogram provides the fresnel entropy of the quantum state @xmath119 it can be readily seen that the fresnel entropy @xmath120 can be easily obtained from the symplectic entropy ( [ t33 ] ) choosing @xmath96 , i.e. , @xmath121 this also means that @xmath122 for the optical tomogram ( [ t18 ] ) , entropy is defined by the formula @xmath123 for the pure state , one has @xmath124 and @xmath125 in view of the expression of tomogram in terms of wave function ( [ t21 ] ) and ( [ t22 ] ) , one has the entropic uncertainty relation in the form @xmath126 since symplectic and optical tomograms are connected as follows : @xmath127 the corresponding entropies are also connected @xmath128 for given symplectic entropy of any pure state @xmath129 , inequality ( [ t46 ] ) reads @xmath130 the optical tomogram @xmath131 and symplectic tomogram @xmath57 connected by ( [ t47 ] ) can be related by another formula @xmath132 this means that for given optical tomogram @xmath131 one can reconstruct symplectic tomogram @xmath57 . inserting eq .
( [ t50 ] ) into basic equation defining the entropy ( [ t33 ] ) yields the equality @xmath133 for symplectic entropies ( [ t49 ] ) , the entropic uncertainty relation yields @xmath134 the extension of this inequality for multimode system reads @xmath135 where entropy @xmath136 is given by ( [ t38 ] ) .
tomogram of the ground state of multimode isotropic harmonic oscillator with unit masses and frequencies has the form @xmath137 entropy associated with this tomogram reads @xmath138 this entropy does not depend on the parameter @xmath139 .
one can check that , if @xmath140 and @xmath141 in formula ( [ t55 ] ) , relation ( [ t53 ] ) yields for @xmath142 the equality @xmath143
entropy of the soliton solution to nonlinear equations was discussed in @xcite . in particular , the soliton solution to gross
pitaevskii equation @xcite was considered in the tomographic - probability representation to study bose
einstein condensate ( bec ) ( see also @xcite ) .
bec soliton under consideration is given as the function @xmath144 where the parameter @xmath145 describes the soliton width .
symplectic tomogram of bec soliton reads @xmath146 where @xmath147 and @xmath148 . since @xmath149 , tomogram ( [ ss2 ] ) is nonnegative normalized probability distribution of random position @xmath48 .
the tomographic entropy of bec soliton equals to @xmath150 we introduce the function @xmath151 according to the entropic uncertainty relation ( [ t52 ] ) this function ( we call it entropic uncertainty function ) must be nonnegative .
equation ( [ t51 ] ) and the additivity property ( [ t34 ] ) mean that the entropic uncertainty function ( [ ss4 ] ) does not depend on parameter @xmath152 .
plots of function ( [ ss4 ] ) for the gaussian state and for the soliton are presented below .
\(i ) the normalized initial gaussian profile is given by @xmath153 where @xmath154 is the waist of gaussian profile .
the corresponding tomogram calculated with the help of eq .
( [ t4 ] ) with @xmath155 and @xmath148 is given by @xmath156.\ ] ] the symplectic gaussian entropy is given as follows : @xmath157.\ ] ] the corresponding entropic uncertainty function @xmath158 in this case can be calculated explicitly @xmath159.\ ] ] note that the positive definite function @xmath158 does not depend on the radial variable @xmath152 , i.e. , it is the same for both the symplectic and optical entropies and it reduces to zero for @xmath160 , whereas for @xmath161 it is periodic with period @xmath162 as can be seen from fig.[figure1 ] .
_ plot of function @xmath158 for two values of the gaussian waist @xmath163 and @xmath164 . _ ] \(ii ) the initial profile of soliton is given by ( [ ss1 ] ) . using the tomographic entropy ( [ ss3 ] ) , one can calculate the entropic uncertainty function ( [ ss4 ] ) for the soliton numerically .
the numerical result was obtained by speeding up the calculation procedure employing fast fourier transform ( fft ) method .
indeed , it was shown ( see @xcite ) that tomogram can be expressed as convolution of the initial profile with a chirp function ( cf ) . here
the convolution was computed via fft , namely , the inverse fourier transform of the product of fft of the initial profile and fft of cf .
plots in fig.[figure2 ] demonstrate the behaviour of entropic uncertainty function @xmath165 for different values of the soliton - width parameter @xmath145 .
one can see that the upper bound of this function depends on the soliton width .
_ plot of entropic uncertainty function @xmath166 for three values of the soliton - width parameter @xmath167 , @xmath168 , and @xmath169 . _ ] before concluding this section , we consider other examples of quantum states with generic gaussian wigner function and the corresponding tomogram @xmath170 where @xmath171 the parameters @xmath172 , @xmath173 , and @xmath174 satisfy the uncertainty relation @xmath175 the state under consideration for @xmath176 is the oscillator quantum thermal state with temperature @xmath177 . for the state ( [ in1 ] )
, the entropic uncertainty function reads @xmath178\nonumber\\ & & \quad + \frac{1}{2}\ln\left[\sigma_{qq}\sin^2t+\sigma_{pp}\cos^2t- 2\sigma_{qp}\sin t\cos t\right]\,.\end{aligned}\ ] ] for squeezed thermal state we have @xmath179 where @xmath180 is squeezing parameter .
to conclude , we point out the main results of this paper . inequalities ( [ t46 ] ) and ( [ t52 ] ) being the generalizations of known entropic inequalities for probability distributions of conjugate position and momentum are obtained for entropies associated with symplectic tomograms .
the new uncertainty relations obtained characterize the behavior of quantum state in quantum mechanics as well as the behavior of analytic signal in signal analysis .
the entropic uncertainty relation for tomographic entropy are obtained also for multimode quantum state .
the uncertainty relation is given by formula ( [ t53 ] ) .
the entropy under study as any shannon entropy provides the informational characteristics of the signal .
the uncertainty relation for tomographic entropies is a new additional property of nonlinear signals including bec solitons obeying the gross pitaevskii equation .
the physical meaning of tomographic entropic uncertainty relations will be deepen in a future work .
v. v. dodonov and v. i. manko , _ invariants and the evolution of nonstationary quantum systems _ ,
_ proceedings of the lebedev physical institute _ , nauka , moscow ( 1987 ) , vol .
183 [ translated by nova science , new york , 1989 ] .
s. de nicola , r. fedele , m. a. manko , and v. i. manko , `` wigner picture and tomographic representation of envelope solitons '' in : _ proceedings of international workshop ` nonlinear physics : theory and experiment .
ii ' _ ( gallipoli , lecce , italy , june
july 2002 ) , eds .
m. ablowitz , m. boiti , f. pempinelli , and b. prinari , world scientific , singapore ( 2003 ) , p. 372 ; _
laser res .
_ , * 25 * , 1 ( 2004 ) ; `` quantum tomography , wave packets and solitons '' in : _ symmetries in science xi _ , eds .
b j gruber , g. marmo , and n. yoshinaga , kluwer academic publishers , london ( 2004 ) , p. 175 .
m. a. manko , s. de nicola , r. fedele , and v. i. manko , `` symplectic group and solitons in bec as probability distributions corresponding to nonlinear signals '' in : g. s. pogosyan , l. e. vicent , and k .- b .
wolf ( eds . ) , _ group theoretical methods in physics , proceedings of the xxv international colloquium on group theoretical methods in physics _
( cocoyoc , mexico , 26 august 2004 ) , institute of physics publishers , bristol ( 2005 ) , conference series number 185 , pp . 385390 . | using the tomographic probability distribution ( symplectic tomogram ) describing the quantum state ( instead of the wave function or density matrix ) and properties of recently introduced tomographic entropy associated with the probability distribution , the new uncertainty relation for the tomographic entropy is obtained .
examples of the entropic uncertainty relation for squeezed states and solitons of the bose
einstein condensate are considered .
pacs number(s ) : 42.50.-p , 42.50.dv , 03.67.-a to be published in european physical journal b |
the famous supernova sn1987a in the large magellanic cloud ( lmc ) brought the field of supernova neutrino astrophysics to life .
two water cherenkov detectors , kamiokande ii and imb , detected 20 events between them @xcite ; two scintillator detectors , baksan and lsd @xcite also reported observations .
the sparse sn1987a neutrino data were sufficient to confirm the baseline model of gravitational collapse causing type ii sne and to put limits on neutrino properties ( such as a @xmath0 mass limit of around 20 ev . ) to make distinctions between different theoretical models of core collapse and supernova explosions and to extract more information about neutrino properties , we await the more copious neutrino signal which the new generation of large neutrino experiments will detect from the next such event in our galaxy . when the core of a massive star at the end of its life collapses , less than 1% of the gravitational binding energy of the neutron star will be released in the forms of optically visible radiation and the kinetic energy of the expanding remnant .
the remainder of the binding energy is radiated in neutrinos , of which @xmath11% will be electron neutrinos from an initial `` neutronization '' burst and the remaining 99% will be neutrinos from the later cooling reactions , roughly equally distributed among flavors .
average neutrino energies are expected to be about 13 - 14 mev for @xmath2 , 14 - 16 mev for @xmath0 , and 20 - 21 mev for all other flavors .
the neutrinos are emitted over a total timescale of tens of seconds , with about half emitted during the first 1 - 2 seconds .
reference @xcite summarizes the expected features of a core collapse neutrino signal ; more recent simulation work can be found in _
e.g_. @xcite .
a core - collapse supernova in our galaxy will bring a wealth of scientific information .
the neutrino signal will provide information about the properties of neutrinos themselves and astrophysicists will learn about the nature of the core collapse .
one unique feature of the neutrino signal is that it is _ prompt _ neutrinos emerge on a timescale of tens of seconds , while the first electromagnetic signal may be hours or days after the stellar collapse .
therefore , neutrino observation can provide an _ early alert _ that could allow astronomers a chance to make unprecedented observations of the very early turn - on of the supernova light curve ; even observations of sne as young as a few days are rare for extra - galactic supernovae .
the environment immediately surrounding the progenitor star is probed by the initial stages of the supernova .
for example , any effects of a close binary companion upon the blast would occur very soon after shock breakout .
uv and soft x - ray flashes are predicted at very early times .
finally , there may be entirely unexpected effects no supernova has ever been observed very soon after its birth . although the neutrino signal will be plentiful in practically all galactic core collapses , it is possible that there will be little or no optical fireworks ( the supernova `` fizzles '' ) ; the nature of any observable remnant would then be very interesting .
this paper focuses on the prompt alert which is possible using the neutrino signal .
we will describe the technical aspects of the system .
section [ overview ] gives an overview of snews , and section [ signal ] briefly covers the expected signal in current detectors .
section [ 3ps ] discusses some issues associated with snews .
section [ indiv ] introduces the individual experiments monitors .
section [ implementation ] covers snews implementation and defines the coincidence conditions and alert scheme .
section [ highrate ] describes the results of the `` high - rate '' system test performed in 2001 .
section [ alert ] describes the alert to the astronomical community .
section [ future ] gives future directions .
the final section summarizes .
the snews ( supernova early warning system ) collaboration is an international group of experimenters from several supernova neutrino - sensitive experiments .
the primary goal of snews is to provide the astronomical community with a prompt alert for a galactic supernova .
an additional goal is to optimize global sensitivity to supernova neutrino physics , by such cooperative work as downtime coordination . the idea of a blind central coincidence computer receiving signals from several experiments has been around for some time ( _ e.g. _ @xcite . )
in addition to the basic early warning advantages of a neutrino detector , there are several benefits from a system involving neutrino signals from two or more different detectors .
first , if the supernova is distant and only weak signals are recorded , a coincidence between signals from different detectors effectively increases the sensitivity by allowing reductions in alarm thresholds and allowing one to impose a minimum of ( possibly model - dependent ) expectations on the form of the signal .
second , even if a highly sensitive detector such as super - k is online , _
_ requiring a coincidence among several detectors effectively reduces the `` non - poissonian '' background present for any given detector and enormously increases the confidence in an alert . _
_ background alarms at widely separated laboratories are highly unlikely to be correlated . without the additional confidence from coincident neutrino observations
, it would be very difficult for any individual detector to provide an _ automated _ alert to astronomers .
finally , using signals from more than one detector , there is some possibility for determining the direction of the source when a single detector alone can provide no information ( see reference @xcite . )
unfortunately triangulation is in practice quite difficult to do promptly , and can not point as well as individual detectors .
an important question for snews is : how often is a galactic supernova likely to occur ?
estimates vary widely , but are typically in the range of about one per 30 years ( _ e.g. _ @xcite . )
this is frequent enough to have a reasonable hope of observing one during the next five or ten years , but rare enough to mean that we must take special care not to miss anything when one occurs .
the charter member experiments of snews are super - kamiokande ( super - k ) in japan , the sudbury neutrino observatory ( sno ) in canada and the large volume detector ( lvd ) in italy .
representatives from amanda , icecube , kamland , borexino , mini - boone , icarus , omnis , and ligo participate in the snews working group , and we hope will eventually join the active coincidence . there is currently a single coincidence server , hosted by brookhaven national laboratory .
we expect that additional machines will be deployed in the future .
the bnl computer continuously runs a coincidence server process , which waits for alarm datagrams from the experiments clients , and provides an alert if there is a coincidence within a specified time window ( 10 seconds for normal running . )
we have implemented a scheme of `` gold '' and `` silver '' alerts : gold alerts are intended for automated dissemination to the community ; silver alerts will be disseminated among the experimenters , and require human checking . as of this writing , no inter - experiment coincidence , real or accidental , has ever occurred ( except in high rate test mode ) , nor has any core collapse event been detected within the lifetimes of the currently active experiments .
there are several classes of detectors capable of observing neutrinos from gravitational collapse .
most supernova neutrino detectors are designed primarily for other purposes , _ e.g. _ for proton decay searches , solar and atmospheric neutrino physics , accelerator neutrino oscillation studies , and high energy neutrino source searches .
.supernova neutrino detector types and their primary capabilities.[tab : detector_types ] [ cols="^,^,^,^,^,^",options="header " , ] although these somewhat non - stationary data , taken at lowered threshold , do not necessarily imply that rates will also be non - stationary when thresholds are raised and running conditions are normal , one can never be completely sure that individual experiment rates will not increase unexpectedly .
this is the motivation for the rate - dependent gold suppression scheme of section [ suppression ] .
the coincidence server now has capability for continuous high rate testing , using tagged test alarms in parallel with normal alarms .
at the supernova early alert workshop of 1998 , the conclusion from the astronomer working group @xcite was that `` the message will spread itself '' and that snews will need to do no more than send out emails to as many astronomers as possible .
snews maintains a mailing list of interested parties , including both professionals and amateurs , to be alerted in the case of a coincidence . in an ideal case
, the coincidence network provides the astronomical community with an event time and an error box on the sky at which interested observers could point their instruments . in a realistic case ,
the size of the error box is dependent on the location of the supernova and the experiments which are online , and may be very large ( and at this time will not be available in the initial alert message . )
however , members of the mailing list with wide - angle viewing capability ( satellites , small telescopes ) should be able to pinpoint an optical event quickly .
although an unknown fraction of galactic supernovae will be obscured by dust , many will be visible to amateurs with modest equipment .
regardless of the quality of neutrino pointing available , however , the advance warning alone gives observers of all kinds valuable time to get to their observatories and prepare to gather data as soon as an accurate position is determined .
a target of opportunity proposal for the hubble space telescope , `` observing the next nearby supernova '' , aiming to take advantage of early supernova light based on an early warning , was approved @xcite for cycle 13 and was operational for cycles 8 through 12 .
the large pool of skilled and well - equipped amateur astronomers is also prepared to help locate a nearby supernova .
the editors of _ sky & telescope _ magazine have set up a clearinghouse for amateur observers in search for first light ( and a precise optical position as early as possible ) @xcite , via their astroalert service @xcite .
this was started by former editor - in - chief leif robinson , and has the continued support of current editor - in - chief rick fienberg . in collaboration with the american association of variable star observers ,
they have developed a set of criteria for evaluating amateur responses to an alert , so that a reliable precise position can be disseminated as early as possible .
for instance : there must be at least two consistent reports , demonstrated lack of motion , lack of identification with known asteroid and variable star databases , variability consistent with supernova light curves and , if the information is available , a spectrum consistent with known supernova types . on february 14 2003 , _ sky & telescope _
performed a test for amateurs . a transient target ( the asteroid vesta at a near - stationary point in its retrograde loop )
was selected , which at the time was about magnitude 6.7 .
_ sky & telescope _ issued an alert ( very carefully tagged as a test ) to their mailing list , with a given 13-degree uncertainty radius .
they received 83 responses via the web response form , and more by email .
the responses were of world - wide distribution , and although many observers experienced poor conditions , six were successful in identifying the target . from this experience , they have suggested refinements to optimize amateur astronomer strategy .
a second test is planned soon , and should be a regular occurrence .
we maintain two alert mailing lists which will be sent to automatically by the snews coincidence software in the case of an alert .
the first is the gold alert list , which includes all astronomers who have signed up , including _ sky & telescope _ and the hst astronomers , and is to be an _ automated _ alert . the second mailing list will be for silver alerts , and is to be sent to neutrino experimenters only .
these alerts will be checked out by shiftworkers at their respective experiments before an alert is issued ; each experiment is responsible for making sure the silver alert messages reach shiftworkers .
each experimental collaboration defines its own protocol for acting on a snews silver or gold alert . for both silver and gold cases , a message containing the following information : * utc time of the coincidence , * all detectors involved in the coincidence , and * the types of alarms ( good , possible ) for each experiment involved in the coincidence will be automatically sent by the server to the snews subgroup members .
the information may also be posted to a restricted snews subgroup page for silver , and a public page for gold . to allow the confirmation of a snews alert as really coming from snews
, any alerts will be public key signed using the snews key .
this key has the i d # 68df93f7 , and is available on the network of public pgp keyservers such as ` http://pgp.mit.edu/ ` note that there is no restriction on individual experiments making any announcement based on individual observation in the case of absence of a snews alert , silver or gold , or preceding or following any snews alert message . any individual experiment may publicly announce a supposed supernova signal following a dispatched silver alert which has not yet been upgraded to gold . in this case the information that a previous silver alert from the snews server(s ) has been received should be cited .
at the time of this writing , silver alerts only between super - k and lvd are activated .
we are working towards having the operational mode described in this paper to be activated in the very short term , comprising automated good alarms from super - k and lvd , but automated possible alarms only from sno , such that sno will participate in a gold alert only if at least two other experiments good alarms are present .
we also expect snews to incorporate more galactic - supernova - sensitive neutrino detectors over the next few years .
in addition , we may expand the network of servers with additional secure sites .
in summary , several supernova neutrino detectors are now online . if a stellar core collapse occurs in our galaxy , these detectors will record signals from which a wealth of physical and astrophysical information can be mined .
an early alert of a gravitational collapse occurrence is essential to give astronomers the best chance possible of observing the physically interesting and previously poorly observed early turn - on of the supernova light curve .
a coincidence of several neutrino experiments is a very powerful technique for reducing `` non - poissonian '' false alarms to the astronomical community , in order to allow a prompt alarm .
we have implemented such a system , currently incorporating several running detectors : lvd , sno and super - k .
we expect to expand the network in the near future , and move to a more automated mode in the near future . | this paper provides a technical description of the supernova early warning system ( snews ) , an international network of experiments with the goal of providing an early warning of a galactic supernova . |
the quantum electrodynamics ( qed ) theory of simple atoms like hydrogen or hydrogen - like ions provides precise predictions for different energy levels @xcite .
particularly , some accurate results were obtained for the lamb shift in the ground state of the hydrogen atom .
the accuracy of the qed calculations of the lamb shift has been limited by unknown higher - order two - loop corrections and inaccuracy of determination of the proton charge radius @xcite .
as far as the proton size is going to be determined very precisely from a new experiment @xcite on the lamb shift in muonic hydrogen , the only theoretical uncertainty is now due to the two - loop contribution .
improvement of the theory is important to determine the rydberg constant with high accuracy @xcite and to test the bound state qed precisely . since the theory seems not to be able to give now any results on higher - order two - loop corrections ( @xmath0 and higher ) we have to look for another way to estimate these terms and so the uncertainty of the hydrogen lamb shift theory .
an opportunity is to study the problem experimentally , measuring the lamb shift in different hydrogen - like ions at not too high value of the nuclear charge @xmath1 . only for two such ions the lamb shift can be available with a high accuracy from experiment at the present time or in the near future .
namely , these are helium @xcite and nitrogen @xcite ions .
the experimental estimation of higher - order two - loop terms is quite of interest also because of recent speculation on a great higher - order term @xcite ( see also refs .
@xcite ) .
the advantage of using @xmath2 is determined by the scaling behaviour of different qed values : * the scaling of the lamb shift is @xmath3 ; * the scaling of the radiative line width of excited states ( e. g. of the @xmath4 and @xmath5 states ) is @xmath3 as well ; * the scaling of the unknown higher - order two - loop corrections to the lamb shift is @xmath6 .
thus , relatively imprecise measurements with higher @xmath1 can nevertheless give some quite accurate data on some qed corrections .
our target is to develop a theory for the lamb shift and the fine structure in these two atomic systems .
eventually we need to determine the @xmath7 splitting in the helium ion ( for comparison with the experiment @xcite ) , difference of the lamb shifts @xmath8 in @xmath9he@xmath10 ( for the project @xcite ) and the @xmath11 interval in hydrogen - like nitrogen .
the difference mentioned is necessary @xcite if one needs to compare the results of the lamb shift ( @xmath12 ) measurement @xcite and the @xmath13 experiment . since the uncertainty of the qed calculations
is determined for these two ions ( he@xmath10 and n@xmath14 ) by the higher - order two - loop terms , we are going to reduce the other sources of uncertainty .
we present results appropriate to provide an interpretation of the experiments mentioned as a direct study of the higher - order two loop corrections .
the results of the ions experiments should afterwards be useful for the hydrogen atom .
the lamb shift is defined throughout the paper as a deviation from an unperturbed energy level . ] @xmath15 -\frac{m_r^2}{2(m+m)}\bigl[f(nj ) -1\bigr]^2 \,,\ ] ] where @xmath16 and @xmath17 are the mass of the nucleus and of the electron , and @xmath18 stands for the reduced mass .
the dimensionless dirac energy of the electron in the external coulomb field is of the form @xmath19 the lamb shift is mainly a qed effect , perturbed by the influence of the nuclear structure @xmath20 the qed contribution @xmath21 includes one- , two- and three - loop terms calculated within the external - field approximation @xmath22\,,\ ] ] and a recoil corrections @xmath23 , which is a sum of pure recoil and radiative recoil contributions depending on the nuclear mass @xmath16 @xmath24 let us start with the one - loop contribution .
the terms of the external - field contribution are usually written in the form of an expansion @xmath25 the dominant contribution comes from the one - loop self energy of the electron .
the known results for some low levels are summarized below . ] : @xmath26 where the state - dependent logarithmic coefficient @xmath27 is known @xmath28 the bethe logarithm is @xcite @xmath29 and higher - order self - energy terms @xmath30 are numerically found in sect .
2.5 . the coefficient of the expansion ( [ exp ] ) for free vacuum polarization can be calculated in any order of @xmath31 in a closed analytic form @xcite .
in particular , the result was found in ref .
@xcite for the circular states ( @xmath32 ) with @xmath33 . for
@xmath34 one can expand at low @xmath31 and find @xmath35 in the case of other states the result has been known only up to the order @xmath36 @xcite .
we present here new results for two other states at @xmath12 and for the @xmath5 state : @xmath37 it is not enough to consider the free vacuum polarization .
the relativistic corrections to the free vacuum polarization in eqs .
( [ relvp1][relvp4 ] ) are of the same order as the so - called wichmann - kroll term due to coulomb effects inside the electronic vacuum - polarization loop . to estimate this term we fitted its numerical values from ref .
@xcite , which are more accurate for some higher @xmath38 , by expression @xcite @xmath39 we found that the contribution of @xmath40 and @xmath41 terms is small enough .
the accuracy of the calculation in ref .
@xcite is not high at @xmath42 and we performed some fitting of higher @xmath1 data . to make a conservative estimate we find two pairs of coefficients which reproduce the result at @xmath43 .
the results are : @xmath44 where the uncertainty comes from inaccuracy of the numerical calculations of @xmath45 @xcite , which is estimated here as a value of a unit in the last digital place .
comparing the results above one can find a conservative estimate : @xmath46 .
the value @xmath45 is valid for both the @xmath47 and @xmath48 states and we use this value for the @xmath5 state as well . on this level of accuracy ( @xmath49 )
there is no shift of the @xmath4 levels @xcite and we use a zero value for them .
we separate from the expression for the self - energy part of the one - loop correction ( [ exp ] ) the function @xmath50 using numerical values of @xmath51 from refs . @xcite and ones of @xmath52 from refs .
@xcite , we performed several types of fitting for these functions .
we started with fitting _
( i ) _ with function @xmath53 minimizing the sum @xmath54 } { \delta g(z ) } \right\}^2 } \,,\ ] ] with respect to @xmath40 and @xmath41 for @xmath47 state ( where @xmath55 ) and to @xmath40 for @xmath48 state ( where @xmath56 ) . in the latter case we used the fact that @xmath57 .
the statistical error of data @xmath58 ^ 2 = \left[\delta_{\rm num }
g(z)\right]^2 + \left[\pi^2 ( z\alpha)^3 a^{(0)}_{70}\right]^2 \,.\ ] ] contains uncertainty of numerical integrations in refs . @xcite and of the fit in ( [ fitlow ] ) due to neglecting of higher - order terms of absolute order @xmath59 ( where @xmath60 is a result of preliminary fitting with @xmath61 ) . to estimate the additional systematic uncertainty which originates from the unknown term of order @xmath62 we studied a sensitivity of the fit _ ( i ) _ to introduction of some perturbation function @xmath63 @xmath64 the final uncertainty of the fit was calculated as a random sum of differences of the fits without function @xmath65 and with @xmath66 from the binomial expansion of the expression @xmath67 for the @xmath47 and @xmath48 states and @xmath68 for the @xmath4-states and for the difference @xmath69 ( see below ) .
the logarithm @xmath70 in the expansion ( [ exp ] ) is a large value ) .
] at very low @xmath1 but it is quite a smooth function of @xmath1 around @xmath42 ( see table [ tablog ] ) . due to that , we can also use a non - logarithmic fitting function @xmath71 with smooth behaviour at @xmath72 .
in particular , we applied the eq .
( [ ii_iii ] ) for numerical data from ref .
@xcite at @xmath73 with central value @xmath74 ( fit _ ii _ ) and @xmath75 with @xmath76 ( fit _ iii _ ) .
.function @xmath70 for small @xmath1 [ cols="^,^,^,^",options="header " , ]
we are grateful to ed myers for stimulating discussions .
the work was supported in part by rfbr ( grant 00 - 02 - 16718 ) and russian state program `` fundamental metrology '' . a support ( sk ) by a nato grant crg 960003 is also acknowledged . w. a. barker and f. n. glover , phys . rev . *
99 * , 317 ( 1955 ) ; + s. j. brodsky and r. g. parsons , phys
. rev . * 163 * , 134 ( 1967 ) ; * 176 * , 423e ( 1968 ) ; + k. pachucki and s. g. karshenboim , j. phys .
b * 28 * , l221 ( 1995 ) | calculation of higher - order two - loop corrections is now a limiting factor in development of the bound state qed theory of the lamb shift in the hydrogen atom and in precision determination of the rydberg constant .
progress in the study of light hydrogen - like ions of helium and nitrogen can be helpful to investigate these uncalculated terms experimentally . to do that it is necessary to develop a theory of such ions .
we present here a theoretical calculation for low energy levels of helium and nitrogen ions . |
an important problem in diophantine approximation is the study of the speed of approach to @xmath5 of a possibly inhomogeneous linear form of several variables evaluated at integers points .
such a linear form is given by @xmath6 @xmath7 more generally , one can consider @xmath1 linear forms for @xmath8 , corresponding to @xmath9 and @xmath10 , where each @xmath11 , is a vector in @xmath12 .
diophantine approximation theory classifies the matrices @xmath13 and vectors @xmath14 according to how `` resonant '' they are ; i.e. , how well the vector @xmath15 approximates @xmath16 as @xmath17 varies over a large ball in @xmath18 one can then fix a sequence of targets converging to @xmath5 , say intervals of radius @xmath19 centered at @xmath5 with @xmath20 , and investigate the integers for which the target is _ hit _ , namely the integers @xmath17 such that and @xmath21 $ ] for every @xmath22
. an important class of targets is given by radii following a power law , @xmath23 for some @xmath24 ( see for example @xcite or @xcite for a nice discussion related to the diophantine properties of linear forms ) .
fix a norm @xmath25 on @xmath26 , and let @xmath27 denote the euclidean norm on @xmath28 . for @xmath29 and @xmath30 or @xmath31 , we define sets @xmath32^r \subset \r^r\end{aligned}\ ] ] for @xmath30 and @xmath33 for @xmath34 .
we then introduce @xmath35 a matrix @xmath36 is said to be _ badly approximable _ if for some @xmath29 , the sequence @xmath37 is bounded . by contrast , matrices @xmath13 for which @xmath38 is unbounded , where @xmath39 is defined as @xmath40 , but with radii @xmath41 instead of @xmath42 are called _ very well approximable _ or vwa .
the obvious direction of the borel - cantelli lemma implies that almost every @xmath43 is not very well approximable ( cf .
the celebrated khintchine - groshev theorem on diophantine approximation implies that badly approximable matrices are also of zero measure @xcite .
analogous definitions apply in the inhomogeneous case of @xmath44 , and similar results hold . for targets given by a power law ,
the radii @xmath42 are thus the smallest ones to yield an infinite number of hits almost surely .
a natural question is then to investigate statistics of these hits , which we call _ resonances_. in the present paper we address in this context the behavior of the resonances on average over @xmath13 and @xmath14 , or on average over @xmath13 while @xmath14 is fixed at @xmath45 or fixed at random . introduce the `` expected '' number of hits @xmath46 when @xmath14 and @xmath13 are uniformly distributed on corresponding tori .
let @xmath47 denote the normal distribution with mean @xmath48 and variance @xmath49 .
[ thclt - u ] suppose @xmath50 and let @xmath13 be uniformly distributed on @xmath51 .
then , @xmath52 converges in distribution to @xmath53 as @xmath54 , where @xmath55 where @xmath56 is the unit ball in @xmath25-norm and vol denotes the euclidean volume .
the restriction @xmath57 above is necessary .
in fact , it is shown in @xcite using continued fractions that in that case the central limit theorem ( clt ) still holds for @xmath58 , but the correct normalization should be @xmath59 rather than @xmath60 [ thclt - v ] let @xmath13 and @xmath14 be uniformly distributed in @xmath61 .
then , @xmath62 converges in distribution to @xmath63 as @xmath54 .
the preceding theorems give clts in the cases of @xmath14 fixed to be @xmath5 or @xmath14 random .
the clt also holds for for almost every fixed @xmath14 .
[ thclt - v - as ] suppose @xmath64 for almost every @xmath14 , if @xmath13 is uniformly distributed in @xmath61 , then @xmath65 converges in distribution to a normal random variable with zero mean and variance one .
using a by now standard approach of dani correspondence ( cf .
@xcite ) we deduce our results about diophantine approximations from appropriate limit theorems for homogeneous flows .
namely we need to prove a clt for siegel transforms of piecewise smooth functions ; these limit theorems are formulated in section [ sclat ] .
the reduction of the theorems of section [ scintro ] to those o section [ sclat ] is given in section [ scda - lat ] .
the clts in the space of lattices are in turn deduced from an abstract central limit theorem ( theorem [ thcltabs ] ) for weakly dependent random variables which is formulated and proven in section [ scabstract ] . in order to verify the conditions of theorem [ thcltabs ] for the problem at hand we need several results about regularity of siegel transforms which are formulated in section [ sch - sl ] and proven in the appendix . in section [ sec.proofs ]
we deduce our central limit theorems for homogeneous flows from the abstract theorem [ thcltabs ] .
section [ sec.variances ] contains the proof of the formula on the variances .
section [ scgengroup ] discusses some applications of theorem [ thcltabs ] beyond the subject of diophantine approximation .
d. d. was partially supported by the nsf .
i.v . appreciates the support of fondation sciences mathmatiques de paris during his stay in paris .
we let @xmath66 the multiplication rule in @xmath67 takes form @xmath68 . we regard @xmath69 as a subgroup of @xmath67 consisting of elements of the form @xmath70 . we let @xmath71 be the abelian subgroup of @xmath69 consisting of matrices @xmath72 and @xmath73 be the abelian subgroup of @xmath67 consisting of matrices @xmath74 where @xmath5 is an origin in @xmath75 @xmath76 is an @xmath1-dimensional vector ,
@xmath13 is a @xmath77 matrix and @xmath78 let @xmath79 be the space of @xmath80 dimensional unimodular lattices and @xmath81 be the space of @xmath80 dimensional unimodular affine lattices .
we identify @xmath79 and @xmath81 respectively with @xmath82 and @xmath83 .
we will need spaces @xmath84 @xmath85 , and @xmath86 of functions which can be well approximated by smooth functions , given @xmath87 .
recall first that the space @xmath88 consists of functions @xmath89 whose derivatives up to order @xmath90 are bounded . to define spaces @xmath91 and @xmath92 , fix bases for @xmath93 and @xmath94 ; then , @xmath91 and @xmath92 consist of functions whose derivatives corresponding to monomials of order up to @xmath90 in the basis elements are bounded ( see appendix for formal definitions ) .
now we define @xmath95-norm on a space equipped with a @xmath96-norm by @xmath97 some properties of these spaces are discussed in the appendix . given a function @xmath98 on @xmath99 we consider its siegel transforms @xmath100 and @xmath101 defined by @xmath102 we emphasize that siegel transforms of smooth functions are never bounded but the growth of their norms at infinity is well understood , see subsection [ sec.truncation ] . in this section we present general central limit theorems for siegel transforms . the reduction of theorems [ thclt - u ] , [ thclt - v ] , [ thclt - v - as ] to the results stated here will be performed in section [ scda - lat ] .
let @xmath103 be a non - negative function supported on a compact set which does not contain @xmath5 .
( the assumption that @xmath98 vanishes at zero is needed to simplify the formulas for the moments of its siegel transform .
see proposition [ lmrogers ] . )
denote @xmath104 we say that a subset @xmath105 is _
@xmath106-regular _ if @xmath107 is a union of codimension 1 submanifolds and there is a one - parameter subgroup @xmath108 such that @xmath109 } \cl\cap
s\neq \emptyset)\leq k\eps^\alpha.\ ] ] we say that a function @xmath110 is _
@xmath106-regular _ if @xmath111 has a @xmath106-regular boundary and the restriction of @xmath112 on @xmath111 belongs to @xmath113 with @xmath114 @xmath106-regular functions on @xmath81 are defined similarly .
let @xmath115 be subgroup of @xmath69 consisting of diagonal matrices .
we use the notation @xmath116 for haar measure on @xmath115 .
we say that @xmath112 is _ @xmath117-centrally smoothable _ if there is a positive function @xmath118 supported in a unit neighborhood of the identity in @xmath115 such that @xmath119 and @xmath120 has @xmath121 norm less than @xmath122 we say that a function @xmath112 on @xmath81 is _ @xmath117-centrally smoothable _ if @xmath123 is a @xmath117-centrally smoothable function on @xmath124 as before , we write @xmath125 for the normal distribution with @xmath48 and variance @xmath49 and `` @xmath126 '' stands for convergence in distribution .
we write @xmath127 for a certain diagonal element of @xmath69 and @xmath67 formally introduced in .
[ thclt - u - lat ] suppose that @xmath64 ( a ) there is a constant @xmath128 such that if @xmath129 is uniformly distributed on @xmath79 then @xmath130 as @xmath54
. \(b ) fix constants @xmath131 with @xmath132 suppose that @xmath129 is distributed according to a density @xmath133 which is @xmath134-regular and @xmath135-centrally smoothable .
then @xmath136 as @xmath54 . [ thclt - v - lat ] ( a ) let @xmath137 be uniformly distributed on @xmath138 then there is a constant @xmath139 such that @xmath140 as @xmath54 .
\(b ) fix constants @xmath131 with @xmath132 suppose that @xmath137 is distributed according to a density @xmath133 which is @xmath134-regular and @xmath135-centrally smoothable .
then @xmath141 as @xmath54 .
let @xmath142 be an unstable rectangle , that is @xmath143 where @xmath144 is a fixed affine lattice and @xmath145 and @xmath146 are rectangles in @xmath147 and @xmath28 respectively .
consider a partition @xmath148 of @xmath142 into @xmath71-rectangles .
thus elements of @xmath148 are of the form @xmath149 for some fixed @xmath150 [ thclt - v - lat - as ] suppose that @xmath64 then for each unstable cube @xmath142 and for almost every @xmath151 , if @xmath137 is uniformly distributed in @xmath152 , then @xmath153 as @xmath54
. the explicit calculation of @xmath128 and @xmath139 when @xmath98 is an indicator functions ( the case needed for theorems [ thclt - u]-[thclt - v - as ] ) will be given in section [ sec.variances ] .
central limit theorems for partially hyperbolic translations on homogeneous spaces are proven in @xcite ( for bounded observables ) and in @xcite ( for @xmath154 observables ) .
( see also @xcite for important special cases ) .
it seems possible to prove theorem [ thclt - u - lat ] for sufficiently large values of @xmath80 by verifying the conditions of @xcite .
instead , we prefer to present in the next section an abstract result which will later be applied to derive theorems [ thclt - u - lat ] , [ thclt - v - lat ] , and [ thclt - v - lat - as ] .
we chose this approach for three reasons .
first , this will make the paper self contained .
second , we replace the @xmath154 assumption of @xcite by a weaker @xmath155 assumption which is important for small @xmath156 third , our approach allows to give a unified proof for theorems [ thclt - u - lat ] , [ thclt - v - lat ] , and [ thclt - v - lat - as ] .
in this section we reduce theorem [ thclt - u ] , [ thclt - v ] , and [ thclt - v - as ] to theorems [ thclt - u - lat][thclt - v - lat - as ] . to fix our notation
we consider @xmath157 and @xmath158 the analyis of @xmath159 and @xmath160 being similar .
we also drop the extra subscript and write @xmath157 and @xmath161 as @xmath58 and @xmath162 , respectively , until the end of this section . in subsections [ sec.dani ] and [ sec.measure ]
we explain how to reduce theorem [ thclt - u ] to theorem [ thclt - u - lat ] .
the reductions of theorems [ thclt - v ] and [ thclt - v - as ] to theorems [ thclt - v - lat ] and [ thclt - v - lat - as ] require only minor modifications which will be detailed in section [ ssred - v ] . in this subsection we use the dani correspondence principle to reduce the problem to a clt for the action of diagonal elements on the space of lattices of the form @xmath163 where @xmath13 is random . let @xmath13 be the matrix with rows @xmath164 , @xmath165 .
for @xmath166 and @xmath167 , we denote the @xmath168 diagonal matrix @xmath169 we then consider the following matrices @xmath170 let @xmath118 be the indicator of the set @xmath171 , |x|^{d / r } y_j \in [ 0,c ] \text { for } j=1,\dots , r\right\}\ ] ] and consider its siegel transform @xmath172 .
now @xmath173 with @xmath174 contributes to @xmath175 ( from ) precisely when there exists @xmath176 such that @xmath177 clearly such a vector @xmath178 is unique .
it is elementary to see that holds if and only if @xmath179 for some integer @xmath180}$ ] . from
we obtain [ lmred1 ] for each @xmath181 @xmath182}^{[\log_2 n ] } \phi\circ g^t(\lambda_{{\mathbf a}}\z^{d+r})+r_n,\ ] ] with @xmath183 and @xmath184 denoting the @xmath185-norm with respect to the lebesgue measure on the unit cube in @xmath186 from , it follows that @xmath187 } \leq |n|<n,\\ \text{or } |n|\leq 2^{\log_2^\eps n } \end{array } : ( l_1(n,0,\a_1 ) , \dots , l_r(n,0,\a_r ) ) \in b(n , d , r , c ) \right\ } .\ ] ] note that for a fixed @xmath188 and @xmath189 , the form @xmath190 is uniformly distributed on the circle . hence @xmath191 and so @xmath192}\le |n|\le n } \frac{1}{|n|^d}+\sum_{|n|\le 2^{\log_2^\eps n } } \frac{1}{|n|^d } \ll ( \log_2 n)^\eps.\ ] ] hence ,
to prove theorem [ thclt - u ] we can replace @xmath175 by @xmath193}^{[\log_2 n ] } \phi\circ g^t(\lambda_{{\mathbf a}}\z^{d+r}).\ ] ] note that the action of @xmath194 on the space of lattices @xmath79 is partially hyperbolic and its unstable manifolds are orbits of the action @xmath195 with @xmath196 ranging in the set of @xmath77 matrices .
this will allow us to reduce the proof of clts to clts for the diagonal action on the space of lattices .
a similar reduction is possible for the @xmath194-action on the space of affine lattices since in that case the unstable manifolds for the action of @xmath194 are given by @xmath197 , with @xmath196 , @xmath198 let @xmath199 where @xmath200 $ ] . for @xmath201 let @xmath202 be independent uniformly distributed in @xmath203 $ ] . also introduce a random matrix @xmath204 where all the entries of @xmath205 are independent uniformly distributed in @xmath206 $ ] .
let @xmath207 and @xmath208 let @xmath209 .
it is clear that if @xmath13 is distributed uniformly in a unit cube , then @xmath210 is distributed according to a @xmath211-regular and @xmath135-centrally smoothable density .
note that @xmath212 observe also that for @xmath213 and @xmath214 , we have @xmath215 and hence for @xmath216 we have @xmath217 where @xmath218 denotes a @xmath219 neighborhood of the boundary of @xmath220 now the same argument as in lemma [ lmred1 ] gives [ lmred2 ] for each @xmath181 , @xmath193}^{[\log_2 n ] } \phi\circ
g^t(\lambda_{{\mathbf a}}\z^{d+r})= \sum_{t=[(\log_2 n)^\eps]}^{[\log_2 n ] } \phi\circ g^t(\tilde{\lambda}({{\mathbf a } } , { { \mathbf b}},{{\mathbf t } } ) \z^{d+r } ) + \tilde{r}_n\ ] ] with @xmath221 now theorem [ thclt - u ] follows from theorem [ thclt - u - lat ] except for the formula for @xmath128 which is derived in section [ sec.variances ] . the reduction of theorems [ thclt - v ] and [ thclt - v - as ] to theorems [ thclt - v - lat ] and [ thclt - v - lat - as ] requires only small changes compared to the preceding section . to wit ,
lemmas [ lmred1 ] and [ lmred2 ] take the following form .
[ lmredv ] ( a ) for each @xmath181 , @xmath222}^{[\log_2 n ] } \tcs(\one_{e_c})(g^t(\lambda_{{\mathbf a}}\z^{d+r}+(0 , { { \bf x } } ) ) ) + r_n\ ] ] where @xmath223 \(b ) let @xmath205 and @xmath224 have the same distribution as in subsection [ sec.measure ] and @xmath225 be uniformly distributed in @xmath226^d$ ] then for each @xmath181 @xmath227}^{[\log_2 n ] } \tcs(\one_{e_c})(g^t(\lambda_{{\mathbf a}}\z^{d+r}+(0 , { { \bf x } } ) ) ) \\
= \sum_{t=[(\log_2 n)^\eps]}^{[\log_2 n ] } \tcs(\one_{e_c})(g^t(\tlambda({{\mathbf a } } , { { \mathbf b } } , { { \mathbf t } } ) \z^{d+r}+({{\bf y } } , { { \bf x } } ) ) ) + \tilde{r}_n \end{gathered}\ ] ] with @xmath228 note that in part ( a ) the error has small @xmath184 norm for each fixed @xmath229 this follows from the fact that for each @xmath14 and @xmath17 , @xmath230 is uniformly distributed on @xmath231 this is useful in the proof of theorem [ thclt - v - as ] since we want to have a control for each ( or at least , most ) @xmath229 we also note that part ( b ) is only needed for theorem [ thclt - v ] since in theorem [ thclt - v - lat - as ] we start with initial conditions supported on a positive codimension submanifold of @xmath138 ( one of the steps in the proof of theorem [ thclt - v - lat - as ] consists of fattening the support of the initial measure , see subsection [ scasclt ] , however lemma [ lmredv](b ) is not needed at the reduction stage ) .
in what follows @xmath232 , and @xmath233 are fixed constants . let @xmath234 be a sequence of random variables satisfying the following conditions .
write @xmath235 for the corresponding centered random variable . 1 . given any @xmath117 ,
there is a sequences @xmath236 of random variables such that 1 .
@xmath237 almost surely ; 2 .
@xmath238 @xmath239 2 .
there exists a filtration @xmath240 defined for @xmath241 such that for every @xmath242 there exists a variable @xmath243 that is @xmath244-measurable , @xmath245 almost surely , @xmath246 , and @xmath247 3 . for @xmath242
, there exists an event @xmath248 such that @xmath249 and for @xmath250 @xmath251 4 .
there exists a numerical sequence @xmath252 for @xmath253 such that for @xmath250 and @xmath254 , @xmath255 [ thcltabs ] ( a ) under conditions ( h1)-(h4 ) if @xmath256 is distributed according to @xmath257 then @xmath258 with @xmath259 \(b ) suppose conditions ( h1)(h4 ) are satisfied .
fix @xmath260 such that @xmath261 and set @xmath262 .
suppose that @xmath256 is distributed according to a measure @xmath263 which has a density @xmath264 satisfying ( d1 ) @xmath265 ( d2 ) for each @xmath17 there is an @xmath266-measurable density @xmath267 such that @xmath268 ( d3 ) for each @xmath269 , @xmath270 then , @xmath271 here we show that the normalized variance converges . under conditions ( h1)(h4 ) we have
that @xmath272 with @xmath128 as in .
first we record a property of cross - terms in the sum that lets us pass from @xmath234 to the truncated sequence .
we have @xmath273 indeed , we have @xmath274 now applying the cauchy - schwarz inequality followed by ( h1a ) and ( h1b ) , we arrive at the bound since @xmath233 . next , applying ( h3 ) and ( h4 ) with @xmath275 gives @xmath276 combing and we get @xmath277 take a small number @xmath278 and assume that @xmath279 @xmath280 taking @xmath281 we see that @xmath282 therefore for each @xmath283 , the following limits exist , @xmath284 and moreover @xmath285 next we claim that under the condition @xmath286 there exists @xmath287 such that @xmath288 indeed , @xmath289 using ( h1a ) and ( h2 ) , we get @xmath290 while ( h3 ) shows that @xmath291 proving . combining and we see that for @xmath117 satisfying we have @xmath292 take a small @xmath293 suppose first that @xmath294 then we can take @xmath295 and holds giving @xmath296 for some @xmath297 . in particular combining , and with @xmath298 @xmath299 we obtain @xmath300 also and show that under , @xmath301 choosing @xmath302 and assuming @xmath303 , @xmath304 to prove the lemma , we need to control the sum @xmath305.\ ] ] using for @xmath306 and using otherwise we see that @xmath307 combining this with we see that the limit in exists and moreover that @xmath308 for all the sequel , we fix @xmath309 and let @xmath310 .
let @xmath311 $ ] , @xmath312:=m_n$ ] .
denote @xmath313 so that @xmath314 we claim that @xmath315 converge to @xmath5 in probability ; it will therefore suffice to prove that @xmath316 to prove the claim , observe first that following the proof of , we have @xmath317 as for the second sum in , we note that @xmath318\ ] ] hence ( h1b ) gives @xmath319 therefore @xmath320 we turn now to the proof of .
we start by defining an exceptional set @xmath321 on which we will not be able to exploit the almost independence of @xmath322 from @xmath323 . the exact reasons for the definition of each condition on @xmath324 will appear during the proof .
let @xmath325 to be able to use ( h2)(h4 ) we let for @xmath326 @xmath327 next define for @xmath328 @xmath329 and @xmath330 for @xmath331 $ ] with @xmath332 define @xmath333 and let @xmath334 and @xmath335 : k'-k\geq n^{\eps^2/2 } } { \bar g}_{m , k , k'}.\ ] ] finally set @xmath336 observe that ( h2)(h4 ) show that @xmath337 the main step in the proof of the clt is the following [ lemma.clt3 ] for @xmath338 , @xmath339 where @xmath340 is uniform in @xmath341 .
to prove we note that @xmath342 where the last step uses that @xmath343 .
next , ( h3 ) implies that on @xmath324 @xmath344 to finish the proof of it suffices to show that for @xmath338 , @xmath345 let us estimate the individual terms in this sum . to fix our notation let us suppose that @xmath346 let @xmath347 be a large constant and consider two cases .
\(a ) @xmath348 in this case ( h4 ) and give @xmath349 \(b ) @xmath350 and hence @xmath351 then @xmath352 the second term is @xmath353 since @xmath354 for the first term use that @xmath355 to obtain @xmath356 so both @xmath357 and @xmath358 are negligible . combining the estimates of cases ( a ) and ( b ) we obtain completing the proof of . to finish the proof of part ( a ) it remains to derive the central limit theorem from . for @xmath359 set @xmath360 then @xmath361 and so @xmath362 therefore , @xmath363 iterating this recurrence relation @xmath364 times we obtain @xmath365 completing the proof of part ( a ) of theorem [ thcltabs ] . part ( b ) can be established by a similar argument and we just briefly describe the necessary changes .
let @xmath366 denote the expectation with respect to @xmath367 that is , @xmath368 to extend the proof of the central limit theorem to the setting of part ( b ) we need to prove and with @xmath366 instead of @xmath369 .
for we need to prove the analogues of and .
we claim the following .
first , still holds with @xmath366 instead of @xmath370 second , @xmath371 ( note that in contrast to the sum here starts with @xmath372 not @xmath373 ) third , @xmath374 where @xmath375 to prove note that @xmath376 so follows from ( d3 ) .
observe that once these three points of the claim are established , the rest of the proof of part ( b ) proceeds exactly as in part ( a ) . to obtain the the other two points of our claim we will need the following [ measurability ] there exists a set @xmath377 with @xmath378 such that for @xmath379 , for @xmath380 and for @xmath381 a bounded random variable we have that @xmath382 let @xmath381 be a bounded random variable , @xmath380 and @xmath256 be such that @xmath383 then @xmath384 we prove now that the set where fails has measure that is exponentially small in @xmath385 .
the proof consists of two steps .
first , it follows form ( d1 ) and ( d2 ) that @xmath386 so @xmath387 is exponentially small by markov s inequality .
second , @xmath388 and so @xmath389 is exponentially small due to .
lemma [ measurability ] , together with lemma [ lemma.clt3 ] , imply that holds for @xmath366 instead of @xmath369 provided that we decrease slightly the set @xmath324 to @xmath390 .
it remains to prove .
note that the argument used to establish and in fact gives for @xmath391 @xmath392 hence lemma [ measurability ] implies that for @xmath393 @xmath394 this estimate implies by direct summation .
the proof of theorem [ thcltabs ] is thus completed . in case @xmath395
are bounded , one can take @xmath396 @xmath397 so the conditions of theorem [ thcltabs ] simplify as follows . 1 .
there exists filtration @xmath398 such that for every @xmath242 there exists a bounded @xmath244-measurable random variable @xmath399 with @xmath400 such that @xmath401 2 . for @xmath242
, there exists @xmath248 such that @xmath249 and for @xmath250 @xmath402 3 . for @xmath250 and @xmath254 @xmath403 if @xmath395 is a bounded sequence satisfying @xmath404@xmath405
then @xmath406 converges as @xmath407 to a normal distribution with zero mean and variance @xmath408
in this section we use the abstract theorem [ thcltabs ] to prove theorem [ thclt - u - lat ] , [ thclt - v - lat ] , [ thclt - v - lat - as ] . for this , we just have to check ( h1)(h4 ) for the case where our probability space is @xmath79 equipped with the haar measure and @xmath409 @xmath410 , where @xmath103 is a positive function supported on a compact set which does not contain @xmath411 before we construct the filtrations and prove ( h1)(h4 ) for the sequence @xmath409 we recall and prove preliminary results about functions defined on the space of lattices , on siegel transforms , and on the action of diagonal matrices .
we will cover this in sections [ sec.rogers ] , [ sec.truncation ] , and [ sec.representative ] respectively
. then we will prove theorems [ thclt - u - lat ] , [ thclt - v - lat ] , [ thclt - v - lat - as ] in section [ sec.proofs ] . in section [ sec.variances ] , we compute the variances in the special case that interests us of @xmath98 being the characteristic function of @xmath412 given in .
this will finish the proof of theorems [ thclt - u ] , [ thclt - v ] , [ thclt - v - as ] .
[ lmrogers ] ( * ? ? ?
* theorems 3.15 and 3.16 ) , ( * ? ? ?
* appendix b ) suppose that @xmath413 is piecewise smooth function that is supported on a compact set which does not contain @xmath411 then , @xmath414 ^ 2 ( \cl ) \ ;
d\mu(\cl)= \left[\int_{\r^{d+r } } f({{\bf x } } ) d{{\bf x}}\right]^2\\&\hspace{5 cm } + \sum_{\overset{(p , q)\in \n^2}{gcd(p , q)=1 } } \int_{\r^{d+r } } \left[f(p{{\bf x } } ) f(q{{\bf x}})+f(p{{\bf x } } ) f(-q{{\bf x}})\right ] dx;\\ \intertext{suppose that $ f\colon \tcm \to \r$ is piecewise smooth function of compact support . then
, } & ( c ) \quad \int_{\tcm } \tcs(f)(\tcl ) \ ; d\mu(\tcl)=\int_{\r^{d+r } } f({{\bf x } } ) d{{\bf x } } ; \\ & ( d)\quad \int_{\tcm } \left[\tcs(f)\right]^2 ( \tcl ) \ ; d\mu(\tcl)= \left[\int_{\r^{d+r } } f({{\bf x } } ) d{{\bf x}}\right]^2+\int_{\r^{d+r } } f^2 ( { { \bf x } } ) d{{\bf x}}. \end{aligned}\ ] ] let @xmath415 be a one parameter subgroup of @xmath416 for example one can take the matrices with ones on the diagonal , an arbitrary number in the upper left corner and zeros elsewhere . note that @xmath417 the filtrations for which we will prove ( h2)(h4 ) for the sequence @xmath409 will consist of small arcs in the direction of the flow of @xmath415 .
the exponential mixing of the @xmath69-action will underly the equidistribution and independence properties that are stated in ( h1)(h4 ) .
we will need the notion of representative partitions that was already used in @xcite
. these will be partitions of @xmath79 whose elements are segments of @xmath415 orbits , whose pushforwards by @xmath418 will become rapidly equidistributed . to guarantee the filtration property
, we would ideally consider an increasing sequence of such partitions with pieces of size @xmath419 , @xmath420 .
however , such partitions with fixed size pieces do not exist because @xmath415 is weak mixing .
we overcome this technical difficulty due to the following observations : 1 .
rudolph s theorem ( see ( * ? ? ?
* section 11.4 ) ) shows that for each @xmath421 we can find a partition @xmath422 into @xmath415-orbits such that the length of each element is either @xmath423 or @xmath424 and if @xmath425 then @xmath426 : \cp_u(\cl ) \text { has length } l/\sqrt{2})\leq \breps.\ ] ] 2 .
given @xmath427 , it suffices to check the properties ( h2)(h4 ) away from a set of measure less than @xmath428 .
having fixed @xmath309 , we will therefore abuse notation and say that a partition is of size @xmath423 if @xmath429 in 1 ) is less than @xmath430 . in light of this , let @xmath422 be a partition of size 1 and @xmath431 be its sub - partition of size @xmath419 . due to
( 1 ) and ( 2 ) we can assume without loss of generality that for every fixed @xmath432 $ ] , the partitions @xmath433 form an increasing sequence and that as a consequence the sequence @xmath434 of @xmath128-algebras generated by @xmath435 forms a filtration .
fix a small constant @xmath436 . given a collection
@xmath437 a set of natural numbers @xmath438 , and a number @xmath423 , we call a partition @xmath422 of size @xmath423 is _ representative _ with respect to @xmath439 if for each @xmath440 and for each @xmath441 , @xmath442 the curve @xmath443 is of the form @xmath444 with @xmath445 $ ] , and we use the notation @xmath446 for the normalized integral @xmath447 . given a finite collection @xmath448 , @xmath438 , and @xmath449 , we let @xmath450 assume that @xmath451 .
then we have as in ( * ? ? ?
* proposition 7.1 ) [ prop.rep ] let @xmath452 $ ] be the set of @xmath453 such that @xmath454 is representative with respect to @xmath439
. then @xmath455 .
we quickly recall how proposition [ prop.rep ] can be deduced , exactly as in ( * ? ? ?
* proposition 7.1 ) , from the polynomial mixing of the unipotent flow @xmath415 .
indeed , assuming that @xmath456 , polynomial mixing implies implies that @xmath457 with @xmath458 thus for curves @xmath459 of the form @xmath460 with @xmath461 $ ] we get that @xmath462 with @xmath463 this implies that if we consider a partition @xmath422 of size @xmath423 and its corresponding shifted partitions @xmath454 , @xmath464 $ ] we get for the measure @xmath465}$ ] that @xmath466 : \left| \int_{\cp_u(\cl ) } a \right| > \ck_a
l^{-\kappa_0}\right ) \leq c l^{-\kappa_0}\ ] ] where @xmath467 denotes the piece of @xmath454 that goes through @xmath129 .
the claim of proposition [ prop.rep ] then follows by markov s inequality .
proposition [ prop.rep ] will be used in the next section to obtain a partition of @xmath79 into pieces of @xmath415 orbits satisfying the condition of theorem [ thcltabs ] .
we could also use a partition into whole @xmath71-orbits .
the proof of proposition [ prop.rep ] in that case would be simpler since we could use effective equidistribution of horospherical subgroups @xcite .
we prefer to use @xmath415 orbits instead since it allows us to give unified proofs of theorems [ thclt - u - lat ] , [ thclt - v - lat ] , and [ thclt - v - lat - as ] ( as well as theorem [ thgenaction ] in section [ scgengroup ] ) . in this section
we give some useful results on truncations of a siegel transform of a compactly supported function @xmath468 .
these bounds are essential to control the truncated @xmath236 that appear in the abstract clt of section [ scabstract ] .
we will leave all the proofs and constructions to appendix a. in particular , we will define @xmath469 , with the properties described in lemma [ crhs - cs ] below .
we will always use the following notation for @xmath470 ( or @xmath471 ) : @xmath472 in the sequel we will consider @xmath473 .
[ crhs - cs ] @xcite there exists a constant @xmath474 such that for each pair of integers @xmath475 and each @xmath347 there is a constant @xmath476 such that the following holds .
let @xmath98 be supported on @xmath477 in @xmath478 \(a ) if @xmath479 then @xmath480 \(b ) if @xmath481 then @xmath482 \(c ) if @xmath481 then @xmath483 [ prop.ha1 ] for every @xmath484 there exists @xmath485 such that @xmath470 or @xmath486 satisfy @xmath487 if @xmath488 , then @xmath486 satisfies @xmath489 in addition , the same inequalities hold if the expectation is considered with respect to a measure that has a @xmath135-centrally smoothable density .
recall that an @xmath71-rectangle is a set of the form @xmath490 where @xmath13 belongs to the box @xmath491 in @xmath492 also define @xmath493 by @xmath494 [ truncation.xfix ] for each @xmath495 there exists a constant @xmath485 such that for any box @xmath491 whose sides are longer than @xmath421 and which is contained in @xmath496^{dr}$ ] and for any @xmath137 with @xmath497 @xmath498 satisfies @xmath499 where @xmath500 is a restriction of the haar measure on @xmath501 to @xmath148 and @xmath502 is expectation with respect to @xmath503
we are ready now to prove theorem [ thclt - u - lat ] , [ thclt - v - lat ] , [ thclt - v - lat - as ] using theorem [ thcltabs ] . for @xmath98 as in the statement of theorem [ thclt - u - lat ] , recall that we defined @xmath410 and @xmath504 recall also the notation @xmath505 and @xmath506 .
we will now prove ( h1)(h4 ) for the sequence @xmath507 .
fix any @xmath508 .
property ( h1 ) follows from inequalities and of lemma [ prop.ha1 ] . the fact that ( h1 ) fails to hold when @xmath509 is the reason why the clt does not hold in this case .
we will use the notion of representative partitions of section [ sec.representative ] to construct the desired filtrations .
first of all , note that to prove ( h4 ) we need to deal with function of the form @xmath510 .
therefore we define for every @xmath511 the collection of functions @xmath512 fix constants @xmath513 and define for every @xmath514 the following collection of functions and sequences of integers @xmath515 next let @xmath422 be a partition of size 1 and @xmath431 be its subpartition of size @xmath516 by proposition [ prop.rep ] and [ crhs - cs ] there is @xmath453 such that for each @xmath517 , @xmath435 is representative with respect to the collections of integers and functions in .
let @xmath434 be the filtration of @xmath128-algebras generated by @xmath435 .
denote @xmath518 we claim that @xmath519 satisfies ( h1)(h4 ) with @xmath520 provided that @xmath521 is sufficiently close to 1 .
since ( h1 ) has been checked above it remains to verify ( h2)(h4 ) . if @xmath522 then ( h2 ) holds if we take @xmath453 sufficiently large . by lemma [ crhs - cs ]
there are functions @xmath523 such that @xmath524 then @xmath525 where @xmath526 and @xmath527 are defined analogously to @xmath528 and @xmath529 with @xmath530 replaced by @xmath531 since @xmath523 are lipschitz , we have @xmath532
so if @xmath533 and @xmath135 is sufficiently large then @xmath534 implies @xmath535 hence markov s inequality gives @xmath536 this proves ( h2 ) provided that @xmath453 is large enough and @xmath537 ( h3 ) follows from the definition of representative partition if @xmath538 while for @xmath539 , @xmath540 provided that @xmath521 is sufficiently close to 1 . likewise , if @xmath541 then ( h4 ) holds by the definition of the representative partition with @xmath542 if @xmath543 we consider two cases \(a ) @xmath544 and so @xmath545 in this case ( h4 ) trivially holds similarly to ( h3 ) .
\(b ) @xmath546 and so @xmath547 accordingly to establish ( h4 ) with @xmath548 it suffices to show that there is a constant @xmath549 such that @xmath550 we are going to show that follows from already established ( h1)(h3 ) .
the argument is similar to the proof of .
namely , denoting by @xmath551 we get @xmath552 ( h1a ) and ( h2 ) imply that @xmath553 next , @xmath554\big|\cf_l\right)\right|\ ] ] and ( h3 ) shows that the expected value of the rhs is @xmath555 now markov s inequality shows that @xmath556 combining the estimates of @xmath357 and @xmath358 we obtain .
having checked ( h1)(h4 ) , we have established theorem [ thclt - u - lat](a ) via theorem [ thcltabs](a ) . to prove theorem [ thclt - u - lat](b )
we just need to check condition ( d1)(d3 ) of theorem [ thcltabs](b ) for @xmath557 property ( d1 ) follows from @xmath134-regularity since @xmath558 to check ( d2 ) let @xmath559 then if @xmath560 } \cl\cap \partial(\supp(\rho))=\emptyset\ ] ] then @xmath133 is holder on the element of @xmath433 containing @xmath129 so @xmath561 so the exceptional set for ( d2 ) consists of points violating .
this set has a small measure since @xmath562 is @xmath134-regular
. finally ( d3 ) follows from inequalities and of lemma [ prop.ha1 ] applied to the centrally smoothable density @xmath133 .
the proof of theorem [ thclt - u - lat ] is thus complete . for @xmath98 as in the statement of theorem [ thclt - v - lat ]
we define @xmath486 and @xmath563 , with @xmath564 distributed according to haar measure .
we also use @xmath505 and @xmath506 . if @xmath57 the analysis is exactly the same as in the case of linear lattices .
if @xmath565 , and of lemma [ prop.ha1 ] are not sufficient anymore to prove property ( h1 ) , and we replace them by and .
the proof of properties ( h2)(h4 ) proceeds exactly as in the case of linear lattices .
theorem [ thclt - v - lat](a ) thus follows from theorem [ thcltabs](a ) .
the changes needed for theorem [ thclt - v - lat](b ) are the same as for for theorem [ thclt - u - lat](b ) .
observe that ( d3 ) and ( h1 ) hold since are valid if haar measure is replaced with measures having centrally - smoothable densities .
here we deduce theorem [ thclt - v - lat - as ] from a refined version of theorem [ thclt - v - lat ] . using
we conclude that it is sufficient to prove the central limit theorem for sums with a shorter range of summation , @xmath566 next , take a large constant @xmath567 and for @xmath568 let @xmath569 let @xmath570 we have a partition @xmath571 of @xmath572 into sets of the form @xmath573 [ lmfat ] if @xmath574 is sufficiently large large then there is a constant @xmath575 such that @xmath576 except possibly for a set of @xmath577 of measure @xmath578 first we use to replace @xmath471 by @xmath579 hence denoting @xmath580 we get functions @xmath581 such that @xmath582 denoting @xmath583 we claim that @xmath584 consider for example the case where @xmath585 the opposite case being similar
. then @xmath586 @xmath587 the second term can be estimated by @xmath588 proving our claim .
next if @xmath574 is sufficiently large then @xmath589 while @xmath590 now the lemma follows from markov s inequality .
lemma [ lmfat ] allows to reduce theorem [ thclt - v - lat - as ] to the following result .
[ thclt - v - lat - as - thick ] suppose that @xmath64 for each @xmath591 and each @xmath181 the following holds .
if @xmath592 denotes the uniform distribution on @xmath593 then @xmath594 the proof of theorem [ thclt - v - lat - as - thick ] is also very similar to the proof of theorem [ thclt - u - lat ] .
let us describe the necessary modifications .
the property ( h1 ) follows from lemma [ truncation.xfix ] instead of and . to define the required filtration of ( h2)(h4 ) we need to adapt proposition [ prop.rep ] as follows .
take @xmath595 going to @xmath5 sufficiently slowly , for example , @xmath596 we let @xmath422 be a partition into segments of @xmath415 orbits of size @xmath595 and @xmath431 the corresponding subpartitions of pieces with length @xmath597 we let @xmath433 be the translates by @xmath415 of these partitions and denote by @xmath598 the collection of pieces of @xmath433 which are contained in @xmath599 we say that @xmath577 is @xmath600-good if there exists @xmath601 $ ] such that for each @xmath602 , @xmath598 is representative with respect to the families .the proof of proposition [ prop.rep ] also shows the following . [ lmmostgood ] given @xmath603 , if we take @xmath604 in sufficiently large then @xmath605 on the other hand if @xmath577 is @xmath600-good then the filtration generated by the partitions @xmath598 satisfies ( h2)-(h4 ) .
theorem [ thclt - v - lat - as - thick ] thus follows from theorem [ thcltabs ] .
[ rkeff ] the argument given above does not tell us for which @xmath14 theorem [ thclt - v - as ] holds .
of course rational @xmath14 have to be excluded due to theorem [ thclt - u ] .
now a simple baire category argument shows that theorem [ thclt - v - as ] also fails for very liouvillian @xmath229 it is of interest to provide explicit diophantine conditions which are sufficient for theorem [ thclt - v - as ] .
the papers @xcite provide tools which may be useful in attacking this question .
the arguments of the previous section are by no means limited to @xmath606 in particular , we have the following result . [ thgenaction ]
let @xmath607 be a @xmath608 diffeomorphism , @xmath609 of a manifold @xmath610 and @xmath611 be a @xmath608 flow on that space .
suppose that both @xmath607 and @xmath611 preserve a probability measure @xmath612 and there exists @xmath29 such that @xmath613 assume that @xmath614 is polynomially mixing , that is , there exist positive numbers @xmath615 and @xmath117 such that if @xmath616 @xmath617 fix @xmath449 and let @xmath618 be a random variable uniformly distributed on @xmath619.$ ] let @xmath620 satisfy .
then for @xmath612 almost all @xmath621 @xmath622 converges as @xmath54 to a normal random @xmath623 variable with zero mean and variance @xmath624 moreover for each @xmath625 there is a constant @xmath135 such that @xmath626 the constant @xmath135 can be chosen uniformly when @xmath423 varies over an interval @xmath627 $ ] for some @xmath628 we note that @xmath610 need not be compact , so different @xmath608 norms on @xmath610 need not be equivalent . in the theorem [ thgenaction ] above we assume that the compositions with @xmath607 preserve @xmath608 norm and moreover @xmath629 the proof of theorem [ thgenaction ] is similar to but easier than the proof of theorem [ thclt - v - lat - as ] .
namely since @xmath620 is bounded we only need to check conditions @xmath404@xmath630 fix a partition @xmath148 of @xmath610 into @xmath611 orbit segments of size @xmath631 @xmath632 denote the partition of @xmath610 of the form @xmath633 which has @xmath634 as the boundary point where @xmath635 is the smallest positive number with this property . as in section [ scasclt ]
we let @xmath636 be the subpartition of @xmath637 into segments of size @xmath638 consider the following collections .
@xmath639\right)\ ] ] and @xmath640 we say that @xmath634 is @xmath600-good if for each @xmath641 the partition @xmath636 is representative with respect to families and .
lemma [ lmmostgood ] is easily extended to show that for each @xmath1 , @xmath642 provided that @xmath643 are large enough .
hence almost every @xmath634 is @xmath600-good for all sufficiently large @xmath644 next let @xmath645 be the filtration corresponding to @xmath646 if @xmath634 is @xmath600-good then @xmath647 satisfies the conditions @xmath404-@xmath405 which implies the clt in view of theorem [ thcltabs ] .
we note the following consequence of theorem [ thgenaction ] .
[ corgengroup ] let @xmath648 be a semisimple lie group without compact factors , and @xmath649 be an irreducible lattice .
let @xmath415 be a unipotent subgroup which is expanded by an element @xmath650 in the sense that @xmath651 for some @xmath29 .
fix @xmath449 and let @xmath618 be a random variable uniformly distributed on @xmath619.$ ] suppose that @xmath1 is large enough and let @xmath620 be a @xmath608 function of zero mean . then for haar almost all @xmath652 @xmath653 converges as @xmath54 to a normal random @xmath623 variable with zero mean and variance @xmath654 moreover for each @xmath625 there is a constant @xmath135 such that @xmath655 the constant @xmath135 can be chosen uniformly when @xmath423 varies over an interval @xmath627 $ ] for some @xmath628 corollary [ corgengroup ] follows from theorem [ thgenaction ] with @xmath656 and @xmath607 and @xmath614 actions of @xmath127 and @xmath657 respectively .
the polynomial mixing follows from ( * ? ? ?
* theorem 3.4 ) .
here we establish . we prove the formula for @xmath658 the computation for @xmath659 is the same . by
it suffices to compute @xmath660 - 1 } \cs(\one_{e_c } ) ( g^n \cl ) \right)\ ] ] where the variance is taken with respect to the haar measure on the space of lattices . note that @xmath661 - 1 } \cs(\one_{e_c } ) ( g^n \cl)$ ] can be replaced by @xmath662 where @xmath663 , |x|{^{d / r } } y_j\in [ 0 , c]\ } .\ ] ] now proposition [ lmrogers](b ) gives @xmath664 since @xmath665 , the last integral equals to @xmath666}(|px| ) \one_{[1 , n]}(|qx| ) \prod_j \left(\one_{[0 , c ] } ( p^{1+d / r}|x|{^{d / r } } y_j ) \one_{[0 , c ] } ( q^{1+d / r}|x|{^{d / r } } y_j ) \right ) dx dy\\ = \int_{\r^{d+r } } \one_{{[1/p , n / q]}}(|x| ) \prod_j \left(\one_{[0 , c ] } ( q^{1+d / r}|x|{^{d / r } } y_j ) \right ) dx dy={\frac{c^r}{q^{d+r}}\ ; \int_{\r^d } \one_{{[1/p , n / q]}}(|x| ) \frac{dx}{|x|^d } } \ ] ] to evaluate the last integral we pass to the polar coordinates @xmath667 where @xmath668 is a unit vector in the euclidean norm . then , @xmath669}}(|x| ) \frac{dx}{|x|^d}= \int_{\mathbb{s}^{d-1 } } ds \int_{1/p|s|}^{n / q|s| } \frac{\rho^{d-1 } d\rho}{|s|^d\rho^d}= \ln \left(\frac{np}{q}\right ) \ ; \int_{\mathbb{s}^{d-1 } } \frac{ds}{|s|^d } .\ ] ] the second factor here equals to @xmath670 therefore , @xmath671 where the last step relies on ( * ? ? ?
* theorem 288 ) .
here we compute the limiting variance for @xmath672 as in section [ ssvaru ] we consider the case of boxes , the computations for balls being similar .
the same computation as in section [ ssvaru ] shows that we need to compute @xmath673 where the variance is taken with respect to the haar measure on the space of affine lattices . by proposition [ lmrogers](d )
this variance equals to @xmath674 ^ 2 ( x , y ) dx dy= \int_{\r^{d+r}}\one_{e_c(n ) } ( x , y ) dx dy=\widehat{v}_n.\ ] ] the fact that the variance of @xmath162 has a simpler form than the variance of @xmath58 has the following explanation .
let @xmath675 so that @xmath676 then @xmath677 s are pairwise independent ( even though triples @xmath678 are strongly dependent ) and hence uncorrelated ( see e.g. @xcite ) while @xmath679 s are not pairwise independent .
for a fixed dimension @xmath166 , we denote by @xmath79 the space of @xmath680 dimensional lattices .
we let @xmath681 denote the space of smooth functions on @xmath79 .
let @xmath682 be a basis in the space of left invariant vector fields on @xmath124 we let @xmath683 the space @xmath684 of smooth functions on the space of @xmath1-dimensional affine lattices is defined similarly . without the loss of generality we may assume that @xmath691 suppose first that @xmath692 given @xmath693 let @xmath694 be the functions such that @xmath695 without the loss of generality we may assume that @xmath696 since otherwise we can replace @xmath694 by @xmath697 where @xmath698 is an appropriate cutoff function
. then @xmath699 this proves the result in case @xmath692 in general , write @xmath700 where @xmath701 @xmath702 and apply the foregoing argument to each term of the product @xmath703 [ lmst - covol ] for each sufficiently large @xmath347 there is a constant @xmath705 such that if @xmath98 is supported on the ball of radius @xmath347 centered at the origin , then @xmath706 and @xmath707 also @xmath708 and are taken from ( ( * ? ? ?
* lemma 5.1 ) ) . follows from .
indeed suppose that @xmath709 then there exists @xmath710 such that @xmath711 now we have @xmath712 where @xmath713 applying to @xmath714 we get .
for example , one can take @xmath727 where @xmath648 is a non negative function with integral one supported on the set @xmath728 we write @xmath729 we can also regard @xmath730 as functions on @xmath81 defined by the formula @xmath731 we are ready now to give the proofs of the statements from section [ sec.truncation ] .
\(c ) we already know from part ( b ) that @xmath740 a similar argument shows that @xmath741\circ g^j \| _ { c^{{{\mathbf s } } , { { \mathbf r}}}}=o(2^{j { { \mathbf s } } } k).\ ] ] now the result follows by lemma [ lmprod ]
. now we deal with the case of affine lattices and @xmath565 .
let @xmath129 be such that @xmath744 we claim that @xmath745 this gives the required improvement of an extra power of @xmath746 that is sufficient to verify and using lemma [ lmcovoltail ] .
to show let @xmath747 be the shortest vector in @xmath748 note that @xmath749 is of order @xmath750 thus @xmath129 is contained in a union of lines going in the direction of @xmath747 so that the distance between the lines is almost @xmath751 if we shift @xmath14 in the direction perpendicular to @xmath747 then the probability that one of the shifted lines intersects the ball of a fixed radius around the origin is @xmath752 since @xmath753 due to , the estimate follows .
we now show that and hold if haar measure is replaced by a measure having a @xmath135-centrally smoothable density with respect to haar measure .
we just prove in the lattice case , since the proofs of as well as the proofs for affine lattices are exactly the same .
@xmath754 where the inequality follows from lemma [ lmst - covol ] since @xmath410 with @xmath98 having a compact support .
next , the @xmath117-central smoothability of @xmath112 and lemma [ lmcovoltail ] imply that 1 .
equation still holds for a measure with density @xmath756 2 .
the tail estimate of lemma [ lmcovoltail ] can be proved for measures with centrally smoothable densities following the same lines as the proof of .
let @xmath501 and @xmath759 denote the haar measures on @xmath71 and @xmath69 respectively .
then @xmath760 similarly for @xmath761 , @xmath762 where the last step follows from .
since the integrand depends only on projection of @xmath763 to @xmath79 the integral can be estimated by @xmath764 thus and follow lemma [ lmcovoltail ] . | we study the distribution modulo @xmath0 of the values taken on the integers of @xmath1 linear forms in @xmath2 variables with random coefficients .
we obtain quenched and annealed central limit theorems for the number of simultaneous hits into shrinking targets of radii @xmath3 . by the khintchine - groshev theorem on diophantine approximations
, @xmath4 is the critical exponent for the infinite number of hits . |
the heavy baryons containing a single heavy quark constitute a perfect laboratory to test the non - perturbative aspects of qcd .
the @xmath2 transition which is based on the flavor changing neutral current ( fcnc ) transition of @xmath3 at quark level has been recently observed in cdf collaboration at fermilab @xcite .
it is also planned to be checked at lhcb collaboration at cern @xcite .
the theoretical studies on the branching ratio of @xmath4 in standard model ( sm ) @xcite show that this decay mode is also possible to be observed at lhc .
it is expected that the experimental studies on the heavy baryons and their decay properties constitute one of the main direction of research program at lhc .
hence , the theoretical calculations can play an essential role in this regard .
although the sm has predictions in perfect agreement with collider data up to now , there are some problems such as neutrino oscillations , baryon asymmetry , unification , dark matter , strong cp violation and the hierarchy problem , etc
. which can not be addressed by the sm and still remain unsolved . to cure these deficiencies
, there are a plenty of new physics ( np ) models such as different extra dimension models ( ed ) , various supersymmetric ( susy ) scenarios , etc .
one of the most important new physics scenarios is @xmath1 model , appears in many grand unified theories , such as @xmath5 or string - inspired @xmath6 models @xcite .
the two @xmath1 models in agenda are family non - universal @xmath1 @xcite and leptophobic @xmath1 scenarios @xcite .
the idea of extra heavy z boson comes from the extension of gauge group @xmath5 , predicted by the grand unification theories to larger group @xmath7 .
the @xmath7 gauge group is the next important one after @xmath5 having one extra rank .
hence , this gauge group requires at least one extra neutral gauge boson @xcite . in general , @xmath1 gauge couplings are family universal @xcite , however , due to different constructions of the different families , in string models it is possible to have family non - universal @xmath1 couplings . in some of them ,
three generation of leptons and also the first and second generation of quarks have different coupling to @xmath1 boson when compared to the third families of quarks @xcite . for more information about this model
see for instance @xcite .
the leptophobic @xmath1 model implies that the new neutral gauge boson does not couple to the ordinary sm charged leptons .
the study of the @xmath1 phenomenology is an important part of the scientific program of every present and future colliders and due to its heaviness , the @xmath1 boson may be used to calibrate the future detectors @xcite . for constraints on the mass of the @xmath1 boson and the mixing parameters of the model
see for example @xcite .
there are direct searches for @xmath8 decay @xcite at tevatron , and the possibility to discover this gauge boson is analyzed in @xcite . in the present work we investigate the fcnc transition of @xmath0 in family non - universal @xmath1 model .
in particular , we analyze the differential branching ratio , forward - backward asymmetry as well as some related asymmetry parameters and polarizations and compare the results with the predictions of the sm .
note that the rare baryonic @xmath9 decay within family non - universal @xmath1 model was analyzed in @xcite within also family non - universal @xmath1 model .
the implications of non - universal @xmath1 model on b meson decays were investigated in @xcite .
the @xmath10 @xcite , @xmath11 , and @xmath12 @xcite were investigated in the same framework as well .
the effects of a family non - universal @xmath1 gauge boson is also searched for @xmath13 decays in @xcite .
recently , the @xmath14 mixing , @xmath15 , @xmath16 , @xmath17 and inclusive @xmath18 decays have been analyzed and the stronger constraints have been put for the family non - universal @xmath1 model parameters in @xcite ( see also @xcite ) .
the outline of the paper is as follows . in section 2 ,
we present the effective hamiltonian and transition matrix elements responsible for the @xmath0 transition . in section 3 , we analyze the differential branching ratio , forward backward asymmetry , double lepton polarizations as well as some other related asymmetries in the @xmath1 model and compare the obtained results with the sm predictions . the last section encompasses our concluding remarks .
neglecting terms proportional to @xmath19 , the effective hamiltonian of the fcnc transition of @xmath20 , proceed via quark level @xmath21 in the sm , can be written as @xcite [ e8401 ] h^eff & = & g_f _ em v_tb v_ts^2 , where @xmath22 is the fine structure constant at z mass scale , @xmath23 is the fermi coupling constant , @xmath24 are elements of the cabibbo - kobayashi - maskawa ( ckm ) matrix , and @xmath25 , @xmath26 , @xmath27 are the wilson coefficients .
when @xmath1 boson is considered and @xmath28 mixing is neglected , the extra part which should be added to the above effective hamiltonian is written as @xcite [ e111203 ] h_eff^z^ - 2 g_f v_tb v_ts^+h.c .
+ where @xmath29 and @xmath30 correspond to the chiral @xmath1 couplings to quarks and leptons , respectively . considering the running effects from @xmath31 to @xmath32 scale @xcite , to get the effective hamiltonian for the transition under consideration in @xmath1 model
, we need to make the following replacements in eq .
( [ e8401 ] ) to include @xmath1 boson contributions besides the z boson : [ e111207 ] c_9^eff & & c_9^eff = c_9^eff - ( 28.82 ) b_sb^l v_tb v_ts^(b_^l + b_^r ) , + c_10 & & c_10^=c_10 + ( 28.82 ) b_sb^l v_tb v_ts^ ( b_^l - b_^r ) , where @xmath33 is the strong coupling constant . here
we should mention that the wilson coefficient @xmath25 remains unchanged . in sm , the wilson coefficient @xmath25 in leading logarithm approximation is written as ( see @xcite ) [ cyedi ] c_7^eff(_b ) ^ c_7(_w)+ ( ^ -^ ) c_8(_w)+c_2 ( _ w ) _
i=1 ^ 8 h_i ^a_i , where c_2(_w)=1 , c_7(_w)=- d_0(x_t ) , c_8(_w)=- e_0(x_t ) .
the functions , @xmath34 and @xmath35 are defined as [ d0 ] d_0(x_t ) - + x_t , + + [ e0 ] e_0(x_t ) - + x_t , where @xmath36 with @xmath37 and @xmath38 being the top quark and w boson masses , respectively .
the coefficients @xmath39 and @xmath40 are given as @xcite [ klar ] rrrrrrrrrl a_i = ( & 1423 , & 1623 , & 623 , & - 1223 , & 0.4086 , & -0.4230 , & -0.8994 , & 0.1456 & ) , + h_i = ( & 2.2996 , & - 1.0880 , & - 37 , & - 114 , & -0.6494 , & -0.0380 , & -0.0186 , & -0.0057 & ) .
the parameter @xmath41 in eq .
( [ cyedi ] ) is also defined as , with _
s(x)=,where @xmath42 and @xmath43 . the wilson coefficient @xmath44 is written as @xcite : [ c9eff ] c_9^eff( ) & = & c_9^ndr(s ) + h(z , s ) ( 3 c_1 + c_2 + 3 c_3 + c_4 + 3 c_5 + c_6 ) + & & - 12 h(1 , s ) ( 4 c_3 + 4 c_4 + 3 c_5 + c_6 ) + & & - 12 h(0 , s ) ( c_3 + 3 c_4 ) + 29 ( 3 c_3 + c_4 + 3 c_5 + c_6 ) , where @xmath45 with @xmath46 , and [ c9tilde]c_9^ndr & = & p_0^ndr + y_0(x_t)^2_w -4 z_0(x_t ) + p_e e(x_t ) , here , ndr stands for the naive dimensional regularization scheme .
we ignore the last term in this equation due to the negligible value of @xmath47 .
the @xmath48 @xcite and the functions @xmath49 and @xmath50 are defined in the following form : [ yo ] y_0(x_t ) , and z_0(x_t ) + x_t .
+ in eq.([c9eff ] ) , the @xmath51 is given as ( s ) & = & 1 + _
s(_b ) ( s ) , with [ omega ] ( s ) & = & - 29 ^2 - 43_2(s ) - 23 s ( 1-s ) - 5 + 4s3(1 + 2s)(1-s ) - + & & 2 s ( 1+s ) ( 1 - 2s)3(1-s)^2 ( 1 + 2s ) s + 5 + 9s-6s^26 ( 1-s ) ( 1 + 2s ) . at @xmath52 scale , for the coefficients @xmath53 we have [ coeffs ] c_j=_i=1 ^ 8 k_ji ^a_i , where @xmath54 are given as : [ klar ] rrrrrrrrrl k_1i = ( & 0 , & 0 , & 12 , & - 12 , & 0 , & 0 , & 0 , & 0 & ) , + k_2i = ( & 0 , & 0 , & 12 , & 12 , & 0 , & 0 , & 0 , & 0 & ) , + k_3i = ( & 0 , & 0 , & - 114 , & 16 , & 0.0510 , & - 0.1403 , & - 0.0113 , & 0.0054 & ) , + k_4i = ( & 0 , & 0 , & - 114 , & - 16 , & 0.0984 , & 0.1214 , & 0.0156 , & 0.0026 & ) , + k_5i = ( & 0 , & 0 , & 0 , & 0 , & - 0.0397 , & 0.0117 , & - 0.0025 , & 0.0304 & ) , + k_6i = ( & 0 , & 0 , & 0 , & 0 , & 0.0335 , & 0.0239 , & - 0.0462 , & -0.0112 & ) . + the other functions in eq . ( [ c9eff ] )
are also given as : [ phasespace ] h(y , s ) & = & -89m_b_b - 89y + 827 + 49 x + & & - 29 ( 2+x ) |1-x|^1/2 \ { ll ( | + 1 - 1| - i ) , & x 4z^2s < 1 + 2 1 , & x 4z^2s > 1 , .
+ where @xmath55 or @xmath56 and , h(0 , s ) & = & 827 -89 m_b_b - 49 s + 49 i.we also consider the long distance contribution ( @xmath57 ) coming from @xmath58 family resonances to @xmath26 and parameterized using breit - wigner ansatz @xcite as y_ld = c^(0)_i=1 ^ 6_i , where @xmath59 , and we consider only two lowest resonances @xmath60 and @xmath61 and choose the corresponding phenomenological factors @xmath62 and @xmath63 .
we use the experimental results on the masses and total decay rates of dilepton decays of the considered vector charmonium states @xcite . for more details about the calculation of long distance contributions , see for instance @xcite .
finally , the explicit expression for @xmath27 is given as : [ con ] c_10= - , where , @xmath64 .
the @xmath0 decay amplitude is obtained by sandwiching the effective hamiltonian between the initial and final states @xmath65 which reads @xmath66 to calculate the amplitude , we need to parameterize the transition matrix elements @xmath67 and @xmath68 in terms of twelve form factors @xmath69 , @xmath70 , @xmath71 and @xmath72 ( @xmath73 ) as follows : @xmath74 u_{\sigma_{b}}(p+q)~ , \nnb \\
\langle \sigma(p)\md \bar s i \sigma_{\mu\nu}q^{\nu } ( 1 + \gamma_5 ) b \mid \sigma_b(p+q)\rangle = \bar{u}_\sigma(p ) \big[\gamma_{\mu}f_{1}^{t}(q^{2})+{i}\sigma_{\mu\nu}q^{\nu}f_{2}^{t}(q^{2})+ q^{\mu}f_{3}^{t}(q^{2 } ) \nnb \\
\ar \gamma_{\mu}\gamma_5 g_{1}^{t}(q^{2})+{i}\sigma_{\mu\nu}\gamma_5q^{\nu}g_{2}^{t}(q^{2 } ) + q^{\mu}\gamma_5 g_{3}^{t}(q^{2 } ) \vphantom{\int_0^{x_2}}\big ] u_{\sigma_{b}}(p+q)~,\end{aligned}\ ] ] where @xmath75 and @xmath76 are spinors of @xmath77 and @xmath78 baryons , respectively . the form factors as the main inputs in the analysis of the @xmath79 have been recently calculated in full theory via light cone qcd sum rules in @xcite ( for details about the light cone qcd sum rules see for instance @xcite )
. the fit function of transition form factors is given as @xmath80=\frac{a}{(1-\frac{q^2}{m_{fit}^2})}+\frac{b}{(1-\frac{q^2}{m_{fit}^2})^2 } , \label{parametrization1}\end{aligned}\ ] ] where the fit parameters @xmath81 , @xmath82 , and @xmath83 are presented in table 1 .
using the amplitude mentioned above , the differential decay rate is found as ( see also @xcite ) = | v_tbv_ts^*|^2 v , [ ddf ] where @xmath85 and @xmath86 with @xmath87 being the angle between the momenta of @xmath77 and @xmath88 in the center of mass of leptons , @xmath89 , @xmath90 and @xmath91 .
the functions , @xmath92 , @xmath93 and @xmath94 are given as _
0(s ) 32 m_^2 m__b^4 s ( 1+r - s ) d_3 ^2 + e_3 ^2 + 64 m_^2 m__b^3 ( 1-r - s ) [ d_1^e_3 + d_3 e_1^ ] + 64 m__b^2 ( 6 m_^2 - m__b^2 s ) re [ d_1^e_1 ] + 64 m_^2 m__b^3 ( 2 m__b s re [ d_3^e_3 ] + ( 1 - r + s ) re [ d_1^d_3 + e_1^e_3 ] ) + 32 m__b^2 ( 2 m_^2 + m__b^2 s ) \ { ( 1 - r + s ) m__b [ a_1^a_2 + b_1^b_2 ] + m__b ( 1 - r - s ) [
a_1^b_2 + a_2^b_1 ] - 2 ( [ a_1^b_1 ] + m__b^2 s [ a_2^b_2 ] ) } + 8 m__b^2 \ { 4 m_^2 ( 1 + r - s ) + m__b^2 } a_1 ^2 + b_1 ^2 + 8 m__b^4 \ { 4 m_^2 + m__b^2 s } a_2 ^2 + b_2 ^2 + 8 m__b^2 \ { 4 m_^2 ( 1 + r - s ) - m__b^2 } d_1 ^2 + e_1 ^2 + 8 m__b^5 s v^2 \ { - 8 m__b s [ d_2^e_2 ] + 4 ( 1 - r + s ) [ d_1^d_2+e_1^e_2 ] + 4 ( 1 - r - s ) [ d_1^e_2+d_2^e_1 ] + m__b d_2 ^2 + e_2 ^2 } , _
1(s,1/r ) & = & -16 m__b^41 v_l \ { 2 re(a_1^ * d_1)-2re(b_1^ * e_1 ) + & + & 2 m _ re(b_1^ * d_2-b_2^ * d_1+a_2^ * e_1-a_1^*e_2 ) } + & + & 32 m__b^5 1 v_l \ { m__b ( 1-r)re(a_2^ * d_2 -b_2^ * e_2 ) + & + & re(a_2^ * d_1+a_1^ * d_2-b_2^*e_1-b_1^ * e_2 ) } , _
2(s,1/r)- 8 m__b^4 v^2 a_1 ^2 + b_1
^2 + d_1 ^2 + e_1 ^2 + 8 m__b^6 s v^2 ( a_2 ^2 + b_2 ^2 + d_2 ^2 + e_2 ^2 ) , where [ nolabel ] a_1 - 2 m_bq^2 c_7^eff f_1^t + g_1^t + c_9^eff f_1-g_1 , + a_2 a_1 1 2 , + a_3 a_1 1 3 , + b_i a_i g_i - g_i ; g_i^t - g_i^t , + d_1 c_10 f_1-g_1 , + d_2 d_1 1 2 , + d_3 d_1 1 3 , + e_i d_i g_i - g_i . to numerically analyze the differential branching ratio with respect to @xmath95 , we perform integral over @xmath96 in the interval @xmath97 $ ] in eq .
( [ ddf ] ) and take the values of input parameters as @xmath98 , @xmath99 , @xmath100 , @xmath101 , @xmath102 and @xmath103 @xcite , @xmath104 , @xmath105 @xcite , @xmath106 , @xmath107 and @xmath108 .
the remaining parameters are related to the family non - universal @xmath109 model .
the modifications on the wilson coefficients in our case are described by the four parameters , @xmath110 and @xmath111 .
constraints to @xmath112 are put fitting the results of the @xmath14 mixing observables to recent measurements performed at tevatron and lhc @xcite .
these parameters are chosen as @xmath113 , @xmath114 to maximize the effects of the additional @xmath109 gauge boson @xcite .
comparing also the theoretical predictions and observational results for the decay channels of @xmath115 @xcite , @xmath116 @xcite and @xmath117 @xcite , the parameters @xmath118 and @xmath119 are obtained @xcite .
we plot the differential branching ratio , forward backward asymmetry , other asymmetry parameters and polarizations with respect to @xmath95 to check the sensitivity of these observables to the model parameters mentioned above at muon channel .
we compare predictions of the @xmath109 and sm models when the uncertainties of the form factors are taken into account . for this aim ,
first we plot the variations of the differential branching ratio in terms of @xmath95 in two models in figure [ dbrmuon.eps ] . in this figure ,
the brown - yellow band surrounded by green lines refers to the family non - universal @xmath109 model while the blue band surrounded by the red lines denotes the sm results . from this figure
, we obtain the following results : cc * the sm and @xmath109 bands intersect each other , the discrepancy between the differential branching ratios obtained from two models is small .
hence the differential branching ratio has not essential sensitivity to @xmath109 gauge boson in @xmath0 channel .
similar result is obtained in @xcite for @xmath2 decay channel . * the order of differential branching ratio indicates a possibility for this decay mode to be checked at @xmath120 in near future .
one of the most promising tools in detecting the np effects is the forward - backward asymmetry of leptons which is defined as _
fb = , where @xmath121 is the number of events that particle is moving `` forward '' with respect to any chosen direction , and @xmath122 is the number of events that particle moves to `` backward '' direction . in technique language , the forward - backward asymmetry @xmath123 is defined in terms of the differential decay rate as _ fb ( s)= , we depict the dependence of @xmath124 on @xmath95 in figure [ afbmuon.eps ] considering the uncertainties of the form factors .
cc from this figure we conclude that the effects of @xmath109 model s parameters to the forward backward asymmetry are considerable compared to the differential branching ratio .
although the errors of the form factors partially kill the discrepancies between two models predictions in some regions , they do not sweep the same areas especially at higher values of @xmath95 .
similar results have been obtained in @xmath2 channel for higher values of @xmath95 @xcite .
however small values of @xmath95 the two baryonic channels show different behaviours when we compare the @xmath109 model predictions with those of the sm .
asymmetry parameters characterize the angular dependence of differential decay width for the cascade decay @xmath126 with polarized and unpolarized heavy baryons . in this part , sensitivity of the baryon polarization asymmetry parameter to new wilson coefficients is analyzed .
the helicity amplitudes for the @xmath127 decay are obtained by analyzing quasi two body decay @xmath128 , followed by the leptonic decay of @xmath129 in @xcite .
these amplitudes are obtained using helicity amplitude formalism and polarization density matrix method , demonstrated in @xcite . considering _
b^1/2 ^+ ^1/2 ^+ ( a + b ) + v^(^+ ^- ) , with @xmath130 being off - shell @xmath131 or @xmath132 boson , the normalized joint angular decay distribution for the two cascade decay is written as @xcite [ e7121 ] v_tb v_ts^^2 v b(_b a + b ) ^2 , + where @xmath133 is calculated in @xcite . for the leptonic part , in the rest frame of the intermediate boson , the angle of the anti - lepton with respect to its helicity axes is shown by @xmath87 . for the hadronic decay , in the rest frame of the @xmath77 , @xmath134 is the angle of the @xmath81 momentum with respect to its helicity axes .
the polar angle distribution of @xmath135 decay is obtained by integrating eq .
( [ e7121 ] ) with respect to @xmath87 and it gives the differential decay rate in terms of @xmath95 and @xmath136 @xcite [ e7126 ] ~1 + ( q^2 ) _ ( q^2 ) _ ( q^2 ) , where the baryon polarization asymmetry parameter @xmath125 is given as [ e7134 ] _ ( q^2 ) \ { 4 m_^2 a_+1/2,+1 ^2 + 2 q^2 ( a_+1/2,+1 ^2 + v^2 b_+1/2,+1 ^2 ) + 2 m_^2 a_+1/2,0 ^2 - q^2 ( a_+1/2,0 ^2 + v^2 b_+1/2,0 ^2 ) - 6 m_^2 b_+1/2,t ^2 + 4 m_^2 a_-1/2,-1 ^2 - 2 q^2 ( a_-1/2,-1 ^2 + v^2 b_-1/2,-1 ^2 ) + 2 m_^2 a_-1/2,0 ^2 + q^2 ( a_-1/2,0 ^2 + v^2 b_-1/2,0 ^2 ) + 6 m_^2 b_-1/2,t ^2 } , and the asymmetry parameter @xmath137 is obtained as [ e7127 ] ( q^2 ) \{4
m_^2 a_+1/2,+1 ^2 + 2 q^2 ( a_+1/2,+1 ^2 + v^2 b_+1/2,+1 ^2 ) + 2 m_^2 a_+1/2,0 ^2 + 6 m_^2 b_+1/2,t ^2 + q^2 ( a_+1/2,0 ^2 + v^2 b_+1/2,0 ^2 ) + 4 m_^2 a_-1/2,-1 ^2 - 2 q^2 ( a_-1/2,-1 ^2 + v^2 b_-1/2,-1 ^2 ) + q^2 ( a_-1/2,0 ^2 + v^2 b_-1/2,0 ^2 ) - 2 m_^2 a_-1/2,0 ^2 - 6 m_^2 b_-1/2,t ^2 } , with [ e7128 ] ( q^2 ) \{4
m_^2 a_+1/2,+1 ^2 + 2 q^2 ( a_+1/2,+1 ^2 + v^2 b_+1/2,+1 ^2 ) + 2 m_^2 a_+1/2,0 ^2 + 6 m_^2 b_+1/2,t ^2 + q^2 ( a_+1/2,0 ^2 + v^2 b_+1/2,0 ^2 ) + 4 m_^2 a_-1/2,-1 ^2 + 2 q^2 ( a_-1/2,-1 ^2 + v^2 b_-1/2,-1 ^2 ) + q^2 ( a_-1/2,0 ^2 + v^2 b_-1/2,0 ^2 ) + 2 m_^2 a_-1/2,0 ^2 + 6 m_^2 b_-1/2,t ^2 } .
the definitions for @xmath138 and @xmath139 are given in @xcite where @xmath140 and @xmath141 are helicities of lepton pairs and vector boson , respectively .
we plot the dependence of the baryon asymmetry parameter @xmath125 on @xmath95 in figure [ alphalambdamuon.eps ] . from this figure , it is clear that cc the family non - universal @xmath109 model s prediction deviates significantly from the sm prediction such that the errors of the form factors can not kill the discrepancies between two model predictions for the baryon asymmetry parameter @xmath125 .
the polar angle distribution of @xmath144 decay is obtained by integrating eq .
( [ e7121 ] ) with respect to @xmath136 which gives the differential decay rate in terms of @xmath95 and @xmath87 [ e7129 ] ~1 + 2 _ + _ ^2 , where [ e7130 ] _ ( q^2 ) 2 v q^2 , and [ e7131 ] _ ( q^2 ) \ { - 4 m_^2 a_+1/2,+1 ^2 + q^2 ( a_+1/2,+1 ^2 + v^2 b_+1/2,+1 ^2 ) + 4 m_^2 a_+1/2,0 ^2 - q^2 ( a_+1/2,0 ^2 + v^2 b_+1/2,0 ^2 ) + 4 m_^2 a_-1/2,-1 ^2 + q^2 ( a_-1/2,-1 ^2 + v^2 b_-1/2,-1 ^2 ) + 4 m_^2 a_-1/2,0 ^2 - q^2 ( a_-1/2,0 ^2 + v^2 b_-1/2,0 ^2 ) } , with [ e7132 ] _ 1(q^2 ) 4 m_^2 a_+1/2,+1 ^2 + q^2 ( a_+1/2,+1 ^2 + v^2 b_+1/2,+1 ^2 ) + 4 m_^2 b_+1/2,t ^2 + q^2 ( a_+1/2,0 ^2 + v^2 b_+1/2,0 ^2 ) + 4 m_^2 a_-1/2,-1 ^2 + q^2 ( a_-1/2,-1 ^2 + v^2 b_-1/2,-1 ^2 ) + 4 m_^2 b_-1/2,t ^2 + q^2 ( a_-1/2,0 ^2 + v^2 b_-1/2,0 ^2 ) .
we plot the dependence of the polar angle distribution parameters @xmath142 and @xmath143 on @xmath95 in figures [ alphathetam.eps ] and [ alphabetamuon.eps ] , respectively . from these figures
, we read that cc cc * as far as the @xmath142 is considered we see a considerable discrepancy between the two model s predictions . in the case of @xmath2
this parameter also shows to be very sensitive to @xmath109 gauge boson .
however in our case this sensitivity is relatively small . *
similar to the @xmath2 channel considered in @xcite , the asymmetry parameter @xmath143 in @xmath145 model gives roughly the same result as the sm .
to analyze the double lepton polarizations , we need to define the following orthogonal unit vectors @xmath146 with @xmath147 or @xmath148 in the rest frame of double leptons ( see for instance @xcite for detailed information ) [ e6616 ] s^-_l 0,_l^-= 0 , , + s^-_n 0,_n^-= 0 , , + s^-_t 0,_t^-= 0,_n^- _
l^- , + s^+_l 0,_l^+= 0 , , + s^+_n 0,_n^+= 0 , , + s^+_t 0,_t^+= 0,_n^+ _ l^+ , where @xmath149 and @xmath150 are the three - momenta of the leptons @xmath151 and @xmath78 baryon .
the symbols l , t and n denote longitudinal , transverse , and normal polarizations , respectively . using the lorentz boost , these unit vectors are transformed from the rest frame of the leptons into the center of mass ( cm ) frame of them along the longitudinal direction . as a result ,
we get [ e6617 ] s^_l _ cm , , where @xmath152 and @xmath153 and @xmath154 are the energy and mass of leptons in the cm frame , respectively . under the above transformation ,
the other two unit vectors , @xmath155 , @xmath156 remain unchanged .
the double lepton polarizations are given as @xcite [ e111213 ] p_ij ( ) .
+ + the first ( second ) subindex of @xmath157 represents polarization of lepton ( anti - lepton ) . using the above definitions , some double lepton polarizations are obtained as [ e111215 ] p_ln ( ) - p_nl ( ) = \ { ( 1- ) ( a_1^d_1 + b_1^e_1 ) + m__b ( a_1^e_3 - a_2^e_1 + b_1^d_3 -b_2^d_1 ) + m__b ( a_1^d_3 + a_2^d_1 + b_1^e_3 + b_2^e_1 ) - m__b^2 ^2 ( b_2^e_3 + a_2^d_3 ) } , [ e111217 ] p_lt ( ) \ { ( 1- ) ( d_1 ^2 + e_1 ^2 ) - ( a_1 d_1 ^ - b_1 e_1^ ) + m__b + m__b^2 ( 1- ) ( a_2 d_2 ^ - b_2 e_2^ ) + m__b + m__b^2 ^2 ( d_2 d_3^+ e_2 e_3^ ) } , [ e111219 ] p_nt ( ) - p_tn = \ { ( a_1 d_1^+b_1 e_1^ ) + m__b^2 ( a_2^d_2 + b_2^e_2 ) } , [ e111221 ] p_nn ( ) \ { 24 _ ^2 ( a_1 b_1^+ d_1 e_1^ ) - 12 m__b _ ^2 ( 1- + ) ( a_1 a_2^+ b_1 b_2^ ) + 6 m__b _ ^2 + 12 m__b _ ^2 ( 1 - - ) ( a_1 b_2^+ a_2 b_1 ^+ d_1 e_3^+ d_3 e_1^ ) + ( a_1 ^2 + b_1 ^2 - d_1 ^2 - e_1 ^2 ) + 24 m__b^2 _ ^2 ^2 ( a_2 b_2^+ d_3 e_3^ ) - m__b^2 ^2 v^2 ( d_2 ^2 + e_2 ^2 ) + m__b^2 \ { - 2 _ ^2 [ 2 ( 1 + ^2 ) - ( 1 + ) - ( 4 + ) ] } ( a_2 ^2 + b_2 ^2 ) } , [ e111222 ] p_tt ( ) \ { - 24 _ ^2 ( a_1 b_1^+ d_1 e_1^ ) - 12 m__b _ ^2 ( 1- + ) ( d_1 d_3^+ e_1 e_3^ ) + 24 m__b^2 _ ^2 ^2 ( a_2 b_2^+ d_3 e_3^ ) + 6 m__b _ ^2 + 12 m__b _ ^2 ( 1 - - ) ( a_1 b_2^+ a_2 b_1 ^+ d_1 e_3^+ d_3 e_1^ ) + ( a_1 ^2 + b_1 ^2 ) + m__b^2 \ { + _ ^2 [ 4 ( 1- ) ^2 - 2 ( 1 + ) - 2 ^2 ] } ( a_2 ^2 + b_2 ^2 ) + \ { - 2 _ ^2 [ 5 ( 1- ) ^2 - 7 ( 1 + ) + 2 ^2 ] } ( d_1 ^2 + e_1 ^2 ) + m__b^2 ^2 v^2 ( d_2 ^2 + e_2 ^2 ) } , where ^(s ) .
we plot the dependence of the double lepton polarizations on @xmath95 in figures [ plnmuon.eps]-[pttm.eps ] .
cc cc cc cc cc from these figures , we observe that * in general , there are considerable discrepancies between the @xmath1 model and the sm predictions on the double lepton polarization asymmetries under consideration except for the @xmath158 .
* the big deviations of the @xmath145 model s results from the sm predictions appear at lower values of @xmath95 .
for the @xmath159 and @xmath160 , the np parameters are more effective at lower values of @xmath95 .
* the maximum discrepancy between two model s predictions belongs to the @xmath161 and @xmath162 .
* compared to the @xmath2 decay channel @xcite , the double lepton polarization asymmetries are more sensitive to the @xmath109 gauge boson in @xmath0 decay channel .
* determination of zero points of the double lepton polarization asymmetries in the experiment and comparison of the obtained results with the theoretical predictions can give valuable information about the existence of the @xmath145 boson .
in the present study , we analyzed the @xmath163 decay channel in both the sm and @xmath1 model considering the errors of form factors .
we discussed the sensitivity of the differential branching ratio , forward - backward asymmetry as well as some asymmetry parameters and polarization asymmetries defining the considered decay channel to @xmath109 gauge boson .
our results on the considered physical quantities overall depict considerable discrepancies between the @xmath145 and sm model s predictions . in the case of the differential branching ratio
the discrepancy between two model s predictions is small such that the uncertainties of the form factors roughly kill the difference .
the maximum discrepancies belong to the double lepton polarization asymmetries and asymmetry parameter @xmath125 which are quite sensitive to @xmath109 gauge boson .
the discrepancies between two model s predictions on some physical observables can be considered as signals for existing the extra @xmath145 gauge boson .
the order of the differential branching ratio indicates that this channel can be studied at lhc in near future .
any measurement on the physical quantities considered in the present work and comparison of the obtained data with the theoretical predictions can give useful information not only about the existence of the @xmath145 gauge boson but also about the nature of participating baryons , @xmath77 and @xmath78 .
we also compared our results with the results of the considered observables for @xmath2 decay channel @xcite which is in agenda of different experiments nowadays .
the study of different fcnc channels can provide us with more data which may help us in searching for @xmath145 gauge boson as new physics effect .
n. katirci thanks boazii university for the financial support provided by the scientific research fund with project no : 7128 .
99 t. aaltonen et al .
[ cdf collaboration ] , phys .
107 , 201802 ( 2011 ) ; arxiv : 1107.3753[hep - ex ] . our personal communications with yasmine sara amhis from lhcb collaboration .
k. azizi , m. bayar , a. ozpineci , y. sarac and h. sundu , phys . rev .
d 85 , 016002 ( 2012 ) ; arxiv : 1112.5147[hep - ph ] . f. grsey , m. serdarolu , lett .
21 , ( 1978 ) .
g. buchalla , g. burdman , c. t. hill and d. kominis , phys . rev .
d 53 , 5185 ( 1996 ) ; arxiv : hep - ph/9510376 .
e. nardi , phys .
d 48 , 1240 ( 1993 ) ; arxiv : hep - ph/9209223
. j. bernabeu , e. nardi and d. tommasini , nucl .
b 409 , 69 ( 1993 ) ; arxiv : hep - ph/9306251 . v. barger , m. berger and r. j. phillips , phys . rev .
d 52 , 1663 ( 1995 ) ; arxiv : hep - ph/9503204 .
e. eichten , i. hinchliffe , k. d. lane and c. quigg , rev .
56 , 579 ( 1984 ) .
p. langacker and m. plumacher , phys .
d 62 , 013006 ( 2000 ) ; arxiv : hep - ph/0001204 . j. l. lopez and d. v. nanopoulos , phys . rev .
d 55 , 397 ( 1997 ) ; arxiv : hep - ph/9605359 . b. b. sirvanli , mod .
. lett . a 23 , 347 ( 2008 ) ; arxiv : hep - ph/0701173 .
a. leike , phys .
317 , 143 ( 1999 ) ; arxiv : hep - ph/9805494 . t. k. kuo and n. nakagawa , phys .
d 31 , 1161 ( 1985 ) ; phys .
d 32 , ( 1985 ) 306 .
k. k. gan , phys .
b 209 , 95 ( 1988 ) .
e. nardi , phys .
d 48 , 1240 ( 1993 ) ; arxiv : hep - ph/9211246
. b. holdom , phys .
b 339 , 114 ( 1994 ) ; arxiv : hep - ph/9407311 . x. zhang and b. l. young , phys
d 51 , 6584 ( 1995 ) .
b. holdom and m. v. ramana , phys .
b 365 , 309 ( 1996 ) ; arxiv : hep - ph/9509272 .
s. chaudhuri , s.w .
chung , g. hockney and j. lykken , nucl .
b 456 , 89 ( 1995 ) . g. cleaver , m. cvetic , j. r. espinosa , l. everett and p. langacker , nucl . phys .
b 525 , 3 ( 1998 ) ; arxiv : hep - th/9711178 . y. zhang , z. cai ; arxiv : 1106.0163[hep - ph ] .
a. kundu , phys .
b 370 , 135 ( 1996 ) ; arxiv : hep - ph/9504417 .
j. erler , p. langacker , phys .
84 , 212 ( 2000 ) ; arxiv : hep - ph/9910315 . c. caso et al .
j. c3 , 1 ( 1998 ) .
p. langacker , rev . mod
81 , 1199 ( 2008 ) ; arxiv : 0801.1345[hep - ph ] . k. cheung , c. w. chiang , n. g. deshpande and j. jiang , phys . lett .
b 652 , 285 ( 2007 ) ; arxiv : hep - ph/0604223 .
x. g. he and g. valencia , phys . rev .
d 74 , 013011 ( 2006 ) ; arxiv : hep - ph/0605202
. f. abe et al .
[ cdf collaboration ] , phys .
79 , 2192 ( 1997 ) .
a. abulencia et al .
[ cdf collaboration ] , phys .
96 , 211801 ( 2006 ) ; arxiv : hep - ex/0602045 . m. carena , a. daleo , b. a. dobrescu , t. m. p. tait , phys . rev .
d 70 , 093009 ( 2004 ) ; arxiv : hep - ph/0408098 . a. k. giri , r. mohanta , eur .
j. c 45 , 151 ( 2006 ) ; arxiv : hep - ph/0510171 . t. m. aliev , m. savci , nuc .
b , 863 , 398 , ( 2012 ) , arxiv:1202.0398[hep - ph ] .
t. m. aliev , m. savci , phys .
b , 718 , 566 , ( 2012 ) , arxiv:1202.5444[hep - ph ] . c. h. chen and h. hatanaka , phys .
d 73 , 075003 ( 2006 ) ; arxiv : hep - ph/0602140
. v. barger , c. w. chiang , p. langacker , h. s. lee , phys .
b 580 , 186 ( 2004 ) ; arxiv : hep - ph/0310073 ; ibid b 598 , 218 ( 2004 ) ; arxiv : hep - ph/0406126 ; v. barger , l. everett , j. jiang , p. langacker , t. liu , c. e. m. wagner , phys . rev .
d 80 , 055008 ( 2009 ) ; arxiv : hep - ph/0902.4507 . c. w. chiang , n. g. deshpande , j. jiang , jhep 0608 , 075 ( 2006 ) ; arxiv : hep - ph/0606122 .
k. cheung , c. w. chiang , n. g. deshpande , j. jiang , phys .
b 652 , 285 ( 2007 ) ; arxiv : hep - ph/0604223 .
r. mohanta and a. k. giri , phys . rev .
d 79 , 057902 ( 2009 ) ; arxiv : 0812.1842[hep - ph ] .
j. hua , c. s. kim and y. li , phys .
b 690 , 508 ( 2010 ) ; arxiv : 1002.2532[hep - ph ] ; q. chang , nucl . phys .
b 845 , 179 ( 2011 ) ; arxiv : 1101.1272[hep - ph ] . q. chang , x. q. li and y. d. yang , jhep 0905 , 056 ( 2009 ) ; arxiv : 0903.0275[hep - ph ] .
y. li , j. hua and k. c. yang , jhep 1002 , 082 ( 2010 ) ; arxiv : 0907.4408[hep - ph ] ; q. chang , x. q. li and y. d. yang , jhep 1004 , 052 ( 2010 ) ; arxiv : 1002.2758[hep - ph ]
. s. w. wang , g. l. sun , x. q. yang and j. s. huang , eur .
j. c 72 , 1852 ( 2012 ) .
r. h. li , c. d. la , w. wang , phys . rev .
d 83 , 034034 ( 2011 ) .
y. li , j. hua , k. c. yang , eur .
j. c 71 , 1775 ( 2011 ) ; arxiv : 1107.0630[hep - ph ] . y. li , x. j. fan , j. hua , e. l. wang , phys .
d 85 , 074010 ( 2012 ) ; arxiv : 1111.7153[hep - ph ]
. q. chang , x. q. li , y. d. yang , int .
phys . a 26 , 1273 ( 2011 ) ; arxiv : 1003.6051[hep - ph ] .
x. qiang li , y. min li , g. r. liu , f. su , ( 2012 ) ; arxiv:1204.5250 [ hep - ph ] .
v. abazov , et al . ,
d0 collaboration , d0 conference note , 6098-conf .
t. aaltonen , et al .
, cdf collaboration , phys .
d 85 072002 , ( 2012 ) .
r. aaij , et al . ,
lhcb collaboration , phys .
108 ( 2012 ) 101803 ; arxiv:1203.4493 [ hep - ph ] .
r. aaij , et al .
, lhcb collaboration , phys .
b 707 497 , ( 2012 ) .
g. buchalla , a. j. buras , m. e. lautenbacher , rev .
68 , 1125 ( 1996 ) ; arxiv : hep - ph/9512380 . g. bobeth , a. j. buras , f. krger , j. urban , nucl .
b 630 , 87 ( 2002 ) ; arxiv : hep - ph/0112305 .
w. altmannshofer , p. ball , a. bharucha , a. buras , d. m. straub , m. wick , jhep 0901 , 019 ( 2009 ) ; arxiv : 0811.1214[hep - ph ] .
a. ghinculov , t. hurth , g. isidori , y. p. yao , nucl .
b 685 , 351 ( 2004 ) ; arxiv : hep - ph/0312128 . k.g .
chetyrkin , m. misiak , m. munz , phys .
b 400 , 206 ( 1997 ) .
a. j. buras , m. misiak , m. muenz and s. pokorski , nucl . phys .
b 424 , 374 ( 1994 ) ; arxiv : hep - ph/9311345 . m. misiak , nucl . phys .
b 393 , 23 ( 1993 ) ; erratum ibid b 439 , 161 ( 1995 ) .
buras , m. muenz , phys . rev .
d 52 , 186 ( 1995 ) ; arxiv : hep - ph/9501281 .
j. beringer et al .
( particle data group ) , phys .
d 86 , 010001 ( 2012 ) .
m. beneke , g. buchalla , m. neubert and c.t .
sachrajda , eur .
j. c 61 , 439 ( 2009 ) ; arxiv : 0902.4446[hep - ph ] .
a. khodjamirian , th .
mannel , a. a. pivovarov and y. m. wang ; arxiv : 1006.4945[hep - ph ] . i. i. balitsky , v. m. braun , and a. v. kolesnichenko , nucl .
b 312 , 509 ( 1989 ) . v. m. braun , arxiv : hep ph/9801222 ( 1998 ) .
i. i. balitsky and v. m. braun , nucl .
b 311 , 541 ( 1989 ) .
t. m. aliev , k. azizi , m. savci , phys .
d 81 , 056006 ( 2010 ) ; arxiv : 1001.0227[hep - ph ] .
p. colangelo , a. khodjamirian , at the frontier of particle physics / handbook of qcd
, edited by m. shifman ( world scientific , singapore , 2001 ) , vol.3 , p. 1495 . b. aubert , et al . ,
babar collaboration , phys .
93 081802 , ( 2004 ) . m. iwasaki , et al .
, belle collaboration , phys .
d 72 092005 , ( 2005 ) .
lee , et al . , babar collaboration , ( 2012 ) ; arxiv:1204.3933 [ hep - ph ] , .
r. aaij , et al . ,
lhcb collaboration , phys .
108 , 181806 ( 2012 ) ; arxiv:1112.3515 [ hep - ph ] .
t. m. aliev , m. savci , jhep 0605 , 001 ( 2006 ) ; arxiv : hep - ph/0507324 .
p. bialas , j. g. krner , m. krmer and k. zalewski , z. phys .
c 57 , 115 ( 1993 ) .
j. g. krner and m. krmer , phys .
b 275 , 495 ( 1992 ) .
t. mannel and g. a. schuler , phys . rev .
d 279 , 194 ( 1992 ) .
m. tanaka , phys .
d 47 , 4969 ( 1993 ) . m. gremm , g. koepp and l. m. sehgal , phys .
d 52 , 1588 ( 1995 ) ; arxiv : hep - ph/9502207 . c. s. huang and h. g. yan , phys .
d 56 , 5981 ( 1997 ) . j. g. krner and d. pirjol , phys .
b 334 , 399 ( 1994 ) ; arxiv : hep - ph/9405360 . t. m. aliev , m. savci , b. b. sirvanli , eur . phys .
j. c 52 , 375 ( 2007 ) , arxiv : hep - ph/0608143 . t. m. aliev , v. bashiry , m. savci , eur .
j. c 38 , 283 ( 2004 ) , arxiv : hep - ph/0409275 . w. bensalem , d. london , n. sinha and r. sinha , phys .
d 67 , 034007 ( 2003 ) , arxiv : hep - ph/0209228 . | using newly available form factors obtained from light cone qcd sum rules in full theory , we study the flavor changing neutral current transition of @xmath0 decay in the family non - universal @xmath1 model . in particular , we evaluate the differential branching ratio , forward - backward asymmetry as well as some related asymmetry parameters and polarizations .
we compare the obtained results with the predictions of the standard model and discuss the sensitivity of the observables under consideration to family non - universal @xmath1 gauge boson .
the order of differential branching ratio shows that this decay mode can be checked at lhc in near future . 0^*0
5_5 s_s^ ? ?
o _ ^0 _ 1s |s s 1_1 8_8 1s 16.3 true cm 23.0 true cm
-0.8 true in 0.00 true in 1_1 8_8 0^*0
5_5 s_s^ ? ?
o _ ^0
_ pacs numbers : 12.60.-i , 12.60.cn , 13.30.-a , 13.30.ce , 14.20.mr |
the electroweak part of the standard model@xcite unequivocally informs us that different fermion chiralities carry distinct properties , when gauged .
the weak isospin ( @xmath0 ) and hypercharge ( @xmath1 ) assignments of the chiral parts of leptonic doublets are carefully designed so that the charge @xmath2 emerges correctly ; however the significant point is that right and left components transform very differently under the associated gauge groups .
the aim of this paper is to see how these features can be incorporated with gravity in a scheme based on _ spacetime _ coordinates augmented by lorentz scalar anticommuting _ property _ coordinates . in two earlier
papers@xcite we investigated such a scheme but paid scant attention to parity as we were only concerned with the unification of gravity with electromagnetism ( using one property , charge ) or with yang - mills theory ( two isotopic properties ) .
we showed how such a unification could be achieved , with the curvature arising from property .
in particular , the lagrangian and energy momentum tensor of the gauge fields that transfer property across spacetime were automatically generated .
the chiral fermions were treated uniformly and assumed to have the same attributes in those works .
it was clear , but more complicated , how to generalize the process for three properties , as needed in qcd , but we shied from carrying out the algebra as it led to a plethora of gauge invariant curvature coefficients in the metric that we are still uncertain how to constrain . with only two properties we have however the full machinery needed to tackle the salient properties of electroweak physics , just by focussing on the subgroup u(1)@xmath3 u(1)@xmath4 , confining ourselves to the charged lepton interactions . before doing so we shall quickly recapitulate the way this is normally handled and our unification goal , if nothing else to establish our conventions ; this we do in section 2 .
we then turn to the implementation with anticommuting scalar attributes in section 3 , starting with the fermions which are , after all , at the root of the entire gauge group ; then we turn to the scalar supermultiplet , which includes the uncharged higgs boson@xcite .
the way that gravity and the gauge bosons enter is through an extended spacetime - property metric ( obtained from frame vectors ) which incorporates the correct gauge group , and this is described in section 4 .
having achieved the goal of combining chirality with anticommuting attributes and curved spacetime , we offer some conclusions in section 5 and what to expect when generalizing to the full su(3)@xmath5su(2)@xmath6@xmath5u(1)@xmath4 of the standard model .
the most significant point of our analysis is the conclusion that any scheme , such as ours , where gravity and other forces are married through some extended metric must lead to a single force strength at some large scale ; extrapolation to low scales via the renormalization group has to then result in the known different forces , given the particle spectrum
. such are the overriding demands of gravitational universality .
before launching into our property scheme for dealing with chiral fermions , let us describe the endpoint attained by conventional theory , ignoring neutrinos and charged vector bosons .
we are dealing therefore with the u(1)@xmath7u(1)@xmath4 subgroup of the electroweak gauge group ; this does at least include electromagnetism and the weak neutral current interactions of the charged leptons . in preserving the same charge @xmath2 for the left and right pieces of the fermion field under local gauge transformations ,
@xmath8 one encounters the covariant derivatives : @xmath9\psi_l,\quad d\psi_r=[\partial+ig'b]\psi_r,\ ] ] where @xmath10 are of course the associated gauge fields with couplings @xmath11 . hence the only way to induce a mass term in addition to the usual kinetic pieces is to introduce a higgs - like scalar @xmath12 which transforms inversely as @xmath13 so as to ensure invariance of the yukawa interaction , @xmath14 the gauge invariant kinetic motion of the scalar field , @xmath15h^\dag.[\partial - i(gw - g'b)/2]h,\ ] ] can then provide a mass to the gauge field combination @xmath16 , provided that the scalar field acquires a vacuum expectation value @xmath17 , by some mechanism .
this is where the weak mixing angle @xmath18 comes in@xcite ; let @xmath19 and as usual define @xmath20 , whereupon the orthogonal fields @xmath21 and @xmath22 emerge from the inverse rotation , @xmath23 with the fermionic interaction lagrangian for the gauge fields devolving to @xmath24 it is worth noting that , because @xmath25 experimentally to a rather good approximation , the interaction ( at _ low _ energy ) almost simplifies to @xmath26 from this perspective we can view @xmath27 and @xmath28 as forming a part of a u(2 ) group within a matrix @xmath29 of which only the @xmath30 component is in play .
however it needs emphasizing that at the level of ( 7 ) the left and right gauge fields are _ not _ rotations amongst one another . [ as we will presently see , we are forced to suppose that @xmath31 and @xmath32 coincide at high energy in order to respect universal gravitational coupling to matter . ]
it only remains to ` curve ' the results ( 4 ) and ( 6 ) by including the vierbein @xmath33 and metric @xmath34 into the contractions over indices and we magically get the interactions of the fermion and boson sources with gravity as well .
the job now is to try to reproduce these conventional constructs through an extended spacetime - property metric .
the very different behaviours of the chiral fermion components under the internal gauge transformations point to the need for distinct property coordinates ; so at the very least we require two independent attributes . in an earlier
paper@xcite we treated the case of two properties but we assumed that they referred to ` up ' and ` down ' parts of an isotopic doublet , which led to the parity - conserving unification of yang - mills u(2 ) with gravity ; there we encountered a property curvature invariant @xmath35 .
now it is no longer necessary to insist upon full u(2 ) invariance and/or parity conservation , for when we identify @xmath36 with left - handedness and @xmath37 with right handedness , we can permit separate u(1)@xmath7u(1)@xmath38 invariants @xmath39 and @xmath40 .
our first task is to check that the flat space results can be reproduced , namely a kinetic term and _ no _ mass term , after we integrate over attribute space ; the next task will be to include the scalar fields to recover induced masses and our third task will be to obtain the gravitational and gauge field interactions via an extended spacetime - property vielbein which ` curves ' the space . at the end of the day we must recognize that the unification of parity violating interactions with gravity will not change an ugly duckling ( the standard model ) into a beautiful swan .
begin with a dirac fermion superfield @xmath41 , expanded in the two chiral properties ( @xmath42 and @xmath43 ) which has been made anti - selfdual@xcite : - @xmath44 as explained in an earlier paper , the adjoint superfield must be defined as @xmath45 because of spinorial orthogonality this means that a mass - like term , @xmath46 will automatically give zero when integrated over property space as there are insufficient numbers of @xmath47s .
however a kinetic term , @xmath48 is perfectly fine as it survives integration over property .
it is very encouraging that we require an extra scalar superfield to induce a mass interaction since it reverberates with gauge invariance requirements .
thus consider the real anti - selfdual superscalar : @xmath49 when interacting with @xmath50 , this will include a piece @xmath51 a vacuum expectation value for the chargeless field @xmath52 does duty for a mass term ; it will also make an appropriate mixture of gauge bosons massive when we introduce curvature into property space .
the most important objective is to get the correct coupling of gauge fields and gravity to the chiral fermions through a generalized spacetime - property metric . the rest ( coupling to bosons , gauge - gravity lagrangian , etc . )
will take care of itself . in a previous paper@xcite , given two properties
, we proved that with a metric field @xmath53 in full @xmath54 space having components @xmath55 + e^2 l^2\bar{\zeta}(w_mw_n+w_nw_m)\zeta(1+c_3\bar{\zeta}\zeta)/2,\\ g_{m\nu}&=&iel^2(\bar{\zeta}w_m)^{\bar{\nu}}(1+c_3\bar{\zeta}\zeta)/2 , \quad{\rm etc.}\end{aligned}\ ] ] we could reproduce all gravitational plus gauge field interactions reliably via the total superscalar curvature @xmath56 , possessing full u(2 ) invariance .
such a metric resulted from frame vectors with components @xmath57 in our case we do not require full u(2 ) invariance but merely u(1)@xmath5u(1 ) symmetry , as dictated by experiment for charged leptons ; so we can relax the condition that the curvature coefficients @xmath58 and the couplings be the same for the two associated gauge fields @xmath21 and @xmath22 .
we have a lot more freedom in constructing the curvatures because @xmath59 and @xmath60 are separately invariant .
[ the appendix contains a full discussion of all property curvatures that do not disturb the u(1)@xmath5u(1 ) symmetry and the complications that ensue . ]
in particular the vielbeins @xmath61 given there have forms which ensure that the couplings of the gauge fields to the fermions correctly sum to @xmath62 . then reinterpreting @xmath63 , @xmath64 , we can indeed reproduce ( 7 ) .
this is all well and good , but we shall presently see that universal gravitational coupling to the stress tensors of the gauge fields imposes a strong constraint on the curvature coefficients .
the appendix contains a full treatment of all possible property curvatures , consistent with the gauge symmetry .
identifying @xmath65 with @xmath66 and @xmath67 with @xmath68 , the metric @xmath69 in the @xmath54 sector equals @xmath70 /2.\ ] ] under the two independent local coordinate transformations , @xmath71 it is readily checked , from the rules of transformation of the metric tensor , that these simply correspond to independent gauge transformations , @xmath72 in turn this translates into @xmath73 which are the independent gauge variations of the more standard gauge fields @xmath21 and @xmath22 .
however we are faced with a serious problem when we introduce gravitation ... the dilemma which confronts us is best seen by working out the integral over the superscalar curvature @xmath56 , as summarized in the appendix : @xmath74}}{l^2\ell^2 } - \frac{e_l^2c_8}{\ell^2}f^{mn}_lf_{lmn } -\frac{e_r^2c_6}{l^2}f^{mn}_rf_{rmn } + c\right]\ ] ] where @xmath75 denotes a cosmological term , @xmath76 and @xmath77}$ ] is the purely gravitational scalar curvature .
if we then make the replacements @xmath78 we need to ensure cancellation of the cross @xmath79 term which , in turn , forces @xmath80 . finally the gauge field lagrangians have to couple universally to gravity through their stress tensors so we must further set @xmath81 as a proper normalization condition .
this is a very strong consequence of attaching force coupling constants to gauge fields in the metric itself .
_ this need for coupling constant equality seems to be an unwelcome result_. but a moment s reflection leads one to the conclusion that it is an inescapable feature which arises from merging the gauge fields and gravity in some kind of enlarged metric , _ not just in our scheme but in any other scheme based on the same concepts_. the significant point is that at a semiclassical level or @xmath1 but only their eigenvalues such as @xmath31 or @xmath32 ] the coupling constants must always accompany the gauge fields , via the frame vectors that latch on to the fermions , and this baggage is inevitably carried into the stress tensors that emerge from @xmath83 .
so the question arises : is this an impasse or not ?
we would suggest , like many others , that the couplings may very well be equal ( or are unified ) at some high energy scale @xmath84 gev or more , associated with @xmath85 , signifying @xmath86 , but that the observed values at low energy @xmath87 are due to running _ down _ in energy to laboratory scales ( corresponding to different gauge groups with their particle contents ) .
we would go so far as to assert that any theory which attempts to unify gravity with the other forces faces the same dilemma and the only reasonable solution is to demand equality of all coupling constants at planckian type scales .
the corollary is that if the continuation of the constants to low energy via the renormalization group does not match experiment , either the particle content associated with the gauge group is wrong or , more drastically , the extended metric theory is doomed , unless it is regarded as fully quantum and contains the chargelike operators within the frame vectors .
we extend spacetime with two property coordinates , @xmath42 and @xmath43 .
this gives superspace coordinates of @xmath88 .
our starting point to building the metric is the following metric distance for a flat 4 + 4 dimensional graded manifold : @xmath89 this results in the extended minkowski metric @xmath90 taking the form : @xmath91 notice that we have introduced two separate length scales @xmath85 and @xmath92 in order to give ourselves complete freedom and wriggle room .
consider spacetime dependent u(1)@xmath5u(1 ) phase transformations to the property coordinates @xmath42 and @xmath43 as follows : @xmath93 now if we want our metric @xmath90 to be a tensor it has to transform correctly . with @xmath94 dependent phase changes in property it does _ not_. therefore , as in the one coordinate case , we need to fix this by including gauge fields .
we do so by introducing the following upper - triangular frame vector : @xmath95 and its inverse or the vielbein ( with @xmath96 : @xmath97 whereupon the metric follows from @xmath98 ; explicitly we get , @xmath99 it is readily checked from the tensorial transformation of the metric that the changes ( b.1 ) correctly yield the gauge variations , @xmath100 [ note , there is no need to include gauge couplings @xmath101 and @xmath102 with the gauge fields @xmath103 and @xmath104 at this stage . in the text @xmath103 and @xmath104
have been replaced by @xmath105 and @xmath106 . ]
representation ( b.4 ) is not the end of the story .
we have the freedom to include terms in the metric which are u(1)@xmath5u(1 ) invariant , involving @xmath107 .
but in so doing we have to ensure that _ all _ components of @xmath83 transform correctly ; after long and careful analysis we find that the nonzero metric elements can be finally reduced to @xmath108\ ! + \ ! l^2
l_{m}l_{n}\bar\zeta_l \zeta_l ( 1\!+\!c_4 \bar\zeta_r \zeta_r ) \!+\ ! \ell^2
r_{m } r_{n}\bar\zeta_r \zeta_r ( 1\!+\ ! c_5 \bar\zeta_l \zeta_l ) , \\
g_{m\zeta_l } = & - i l^2 \bar\zeta_l l_{m } ( 1+c_4 \bar\zeta_r \zeta_r)/2,\\ g_{m\zeta_r } = & -i \ell^2 \bar\zeta_r r_{m}(1+c_5 \bar\zeta_l \zeta_l)/2,\\ g_{m\bar\zeta_l } = & -i l^2 l_{m}\zeta_l(1+c_4 \bar\zeta_r \zeta_r)/2,\\ g_{m\bar\zeta_r } = & -i \ell^2 r_{m}\zeta_r(1+c_5 \bar\zeta_l\zeta_l)/2 , \\
g_{\zeta_l\bar\zeta_l } = & l^2(1 + c_{6 } \bar\zeta_l \zeta_l + c_{4 }
\bar\zeta_r \zeta_r + c_{7 } \bar \zeta_l \zeta_l \bar\zeta_r \zeta_r)/2,\\
g_{\zeta_r \bar\zeta_r } = & \ell^2 ( 1 + c_{5 } \bar\zeta_l \zeta_l + c_{8 } \bar\zeta_r \zeta_r + c_{9 } \bar \zeta_l \zeta_l \bar\zeta_r \zeta_r)/2.\end{aligned}\ ] ] the ensuing inverse metric components read : @xmath109,\\ g^{m \zeta_l } & = i l^m \zeta_l ( 1 - c_2 \bar\zeta_r\zeta_r),\\ g^{m \zeta_r } & = i r^m \zeta_r ( 1 - c_1 \bar\zeta_l \zeta_l),\\ g^{m \bar\zeta_l } & = - i l^{m } \bar\zeta_l ( 1-c_2 \bar\zeta_r \zeta_r),\\ g^{m \bar\zeta_r } & = - i r^m \bar\zeta_r ( 1-c_1 \bar\zeta_l \zeta_l),\\ g^{\zeta_l\bar\zeta_l } & = 2[1-c_6 \bar\zeta_l\zeta_l - c_4 \bar\zeta_r\zeta_r + ( 2 c_6 c_4 - c_7 ) \bar\zeta_l
\zeta_l \bar\zeta_r \zeta_r]/l^2 - l^m l_m \bar\zeta_l\zeta_l ( 1- c_2 \bar\zeta_r\zeta_r),\\ g^{\zeta_r\bar\zeta_r } & = \!2 [ 1-c_5 \bar\zeta_l \zeta_l -c_8\bar\zeta_r\zeta_r+(2c_5 c_8-c_9)\bar\zeta_l\zeta_l\bar\zeta_r\zeta_r]/\ell^2 -r^m r_m \bar\zeta_r\zeta_r ( 1 - c_1 \bar\zeta_l \zeta_l),\\ g^{\zeta_l \zeta_r } & = - l^m r_m \zeta_l\zeta_r,\\ g^{\bar\zeta_l\bar\zeta_r } & = - l^m r_m \bar\zeta_l \bar\zeta_r,\\ g^{\zeta_l \bar\zeta_r } & = l^m r_m \zeta_l\bar\zeta_r,\\ g^{\zeta_r \bar\zeta_l } & = l^m r_m \zeta_r \bar\zeta_l.\\\end{aligned}\ ] ] in the text we have substituted @xmath65 by @xmath66 and @xmath67 by @xmath68 ; on the other hand , @xmath110 enter the normalization of the lagrangians for the left and right gauge fields separately .
the berezinian or superdeterminant requires some more work . following standard procedures the superdeterminant
is obtained from the graded supermatrix , @xmath111.\end{aligned}\ ] ] the frame vectors @xmath112 which produce the metric have the components : @xmath113,\\ { \cal e}_{m}{}^{\zeta_l } & = - i l_m \zeta_l(1+c_4\bar\zeta_r\zeta_r/2),\\ { \cal e}_{m}{}^{\zeta_r } & = - i r_m \zeta_r(1+c_5\bar\zeta_l\zeta_l/2),\\ { \cal e}_{m}{}^{\bar\zeta_l } & = i l_m \bar\zeta_l(1+c_4\bar\zeta_r\zeta_r/2),\\ { \cal e}_{m}{}^{\bar\zeta_r } & = i r_m \bar\zeta_r(1+c_5\bar\zeta_l\zeta_l/2),\\ { \cal e}_{\zeta_l}{}^{\zeta_l } & = { \cal e}_{\bar\zeta_l}{}^{\bar\zeta_l } = 1+\frac{c_6}{2 } \bar\zeta_l\zeta_l + \frac{c_4}{2 } \bar\zeta_r\zeta_r + ( \frac{c_7}{2 } - \frac{c_6 c_4}{4 } ) \bar\zeta_l\zeta_l\bar\zeta_r\zeta_r,\\ { \cal e}_{\zeta_r}{}^{\zeta_r } & = { \cal e}_{\bar\zeta_r}{}^{\bar\zeta_r } = 1+\frac{c_5}{2 } \bar\zeta_l\zeta_l + \frac{c_8}{2 } \bar\zeta_r\zeta_r + ( \frac{c_9}{2 } - \frac{c_5 c_8}{4 } ) \bar\zeta_l\zeta_l\bar\zeta_r\zeta_r.\\\end{aligned}\ ] ] also needed are : @xmath114 as a useful crosscheck , s@xmath115 , so all is as it should be .
now to find the lagrangian . using the palatini form of the ricci scalar@xcite we get : @xmath116 } - \frac{c_6}{l^2 } r{}^{m n } r{}_{m n } - \frac{c_8}{\ell ^2 } l^{mn } l_{m n } + c,\end{aligned}\ ] ] where @xmath117 , @xmath118 and + @xmath119 $ ] .
+ this is all very complicated but it demonstrates how many gauge invariant property curvature terms work their way into the system 9 in our case .
one therefore needs some guiding principle for cutting them down .
one possible suggestion is to make all the curvature pieces _ fully _
u(@xmath120 ) invariant with @xmath120 properties , _ except _ for those products that involve the gauge fields embedded in the frame vectors ; so in our instance one could think of setting @xmath121 leaving us with only four constants , we could reduce the whole down to @xmath122 and there would just be an overall property factor @xmath123 $ ] multiplying the property flat elements , including the gauge field elements @xmath124 in the @xmath54 sector . ] , one scale and one coupling constant . although this simplification really needs some fundamental justification , it can be used to make the four - property case of su(2)@xmath7u(1 ) more manageable and is even extensible to the direct product with chromodynamics .
the main point is this : the property invariant curvatures must be severely constrained before one can handle the full standard model ; we are uncertain how to do so at this stage .
99 s.l glashow , nucl .
* 22 * , 579 ( 1961 ) .
a. salam and j.c .
ward , phys .
* 13*,168 ( 1964 ) .
s. weinberg , phys .
lett . * 19*,1264 ( 1967 ) .
a. salam , _ eighth nobel symposium _ , ed .
n. svartholm , almquist and wiksell , ( 1968 ) .
r. delbourgo and p.d . stack , int . j. mod
29a * , 50023 ( 2014 ) .
stack and r. delbourgo , int .
phys . * 30a*,1550005 ( 2015 ) .
f. englert and r. brout , phys .
* 13 * , 321 ( 1964 ) .
higgs , phys .
* 13 * , 508 ( 1964 ) .
guralnik , c.r .
hagen and t.w.b .
kibble , phys .
* 13 * , 585 ( 1964 ) .
kibble , phys . rev . * 155 * , 1554 ( 1967 ) .
r. delbourgo , j. phys . *
a39 * , 14735 ( 2006 ) . | the standard model ascribes distinct properties to different chiralities of fermions .
we show how to incorporate this aspect in an extended spacetime - property framework involving two different attributes using a generalized metric which includes gauge fields as well as gravitation . because the gauge fields are accompanied by coupling constants , all such schemes , including ours , necessitate coupling unification at high energy to ensure universality of gravtitational interactions with matter . |
the notion that the fundamental structure of spacetime might be something other than a continuum has been around for many decades .
a particular type of modified spacetime structure , which in modern parlance has come to be known as non - commutative geometry , had in fact been studied as far back as the 1940 s @xcite .
recent developments in string theory suggest that this modification to geometry at high energy scales can arise in certain d - brane backgrounds with a non - trivial rr flux , resulting in a resurgence of interest in the phenomenological consequences of physics on non - commutative spaces .
the basic idea of non - commutative geometry is rather simple : instead of considering spacetime coordinates @xmath0 ( with @xmath1 ) to be described by c - number functions on some manifold , we consider our spacetime coordinates to satisfy some sort of structure relations @xmath2 = i o^{\mu\nu},}\ ] ] where the @xmath3 are hermitian operators , and @xmath4 is a hermitean operator antisymmetric in its indices .
several widely studied examples include the type of non - commutativity first introduced in @xcite @xmath5 = i i\theta^{\mu\nu},}\ ] ] where @xmath6 is the identity operator and @xmath7 is a non - degenerate antisymmetric tensor , and @xmath8 = i\kappa \epsilon_{ijk}\hat x^k,}\ ] ] ( with @xmath9 denoting spatial components ) where @xmath10 has dimensions of length , and corresponds to the scale of non - commutativity .
both examples arise in string theory when describing d - brane dynamics in non - trivial flux backgrounds @xcite . in this report
, we focus on the former relation ( [ ncg ] ) , and explore its consequences for inflationary cosmology , where we will find that we can tightly constrain the amount of spacetime non - commutativity present during inflation . the initial intuition that motivated this study , was that the uv / ir mode mixing present in theories such as those defined on non - commutative ( moyal ) spaces defined by ( [ ncg ] ) , could provide us a constraining lever on physics at high energy scales complimentary to the lever offered by the effect of modes being stretched by inflation .
what we find is that the presence of this effect present in this example is so severe as to already be ruled out by present observations ( if it had nt already been severely constrained for other reasons @xcite ) . in this sense
, the spirit in which we undertake this study is that of a first pass ( within a concrete and widely studied context ) , on the notion that uv / ir mode mixing may offer us a complimentary lever on very high energy physics in the cmb , which in this example , is used to rule out a class of modifications to space time structure at such energy scales .
we begin by reviewing those aspects of non - commutative field theory relevant for our study , after which we will consider inflation in the presence of background non - commutativity . after showing that the background inflaton dynamics are identical for the commutative vs the noncommutative cases ( in the event that non - commutativity is purely spatial ) , we derive the action for the scalar perturbations , where we find that loop effects render a non - local effective action .
we show that the net effect of spatial non - commutativity is to modify the dispersion relations for the propagation of the perturbations , such that mode propagation differs from the usual case for all wavelengths _ less _ than the critical wavelength @xmath11 , where @xmath12 is the non - commutativity length scale , and @xmath13 is the slow roll parameter .
we see that uv / ir mode mixing induced by loop effects communicates signatures of non - commutativty to scales far below the no n - commutativity scale independent of whether or not the theta tensor redshifts during inflation .
we will then compute the two point correlation functions for the metric fluctuations , and find that the presence of any amount of spatial non - commutativity to be inconsistent with the observed cmb spectrum .
in what follows , we offer a brief overview of those aspects of non - commutative field theory relevant for our investigation . we refer to the excellent review of ref .
@xcite for details not elaborated upon here . to start with ,
once we assume the structure relations @xmath14 = i i\theta^{\mu\nu}$ ] for our spacetime coordinates , we find that the net effect of spacetime non - commutativity is to deform the usual product on our manifold to the so - called star product , defined by @xmath15 for readers unfamiliar with this aspect of non - commutative physics , we offer a quick review of why this is so in appendix [ app - a ] .
field theories on non - commutative spaces can be defined in the usual way , except that now all products in the action are to be replaced with star products @xmath16,}\ ] ] where @xmath17 indicates that all fields are multiplied with star products .
it is a property of the star product that the quadratic part of any action remains unaffected @xmath18 as is easily checked by fourier expanding @xmath19 and @xmath20 and performing the spacetime integral .
therefore the quadratic part of the action is unaffected by the underlying non - commutative structure .
in particular , this means that the green s functions for the commutative and the non - commutative theories are identical , and suffer from the same short distance divergences @xcite .
consider now the @xmath21 theory in four dimensions : @xmath22.}\ ] ] as we see immediately , the underlying non - commutative structure affects only the interaction terms . working in fourier space , it is straightforward to show that the net effect of non - commutativity on the interaction vertices is to introduce an additional phase factor relative to the commutative case @xcite @xmath23 where @xmath24 is the @xmath25 momentum flowing into the vertex , and we employ the shorthand notation : @xmath26 one should note that @xmath27 is not invariant under arbitrary permutations of the @xmath24 , but only cyclic permutations .
this results in a relative phase factor between the two one loop self energy diagrams shown in fig . 1 .
these two diagrams make the 1-loop 1pi two point function to differ by a phase of the form @xcite @xmath28 where they would have yielded identical contributions in the commutative case ( up to symmetry factors ) . following the treatment of ref .
@xcite , we observe that non - commutativity appears to regulate the non - planar sector of the theory ( through the oscillating phase factor ) , whilst leaving the planar sector unaffected . in order to see the effects of this phase factor
, we use the schwinger reparameterization of the propagator @xmath29 which renders the loop momentum integrations gaussian .
we can evaluate these straight away to yield @xmath30 where we adopt the notation : @xmath31 regulating the small @xmath32 divergence with the cut - off @xmath33 $ ] yields @xmath34 from which we can immediately read that if @xmath35 is the momentum cut - off for the planar sector , then the non - planar sector has a cut - off : @xmath36 in the event that not all coordinates are non - commutative , @xmath37 is to be understood as being only constructed out of the non - commutative components of the 4 momenta .
we evaluate the contributions to the one loop effective actions as : @xmath38 we note from the form of eq .
( [ effc ] ) that in the limit @xmath39 , the non - planar contribution to the effective action diverges in the limit @xmath40 .
that is to say , the non - planar sector appears to inherit an ir divergence at low momenta from the uv divergence of the planar sector . in momentum space ,
the resulting effective action from the contributions of all self energy graphs is given by @xmath41 where @xmath42 is the renormalized mass given by : @xmath43 thus we see from eq .
( [ 1lea ] ) , that loop effects can communicate even high scale non - commutativity to large wavelengths , as is evidenced by the non - local term proportional to the self coupling , which becomes significant for low momenta .
we take note of a similar effect for @xmath44 theory in 4-d .
if we denote the ( now dimensionful ) coupling constant as @xmath20 , we see from fig . 2 that again there is a planar and a non - planar sector to the one loop self energy .
it follows for similar reasons that the planar and the non - planar sectors contribute differently to the one loop effective actions and yield @xcite @xmath45 where now , unlike the case of @xmath21 theory , the leading singular term at low momenta for the one loop effective action for @xmath44 theory is only logarithmically divergent .
we note in passing that although the ir divergent terms in the low energy effective action were derived in the context of imposing a cut - off regularization and sending this cut - off to infinity , the ir divergence derived above are to be taken to be as generic features of theories which exhibit strong uv / ir mixing .
in fact , in @xcite it was shown in the context of @xmath21 theory that quadratic ir divergences result independent of the scale of the cut - off @xmath35 . in the next section ,
we investigate the consequences of such loop effects ( uv / ir mixing ) for inflation in the presence of any underlying non - commutativity , and compute two point functions for the scalar fluctuations .
we now commence our analysis of the effects of non - commutativity on cmb observables .
first , we posit that there be no temporal non - commutativity ( @xmath46 ) and consider = i^ij , [ spat - non - commutativity ] ( with @xmath9 ) which is necessary if we are to preserve unitarity @xcite as well as requiring that our construction be consistently realized in string theory . with this in mind , it is possible to see that the dynamics for homogeneous field configurations @xmath47 remain unaffected by non - commutativity . in particular , in a friedmann - robertson - walker ( frw ) background ds^2 = - dt^2 + a^2(t ) _ ij dx^i dx^j , [ frw - metric ] the equations of motion @xmath48 reduce to the usual commutative equations ( with @xmath49 ) @xmath50 for homogeneous field configurations @xmath47 , given eq .
( [ sp ] ) and the fact that @xmath51 : @xmath52 an alternative way to see this is to consider the corresponding weyl operator equations of motion ( see appendix [ app - a ] ) + 3h + v ( ) = 0 , and realizing that any commutative homogeneous solution @xmath53 is a solution at finite theta by considering the matrix solutions @xmath54 . hence given any purely spatial non - commutativity during inflation
, we will have the usual solutions for the background .
inflaton perturbations @xmath55 , on the other hand , will feel the underlying non - commutativity and as shown in appendix b , will result in the following action for the comoving curvature perturbations @xmath56 : @xmath57\end{aligned}\ ] ] where @xmath58 involves only spatial derivatives and @xmath59 is a dimensionless parameter given by : @xmath60 where @xmath61 is taken to be commensurate to the slow roll parameters @xmath13 and @xmath62 , and its precise value will depend on the details of the background solution . in the previous expression @xmath63
is a number of order unity which depends on the specifics of the underlying inflationary model , whereas @xmath13 and @xmath62 are the usual slow roll parameters : @xmath64 the action of eq .
( [ pertac ] ) is derived in appendix b [ eq .
( [ pertactss ] ) ] and includes one - loop effects arising from non - commutativity as a consequence of higher order interactions in the inflaton perturbations proportional to @xmath13 . at short wavelengths ,
the order of non - commutative effects comes determined by the size of the lengthscale @xmath65 , entering the non - commutative structure tensor @xmath66 .
it is rather clear that the validity of slow - roll inflationary framework in a non - commutative background requires us to have l _ h 1 , [ cond - desitter ] which simply means that the de sitter radius @xmath67 determining the size of causally connected regions during inflation has to be much larger than the non - commutative lengthscale @xmath65 .
we also note here that eq .
( [ pertac ] ) is a one loop effective action , which diverges at low momenta ( and where higher loop effects become significant ) , has a domain of validity which does not stretch arbitrarily far into the ir .
as demonstrated in @xcite , this domain of validity is given by @xmath68 where @xmath69 is of order unity , and as shown in the appendix , @xmath70 .
for any reasonable model of inflation , both @xmath71 and @xmath61 are smaller than @xmath72 , therefore making the exponent on the right hand side almost vanishing , and certainly all scales of cosmological interest fall within the domain of validity of our approximation .
we now study the evolution of perturbations during inflation in the presence of non - commutativity , and its consequences for the spectrum of scalar fluctuations . in what follows ,
we rewrite the metric of eq .
( [ frw - metric ] ) and work in conformal coordinates @xmath73 where @xmath62 is the usual conformal time satisfying @xmath74 . furthermore , to simplify our analysis , we will consider the evolution of perturbations on a de sitter background @xmath75 , where @xmath67 corresponds to the de sitter radius , and consider the approximation whereby @xmath76 evolves adiabatically . from eq .
( [ pertac ] ) the equation of motion for the curvature perturbations is found to be and the frw form of the metric tensor ( [ ccs ] ) , the operator @xmath77 is self - adjoint with respect to the scalar product @xmath78 , and can hence be brought over to act on the partner field in any bi - linear product .
( [ peom ] ) then follows directly .
] @xmath79 by using the metric of eq .
( [ ccs ] ) and realizing that @xmath80 is purely spatial , we can write the above in momentum space : - + ( k^2 + ) = 0 . [ mode - eq ] since @xmath81 has to have even rank @xcite , we pick two directions in which @xmath82 .
that this choice breaks isotropy is inevitable . in @xcite the consequences of this preferred direction for the cmb was explored , and in @xcite , its consequences for non - gaussian imprints in the cmb were considered .
we wish to show in this report is that if indeed non - commutativity is present , these effects are subsidiary , and are dwarfed by the fact that cmb power is greatly suppressed at long wavelengths by the effects of non - planar uv / ir mixing .
that is , we are not concerned with these aspects of what we view as model specific features .
instead , we are more interested in exploring what we view as a generic consequences of uv / ir mixing in the context of a concrete model , and in doing so , deriving constraints on the presence of non - comutativity before , or during inflation . to make our analysis as exhaustive as possible , in this report we consider various possible time dependences of the tensor @xmath81 at the right hand side of eq .
( [ spat - non - commutativity ] ) . to be more precise and without loss of generality
we will take the coordinates @xmath83 in eq .
( [ spat - non - commutativity ] ) to coincide with the co - moving coordinates of the metric in eq .
( [ ccs ] ) and consider , separately , the three different possibilities whereby either @xmath81 , @xmath84 or @xmath85 are constant tensors ( i.e. independent of the metric ) , and indices are raised with the spatial part of the metric @xmath86 . in the context of string theory , different time dependences of @xmath81 can be traced back to different realizations of inflation .
supposing , for instance , that the inflaton is a transverse modulus describing the end points of open strings ( d - branes ) in non - trivial @xmath87-field backgrounds , the open string endpoints ( brane coordinates ) will satisfy the moyal type non - commutative structure relations @xcite : @xmath88 where @xmath20 is the string worldsheet metric and @xmath87 is the worldsheet anti - symmetric field ( with dimensions of inverse length squared ) , and @xmath89 denotes anti - symmetrization .
we note that expanding in powers of @xmath90 , to leading order , two factors of the metric will contract @xmath91 , leading to @xmath92 , and thus we see that in the particular case where the background is characterized by a constant @xmath91 field , we recover a constant @xmath85 .
one can imagine , however , a situation in which inflation is being realized as the result of a moving d - brane on a warped background with @xmath91 scaling spatially accordingly . in such a case ,
the effective time variation of @xmath91 on the brane would be such that @xmath93 ( instead of @xmath85 ) remains constant .
independently of the details of how inflation is actually realized , in what follows consider the three aforementioned cases .
the first case of interest consists of the choice in which @xmath81 is independent of the metric . by writing @xmath94 we
then obtain ^i _
j ^j _ m k^m k_i = a^2 ^2 |k^2 = , [ theta - case-1 ] where @xmath95 is the momentum squared along the non - commuting directions .
notice that we can rewrite the non - commutative relation in terms of _ proper _ coordinates @xmath96 in the following way : = i a^2 ( ) ^ij .
this means that during inflation the proper noncommutative length scale @xmath65 grows with time as @xmath97 where @xmath98 is a co - moving length scale . by inserting eq .
( [ theta - case-1 ] ) in eq .
( [ mode - eq ] ) we arrive to the following equation of motion : @xmath99 we immediately read off the solutions to this equation , as they only differ from the standard canonically normalized mode functions in their @xmath100 dependence @xmath101 with @xmath102 defined as @xmath103 we take note of the following duality in the above @xmath104 which has a simple interpretation in terms of the scattering of localized @xmath105 wave packets upon self - interaction @xcite .
we can immediately see how this will lead to an unacceptable modification to the cmb spectrum , namely , that any such duality will cause the cmb to be symmetric around the pivot wavelength : @xmath106 note that even in the case of very high scale non - commutativity , even as high as planckian , this pivot wavelength will be below this scale ( for @xmath13 and @xmath107 , this scale becomes the gut scale ) .
any feature of the spectrum at short wavelengths will be exactly mirrored at long wavelengths .
hence , if the normalization of the mode functions vanishes towards the uv , it will also do so in the ir .
this will result in suppression of power at low multipoles , as is easily seen by using standard techniques to compute the resulting power spectrum for curvature perturbations ( defined in appendix [ app - b ] ) @xmath108 which results in the power spectrum : @xmath109 we easily read off that the two point function will start to receive significant corrections _
below _ the wavlength ( [ kpiv ] ) .
if we impose cobe normalization at the comoving wavelength @xmath110 via observational input , the above implies in order for non - commutativty to not affect this normalization , that @xmath111 in other words , non - commutativity at cosmological scales , a clearly untenable state of affairs . viewed from another perspective ,
if non - commutativity were present at some high energy scale before , or during inflation , the resulting spectrum of metric fluctuations would have vanishing power at all observable scales that is , the main feature of inflation that inspires our belief in it would be nullified by non - commutativity and we d have to search elsewhere for the origins of large scale structure .
this ` stiffening ' of long wavelength fluctuations is a well known and widely studied feature of non - commutative field theories ( see @xcite and references therein ) , and we see here that in the context of inflation , would nullify the mechanism with which it generates the primordial fluctutations responsible for large scale structure . if the reader as doubts at this point that this may be an artifact of the fact that we considered the case where the theta tensor is independent of the metric tensor , in the next section we demonstrate that this stiffening of long wavelength modes is even more severe in the opposite case .
the second case of interest corresponds to the situation in which @xmath80 ( with both upper and lower indices ) is independent of the metric . notice that in this case one has = i g^i k _ k^ j = i a^-2 ( t ) _
i^ j , and therefore the non - commutative relation expressed in _ proper _ coordinates @xmath96 is given by : = i _
i^ j , which means that the proper noncommutative length scale @xmath112 remains constant during the inflationary stage .
if , similarly to the previous case , we write @xmath113 , we are then left with @xmath114 and the equations of motion ( [ mode - eq ] ) for the mode functions take the form - + ( k^2 + ) = 0 .
[ second - case - eq ] notice that the term arising from the 1-loop contribution becomes dominant as @xmath115 .
it is rather clear that this term can not dominate the dynamics for modes at any stage during inflation , if these are to be held responsible for the cmb anisotropies observed today . to appreciate this ,
let us suppose that this term indeed dominates .
then , in the long wavelength limit , the canonically normalised solution to the equations are found to be _ = e^ e^k x , [ second - case - sol ] where @xmath116 is given by c^2(k ) = . from this
, one can immediately evaluate the two point correlation function of curvature perturbations @xmath117 , which reads : [ 2pcc2 ] ( k)(k ) = ( 2)^3 ^3(k - k ) . at this point
, it should be clear that the power spectrum deduced out of this quantity is strongly dependent on the scales @xmath118 via @xmath116 , loosing one of the most important features found in slow - roll inflation , namely the prediction of a nearly scale invariant power spectrum .
furthermore , we note from ( [ second - case - eq ] ) that the curvature perturbation no longer ceases to evolve after horizon crossing .
instead , it is damped until the end of inflation , after which non - commutative effects are no longer felt and the curvature perturbation proceeds conserved until it re - enters our horizon as imprints on the cmb .
hence the observed power spectrum , on top of its shape being highly distorted , is extremely suppressed : @xmath119,}\ ] ] where @xmath120 is the number of e - folds the mode @xmath100 undergoes after it exits the horizon until inflation ends , whence it is the appearance of the factors of @xmath62 in ( [ 2pcc2 ] ) that result in this drastic suppression ( and blueshifting of the spectrum ) .
it is important to notice that , as @xmath115 , the momenta of the mode solutions is redshifted until eq .
( [ domval ] ) , determining the validity of this regime , is no longer satisfied . at this point ,
higher loop contributions start kicking in , modifying the asymptotic solution expressed in eq .
( [ second - case - sol ] ) .
it is nevertheless clear that these higher order loop effects will not restore the scale invariance of the power spectrum .
we can in fact deduce a strong constraint on the parameter @xmath121 by requiring that 1-loop contributions to the equation of motion ( [ second - case - eq ] ) do not affect the evolution of modes presently observed at the cmb . for this
, let us consider those modes observed in the cmb with the longest wavelengths
. roughly speaking , these modes crossed the horizon about @xmath122 @xmath123-folds before the end of inflation , the precise number depending on the specific model under consideration .
as we have seen , 1-loop corrections can drastically affect the evolution of these modes outside the horizon , and therefore the term @xmath124 in eq .
( [ second - case - eq ] ) has to remain suppressed during the whole @xmath122 @xmath123-folds following the horizon crossing @xmath125 . in order to get an estimate of how suppressed this term has to be ,
let us consider the equations of motions ( [ second - case - eq ] ) in terms of the re - escaled field @xmath126 + ( k^2 - + ) = 0 , [ second - case - eq - chi ] where , in order to obtain a conservative bound , we have for simplicity replaced @xmath127 .
then , the requirement of having a suppressed 1-loop contribution affecting the evolution of modes during the last 55 @xmath123-folds after horizon crossing @xmath128 is given by < e^- 2
n , [ bound-1 ] where @xmath122 . to proceed ,
let us recall that @xmath129 .
moreover , as shown in the appendix b , there is at least one contributing term in @xmath59 of order @xmath130 , which for the present discussion we treat as the leading piece . to be of order @xmath131 @xcite , therefore implying that @xmath62 is at least of that same order ( recall that @xmath132 ) .
@xmath13 on the other hand is not well constrained yet and might be much smaller than @xmath133 . ] then , since the spectral index is given in terms of the slow roll parameters as @xmath134 , we can therefore write @xmath135
. moreover , provided that eq .
( [ bound-1 ] ) holds , one can use the cobe normalization of the spectrum and write @xmath136 to eliminate @xmath13 in terms of @xmath76 .
putting all of this together , we then obtain : l_^-1
< 10 ^ -2 e^-n/2 |n_s-1|^-1/4 m_pl .
given that the spectral index is measured to be @xmath137 @xcite , we obtain the following bound on the non - commutative lengthscale : @xmath138 that is , the non - commutative energy scale is required to be close to current high energy experiments .
moreover , from eq .
( [ cond - desitter ] ) we see that the inflationary scale had to be well within reach of current observations .
finally , consider now the non - commutative relation ( [ mode - eq ] ) in the case where @xmath139 with @xmath85 a constant tensor .
as before , we introduce @xmath121 by writing @xmath140 , which allows us to deduce : ^i _ j ^j _ m k^m k_i = h^6 ^6 ^2 |k^2 , [ theta - case-2 ] notice that in this case the proper coordinates @xmath141 satisfy the following non - commutative relation = i a^-2 ( ) _ ij , and therefore the proper non - commutative lengthscale decreases during inflation as @xmath142 , where @xmath143 .
interestingly , even though this lengthscale is becoming smaller as inflation progresses , due to the way in which ir effects are accentuated by inflation , a non - zero @xmath144 produces drastic modifications of the spectrum , as we see in the following .
first , by replacing eq .
( [ theta - case-2 ] ) in eq .
( [ mode - eq ] ) we obtain @xmath145 as before , if the 1-loop contribution dominates ( @xmath146 ) , we can ignore the first term to obtain the canonically normalized solutions @xmath147 with @xmath116 ( which has dimensions inverse mass cubed ) is defined by : @xmath148 from this , one can immediately evaluate the two point correlation function in the long wavelength limit to find : [ fc2p ] ( k)(k ) = ( 2)^3 ^3(k - k ) .
as before , we see that power spectrum deduced from this expression is not only heavily suppressed , but it is heavily dependent on the scales due to the scale dependence posed by @xmath116 .
we can now repeat the previous analysis and ask ourselves what the value of @xmath121 should be in order to avoid modifications to the evolutions of modes observed at the cmb after they left the horizon . to this
extent , notice that since the proper lengthscale @xmath149 is varying with time , then @xmath150 determines the value of @xmath65 at the time @xmath151 where @xmath152 , which we are free to specify .
let us choose @xmath151 to correspond to the epoch in which the longest wavelengths present in the cmb crossed the horizon [ recall the discussion immediately before eq .
( [ second - case - eq - chi ] ) ] . at the end of inflation @xmath153 , with @xmath154 , and therefore , in order to avoid having modifications to the power spectrum coming from non - commutative effects we must require l_0 ^ -1 <
10 ^ -2 e^-3n/2 |n_s-1|^-1/4 m_pl , which , after replacing @xmath155 and the value of the spectral index , gives : l_0 ^ -1 < 10 ^ -20 gev .
since eq .
( [ cond - desitter ] ) must be valid at any time during inflation , and in particular at the epoch @xmath151 where our reference perturbations are crossing the horizon , the previous bound additionally implies @xmath156 .
we therefore conclude that this type of non - commutativity is incompatible with slow - roll inflation as a mechanism to generate the anisotropies presently observed at the cmb .
in addition , in the present case as well , we see from ( [ mfeq2 ] ) that the curvature perturbations continue to be damped after horizon crossing until inflation ends .
hence we compute in an entirely analogous manner to ( [ psc2 ] ) , the power spectrum observed in the cmb to be : @xmath157 where again we note an extremely damped , blue tilted spectrum ( @xmath158 as @xmath159 , since @xmath120 is the number of e - folds inflation lasts after the mode @xmath100 crosses the horizon ) .
hence long wavelength power is greatly suppressed by the stiffening of modes even in the case where the non - commutativity tensor is redshifted by the expansion of spacetime .
this is to be viewed as an analogous situation to the discontinuity in @xmath121 found in other contexts ( see discussion surrounding ( [ bfncq ] ) ) . in summary , we have seen how uv / ir mode mixing due to the underlying non - commutative structure of spacetime drastically modifies the predictions of inflation regardless of the metric dependences of the non - commutativity tensor .
the stiffening of long wavelength modes is a well known feature of non - commutative theories @xcite , and here we have seen how it can nullify one of the more successful features of inflation .
slow - roll inflation provides a powerful framework to address speculative physics at very high energies . in this paper
we have investigated the particular case of non - commutative geometry and its implications for the generation of scalar perturbations during inflation .
we have shown that the phenomenon of uv / ir mode mixing , arising from one - loop corrections to the inflaton quadratic action , affects the evolution of fluctuations after horizon crossing in an unacceptable way . to make our study as general as possible ,
we have considered three different situations , characterized by different time evolutions of the non - commutative lengthscale @xmath65 .
interestingly , in the cases where @xmath84 and @xmath85 are constant , we find that counter to the intuition that during inflation , the rapidly decaying theta tensor in the right hand side of eq .
( [ spat - non - commutativity ] ) would render any effects of non - commutativity null , we find the effects of uv / ir mode mixing to be _ even more pronounced_. the reason for this is well understood , and comes down to the fact that many features of non - commutative theories are different in the @xmath160 limit than if there was no commutativity to begin with , such as various amplitudes @xcite and the phase structure of the theory @xcite among others .
a rather stark demonstration of this comes from when one considers the beta function for non - commutative qed @xcite , @xmath161 with @xmath162 the number of dirac fermion species .
the new negative term arises from non - commutative photon self interactions , and is independent of @xmath121 , provided non - commutativity exists in the first place . in an analogous manner , we find that even if the theta tensor is being rapidly redshifted , the consequence for the cmb two point function is rather pronounced .
we thus conclude , echoing the sentiment expressed in the preliminaries , that as far as very high energy physics is concerned , the phenomenon of uv / ir mode mixing offers us a complimentary lever ( on high energy physics ) to the effects of inflation blowing up modes .
however as we have seen in this context , the effects are so pronounced and in conflict with observations as to strongly disfavour the underlying model with manifests this mode mixing .
s.p . wishes to thank the lorentz institute at leiden university , and the ccpp at nyu , and gregory gabadadze in particular for hospitality during the preparation of this report .
g.a.p . wishes to thank the theory group at the humboldt university for similar hospitality .
is supported by the netherlands organization for scientific research ( n.w.o . ) under the vici and vidi programmes .
is supported at the humboldt university with funds from the volkswagon foundation , and is grateful to jan plefka for this .
the purpose of this appendix is to demonstrate why ( [ sp ] ) represents the structure relations ( [ ncg ] ) on function space . starting with relations @xmath163
= i\theta^{\mu\nu},}\ ] ] where each @xmath3 is a hermitean operator , we define the bounded operators @xmath164 these operators satisfy the relations : t^+(k ) & = & t(-k ) , [ rel1 ] + t(k)t(k ) & = & t(k + k)e^k_k_^ , [ rel2 ] + tr t(k ) & = & _ ( k _ ) , [ rel3 ] where ( [ rel2 ] ) follows from the campbell - baker - hausdorf relations , and the trace in ( [ rel3 ] ) is taken over the appropriate hilbert space spanned by the representation ( in 2-d , it would correspond to the usual hilbert space used to represent the canonical pair @xmath165 and @xmath166 , where @xmath167 = i$ ] ) .
next , we associate to any classical function @xmath168 , the operator @xmath169 @xmath170 where @xmath171 is the fourier transform of the function @xmath168 . in this sense
, we can consider the operator @xmath169 as the object that would result from fourier transforming back from @xmath171 , except that we convolute with ( [ opdef ] ) instead of the usual c - number phase .
using the trace relation ( [ rel3 ] ) , we can recover @xmath168 from @xmath169 : @xmath172 so that we have established a one to one and onto map between the space of c - number functions and operators over some hilbert space .
this map effects the standard prescription of ` putting hats ' on the c - number functions : @xmath173 when products of @xmath174 occur that are ambiguous , the above map returns the ` weyl ordered ' or totally symmetrized product of operators , hence this mapping is referred to as the weyl correspondence .
now we know that on the hilbert space side of this correspondence , there exists the usual algebra of linear operators .
that is , any two operators @xmath175 and @xmath176 multiply as @xmath177 .
we can then ask what c - number function this operator maps back onto , and denoting this function as @xmath178 , we find after some straightforward manipulations that : @xmath179\\ & = & e^{\frac{i\theta^{\mu\nu}}{2}\frac{\partial}{\partial x^\mu}\frac{\partial}{\partial y^\nu}}\phi_1(x ) \phi_2(y)|_{y = x } .\end{aligned}\ ] ] hence in order for this one to one map between c - numbers and operators to be a homomorphism ( such that we reproduce the operator algebra on the function space ) , we see that we have to deform the usual product into the so called star product as outlined above . in this way
, we can encode the non - commutative spacetime structure implied by ( [ rcomm ] ) on our function space by simply replacing all commutative products with the star product ( [ sp ] ) .
in this appendix , we derive the relevant quadratic terms of the effective action for cosmological perturbations in the presence of spatial non - commutativity . as we have seen in the preliminaries for this report
, loop effects can drastically modify the structure of the quadratic part of the one loop effective action at low momenta relative to the commutative case .
we shall demonstrate a similar phenomenon for cosmological perturbations _ even if loop effects are insignificant in the corresponding commutative situation_. in order to calculate this effect , we refer to the formalism and results of @xcite . we begin by
considering the action for the metric perturbations up to third and fourth order , after which we consider which terms contribute the leading contributions to the ( quadratic ) one loop effective action . since we are interested in computing corrections to the cmb observables that arise from non - planar uv / ir mixing , we focus on the terms which are leadingly singular in the ir , as it is these term which determines the leading corrections to the mode functions that would imprint themselves on the cmb . following @xcite , we begin with the adm decomposition @xmath180 and work in the uniform curvature gauge : @xmath181 we start with the action : @xmath182,}\ ] ] where we note that in this convention , the field @xmath183 is dimensionless . upon expanding the einstein - hilbert action coupled to the inflaton field to first order , the shift and the lapse functions @xmath184 and @xmath185 turn out to be lagrange multipliers .
writing @xmath186 and @xmath187 where the subscripts denote the order of the perturbation , we find that : @xmath188 which when solved for and plugged back into the action , results ( after many integrations by parts ) in @xcite @xmath189,}\ ] ] where @xmath190 and the background metric @xmath191 is given by the associated line element @xmath192 and the slow roll parameter @xmath13 has been defined as @xmath193 [ recall that now @xmath194 appears explicitly in the definition of the scalar potential in eq .
( [ saction ] ) ] . the action to third order
is found to be : @xmath195 .
\label{3pa}\end{aligned}\ ] ] this is expression ( 3.9 ) in @xcite after we employ the field redefinition ( 3.10 ) in the same reference .
we note from ( [ chi ] ) that @xmath196 has mass dimension @xmath197 . in the non - commutative case , where the non - commutativity is a constant purely spatial tensor ( so integrating by parts over time yields no extra terms )
, we simply replace all products in the above with the star product ( [ sp ] ) , as inflaton field fluctuations feel this non - commutativity being the modulus parametrizing the locus of open string endpoints .
background quantities do not feel this non - commutativity as it is purely spatial . in the resulting expression
, we would like to extract the terms that contributes to the leading divergence in the ir for the quadratic part of the one loop effective action .
as seen in the preliminaries , non - planar cubic vertices contribute at most logarithmic ir divergences , unless there are derivatives acting on the vertex . to this end , we begin with the first term in the above as it has the most derivatives that could run along a loop ( recalling from ( [ chi ] ) that @xmath198 ) to leading order in @xmath13 ( the only other terms which have as many derivatives are subleading in @xmath13 ) .
that is , we consider the interaction vertex @xmath199 where @xmath200 is the metric tensor multiplied by the matrix @xmath201 .
however , it is easily seen that integrating this term by parts yields terms proportional to @xmath202 , @xmath203 and terms of fourth order . hence at third order , we only have to consider terms formally corresponding to the to the last three lines of ( [ 3pa ] ) . to leading order
, these terms will only contribute logarithmic ir divergences , as can also be seen by way of the field redefinition : @xmath204 + ... } \ ] ] as shown in @xcite , in terms of @xmath205 , ( [ 3pa ] ) becomes : @xmath206 given that we need two vertices of this interaction to construct the non - planar diagram in fig 2 , with the relevant diagrams involving only time derivatives acting on the external legs , we see that this diagram yield at most logarithmically ir divergent contributions . at fourth order
, we consider the action for @xmath105 as computed by @xcite , where through similar considerations as above , we extract the leading relevant terms : @xmath207\varphi^4 + ... \end{aligned}\ ] ] we symbolically denote the above interaction vertex as @xmath208 where @xmath61 is taken to be commensurate to the slow roll parameters @xmath13 , @xmath62 , etc ... , and its precise value will depend on the details of the background solution .
we now realize that non - planar one loop diagrams such as those considered in fig 1 are generated by this vertex .
this will contribute a non - local counter term of order @xmath209 from the loop integral : @xmath210 with the resulting one loop effective action for the perturbations ( up to a factor of order unity @xmath63 , which depends on the precise details of the background dynamics ) : @xmath211.}\ ] ] converting the above expression into the action for the curvature perturbation @xmath212 using the relation @xcite ( which is valid at the quadratic level in cour case as well as the non - commutativty tensor does not contract any time derivatives ) : @xmath213 we finally obtain the quadratic action for the curvature perturbation as : @xmath214,}\ ] ] where = .
to finish , we notice that disregarding the last two terms in eq .
( [ s4r ] ) , we are left with @xmath215 .
h. s. snyder , `` quantized space - time , '' phys .
* 71 * , 38 ( 1947 ) .
v. schomerus , `` d - branes and deformation quantization , '' jhep * 9906 * , 030 ( 1999 ) [ arxiv : hep - th/9903205 ] .
r. c. myers , `` dielectric - branes , '' jhep * 9912 * , 022 ( 1999 ) [ arxiv : hep - th/9910053 ] .
r. h. brandenberger , `` trans - planckian physics and inflationary cosmology , '' arxiv : hep - th/0210186 .
r. c. helling and j. you , `` macroscopic screening of coulomb potentials from uv / ir - mixing , '' jhep * 0806 * , 067 ( 2008 ) [ arxiv:0707.1885 [ hep - th ] ] .
s. a. abel , j. jaeckel , v. v. khoze and a. ringwald , `` photon defects in noncommutative standard model candidates , '' j. phys .
53 * , 843 ( 2006 ) [ arxiv : hep - ph/0606106 ] . m. r. douglas and n. a. nekrasov , `` noncommutative field theory , '' rev .
phys . * 73 * , 977 ( 2001 ) [ arxiv : hep - th/0106048 ] .
t. filk , `` divergencies in a field theory on quantum space , '' phys .
b * 376 * , 53 ( 1996 ) .
s. minwalla , m. van raamsdonk and n. seiberg , `` noncommutative perturbative dynamics , '' jhep * 0002 * , 020 ( 2000 ) [ arxiv : hep - th/9912072 ] . j. gomis and t. mehen , `` space - time noncommutative field theories and unitarity , '' nucl .
b * 591 * , 265 ( 2000 ) [ arxiv : hep - th/0005129 ] .
s. s. gubser and s. l. sondhi , `` phase structure of non - commutative scalar field theories , '' nucl .
b * 605 * , 395 ( 2001 ) [ arxiv : hep - th/0006119 ] .
m. panero , `` numerical simulations of a non - commutative theory : the scalar model on the fuzzy sphere , '' jhep * 0705 * , 082 ( 2007 ) [ arxiv : hep - th/0608202 ] ; m. panero , `` quantum field theory in a non - commutative space : theoretical predictions and numerical results on the fuzzy sphere , '' sigma * 2 * , 081 ( 2006 ) [ arxiv : hep - th/0609205 ] .
e. akofor , a. p. balachandran , s. g. jo , a. joseph and b. a. qureshi , `` direction - dependent cmb power spectrum and statistical anisotropy from noncommutative geometry , '' jhep * 0805 * , 092 ( 2008 ) [ arxiv:0710.5897 [ astro - ph ] ] .
e. akofor , a. p. balachandran , a. joseph , l. pekowsky and b. a. qureshi , `` constraints from cmb on spacetime noncommutativity and causality violation , '' phys .
d * 79 * , 063004 ( 2009 ) [ arxiv:0806.2458 [ astro - ph ] ] . c. s. chu , b. r. greene and g. shiu , `` remarks on inflation and noncommutative geometry , '' mod .
a * 16 * , 2231 ( 2001 ) [ arxiv : hep - th/0011241 ] . f. lizzi , g. mangano , g. miele and m. peloso , `` cosmological perturbations and short distance physics from noncommutative geometry , '' jhep * 0206 * , 049 ( 2002 ) [ arxiv : hep - th/0203099 ] . k. fang , b. chen and w. xue , `` non - commutative geometry modified non - gaussianities of cosmological perturbation , '' phys .
d * 77 * , 063523 ( 2008 ) [ arxiv:0707.1970 [ astro - ph ] ] .
e. komatsu _ et al .
_ [ wmap collaboration ] , `` five - year wilkinson microwave anisotropy probe ( wmap ) observations : cosmological interpretation , '' astrophys . j. suppl . * 180 * , 330 ( 2009 ) [ arxiv:0803.0547 [ astro - ph ] ] .
j. m. maldacena , `` non - gaussian features of primordial fluctuations in single field inflationary models , '' jhep * 0305 * , 013 ( 2003 ) [ arxiv : astro - ph/0210603 ] . m. s. sloth , `` on the one loop corrections to inflation and the cmb anisotropies , '' nucl . phys .
b * 748 * , 149 ( 2006 ) [ arxiv : astro - ph/0604488 ] ; d. seery , j. e. lidsey and m. s. sloth , `` the inflationary trispectrum , '' jcap * 0701 * , 027 ( 2007 ) [ arxiv : astro - ph/0610210 ] .
c. p. martin and d. sanchez - ruiz , `` the one - loop uv divergent structure of u(1 ) yang - mills theory on noncommutative r**4 , '' phys .
lett . * 83 * , 476 ( 1999 ) [ arxiv : hep - th/9903077 ] . | it is well understood that spatial non - commutativity , if indeed realized in nature , is a phenomenon whose effects are not just felt at energy scales comparable to the non - commutativity scale .
loop effects can transmit signatures of any underlying non - commutativity to macroscopic scales ( a manifestation of a phenomenon that has come to be known as uv / ir mode mixing ) and offer a potential lever to constrain the amount of non - commutativity present in nature , if present at all .
field theories defined on non - commutative spaces ( realized in string theory when d - branes are coupled to backgrounds of non - trivial rr background flux ) , can exhibit strong uv / ir mode mixing , manifesting in a non - local one loop effective action . in the context of inflation in the presence of any background non - commutativity , we demonstrate how this uv / ir mixing at the loop level can allow us to place severe constraints on the scale of non - commutativity if we presume inflation is responsible for large scale structure .
we demonstrate that any amount of non - commutativity greatly suppresses the cmb power at all observable scales , _ independent _ of the scale of inflation , and independent of whether or not the non - commutativity tensor redshifts during inflation , therefore nullifying a very salient and successful prediction of inflation . |
clusters of galaxies are the largest evolved structures in the universe and , as such , qualify for being important cosmological probes .
the abundance of galaxy clusters provides sensitive constraints on the cosmological parameters that govern the growth of structures in the universe @xcite . in this regard ,
x - ray observations provide a useful tool for identifying and studying galaxy clusters @xcite .
galaxy clusters can also be observed in the microwave band through sunyaev - zeldovich ( sz ) effect @xcite , which results from the up - scatting of cosmic microwave background ( cmb ) photons by hot electrons in the intracluster medium ( icm ) .
the sz effect has a unique property that unlike x - ray emission it is independent of redshift and does not suffer from cosmological dimming . with the current and upcoming data from planck ,
the sz cluster surveys have become a robust probe for determining cosmological parameters and global properties of icm @xcite .
however , in order to obtain robust cosmological estimates using such techniques one requires the precise knowledge of the evolution of galaxy clusters with redshift and the thermodynamical properties of icm . in the simplest case , where one considers a pure gravitational collapse , the cluster scaling relations are expected to follow simple self - similarity @xcite .
correlations between the x - ray properties are widely used to probe the self - similarity in the galaxy clusters .
for example , the luminosity - temperature ( @xmath18 ) relation for self - similar models predict a shallower slope ( @xmath19 ) than observed ( @xmath20 ) @xcite implying a break in the self - similarity in galaxy clusters .
similarly , studies of the scaling relations involving sz effect also show discrepancies between observations and predictions from a pure gravitational model @xcite .
such studies have revealed the importance of the complex non - gravitational processes , such as injection of energy feedback from active galactic nuclei , radiative cooling , supernovae , and star formation , influencing the thermal structure of icm , particularly in low mass ( temperature ) clusters @xcite .
the first direct evidence for non - gravitational entropy in galaxy clusters and galaxy groups was given by @xcite using rosat position sensitive proportional counter ( pspc ) observations .
the observations showed that there was excess entropy compared to that expected from gas collapsing in a gravitational potential .
motivated by these findings , several groups have drawn similar conclusions using both numerical and semi analytical models with an entropy floor of the order of @xmath21 kev @xmath6 @xcite .
although , sne feedback are essential to explain the enrichment of the icm to the observed metallicity level and heavy - element abundances , they provide insufficient amount of energy per particle as compared to recent observations .
moreover , they are also inefficient to quench cooling in massive galaxies @xcite .
there is a growing evidence that agn feedback mechanism provides a major source of heating for the icm gas , thereby reducing number of cooling flow clusters @xcite .
the agn - jet simulations show that such mechanisms can overcome the cooling flow over a cosmological timescale and produce results similar to observations @xcite . [
cols="<,^,^,^,^,^,^,^ " , ] columns ( 1 ) , ( 2 ) , ( 3 ) & ( 4 ) shows cluster names , redshift , state , and @xmath4 values from @xcite respectively .
columns ( 5 ) & ( 6 ) shows values @xmath4 and @xmath3 for @xmath22 case . columns ( 7 ) & ( 8) shows values of @xmath4 , @xmath3 for @xmath23 case.the numbers in brackets indicate increase in @xmath4 and @xmath3 if clumping is taken into account .
moreover , it has been found with suzaku observations that the entropy profile flattens out at large radii @xcite .
this entropy decrement can be related to the gas clumping @xcite , presence of non - thermal pressure @xcite , accretion / merger shocks in outskirts of clusters @xcite , loss of kinetic energy of gas due to the cosmic ray acceleration @xcite or due to the rapid temperature decrease in the outskirts of clusters as a result of the non - gravitational processes @xcite .
it is convenient to define an entropy profile constant .
] of gas as , @xmath24 , where @xmath25 is the boltzmann constant , and the exponent @xmath26 refers to the adiabatic index . with this definition , @xmath27 remains unchanged for all adiabatic processes and can therefore probe the thermal history of gas .
purely gravitational models predict entropy profiles in clusters of the form @xmath28 @xcite .
however , as pointed , several recent observations found deviations from this expected entropy profile , especially at inner and outer radii @xcite as a result of non - gravitational feedback . to allow a meaningful interpretation and to estimate degree of feedback
, one needs to compare recent observations with theoretically expected profiles with no feedback .
previously , @xcite estimated the non - gravitational energy deposition profile up to @xmath4 by comparing the observed entropy profiles with a benchmark entropy profile without feedback @xcite for the rexcess sample of 31 clusters @xcite observed with xmm - newton .
they found an excess mean energy per particle of @xmath29 kev and @xmath30 kev using benchmark entropy from adaptive mesh refinement ( amr ) and smoothed particle hydrodynamics ( sph ) simulations respectively , along with a strong correlation for agn feedback .
our study extends their work by going beyond @xmath4 . here , we consider the joint data set of planck sz pressure profile and rosat gas density profiles of 17 clusters @xcite to estimate the excess entropy and feedback energy profiles up to @xmath3 .
recent simulations show significant level of non - thermal pressure from bulk motion @xcite and gas clumping ( which by definition is measured by @xmath31=<@xmath32>/<@xmath33>@xmath34 ) @xcite in the outer regions of icm .
we therefore , incorporate both these factors in our calculations .
this paper is a continuation of our recent work @xcite wherein we showed that pre - heating scenarios are ruled out at more than 3@xmath17 statistical level . in this work ,
we present a detailed study of the excess entropy profiles along with feedback energy profiles and discuss the effect of non - thermal pressure and clumping in our estimates .
we also look at sample selection , cool - core vs non cool - core clusters , effects of boundary conditions , choice of benchmark theoretical entropy profiles , choice of x - ray methodology , etc .
the paper is organized as follows : in section 2 , we describe the cluster sample used in this work . in section 3
, we describe the self - similar non - radiative model for galaxy clusters .
section 4 is dedicated to the determination of excess entropy and energy profiles , and the effects of the non - thermal pressure and gas clumping on their estimates . in section 5
, we check the robustness of our results . in section 6
, we compare the feedback profiles for amr and sph benchmark entropy profiles .
section 7 gives the comparison of our results with the previous estimates .
finally , the conclusions of our work is given in last section . throughout this paper ,
we assume cosmology where ( @xmath35 , @xmath36 , @xmath37 ) = ( 0.3 , 0.7 , 70 ) .
in this work , we have studied a sample of 17 clusters in the redshift range ( @xmath38 ) that are common in @xcite and @xcite . based on their central entropy @xcite ,
six of these clusters are classified as cool - core ( @xmath39 kev @xmath6 ) , while the remaining 11 are non - cool core .
this sample was earlier used by @xcitedeckert / newsite / dominique_eckerts_hom + epage.html . ] where they have shown that the thermodynamic state of the icm can be accurately recovered by using the planck sz pressure profile @xcite and rosat position sensitive proportional counter ( pspc ) gas density profile @xcite .
since the sz signal is proportional to the integrated pressure ( gas density ) , unlike the x - ray signal , which is proportional to the square of density , it decreases more gently at large radius and therefore can provide more accurate pressure profiles in the cluster outskirts .
this allows us to accurately recover the temperature profile beyond @xmath4 .
we use parametric profiles obtained by them in this work ( see @xcite for more details ) which were obtained by fitting a functional form to the projected emission - measure data @xcite and planck sz pressure data @xcite .
the errors were obtained through monte carlo markov chain , with direct sampling of the posterior temperature , entropy , and gas fraction distributions .
@xcite also obtained deprojected profiles by estimating density profile through `` onion peeling '' technique @xcite and interpolating the sz pressure profile .
correction for edge effects were also applied along with median smoothing regularization for minimizing the roller - coaster effect @xcite .
the error bars were recovered by perturbing the original profile using a monte carlo and recomputing the deprojected profiles each time .
since the parametric profiles are forced to be regular , this reduces the cluster to cluster scatter and the errors . at smaller radii
the angular resolution of both planck and rosat is insufficient to obtain reliable constraints .
therefore , the parametric fitting was only performed on the data beyond @xmath40 and were found to be consistent with the deprojected profiles .
we use the hydrostatic equation to obtain the total mass profile @xmath41 in the galaxy clusters , @xmath42 where @xmath33 and @xmath43 are density and thermal pressure of the icm respectively .
@xmath44 is calculated by using the best fit generalized nfw ( gnfw ) pressure profile @xcite .
the quantities @xmath4 and @xmath3 where obtained by first interpolating the @xmath41 profile and then iteratively solving is defined such that @xmath45 is the radius out to which mean matter density is @xmath46 , where @xmath47 being critical density of the universe at redshift @xmath48 . ] , @xmath49 the virial radius , @xmath50 , is calculated with the help of the spherical collapse model @xcite .
if required , the mass profile is obtained by linear extrapolation in logarithmic space .
recent simulations suggest that a significant non - thermal pressure contributes to the total energy of the icm gas , mainly due to bulk gas motions and turbulence in the icm gas @xcite .
while non - thermal pressure is small in the inner region , its relative importance steadily increases with radius , becoming a significant fraction of the total pressure in the outer region @xcite .
it has been found from both observations and simulations that @xmath51 is underestimated by about @xmath52 , if one uses the hydrostatic equation without non - thermal pressure @xcite . from the recent numerical simulations , we model the non - thermal pressure fraction in the power law form similar to that given in @xcite , @xmath53 where @xmath54 is the total gas pressure , @xmath55 , @xmath56 for low redshift clusters ( @xmath57 ) with @xmath58 , @xmath59 and @xmath60 @xcite . in upper panel of fig .
[ pressureanddensity ] , we have plotted @xmath61 as a function @xmath62 .
it can be seen that @xmath63 becomes @xmath64 of the thermal gas pressure @xmath65 in the cluster outskirts .
since non - thermal pressure is not negligible beyond @xmath4 , one should take it into account in order to properly study the cluster physics in the outer regions .
similarly , it has been seen that gas clumping results in an overestimation of the observed gas density ( @xmath66 ) and hence underestimation of the entropy and total mass profiles .
it has been found from various observations and hydrodynamical simulations that the clumping factor is negligible in the innermost cluster regions but radially increases with @xmath67 around @xmath3 @xcite .
however , few works have also reported either smaller or higher values of clumping factor @xcite .
@xcite found that the azimuthal median is a good tracer of the true 3d density ( @xmath68 ) and showed from both hydrodynamical simulations that their method recovered the @xmath68 profiles with deviations less than 10% at all radii .
they recovered the average @xmath69 at @xmath3 which is consistent with the recent results .
lower panel of fig .
[ pressureanddensity ] shows the consistency of @xcite clumping profile with that of @xcite .
we have also calculated the total mass profile @xmath70 by including non - thermal pressure @xmath63 in eq .
[ eqn1:eq1 ] and correcting the density profile using @xcite clumping profile . in tab .
[ fit ] , we give estimates of @xmath4 and @xmath3 obtained by using parametric density profiles along with _ planck _ best fit gnfw pressure profile @xcite . for comparison ,
we have also given planck @xmath4 values from @xcite which are consistent with our estimates within @xmath71 for most of the clusters .
moreover , we find that the average scaling @xmath72 @xcite is in excellent agreement with our results with small scatter . as a function @xmath73 for all the clusters ( considering clumping and @xmath74 ) .
solid lines represent ncc clusters and dashed lines represent cc clusters.,width=321 ] standard models of large scale structure show that matter is shock heated as it falls into clusters under the influence of gravity and predict that the entropy of gas ( @xmath75 ) should behave as a power - law , with entropy profiles flattening near cluster cores .
@xcite performed several non - radiative sph and amr simulations in order to study the main features of the entropy profiles .
they found differences in entropy profiles in the inner cores but the differences were small for @xmath76 in sph and amr simulations .
they found that the simulated non - radiative scaled entropy profile can be described by a simple power law form in the range @xmath77 with a slightly higher normalization for amr case .
for the inner radii , @xcite found a large discrepancy in the scaled entropy profiles between the sph and amr simulations .
a flat entropy core has been observed in the center of non - radiative galaxy clusters in eulerian grid codes ( amr ) which is absent in lagrangian approaches ( sph ) . however , after accounting for certain hydrodynamical processes ( i.e shocks and mixing motions ) the results of sph simulations match with that of amr case @xcite .
we use amr and sph median entropy profiles obtained by @xcite for our baseline model described by , @xmath78 in the range @xmath77 plus by a flatter core below @xmath79 which is much more pronounced in case of amr simulations .
@xmath80 is equal to 1.32 and 1.41 for sph and amr respectively and @xmath81 is given by , @xmath82 where @xmath83 is the universal baryonic fraction . in order to calculate the initial ( without feedback ) density ( or gas mass ) and temperature profiles ,
one solves the the hydrostatic equation with appropriate boundary condition @xcite .
considering non - thermal pressure component , we rewrite the hydrostatic equation as , @xmath84 where @xmath85 is the initial ( theoretical ) thermal pressure of the icm and @xmath70 is the sum of two terms , @xmath86 . since energy injection only effects the gas mass profile , one can assume the dark matter profile and hence total mass profile to remain constant during the feedback processes .
for the boundary condition we assume the gas fraction ( @xmath87 ) to be @xmath88 at virial radius @xcite . on the addition of non - thermal pressure
, the initial entropy profile is increased due the overall increase in the normalization and therefore , the deviation from the observed entropy decreases .
it is important to note that initial entropy profile also depends on the baryonic fraction through @xmath81 .
most of the previous estimates of the entropy floor were based on the wmap7 estimates of @xmath89 and since the planck predicts relatively lower value of @xmath90 @xcite , this will further increase the initial entropy profile thereby decreasing the estimates of excess entropy . in this section
, we estimate the total mechanical feedback energy .
it is important to note that for a meaningful interpretation , one should compare the theoretical and observed entropy profiles at the same gas mass ( @xmath91 ) instead of same radii in order to provide an allowance for redistribution of gas on account of feedback processes @xcite . considering a transformation from the baseline configuration to new configuration
i.e @xmath92 , the additional thermal energy per particle in icm corresponding to the transformation is given by , @xmath93 where @xmath94 .
for a value of @xmath95 , the ratio between @xmath96 in two cases is @xmath97 .
this implies that if the observed temperature @xmath98 deviates from the theoretically calculated value @xmath99 by a factor @xmath100 , then the two above mentioned estimates of energy input per unit mass differ by only a factor of @xmath101 .
[ tobs_by_tth ] shows the ratio @xmath94 for all the clusters and is mostly in the range @xmath102 .
we choose the expression for the isobaric process in our estimates .
moreover , we find using isochoric expression instead does not make any notable difference .
the total excess energy per particle in icm can be found by including the change in potential energy term in last equation , @xmath103 where @xmath104 and @xmath105 are theoretical and observed radii respectively enclosing the same gas mass .
the total feedback energy per particle in icm can be found after adding the energy lost due to cooling i.e , @xmath106 we approximate the energy lost in icm in a given mass shell as , @xmath107 where @xmath108 is the bolometric luminosity emitted by the icm in a given shell which is obtained by considering cooling function @xmath109 given in @xcite and @xmath110 is the age of the cluster which we have fixed at @xmath111 gyr @xcite .
@xmath109 is calculated using theoretical ( initial ) temperature and density profiles .
we found using observed profiles instead of theoretical profiles does not make any notable difference in our estimates .
the total amount of energy deposited , for the whole cluster is , @xmath112 where @xmath113 is the mean molecular weight of gas and @xmath114 is mass of proton . dividing the total energy in the icm by the total number of particles in the icm , we estimate the average energy per particle ( @xmath115 ) .
in this section , we study the entropy and energy deposition profiles ( in terms of @xmath116 and @xmath117 profiles ) in the galaxy clusters up to @xmath3 ( @xmath118 ) using the methodology discussed in the previous section .
we also investigate the impact of the non - thermal pressure , gas clumping and baryonic fraction on our estimates .
all the figures that follow assume amr entropy profiles and planck estimates of the universal baryonic fraction @xmath90 unless stated otherwise .
we shall refer to the case where we assume @xmath90 and consider both non - thermal pressure and clumping as a fiducial case .
the figures obtained using sph baseline entropy profiles are shown in appendix . in fig .
[ fig : amrdeltak ] , we show @xmath116 profiles of all the individual clusters along with weighted average of full sample as a function of @xmath73 and fig .
[ fig : amrdeltae ] shows the corresponding @xmath117 profiles .
in general , we find that for both @xmath116 and @xmath117 , * the average profiles for full sample are positive in the inner regions , but it becomes negative in the outer regions . * the inclusion of clumping factor , increases overall profiles , due to the increase of observed entropy profiles .
* the inclusion of non - thermal pressure decreases overall profile up to @xmath4 and in the outer radii the profiles actually unexpectedly increase . * the average profiles of cc and ncc clusters differ significantly .
cc clusters have much higher values compared to the average .
profile for the entire sample with and without adding energy lost due to cooling.,width=321 ] given the negative values of @xmath116 at the outer radii , the corresponding profiles of @xmath119 per particle would also be negative . on the face of it
, the result would be physically meaningless .
however , one should note that icm gas loses energy due to radiation and the amount of energy lost due to radiation can be added to offset offset the negative values . solid red line and dashed red line in fig . [ cooling ] shows the average profiles with and without taking into account energy lost due to cooling respectively .
the difference in these two curves is small beyond @xmath4 because of the fact that gas density is small in those regions and therefore radiative cooling can not explain the profiles going below zero in the outer region
. tabs .
[ amrplanck ] & [ amrwmap ] give the estimates of average feedback energy per particle @xmath115 in the ranges @xmath120 , @xmath121 and @xmath122 using planck and wmap estimates of the baryonic fraction .
it can be seen that the inclusion of non - thermal pressure affects the estimates of @xmath115 both in the inner and outer regions of the cluster and that clumping has a substantial effect only in the cluster outer regions . in the next two subsections , we show as to how the proper incorporation of both clumping and non - thermal pressure can lead to meaningful estimates of the feedback profiles . in the outer regions
the level of clumping in gas profile can be significant which can also lead to biased estimates of density and hence in entropy measurements .
[ fig : average ] shows the comparison of average @xmath116 and @xmath117 profiles for various cases .
one can clearly see the implications of correcting the entropy by using the clumping profile from @xcite on our estimates .
the addition of clumping factor raises @xmath116 and hence @xmath117 profiles as expected . here
, the increase is not negligible and makes the average profile become more or less consistent with zero in the outer regions for @xmath23 case .
however , for the pure thermal case , the profiles are still negative in the outer regions to a significant level which we ascribe to the neglect of non - thermal pressure in the next subsection . comparing the average @xmath116 feedback profiles with and without clumping
, one can see that ignoring the clumping correction leads to a decrease of @xmath5 kev @xmath6 at @xmath4 and @xmath7 kev @xmath6 at @xmath3 .
the energy feedback profiles on other hand are under - estimated by @xmath123 kev at @xmath4 and @xmath124 kev at @xmath3 .
similarly , from tabs . [ amrplanck ] & [ amrwmap ] , it is also evident that the the average feedback energy per particle , @xmath115 , is under - estimated by @xmath125 kev in the region @xmath120 and @xmath101 kev in the region @xmath122 if clumping correction is neglected .
although , the inclusion of the non - thermal pressure decreases the feedback profiles up to @xmath4 due to the overall increase in the normalization ( @xmath81 ) in the benchmark entropy profiles , however , as seen from fig .
[ fig : average ] , it unexpectedly increases beyond that .
the cross - over occurs around @xmath126 .
this can be understood as follows : due to the neglect of the non - pressure , the @xmath70 profile is underestimated which in turn results in the under - estimation of theoretical gas mass as @xmath127 is fixed at the virial radius .
this implies for the given observed gas mass shell the corresponding theoretical gas mass for @xmath22 case will occur at larger radius as compared to @xmath23 case and will , therefore , have higher theoretical entropy leading to decrease in feedback profiles . below @xmath4 ,
the increase in @xmath81 term dominates ( since non - thermal pressure is small ) and therefore , there is an overall decrease in the feedback profiles for @xmath23 case . from fig .
[ fig : average ] , it is evident that ignoring the non - thermal pressure leads to an over - estimation of @xmath8 kev @xmath6 at @xmath4 and under - estimation of @xmath9 kev @xmath6 at @xmath3 .
this in turn leads to an over - estimation of @xmath14 kev at @xmath4 and under - estimation of @xmath15 kev at @xmath3 .
similarly , one can see from tabs .
[ amrplanck ] & [ amrwmap ] , if non - thermal pressure is ignored then the @xmath115 is over - estimated by @xmath128 kev in the region @xmath120 while it has a negligible affect in the region @xmath122 . in fig .
[ nonthermalnor ] , we show the effect on the feedback profile by changing the normalization @xmath80 in the non - thermal pressure .
we find that changing the normalization from 0.18 to ( 0.10 , 0.26 ) gives around ( 10% , 30% ) mass difference at @xmath4 .
we find that a change in the normalization has a small effect on the profiles and our results are still consistent with zero line given the error bars .
we also observe that other parameterization of non - thermal pressure such as that of @xcite lie within the normalization @xmath129 band of our non - thermal pressure model .
therefore , our results are independent of the non - thermal parameterization .
fig [ fig : ccandncc ] shows the @xmath116 and @xmath117 profiles for cc and ncc populations .
the higher value of @xmath116 and @xmath117 profiles for cc clusters can be interpreted in terms of the gas mass fraction .
@xcite found that the observed gas mass fraction profile in the cc clusters was systematically lower than ncc clusters . since
, the theoretical value @xmath87 is fixed at @xmath88 at virial radius and the observed value of @xmath130 for cc clusters is relatively smaller than ncc clusters , this means that for a given observed gas mass @xmath131 , the corresponding theoretical gas @xmath132 for cc clusters will also occur at a relatively smaller radius with a smaller value of theoretical entropy .
this will thus result in a relatively higher degree of feedback and up to a much larger radii in cc clusters compared to ncc clusters .
there are four clusters ( i.e clusters 1 , 6 , 16 and 17 ) which have relatively large value of the @xmath116 profile in the outer regions ( particularly for the clumping case ) which correspondingly give a large thermal energy profile @xmath96 .
however , we see that after taking into account a potential energy term the @xmath117 profiles for such clusters become close to zero ( or even negative ) . in order to see the effect of such clusters , we have plotted in fig .
[ subsample ] average @xmath116 and @xmath117 profiles for the full sample along with the sub sample which do not include theses clusters .
we find that the average feedback entropy become consistent with @xmath133 line at 1@xmath17 beyond @xmath4 for the sub - sample .
moreover , average @xmath116 and @xmath117 profiles for the sub - sample and full sample are always consistent with one another . in fig .
[ fig : deprojected ] , we show the comparison of the entropy feedback profiles for two different boundary conditions , i.e universal baryonic fraction @xmath83 at virial radius to be @xmath134 ( from planck ) and @xmath135 ( from wmap ) . the larger value of the @xmath83 from wmap , would result in an overall increase of feedback profiles .
this is because higher value of @xmath83 at the virial radius will increase the total theoretical gas mass profile ( as total mass remains constant ) .
therefore , a given theoretical gas mass shell would occur at a smaller radius having a lower value of entropy leading to an increase in feedback profiles .
it is clear from fig .
[ fig : deprojected ] , that the entropy feedback profiles for the wmap boundary conditions is significantly higher at the outer regions ( @xmath136 kev @xmath6 ) .
@xcite found that the parametric and deprojected density profiles are similar and the difference is less than 10% .
we show the @xmath116 profiles for all the clusters using deprojected data in fig .
[ fig : withfg ] .
we find parametric and deprojected profiles have similar values of entropy difference from the base theoretical entropy profiles except at the cluster outskirts .
feedback profiles as a function of @xmath73 for two boundary conditions of gas fraction at virial radius ( i.e @xmath137 from planck and @xmath138 from wmap).,width=321 ] feedback profiles as a function of @xmath73 for all the clusters considering the deprojected case.,width=321 ]
given the current status , it is difficult to judge whether sph or amr is more accurate since both these methods are known to have some demerits . for example
sph suffers from a relatively poor shock resolution and noise on the scale of the smoothing kernel .
amr simulations may suffer from over - mixing due to advection errors in the presence of bulk flows .
apart from the slightly smaller normalization of the entropy profile @xmath80 , the amr simulations predict a much higher flatter entropy core than the corresponding sph simulations @xcite .
however , it has been recently pointed out by many authors that , after resolving certain hydrodynamic processes , the results of sph simulations exactly match with amr simulations @xcite .
we nevertheless use sph simulated entropy in our estimates to compare it with that of amr case . in fig .
[ amrsphcomp ] , we show the average excess entropy and energy profiles for both amr and sph cases .
as can be seen , sph case predict much higher values of @xmath119 in the inner regions of the cluster because of the absence of flatter entropy profile as is present in amr case .
the @xmath116 and @xmath117 feedback profiles at the outer radii contain information of past events in the cluster and hence are ideal to probe for any signature of pre - heating that may have taken place at high redshifts much before the cluster formation .
this is because of the fact that feedback processes from agn or supernovae are unlikely to affect the gas properties there .
however , we should also point that our estimated @xmath117 profiles corresponds to the change from the initial theoretical model to the observed configuration in the collapsed systems . pre - heating ( if any ) at hight redshift when the density of gas was small would actually require much smaller energy input to bring it to final observed state @xcite .
therefore , @xmath117 would represent an upper limit on pre - heating energy .
it has been found that pre - heating scenarios ( at @xmath139 ) typically require feedback energy of @xmath140 kev per particle or an entropy floor of @xmath141 kev @xmath6 to explain break in the self - similarity scaling relations @xcite .
our results show that given the uncertainties , the values of @xmath119 at the outer radii are comparable to zero for both amr and sph cases ( see fig . [ amrsphcomp ] ) . for our fiducial case
, we see that in the range @xmath122 , the average energy per particle @xmath142 for the amr case and @xmath143 for sph case .
this implies that pre - heating scenarios which predict 1 kev energy per particle are ruled out with more than 3@xmath17 for amr case and at around 2@xmath17 for sph case .
considering @xmath116 profiles , we find for most of the cluster region an entropy floor @xmath141 kev @xmath6 is ruled out at @xmath11 for the amr case .
however , no such strong constrains are possible for the sph case and that @xmath136 kev @xmath6 is consistent with @xmath144 as seen in the left side of fig .
[ amrsphcomp ] .
@lccccccc & & + & & & + @xmath31 & @xmath63 & @xmath145 & @xmath146 & @xmath122 & @xmath145&@xmath146&@xmath122 + + 0 & 0 & @xmath147 & @xmath148 & @xmath149 & @xmath150 & @xmath151 & @xmath152 + 0 & nonzero & @xmath153 & @xmath154 & @xmath155 & @xmath156 & @xmath157 & @xmath158 + nonzero & 0 & @xmath159 & @xmath160 & @xmath161 & @xmath162 & @xmath163 & @xmath164 + nonzero & nonzero & @xmath165 & @xmath166 & @xmath167 & @xmath168 & @xmath169 & @xmath170 + + 0 & 0 & @xmath171 & @xmath172 & @xmath173 & @xmath174 & @xmath175 & @xmath176 + 0 & nonzero & @xmath177 & @xmath178 & @xmath179 & @xmath180 & @xmath181 & @xmath182 + nonzero & 0 & @xmath183 & @xmath184 & @xmath185 & @xmath186 & @xmath187 & @xmath188 + nonzero & nonzero & @xmath189 & @xmath190 & @xmath191 & @xmath192 & @xmath193 & @xmath194 + + 0 & 0 & @xmath195 & @xmath196 & @xmath197 & @xmath198 & @xmath199 & @xmath200 + 0 & nonzero & @xmath201 & @xmath202 & @xmath203 & @xmath204 & @xmath205 & @xmath206 + nonzero & 0 & @xmath207 & @xmath208 & @xmath209 & @xmath210 & @xmath211 & @xmath212 + nonzero & nonzero & @xmath213 & @xmath214 & @xmath215 & @xmath211 & @xmath216 & @xmath217 + columns ( 3 ) , ( 4 ) & ( 5 ) : @xmath115 in the ranges @xmath218 , @xmath77 and @xmath122 respectively without taking into account energy lost due to cooling . columns ( 6 ) , ( 7 ) & ( 8) : @xmath115 in the ranges @xmath218 , @xmath77 and @xmath122 respectively after taking into account energy lost due to cooling .
the errors are given at 1@xmath17 level .
for meaningful comparison , @xmath115 for @xmath22 case are also calculated up to same radii as that of non - thermal case ( i.e @xmath4 and @xmath3 of @xmath23 ) .
@lccccccc & & + & & & + @xmath31 & @xmath63 & @xmath145 & @xmath146 & @xmath122 & @xmath145&@xmath146&@xmath122 + + 0 & 0 & @xmath219 & @xmath220 & @xmath221 & @xmath222 & @xmath223 & @xmath224 + 0 & nonzero & @xmath225 & @xmath226 & @xmath227 & @xmath228 & @xmath229 & @xmath230 + nonzero & 0 & @xmath231 & @xmath232 & @xmath233 & @xmath234 & @xmath235 & @xmath236 + nonzero & nonzero & @xmath237 & @xmath238 & @xmath239 & @xmath240 & @xmath241 & @xmath242 + + 0 & 0 & @xmath243 & @xmath244 & @xmath245 & @xmath246 & @xmath247 & @xmath248 + 0 & nonzero & @xmath249 & @xmath250 & @xmath251 & @xmath252 & @xmath253 & @xmath254 + nonzero & 0 & @xmath255 & @xmath256 & @xmath257 & @xmath258 & @xmath259 & @xmath260 + nonzero & nonzero & @xmath174 & @xmath261 & @xmath262 & @xmath263 & @xmath264 & @xmath265 + + 0 & 0 & @xmath266 & @xmath267 & @xmath268 & @xmath269 & @xmath270 & @xmath271 + 0 & nonzero & @xmath272 & @xmath273 & @xmath274 & @xmath275 & @xmath276 & @xmath277 + nonzero & 0 & @xmath278 & @xmath279 & @xmath280 & @xmath281 & @xmath282&@xmath283 + nonzero & nonzero & @xmath284 & @xmath285 & @xmath286&@xmath287 & @xmath288 & @xmath289 + columns ( 3 ) , ( 4 ) & ( 5 ) : @xmath115 in the ranges @xmath218 , @xmath77 and @xmath122 respectively without taking into account energy lost due to cooling . columns ( 6 ) , ( 7 ) & ( 8) : @xmath115 in the ranges @xmath218 , @xmath77 and @xmath122 respectively after taking into account energy lost due to cooling .
the errors are given at 1@xmath17 level .
for meaningful comparison , @xmath115 for @xmath22 case are also calculated up to same radii as that of non - thermal case ( i.e @xmath4 and @xmath3 of @xmath23 ) .
it is important to distinguish between the profiles of @xmath290 with respect to shells at fixed position @xcite , and with respect to shells with a given gas mass interior to it @xcite .
this aspect is demonstrated in fig .
[ fig : comp ] .
the right panel shows the entropy profiles with respect to fixed radii , and the left panel shows the entropy with respect to shells with a given gas mass interior to it .
the differences between the profiles ( w.r.t . @xmath62 and @xmath73 ) are striking and noteworthy . the observed entropy profiles ( solid brown line )
shows an enhancement of entropy in the inner region , but drops below the theoretical profile in the outer region . in this case
, the inclusion of non - thermal pressure would seem to exacerbate the situation and the deviation of observed profile becomes acute .
however , one see that after accounting for the clumping correction , the recovered entropy profiles ( solid black line ) shows an excess compared to theoretical entropy profile with respect to gas mass for most of the cluster region .
comparing entropy profiles at the same radii , we find the deviations between theoretical and clumping corrected entropy profiles become negligible beyond @xmath291 .
it is also worth mentioning here that for theoretical entropy profiles the cross - over between @xmath22 and @xmath23 cases is around @xmath126 as also seen in average @xmath116 and @xmath116 profiles ( see fig .
[ fig : average ] ) .
earlier , @xcite determined the feedback profile up to @xmath4 .
they found total feedback energy @xmath292 scales with the mean spectroscopic temperature as @xmath293 and @xmath294 for the sph and amr baseline profiles respectively .
they showed that @xmath292 correlates strongly with the radio luminosity @xmath295 of the central radio sources and estimated energy per particle to be @xmath296 kev for the sph simulations and @xmath297 kev for the amr simulations which is much greater than our estimate . notice that they did not consider non - thermal pressure and clumping and their calculations included cluster cores which results in the higher estimates of energy per particle .
recent studies have revealed that non - gravitational processes play an important role in modifying the thermodynamic properties of the icm .
it has also been observed that there is an entropy enhancement in galaxy clusters which is believed to be a result of the non - gravitational feedback from active galactic nuclei , radiative cooling , supernovae etc .
we have studied the fractional entropy enhancement and the corresponding feedback energy in the icm for a sample of 17 galaxy clusters by comparing the observed entropy profiles with that of amr and sph non - radiative simulations .
unlike , previous work by @xcite where they estimated the feedback energetics up to radius @xmath298 , our study goes up to @xmath3 . since around 80% of the cluster volume
is beyond @xmath4 this study has important implication on the icm thermodynamics and feedback processes .
the cluster outskirts have many features which are not significant into cluster cores .
these include , deviation from the hydrostatic equilibrium and gas clumping .
therefore , our analysis takes both non - thermal pressure and clumping into account which are important to study the energetics of icm in the outer regions .
we show that neglect of clumping and non - thermal pressure can lead to an under - estimation of @xmath4 and @xmath3 by @xmath299 .
similarly , we find an under / over - estimation of feedback profiles .
we find that the effect of clumping is much more pronounced than the non - thermal pressure and neglecting it always leads to an under - estimation of feedback profiles .
the neglect of clumping leads to an under - estimation of entropy @xmath7 kev @xmath6 and feedback energy per particle @xmath123 kev at @xmath3 .
the neglect of non - thermal pressure on the other hand lead to an over - estimation in the inner regions and under - estimation in the outer regions .
the omission of the non - thermal pressure results in an under - estimation of entropy @xmath9 kev @xmath6 and feedback energy per particle @xmath300 kev at @xmath301 further , we find that the feedback energy profiles are centrally peaked which can be related with agn feedback models and more or less flatten out in the outer regions becoming consistent with zero . finally , to check the robustness of our results , we compared the feedback profiles for different cases : cc and ncc clusters , the full sample and a sub - sample , parametric and deprojected cases .
we find the much higher value of feedback profiles for the cc clusters compared to ncc clusters .
however , since cc clusters comprise a much smaller sample one , therefore , needs to improve / verify the estimates by considering a larger sample .
we also find that the choice of the universal baryonic fraction from wmap and planck can have implications on the estimates of the feedback profiles . in order to obtain any robust estimates of cosmological parameters from large sz surveys and sz power spectrum
, one needs to properly incorporate the non - gravitational feedback .
therefore , it is utmost important to understand the nature and extent of the non - gravitational feedback in galaxy clusters , out to the virial radius , so as to properly calibrate the scaling relations and theoretical cluster models . in principle , one can consider different parameterizations of excess entropy and feedback energy to see the affects of the non - gravitational feedback on icm thermodynamics .
this work was supported by serb ( dst ) project grant no .
sr / s2/hep-29/2012 .
ai would like to thank tata institute of fundamental research ( tifr ) , mumbai and raman research institute ( rri ) , banglore for hospitality .
[ fig : sphdeltak ] @lccccccc & & + & & & + @xmath31 & @xmath63 & @xmath145 & @xmath146 & @xmath122 & @xmath145&@xmath146&@xmath122 + + 0 & 0 & @xmath302 & @xmath303 & @xmath304 & @xmath305 & @xmath306 & @xmath307 + 0 & nonzero & @xmath308 & @xmath309 & @xmath221 & @xmath310 & @xmath311 & @xmath312 + nonzero & 0 & @xmath313 & @xmath314 & @xmath315 & @xmath316 & @xmath317 & @xmath318 + nonzero & nonzero & @xmath319 & @xmath320 & @xmath321 & @xmath322 & @xmath323 & @xmath324 + + + 0 & 0 & @xmath325 & @xmath326 & @xmath327 & @xmath328 & @xmath329 & @xmath330 + 0 & nonzero & @xmath331 & @xmath332 & @xmath333 & @xmath334 & @xmath335 & @xmath336 + nonzero & 0 & @xmath337 & @xmath338 & @xmath339 & @xmath340 & @xmath341 & @xmath342 + nonzero & nonzero & @xmath343 & @xmath344 & @xmath345 & @xmath346 & @xmath347 & @xmath348 + + 0 & 0 & @xmath349 & @xmath350 & @xmath351 & @xmath352 & @xmath353 & @xmath354 + 0 & nonzero & @xmath355 & @xmath356 & @xmath357 & @xmath358 & @xmath359 & @xmath360 + nonzero & 0 & @xmath361 & @xmath362 & @xmath363 & @xmath364 & @xmath365 & @xmath366 + nonzero & nonzero & @xmath367 & @xmath368 & @xmath369 & @xmath370 & @xmath371 & @xmath372 + @lccccccc & & + & & & + @xmath31 & @xmath63 & @xmath145 & @xmath146 & @xmath122 & @xmath145&@xmath146&@xmath122 + + 0 & 0 & @xmath373 & @xmath374 & @xmath375 & @xmath376 & @xmath377 & @xmath378 + 0 & nonzero & @xmath310 & @xmath379 & @xmath380 & @xmath381 & @xmath382 & @xmath383 + nonzero & 0 & @xmath384 & @xmath385 & @xmath386&@xmath387 & @xmath388 & @xmath389 + nonzero & nonzero & @xmath390 & @xmath391 & @xmath392 & @xmath393 & @xmath234 & @xmath394 + + 0 & 0 & @xmath395 & @xmath396 & @xmath397 & @xmath398&@xmath399 & @xmath400 + 0 & nonzero & @xmath401 & @xmath402 & @xmath403 & @xmath404 & @xmath405 & @xmath406 + nonzero & 0 & @xmath407 & @xmath408 & @xmath409&@xmath410 & @xmath411 & @xmath412 + nonzero & nonzero & @xmath413 & @xmath414 & @xmath415&@xmath416 & @xmath417 & @xmath418 + + 0 & 0 & @xmath419 & @xmath420 & @xmath421 & @xmath422 & @xmath423 & @xmath424 + 0 & nonzero & @xmath425 & @xmath426 & @xmath427 & @xmath428 & @xmath429 & @xmath215 + nonzero & 0 & @xmath430 & @xmath431 & @xmath432 & @xmath433 & @xmath434 & @xmath435 + nonzero & nonzero & @xmath436 & @xmath437 & @xmath438 & @xmath439 & @xmath440 & @xmath441 + | we estimate the non - gravitational entropy injection profiles , @xmath0 , and resulting non - gravitational energy feedback profiles , @xmath1 , of the intracluster medium ( icm ) for a sample of 17 clusters using the joint data sets of planck sz pressure profiles and rosat x - ray observations .
the clusters are chosen such that all observed profiles span a large radial range from @xmath2 ( i.e. , from inside the cluster cores ) up to @xmath3 .
the feedback profiles are estimated by comparing the observed entropy , at fixed gas mass shells , with theoretical entropy profiles predicted from non - radiative ( i.e , without feedback ) hydrodynamic simulations .
we include non - thermal pressure and gas clumping in our estimates of the feedback profiles since they become important at larger radii , typically beyond @xmath4 . the inclusion of non - thermal pressure and clumping results in changing the estimates for @xmath4 and @xmath3 by 10%-20% .
we show that neglect of clumping leads to an under - estimation of @xmath5 kev @xmath6 at @xmath4 and @xmath7 kev @xmath6 at @xmath3 . on the other hand , neglecting non - thermal pressure results in the over - estimation of @xmath8 kev @xmath6 at @xmath4 and under - estimation of @xmath9 kev @xmath6 at @xmath3 . combining both in our analysis
, we conclusively show that for the sample as a whole , an entropy floor of @xmath10 is ruled out at @xmath11 throughout the entire radial range and hence strongly constraining all icm pre - heating scenarios . for the estimated feedback energy profiles , we find that the neglect of clumping leads to an under - estimation of energy per particle @xmath12 kev at @xmath4 and @xmath13 kev at @xmath3 . similarly , neglect of the non - thermal pressure results in an over - estimation of @xmath14 kev at @xmath4 and under - estimation of @xmath15 kev at @xmath3 .
we find that the average feedback energy per particle of @xmath16 kev , reminiscent of pre - heating , is also ruled out at more than 3@xmath17 beyond @xmath4 .
subdividing the sample into cool - core ( cc ) and non cool - core ( ncc ) fractions , we find that cc clusters feedback profiles have significantly higher profiles compared to ncc clusters ; this can be attributed to the lower values of gas mass fraction of cc compared to ncc clusters .
we explore in details the effect of changes in non - thermal pressure modeling as well as the choice of choosing either amr vs sph simulations to get the pre - feedback profiles .
we also demonstrate that our non - gravitational feedback constraints are robust w.r.t cluster sample selection , x - ray analysis procedures and entropy modeling etc .
this study has important implications for the understanding of the icm physics at cluster outskirts as well as for any plausible energy feedback mechanism over a large cluster radial range .
galaxies : clusters : intracluster medium - cosmological parameters . |
finding the earliest galaxies in the universe and characterizing their properties and contribution to the reionization of the universe are some of the most import goals of extragalactic astronomy . the wide field camera 3 ( wfc3 ) aboard the _ hubble space telescope _ ( _ hst _ )
has significantly expanded the high - redshift frontier with the detection and study of galaxies at @xmath22 .
the ultradeep wfc3/ir observations of the hubble ultra - deep field ( hudf09 ) and nearby fields have so far yielded 73 @xmath23 and 59 @xmath8 lyman - break galaxy ( lbg ) candidates ( * ? ? ?
* see also @xcite ) , including one at @xmath24 @xcite .
these ultradeep observations show tantalizing evidence for a rapid evolution of the rest - frame ultraviolet ( uv ) galaxy luminosity function ( lf ) from @xmath25 to @xmath26 and a declining star - formation rate with increasing redshift @xcite , as expected on the basis of dark matter halo assembly @xcite .
however , because the ultradeep datasets cover only a few wfc3/ir fields , their dynamic range is limited by the small volume that they probe .
in particular they provide poor constraints on the population of bright galaxies at @xmath8 , which are very rare and highly clustered @xcite .
a complete understanding of the number density and overall shape of the lf at @xmath8 requires a large - area survey to search for the brightest galaxies at these epochs .
identifying the brightest @xmath8 candidates from broad - band photometry is also of fundamental importance to provide the best targets for spectroscopic follow - up studies aimed both at confirming the redshift of the sources as well as inferring the properties of the intergalactic medium ( igm ) in proximity of these galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
spectroscopic confirmation of the redshift of lbg galaxies has been carried out for large samples at @xmath27 ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , and recently extended out to @xmath28 @xcite .
these studies show that photometrically selected samples of lbgs have very low contamination ( @xmath29 , see @xcite ) .
it is necessary to extend the frontier of spectroscopy further into the epoch of reionization , at @xmath0 , not only to provide definitive proof that the lbg selection continues to be reliable for candidates into the reionization epoch , but more importantly to infer the ionization state of the igm from the study of the ly@xmath30 equivalent width distribution of lbg sources ( e.g. , * ? ? ?
at present there are several ongoing wfc3/ir surveys designed to identify a relatively large sample of bright ( @xmath31 ; @xmath32 ) @xmath8 galaxies , namely the candels multi - cycle treasury ( mct ) program @xcite , the hubble infrared pure parallel imaging extragalactic survey ( hippies ) @xcite , and the brightest of reionizing galaxies ( borg ) @xcite . in the near term , the sample size of bright @xmath22 galaxies
will also be augmented by cluster lensing surveys ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , including the cluster lensing and supernova survey with hubble ( clash ; @xcite ) mct program , which recently reported the discovery of a magnified lbg candidate at @xmath33 @xcite .
all of these programs take advantage of multi - band _ hst _ optical nir data to search for @xmath8 galaxies as @xmath3 or @xmath4 dropouts using the lyman - break technique @xcite .
currently , the shape of the bright - end of the galaxy uv lf at @xmath34 is debated .
theoretical and numerical investigations predict that it should remain schechter - like @xcite , of the form @xmath35 , at @xmath36 @xcite .
conversely , some numerical studies have suggested a possible excess of sources at the bright end , with a non - schechter behavior tied to the inefficient onset of agn feedback at early times ( e.g. , * ? ? ?
observationally , two recent papers from the candels dataset in goods - south yielded different results . @xcite found a significant excess of bright @xmath8 candidates with @xmath37 over the expectation from a schechter function using data from @xmath38 arcmin@xmath39 of area in the goods - south field .
the shape of their @xmath8 luminosity function resembles a step function ( see their figure 4 ) . on the other hand
, @xcite analyzed the same dataset and derive a well - behaved schechter function for bright @xmath8 sources , consistent with the extrapolation of the lf evolution from @xmath23 and with the predictions of the lf model of @xcite .
the @xcite result ( @xmath40 ) is also consistent with the first - epoch borg ( borg09 ) determination of the knee of the schechter function ( @xmath41 , with the error bar at a fixed @xmath30 and @xmath19 ) .
our initial borg data were about @xmath42 mag shallower than the current candels dataset , but we already had larger sky coverage ( @xmath43 arcmin@xmath44 vs. @xmath38 arcmin@xmath44 ) . here we take advantage of the additional fields recently acquired as part of the borg survey and update the sample of @xmath3 dropouts described in @xcite to include an additional 139 arcmin@xmath2 of area .
our latest catalog now contains eight very bright ( @xmath45 ) @xmath8 candidates and an additional 25 bright @xmath8 candidates detected at lower significance ( @xmath46 ) , with @xmath47 .
this is currently the largest - area search for @xmath8 candidates , now totaling @xmath48 arcmin@xmath44 .
additionally , given the random pointing nature of our pure - parallel _ hst _ program , the borg dataset is distinct in that it is minimally affected by cosmic variance @xcite .
therefore , this catalog is uniquely positioned to set the tightest constraints on the number density of the brightest @xmath8 galaxies and the bright end of the rest - frame ultraviolet galaxy luminosity function ( @xmath49 ; @xmath50 ) .
this paper is organized as follows .
we describe the borg survey in section [ sect : survey ] .
we present the observations and photometry in section [ sect : obs ] and dropout selection in section [ sect : select ] .
the results and constraints on the @xmath8 uv lf are discussed in section [ sect : uvlf ] .
we summarize our results and conclusions in section [ sect : concl ] . throughout this work ,
we assume a cosmology with @xmath51 , @xmath52 , and @xmath53 @xmath54 .
this provides an angular scale of 4.8 ( proper ) kpc @xmath55 at @xmath56 .
all magnitudes are expressed in the ab photometric system @xcite .
in 2009 we initiated the borg survey , a pure - parallel wfc3 survey that complements deep and ultradeep wfc3/ir observations by looking for very bright ( @xmath31 ; @xmath57 ) galaxies at @xmath58 @xcite by obtaining wfc3 imaging in four filters ( f606w , f098 m , f125w , f160w ) on random sightlines at high galactic latitudes ( @xmath59 ) . because luminous massive galaxies at these redshifts are expected to be clustered @xcite , the random pointing nature of our borg survey is ideal to mitigate the severe effects of cosmic variance @xcite .
in fact , our current survey geometry of 59 independent wfc3/ir fields makes our number counts of @xmath8 galaxies essentially follow a poisson distribution and can be used to constrain the luminosity function as well as a contiguous single - field survey with about two times more area at the same depth .
if all the borg area had been in a single field of @xmath60 arcmin@xmath2 , we derive from @xcite that cosmic variance would have dominated over poisson uncertainty for a sample of 33 candidates ( @xmath61 vs. @xmath62 respectively , for a total fractional error of @xmath63 ) . by reducing the uncertainty in the number density of bright sources due to cosmic variance
, we can place stronger observational constraints on @xmath64 ( or @xmath20 ) , the characteristic luminosity of the lf .
this , in turn , can help break the well - known degeneracy between @xmath20 and @xmath30 , the faint - end slope , when fitting a schechter @xcite luminosity function to the data . placing tighter constraints on the value of @xmath30 at redshifts
@xmath65 is crucial in determining the contribution of galaxies to the reionization of the universe .
further , the relative brightness of the lbgs discovered in the borg survey can possibly place @xmath8 galaxies within reach of ground - based spectroscopic follow - up observations , as has been attempted on some of our brighter candidates @xcite .
our ongoing borg survey has covered 274 arcmin@xmath44 to date thanks both to its continuation in cycle 19 ( go 12752 , pi trenti ) and by assimilating data from the similar hippies pure - parallel program @xcite as well as the coordinated parallel observations acquired as part of the cosmic origins spectrograph ( cos ) gto program .
as illustrated in figure [ fig : survey ] , the borg survey is @xmath66 mag shallower than the `` wide '' part of the candels program that has @xmath4-band data , but has a larger area with @xmath3 ( or @xmath4 ) , @xmath5 , and @xmath6 band coverage that is needed to search for @xmath8 galaxies . in this paper
we focus on @xmath8 lbg candidates selected as @xmath3-band dropouts . to use a homogeneous selection of @xmath8 candidates ,
we specifically exclude the hippies cycle 18 dataset , which replaced the f098 m band with the f105w band .
we leave the identification of @xmath8 @xmath4-band dropouts for discussion in a future paper .
the primary goal of the borg survey is to identify galaxies at @xmath8 and measure the bright end of the lbg rest - frame ultraviolet lf at this redshift .
our observations are designed to acquire wfc3 imaging in four filters ( f606w , f098 m , f125w , f160w ) , which are obtained on random and discrete sightlines inherent to this being a pure - parallel program .
we identify bright ( @xmath67 ) high - redshift galaxies by searching for @xmath3 dropouts using the well - known lyman - break dropout technique @xcite .
the @xmath5 and @xmath6 bands are used for source detection and to measure the rest - frame uv color , while the wfc3/uvis @xmath68 band is used to reject low - redshift ( @xmath69 ) interloper galaxies , which are the primary source of contamination to our @xmath58 sample . as discussed in section [ sect :
select ] , our borg survey is optimized to minimize the probability of contamination from both low - redshift interlopers and cool dwarf stars . we also included data from the similar hippies wfc3 pure - parallel program @xcite and the coordinated parallel observations from the cos gto program .
the filter selection of the hippies program and the earlier part of cos gto program differs from borg in that those datasets used the f600lp filter instead of f606w to control for contamination from reddened low - redshift sources . @xcite
discuss the benefits of f606w compared to f600lp .
here we reiterate that , because f606w has a larger transmission efficiency integrated in frequency space over the passband , it reaches deeper at fixed integration time .
this also becomes apparent in the data , for example in field borg_@xmath70 ( table [ tbl : borg09 ] ) , which has deeper data in f606w ( @xmath71 with @xmath72 s ) compared to f600lp ( @xmath73 with @xmath74 s ) despite the shorter integration time .
pure - parallel observations come with a unique set of challenges .
because the primary observations are spectroscopic in nature , dithering is mostly absent in our pure - parallel wfc3 data . as a result ,
detector artifacts ( e.g. , hot / warm pixels ) , uncorrected cosmic rays , and wfc3/ir detector persistence must be carefully considered in the data reduction and data analysis .
to mitigate detector hot / warm pixels and cosmic rays , we employed a robust algorithm based on a variation of laplacian edge detection @xcite to filter the individual flt files prior to combining them with multidrizzle @xcite .
the parameters for the algorithm were chosen to remove most sharp - edged artifacts from the data , but to also be rather conservative so as not to remove any real sources .
as detailed in @xcite , our observing strategy is also unique because it is designed to minimize the impact of wfc3/ir detector persistence from observations taken in orbits immediately prior to our exposures .
we take advantage of the property that the amount of residual image persistence appears to decay roughly as a power law with time .
briefly , observations in either the @xmath5 band or @xmath6 band are preceded in the same orbit by a comparably long @xmath3 band exposure , and whenever possible , @xmath5-band exposures precede those in @xmath6 band .
therefore any persistent residual images from prior observations will be brighter in the bluer bands and thus will not contaminate our high - redshift dropout sample .
multidrizzle @xcite was used to combine individual exposures in a given filter to produce the final science images as well as the associated inverse - variance weight maps .
the images are drizzled to a final pixel scale of 0.08/pixel . because borg is a pure - parallel program , the image depths vary among our fields . for a typical 4-orbit parallel field ,
we obtain exposures of 2200 s , 3800 s , 1800 s , and 1800 s in @xmath68 , @xmath3 , @xmath5 , and @xmath6 , respectively . in a @xmath75 diameter aperture , the corresponding @xmath76 limiting magnitudes are approximately 27.4 , 27.1 , 26.9 , and 26.6 , respectively . in table
[ tbl : borg09 ] , we list the survey fields , exposure times , and @xmath76 limiting magnitudes ( @xmath77 aperture ) of the initial borg09 dataset @xcite , and in table [ tbl : borg12 ] we list the same properties for the new fields in the significantly expanded borg12 dataset .
we have released the reduced drizzled science and rms images for the borg dataset , which is publicly available via the borg website at http://wolf359.colorado.edu/[wolf359.colorado.edu ] and the high level science products on the mast website ] we used sextractor @xcite in dual - image mode for object detection and photometry . for the detection image we used the wfc3/ir @xmath5 data .
the inverse - variance weight images produced by multidrizzle were used to generate rms maps for each image .
we subsequently normalized the rms maps to account for correlated noise introduced by the drizzling procedure ( see * ? ? ? * ) and used them in sextractor for both source detection and photometry .
sources were required to have at least 9 contiguous pixels , each detected at a threshold of 0.7@xmath78 above the background .
object colors and signal - to - noise values were measured using isophotal magnitudes ( isomag ) and total magnitudes were measured within scalable kron apertures ( automag ) .
the photometry in each field has been corrected for galactic extinction using the @xcite dust maps .
all measurements given in this paper include this correction , which is typically modest @xmath79 because the survey primarily targets lines of sight at high galactic latitudes .
we searched for @xmath58 galaxies using a @xmath3-dropout selection criteria in two broadband colors .
the general criteria are a strong break between the @xmath3 and @xmath5 filters , a relatively blue or flat @xmath80 color , and a non - detection in the optical @xmath68 ( or @xmath81 ) band .
specifically , following @xcite we require : @xmath82 . \\\end{aligned}\ ] ] these criteria select galaxies with redshifts in the range @xmath83 ( see section [ sect : uvlf ] ) . to minimize the probability of contamination of our sample by low - redshift interlopers
, we impose a conservative non - detection threshold of @xmath84 on the optical - band data .
the non - detection in the optical @xmath85 band is fundamental to produce a clean sample of @xmath8 candidates ( see @xcite ) .
if a source is not detected in a given filter , we set its magnitude to the corresponding @xmath86 upper limit to calculate its colors . for our final catalog ,
we require sources to be detected with a signal - to - noise ( @xmath87 ) threshold of @xmath88 in @xmath5 and @xmath89 in @xmath6 , as measured in isophotal apertures ( isomag ) .
for the bright @xmath90 catalog , we require sources to be detected with a signal - to - noise threshold of @xmath91 in @xmath5 and @xmath92 in @xmath6 . using these color criteria
, we identified 33 relatively bright @xmath3-dropout galaxy candidates with observed @xmath5 band magnitudes between @xmath9 and @xmath10 mag over our 274 arcmin@xmath44 search area .
eight of these @xmath8 lbg candidates are brighter than @xmath93 mag and are detected at a significance level of @xmath45 in @xmath5 .
this is the largest sample of bright ( @xmath11 ) @xmath8 galaxies presented to date .
we list the properties of the candidates in the @xmath90 and @xmath76 catalogs in tables [ tbl:8sig ] and [ tbl:5sig ] , respectively .
the postage - stamp cutout images of the eight @xmath90 candidates are shown in figures [ fig : stamps1 ] and [ fig : stamps2 ] .
the ( @xmath94 ) and ( @xmath80 ) colors of our @xmath8 candidates along with the color - color selection criteria are illustrated in figure [ fig : colcol ] .
we also show the expected colors of galaxies simulated over a wide range of redshifts and also those of low - mass dwarf stars ( e.g. , @xcite ; holwerda et al .
2012 ( in prep ) ) . as figure
[ fig : colcol ] shows , our conservative @xmath95 color criterion is effective in minimizing contamination from cool dwarf stars and low - redshift interlopers .
in this paper we employ an improved data reduction that includes laplacian filtering ( see section [ sect : obs ] ) .
therefore , we expect some photometric scatter within the measurement uncertainty with respect to the previous catalogs of @xmath3 dropouts published in @xcite . indeed , this is the case : one of the four bright dropouts identified in @xcite , source `` borg1k '' , is now just marginally out of the catalog with @xmath96 .
this is not surprising , as that candidate was at the edge of the selection window in the previous photometry and we were already considering it likely ( @xmath97 ) to be a contaminant that scattered into the dropout selection ( see section 5.2 in @xcite ) . while it was , and still is , uncertain whether this source is at @xmath8 , our artificial source recovery simulations ( see section [ sect : uvlf ] ) take photometric scatter into account when determining the effective volume of the survey , making the derivation of the luminosity function robust . with the new data reduction we also verified the photometry for field borg58 ( here borg_@xmath98 ) , where we identified a @xmath8 protocluster candidate @xcite
the improved photometry confirms the previous measurements , although we note that two of the fainter @xmath8 candidates have scattered out of our current catalog . in this new reduction and catalog , they are detected in the @xmath5band at @xmath99 and @xmath100 , slightly below our formal threshold of @xmath76 detections ( see table [ tbl : borg58 ] ) .
it is worth noting that two new @xmath3-dropout sources are also detected in this field , just below the detection threshold , at @xmath101 in @xmath5 , providing circumstantial evidence that the overdensity extends to fainter luminosities in line with our theoretical and numerical predictions @xcite .
deeper _ hst _ imaging would be very useful to further investigate the nature of this overdensity .
while two additional candidates pass our @xmath90 selection criteria ( see table [ tbl : other ] ) , we exclude them from our final catalog .
one of these sources ( borg_0240@xmath1021857_392 ) is extremely compact and clearly unresolved . given its compactness , the likely explanation is that source is a galactic star with either a significant fluctuation in the photometric measurement or with an unusually red @xmath103 intrinsic color ( at the @xmath104 level ) . because this object is located right above the galactic center ( @xmath105 , @xmath106 ) , the source is most likely a reddened l or t dwarf star .
an exciting , but much less likely , alternative would be a @xmath8 qso , although the limited area of borg implies that this occurs with @xmath107 based on the @xcite qso luminosity function predictions .
the other source we excluded ( borg_1632@xmath1083737_386 ) is extremely bright at 25.1 ab mag in the @xmath5 band and located immediately adjacent ( sextractor extraction flag `` 2 '' ) to an even brighter ( @xmath109 mag ) foreground spiral galaxy .
the spiral galaxy is tidally interacting with a second spiral and has colors similar to the dropout candidate ( @xmath110 , @xmath111 , @xmath112 ) , but it is clearly detected in @xmath68 . in principle , the @xmath3-dropout could have been lensed ( e.g. , see * ? ? ?
* ) , but because of the similar colors with the foreground source , we consider it much more likely that this dropout is part of the foreground system , with its @xmath5-band photometry partially contaminated by emission - line flux ( see * ? ? ?
* ) and a @xmath68 continuum that can not be detected in the current data assuming its @xmath113 is similar to that of the spiral galaxy .
we derive the completeness , @xmath114 , and magnitude - dependent redshift selection function , @xmath115 , of our dataset from simulations as described in @xcite . briefly
, artificial galaxies with a range of spectral energy distributions ( seds ) , luminosities , redshifts , and sizes are added to the real images .
we then rerun our detection and selection procedure on the data in each individual field to determine both @xmath114 and @xmath115 ( see figure [ fig : szm ] ) .
the simulations are based on @xmath116 galaxy images rescaled to the desired input magnitude and to higher redshift using standard evolutionary relations . in particular
we use a size scaling of @xmath117 as determined from lbgs in the range @xmath118 @xcite and adopt a uv - continuum slope distribution of @xmath119 , motivated by the recent measurements at @xmath120 @xcite . from these simulations
we derive the effective volume for our @xmath8 selection as a function of @xmath5magnitude as shown in figure [ fig : effvol ] . by design these simulations
take into account the reduction of the effective area and volume probed by the survey because of photometric scatter and presence of foreground sources , including any persistence images in the @xmath3 data .
we also note that because we are selecting sources from the @xmath5 images , the @xmath121 color is not an unbiased measure of the rest - frame uv slope @xmath122 , so the simulations we perform are crucial to derive the effective volume of our selection , and show that we retain high efficiency , with @xmath123 ( figure [ fig : effvol ] ) .
we construct the stepwise luminosity function from our borg observations in @xmath42 mag bins for the sample of all @xmath124 @xmath3-dropouts and for the smaller , but more robust , sample of @xmath125 sources with high @xmath87 ( @xmath126 ) detections .
the results are shown in table [ tbl : swlf ] and figure [ fig : uvlf ] and take into account @xmath127 contamination for the @xmath128 sample and @xmath129 contamination for the @xmath130 sample based on improved estimates following @xcite .
we take into account the contamination levels of the legacy fields following @xcite and estimate the borg contamination rate by using the early release science ( ers ) wfc3/ir data @xcite , which also used the f098 m filter as the @xmath131 band , in combination with the goods f606w data degraded to match our relative @xmath68 and @xmath3 depths .
the main limitation of this approach is the limited area of the ers data , which is a factor @xmath132 smaller than the borg area .
also , we assume that contamination does not vary as a function of luminosity for a given dataset , which is the working assumption of previous studies deriving the lf at high redshift , both in legacy and pure - parallel fields . in principle
, the contamination might vary with source brightness , depending on the lf difference for contaminants and @xmath8 galaxies with the same near - ir colors . without a spectroscopic follow - up survey , and given the lack of confirmed @xmath8 galaxies even in legacy surveys @xcite , it is challenging , if at all possible , to investigate this in more detail .
yet , it is reassuring that the spectra of the two borg @xmath8 candidates that have been observed at keck with nircam by @xcite and @xcite allow us to exclude the scenario where these sources are @xmath133 galaxies with strong emission lines , which are expected to be the main contaminants of our selection ( see @xcite and @xcite ) . to explore the effect of contamination in borg
, we performed the experiment of leaving the contamination fraction as a free parameter in the lf fit , as discussed below ( see table [ tbl : contam ] ) .
a different contamination fraction would , under our assumptions , simply scale up or down the borg lf measurement .
shifting up our data by @xmath134 dex , assuming no contamination , can be interpreted very conservatively as an upper limit to the bright end of the @xmath8 galaxy luminosity function . as figure
[ fig : uvlf ] shows , the low- and high-@xmath87 samples provide a consistent determination of the luminosity function of @xmath3 dropouts in the borg dataset and the data in our faintest bins agree with ( and are actually slightly lower than ) the brightest bins in @xcite from ers and hudf09 data .
finally , the measure of a strong clustering signal from the borg survey argues against a strong contamination fraction @xcite .
we derive schechter function parameters by performing a maximum - likelihood fit to the data , assuming a poisson distribution of the galaxy number counts in each magnitude bin .
specifically , we maximize the poisson likelihood for observing @xmath135 sources in a given magnitude bin when @xmath136 are expected to be observed from a given schechter lf .
the likelihood @xmath137 is expressed as @xmath138 , where @xmath139 is the poisson probability distribution and the products are taken over all the fields , @xmath140 , and magnitude bins , @xmath141 . to establish the best estimate of the overall shape of the uv lf
, we combine our borg dataset with those from the deeper ers+hudf09 data of @xcite , providing the largest dynamic range in luminosity currently possible .
given that the candels observations yield very different preliminary results depending on the team that analyzed the data to search for @xmath8 galaxies ( see * ? ? ?
* ; * ? ? ?
* ) , we exclude those data from our analysis . at @xmath13 confidence , for @xmath14 we derive @xmath15 mpc@xmath16 , @xmath17 , and a very steep faint - end slope @xmath18 . as observed in figure [ fig : uvlf ] , overall the best fit provides a very good description of the data .
the covariance in the parameter values is smaller than that in the candels+ers+hudf analysis by @xcite ( @xcite do not fit a lf to their data ) , but it is still significant , as shown in figure [ fig : lfparam ] . from this figure
it is also clearly evident that our @xmath20 versus @xmath30 parameter values have much better constraints than previous studies at @xmath8 @xcite .
formally the bright end of the luminosity function has a marginally low @xmath142 value .
however , this is compensated by the low @xmath19 such that the best - fit schechter function ( @xmath143 mpc@xmath144 ) derived by @xcite is fully consistent with our determination at @xmath13 confidence .
as can be seen in figure [ fig : uvlf ] , the faintest three bins of the borg dataset match very closely the brightest bins in @xcite .
the latest borg lf is also very similar to our earlier fit from @xcite , where we had kept @xmath19 and @xmath30 fixed ( see table [ tbl : lfparam ] ) .
table [ tbl : contam ] shows the impact of varying the contamination fraction @xmath145 for the borg dataset and its associated contamination correction @xmath146 that is applied to the intrinsic counts per magnitude bin to obtain the estimated observed counts .
the table shows that the impact of contamination around our fiducial value @xmath147 is relatively modest .
we note that the highest likelihood is associated to a contamination fraction just below our fiducial value , providing indirect evidence to suggest that our derived contamination fraction is robust . varying the contamination fraction around @xmath148 has almost no impact on the determination of @xmath20 , as the lf is well sampled by the borg data in this luminosity range , but affects mostly @xmath30 ; if we have underestimated @xmath145 , then the faint - end slope would be even steeper .
as shown in table [ tbl : lfparam ] , our determination of the @xmath8 luminosity function is consistent with previous work @xcite , taking into account the significant uncertainties in all these measures .
formally , our best fit prefers a low normalization for @xmath19 and a brighter @xmath20 ( similar to the @xmath23 value ) , but the covariance between these parameters is very large ( see figure [ tbl : lfparam ] ) .
in particular our borg+ers+hudf lf fit provides a better constraint on the @xmath20vs .
@xmath30 parameter uncertainties than the candels+ers+hudf determination at @xmath8 @xcite . in comparing different datasets ,
it is important to note that the bright end is still sparsely sampled ; for example , the brightest borg magnitude bin includes only one source .
therefore , small - number fluctuations are very significant and poisson noise in the brightest bins can have a large impact on the fit
. however , our maximum likelihood fit shows , independent of the @xmath20 versus @xmath19 degeneracy , that the faint - end slope @xmath30 is being constrained with growing accuracy . of note , we reduce its @xmath86 uncertainty at @xmath8 to @xmath149 and show in figure [ fig : alpha ] that the steepening trend suggested by previous studies ( e.g. , * ? ? ?
* ) is also supported by our lf fit .
the steep faint - end slope derived in our best - fit lf implies that faint galaxies are playing a key role as major producers of ionizing photons . for a schechter lf ,
most of the luminosity density contribution is at the faint end ( e.g. , see left panel of fig . 3 in @xcite ) .
furthermore , in the case of @xmath150 , there is a logarithmically divergent contribution from sources below the detection limit of the survey .
this implies that the total ionizing flux produced by galaxies is sensitive to the exact measure of @xmath30 ( as well as on the extrapolation beyond the detection limit ) , as shown , for example , in the left panel of fig . 4 of @xcite .
assuming that star formation continues to be efficient in lower luminosity galaxies down to @xmath151 at @xmath0 ( the limit of atomic hydrogen cooling halos , e.g. see @xcite ) , then we expect that the current _ hst _
observations are only observing @xmath152 of the light present at @xmath153 . a direct proof that we are detecting only the tip of the iceberg of star formation during the epoch of reionization
is provided by _ hst
_ observations of high - redshift , spectroscopically confirmed grbs that have failed to detect host galaxies despite reaching ultrafaint sensitivity ( @xmath154 , see @xcite ) .
the non - detection of galaxies at locations in the sky where it is confirmed that star formation is happening ( because a grb explosion has occurred ) implies that @xmath155 galaxies were indeed the main ionizing sources @xcite , in full agreement with the interpretation of the lf fit derived in this paper .
interestingly , our determination of the @xmath8 luminosity function at the bright end is between the debated measurements from the candels dataset obtained by @xcite and @xcite .
however , our data clearly show a well - behaved schechter function , similar to @xcite , without the unusual shape derived by @xcite . our lf determination and
that by @xcite are consistent at the @xmath104 level .
the observed differences could be partially related to the environment ( i.e. cosmic variance @xcite ) .
@xcite discuss that the ers measurement by @xcite could have been affected by an overdensity .
however , given that the borg dataset is not affected by cosmic variance and agrees with @xcite , it may well be that the candels - south field is underdense , which would explain the lower @xmath20 value derived from that dataset .
the upcoming borg observations and those in the candels - north field will help clarify the nature of the discrepancy and provide further improvements to the determination of the bright end of the @xmath8 luminosity function .
we present the discovery of 33 lyman break galaxy ( lbg ) candidates at @xmath8 detected in _
hst_/wfc3 imaging as part of the brightest of reionizing galaxies ( borg ) pure - parallel survey .
our sample of bright @xmath3-dropout galaxy candidates have @xmath5-band magnitudes between @xmath9 and @xmath10 mag , obtained from @xmath156 independent lines of sight over a total area of @xmath1 arcmin@xmath2 .
the pure - parallel nature of borg allows us to obtain an estimate of the galaxy luminosity function which is unaffected by large - scale structure uncertainty and distinct from determinations using legacy surveys limited to a single , or a few , contiguous fields . with our new data we detect galaxies between @xmath157 , demonstrating the bright end of the @xmath8 luminosity function
is well described by a schechter form , similar to that found at lower redshifts .
our measurement of the number density of galaxies near the borg detection limit ( @xmath158 ) is in agreement with the results by @xcite .
their measurement is based on ers+hudf09 observations that extend to much fainter luminosities , but they lack the area to detect the brighter galaxies at @xmath159 we identify in the borg survey .
the two datasets are complementary and allow us to obtain the best determination yet of the galaxy luminosity function at @xmath8 , with @xmath160 mpc@xmath16 , @xmath17 , and @xmath161 for @xmath14 .
covariance in the parameters , albeit reduced thanks to the increase in the dynamic range of the fit , is still very significant . however , this affects primarily @xmath20 versus @xmath19 .
we find that our borg+ers+hudf lf fit provides a better constraint on the @xmath20 versus @xmath30 parameter uncertainties than the candels+ers+hudf determination at @xmath8 @xcite .
the faint - end slope uncertainty is starting to be reduced to a point where there is a hint of steepening compared to @xmath162 .
this steepening is expected based on numerical and theoretical models of galaxy formation in the epoch of reionization @xcite . as a consequence of this large abundance of faint galaxies
, it is expected that such systems will dominate the total star formation rate and ionizing photon production @xcite , which has recently been confirmed observationally by the non - detection of host galaxies in a sample of six @xmath163 gamma - ray bursts @xcite .
our determination of a steep @xmath30 for a schechter fit of the lf is robust against our estimate of the contamination fraction of the borg survey ( @xmath147 for the @xmath76 sample ) ; as shown in table [ tbl : contam ] , if @xmath145 were higher , @xmath30 would be steeper . in the near future ,
the upcoming ultradeep observations of go 12498 ( pi ellis ) will improve the determination of the number density of the faintest galaxies observable by _ hst _ , while scheduled borg and candels - north observations will further increase the search area for the brightest galaxies at @xmath8 , allowing us to further tighten the constrains on the lf .
the @xmath8 candidates identified here are good targets for follow - up observations from ground and space observatories .
we have started a spectroscopic campaign to confirm their redshift , to measure the distribution of ly@xmath30 equivalent width ( which is related to the igm ionization state ) , and also to rule out contamination from emission - line galaxies at @xmath133 @xcite .
the availability of multi - object spectrographs both on keck ( mosfire ) and gemini south ( flamingos-2 ) , will enable faster progress , especially in fields that have overdensities of sources similar to the protocluster candidate we reported in @xcite .
in addition , the brightest borg sources are expected to have star formation rates of @xmath164 @xmath165 @xmath54 based on their rest - frame uv luminosity , making them prime targets for alma observations as they are expected to fall well within the telescope sensitivity based on the @xcite predictions .
if these candidates are spectroscopically confirmed , this would allow us to possibly extend the detection of 158 @xmath166 m , high-@xmath167 co emission lines , and perhaps dust emission as well , to @xmath8 , after the recent record established at @xmath168 from a qso host galaxy @xcite . finally , _
spitzer_/irac observations , combined with spectroscopic redshifts , would also be very useful to quantify the rest - frame optical properties of these sources before the advent of the _ james webb space telescope_. the space telescope science institute is operated by aura inc .
, under nasa contract nas5 - 26555 .
this work was supported by grants hst - go-11563 , hst - go-11700 , and hst - go-12572 .
this work included coordinated parallel data taken as part of the cos gto program ( pi green ) .
support for po was provided by nasa through hubble fellowship grant hf-51278.01 .
js acknowledges support from grant nnx08-ac14 g .
mt gratefully acknowledges the hospitality of the kavli institute for theoretical physics , where part of this research was carried out with support by the national science foundation under grant no .
phy11 - 25915 .
part of the work presented in this paper was performed by mt and tt while attending the program
first galaxies and faint dwarfs : clues to the small scale structure of cold dark matter " at the kavli institute of theoretical physics at the university of california santa barbara . , a. m. , fruchter , a. s. , hook , r. n. , & hack , w. 2002 , in the 2002 hst calibration workshop : hubble after the installation of the acs and the nicmos cooling system , ed .
s. arribas , a. koekemoer , & b. whitmore , 337 | we report the discovery of 33 lyman break galaxy ( lbg ) candidates at @xmath0 detected in _ hubble space telescope ( hst ) _ wide field camera 3 ( wfc3 ) imaging as part of the brightest of reionizing galaxies ( borg ) pure - parallel survey .
the ongoing borg survey currently has the largest area ( @xmath1 arcmin@xmath2 ) with @xmath3 ( or @xmath4 ) , @xmath5 , and @xmath6 band coverage needed to search for @xmath7 galaxies , about three times the current candels area , and slightly larger than what will be the final candels wide component with @xmath4 data ( required to select @xmath8 sources ) .
our sample of 33 relatively bright @xmath3-dropout galaxies have @xmath5 band magnitudes between @xmath9 and @xmath10 mag .
this is the largest sample of bright ( @xmath11 ) @xmath8 galaxy candidates presented to date . combining our dataset with the hubble ultra - deep field ( hudf09 ) dataset
, we constrain the rest - frame ultraviolet galaxy luminosity function at @xmath8 over the widest dynamic range currently available .
the combined datasets are well fitted by a schechter function , i.e. @xmath12 , without evidence for an excess of sources at the bright end . at @xmath13 confidence , for @xmath14 we derive @xmath15 mpc@xmath16 , @xmath17 , and a very steep faint - end slope @xmath18 . while the best - fit parameters still have a strong degeneracy , especially between @xmath19 and @xmath20 , our improved coverage at the bright end
has reduced the uncertainty of the faint - end power - law slope at @xmath8 compared to the best previous determination at @xmath21 . with a future expansion of the borg survey , combined with planned ultradeep wfc3/ir observations , it will be possible to further reduce this uncertainty and clearly demonstrate the steepening of the faint - end slope compared to measurements at lower redshift , thereby confirming the key role played by small galaxies in the reionization of the universe . |
within the quantum world , the question what quantities can be measured in an experiment is often a non - trivial one , e.g. , measuring time correlators ( with times @xmath0 ) requires finding the correct ordering of operators .
concrete examples in mesoscopic physics and quantum optics are the measurements of charge correlators @xcite or full counting statistics @xcite and that of photon correlators @xcite . the question is usually resolved by including the measurement apparatus in the description and its internal workings decide upon the form of the measured correlator .
examples are the ampremeter , double - dot detector , and spin counter used in refs . , , and , or the different photodetectors introduced by glauber and by mandel @xcite .
these detectors then act back on the system , thereby influencing the measurement outcome , i.e. , the specific form of the correlator .
e.g. , a weak measurement as used in ref . , see also ref . , leads to symmetrized ( @xmath1 ) and antisymmetrized ( @xmath2 ) correlators weighted with different detector response functions @xcite , while a strong measurement produces a projected correlator @xmath3 .
these different forms of measured correlators can be derived @xcite by invoking the von neumann projection postulate @xcite , telling how to restart the system after the first measurement at @xmath4 ( or after previous measurements at times @xmath5 in a @xmath6-th order correlator ) .
the measurement with a weakly coupled detector can be treated perturbatively , with the von neumann projection excerted on the detector and no back action on the system @xcite . in a strong measurement with a large system detector coupling
, the projection formally can be applied directly to the system , therefore producing a maximal back action @xcite .
schematic illustration of different measurement procedures for a two - time ( @xmath4 , @xmath7 ) correlator : ( a ) strong measurement described by von neumann projection acting directly on the system at time @xmath4 and providing a projected correlator @xmath8 after readout at @xmath9 ; for a weak measurement , the von neumann projection at time @xmath4 acts on the weakly coupled detector .
( b ) repeated measurement without von neumann projection at @xmath4 : unitary co - evolution of system and quantum memories which are entangled at times @xmath4 and @xmath9 and final readout after @xmath9 .
the coupling strength between the system and the quantum memories determines the degree of entanglement.,width=302 ] in this paper , we invoke the principle of deferred measurement @xcite known from quantum information theory , where it can be used for quantum computing , and apply it to the problem of repeated measurements , specifically , of time - correlators .
we replace the von neumann projection by entangling the measured system with co - evolving quantum memories , see fig.[fig : comp ] , thereby ( effectively ) expanding the hilbert space of the total system in every measurement step .
the desired correlator then is derived from a final measurement of all the quantum memories by invoking born s rule @xcite .
hence the entire system plus memories undergoes a unitary quantum evolution until the very end , where the born rule takes us from the quantum to the classical world .
the new scheme captures the cases of weak and strong measurement within a unique formalism by merely changing the degree of entanglement between the system and the quantum memory . in the limits of weak and strong entanglement
, we reproduce the results previously derived via use of the projection postulate .
no simple physical form for the time - correlators could be found so far in the intermediate coupling regime . describing a measurement by entangling the system with a detector and including a ( dissipative ) bath in the evolution of the density matrix is a concept that has been well developed over the past two decades @xcite . here
, we extend the idea of system detector entanglement to the case of repeated measurement . thereby , the quantum memories evolve coherently during and after the information transfer from the system due to entanglement , with the ( dissipative ) measurement deferred to the very end of the process .
the proposed scheme for the measurement of time - correlated observables finds an interesting application in mesoscopic physics .
in particular , measuring the charge @xmath10 dynamics of a quantum dot with the help of a nearby quantum point contact is a classic problem by now@xcite . in such a setup , single electron pulses
@xcite assume the role of flying qubit memories which are either transmitted or reflected by the quantum point contact ( qpc ) , depending on the dot s charge state , see fig.[fig : meso_setup ] .
analyzing the charge transmitted across the qpc then provides the desired information on the dot s charge correlator . making use of recent developments in electron quantum optics @xcite
, we propose a setup , see fig .
[ fig : exp ] , that shifts the task of resolving short times , typically done on the level of the detection , to the proper timing of electron pulses and gate operations .
we briefly sketch the main idea of the paper : consider a system observable @xmath11 with an eigenbasis @xmath12 that is to be measured .
we start with a system state at the initial time @xmath13 @xmath14 and a quantum memory in the initial state @xmath15 and have them transiently interact at time @xmath4 with the help of an externally controlled interaction ( to be identified with a quantum detector ) .
the system and memory then become entangled , @xmath16 |\phi_\mathrm{in}^{\scriptscriptstyle ( 1 ) } \rangle \to \sum_n \psi_n(\tau_1 ) |n\rangle |\phi_n^{\scriptscriptstyle ( 1)}\rangle,\end{aligned}\ ] ] where @xmath17 denote memory states after interaction with the system in state @xmath18 and we assume a negligible evolution of the system during the time of interaction . evolving the system to the later time @xmath9 and entangling it with a second memory ,
we obtain the state @xmath19 with @xmath20 the matrix elements of the system propagator @xmath21 and @xmath22 . the unitary evolution of the memory states @xmath23 preserves the system information gained at times @xmath0 . at time
@xmath24 we measure the memory observables @xmath25 and @xmath26 with discrete spectra @xmath27 .
making use of born s rule , we find the probability distribution function @xmath28 for the measurement outcomes @xmath29 and @xmath30 on the two memory observables ; this probability distribution contains the desired information on the system s two - time correlator .
the specific relation between the distribution function @xmath28 of measurement outcomes and the correlator of system observables @xmath31 and @xmath32 depends on the system detector coupling and the observables measured on the memories ; we will show below how to extract the well known ( anti-)symmetrized and projected correlators from the probabilities @xmath28 in the limits of weak and strong measurements . in the end , by making use of quantum memories which store the information acquired from the system at the quantum level at earlier times @xmath4 and @xmath9 until the final time @xmath33
, we have avoided the intermediate readout which requires the use of the projection postulate .
hence the entire measurement process follows a unitary quantum evolution until the transition to the classical world is done via born s rule . in the following
, we will first derive the general framework describing the deferred measurement of a correlator with the intermediate von neumann projection replaced by a system detector entanglement ( sec .
[ sec : uni ] ) . in section
[ sec : gen_wc ] , we discuss the limit of weak measurement and use qubits as quantum memories to arrive at a simple relation between the measurement outcome on the qubits and the ( anti- ) symmetrized correlators of the system . in sec .
[ sec : gen_sc ] , we first discuss a strong measurement at strong coupling using qudit memories and then invoke ( weakly coupled ) qubit registers to show that both types of strong measurements produce the projected time correlator of the system .
an illustration of our formalism is given in sec .
[ sec : charge ] , where we describe the measurement of the charge correlator in a mesoscopic setting , specifically , the two - time charge correlator of a quantum dot ( qd ) as measured by a quantum point contact ( qpc ) .
[ sec : charge_exp ] describes a possible experimental implementation and in sec .
[ sec : con ] we summarize our results .
we consider the situation where a two - time correlator of a system operator @xmath11 is measured with the help of two quantum memories ; the system s initial state @xmath34 is given by eq .
( [ eq : psi ] ) , while the memories are described by initial states @xmath35 , @xmath36 ( see below for the discussion of an open system described by a density matrix @xmath37 ) .
the memories interact with the system at times @xmath38 during a small time intervall @xmath39 . after this interaction , the resulting memory states @xmath40
depend on the system state @xmath18 , where @xmath41 describes the time evolution of the quantum memories during the interaction with the system ( we assume a trivial evolution of the free memories ) .
the system state @xmath42 is assumed to remain unchanged during the time @xmath39 of the individual interaction events , see sec .
[ sec : finitewidth ] for an extended discussion of this point .
after the second interaction event at @xmath9 , the wave function @xmath43 of the system is entangled with the states @xmath44 of the memories and the combined wave function @xmath45 reads @xmath46 with @xmath47 .
the quantum memories are supposed to keep their system information after their interaction . at time @xmath33
, we measure the operators @xmath48 and @xmath26 on the first and second memory , respectively . denoting the ( discrete ) eigenvalues and eigenstates of @xmath49 by @xmath50 and @xmath51 , we rewrite the memory states @xmath52 .
applying born s rule to the final state ( [ eq : psif ] ) provides us with the probability distribution @xmath53 where @xmath54 is the projector onto the eigenstate @xmath29 of the @xmath55-th memory , i.e. , @xmath56 . making use of the unitarity condition @xmath57 ( rendering the evolution @xmath58 in eq .
( [ eq : psif ] ) irrelevant ) , we obtain the probabilities @xmath59 the above expressions ( [ eq : pabdef ] ) and ( [ eq : pab ] ) for the probabilities @xmath60 provide us with the desired information on the two - time correlator of the system .
they are easily generalized to the case of open systems by introducing the combined system plus bath ( the open system ) density matrix @xmath61 and evolve it in time including the subsequent entanglement with the quantum memories : starting from the initial density matrix @xmath62 describing the open system plus memories at time @xmath63 , we proceed as in the case of isolated systems by conditioning the time evolution of the memory states on the corresponding system states and obtain the final density matrix at time @xmath64 @xmath65 with the open system s density matrix @xmath66 at time @xmath4 , its reduced part @xmath67 , and the reduced operators @xmath68 with @xmath69 the evolution operator of the open system .
note that here , the outgoing states @xmath70 and @xmath71 are conditioned on the system states @xmath72 and @xmath73 at times @xmath4 and @xmath9 .
we define the probabilities @xmath28 as @xmath74\end{aligned}\ ] ] with the trace taken over both the open system and the memory states . calculating this expression with the final density matrix eq .
, we obtain @xmath75 , \nonumber\end{aligned}\ ] ] with the remaining trace taken over the bath degrees of freedom . this result is the direct generalization of to the case of open systems .
the expressions ( [ eq : pabdef ] ) and ( [ eq : pab ] ) as well as ( [ eq : pabdef_o ] ) and ( [ eq : pab_o ] ) constitute the basic formulas which we will further develop in the following sections .
indeed , as expressed in terms of evolution amplitudes of system and detectors , it is difficult to appreciate the physical meaning and content of these results . in order to make progress , we consider next the two cases of weak and strong measurements . given a _ weak system detector coupling _ , the most direct way to find the probabilities @xmath28 in terms of physically transparent quantities is to start from eq .
( [ eq : pabdef ] ) and evaluate this expression perturbatively in the linear system detector coupling @xmath76 , where the time - dependent coupling @xmath77 acts on the @xmath55-th memory during a time @xmath78 around @xmath0 . the unperturbed evolution of the memories is described by the hamiltonian @xmath79 and we make use of the interaction representation .
we go over to irreducible quantities by subtracting the uncorrelated contribution , @xmath80 with @xmath81 and latexmath:[$p_\beta^{\scriptscriptstyle ( 2 ) } = \langle \psi_f^{\scriptscriptstyle ( 2)}| { \hat{p}}_\beta^{\scriptscriptstyle ( 2 ) } involving a single entanglement at time @xmath4 or @xmath9 with only one memory , respectively .
the quantity @xmath83 can be obtained by a simple summation of @xmath28 , @xmath84 and @xmath85 for a time - independent problem ( otherwise , the determination of @xmath86 necessitates a second measurement ) .
note that the sum over the first index of @xmath28 already includes correlations , see eq .
( [ eq : pab ] ) , and hence @xmath87 .
the task then is to evaluate the irreducible expression @xmath88 with the expectation value to be taken over the initial system state @xmath89 , @xmath90 refers to the irreducible part , and the time evolution operator reads @xmath91,\ ] ] with @xmath92 denoting time - ordering .
evaluating to lowest relevant order in the coupling , we find @xmath93 , \hat{h}_\mathrm{sd}(\tau '' ) ] | \psi \rrangle \nonumber\\ & = \frac{(-i)^2}{\hbar^2 } \int_{\tau_{\rm in}}^{\tau } \!\!d\tau ' \int_{\tau_{\rm in}}^{\tau ' } \!\!d\tau '' \bigl [ \llangle \psi| { \hat{o}}(\tau ' ) { \hat{o}}(\tau '' )
|\psi\rrangle \ , \langle \phi_\mathrm{in}^{\scriptscriptstyle ( 1 ) } |{\hat{p}}_{\alpha}^{\scriptscriptstyle ( 1)}(\tau_f ) { \hat{b}}^{\scriptscriptstyle ( 1)}(\tau '' ) |\phi_\mathrm{in}^{\scriptscriptstyle ( 1)}\rangle \ , \langle \phi_\mathrm{in}^{\scriptscriptstyle ( 2 ) } |{\hat{p}}_{\beta}^{\scriptscriptstyle ( 2)}(\tau_f ) { \hat{b}}^{\scriptscriptstyle ( 2)}(\tau ' ) |\phi_\mathrm{in}^{\scriptscriptstyle ( 2)}\rangle \nonumber\\ \noalign{\vspace{-8pt } } & \hspace{97pt}- \llangle \psi| { \hat{o}}(\tau '' ) { \hat{o}}(\tau ' ) |\psi\rrangle \ , \langle \phi_\mathrm{in}^{\scriptscriptstyle ( 1)}|{\hat{b}}^{\scriptscriptstyle ( 1)}(\tau '' ) { \hat{p}}_{\alpha}^{\scriptscriptstyle ( 1)}(\tau_f ) |\phi_\mathrm{in}^{\scriptscriptstyle ( 1)}\rangle \ , \langle \phi_\mathrm{in}^{\scriptscriptstyle ( 2)}i |{\hat{p}}_{\beta}^{\scriptscriptstyle ( 2)}(\tau_f ) { \hat{b}}^{\scriptscriptstyle ( 2)}(\tau ' ) |\phi_\mathrm{in}^{\scriptscriptstyle ( 2)}\rangle \nonumber\\ \noalign{\vspace{-4pt } } & \hspace{97pt}- \llangle \psi| { \hat{o}}(\tau ' ) { \hat{o}}(\tau '' ) |\psi\rrangle \ , \langle \phi_\mathrm{in}^{\scriptscriptstyle ( 1)}i |{\hat{p}}_{\alpha}^{\scriptscriptstyle ( 1)}(\tau_f ) { \hat{b}}^{\scriptscriptstyle ( 1)}(\tau '' ) |\phi_\mathrm{in}^{\scriptscriptstyle ( 1)}\rangle \ , \langle \phi_\mathrm{in}^{\scriptscriptstyle ( 2)}|{\hat{b}}^{\scriptscriptstyle ( 2)}(\tau ' ) { \hat{p}}_{\beta}^{\scriptscriptstyle ( 2)}(\tau_f ) i |\phi_\mathrm{in}^{\scriptscriptstyle ( 2)}\rangle \nonumber\\ \noalign{\vspace{-4pt } } & \hspace{97pt}+ \llangle \psi| { \hat{o}}(\tau '' ) { \hat{o}}(\tau ' ) |\psi\rrangle \ , \langle \phi_\mathrm{in}^{\scriptscriptstyle ( 1)}|{\hat{b}}^{\scriptscriptstyle ( 1)}(\tau '' ) { \hat{p}}_{\alpha}^{\scriptscriptstyle ( 1)}(\tau_f ) |\phi_\mathrm{in}^{\scriptscriptstyle ( 1)}\rangle \ , \langle \phi_\mathrm{in}^{\scriptscriptstyle ( 2)}|{\hat{b}}^{\scriptscriptstyle ( 2)}(\tau ' ) { \hat{p}}_{\beta}^{\scriptscriptstyle ( 2)}(\tau_f ) |\phi_\mathrm{in}^{\scriptscriptstyle ( 2)}\rangle \bigr ] , \label{eq : pirrexpr}\end{aligned}\ ] ] where we made sure that the first memory interacts with the system at the earlier time @xmath94 . for a slow system dynamics and exploiting that @xmath95 only for @xmath96 , we can replace @xmath97 and @xmath98 .
we make use of the standard definitions for the symmetrized and anti - symmetrized irreducible correlators ( with @xmath99 $ ] and @xmath100 denoting the usual commutator and anti - commutator ) @xmath101\bigr\rrangle,\end{aligned}\ ] ] to arrive at the final result @xmath102 with the detector response functions @xmath103| \phi^{\scriptscriptstyle ( j)}_\mathrm{in } \rangle .
\label{eq : isdet}\end{aligned}\ ] ] the symbols @xmath104 and @xmath105 address symmetrized and anti - symmetrized quantities ( or equivalently , up to factors of 2 , real- and imaginary parts ) . in a situation where the full information @xmath28 can be extracted from the memories , the individual correlators @xmath106 and
@xmath107 can be obtained from ( [ eq : pabweak ] ) by combining two different probabilities , e.g. , using @xmath108 and @xmath109 one obtains @xmath110/ d^{\scriptscriptstyle ( 1)}\ , \mathcal{i}s^{\scriptscriptstyle ( 2)}_{{\rm det},\beta } , \\ \label{eq : isoof } \mathcal{i}s^{\rm irr}_{{o}{o } } & = [ -\mathcal{i}s^{\scriptscriptstyle ( 1)}_{{\rm det},\bar\alpha } p^{\rm irr}_{\alpha\beta } + \mathcal{i}s^{\scriptscriptstyle ( 1)}_{{\rm det},\alpha } p^{\rm irr}_{\bar\alpha\beta } ] / d^{\scriptscriptstyle ( 1)}\ , \mathcal{i}s^{\scriptscriptstyle ( 2)}_{{\rm det},\beta},\end{aligned}\ ] ] with @xmath111 .
alternatively , one may have preferential access to combinations of probabilities @xmath28 ( see sec .
[ sec : charge_exp ] for an example ) or make use of specific detector properties , see below and the appendix for examples . a generic choice for the quantum memories
are qubit devices that couple to the system via the hamiltonian @xmath112 , where the coupling @xmath113 is switched on during a short time @xmath39 around @xmath0 .
assuming initial states @xmath114/\sqrt{2},\end{aligned}\ ] ] a system residing in a state @xmath18 with eigenvalue @xmath115 will rotate the qubit around the @xmath116-axis by @xmath117 with @xmath118 .
the response functions ( [ eq : rsdet ] ) and ( [ eq : isdet ] ) involve integrals of the type @xmath119 and we obtain the response functions @xmath120 choosing @xmath121 , i.e. , polarizing the second memory along the @xmath122-axis , we can directly find the correlator @xmath123 ( @xmath124 ) from the memory correlator @xmath125 by choosing @xmath126 ( @xmath127 ) .
alternatively , we may use the results ( [ eq : rsoof ] ) and ( [ eq : isoof ] ) and @xmath128 to find ( we choose @xmath129 , @xmath130 , and @xmath131 ) @xmath132 note that the sums @xmath133 and @xmath134 contain very different types of information , once a correlation , the other time only a mean value , as discussed in more detail in sec .
[ sec : charge_exp ] .
hence , we find that the delayed measurement of the two quantum memories provides the symmetrized and anti - symmetrized correlators ( [ eq : rsoo ] ) and ( [ eq : isoo ] ) .
finally , we note that a weak linear coupling between the system observable and the detector / memory variable canonically conjugated to the detector readout provides a more effective entanglement .
such a von neumann like interaction allows to produce a strong entanglement and a strong measurement even for weak coupling if sufficient time is available for the entanglement process@xcite .
we are now going to show that a _ strong system detector coupling _ naturally leads to projective correlators .
consider a situation where the system operator @xmath11 is measured via entanglement with two quantum memories .
we first consider an operator @xmath11 with a non - degenerate spectrum and comment on the general case in the end .
a strong coupling between the system and the memory implies that the memory states @xmath135 after interaction with the system in state @xmath18 are fully distiguishable , i.e , @xmath136 the observables @xmath49 distinguish between memory eigenstates @xmath51 and we assume a one - to - one relation with the evoluted memory states @xmath44 , @xmath137 with @xmath138 for @xmath139 ( otherwise , the observable @xmath49 measures linear combinations of eigenstates of @xmath11 and thus is not suitable for a measurement of this observable ) . under these ( strong coupling ) conditions ,
the amplitudes @xmath140 reduce to @xmath141 ( for @xmath142 , @xmath143 for the second memory ) . in order to describe the strong coupling situation
it is favorable to proceed with the expression ( [ eq : pab ] ) and we obtain the result @xmath144 the right hand side of the above expression is nothing but the projected system correlator @xmath145 with the projected density matrix @xmath146 @xmath147 and the projectors @xmath148 onto different system states , see eq.([eq : psi ] ) .
the projected correlators @xmath149 are easily combined into the desired two - time correlator @xmath150 @xmath151 \\
\nonumber & = \sum_{nm } o_n o_m s^{p}_{{p}_{n}{p}_{m}}(\tau_{21})\\ \nonumber & = \sum_{nm } o_n o_m p_{\alpha_n \beta_m}(\tau_{21}),\end{aligned}\ ] ] thereby establishing the general result relating the system correlator to the memory readings .
a further simplification can be achieved if the eigenvalues of @xmath49 and @xmath11 obey a linear relation @xmath152 , in which case we obtain the simple result @xmath153 directly relating the projected system correlator to the memory correlator @xmath154 .
these results can easily be generalized to the case where the observable @xmath11 involves a degenerate spectrum : the system evolution of the memories is conditioned on the eigenvalue @xmath115 rather than the state @xmath18 of the system .
degenerate eigenstates produce equal evolutions of the memories , implying that @xmath155 has to be replaced by @xmath156 and @xmath157 is substituted by @xmath158 with the projection operator @xmath159 . with these replacements , the above results remain valid also for a degenerate spectrum . hence , the perfect entanglement between the system and the memories arising due to strong coupling is equivalent to a von neumann projection applied to the system . while no back action is apparent on the level of the system dynamics , the strong back action of this maximal entanglement with the memories manifests itself in a strong change of the system s density matrix when tracing over the memories .
the realization of a strong measurement as described above requires equal or more memory states @xmath160 than system states @xmath18 , hence quantum memories with dimensionality @xmath161 or qudits are required .
besides the obvious difficulty in realizing qudit memories , their coupling to the system in order to serve as a measurement device is equally nontrivial . as an alternative
, we discuss a measurement scheme involving qubit registers , instead .
such an alternative setup implementing a strong measurement involves a weak system detector coupling but invokes multiple measurements .
we replace the strongly coupled qudit memories by weakly coupled qubit registers with @xmath162 qubits each , probing the system state close to @xmath4 and @xmath9 within a short time interval @xmath39 . for a strong measurement , @xmath162
is chosen sufficiently large to distinguish the different eigenvalues @xmath115 of the system . assuming the system not to change during the interaction time @xmath163 with one register , the final state of system and memory registers is of the same form as in eq . with the outgoing individual memory states
@xmath164 and @xmath165 replaced by the outgoing register states @xmath166 .
making use of born s rule , we obtain the probabilities @xmath167 for finding @xmath168 ( @xmath169 ) qubits of the first ( second ) register in states @xmath170 ( @xmath171 ) , @xmath172 with @xmath173 and @xmath174 . introducing the conditional probability @xmath175 for measuring @xmath168 of the @xmath162 qubits in the state @xmath29 after interaction with the system in state @xmath18 ,
we can separate the system- and detector response in the above equation , @xmath176 p_\alpha^\mu(n , n ' ) p_{\beta}^\nu(m)\end{aligned}\ ] ] with the ` off - diagonal ' conditional probabilities @xmath177 ( note that @xmath178 ) .
the conditional probabilities @xmath179 depend only on the eigenvalue @xmath115 of the state @xmath18 ( and not on the state @xmath18 itself ) .
we then have to distinguish two cases : ( i ) all system eigenvalues are non - degenerate , i.e. , @xmath180 for @xmath181 , ( ii ) there are degenerate system states . in the non - degenerate case ( i ) and for a strong measurement , the probability distributions @xmath179 and @xmath182 for different @xmath183 do not overlap as functions of @xmath168 ( this is the very definition of this measurement being a strong one ) and
the ` off - diagonal ' elements @xmath184 for @xmath185 are suppressed , as follows from the relation @xmath186 .
we then can simplify the expression ( [ eq : pabmnrp ] ) to the form @xmath187 .
the register correlators ( replacing the distribution functions @xmath28 ) take the form @xmath188 assuming again a linear system qubit coupling @xmath189 that rotates the qubits by an angle @xmath190 around the @xmath116-axis , the evolution @xmath191 produces the memory states @xmath192 , where we again assume initial states polarized in the @xmath193-plane , see ( [ eq : q_in ] ) .
the probabilities @xmath194 for an individual qubit to reside in state @xmath195 are given by @xmath196 with the qubits initially polarized along the @xmath122-axis , i.e. , @xmath197 , we define the register s ` magnetizations ' @xmath198 = 2 j \delta\vartheta\ , o_n,\end{aligned}\ ] ] where we have made use of ( [ eq : prob_q ] ) in the last equation .
the combination @xmath199 then involves the product of register polarizations @xmath200 and using the relation @xmath201 , see eq.([eq : spoo ] ) , we can relate the system correlator to the register correlators via @xmath202 note the normalization @xmath203 , which follows from replacing @xmath204 $ ] by @xmath205 = \langle \mu \rangle_0 + \langle \mu \rangle_1 = j$ ] .
in the degenerate case ( ii ) , the distribution functions separate only for states with different eigenvalues @xmath206 , implying that @xmath207 , while for degenerate eigenvalues with @xmath208 , @xmath209 .
the probabilities @xmath210 then are given by @xmath211 p_\alpha^\mu(n ) p_{\beta}^\nu(m).\end{aligned}\ ] ] on the other hand , the projected density matrix @xmath157 appearing in the system correlator @xmath212 involves the projectors @xmath213 , such that the result ( [ eq : spooreg ] ) remains unchanged .
summarizing , we have seen that the measurement scheme invoking entanglement with quantum memories and their delayed measurement provides us , in the weak- and strong measurement limits , with the same results as obtained via the traditional route using an intermediate von neumann projection .
furthermore , these results have been obtained within a unified description starting from the same initial formula in the form ( [ eq : pabdef ] ) or ( [ eq : pab ] ) .
the above general theoretical considerations are rather non - trivial to implement in a practical situation , as the preparation , entanglement , and measurement of many quantum memories is often a non - trivial task .
the implementation of these ideas is less demanding , though still challenging , when considering specific examples . indeed , quantum memories for delayed measurement are naturally provided in a scattering geometry , where individual scattered particles take the role of flying qubits . in the following , we focus on a specific example in mesoscopic physics , the measurement of a charge correlator of a quantum dot by scattering electrons in a nearby quantum point contact .
we first analyze the situation for a simple two - state system with charges @xmath214 , @xmath215 , and then extend these considerations to arbitrary charge states .
we consider a classic problem @xcite , the charge @xmath10 dynamics of a quantum dot ( qd ) ( attached to leads or coupled to another dot in an isolated double - dot system ) measured by a nearby quantum point contact ( qpc ) , see fig . [
fig : meso_setup ] and ref .
for a recent discussion of this system . here , we want to characterize the dot s dynamics by its two - time charge correlator . in this system
, the measurement is executed by the electrons which are transmitted across / reflected by the qpc with probabilities that depend on the dot s charge state . for the present discussion ,
it is convenient to view the qpc current as a sequence of individual electron pulses ; during recent years , this theoretical idea @xcite has progressed to an experimental reality @xcite , opening the new field of electron quantum optics @xcite .
the quantum memories then can be viewed as flying qubits , individual electron pulses arriving at the qpc at times @xmath4 and @xmath9 that probe the charge state of the qd through the capacitive coupling between the qd and the qpc , see fig .
[ fig : meso_setup ] . quantum dot system ( qd ) measured by a capacitively coupled quantum point contact ( qpc ) : single - electron pulses incident on the qpc from the left ( at @xmath13 ) are either transmitted ( with amplitude @xmath216 ) or reflected ( amplitude @xmath217 ) ; the outgoing lippmann - schwinger wave functions describe flying qubits without own dynamics and serve as quantum memories .
two pulses separated in time by @xmath218 are needed to measure the two - time correlator of the dot s charge .
after scattering at the qpc , the two electrons ( flying qubits ) are entangled with the quantum dot system and carry information on its dynamics .
simultaneous detection of the two scattered electrons ( at @xmath33 ) , e.g. , a distance @xmath219 away with both positions on the right of the qpc , provides information on the two - time charge correlator . , width=264 ] the memory states @xmath220 are the two scattered states where the electron is reflected ( @xmath221 ) or transmitted ( @xmath222 ) , i.e. , the outgoing state is given by @xmath223 and @xmath224 depending on the charge state of the system .
we assume well separated single - electron pulses and an evolution of the scattered waves @xmath225 emanating from the qpc at times @xmath0 that preserves the corresponding system information , in particular , @xmath226 .
this allows us to envision an individual detection of the electrons @xcite . in the final readout ,
the flying qubits are detected on the right or left side of the qpc , telling whether the two electrons have been transmitted ( with probability @xmath227 ) , reflected ( @xmath228 ) or mixed ( @xmath229 and @xmath230 ) .
these probabilities then contain the information about the two - time charge correlator of the qd .
formally , such a final - state analysis corresponds to measuring the ( charge ) operators @xmath231 , with the projectors @xmath232 providing the transmitted ( @xmath233 ) or reflected ( @xmath234 ) components of the @xmath55-th electron .
the eigenvalues @xmath235 depend on the measured charge , e.g. , @xmath236 and @xmath237 if the transmitted charge is measured on the right of the qpc , while @xmath238 and @xmath239 if the reflected charge is measured on the left
. both types of measurements provide us with the same probability distributions ( [ eq : pabdef ] ) or ( [ eq : pab ] ) with @xmath240 . in the following ,
we first analyze the case of a qd with binary charge states @xmath241 and @xmath242 and eigenvalues @xmath243 , @xmath244 , @xmath245 , see eq .
( [ eq : psi ] ) .
second , we generalize the discussion to a qd with multiple charge states as described by the charge operator @xmath246 . while two - state quantum memories ( qubits ) are always sufficient for a complete description when the measurement is weak , for a strong measurement , a qd with multiple charge states will require the use of qubit registers , i.e. , finite trains of electron pulses . for the case of a weak measurement , we can make immediate use of the general results eqs .
( [ eq : pabweak ] ) , ( [ eq : rsoof ] ) , and ( [ eq : isoof ] ) by replacing @xmath247 and choosing the values @xmath248 equal to @xmath249 and @xmath250 .
it remains to determine the detector response functions ( [ eq : isdet ] ) and ( [ eq : rsdet ] ) .
we consider a linear system detector coupling of the form @xmath251 with @xmath252 the charge on the qpc .
furthermore , we parametrize the scattering matrices for the system in states @xmath241 and @xmath242 by @xmath253 , @xmath254 , @xmath255 , @xmath256 , with small corrections @xmath257 , @xmath258 , and @xmath259 . in order to find the detector response functions ( [ eq : rsdet ] ) and ( [ eq : isdet ] ) we replace @xmath260 by @xmath261 and determine the integral @xmath262 ( we drop the memory index @xmath263 as the electrons are scattered by the same qpc ) .
to lowest order in @xmath264 , this can be written in the form @xmath265 with @xmath266 and @xmath267 describing the dynamics of the detector in the heisenberg representation in the absence and presence of a charge on the dot , respectively .
furthermore , we have used that @xmath268 with @xmath269 and @xmath270 in the interaction representation .
the initial memory states @xmath271 evolve under @xmath272 and @xmath273 according to @xmath274 and hence @xmath275 , where we have used that @xmath276 .
expanding @xmath277 in @xmath257 , @xmath278 , and @xmath279 , we find that @xmath280 and taking real and imaginary parts , we arrive at the final results @xmath281 using these detector response functions in the general expressions ( [ eq : rsoof ] ) and ( [ eq : isoof ] ) one easily arrives at the system correlators ( we choose @xmath282 , @xmath283 , and @xmath284 ) , @xmath285 alternatively , the qpc can be tuned to deliver the individual system correlators @xmath286 or @xmath287 .
indeed , using a detector with high transmission , e.g. , a qpc with energetic ( @xmath288 ) single - electron pulses @xmath289 , @xmath290 the qpc barrier , one easily finds that ( although @xmath291 and therefore @xmath292 measures @xmath287 , while @xmath293 measures @xmath294 .
when the detector predominantly reflects particles , e.g. , for low energy single electron pulses with @xmath295 , the situation is reverse , @xmath296 ( @xmath286 ) is measured by @xmath297 ( @xmath292 ) , see the appendix for further details on the qpc detector response . for a quantum dot with multiple charge states we have to require that the expansion of the scattering amplitudes @xmath298 and @xmath299 of the qpc detector scale linearly in the charge @xmath300 of the dot , i.e. , @xmath301 , @xmath302 , and @xmath303 .
a straightforward calculation then shows that the results ( [ eq : det_r ] ) and ( [ eq : det_i ] ) for the detector response functions as well as the final results ( [ eq : rs_c ] ) and ( [ eq : is_c ] ) remain unchanged .
when performing a strong measurement of a quantum dot with a binary charge it is sufficient to invoke individual electron pulses as quantum memories . for a strong dot qpc coupling ,
we require a one - to - one relation between the presence of a charge on the dot and the outcome of the measurement , i.e. , @xmath304 and @xmath305 and @xmath306 .
this is achieved by tuning the qpc such as to generate a unique scattering outcome with @xmath307 , @xmath308 and @xmath309 , @xmath310 , i.e. , the presence of a charge @xmath244 on the dot reflects the qpc electron back to the left . in this case , it is the reflection probability @xmath311 that directly traces the charge @xmath10 and according to ( [ eq : pab ] ) , we have to determine the expression @xmath312 with @xmath313 , we obtain the simple result @xmath314 ( see also ( [ eq : panbm ] ) with @xmath315 and @xmath316 ) and since @xmath317 , we find the projected correlator ( see eq .
( [ eq : spoo ] ) ) @xmath318 similarly , the probability to find no charge on the dot in either of the two measurements is @xmath319 , while the mixed results are @xmath320 and @xmath321 .
the strong measurement of the charge correlator for a multi - charge quantum dot quite naturally involves trains of electron pulses @xcite , with the number @xmath162 of electrons in each train sufficiently large to distinguish the different charge eigenvalues @xmath300 of the dot .
the separation @xmath78 between electron pulses within a train has to be sufficiently long in order to allow for their separate detection ( i.e. , counting ) , while the train duration @xmath322 must remain small on the scale @xmath323 of the dot s dynamics .
when going over from qubit registers to electron trains scattered at the qpc we replace the ` magnetization ' ( [ eq : mn ] ) by the disbalance between reflected and transmitted electrons , @xmath324 = j[r - t + 2 q_n \delta r],\end{aligned}\ ] ] where we have assumed a linear qpc characteristic with a reflection probability scaling linearly with the dot s charge @xmath300 , @xmath325 .
operating the qpc at the symmetry point @xmath326 , we can determine the combination @xmath327 and relate this quantity to the projected charge correlator @xmath328 , @xmath329 operating the qpc away from the symmetric point , one has to determine the weighted sum @xmath330 instead and divide by @xmath331 to arrive at the projected correlator @xmath332 .
above we have assumed an instantaneous ( within a short time @xmath78 ) entanglement between the system and the memory states , requiring that both the width @xmath333 of the wave - packet and the scattering time @xmath334 at the qpc satisfy @xmath335 , where @xmath336 denotes the characteristic time scale of the system . here , we allow for a spread in time of the detector s electron wave function and drop the condition @xmath337 , i.e.
, we assume that @xmath338 while the scattering event itself remains fast , @xmath339 .
in general terms , this corresponds to a measurement which probes the system sharply ( @xmath340 ) during some finite time ( @xmath341 ; longer measurement times @xmath342 do not provide meaningful results ) .
let us suppose that the @xmath55-th wave - packet incident on the qpc around @xmath0 is described by the wave function @xmath343 which is normalized ( @xmath344 ) and peaked at the time @xmath0 .
assuming instantaneous scattering , we obtain the final state @xmath345 with @xmath346 and @xmath347 and the outgoing memory states @xmath348 that have been scattered at the qpc at time @xmath349 . making use of @xmath350 , we obtain the smeared probabilities @xmath351 with @xmath28 given by eq . .
the finite width wave - packets enter as integration kernels . while for sharp wave - packets
, the entanglement between system and memories ( i.e. , the ` measurement ' ) takes place at times @xmath4 and @xmath9 , for broader wave - packets , the entanglement arises within a finite time @xmath333 around @xmath4 and @xmath9 with distribution functions @xmath352 and @xmath353 .
as a consequence , for the case of strong coupling where the entanglement gives rise to projective measurements , the times of projection are not fixed but distributed with the distribution functions above .
note that the result is only valid for negligible scattering time @xmath354 in the qpc .
if the scattering time @xmath354 is finite compared to the system time @xmath336 , the effect of interaction can not be accounted for by the scattering matrices @xmath355 of the qpc depending on the system state , but the interaction during the scattering has to be treated in more detail .
the limit of @xmath356 considered above has also implications for the resulting back - action : while the back - action for @xmath356 only consists of dephasing , i.e. , a suppression of off - diagonal elements of the system s density matrix , for finite @xmath354 , the back - action of the measurement on the system goes beyond pure dephasing and alters the system s dynamics .
the study of higher - order correlators is straightforward , e.g. , to measure a third - order charge correlator , we send three electrons scattering from the qpc at times @xmath357 and obtain the probabilities @xmath358 describing electrons transmitted across ( @xmath359 ) or reflected from ( @xmath360 ) the qpc . for weak coupling , its irreducible part can be recast in the form @xmath361 with @xmath362 , the detector responses @xmath363 and @xmath364 , and the third - order correlators @xmath365_{\sigma'}]_{\sigma}\rrangle\end{aligned}\ ] ] with the constants @xmath366 and @xmath367 and @xmath99_-=[\cdot,\cdot]$ ] resp .
@xmath99_+=\{\cdot,\cdot\}$ ] ( note that ( anti-)symmetrized charges in @xmath368 ( encoded in @xmath369 ) relate to opposite detector response functions ( encoded by @xmath370 ) . this result agrees with the one in ref .
obtained with the help of the von neumann projection postulate and shows that only _
keldysh time - ordered charge correlators _ are measurable . our general concept of deferred repeated measurements has been formulated with quantum memories , e.g. , qubits , qubit registers , and qudits , see sec.[sec : uni ] . in our application of these general considerations , the measurement of a charge correlator with the help of a quantum point contact ( sec .
[ sec : charge ] ) , the quantum memories have been replaced by scattered electrons ( flying qubits ) . it then is natural to seek for an experimental implementation , where the final measurement of the quantum memories , i.e. , the scattered electrons , can be cast into a measurement of currents and noise rather than an individual detection of qubit states . such an implementation is proposed below . by now
, several experiments have demonstrated the controlled generation of individual electron pulses @xcite and even the detection of such pulses via qubit detectors seems to be in reach @xcite .
a setup particularly well suited within the current context is the one of fletcher _
@xcite , involving a single electron pump and a time - correlated detector setup in a quantum hall device , see fig .
[ fig : exp ] for an illustration .
pairs of electron pulses of width @xmath371 ps are injected at a typical rate @xmath372 and are well suited to probe the dynamics of a dot with a system time @xmath373 ns .
the detector involves a dynamically switchable barrier ( dsb in fig .
[ fig : exp ] ) that can be switched on or off with a nanosecond precision in time .
schematic illustration of a quantum hall device with a single electron pump ( left ) , the qpc
qd setup ( middle ) and a detector arrangement ( right ) , inspired by refs .. pairs of electrons are injected by the single electron pump with a time delay @xmath374 and scattered by the quantum point contact with scattering coefficients @xmath217 and @xmath216 depending on the charge state of the quantum dot .
the dynamically switchable barrier@xcite ( dsb ) separates the first and second electron ( flying qubits @xmath142 and @xmath375 ) into arms 1 and 2 . closing the static switchable barrier ( ssb ) and measuring the currents @xmath376 and @xmath377 provides the anti - symmetrized charge - charge correlator @xmath378 .
opening the static switchable barrier ( ssb ) , having the two particle streams along the trajectories 1 and 2 interfere at the 50/50 beam splitter , and measuring the current noise @xmath379 as a function of the tunable delay @xmath380 in the detector 3 provides the symmetrized correlator @xmath381.,width=302 ] we propose an experiment implementing a weak measurement of a charge correlator along the lines of sec .
[ sec : charge_wc ] , where electron pairs with a tunable delay time in the range @xmath382 nanoseconds are injected by the single electron pump ( with injection currents in the @xmath383 pa regime ) and scattered at the qpc probing the quantum dot qd .
the transmitted electrons arrive at the detector dsb after a delay time @xmath384 .
the detector barrier dsb is switched on at a time @xmath385 in order to let the first electron pass along the path @xmath386 and deflect the second electron along @xmath387 , see figure [ fig : exp ] . with the static barrier ( ssb )
closed , the measured currents @xmath388 and @xmath389 resulting from the transmission of the first and second electron , respectively , can be related to the probabilities @xmath28 via @xmath390 the two sums @xmath391 and @xmath392 in the above equations are fundamentally different in a subtle way : summing over the second particle @xmath30 produces the trivial probability latexmath:[$p_\alpha^{\scriptscriptstyle ( 1 ) } = \langle \psi_f ( independent ) transmission ( @xmath282 ) or reflection ( @xmath234 ) of the first electron by the qpc , @xmath394 .
hence , the first equation ( [ eq : i1 ] ) reduces to @xmath395 and thus contains only information about the mean charge on the dot . on the contrary , summing over the first particle @xmath29 _ does not _ generate @xmath396 as the interaction of the first electron at the earlier time @xmath4 introduces nontrivial correlations , see eq .
( [ eq : pab ] ) .
this becomes apparent when going over to irreducible probabilities @xmath397 and using the normalization @xmath398 .
then , the second equation ( [ eq : i2 ] ) becomes @xmath399 which includes information about the dot s charge correlator . assuming a time - independent mean charge on the dot
, we have @xmath400 and using the general result ( [ eq : pabweak ] ) , the measured currents are easily transformed to provide the antisymmetrized correlator @xmath401 note that the evaluation of the irreducible probability @xmath402 with the help of ( [ eq : pabweak ] ) indeed provides a vanishing result , @xmath403 , explicitly demonstrating that the sum @xmath404 contains no correlations . in order to find the symmetric correlator @xmath405 one has to measure a time correlator on the transmitted channel .
this can be conveniently done with the help of a hong - ou - mandel type splitter as implemented in the experiment of bocquillon _ et al .
_ @xcite and sketched in fig .
[ fig : exp ] . in this experiment ,
the dynamically switchable barrier ( dsb ) again splits the two electrons in each pair to propagate along the paths 1 and 2 @xmath406 , respectively . the static barrier ( ssb )
is left open , such that the two particles interfere in the splitter .
measuring the current noise @xmath407 in channel 3 as a function of mutual delay @xmath380 ( tuned via an additional gate in loop 1 , see figure ) then provides all information needed to construct @xmath405 .
we can calculate the evolution of the state through the hong - ou - mandel setup using the wave function @xmath408 , i.e. , the individual components are not normalized and only @xmath409 . due to the orthogonality of these four components , the particle numbers @xmath410 , @xmath411 , emerging from our hong - ou - mandel splitter
can be analyzed term by term .
in particular , the particle number fluctuations @xmath412 in channel 3 involve single - particle and two - particle contributions , @xmath413 and @xmath414 , hence , @xmath415 and the same result holds true for the @xmath416 scattering component .
while there is no contribution from @xmath417 , the one originating from @xmath418 depends on the time delay @xmath380 .
let @xmath419 and @xmath420 denote the two electron wave - packets propagating along the incoming paths @xmath386 and @xmath387 of the hom - interferometer . as shown in ref .
, we can relate the number fluctuation in the component @xmath418 to the overlap of wavefunctions as @xmath421 where @xmath422 . for
a large time delay between the first and second electron , these do not interfere , @xmath423 , such that the fluctuations are just the double of ( [ eq : dn32 ] ) , @xmath424 compensating the original time delay @xmath425 ( such that the two electrons arrive simultaneously at the splitter ) , we have @xmath426 since the pauli exclusion forces the two particles to propagate to different channels .
the final result @xmath427 and @xmath227 , @xmath428 when the delay between electron pulses is not compensated and @xmath429 when the time delay is properly compensated , @xmath425 . in a next step , we express the charge fluctuations in channel @xmath430 through the irreducible current - current correlator,@xcite @xmath431 where the time - window @xmath432 is centered around the arrival time of the wave - packets at the detector @xmath430 and has a width of the order of @xmath433 , with @xmath434 the rate of pair injection by the pump .
the particle number fluctuations then can be expressed through the low - frequency current noise @xmath435 , @xmath436 with @xmath437 . neglecting the frequency dependence of the noise at small @xmath438 , @xmath439
, we obtain @xmath440 and hence @xmath441 using eq .
, we can derive an alternative ( more symmetrized ) version of eq .
which involves just the combinations @xmath442 and @xmath443 , @xmath444 defining the detector parameters @xmath445 and going over to irreducible probabilities ( with @xmath446 independent of time ) , we obtain the final result @xmath447 note that in this setup , the final measurement of qubit memories does not require fast or time resolved detection schemes , but merely relies on the measurement of average currents and low - frequency noise .
this is due to the fact , that all timing tasks are realized by the properly time - delayed electron pulses in the incoming channel and the dynamically switchable gate ( dsb ) which separates the electron pairs ; both elements have been realized in an experiment@xcite .
hence , the new measurement scheme , combined with novel elements from electron quantum optics , allows to shift the ( difficult ) timing issues in the measurement of a time - correlator from the detector to the source . a _ strong coupling _ between the quantum dot and the quantum point contact provides us with a projected correlator . for the simplest case of a binary charge on the dot with values @xmath448 ,
the projected charge correlator is given by eq .
, @xmath449 . making use of the normalization @xmath450 and the result eq . ( [ eq : prts ] ) , we find that @xmath451
in conclusion , we have applied the principle of deferred measurement to the problem of repeated measurement and have derived physical expressions for the two- and multi - time correlators .
the measurement involves the inclusion of quantum memories which are entangled with the system at specific times @xmath0 where the system observable is to be probed .
the expanded system plus memories undergoes a unitary evolution until the very end , where the result is extracted via application of born s rule to the memories .
the measured probabilities @xmath28 ( see eq .
( [ eq : pab ] ) ) or memory correlators @xmath452 ( see eq .
( [ eq : sab ] ) ) then can be combined to extract the desired system correlators .
the limits of weak and strong measurements provide the standard ( anti- ) symmetrized and projected time correlators previously obtained by invoking the ( non - unitary ) von neumann projection .
the general results have been illustrated by using qubits and qubit registers as quantum memories .
our analysis sheds new light on the problem of repeated measurement and illustrates the usefulness of qubits as sensitive measurement devices .
although our paper s main results are rather on the conceptual side , one could imagine an implementation of such a deferred measurement in an experiment .
a system that naturally lends itself for a realization of these ideas is the classic mesoscopic setup which probes the charge of a quantum dot through a quantum point contact .
the individual scattered electrons in the qpc can be understood as flying qubits which are either transmitted or reflected , with amplitudes depending on the charge state of the quantum dot .
in particular , the qubit register required in the strong measurement of a dot with a multi - valued charge is easily implemented in terms of finite trains of electrons .
we have applied our formalism to this situation and derived the corresponding expressions for a weak and strong measurement .
the experimental implementation of these ideas requires a system control that can only be met with a modern quantum engineering approach .
recent developments in electron quantum optics provide controlled single electron pulses and allow for their time resolved manipulation / detection on a sub - nanosecond time scale @xcite . using the setup of ref . as a base and augmenting it with a hong - ou - mandel type analyzer @xcite
, we propose to include a quantum point contact and a quantum dot in order to realize the principle of repeated measurement by a deferred measurement of quantum memories . as a final remark
, one may appreciate the relation of the deferred measurement principle to everett s idea of a multiverse @xcite . rather than applying a projection after the first measurement and pursuing a single further evolution ( of the system = ` universe ' )
, the principle of a deferred measurement involving the system s entanglement with a quantum memory enhances the overall dimensionality , e.g. , for a qubit memory the dimensionality is doubled ( with two ` universes ' evolving in parallel ) .
it is then only the final measurement which determines which evolution ( i.e. , which ` universe ' ) actually has been realized .
we thank renato renner for discussions and acknowledge financial support from the swiss national science foundation through the national center of competence in research on quantum science and technology ( qsit ) , the pauli center for theoretical studies at eth zurich , and the rfbr grant no.14 - 02 - 01287 .
a quick overview is provided by the example of a _ @xmath453-function scatterer : _ expressing the strength of the scatterer @xmath454 for the two charge states by @xmath455 and @xmath456 , an incoming state with wave vector @xmath457 is transmitted with amplitude @xmath458 , @xmath459 .
expanding the transmission @xmath460 and phase @xmath461 , we find the modifications @xmath257 and @xmath462 ( for a symmetric scatterer ) in the scattering characteristic of the qpc upon charging the dot @xmath463 in the limit of a large incoming energy , i.e. , @xmath464 , we find @xmath465 and @xmath466 and hence @xmath467 ; with @xmath468 and @xmath469 , we have @xmath470 and therefore @xmath471 . for a small incoming energy @xmath472 , we obtain @xmath473 and @xmath474 and using @xmath475 and @xmath476 , we find @xmath477 . when @xmath478 , all response functions are of the same order . alternatively
, we can consider a _
single electron transistor ( set ) _ with the level position @xmath479 affected by the capacitive coupling and depending on the dot s charge state @xmath18 , i.e. , @xmath480 and @xmath481 .
the transmission coefficient is given by @xmath482 , where @xmath483 is the level width .
again expanding @xmath484 and @xmath485 for small @xmath486 , we find @xmath487 ^ 2},\\ \delta\theta = \delta \chi & = -\frac{\gamma \delta k_{\rm res } } { ( k - k_{\rm res})^2+\gamma^2}.\end{aligned}\ ] ] for incoming electrons on resonance with the level , i.e. , @xmath488 , we obtain @xmath489 and @xmath490 , such that @xmath467 and using @xmath468 and @xmath491 , we find that @xmath492 and @xmath493 , such that @xmath467 and using @xmath494 and @xmath495 , we find @xmath496 , all response functions are of the same order . a more realistic description for the _ quantum point contact ( qpc ) _ is achieved by considering a parabolic scattering potential @xmath497 where the offset @xmath498 is the qpc barrier height when the dot is in the charge state @xmath18 . here , we assume a quasi - classical description and consider the two limits of electrons with energy @xmath499 resp .
@xmath500 , see fig .
[ fig : qpc_qc ] .
using the kemble formula @xcite , we obtain the transmission @xmath501)$ ] , where we have chosen @xmath502 . for weak coupling @xmath503
, we obtain the shift ( we define the energy scale @xmath504 ) @xmath505 which is suppressed exponentially for @xmath506 and @xmath295 due to an exponentially small reflection or transmission .
the change in phase at large energies @xmath506 is determined by the transmission phase accumulated in the region @xmath507 $ ] ; within a quasi - classical description , this is given by ( @xmath508 ) @xmath509.\end{aligned}\ ] ] expanding this result for small @xmath510 , we obtain the change in phase @xmath511 , @xmath512 given the exponential suppression of @xmath257 at large energies @xmath506 , we find that @xmath513 and for large transmission we have @xmath514 . in the opposite regime of small energies
@xmath295 , we determine the change in phase ( within quasi - classics ) from the phase of the reflection amplitude , @xmath515 where the reversal point @xmath516 is characterized by @xmath517 . to leading order in @xmath510
we find that @xmath518 once more , it follows that @xmath519 due to the exponential suppression of @xmath520 and the response functions respect the order latexmath:[$|\mathcal{r}s_\mathrm{det , r}| \gg |\mathcal{i}s_\mathrm{det , r / t}|\gg functions are of similar magnitude . summarizing , we find that a scatterer with large transmission is characterized by the response functions satisfying latexmath:[$|\mathcal{r}s_\mathrm{det , t}|\gg |\mathcal{i}s_\mathrm{det , r / t}|\gg m. field , c.g .
smith , m. pepper , d.a .
ritchie , j.e.f.frost , g.a.c .
jones , and d.g .
hasko phys .
lett . * 70 * , 1311 ( 1993 ) ; l.m.k .
vandersypen , j.m .
elzerman , r.n .
schouten , l.h .
willems van beveren , r. hanson , and l.p .
kouwenhoven , appl .
lett . , 4394 ( 2004 ) ; s. gustavsson , r. leturcq , b. simovi c , r. schleser , t.ihn , p. studerus , k. ensslin , d.c .
driscoll , and a.c .
gossard , phys .
96 , 076605 ( 2006 ) .
g. feve , a. mahe , j .-
berroir , t. kontos , b. placais , d.c .
glattli , a. cavanna , b. etienne , and y. jin , science * 316 * , 1169 ( 2007 ) ; j. dubois , t. jullien , f. portier , p. roche , a. cavanna , y. jin , w. wegscheider , p. roulleau , and d.c .
glattli , nature * 502 * , 659 ( 2013 ) . | repeated measurements as typically occurring in two- or multi - time correlators rely on von neumann s projection postulate , telling how to restart the system after an intermediate measurement .
we invoke the principle of deferred measurement to describe an alternative procedure where co - evolving quantum memories extract system information through entanglement , combined with a final readout of the memories described by born s rule .
the new approach to repeated quantum measurements respects the unitary evolution of quantum mechanics during intermediate times , unifies the treatment of strong and weak measurements , and reproduces the projected and ( anti- ) symmetrized correlators in the two limits . as an illustration
, we apply our formalism to the calculation of the electron charge correlator in a mesoscopic physics setting , where single electron pulses assume the role of flying memory qubits .
we propose an experimental setup which reduces the measurement of the time correlator to the measurement of currents and noise , exploiting the ( pulsed ) injection of electrons to cope with the challenge of performing short - time measurements . |
spin effects are known to provide sensitive tests for the underlying particle theories in many cases .
one may recall the spin asymmetries in pion - nucleon scattering which can not be described in the simplest pole approximation of the regge theory and require to take into account cuts . in the case of qcd
the so - called emc spin crisis was a signal of a breakdown of the naive parton model and of the importance of the incorporation of such a subtle field theory effect as the axial anomaly @xcite .
a less popular , but also extremely interesting example is the problem of single transverse spin asymmetries in reactions mediated by strong interactions .
as these effects require a mass parameter and an imaginary part in the scattering amplitude , they are proportional to @xmath0 in the naive perturbative approach . however
, a careful application of qcd factorization at twist-3 level removes both of these small parameters @xcite .
the collective gluon field of the hadron is shifting the current quark mass @xmath1 to a mass parameter of the order of the hadron mass and , simultaneously , provides the imaginary part .
technically , the latter appears when the pole of the quark ( or gluon @xcite ) propagator is integrated over the light - cone momentum fraction .
a similar imaginary part is present in the qcd higher twist contributions to dilepton production in pion - nucleon scattering , due to the integration over the light - cone momentum fraction of the quark in the pion .
this effect leads in the case of unpolarized nucleon targets to a significant contribution to the dilepton angular distribution @xcite , which is sensitive to different anstze for the pion distribution function .
it is interesting to look for single spin asymmetries for which the imaginary part does not constitute a ( large ) correction , but the whole contribution .
the natural candidate is the spin asymmetry in the production of dileptons from the scattering of pions on a _ longitudinally polarized _ nucleon .
this spin asymmetry has been thoroughly studied in perturbative qcd @xcite .
the analogous mesonic processes have been discussed in @xcite . in the present work we study this single spin asymmetry , taking into account pion bound state effects .
we will consider the inclusive production of muon pairs from the scattering of a pion on a polarized proton ( the drell - yan process ) , @xmath2 where @xmath3 denotes the degree of longitudinal polarization of the proton , and all momenta refer to the overall c.m . system .
it is convenient to write the differential cross section for ( [ react ] ) in terms of a hadronic and a leptonic tensor , @xmath4 where @xmath5 is the hadronic c.m . energy .
the leptonic tensor has the well - known form @xmath6 note that @xmath7 is symmetric , since we only consider the exchange of a virtual photon ( not a @xmath8 boson ) and since we sum over the polarizations of @xmath9 .
the hadronic tensor @xmath10 may be written as @xmath11 only the symmetric part of @xmath10 is relevant after contraction with @xmath12 of eq .
( [ leptonictensor ] ) .
therefore @xmath10 in the following is implicitly assumed to be symmetrized .
the angular distribution of @xmath13 in ( [ react ] ) may be parametrized as @xmath14 here , @xmath15 and @xmath16 are angles defined in the muon pair rest frame , and @xmath17 are angle independent coefficients .
they depend on @xmath18 ( the virtuality of the photon ) , @xmath19 ( the squared transverse momentum of the photon in the hadronic c.m.s . ) , and on @xmath20 ( the longitudinal momentum fraction of the photon ) .
we take @xmath21 for a photon moving forward with respect to the pion in the hadronic c.m .
the normalization of the cross section can be determined from @xmath22 the normalization @xmath23 and the angular coefficients are related to the helicity amplitudes of the hadronic tensor : @xmath24 where the polarization vectors @xmath25 and @xmath26 are determined by specifying the coordinate axes @xmath27 in the muon rest frame .
the coefficients @xmath28 and @xmath29 are nonzero only in the polarized drell - yan process and are induced by absorptive parts in the scattering amplitude @xmath30 . in the parton model ,
the leading contributions to @xmath28 and @xmath29 come from the interference of born diagrams with the absorptive parts of one - loop diagrams @xcite .
the relation of the single spin asymmetry @xmath31 to the angular coefficients is simply given by @xmath32 for the unpolarized drell - yan process the parton model predictions @xcite fail to describe the data on @xmath33 and @xmath34 @xcite .
in contrast , taking into account pion bound state effects yields reasonable fits to the data @xcite .
an interesting feature of this bound state model is a nonvanishing absorptive part in _ leading _ order which is proportional to the pion distribution amplitude @xmath35 .
this property directly leads to nonvanishing @xmath28 and @xmath29 which are strongly sensitive to @xmath35 . in the following
we will shortly describe the salient features of the bound state model for the pion.further details may be found , e.g. , in @xcite , @xcite . the bound state problem may be treated perturbatively if the momentum fraction of the quark from the pion is large , which is the case for large @xmath36 . the dominant contribution to
the drell - yan process comes from the annihilation of a quark with an antiquark @xcite . in the region of large @xmath36 ,
the diagrams of fig .
then give the leading contributions .
we will explain the formalism for the annihilation of a @xmath37 quark from the proton with the @xmath38 quark from the pion ; the other possible annihilation process of a @xmath39 quark from the proton with the @xmath40 quark from the pion is treated in complete analogy and its contribution will later be added incoherently . in diagram 1a
we see that the @xmath38 quark propagator is far off - shell , @xmath41 , if the light cone momentum fraction of the quark from the pion @xmath42 is close to 1 @xcite .
thus , the pion can be resolved by a single hard gluon exchange @xcite . the second diagram , fig .
1b , is required by gauge invariance .
the hadronic tensor @xmath10 is obtained by a convolution of the partonic tensor @xmath43 with the corresponding parton distribution functions for the polarized proton , @xmath44 . ] .
the partonic tensor @xmath43 is computed from the product @xmath45 .
@xmath46 is the amplitude for the reaction @xmath47 where the arrow indicates the polarization of the parton .
this amplitude @xmath46 is obtained by convoluting the partonic amplitude @xmath48 with the pion distribution amplitude @xmath49 @xcite , @xmath50 where @xmath51 is the cutoff for the integration over soft momenta in the definition of @xmath35 , @xmath52 is the pion decay constant , and @xmath53 is the number of colors . a detailed discussion of the scale setting problem for exclusive processes is given in @xcite .
using the commensurate scale relations in the calculation of ( [ convdist ] ) would allow to eliminate ambiguities caused by the factorization and renormalization .
however , this does not change the qualitative features of our predictions and thus it will not be discussed in this paper .
the imaginary part in this model arises because the internal quark line of fig .
1b can go on - shell .
after integrating the hard scattering amplitude @xmath54 according to equation ( [ convdist ] ) , we are left with a regular amplitude @xmath46 which has an imaginary part proportional to the pion distribution amplitude at the point where the quark propagator gets singular .
we are now ready to present our analytic results .
we relegate the full result for the hadronic tensor to the appendix , and confine ourselves here to a discussion of the angular coefficients @xmath28 and @xmath29 . to present our results for the angular distribution we choose the gottfried - jackson frame where the @xmath55 axis is taken to be the pion direction in the muon rest frame and the @xmath56 axis lies in the @xmath57 plane such that the proton momentum has a negative @xmath58 component . from eqs .
( [ angularcoef ] ) , ( [ hadronictensordecomposition ] ) , and ( [ structurfunction ] ) we get the angular coefficients @xmath59 + \rho^2\tilde{x}^2 f^2(4+\rho^2 ) } \nonumber\\ & & \qquad \times \ { { 4\over 9 } \delta q_u^v(x_p ) + { 4\over 9 } \delta q_u^s(x_p ) + { 1\over 9 } \delta q_d^s(x_p)\over { 4\over 9 } q_u^v(x_p ) + { 4\over 9 } q_u^s(x_p ) + { 1\over 9 } q_d^s(x_p ) } , \nonumber\\ \bar{\nu } & = & 2 \rho \bar{\mu}.\end{aligned}\ ] ] here , @xmath60 and @xmath61 is a function of @xmath62 and @xmath63 @xmath64 furthermore , @xmath65 denote integrals over the pion distribution amplitude .
finally , @xmath66 and @xmath67 are the unpolarized and polarized valence and sea quark distribution functions evaluated at the point @xmath68 .
the results for @xmath69 within the pion bound state model are given elsewhere @xcite .
the numerical results for the angular coefficient functions @xmath28 and @xmath29 , eq . ( [ mubarnubar ] ) , depend on @xmath61 , @xmath63 , the pion distribution amplitude and on the ratio of the polarized and unpolarized quark distribution functions of the proton . whereas the unpolarized quark distributions are known quite well , the polarized ones still contain large uncertainties . for our results
we choose the parametrizations given in @xcite at @xmath70 . in fig . 2 we present our results for several distribution amplitudes @xmath35 that are quite different in shape
2a shows our choices for @xmath35 : the solid line represents the two - humped function @xcite which gave a good fit to the data on @xmath33 and @xmath34 @xcite .
the effective evolution parameter is set to @xmath71 . since the predictions are also very sensitive to the behavior of the distribution amplitude in the endpoint region , we choose two different parametrizations @xcite for the convex distribution amplitude @xmath72 . the dashed line in fig .
2a represents the asymptotical amplitude , i.e. , @xmath73 , and the dotted line shows a narrow distribution amplitude with @xmath74 . in fig .
2b the moment @xmath75 is plotted in arbitrary units versus @xmath36 for the values @xmath76 , @xmath77 gev and @xmath78 gev .
it can be seen that different shapes of the distribution amplitude could be distinguished by a measurement of this quantity , which is proportional to the pion distribution amplitude evaluated at @xmath79 ( cf . the numerator of ( [ mubarnubar ] ) ) .
it is also demonstrated that narrow distribution amplitudes give a moment which vanishes in the large @xmath36 region . in figs . 2c and 2d the angular coefficient @xmath80
is plotted versus @xmath36 for two different choices of @xmath81 and the same values for @xmath5 and @xmath82 as in fig .
2b . in fig .
2c @xmath81 was set again to @xmath83 gev , which gives the moderate value @xmath84 .
it is demonstrated that the two - humped form for @xmath35 induces a minimum at @xmath85 which , however , vanishes for the smaller value @xmath86 used in fig .
2d . for very narrow distribution amplitudes @xmath28
is strongly suppressed . since @xmath87 ( cf .
( [ mubarnubar ] ) ) in the bound state model , we do not show separate plots for this quantity .
note that the effects of a variation of the pion distribution amplitude in the spin - dependent and spin - averaged cross sections
i.e. in the numerator and denominator of ( [ mubarnubar ] ) partially compensate each other . as a consequence
, the result for the two - humped distribution amplitude appears in between the results for the two convex distributions .
the spin - dependent part of the cross section is a more sensitive `` partonometer '' , than the dimensionless coefficients @xmath88 .
although the former quantity suffers from larger uncertainties due to higher order corrections , , one may expect only a small uncertainty for the determination of the shape of pion distribution amplitude , if the @xmath36 dependence of the corrections is weak .
although the naive parton model is at variance with the unpolarized angular distribution @xcite , we would like to present for completeness and comparison also the predictions of this model for @xmath28 and @xmath29 .
we took the analytic results from @xcite and again used the parton distribution functions of @xcite . in the range @xmath89 the parton model yields coefficients @xmath80 and @xmath90 which are positive and to a very good approximation independent of @xmath36 .
for the same values used in fig2b and 2c , i.e. @xmath91 and @xmath92 we get a value of @xmath93 .
this means that bound state effects are roughly of the same magnitude but opposite in sign compared to the parton model at moderate values of @xmath63 . for the smaller value @xmath94
we find @xmath95 , again independent of @xmath36 and positive .
at such a small value of @xmath63 the bound state model predicts that @xmath28 and @xmath29 are concentrated at large @xmath36 for not too narrow distribution amplitudes . in the kinematical range
considered , the relation between @xmath29 and @xmath28 calculated within the parton model takes a similar form as in the bound state model , namely , @xmath96 .
in conclusion we have shown that experiments on dimuon production from the scattering of pions on polarized targets will give detailed information both on the shape and on the endpoint behavior of the pion distribution amplitude .
we would like to thank s. j. brodsky , s. v. mikhailov , j. soffer and a. v. radyushkin for valuable discussions .
b. and o. v. t. would like to thank also the naturwissenschaftlich - theoretisches zentrum and the graduiertenkolleg `` quantenfeldtheorie '' at leipzig university for the kind hospitality during their stay , where substantial parts of this work were done .
was financially supported by the deutsche forschungsgemeinschaft .
o.v.t . was supported in part by russian foundation for fundamental researches , international science foundation and intas .
using the kinematic decomposition of the ( symmetric part of the ) hadronic tensor @xmath97 where @xmath98 and @xmath99 , we obtain from the bound state model the following predictions for the structure functions : @xmath100f + ( 1-\tilde{x}){\rm re\ } i(\tilde{x})\right)^2 + \pi^2 \varphi(\tilde{x})^2\right\ } f\left(x_p\right ) \nonumber\\ w_2 & = & -n \tilde{x}^2 f\left(x_p\right ) \nonumber \\ \ & & \left\{(f+(1-\tilde{x } ) { \rm re\ } i(\tilde{x } ) ) \left[(1-\tilde{x}(1+\rho^2)^2 ) f + ( 1-\tilde{x } ) { \rm re\ } i(\tilde{x}))\right]+ \pi^2 \varphi(\tilde{x})^2\right\ } \nonumber\\ w_3 & = & { n x_p^2 \tilde{x}\over 1-\tilde{x } } \left[1-\tilde{x}(1+\rho^2)\right]^2 f ( f+{\rm re\ } i(\tilde{x } ) ) f\left(x_p\right ) \nonumber\\ w_4 & = & - { n x_p \tilde{x}^2 \over 1-\tilde{x } } \rho^2 \left[1-\tilde{x}(1+\rho^2)\right ] f(f+(1-\tilde{x } ) { \rm re\ } i(\tilde{x } ) ) f\left(x_p\right ) \nonumber\\ w_5 & = & -{n x_p \tilde{x}^3 \over 1-\tilde{x } } ( 1+\rho^2 ) \left[1-\tilde{x}(1+\rho^2)\right ] \pi\varphi(\tilde{x } ) f \delta
f\left(x_p\right ) \nonumber\\ w_6 & = & { n x_p^2 \tilde{x}^2 \over ( 1-\tilde{x})^2 } \left[1-\tilde{x}(1+\rho^2)\right]^2 \pi\varphi(\tilde{x } ) f \delta
f\left(x_p\right).\end{aligned}\ ] ] here , @xmath101 is a normalization factor , and @xmath102 are given by the unpolarized and polarized valence and sea quark distribution functions .
the variable @xmath61 and @xmath103 as well as the integrals @xmath104 and @xmath105 containing the pion distribution amplitude are defined below eq .
( [ mubarnubar ] ) .
m. anselmino , a.v .
efremov , and e. leader , phys . rep .
* 261 * ( 1995 ) 1 .
efremov and o.v .
teryaev , phys .
* b 150 * ( 1985 ) 383 . j. qiu and g. sterman , nucl . phys .
* b 378 * ( 1992 ) 52 .
a. brandenburg , s.j .
brodsky , v.v .
khoze , and d. mller , phys .
* 73 * ( 1994 ) 939 .
b. pire and j.p .
ralston , phys . rev .
* d 28 * ( 1983 ) 260 .
r. d. carlitz and r. s. willey , phys . rev . *
d 45 * ( 1992 ) 2323 .
teryaev , in aip conference proceedings 343 on high energy spin physics , eds .
k. j. heller and s. l. smith , woodbury , ny ( 1995 ) , p. 467 .
p. chiappetta and m. le bellac , z. phys * c 32 * ( 1986 ) 521 .
na10 collab .
s. falciano et al . , z. phys . * c 31 * ( 1986 ) 513 ; na10 collab . m. guanziroli et al . , z. phys . *
c 37 * ( 1988 ) 545 ; j.s .
conway et al . , phys . rev . *
d 39 * ( 1989 ) 92 .
berger and s.j .
brodsky , phys .
* 42 * ( 1979 ) 940 ; s.j .
brodsky , e.l .
berger , and g.p .
lepage in proceedings of the workshop on drell yan processes ( fermilab , batavia , 1982 , p. 187 ) ; e.l .
berger , z. phys .
* c 4 * ( 1980 ) 289 .
s.d . drell and t.m .
yan , phys .
* 25 * ( 1970 ) 316 .
lepage and s.j .
brodsky , phys .
* d 22 * ( 1980 ) .
s.j . brodsky and h.j .
lu , in proceedings of the international symposium on radiative corrections : status and outlook ( gatlinburg , tennessee , 1994 ) ; s.j . brodsky and h.j .
lu , phys . rev . *
d51 * ( 1995 ) 3652 .
t. gehrmann and w.j .
stirling , z. phys .
* c 65 * , ( 1995 ) 461 . v.l .
chernyak and a.r .
zhitnitsky , nucl . phys .
* b 201 * ( 1982 ) 492 ; ( e ) nucl . phys . * b 214 * ( 1983 ) 547 .
s.v . mikhailov and a.v .
radyushkin , phys . rev .
* d 45 * ( 1992 ) 1754 . | we consider the production of muon pairs from the scattering of pions on longitudinally polarized protons .
we calculate the cross section and the single spin asymmetry for this process , taking into account pion bound state effects .
we work in the kinematic region where the photon has a large longitudinal momentum fraction , which allows us to treat the bound state problem perturbatively .
our predictions are directly proportional to the pion distribution amplitude .
a measurement of the polarized drell - yan cross section thus allows the determination of the shape of the pion distribution amplitude .
preprint # 1#2#3#4 preprint + |
the work was supported by the doe grant de - f05 - 96er-40983 .
the help of y. gu in performing cns calculations and the discussions with i. ragnarsson are highly appreciated .
proton ( top ) and neutron ( bottom ) single - particle energies ( routhians ) in the self - consistent rotating potential as a function of the rotational frequency @xmath99 .
they are given along the deformation path of the configuration ( @xmath27 above the band crossing ) assigned to the band 1 .
solid , short - dashed , dot - dashed and dotted lines indicate @xmath100 , @xmath101 , @xmath102 and @xmath103 orbitals , respectively . at @xmath104 mev , the single - particle orbitals are labeled by means of the asymptotic quantum numbers @xmath105\omega$ ] ( nilsson quantum numbers ) of the dominant component of the wave function .
large nilsson labels are used to indicate the proton orbitals which participate in unpaired band crossings seen in the bands 1 and 2 .
these unpaired band crossings are indicated by the arrow a ( band 1 ) and ellipse b ( band 2 ) .
solid ( open ) circles indicate the orbitals occupied ( emptied ) in the configuration assigned to the band 1 .
solid ( open ) squares show the orbitals occupied ( emptied ) in the configuration of the band 2 as compared with the one of the band 1.,width=226 ] kinematic ( panel ( a ) ) and dynamic ( panel ( b ) ) moments of inertia .
circles and squares are used for the bands 1 and 2 , respectively .
their theoretical counterparts are shown by solid and dashed lines , respectively .
kinematic moment of inertia of the band 1 is shown under assumption that the spin of its initial state is @xmath31 .
the comparison of calculated and experimental kinematic moments of inertia for band 2 as well as relative alignment analysis suggest spin @xmath52 for its lowest state . with this spin assignment @xmath24 moment of inertia of band 2
is lower by only @xmath106 mev@xmath79 than @xmath24 of band 1 at @xmath107 mev and , thus , for simplicity it is not shown in panel ( a ) .
the insert compares calculated and experimental ( with suggested spin assignments ) relative alignments of bands 2 and 1 , defined as @xmath108 @xcite .
, width=302 ] the same as fig .
[ routh ] , but along the deformation path of hyperdeformed @xmath109 configuration .
dashed box indicates the frequency range corresponding to the spin range @xmath110 in this configuration .
solid ( open ) circles indicate the occupied ( emptied ) orbitals .
the labels @xmath111 are used to show two lowest @xmath72 orbitals.,width=264 ] the self - consistent neutron density @xmath112 as a function of @xmath113 and @xmath114 coordinates for lowest hd configuration ( @xmath70 ) at rotational frequency @xmath115 mev .
the densities are shown in steps of 0.01 @xmath116 starting from @xmath117 @xmath116 .
, width=302 ] | the superdeformation and hyperdeformation in @xmath0cd have been studied for the first time within the framework of the fully self - consistent cranked mean field theory , namely , cranked relativistic mean field theory .
the structure of observed superdeformed bands 1 and 2 have been analyzed in detail .
the bumps seen in their dynamic moments of inertia are explained as arising from unpaired band crossings .
this is contrary to an explanation given earlier within the framework of projected shell model .
it was also concluded that this nucleus is not doubly magic sd nucleus . since the discovery of superdeformation in @xmath1dy two decades ago @xcite , nuclear super- and hyperdeformation have been in the focus of attention of nuclear structure community . at present , superdeformation ( sd ) has been discovered in different mass regions and extensively studied experimentally @xcite and theoretically ( see , for example , ref .
@xcite and references therein ) . on the other side , nothing is known experimentally about hyperdeformation ( hd ) apart from some indications of this phenomenon at low spin in the uranium nuclei @xcite .
recent observation of the very extended shapes in @xmath0cd @xcite has opened new region of superdeformation and renewed interest to the study of hyperdeformation @xcite .
two sd bands have been found in this nucleus . in the present manuscript , they are labeled according to ref .
@xcite , namely , the lowest sd band as band 1 and the excited sd band as band 2 .
these experiments were strongly motivated by earlier calculations of ref .
@xcite and more recent studies of ref .
@xcite .
so far theoretical interpretation has been performed only for band 1 within the framework of projected shell model ( psm ) ( ref .
@xcite ) .
clear deficiency of this description is the fact that the equilibrium deformation is not defined within some `` variational '' procedure , but is adjusted for a better description of moments of inertia .
it was concluded in ref .
@xcite that the low-@xmath2 @xmath3 proton orbitals are responsible for the observed bump in the dynamic moment of inertia @xmath4 at low rotational frequencies and that the 2-quasiparticle configurations from these orbitals dominate the structure of the observed states .
however , it is well known that the pairing is considerably quenched at high rotational frequencies and superdeformation @xcite . for example , no paired band crossing is observed above rotational frequencies @xmath5 mev in the @xmath6 @xcite and @xmath7 mev in the @xmath8 @xcite mass regions of superdeformation and the experimental data above these frequencies are well ( within @xmath9 accuracy for the moments of inertia ) described by the cranked relativistic mean field ( crmf ) theory with no pairing . considering that the @xmath10 mass region of superdeformation is located between these two regions ,
it is reasonable to expect that the influence of pairing will be negligible above @xmath11 mev .
thus , rephrasing the results of the psm studies in ref .
@xcite to the language of the calculations without pairing , one can conclude that the @xmath12 configuration has to be assigned to the band 1 in @xmath0cd above the band crossing .
the @xmath12 sd configurations are active in the medium - mass part of the rare - earth region of superdeformation [ gd ( @xmath13 ) nuclei ] @xcite .
however , the fermi level is located at much lower @xmath14 in @xmath0cd as compared with the @xmath6 mass region of superdeformation , and , thus , considerably larger deformation and higher rotational frequencies than in the rare - earth region of sd will be required to have the two lowest @xmath15 proton orbitals occupied . in the present manuscript ,
the crmf theory @xcite is used for detailed investigation of the structure of observed sd bands in @xmath0cd and the nature of hyperdeformation in the @xmath10 mass region .
additional insight has been provided by the cranked nilsson - strutinsky ( cns @xcite ) calculations performed with the standard nilsson parameters @xcite . in both calculations , the pairing is neglected .
the crmf equations are solved in the basis of an anisotropic three - dimensional harmonic oscillator in cartesian coordinates with the deformation parameters @xmath16 , @xmath17 and oscillator frequency @xmath18 mev .
all fermionic and bosonic states belonging to the shells up to @xmath19 and @xmath20 are taken into account in the diagonalization of the dirac equation and the matrix inversion of the klein - gordon equations , respectively .
the detailed investigation indicates that this truncation scheme provides good numerical accuracy .
the nl1 set @xcite is used for the rmf lagrangian .
as follows from our experience @xcite , this set provides reasonable description of the single - particle energies .
the results of the crmf calculations for the configurations forming the yrast line or located close to it in energy are shown in fig .
[ eld ] .
according to these calculations , normal - deformed bands , many of which show high triaxiality indicative of approaching band termination , dominate the yrast line up to @xmath21 . at higher spin hyperdeformed bands
become yrast .
the sd bands are never yrast , but come close to the yrast line at @xmath22 . the lowest sd configuration has @xmath23 ( even spins ) quantum numbers and it is assigned to band 1 .
we assign bands 1 and 2 to sd configurations , because the calculated kinematic @xmath24 and dynamic @xmath4 moments of inertia agree well with experiment .
the configurations with normal deformation have too small and the configurations with hyperdeformation too large values of moments of inertia .
the experimental values of transition quadrupole moment @xmath25 are best reproduced by the sd configurations .
the proton and neutron single - routhian diagrams along the deformation path of this configuration are shown in fig .
[ routh ] . before unpaired band
crossing it has the @xmath26 structure , while after band crossing the @xmath27 structure .
the unpaired band crossing is due to the crossing of the @xmath281/2(r =- i)$ ] and @xmath291/2(r =- i)$ ] orbitals ( arrow a in fig . [ routh ] ) .
the exact band crossing frequencies are not known for bands 1 and 2 because downsloping branches of @xmath4 below the band crossing have not been observed .
the comparison of the experimental and calculated bumps in @xmath4 suggests that the crossing takes place @xmath30 kev earlier in the calculations than in experiment ( see fig .
[ j1 ] ) .
the frequency of this band crossing depends on the relative position of the above mentioned orbitals and thus the discrepancy between experiment and theory suggests that the relative energy distance between these orbitals is underestimated in the calculations by approximately 0.7 mev .
the crmf calculations well reproduce the absolute value of dynamic moment of inertia @xmath4 above the band crossing , but underestimate somewhat the height of the bump in dynamic moment of inertia @xmath4 at the band crossing ( fig . [ j1]b ) . the best agreement between calculated and experimental kinematic moments of inertia
@xmath24 is seen if the lowest state in band 1 has spin @xmath31 ( fig .
[ j1]a ) .
this suggests that the band in @xmath0cd has been observed in the spin range from @xmath32 up to @xmath33 .
however , taking into account that typical accuracy of the description of the moments of inertia in the crmf calculations is around 5% @xcite and possible minor impact of pairing at high rotational frequencies , which would lead to slight decrease of calculated @xmath24 @xcite , one can not completely exclude that the lowest sd band in @xmath0cd has been observed from @xmath34 up to @xmath35 .
the spin range @xmath36 has been suggested in ref .
@xcite using the assumption that @xmath37 . however , this assumption is not supported by our calculations where @xmath38 for assigned configuration above the band crossing .
the average transition quadrupole moment of the assigned configuration in the suggested spin range is @xmath39 @xmath40b , which agrees with the lower limit of @xmath41 @xmath40b obtained in experiment @xcite .
the occupation of the first proton @xmath15 orbital at the band crossing has only minor impact on @xmath25 : an increase of @xmath25 by approximately 0.2 @xmath40b .
band 2 has the features similar to the band 1 except that the band crossing takes place @xmath30 kev earlier ( fig . [ j1]b ) .
especially interesting is the observation that the @xmath4 moments of inertia of these two bands are very similar above the band crossing .
similar to the @xmath6 region of superdeformation @xcite , this suggests that the intruder content of observed bands should be the same .
the only possible explanation found in our calculations is related to the excitation of proton from the @xmath281/2(r=+i)$ ] orbital into the lowest positive parity orbital with signature @xmath42 above the @xmath14 shell gap ( see fig . [ routh ] ) . in fig .
[ routh ] above @xmath43 mev , this orbital has the @xmath443/2 $ ] structure .
however , contrary to the situation shown in fig .
[ routh ] at equilibrium deformation of the configuration of interest the @xmath453/2(r=\pm i)$ ] orbitals are located below the @xmath467/2(r=\pm i)$ ] orbitals even at zero rotational frequency . at @xmath47 mev ,
the @xmath443/2(r =- i)$ ] and the @xmath291/2(r =- i)$ ] orbitals interact strongly .
this interaction creates the bump in @xmath4 ( fig .
[ j1]b ) . above the band
crossing , the calculated @xmath4 reproduces well the experimental one and is very close to the @xmath4 values of the configuration assigned to the lowest sd band , the feature observed also in experiment .
when interpreting band 2 , we also considered particle - hole excitations in neutron subsystem , namely from the @xmath481/2(r=+i)$ ] orbital to the @xmath495/2(r=\pm i)$ ] orbitals ( fig .
[ routh ] ) .
such excitations lead to smaller deformation , and thus to the delay of the unpaired band crossing ( originating from interaction of the @xmath281/2(r=+i)$ ] and the @xmath291/2(r=+i)$ ] orbitals ) .
for example , in the configuration based on the occupied @xmath495/2(r =- i)$ ] orbital this crossing is delayed by @xmath50 mev , which is in contradiction with experiment . therefore , similar to the configuration assigned to the band 1 , the structure of the configuration assigned to the band 2 changes from @xmath26 to @xmath27 at the band crossing .
however , it has total signature @xmath51 ( odd spins ) and negative parity . the band crossing
a , active in the configuration of the band 1 , is not active in the configuration of band 2 since both interacting orbitals ( @xmath281/2(r =- i)$ ] and @xmath291/2(r =- i)$ ] ) are occupied ( see fig .
[ routh ] ) . since the @xmath467/2(r =- i)$ ] and
@xmath291/2(r =- i)$ ] orbitals interact very weakly ( see inside of ellipse b in fig . [ routh ] ) , their interaction can not explain the observed features of band 2 in the band crossing region . the comparison of relative alignments of bands 2 and 1 ( fig .
[ j1 ] ) suggests that with @xmath31 assigned for initial state of band 1 , the band 2 has been observed in the spin range from @xmath52 up to @xmath53 . in the region beyond band
crossing ( @xmath54 mev ) , the calculated relative alignments differ from the experimental ones only by @xmath55 : this is within the typical uncertainty of the description of relative alignments in the crmf theory @xcite .
the calculations show that the quadrupole deformations of the configurations assigned to the bands 1 and 2 are almost the same , but they differ somewhat in the @xmath56-deformation .
while the configuration of band 1 has @xmath57 , the one of the band 2 has @xmath58 . as a consequence ,
the configuration of the band 2 is more collective ( @xmath59 @xmath40b ) as compared with the one of band 1 .
although the experimental analysis for band 2 yielded a lower value for the transition quadrupole moment @xmath60 @xmath40b , it suffered from the large errors on the @xmath61 values which did not allow an accurate measurement of the @xmath25 values @xcite .
thus based on the similarity of rotational properties of bands 1 and 2 above the band crossing and the results of the crmf calculations , it is reasonable to believe that the deformation for band 2 is comparable to the one for band 1 ( see also ref .
@xcite ) .
the earlier band crossing in the configuration of band 2 as compared with experiment is due to small energy distance between the @xmath443/2 $ ] and @xmath291/2 $ ] orbitals in the calculations . in order to have this crossing at
experimentally observed frequency , one should increase this distance by @xmath62 mev .
this value is similar to the underestimate of the energy distance between the @xmath281/2 $ ] and @xmath631/2 $ ] orbitals deduced earlier .
this suggests that the energy distance between the @xmath443/2 $ ] and @xmath281/2 $ ] orbitals , and , consequently the @xmath14 sd shell gap is reasonably well reproduced in the calculations .
thus , one concludes that the @xmath14 sd shell gap is smaller than the @xmath64 sd shell gap in the @xmath8 region of superdeformation ( see fig . 1 in ref .
@xcite ) and @xmath65 sd shell gap in the rare - earth region of superdeformation ( see fig . 3 in ref .
@xcite ) .
this maybe a possible reason why the search for sd bands in neighbouring nuclei has been unsuccesful so far @xcite .
in addition , it indicates that in reality the @xmath631/2 $ ] orbital is located @xmath62 mev higher in energy with respect of the @xmath281/2 $ ] and @xmath443/2 $ ] orbitals than suggested by the present calculations .
this is not far away from the typical accuracy with which the energies of the single - particle states are reproduced at normal- @xcite and superdeformation @xcite in the rmf calculations with the nl1 set .
unpaired band crossings at superdeformation were reported earlier , for example , in @xmath66gd @xcite and @xmath67ho @xcite .
the band crossing in the sd band 1 of @xmath67ho is similar to the ones seen in @xmath0cd since it involves the interaction of the routhians with @xmath68 .
the hyperdeformed configurations become yrast above @xmath69 ( fig .
[ eld ] ) .
these are the signature partner @xmath70 configurations of negative parity which are degenerated in energy .
this energy degeneracy is due to the excitation of one neutron from the @xmath717/2(r=\pm i)$ ] orbitals into the lowest @xmath72 orbital ( see fig .
[ routh - hd ] ) .
the hd configurations are favoured in energy at these spins due to the @xmath14 hd shell gap and low neutron level density at @xmath73 .
the transition quadrupole moment @xmath25 increases from @xmath74 @xmath40b at @xmath75 up to @xmath76 at @xmath77 .
this fact is related to the stretching of the nucleus due to centrifugal force .
the shape of nucleus corresponds to the 2.3:1 semiaxis ratio ( fig . [ density ] ) .
the kinematic moment of inertia drops down from @xmath78 mev@xmath79 at @xmath75 to @xmath80 mev@xmath79 at @xmath77 . in the frequency range where
the experimental bands 1 and 2 have been observed , the dynamic moments of inertia of calculated configurations are equal to @xmath81 mev@xmath79 ; the quantity which is much larger than in experiment . at higher frequencies
there is smooth increase of @xmath4 up to @xmath82 mev@xmath79 at @xmath83 mev , possibly caused by unpaired band interaction .
no experimental counterparts of these configurations have been observed so far .
it is interesting to mention the similarity of the single - particle spectra at hyperdeformation obtained in the crmf calculations ( fig .
[ routh - hd ] ) and in the wood - saxon - strutinsky calculations of ref .
@xcite ( fig . 4 and 5 of ref .
@xcite )
. however , the @xmath84 ( @xmath85 and @xmath86 ) shell gaps are somewhat smaller ( larger ) in the crmf calculations . in the later calculations ,
the quadrupole , octupole , hexadecapole and necking degrees of freedom were taken into account .
there are , however , no indications that octupole deformation should play a role at hd in @xmath0cd @xcite . on the other side ,
the crmf calculations take into account all even - multipole deformations and triaxiality in a fully self - consistent way .
the basis of the crmf model is sufficiently large to see if there is a pronounced tendency for the development of necking .
fig .
[ density ] , showing the densities of lowest hd configuration , does not indicate such a tendency .
the cns calculations , which will not be discussed here in detail , also indicate that the band 1 in @xmath0cd is associated with the @xmath27 configuration having @xmath87 @xmath40b and located in the @xmath88 minimum .
similar to earlier calculations ( ref .
@xcite ) , the triaxiality is more pronounced in the cns calculations for @xmath0cd as compared with the crmf ones . in the cns calculations ,
the hd configurations are yrast above @xmath89 , and the @xmath70 configurations with @xmath90 @xmath40b are the lowest ones .
thus , one can conclude that general properties of sd and hd bands are similar in both approaches .
the present crmf calculations suggest that in the spin range where the feeding of sd bands takes place ( i ) the sd bands are non - yrast and ( ii ) the hd bands compete in energy with the sd bands
. however , only sd bands have been reported at high spin in refs . @xcite . because of limited experimental information it is difficult to find unique explanation for this discrepancy .
on one side , one would expect that the calculated relative energies of different minima deviate somewhat from reality .
alternative explanation is related to the differences in feeding mechanism of sd and hd bands @xcite .
although the hd and sd states are comparable in energy at @xmath91 at zero temperature ( fig .
[ eld ] ) , these high spin nuclear states are populated at somewhat higher spin at temperatures @xmath92 mev . at these temperatures ,
the deformed shell structure melts , and thus sd minimum becomes lower in energy than hd one . as a consequence , the feeding of the sd minimum maybe the dominant population channel @xcite .
in summary , the detailed analysis of the structure of observed superdeformed bands 1 and 2 in @xmath0cd within the cranked relativistic mean field theory led to new configuration assignment : both bands have been assigned the @xmath27 configurations .
the band crossings seen in these bands have been related to the unpaired crossings of the lowest proton @xmath15 orbital and the @xmath93 orbitals .
the size of the @xmath14 sd shell gap , which is smaller than that of doubly magic @xmath94zn and @xmath1dy sd nuclei , and the absence of large neutron shell gap at @xmath95 indicate that the @xmath0cd nucleus can not be considered as a doubly magic sd nucleus .
the properties of hyperdeformed configurations , which are expected to be yrast above @xmath96 , have also been discussed . |
the decoherence of a center spin in a spin bath@xcite has been of interest in spin resonance spectroscopy for a long historykubo1954_jpsj , pines1955,feher1959,klauder1962 and is also a paradigmatic model in studying the state collapse in quantum mechanics .
recent revisited interest in this problem is mostly due to the decoherence issue in quantum computing.@xcite most relevant are single electron spins in quantum dots or impurity centers where the dominating decoherence mechanism at low temperatures ( such as below a few kelvins ) is the nuclear spins of the host lattice.@xcite in a small system such as a quantum dot , the center spin ( hereafter referred to as qubit for clarity ) and the spin bath , in the time - scale of decoherence , is a relatively isolated subsystem in the whole environment .
thus the qubit decoherence is due to the entanglement with the bath during the coherent evolution of the whole system.@xcite in this paper , we are interested in the so - called pure dephasing in which the qubit experiences no longitudinal relaxation but only loses its off - diagonal phase coherence .
the pure dephasing is relevant to an electron spin under a moderate or strong magnetic field ( @xmath0 t for a typical gaas dot ) , where the electron spin flip due to the hyperfine interaction is largely suppressed by the zeeman energy mismatch between the electron and the nuclei . in the absence of qubit flip ,
a qubit - bath system has a hamiltonian as @xmath1 by which the spin bath is driven by different hamiltonians @xmath2 depending on the qubit states @xmath3 .
when a coherent qubit state @xmath4 is prepared , the initial state of the qubit - bath system is the product state @xmath5 . at time @xmath6
, the bath evolution @xmath7 predicated on the qubit state @xmath8 establishes an entangled state @xmath9 .
the qubit coherence is reduced from @xmath10 to @xmath11 .
the decoherence is characterized by the bath state overlap @xmath12 .
thus the key is the many - body bath dynamics caused by the interaction within the bath . without many - body interactions , the bath would not evolve ( except for a trivial phase factor that can be eliminated by standard spin - echo@xcite ) and the qubit coherence would not decay . recently , a variety of quantum many - body theories for nuclear spin bath dynamics have been developed including the density matrix cluster expansion ( ce),@xcite the pair - correlation approximation,@xcite and the linked - cluster expansion ( lce).@xcite in the pair - correlation approximation , each pair - wise flip - flop of nuclear spins is identified as an elementary excitation mode and is taken as independent of each other . to study the higher oder correlations , the feynman diagram lce is developed .
the evaluation of higher - order lce , however , is rather tedious due to the increasing number of diagrams , especially for spins higher than 1/2 ( see the appendix for details ) .
the density matrix ce is developed in the spirit of the standard cluster expansion or virial expansion for interacting gases in grand canonical ensembles.@xcite it serves as a simple method ( without the need to count or evaluate feynman diagrams ) to include the higher - order spin interaction effects beyond the pair - correlation approximation .
the ce calculations show that when the pair - correlation contribution to the decoherence is suppressed by pulse control,@xcite the higher - order correlations are not negligible.@xcite in order to obtain the decoherence exponent from the ce , however , the terms involving overlapping clusters have to be neglected , which make the accuracy problematic even when the expansion converges .
this problem limits the ce to applications in large baths in which the contribution of each individual cluster to the decoherence is small and the overlapping correction is unimportant . in this work ,
we develop a cluster - correlation expansion ( cce ) method in which the bath spin evolution are factorized into cluster correlations .
each cluster correlation term is equivalent to the sum of all the lce seriessaikin2007 consisting of a given set of bath spins flip - flopped . in particular , the two - spin cluster correlations include all diagrams with two spins flip - flopped and is equivalent to the pair - correlation approximation.@xcite the cce bears the accuracy of the lce ( the results are accurate whenever converge ) and the simplicity of the ce ( without the need to count or evaluate feynman diagrams ) , while free from the large - bath restriction of the ce .
the cce coincides with the ce in the leading order of the short - time expansion , which is applicable in large spin baths where the decoherence completes well within the bath spin flip - flop time . for small baths , however
, the qubit decoherence may not complete within the bath spin flip - flop time and individual higher - order cluster correlations could grow significant . in this case
only the cce converges to the exact coherent dynamics of multi - spin clusters .
such coherent dynamics of small clusters of bath spins is of special interest in systems with randomized qubit - bath couplings .
an interesting example is nitrogen - vacancy centers in diamondsjelezko2004_prlnv , childress2006_sciencenv , dutt2007_sciencenv , hanson2008_sciencenv which are coupled to randomly located bath spins ( carbon-13 and nitrogen nuclear spins ) in the proximity .
this paper is organized as follows . in section
ii , we derive , for a generic spin bath hamiltonian , the cce from a recursive cluster factorization procedure .
we show that the cce is equivalent to an infinite resummation of the lce series and also compare it to the ce . in section iii
, we check the accuracy and convergence of the truncated cce in an exactly solvable model ( the one - dimensional spin-1/2 xy model ) .
section iv draws the conclusions .
the appendix contains the details of the lce for spin bath dynamics , extended to spins higher than @xmath13 .
as discussed in the introduction , the hamiltonian for the pure dephasing problem has the form of eq .
( [ pure_dephasing_h ] ) . for a given initial bath state @xmath14 ( which could be considered as one sample chosen from a thermal ensemble ) ,
the qubit coherence is characterized by @xmath15for a thermal ensemble of baths , a further ensemble average should be processed .
the thermal fluctuation leads to the inhomogeneous broadening , which can be eliminated by spin echo . to focus on the qubit decoherence due to the quantum dynamics of the bath , throughout this paper
, we consider the decoherence for a single bath state without the ensemble average . for temperatures much higher than the bath spin flip - flop rates ( @xmath1610@xmath17 k for nuclear spins in gaas ) ,
the thermal ensemble has no off - diagonal coherence and @xmath14 can be taken as a noninteracting product state @xmath18 , where @xmath19 denotes the quantum number of the @xmath20th bath spin quantized along the external magnetic field . in this subsection
we illustrate the central idea of the cce method ( as an extension of the previously developed pair - correlation approximationyao2006_prb , yao2007_restorecoherence , liu2007_njp ) using a bath consisting of @xmath21 spins @xmath22 , @xmath23 , @xmath24 , and @xmath25 with only pairwise secular interactions @xmath26where @xmath27 @xmath28 is the qubit ( bath spin ) splitting energy , @xmath29 is the diagonal qubit - bath spin interaction constant ( corresponding to the hyperfine interaction strength in an electron - nuclear spin system ) , @xmath30 ( @xmath31 ) is the diagonal ( off - diagonal ) intrinsic bath interaction strength , and @xmath32 ( @xmath33 ) is the diagonal ( off - diagonal ) extrinsic bath interaction depending on the qubit states ( which could result from the interaction mediated by virtual flips of the qubit spin while real flips are suppressed by the large energy mismatch@xcite ) . in the pair - correlation approximation ,
the qubit coherence is given by the product of all possible spin pair contributions up to a phase factor , yao2006_prb @xmath34where the contribution due to the flip - flops of a spin pair @xmath35 [ see fig .
[ g_cluster23](a ) ] is @xmath36with the hamiltonian governing the pair dynamics obtained from the full hamiltonian @xmath2 by excluding the flip - flops of all spins other than @xmath37 and @xmath38 @xmath39which is equivalent to replacing in the full hamiltonian the spins outside the pair with their mean - field averages , @xmath40the pair - correlation approximation is valid for situations where pair correlations dominate . when the collective flip - flops of more spins become important
, we need to consider the higher - order correlation correction @xmath41 defined by @xmath42to illustrate how the high - order correction could be evaluated , let us consider a spin bath of only three spins @xmath43 .
obviously , the result is @xmath44this result motivates a definition of non - factorizable three - spin correlations due to collective flip - flops . for a three - spin cluster @xmath45 in a bath
[ see fig .
[ g_cluster23](b ) ] , when all the spins outside the cluster are frozen , the qubit coherence is @xmath46with the cluster hamiltonian @xmath47obtained from the full hamiltonian @xmath2 by replacing the spins outside the cluster with their mean - field averages .
the authentic ( or non - factorizable ) three - spin correlation is singled out by excluding all the pair - correlations @xmath48if all such three - spin correlations are picked up , the qubit coherence is given by ( up to a global phase factor ) @xmath49thus a systematic cluster correlation expansion is motivated .
pulses at @xmath50,@xmath51,@xmath52 , ( indicated by vertical lines ) and the corresponding contour - time - dependent hamiltonian for electron spin decoherence .
( b ) , ( c ) , and ( d ) exemplify the cases of free - induction decay , hahn echo , and carr - purcell echo , respectively . ] we consider a generic bath hamiltonian for the pure dephasing problem @xmath53which need not contain only pairwise interactions , or conserve the spin angular momentum along any direction .
for instance , multi - spin interaction terms like @xmath54 and non - secular terms like @xmath55 could be present . for qubit decoherence under the control of an arbitrary sequence of @xmath56-pulses applied at @xmath50 , @xmath51 , @xmath24 , as shown schematically in fig .
[ g_contour](a ) , the bath evolution predicated on the qubit state is given by @xmath57where @xmath58the qubit coherence at the end of the evolution is @xmath59it can be written in the contour time - ordered form as@xcite @xmath60where @xmath61 is the time - ordering operator along the contour @xmath62 . as illustrated in fig .
g_contour(a ) , the contour time - dependent hamiltonian @xmath63 alternatively switches between @xmath64 and @xmath65 each time the qubit state is flipped by a @xmath56-pulse or when the time direction is reverted at @xmath66 .
the examples for free - induction decay , single - pulse hahn echo , and carr - purcell echo are illustrated in figs .
[ g_contour ] ( b ) , ( c ) , and ( d ) , respectively .
following the idea illustrated in the previous subsection , the cluster correlations are recursively defined as follows .
the empty - cluster correlation @xmath67is a pure phase factor obtained from eq .
( [ w0timeordered ] ) by replacing the bath hamiltonian @xmath63 with its mean - field average @xmath68 .
2 . the single - spin correlation @xmath69 where @xmath70 is obtained from eq .
( [ w0timeordered ] ) by replacing the bath hamiltonian @xmath71 with @xmath72 in which all the spin operators except @xmath73 are mean - field averaged .
3 . the two - spin ( pair ) correlation @xmath74where @xmath75is obtained from eq .
( [ w0timeordered ] ) by replacing the bath hamiltonian @xmath76 with @xmath77 in which all the spin operators except @xmath73 and @xmath78 are mean - field averaged .
so on and so forth , the cluster correlation for an arbitrary set of bath spins @xmath79 is defined as @xmath80 where @xmath81 is obtained from eq .
( [ w0timeordered ] ) by replacing the bath hamiltonian @xmath71 with @xmath82 in which all the spin operators outside the cluster are mean - field averaged or their flip - flops are frozen . thus , by definition , the qubit coherence is factorized into all possible cluster correlations as @xmath83calculating the cce to the maximum order @xmath84 amounts to solving the exact bath propagator , which is in general not possible . in the decoherence problem
, we consider a finite - time evolution and it often suffices to truncate the expansion by keeping cluster correlations up to a certain size @xmath85 , as the @xmath85th - order truncation of the cce ( @xmath85-cce for short ) , @xmath86where @xmath87 is the number of spins contained in the cluster @xmath79 . as an example , for the pairwise hamiltonian in eq .
( [ hpn ] ) , the lowest nontrivial order of truncation is @xmath88which is the pair - correlation approximation .
saikin _ et al .
_ have recently developed an lce method for the qubit decoherence in a spin-1/2 bath.@xcite the detailed descriptions of the lce for a generic spin bath are given in the appendix .
in general , the bath evolution can be factorized using feynman diagrams so that @xmath89where @xmath90is the sum of all connected feynman diagrams obtained by using wick s theorem on the series expansion of eq .
( [ w0timeordered ] ) .
some of the connected diagrams up to the 4th - order have been evaluated in ref .
[ ] for free - induction decay and single - pulse hahn echo with a spin-1/2 bath hamiltonian in the form of eq .
( [ hpn ] ) .
the complexity for diagram counting and evaluation increases dramatically when considering higher - order diagrams ( see fig .
[ g_d123 ] in appendix append_lcespin12 ) or higher spins [ see fig .
[ g_spin1ld12](b ) in appendix [ append_lcegeneral ] for the second - order diagrams for a spin-1 bath ] .
we notice that each diagram can be expanded as the sum of diagrams involving the flip - flops of different clusters of spins . as an example shown in fig .
[ g_cceu02](a ) , a third - order diagram involving the flip - flop of a spin pair contains diagrams for spin clusters @xmath91 , @xmath92 , @xmath24 , where the numbers stand for the indices of the spins flip - flopped ( i.e. , the @xmath93 quantum number changed ) . thus all the connected diagrams can be classified according to the spin clusters instead of the interaction orders . for an arbitrary cluster @xmath79 , we define @xmath94 as the sum of all connected diagrams in which all and only the spins in cluster @xmath79 have been flip - flopped . for instance , some of the diagrams constituting @xmath95 and @xmath96 for a spin-1/2 hamiltonian in eq .
( [ hpn ] ) are shown in figs .
[ g_cceu02](b ) and ( c ) , respectively . with these @xmath97 functions ,
the lce is expressed as @xmath98 in particular , the infinite summation of all the connected diagrams for a certain cluster @xmath99 and all its subsets @xmath100can be obtained from the series expansion eq .
( [ lce_vk ] ) by dropping all the terms involving the flip - flop of spins outside the cluster @xmath101 , or , equivalently , by reducing the bath hamiltonian @xmath63 to the cluster hamiltonian @xmath102 in which the spins outside the cluster are mean - field averaged .
thus we have @xmath103comparing this to eqs .
( [ uc_expansion])-([h_cluster ] ) , we immediately have @xmath104i.e .
, a cluster correlation term corresponds to the infinite partial summation of all the connected diagrams in which all and only the spins in the cluster have been flip - flopped . and @xmath105 , respectively . in the diagrams ,
a solid arrow denotes the propagation of a spin .
a wavy ( dotted ) line connecting two solid arrows / spins denotes a pairwise off - diagonal ( diagonal ) interaction , and an open - ended line represents the interaction with the qubit spin or an external field .
each filled circle / empty circle / empty square on a solid arrow denotes a @xmath106 operator , which raises / lowers / keeps invariant the @xmath107 quantum number of that spin . ] from the lce expression of the cce in eq .
( [ cce_lce ] ) , it is obvious that the short - time profile of the decoherence due to clusters of a certain size is determined by the lowest - order diagram .
in particular , for free - induction decay with the secular pair - interaction hamiltonian in eq .
( [ hpn ] ) , the leading order contribution from a cluster of size @xmath85 is @xmath108where @xmath109 is the typical magnitude of the pair flip - flop interaction strength @xmath110 .
if each spin interacts , on average , with @xmath111 spins , then the number of size-@xmath85 clusters is @xmath112 with @xmath21 the total number of bath spins .
the sum of all the leading order @xmath85-spin connected diagrams is @xmath113 .
thus for @xmath114 i.e. , for a time @xmath66 much shorter than the bath spin flip - flop timescale @xmath115 , a truncated cce converges .
the short time condition @xmath116 is usually satisfied for electron spin decoherence caused by nuclear spins in typical quantum dots .
the convergence of the truncated cce , however , could go well beyond the short time restriction .
one such scenario is small spin baths with disorder in qubit - bath couplings or in spin splitting energies , in which multi - spin correlation could develop at a time well beyond the short - time limit but the size of the contributing clusters could remain bounded due to the localization effect in a disordered system , as will be verified later in this paper by numerical simulations .
witzel _ et al .
_ recently developed a density matrix ce approach to solving the nuclear spin dynamics in the electron spin decoherence problem,@xcite in the spirit of the cluster or virial expansion for interacting gases in grand canonical ensembles.beth1936_ce,beth1937_ce,kahn1938_ce below we reproduce the basic procedure of the ce and compare it to the cce .
instead of the ensemble ce in ref .
[ ] , we consider a single sample state of the bath for a direct comparison .
defining the qubit decoherence due to a cluster @xmath117 of bath spins as @xmath118a hierarchy of cluster terms @xmath119 are recursively defined as [ ce_summation ] @xmath120where in the last line the sum runs over all possible partitions of the cluster @xmath79 into nonoverlapping nonempty subsets @xmath121 . the @xmath85th - order
truncated ce ( @xmath85-ce for short ) is @xmath122 with the sum running over all possible partitions of the bath into nonoverlapping nonempty clusters @xmath123 of size up to @xmath85 . in the cluster expansion for interacting gases in grand canonical ensembles with translational symmetry
, the contribution from different clusters can be factorized@xcite and the evaluation of a truncated ce amounts to the calculation of a finite number of finite - size cluster contributions , similar to the cce in this paper . for a finite - size spin bath or for a bath without translational symmetry , however , such factorization of different clusters does not exist , which makes it essentially impossible to calculate the sum in eq .
( [ ce_mth ] ) even for a small-@xmath85-ce .
for example , for a bath of @xmath21 spins , the number of all terms containing only pair clusters is @xmath124 and all such terms have to be individually calculated and summed in the 2-ce , which is practically impossible .
a remedy is possible when all the cluster terms @xmath125 are individually small ( but the sum could still be substantial ) .
under such a condition , the ce can be approximated by a factorized form by adding some overlapping terms which are higher - order small.@xcite for example , with the secular pair - interaction hamiltonian in eq .
( [ hpn ] ) , for which @xmath126 , the @xmath85-ce is @xmath127 , \label{ce_factorized}\]]under the small - term condition@xmath128comparing the factorization approximation in eq .
( [ ce_factorized ] ) to the exact @xmath85-ce in eq .
( [ ce_mth ] ) , the error added is @xmath129containing products of any set of cluster terms sharing at least one spin , i.e. , the overlapping terms .
such overlapping terms are higher - order small if each individual cluster term ( for @xmath130 ) is small . furthermore , under the small - term condition ,
@xmath131 $ ] , and the ce for an arbitrary cluster becomes @xmath132which takes the same form as the cce .
thus the cce coincides with the ce under the small - term condition eq .
( smallterm ) , which is justified for large spin baths where the number of contributing clusters is large and hence the contribution from each individual cluster remains small within the timescale of decoherence .
the problem with the neglected overlapping terms is relevant for small spin baths where the coherent dynamics of a small number of multi - spin clusters dominating the decoherence may persist well beyond the bath spin flip - flop time and the small - term condition is no longer satisfied . in this case
the ce will not converge to the exact multi - spin cluster dynamics , as will be seen in the numerical check in the next section .
here we consider an exactly solvable spin bath model ( the one - dimensional spin-1/2 xy model ) and compare the exact solution to the results obtained with the cce and the ce .
the @xmath21-spin bath hamiltonian conditioned on the qubit state @xmath133 ( with spin splitting constants dropped ) is @xmath134where @xmath29 denotes the qubit - bath spin interaction coefficient , @xmath135 is the intrinsic bath interaction strength , and @xmath136 is the interaction dependent on the qubit state . the initial bath state @xmath137 is taken as a product state of all bath spins , in which the orientation of each bath spin is randomly chosen as up or down .
the qubit - bath interaction coefficients @xmath138 are either taken from a sinusoidal distribution @xmath139 ( referred to as a `` sinusoidal '' spin chain ) or randomly chosen from @xmath140 $ ] ( referred to as a `` random '' spin chain ) . hereafter @xmath141 is taken as the unit of energy .
the spin - flip interaction strengths @xmath142 and @xmath143 are randomly chosen from @xmath144 $ ] , corresponding to typical bath spin flip - flop time @xmath145 .
the exact solution is obtained by the jordan - wigner transformation , which transforms the interacting spin-1/2 chain to a noninteracting fermion system.@xcite cluster contributions ( denoted as cce@xmath146 ) to the exponential decoherence factor @xmath147 in ( a ) free - induction decay ( fid ) , ( b ) hahn echo , and ( c ) carr - purcell echo ( cp ) for a long `` sinusoidal '' spin chain with @xmath148 spins .
a solid ( dashed ) line indicates that the cluster contribution to @xmath149 is positive ( negative ) .
the exact @xmath150 is also shown ( empty squares ) for comparison . ] as the first example , we consider a sinusoisal chain of @xmath148 spins . in this case , the energy cost @xmath151 for a pairwise flip - flop between neighboring spins varies smoothly from one end to the other of the chain , so that a correlated cluster can grow to larger and larger size as time passes by . for a time greater than the bath spin flip - flop time @xmath152 ,
the whole bath could become correlated .
but the qubit decoherence would be completed within a time @xmath66 much shorter than @xmath152 , if the bath size is relatively large , or @xmath153due to the large number of small clusters contributing to the decoherence . under the short - time condition ,
the cce can be truncated with a rather small cut - off size @xmath85 .
the total contributions from clusters of various sizes to the decay of the qubit coherence as functions of the pulse delay time @xmath154 [ see figs .
[ g_contour ] ( b)-(d ) ] are shown in fig .
[ g_sine1000log ] . for the free - induction decay in fig .
[ g_sine1000log](a ) , the cluster contributions decrease rapidly with increasing the cluster size @xmath146 at a time much shorter than @xmath152 , so the 2-cce ( pair correlation approximation ) already converges to the exact result . in the single - pulse hahn echo [ fig .
[ g_sine1000log](b ) ] , the higher - order correlations are more noticeable than in the free - induction decay , but the pair correlations still dominate for @xmath155 . for the two - pulse carr - purcell echo [ fig . [ g_sine1000log](c ) ] , as the decoherence due to the pair correlations is eliminated in the leading order of the spin - flip interactions,@xcite the larger - size cluster correlations become important and a 6-cce is required to reproduce the exact solution .
the results from the 6-cce ( dotted lines ) and 6-ce ( solid lines ) are compared with the exact solutions ( empty squares ) . ] in the relatively large bath , due to the large number of contributing clusters , the qubit decoherence completes when each individual cluster contribution is small , i.e. , the small - term condition in eq .
( [ smallterm ] ) is satisfied .
thus the cce coincides with the ce .
this is verified in fig .
[ g_fidhahncp ] , where the exact solution for qubit decoherence agrees with both the cce and the ce truncated at the 6th order .
spins in free - induction decay ( fid ) , hahn echo , and carr - purcell echo ( cp ) .
the results from ( a ) cce and ( b ) ce truncated to the 2nd ( dashed lines ) , the 4th ( solid lines ) and the 6th ( dotted lines ) order are compared to the exact solution ( empty squares ) . ] as the second example , we consider a short `` sinusoidal '' spin chain consisting of @xmath156 spins . in this case
the number of contributing clusters is small and the decoherence proceeds much slower
[ g_comparesine100 ] shows that both the cce and the ce converge to the exact result for a time much shorter than @xmath157 .
as the time approaches and goes beyond the bath spin flip - flop time @xmath158 , the deviation from the exact solution is noticeable in both the cce and the ce , indicating the emergence of correlations for clusters larger than the truncation cut - off size .
but the cce agrees with the exact result as long as it converges , while the ce converges to a different result ( for @xmath159 in free - induction decay , @xmath160 in hahn echo , and @xmath161 in carr - purcell echo ) .
the deviation of the converged ce from the exact result is due to the overlapping correction [ see eq .
( [ ce_overlap ] ) ] neglected in the ce : for @xmath162 , the contribution of an individual cluster could be sizable and the small - term condition in eq .
( smallterm ) is violated , so the overlapping correction has to be taken into account . in the example
discussed above , the cce may not converge at a long time @xmath163 .
this is because the qubit - bath spin coupling assumes a smooth sinusoidal distribution and hence the energy cost of each neighboring pair is small and slow - varying as a function of the pair s position along the chain .
the small and slow - varying energy cost of a pair - flip makes it possible for one pair - flip to affect its neighboring pairs , then the next neighbors , and so on and so forth .
this way large - size cluster correlations may grow rapidly after the time goes beyond the pair - flip time @xmath152 . in a relatively small bath of smooth qubit - bath coupling distribution , the cce may fail to converge in the long time limit which is of interest in elongating the qubit coherence time by pulse control
. spins .
( a ) @xmath85-cce and ( b ) @xmath85-ce are compared to the exact solution ( empty squares ) . ] except that the qubit - bath coupling constants @xmath138 are such that there are four neighboring spins in near - resonance .
notice that the 4th - order and the 6th - order truncations are not distinguishable by eyes in both cce and ce . ]
the convergence problem of the cce for small spin baths may be avoided if the qubit - bath coupling is random so that the bath correlation is localized and the size of correlated clusters is upper bounded . as the last example , we consider a short `` random '' spin chain consisting of @xmath156 spins . in this case , the pair - flip energy cost is usually much larger than the spin - flip strength unless two neighboring spins are accidentally in near - resonance .
thus large - size cluster correlation can hardly grow significant even for a time well beyond the bath spin flip - flop timescale @xmath152 .
figure [ g_compareran1 ] shows that the pair - correlation approximation already suffices for the qubit decoherence for an arbitrarily long time .
since higher - order correlations are small , the overlapping correction to the ce ( which corresponds to at least three - spin cluster correlations ) is unimportant , good agreement between the ce and the exact solution is also seen .
interestingly , in the small bath of random qubit - bath coupling , the qubit decoherence is not fully developed but coherent oscillations persist over a long time .
these coherent qubit oscillations have indeed been observed in a system of a similar nature , namely , nitrogen - vacancy centers in diamonds coupled to nuclear spins in the proximity.@xcite we would like to point out that the coherent oscillations can not be reproduced by the lce truncated as any finite interaction order . as discussed in sec .
[ cce_ce ] , the cce and the ce differ in describing the multi - spin correlations , which may be important for a small spin bath .
such difference can indeed be seen in fig .
[ g_comparesine100 ] for a short `` sinusoidal '' spin chain .
it is interesting to see the difference between the two theories in studying the coherent multi - spin dynamics in a small bath . for this purpose
, we consider a short ( @xmath156 ) `` random '' spin chain with accidental near - resonance between four neighboring spins .
the pair - flip between the four spins in near - resonance may develop up to four - spin cluster correlations as the evolution time goes beyond the pair - flip timescale @xmath152 .
indeed , fig .
[ g_compareran2](a ) shows coherent oscillations that are correctly reproduced by the cce at the 4th or higher - order truncation .
the correction by the 5th and higher - order clusters is actually negligible , verifying that the dominating cause of the coherent qubit oscillation is due to up to four - spin cluster dynamics in the bath .
in contrast , the ce , though already converges at the 4th - order truncation , does not reproduce the exact solution [ see fig .
[ g_compareran2](b ) ] , due to the neglect of the overlapping terms [ see eq .
( [ ce_overlap ] ) ] .
we have developed a cce approach to solving the many - body dynamics of a generic interacting spin bath relevant to the center spin decoherence problem . in this approach
, the bath propagator is factorized exactly into the product of cluster correlation terms , each of which accounts for the correlated flip - flops of a group of bath spins . in terms of the standard lce
, a cluster correlation term corresponds to the infinite summation of all the connected diagrams with all and only the spins in the cluster flip - flopped . for a finite - time evolution as in qubit decoherence
, a convergent result can be obtained by truncating the expansion up to a certain cluster size .
the cce gives exact results whenever it converges .
the lowest nontrivial order of the cce corresponds to the previously developed pair - correlation approximation . compared to the ce method ,
the two theories yield similar results for large spin baths , but for small spin baths only the cce accurately takes into account the multi - spin cluster correlations . as a simple method to sum over an infinite series of lce diagrams , the cce method can be readily applied to the case of interacting bosons or fermions .
this work was supported by hong kong rgc project 2160285 .
the lce for the propagator of an arbitrary spin bath in a thermal ensemble has been derived in ref .
[ ] . here we summarize the main results including the feynman diagram representation , as the basis of the extension to the lce for an arbitrary noninteracting bath state . for a thermal ensemble of spin baths characterized by the noninteracting density matrix @xmath164 with @xmath165 , the evolution in the contour time - ordered form is @xmath166 .
\label{propagator_ensemble}\]]here the contour time - dependent hamiltonian @xmath63 switches alternatively between @xmath64 and @xmath65 on the contour [ see fig .
[ g_contour ] ( a ) ] . as an example , the pairwise bath hamiltonian in eq .
( [ hpn ] ) leads to @xmath167where @xmath168 , @xmath169 , @xmath170 , and @xmath171 depend on which contour time segment @xmath172 is in , and the spin operators @xmath173 , @xmath174 are time - independent but are written with explicit time - dependence to keep track of the time - ordering .
( [ propagator_ensemble ] ) can be expanded into series as @xmath175 , \label{series_ensemble}\]]where each term is the sum of ensemble - averaged @xmath61-products of spin operators . consider the ensemble average @xmath176 $ ] of an arbitrary product @xmath177 , where the subscripts @xmath178 label bath spins and the superscripts @xmath179 label the spin operators . in order for @xmath180 not to vanish
, @xmath181 and @xmath182 operators must make pairs .
an @xmath183 consisting of @xmath107 operators only is called a fully contracted product , whose ensemble average is trivially evaluated .
so next we consider the case with at least one @xmath181 operator , @xmath184for the moment , we assume that the product in @xmath183 is already time - ordered with @xmath185 and use the following procedure to reduce @xmath180 to the sum of ensemble - averaged fully contracted products .
first we move the spin raising operator @xmath186 to the right , generating commutators between @xmath186 and the operators on its right , @xmath187 \cdots j_{m_{n}}^{\alpha _ { n}}(t_{n})\right ] \\ & + \mathrm{tr}\left [ \rho _ { 0}j_{m_{1}}^{\alpha _ { 1}}(t_{1})\cdots j_{m_{k-1}}^{\alpha _ { k-1}}(t_{k-1})j_{m_{k+1}}^{\alpha _ { k+1}}(t_{k+1})\cdots j_{m_{n}}^{\alpha _ { n}}(t_{n})j_{m_{k}}^{+}(t_{k})% \right ] .\end{aligned}\]]then in the second term , we use the cyclic invariance of the trace to move @xmath188 to the left ( before @xmath189 ) , and use @xmath190 to move @xmath188 across @xmath189 back to its original position to get @xmath191 . during this process , we get additional commutators @xmath192 \cdots j_{m_{n}}^{\alpha _ { n}}(t_{n})\right ] .$ ] collecting all terms , we obtain @xmath193 ^{\bullet } \left [ j_{m_{k}}^{+}(t_{k})\right ] ^{\bullet } \cdots j_{m_{n}}^{\alpha _ { n}}(t_{n})\right ] , \label{lemma_ensemblea}\]]where the contraction between a spin raising operator @xmath194 and an arbitrary spin operator @xmath195 is defined as @xmath196 ^{\bullet } \left [ j_{m}^{+}(t)% \right ] ^{\bullet } & \equiv \left [ j_{m}^{+}(t)\right ] ^{\bullet } \left [ j_{n}^{\alpha } ( t_{\alpha } ) \right ] ^{\bullet } \\ & \equiv \delta _ { m , n}g_{m}(t_{\alpha } -t)\left [ j_{n}^{\alpha } , j_{m}^{+}% \right ] ( t_{\alpha } ) , \end{aligned}\]]with @xmath197 + \theta ( -t)f(\omega _ { m})$ ] being a green s function and @xmath198 .
note that the contraction @xmath199^{\bullet } [ j_{m}^{+}(t)]^{\bullet } \sim \left [ j_{n}^{\alpha } , j_{m}^{+}% \right ] ( t_{\alpha } ) $ ] is still an operator associated with a contour time @xmath200 and should be used in subsequent contractions .
with @xmath201 = j^{+}$ ] and @xmath202 = -2j^{z},$ ] the contraction between @xmath194 and @xmath203 [ or @xmath204 eliminates @xmath194 and converts @xmath203 [ or @xmath204 to @xmath205 [ or @xmath206 , reducing the number of spin operators by one .
the resulting product in eq . ( [ lemma_ensemblea ] ) is still in a time - ordered sequence , so the time - ordering operator can be recovered so that @xmath207 \notag \\ & = \mathrm{tr}\left [ \rho _ { 0}\sum_{p(\neq k)}\mathscr{t}_{\mathrm{c}% } j_{m_{1}}^{\alpha _ { 1}}(t_{1})\cdots \left [ j_{m_{k}}^{+}(t_{k})\right ] ^{\bullet } \left [ j_{m_{p}}^{\alpha _ { p}}(t_{p})\right ] ^{\bullet } \cdots j_{m_{n}}^{\alpha _ { n}}(t_{n})\right ] .
\label{lemma_ensemble}\end{aligned}\]]in the above equation the time ordering @xmath185 is no longer assumed , since all the spin operators commute in the @xmath61-product . the contraction procedure in eq .
( [ lemma_ensemble ] ) can be repeated whenever there is still a spin raising operator left .
thus wick s theorem follows : under the average over a noninteracting thermal ensemble , a @xmath208-product of spin operators can be replaced by the sum of all possible fully contracted products which contains only @xmath107 operators . according to wick s theorem , each term in eq .
( [ series_ensemble ] ) generates a series of fully contracted products .
the fully contracted products can be visualized by feynman diagrams with the following definition of constituent elements and construction rules . 1 . a spin operator @xmath181 , @xmath182 , @xmath107 is represented by a vertex as a filled circle , an empty circle , or an empty square , respectively ; 2 .
each diagonal ( off - diagonal ) interaction term containing @xmath20 spin operators in the hamiltonian is represented by a dashed ( wavy ) interaction line connecting @xmath20 vertices ; 3 .
each contraction @xmath211^{\bullet}\left[% j_{m}^{\alpha}(t_{\alpha})\right]^{\bullet}$ ] is represented by a solid arrow starting from the vertex @xmath194 and ending at the vertex @xmath212 . at the end of the propagating arrow
the commutator @xmath213 $ ] is to be taken ; 4 .
each @xmath214 vertex denoted by a filled circle is connected with one outgoing propagating arrow , each @xmath215 vertex denoted by an empty square is either free - standing or connected to one incoming arrow [ converting @xmath215 to @xmath214 ] and one outgoing arrow [ from the resulting operator @xmath214 ] , and each @xmath217 vertex denoted by an empty circle is connected to one incoming arrow [ converting @xmath217 to @xmath218 or two incoming arrows [ the first arrow converting @xmath217 to @xmath219 , and the second arrow converting @xmath219 to @xmath220 and one outgoing arrow [ from the resulting @xmath214 ] . taking the hamiltonian in eq .
( [ htexample ] ) for example , the diagonal single - spin term @xmath221 the diagonal interaction @xmath209 and the off - diagonal interaction @xmath222 are visualized in fig . [ g_d12](a ) .
the first - order expansion @xmath223 in eq .
( [ series_ensemble ] ) gives two fully contracted products [ the first two diagrams in fig .
[ g_d12](a ) ] .
the second - order expansion of eq .
( [ series_ensemble ] ) gives four fully contracted products shown in fig .
[ g_d12](b ) .
the first three diagrams @xmath224come from diagonal terms and involve no contractions . here , as a convention
, we have suppressed the sum over spin indices and the contour time integrals .
the last diagram @xmath225 ^{\diamond } \left [ j_{n}^{-}(t_{1})\right ] ^{\bullet } \left [ j_{n}^{+}(t_{2})\right ] ^{\bullet } \left [ j_{m}^{-}(t_{2})% \right ] ^{\diamond } \label{secondorderexp}\end{aligned}\ ] ] comes from two off - diagonal interaction terms and involves two contractions @xmath226 ^{\diamond } \left [ j_{m}^{-}(t_{2})\right ] ^{\diamond } $ ] and @xmath227 ^{\bullet } \left [ j_{n}^{+}(t_{2})\right ] ^{\bullet } $ ] , as indicated by the two solid arrows in fig .
[ g_d12 ] ( b ) ( iv ) .
similarly , higher - order diagrams can be constructed by using the above feynman rules .
figure [ g_contractionexample ] gives two examples .
the 3rd - order diagram in fig .
[ g_contractionexample](a ) consists of two off - diagonal interactions @xmath228 and @xmath229 and one diagonal interaction @xmath230 it contains three contractions , one on spin @xmath231 and two on spin @xmath232 , corresponding to the three solid arrows .
the 4th - order one in fig .
g_contractionexample(b ) consists of four off - diagonal interactions and contains five contractions , three on spin @xmath233 and one on each of the other two spins . to illustrate the evaluation of the diagrams , we consider again the secular pair - interaction hamiltonian in eq .
( [ htexample ] ) as the example .
the rules for constructing the analytical formula for a diagram are : 1 . a contour - time dependent constant
is associated with each interaction line , namely , @xmath169 for an open - ended dashed line representing the spin splitting , @xmath170 for a dashed line connecting two diagonal spin operators representing the diagonal interaction , and @xmath171 for a wavy line representing the off - diagonal interaction ; 2 .
each solid arrow from @xmath194 to @xmath234 gives the green s function @xmath235 , each freestanding @xmath236 vertex gives @xmath237 , each @xmath238 vertex connected to one incoming arrow gives @xmath239 , and each @xmath238 vertex connected with two incoming arrows and one outgoing arrow gives @xmath240 ; 3 .
a global factor @xmath241 is associated with a diagram containing @xmath242 interaction lines ; 4 .
the spin indices are summed over and the contour times are integrated over .
summation of all the fully contracted diagrams leads to the lce of the ensemble - averaged evolution @xmath244where @xmath245 is the sum of all the _ connected _ diagrams , such as the first two diagrams in fig .
[ g_d12](a ) and the last diagram in fig .
[ g_d12](b ) , but does not include the disconnected ones such as the first three diagrams in fig .
[ g_d12](b ) . as an example , for the hamiltonian in eq .
( [ htexample ] ) , all the topologically inequivalent connected diagrams up to the 4th order are shown in fig .
[ g_d123 ] .
we see that the number of diagrams increases significantly with increasing perturbation order . for a single noninteracting bath state @xmath246 ( in contrast to the thermal ensemble )
, the lce for spin-1/2 baths has been outlined in ref .
here we reproduce the main results using the ensemble lce depicted in the previous subsection .
the key is that for a spin-1/2 bath , any noninteracting state @xmath247 can be taken as the ground state of a corresponding noninteracting hamiltonian , @xmath248where @xmath249 ( or @xmath250 ) for @xmath251 ( or @xmath252 ) .
so the noninteracting single - sample average @xmath253 becomes the zero - temperature limit ( @xmath254 ) of the corresponding noninteracting ensemble average @xmath255 with the density matrix @xmath256 .
in particular , the single - sample expectation value of the bath propagator @xmath257 .\]]thus the single - sample lce is obtained by simply setting the green s function @xmath258 and replacing the ensemble average @xmath259 with @xmath260 , i.e. , @xmath261note that the connected diagrams in @xmath262 contain only @xmath107 operators , which commute with each other .
thus the single - sample average can be performed for each diagram to convert it into a @xmath263-number , so that @xmath264 for a higher - spin bath , a noninteracting single sample state @xmath267 in general is not the eigen state of a noninteracting hamiltonian as in eq .
( [ h_j ] ) . thus the single - sample lce for higher - spin baths may not be derived from the ensemble lce directly .
here we provide a solution by mapping a higher spin to a composite of pseudo - spin-1/2 s .
the mapping is such that the physical states form an invariant subspace in the much larger pseudo - spin hilbert space . without loss of generality
, we consider a spin-1 @xmath233 .
the mapping from the spin-1 states to the states of two pseudo - spin-1/2 s @xmath268 ( @xmath269 ) is @xmath270as schematically shown in fig . [ g_spin1mapping ] .
the two pseudo - spin-1/2 s have more basis states than the spin-1 . to restrict the evolution of the spin state within the physical hilbert space , we map the spin-1 operators to pseudo - spin-1/2 operators as ( see fig .
g_spin1mapping ) @xmath271 in which the flip - flop of one pseudo - spin-1/2 is conditioned on the state of the other pseudo - spin-1/2 .
thus each spin-1 operator is mapped to an interaction term of the two pseudo - spin-1/2 s .
an interaction term between two spin-1 s would contain up to four pseudo - spin-1/2 operators , making the hamiltonian rather complicated .
nonetheless , the lce can be readily applied to the spin-1 bath mapped to a pseudo - spin-1/2 one . for a spin-1 hamiltonian as in eq .
( [ htexample ] ) , the different terms @xmath272 and @xmath273 converted to pseudo - spin-1/2 operators , are represented in turn by the three diagrams in fig .
[ g_spin1ld12](a ) . there
each vertex ( denoting a pseudo - spin-1/2 operator according to the same feynmann rule depicted in appendix [ append_lcegaussian ] ) is associated with a physical spin index @xmath274 and a pseudo - spin index @xmath275 .
the mapping leads to three different types of off - diagonal interactions ( including simultaneous interaction involving up to four pseudo - spin-1/2 s ) . as a result ,
the number of connected diagrams increases dramatically , e.g. , we have six topologically inequivalent second - order connected diagrams for spin-1 baths , as shown in fig .
g_spin1ld12(b ) , while for a spin-1/2 bath we have only one such diagram ( the third diagram in fig .
[ g_d123 ] ) . | decoherence of a center spin or qubit in a spin bath is essentially determined by the many - body bath evolution .
we develop a cluster - correlation expansion ( cce ) theory for the spin bath dynamics relevant to the qubit decoherence problem .
a cluster correlation term is recursively defined as the evolution of a group of bath spins divided by the cluster correlations of all the subgroups .
the so - defined correlation accounts for the authentic ( non - factorizable ) collective excitations within a given group .
the bath propagator is the product of all possible cluster correlation terms . for a finite - time evolution as in the qubit decoherence problem ,
a convergent result can be obtained by truncating the expansion up to a certain cluster size .
the two - spin cluster truncation of the cce corresponds to the pair - correlation approximation developed previously [ phys .
rev .
b * 74 * , 195301 ( 2006 ) ] . in terms of the standard linked cluster expansion , a cluster correlation term is the infinite summation of all the connected diagrams with all and only the spins in the group flip - flopped , and thus the expansion is exact whenever converges . when the individual contribution of each higher - order correlation term to the decoherence is small ( while all the terms combined in product could still contribute substantially ) , as the usual case for relatively large baths where the decoherence could complete well within the bath spin flip - flop time , the cce coincides with the cluster expansion [ phys .
rev .
b * 74 * , 035322 ( 2006 ) ] . for small baths ,
however , the qubit decoherence may not complete within the bath spin flip - flop timescale and thus individual higher - order cluster correlations could grow significant . in such cases ,
only the cce converges to the exact coherent dynamics of multi - spin clusters .
we check the accuracy of the cce in an exactly solvable spin - chain model . |
the gross - pitaevskii equation is the nonlinear schrdinger equation with an external potential : @xmath6 in @xcite we investigated the case of @xmath7 where @xmath8 and @xmath9 . because of our earlier work on high velocity scattering @xcite,@xcite , which is potential specific ,
the paper presented the case of @xmath10 , but as was pointed out there the method applies verbatim to the more general case , @xmath11 .
motivated by the approach of @xcite we now revisit the case of slowly varying potentials , @xmath12 , and show that using the effective hamiltonian approach we can describe the evolution of the soliton with errors of size @xmath2 . in particular , in this setting , we improve the results of frhlich - gustafson - jonsson - sigal @xcite .
[ t:1 ] suppose that in , @xmath13 , where @xmath14 .
let @xmath15 . then on the time interval @xmath16 , @xmath17 we have @xmath18 where @xmath19 , @xmath20 , @xmath21 , solve the following system of equations @xmath22 with initial data @xmath23 .
the constants @xmath24 and @xmath25 depend on @xmath26 , @xmath27 , and @xmath28 only . ] and initial condition in with @xmath29 , @xmath30 .
the dashed red curve shows the solution to newton s equations used in @xcite and @xcite , the blue curve shows the center of the approximate soliton @xmath31 , and the black dashed curve is given by the equations of motion of the effective hamiltonian @xmath32 the improvement of the approximation given by the effective hamiltonian is remarkable even in the case of @xmath33 in which we already see radiative dissipation in the first cycle.,width=576 ] as in @xcite the proof of our theorem follows the long tradition of the study of stability of solitons which started with the work of m.i .
weinstein @xcite .
the interaction of solitons with external potentials was studied in the stationary semiclassical setting by floer and a. weinstein @xcite and oh @xcite , and the first dynamical result belongs to bronski and jerrard @xcite , see also @xcite,@xcite .
the semiclassical regime is equivalent to considering slowly varying potentials , and the dynamics in that case was studied in @xcite,@xcite,@xcite,@xcite ( see also numerous references given there ) . the results of @xcite in the special case of give@xmath34 where @xmath35 with initial data @xmath23 . in other words ,
the motion of the soliton is approximately given by newton s equations , @xmath36 , @xmath37 .
however it is not clear if , as in , the errors @xmath38 in can be removed without affecting see [ ode ] . since imply equations , theorem [ t:1 ] shows that we can replace @xmath39 by @xmath40 in , keeping . when @xmath41 , and @xmath9 @xcite , newton s equations are clearly not applicable .
to describe evolution up to time @xmath42 we introduced a natural effective hamiltonian . a numerical experiment shown in ( * ? ? ?
* fig.2 ) and reproduced here in fig.[f : drama ] suggested that the effective hamiltonian approach gives a dramatic improvement for slowly varying potentials .
theorem [ t:1 ] ( and a more precise theorem [ t:2 ] below ) quantify that improvement by changing the errors from @xmath43 in to @xmath2 in . to describe the natural effective hamiltonian we recall that the gross - pitaevski equation is the equation for the hamiltonian flow of @xmath44 with respect to the symplectic form on @xmath45 ( considered as a real hilbert space ) : @xmath46 when @xmath47
, @xmath48 is a minimizer of @xmath49 with the prescribed @xmath50 norm ( @xmath51 ) : @xmath52 here @xmath53 is the differential of @xmath54 , see [ s : sy ] . the flow of @xmath49 is tangent to the manifold of solitons , @xmath55 which of course corresponds to the fact that the solution of with @xmath56 and @xmath57 , is @xmath58 the symplectic form restricted to @xmath59 is @xmath60 see [ ssms ] .
the evolution of the parameters @xmath61 in the solution follows the hamilton flow of @xmath62 with respect to the symplectic form @xmath63 .
the system of equations is obtained using the following basic idea : if a hamilton flow of @xmath64 , with initial condition on a symplectic submanifold , @xmath59 , stays close to @xmath59 , then the flow is close to the hamilton flow of @xmath65 . in our case
@xmath66 is the manifold of solitons and @xmath64 is given by @xmath67 the equations are simply the equations of the flow of @xmath68 see [ s : gph ] .
they are easily seen to imply but some @xmath69 corrections are built into the classical motion leading to the improvement in theorem [ t:1 ] , see also fig.[f : drama ] . as in previous works
all of this hinges on the proximity of @xmath70 to the manifold of solitons , @xmath59 . in @xcite
we followed @xcite and used weinstein s lyapunov function @xcite , @xmath71 where @xmath72 is given by . here ,
in the notation similar to , @xmath73 for an optimally chosen @xmath74 see lemma [ l : stan ] .
the use of @xmath75 seems essential for the all - time orbital stability of solitons . up to times
@xmath76 we found that it is easier to use its quadratic approxition @xmath77 rather than use conservation of energy , @xmath78 , we use the nonlinear equation for @xmath79 in estimating @xmath80 see [ pr ] . that involves solving a nonhomogeneous linear equation approximately using the spectral properties of @xmath81 see and proposition [ p : new ] .
the fact that the solution of that equation is of size @xmath2 in @xmath3 gives the first indication of the improvement based on using the effective hamiltonian .
ultimately , this makes the argument simpler than in the case of @xmath82 ( or the special case of @xmath83 ) .
the paper is organized as follows . in [ rhs ] we recall the hamiltonian structure of the nonlinear flow of and describe the manifold of solitons . as in @xcite , its identification with the lie group @xmath84 , where @xmath85 is the heisenberg group ,
provides useful notational shortcuts . in [ re ] we describe the reparametrized evolution .
the starting point there is an application of the implicit function theorem and a decomposition of the solution into symplectically orthogonal components .
that method has a long tradition in soliton stability and we learned it from @xcite .
the analysis of the orthogonal component using an approximate solution to a linear equation and a bootstrap argument are presented in [ pr ] .
this results in a somewhat more precise version of theorem [ t:1 ] see theorem [ t:2 ] . the ode estimates needed for the exact evolution
are given in
finally we show how theorem [ t:2 ] implies theorem [ t:1 ] . except for basic material such as properties of sobolev spaces or elementary symplectic geometry , and a reference to the proof of proposition [ p : coer ]
, the paper is meant to be self contained . acknowledgments .
the work of the first author was supported in part by an nsf postdoctoral fellowship , and that of the second second author by the nsf grant dms-0654436 .
in this section we recall the well known facts about the hamiltonian structure of the nonlinear schrdinger equation .
the manifold of solitons is given as an orbit of a semidirect product of the heisenberg group and @xmath86 . in our work
, we consider @xmath87 viewed as a _ real _ hilbert space .
the inner product and the symplectic form are given by @xmath88 let @xmath89 be a function , a hamiltonian .
the associated hamiltonian vector field is a map @xmath90 , which means that for a particular point @xmath91 , we have @xmath92 .
the vector field @xmath93 is defined by the relation @xmath94 where @xmath95 , and @xmath96 is defined by @xmath97 in the notation above @xmath98 if we take @xmath99 with the symplectic form , and @xmath100 then we can compute @xmath101 thus , in view of and , @xmath102 the flow associated to this vector field ( hamiltonian flow ) is @xmath103 for future reference we state two general lemmas of symplectic geometry . the simple proofs can be found in @xcite .
[ l : gen1 ] suppose that @xmath104 is a diffeomorphism such that @xmath105 , where @xmath106 .
then for @xmath107 , @xmath108 suppose that @xmath109 and that @xmath110 .
then the hessian of @xmath111 at @xmath112 , @xmath113 , is well defined .
we can identify @xmath114 with @xmath115 using the innner product , and define the hamiltonian map @xmath116 by @xmath117 in this notation we have [ l : gen2 ] suppose that @xmath118 is a finite dimensional symplectic submanifold of @xmath119 , and @xmath120 satisfies @xmath121 if at @xmath122 , @xmath123 , then the hamiltonian map defined by satisfies @xmath124 the hamiltonian map , @xmath125 , is simply the linearization of @xmath126 at a critical point of @xmath127 .
an example relevant to this paper is @xmath128 see .
the soliton @xmath129 is a critical point of @xmath72 and the hessian of @xmath130 is given by @xmath131 in .
the hamiltonian map @xmath132 is the linearization of @xmath126 at @xmath129 .
in other words , @xmath133 is the linearization of ( with @xmath56 ) at @xmath134 .
the @xmath135 term comes from @xmath136 in the definition of @xmath137 . in lemma
[ l : gen2 ] we can take @xmath138 to be the four dimensional manifold of solitons and @xmath139
. it then says that @xmath140 preserves the symplectic orthogonality of @xmath141 to @xmath142 . for @xmath143
we define the following map @xmath144 this action gives a group structure on @xmath145 and it is easy to check that this transformation group is a semidirect product of the heisenberg group @xmath85 and @xmath86 : @xmath146 explicitly , the group law on @xmath147 is given by @xmath148 where @xmath149 the action of @xmath147 is not symplectic but it is _ conformally symplectic _ in the sense that @xmath150 as is easily seen from .
the lie algebra of @xmath151 , denoted by @xmath152 , is generated by @xmath153 , @xmath154 and the bracket acts as follows : @xmath155=e_1 , \quad [ e_2,e_4 ] = -e_2 , \quad [ e_1,e_2]=-e_3 , \quad [ e_3,\bullet ] = 0 \,,\ ] ] so @xmath156 is in the center .
the infinitesimal representation obtained from is given by @xmath157 it acts , for instance on @xmath158 , and by @xmath159 we will denote a linear combination of the operators @xmath160 .
the proof of the following standard lemma can be again found in @xcite : [ l : stan ] suppose @xmath161 is a @xmath162 function and that @xmath163 . then , in the notation of , @xmath164 where @xmath165 is given by @xmath166 where @xmath167 .
the manifold of solitons is an orbit of this group , @xmath168 , to which @xmath169 , defined in , is tangent . in view of that
means that @xmath170 for some @xmath171 .
the simplest choice is given by taking @xmath172 , @xmath173 , so that @xmath129 solves a nonlinear elliptic equation @xmath174 this has a solution in @xmath3 if @xmath175 and it then is @xmath176 .
we will fix @xmath177 so that @xmath178 using lemma [ l : gen1 ] we can check that @xmath179 is the _ only _ orbit of @xmath147 to which @xmath169 is tangent .
we define the submanifold of solitons , @xmath180 , as the orbit of @xmath181 under @xmath182 , @xmath183 and thus we have the identifications @xmath184 the quotient corresponds to the @xmath185-action @xmath186 we first compute the symplectic form @xmath187 on @xmath142 using the identification : @xmath188 since @xmath189 we obtain from that @xmath190 and all the other @xmath191 s vanish . in other words , @xmath192 using we conclude that @xmath193 see ( * ? ? ?
* lemma 2.3 ) .
now let @xmath127 be a function defined on @xmath194 , @xmath195 .
the associated hamiltonian vectorfield , @xmath196 , is defined by @xmath197 using we obtain @xmath198 the hamilton flow is obtained by solving @xmath199 the restriction of @xmath200 to @xmath59 is given by computing by @xmath201 the flow of for this @xmath111 describes the evolution of a soliton .
we now consider the gross - pitaevski hamiltonian , @xmath202 and its restriction to @xmath203 : @xmath204 the flow of @xmath205 can be read off from : @xmath206 this are the same equations as .
the evolution of @xmath207 and @xmath208 is simply the hamiltonian evolution of @xmath209 , @xmath210 .
the more mysterious evolution of the phase @xmath211 is now explained by .
we can also rewrite as follows : @xmath212 so that @xmath213 and @xmath214 where we used that @xmath215 and @xmath216 in the taylor expansions . here for the purpose of presentation we assumed that @xmath217 but we will never use taylor s formula with more than four terms , for which @xmath218 is only needed .
to see the effective dynamics described in
[ s : gph ] we write the solution of as @xmath219 where @xmath220 satisfies @xmath221 to see that this decomposition is possible , initially for small times , we apply the following consequence of the implicit function theorem and the nondegeneracy of @xmath63 ( see ( * ? ? ?
* proposition 5.1 ) for a more general statement ) : [ l : fgjs ] for @xmath222 ( where the topology on @xmath223 is given by the identification with @xmath224 ) let @xmath225 if @xmath226 then for any @xmath227 , there exists a unique @xmath228 such that @xmath229 moreover , the map @xmath230 is in @xmath231 .
we define the transformation @xmath232 ( x ) { \stackrel{\rm{def}}{=}}\omega ( h \cdot u - \eta , x \cdot \eta ) \,,\ ] ] and want to solve @xmath233 for @xmath234 . by the implicit fuction theorem that follows for @xmath31 near @xmath168 if for any @xmath235 the linear transformation @xmath236 is invertible .
clearly we only need to check it for @xmath237 , that is that @xmath238 is invertible . but as an element of @xmath239 , @xmath240 , which is nondegenerate .
from [ s : sy ] and [ s : gph ] we recall that the equation for @xmath31 can be written as @xmath241 using lemma [ l : fgjs ] we define @xmath242 and we want to to derive an equation for @xmath243 .
let @xmath244 note that dependence on @xmath245 , @xmath246 makes @xmath247 , @xmath248 into time - dependent parameters .
they are however independent of @xmath249 . the proper motivation for the choice of @xmath250 and @xmath251 will come in lemma [ l : new ] below .
set @xmath252 where @xmath160 s are given by .
let also @xmath253 we now have [ l : wt ] the equation for @xmath254 , defined by , is @xmath255 & x\eta + i\big[-v\big(\frac{x}{\mu}+a\big)+\alpha + \beta x\big]\eta \\ & + xw+i\big[-v\big(\frac{x}{\mu}+a\big)+\alpha + \beta x\big]w -i\mu^2\mathcal{l}w + i\mu^2\mathcal{n}w\ , , \end{aligned}\ ] ] where @xmath256 is given by , @xmath257 by , @xmath258 by .
we compute , by the chain rule @xmath259 where @xmath260 from this , and the fact that the equation for @xmath261 can be written @xmath262 , we get @xmath263 we apply lemma [ l : gen1 ] to obtain @xmath264 we compute @xmath265 therefore , @xmath266 substituting and expanding the cubic term , @xmath267 & -y(\eta+w ) - \frac12iv^2(\eta+w ) + \frac12i\mu^2\partial_x^2(\eta+w ) + i\mu^2[\eta^3 + 2\eta^2w+\eta^2\bar w \\ & + 2|w|^2\eta+\eta w^2 + |w|^2w ] - iv\big ( \frac{x}{\mu}+a\big)(\eta+w ) - v\mu\partial_x(\eta+w ) \end{aligned}\ ] ] using that @xmath268 , we obtain @xmath269 & -y\eta - \frac12iv^2\eta+\frac12i\mu^2\eta - iv\big(\frac{x}{\mu}+a\big)\eta -v\mu\partial_x \eta\\ & -yw-\frac12iv^2w+\frac12i\mu^2w - i\mu^2\mathcal{l}w+i\mu^2\mathcal{n}w -iv\big(\frac{x}{\mu}+a\big)w - v\mu\partial_xw \end{aligned}\ ] ] now set @xmath270e_3\ ] ] from this , we get the claimed equation .
let us make some remarks about the lemma .
first , note that if @xmath271 , then @xmath272 which are exactly the equations of motion of the effective hamiltonian see
[ s : gph ] .
second , note that the term @xmath273\eta\ ] ] projects symplectically to @xmath274 ( used in the next lemma ) , in the sense that @xmath275 where @xmath276 is defined by the condition that @xmath277 see ( * ? ? ? * ( 3.7 ) ) for an explicit expression of @xmath278 .
to see we use the following simple [ l : new ] let @xmath278 be the symplectic projection defined above and @xmath279 ( that is @xmath111 is real valued tempered distribution ) . then , @xmath280 this follows from a straightforward calculation based on .
if @xmath281 , we obtain with @xmath250 and @xmath251 given by .
finally , note the following taylor expansions : @xmath282 and thus @xmath283 where all the errors are polynomially bounded in @xmath284 . if we assumed that @xmath285 then he expansions for @xmath250 would be valid with an error @xmath286 .
[ l : x ] let @xmath287 be given by , with @xmath288 obtained from lemma [ l : fgjs ] , and @xmath289 by .
suppose @xmath290 .
then @xmath291 we use the symplectic orthogonality of @xmath292 , @xmath293 and @xmath287 .
since @xmath294 , lemma [ l : wt ] gives @xmath295 we recall from ( * ? ? ?
* lemma 3.3 ) the following straightforward estimates : @xmath296 we already observed in that the first term on the right hand side vanishes . from we see that the second term is @xmath297 .
the third and fourth term are estimated using by @xmath298 the last term vanishes : the linear operator @xmath299 is the hessian of @xmath300 , given in , at the critical point @xmath129 . the fact that @xmath301 is tangent to @xmath66 and lemma [ l : gen2 ] ( or a direct computation ) show that @xmath302 summarizing , @xmath303 the smallness of @xmath304 concludes the proof .
we conclude this section which two lemmas which effectively eliminate @xmath305 from the coefficients of @xmath306 .
[ l : gath ] suppose that @xmath307 and that @xmath308 , for every @xmath309 .
then @xmath310 we first compute @xmath311 where we used the conservation of the @xmath50 norm . by the symplectic orthogonality assumption statement of the lemma @xmath312 and
hence @xmath313 from which the conclusion follows . from this
we immediately deduce the following [ c : nomus ] suppose @xmath314 and @xmath315 .
let @xmath316 where @xmath256 is given by : @xmath317 where @xmath250 and @xmath251 are given by ( and depend on @xmath305 ) . the the conclusions of lemma
[ l : x ] hold for @xmath318 : @xmath319
in this section we will recall the now standard estimates on the operator @xmath131 which arises as hessian of @xmath72 at @xmath129 : @xmath320 or @xmath321 in our special case we can be more precise than in the general case ( see @xcite , and also ( * ? ? ?
* appendix d ) ) .
the self - adjoint operators @xmath322 belong to the class of schrdinger operators with _ pschl - teller _ potentials and their spectra can be explicitly computed : @xmath323 the eigenfuctions can computed by the same method but a straightforward verification is sufficient to see that @xmath324 from @xcite we recall the following [ p : coer ] let @xmath307 and suppose that for any @xmath309 , @xmath308 . then , @xmath325 the next proposition will be useful in solving a linear equation for an approximation of @xmath287 . [
p : new ] the equation @xmath326 has a unique solution in @xmath327 . in addition , @xmath328 and @xmath329 since @xmath274 is an isolated point of the spectrum of @xmath330 and @xmath330 is self - adjoint , @xmath330 has a bounded inverse on @xmath331 hence , @xmath332 by elliptic regularity , @xmath127 is smooth .
moreover , since @xmath333 commutes with @xmath334 and @xmath335 is even , we conclude that @xmath127 is even .
next we argue that @xmath127 has exponential decay .
for that we conjugate the equation by @xmath336 : @xmath337 and apply both sides to @xmath338 : @xmath339 taking the fourier transform this means that we have @xmath340^{\widehat{\;}}(\xi),\ ] ] which makes sense if @xmath341 as then the right hand side is in @xmath50 .
since the multiplier on the left - hand side is bounded away from @xmath274 , we can invert the expression to obtain that @xmath342 . from this
, we deduce by sobolev embedding that @xmath343 . by applying derivatives to and then repeating this argument ,
we in fact obtain that all derivatives are pointwise exponentially localized in space . finally , we prove the orthogonality condition , that is , @xmath344 where @xmath160 s are given by .
since @xmath127 is real , we clearly have @xmath345 since @xmath127 is even , we also have @xmath346 .
it remains to show that @xmath347 , that is that @xmath348 .
note that @xmath349=\eta$ ] ( by direct computation or from the structure of the generalized kernel of @xmath350 ) .
hence @xmath351 \,dx = \int l_+f \ ; ( x\eta ) ' \,dx = \int \big(\frac{\pi^2}{12}+x^2\big)\eta ( \eta+ x\eta')\ , dx\,.\ ] ] integration by parts and @xmath352 , @xmath353 , show that this is @xmath274 .
in the arguments that follow , we will assume that @xmath354 satisfying is always defined on any time interval under consideration , and thus @xmath254 given by is also always defined .
this is , however , not known _ a priori _ and lemma [ l : fgjs ] must be considered as part of the bootstrap argument , together with the lemmas that follow .
however , since this aspect of the argument is standard in papers on this subject , we will not make mention of it
. recall from lemma [ l : wt ] that the equation for @xmath254 is @xmath355 & x\eta + i\big[-v\big(\frac{x}{\mu}+a\big)+\alpha + \beta x\big]\eta \\ & + xw+i\big[-v\big(\frac{x}{\mu}+a\big)+\alpha + \beta x\big]w -i\mu^2\mathcal{l}w + i\mu^2\mathcal{n}w\ , , \end{aligned}\ ] ] where @xmath250 and @xmath251 are time dependent parameters given by .
note that by taylor s theorem the forcing term in this equation has second - order expansion @xmath356\eta = -i\frac{v''(a)}{2\mu^2}\big(\frac{\pi^2}{12}+x^2\big)\eta + h^3{\langle}x{\rangle}^3\eta f_2\ ] ] where , provided @xmath357 , we have @xmath358 . we first consider the forced linear evolution obtained from by discarding all terms we expect will be of order @xmath359 or higher : @xmath360 we now introduce a natural approximate solution to this forced linear equation .
let @xmath361 where @xmath111 is given in proposition [ p : new ] .
then @xmath362 satisfies to second - order , i.e. @xmath363 where @xmath364\ ] ] [ l : thetabound ] suppose there is a constant @xmath365 such that @xmath366}^\infty h_x^1 } \leq c_1h^{3/2}\ ] ] then provided @xmath367 , we have @xmath368 where @xmath369 is a constant depending only upon @xmath365 , @xmath26 for @xmath370 and @xmath371 . by , lemma [ c : nomus ] and taylor expansions ( see ) ,
we have @xmath372 we have @xmath373 and thus , by and , @xmath374 integrating this bound , we obtain a near conservation of classical energy , @xmath375 from which we obtain @xmath376 .
this and imply that @xmath377 and thus @xmath378 .
this approximate solution @xmath362 provides heuristic evidence that @xmath254 should be of order @xmath379 , but it will also play a key rle in our rigorous argument establishing this fact . [
l : lyapunov ] suppose that , for some constant @xmath365 , @xmath380}^\infty h_x^1 } \leq c_1h^{3/2}\ ] ] then , provided @xmath381 } h^1 } \leq 4\sqrt{c_2}\|w(t_1)\|_{h^1 } + ch^2\ ] ] the constants @xmath369 in and depend upon @xmath365 , @xmath26 for @xmath370 , and @xmath371 , although are independent of @xmath382 .
the constant @xmath383 , with @xmath384 the absolute constant appearing in
. we will postpone the proof of lemma [ l : lyapunov ] to the end of the section . in the next theorem
we iterate the above bound , and close the bootstrap argument .
[ t:2 ] let @xmath385 .
let @xmath386 , and suppose that there is a constant @xmath365 such that @xmath387 then , provided @xmath388}^\infty h_x^1 } \leq 4\sqrt{c_2}h^{-2\delta}\|w_0\|_{h^1 } + ch^{2(1-\delta)}\ ] ] the constant @xmath369 in and and the constant @xmath389 in depend upon @xmath365 , @xmath26 for @xmath370 , and @xmath390 , although are independent of @xmath391 and @xmath392 .
the constant @xmath383 , with @xmath384 the absolute constant appearing in .
we apply lemma [ l : lyapunov ] @xmath393 times on successive intervals each of size @xmath394 ( where @xmath369 is as given in lemma [
l : lyapunov ] ) to obtain the bound @xmath395}^\infty h_x^1 } \leq ( 4\sqrt c_2)^k\|w_0\|_{h^1 } + \left ( \sum_{j=0}^{k-1 } ( 4\sqrt c_2)^j \right)ch^2\ ] ] this is only valid provided that the hypothesis of lemma [ l : lyapunov ] is satisfied over the whole collection of time intervals : @xmath396 by taking @xmath397 we obtain that @xmath398 thus , it suffices to ensure that @xmath399 and this is accomplished provided @xmath400 , for a suitable @xmath389 with the dependence as stated in the proposition .
we note that the constant @xmath369 from lemma [ l : lyapunov ] did not change from one iteration to the next , since @xmath371 remains uniformly bounded by lemma [ l : thetabound ] , which applies on a time interval of size @xmath401 .
we now conclude the section with the proof of lemma [ l : lyapunov ] . for the proof , we shall assume , in place of , the bound @xmath402}^\infty h^1 } \leq c_1h^{2(1-\delta')}\ ] ] we will conclude at the end that it suffices to take @xmath403 .
let @xmath404 where @xmath405 is an approximate solution to given by . from and , we derive the equation satisfied by @xmath406 : @xmath407 & -i\mu^2\mathcal{l}w_1 + x\eta + i\big[- v\big(\frac{x}{\mu}+a\big)+\alpha + \beta x + \frac{v''(a)}{2\mu^2}\big(\frac{\pi^2}{12}+x^2\big)\big]\eta - h^3\theta f\\ & + xw+i\big[-v\big(\frac{x}{\mu}+a\big)+\alpha + \beta x\big]w + i\mu^2\mathcal{n}w \ , .
\end{aligned}\ ] ] by grouping forcing terms of size @xmath408 , we rewrite the above as @xmath409w + i\mu^2\mathcal{n}w\ ] ] where @xmath410 ( uniformly in @xmath411 ) .
note that here we have applied lemma [ l : thetabound ] to conclude that @xmath378 .
we recall that @xmath412 is self - adjoint with respect to @xmath413 and hence , writing @xmath414 , we compute @xmath415 & -\mu^2{\langle}\mathcal{l}w_1 , i\mathcal{l}w_1{\rangle}+ { \langle}\mathcal{l}w_1,h^3f_1{\rangle}+ { \langle}\mathcal{l}w_1 , xw_1{\rangle}+ { \langle}\mathcal{l}w_1 , x\tilde w{\rangle}\\ & + { \langle}\mathcal{l}w_1 , i\big[-v\big(\frac{x}{\mu}+a\big)+\alpha + \beta x\big]w_1 { \rangle}\\ & + { \langle}\mathcal{l}w_1 , i\big[-v\big(\frac{x}{\mu}+a\big)+\alpha + \beta x\big]\tilde w { \rangle}+ \mu^2{\langle}\mathcal{l}w_1 , i\mathcal{n}w { \rangle}\end{aligned}\\ & = \text{i}+\text{ii}+\text{iii}+\text{iv}+\text{v}+\text{vi}+\text{vii}\end{aligned}\ ] ] now we analyse these terms one - by - one .
first , @xmath416 for ii , we use integration by parts for the @xmath417 term to move one @xmath418 onto @xmath419 and then apply the cauchy - schwarz inequality ; for the other terms we use a direct application of the cauchy - schwarz inequality , and together these give @xmath420 , requires more care : @xmath421 where @xmath422 , the components of @xmath289 , are time dependent but space independent . by the bootstrap assumption and lemma [ l : x ] , @xmath423 . to proceed , we do some further calculations : @xmath424 @xmath425 and thus the above two terms are bounded by @xmath426 . for the terms involving @xmath129
, we use the cauchy - schwarz inequality and the fact that @xmath427 : @xmath428 altogether then , we have @xmath429 for this term , we are forced to directly estimate by the cauchy - schwarz inequality ( although for the @xmath430 term , we integrate by parts one @xmath418 factor ) .
the bound that we get is @xmath431 from the definition of @xmath362 and proposition [ p : new ] , we obtain that all the norms involving @xmath362 are bounded by @xmath379 . thus , @xmath432w_1 { \rangle}\\ & = \frac 12 { \langle}(w_1-\partial_x^2w_1- 4\eta^2w_1 - 2 \eta^2\bar w_1 ) , i\big[-v\big(\frac{x}{\mu}+a\big)+\alpha + \beta x\big]w_1 { \rangle}\\ & = \frac 12 { \langle}(-\partial_x^2w_1-\eta^2\bar w_1 ) , i\big[-v\big(\frac{x}{\mu}+a\big ) + \alpha + \beta x\big]w_1 { \rangle}\ , .
\end{aligned}\ ] ] to estimate the first term , we integrate by parts and use that @xmath433 ( note that other estimates are available , like @xmath434 , but we do not want an @xmath249 coefficient here ) . for the second term , we use , the cauchy - schwarz inequality and the rapid decay of @xmath435 : @xmath436\eta^2 \big| \leq ch^2\ ] ] this gives the bound as a coefficient here , and thus this seems to be what limits us ultimately to time @xmath437 . ]
@xmath438\tilde w { \rangle}\ ] ] in the @xmath417 term of @xmath412 , we integrate by parts one @xmath418 , and then estimate by the cauchy - schwarz inequality .
all other terms , are estimated by a direct application of the cauchy - schwarz inequality .
the bound obtained is @xmath439\tilde w\right \|_{l^2 } \\
\notag & \leq ch^2\|w_1\|_{h^1}(\| { \langle}x{\rangle}^2\tilde w\|_{l^2 } + \|{\langle}x{\rangle}^2\partial_x\tilde w\|_{l^2})\\ \label{e : bdvi } & \leq ch^4\|w_1\|_{h^1}\end{aligned}\ ] ] by the localization of @xmath362 .
for the last term , vii , we use integration by parts once for the @xmath417 term , and then apply the cauchy - schwarz inequality to all terms . since we are in one - dimension , we have the embedding @xmath440 . @xmath441 @xmath442 by the bootstrap assumption .
this completes the step - by - step estimation process , and the bound we get from , , , , , , is @xmath443 we see that it suffices to take @xmath403 . integrating in time ,
we get @xmath444 & { \langle}\mathcal{l}w_1(t_1),w_1(t_1){\rangle}+ c(t - t_1)h^3\|w_1\|_{l_{[t_1,t_2]}^\infty h^1 } \\ & + c(t - t_1)h\|w_1\|_{l_{[t_1,t_2]}^\infty h^1}^2 \end{aligned}\ ] ] by and , @xmath445 satisfies the hypothesis of proposition [ p : coer ] , and we have @xmath446 by direct estimation , we have the upper bound @xmath447 combining this with we get the bound @xmath448 & 4c_2\|w_1(t_1)\|_{h^1}^2 + c(t - t_1)h^3\|w_1\|_{l_{[t_1,t_2]}^\infty h^1 } \\ & + c(t - t_1)h\|w_1\|_{l_{[t_1,t_2]}^\infty h^1}^2 \end{aligned}\ ] ] from this , we infer from the monotonicity of the right side that @xmath449}^\infty h^1}^2 \leq \begin{aligned}[t ] & 4c_2\|w_1(t_1)\|_{h^1}^2 + c(t_2-t_1)h^3\|w_1\|_{l_{[t_1,t_2]}^\infty h^1 } \\ & + c(t_2-t_1)h\|w_1\|_{l_{[t_1,t_2]}^\infty h^1}^2 \end{aligned}\ ] ] requiring that @xmath450 implies @xmath451}^\infty h^1}^2 & \leq 8c_2\|w_1(t_1)\|_{h^1}^2+ch^2\|w_1\|_{l_{[t_1,t_2]}^\infty h^1}\\ & \leq 16c_2\|w_1(t_1)\|_{h^1}^2+ch^4 \ , .
\end{split}\ ] ] since @xmath452 and @xmath453 , @xmath454 } h^1 } \leq 4\sqrt{c_2}\|w(t_1)\|_{h^1 } + ch^2 \,,\ ] ] which is the claimed estimate .
to pass from an approximate equations for the parameters of the soliton , @xmath455 given in lemma [ c : nomus ] , to the odes we need some elementary estimates which we present in this section .
they are similar to those in @xcite .
[ l : odecompare ] suppose that @xmath456 , and @xmath457 , @xmath458 , @xmath459 , @xmath460 are @xmath461 real - valued functions .
suppose @xmath462 is a @xmath463 mapping such that @xmath464 and @xmath465 are uniformly bounded .
suppose that on @xmath466 $ ] , @xmath467 let @xmath468 and @xmath469 be the @xmath461 real - valued functions satisfying the exact equations @xmath470 with the same initial data .
suppose that on @xmath466 $ ] , we have @xmath471 for @xmath472 . then provided @xmath473 , we have on @xmath466 $ ] the estimates @xmath474 before proceeding to the proof , we recall some basic tools . _ gronwall estimate_. suppose @xmath475 and @xmath476 are @xmath461 real - valued functions , @xmath477 is a constant , and @xmath478 satisfy the differential inequality : @xmath479 let @xmath480 , @xmath481 . then @xmath482 let @xmath483 .
then @xmath484 , and hence @xmath485 .
thus @xmath486 converting from @xmath487 back to @xmath478 , we obtain the gronwall estimate @xmath488 _ duhamel s formula_. for a two - vector function @xmath489 , a two - vector @xmath490 , and a @xmath491 matrix function @xmath492 , let @xmath493 denote the solution to the ode system @xmath494 with @xmath495 : @xmath496 then , for a given two - vector function @xmath497 , the solution to the inhomogeneous ode system @xmath498 with initial condition @xmath499 is given by duhamel s formula @xmath500 let @xmath501 and @xmath502 ; these perturbative functions satisfy @xmath503 where @xmath504 is given by @xmath505 which is @xmath461 ( in particular , uniformly bounded ) .
set @xmath506 in , and appeal to duhamel s formula to obtain @xmath507 apply the gronwall estimate with @xmath508 to conclude that @xmath509 feed this into to obtain that on @xmath466 $ ] @xmath510 taking @xmath511 , we obtain the claimed bounds .
we can now put all the components of the proof together .
lemma [ c : nomus ] and theorem [ t:2 ] show that on the time interval @xmath512 we have with the parameters satisfying @xmath513 lemma [ l : odecompare ] can be applied to replace @xmath207 and @xmath208 with solutions of and the direct integration of the error terms shows that the same is true for @xmath305 and @xmath211 . in particular we can drop @xmath305 altogether .
we conclude the paper with some remarks .
the proof above and theorem [ t:2 ] show that the conclusions of theorem [ t:1 ] remain unchanged if instead of taking @xmath514 as initial condition , we took @xmath515 we could go down to @xmath516 at the expense of complicating the final statement to .
a more general condition on the initial value would make the bootstrap argument in [ pr ] so unwieldy that we opted out of pursuing that technical issue . in higher dimensions
similar methods are clearly applicable for weaker nonlinearities and under further spectral assumptions
see @xcite for examples . at this early stage
we restrict ourselves to the physically relevant cubic nonlinearity which at the moment is tracktable only in dimension one . | we study the gross - pitaevskii equation with a slowly varying smooth potential , @xmath0 .
we show that up to time @xmath1 and errors of size @xmath2 in @xmath3 , the solution is a soliton evolving according to the classical dynamics of a natural effective hamiltonian , @xmath4 .
this provides an improvement ( @xmath5 ) compared to previous works , and is strikingly confirmed by numerical simulations
see fig.[f : drama ] . |
many different mathematical models have been proposed to study an evolutionary self - governing system . in the last several decades , a particular attention was devoted to those that are based on simple generating rules that produce complex outcomes .
such an example is the growing model based on the numerical @xmath3-rule introduced in @xcite @xmath6 where @xmath7 .
the numbers are recorded in cells and , for simplicity , we keep the unidirectional development of future generations , composed of children @xmath8 born from parents @xmath9 and @xmath10 , which are neighbor cells in the previous generation . for a plastic representation of the @xmath3-rule , one can think that any cell containing a positive integer @xmath11 is a citadel composed of towers .
there are as many towers in the citadel as prime factors @xmath11 has .
each tower is associated to a prime and the height of the tower corresponding to a prime @xmath12 that divides @xmath11 equals the power of @xmath12 in the factor decomposition of @xmath11 . in particular , the citadel of a cell containing the number @xmath13 has no towers at all .
likewise , one may think that the citadel @xmath11 has towers associated to the primes that do not divide @xmath11 also , but these towers have zero height .
then , the @xmath3-rule topples the towers of the neighbor citadels @xmath9 and @xmath10 creating a new citadel @xmath8 in the next generation .
the towers of the new citadel have heights equal with the absolute difference of the heights of towers corresponding to the same prime in @xmath9 and @xmath10 and , if a prime divides only one of @xmath9 and @xmath10 , then this tower is reproduced unchanged in the new citadel .
the process starts with a sequence of numbers @xmath2 , which may be finite or not , which are placed in a row of cells .
this sequence is called the initial generation and the @xmath3-rule is applied sequentially on each two consecutive terms of @xmath2 .
whence , a new generation is born and its cells are placed in the following row . usually , in graphic representations , we slightly shift to the right the new generation such that any new cell is placed in the middle under its parents . repeating the process
, we obtain a matrix @xmath14 with infinitely many rows if @xmath2 is infinite . as the reproducing rule remains unchanged , the results depend only on the initial generations and we shall see that in this way a large variety of outcomes are produced .
the matrices of numbers @xmath14 have lots of features of which some are similar to the objects created by the abelian sandpile model proposed by bak , tang and wiesenfeld @xcite .
the intensely studied model , also called the chip - firing game , was surveyed by levine and propp @xcite .
our @xmath3-model also captures features of other evolutionary systems such as the ducci - type game @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , the numerical ensembles created by median insertions , such as those related to pascal triangle @xcite @xcite , @xcite , @xcite , @xcite , @xcite , @xcite or the farey sequences @xcite , @xcite . for many initial generations @xmath2 ,
the matrices @xmath14 show complex self - similar structures , like those of some particular abelian sandpile states @xcite,@xcite , @xcite,@xcite or the outcomes produced in the related rotor - router model @xcite , @xcite , @xcite .
a special feature of matrices @xmath4 is the fact that they can be localized . for any prime @xmath12 , the _
@xmath12-tomography _ is the matrix of citadels of @xmath4 in which all towers , except the towers associated to @xmath12 , are deleted .
then @xmath14 is the superposition ( the element - wise multiplication ) of the @xmath12-tomographies for all primes @xmath12 , since the evolution according to the @xmath3-rule is independent to one another .
we have already proved in @xcite that the @xmath3-rule produces objects with a fractal aspect if @xmath2 is the sequence of prime numbers or its localized slice , the sequence of zeros except one term that is equal to a prime @xmath12 . in this article
we show that this also happens if the initial generation is the @xmath12-spaced sequence @xmath15 , where @xmath16 or @xmath17 , the @xmath12-section of positive integers , @xmath18 where @xmath19 is the @xmath12__-valuation _ _ of @xmath11 , which is , the power of @xmath12 into the prime decomposition .
the classic sierpinski fractal appears if @xmath20 ( a finite fragment is shown in figure [ figtomo2 ] ) , while more complex self - similar patterns are typical for larger primes ( see figures [ figtomo13 - 19][figtomo5 ] ) .
always , in a graphical representation , we present only a triangular region of @xmath4 , the one composed by the cells born in future generations from the part of @xmath2 shown on the first row .
@xmath21{$$ } } & { \makebox[1.05em]{$$ } } & { \makebox[1.05em]{$$ } } & { \makebox[1.05em]{$$ } } & { \makebox[1.05em]{$$ } } & { \makebox[1.05em]{$$ } } & { \makebox[1.05em]{$$ } } & { \makebox[1.05em]{$$ } } & { \makebox[1.05em]{$$ } } & { \makebox[1.05em]{$$ } } & { \makebox[1.05em]{$$ } } & { \makebox[1.05em]{$$ } } & { \makebox[1.05em]{$$ } } & { \makebox[1.05em]{$$ } } & { \makebox[1.05em]{$$ } } & { \makebox[1.05em]{$$ } } & { \makebox[1.05em]{$$ } } & { \makebox[1.05em]{$$ } } & { \makebox[1.05em]{$$ } } & { \makebox[1.05em]{$$ } } & { \makebox[1.05em]{$$ } } & { \makebox[1.05em]{$$ } } & { \makebox[1.05em]{$$}}\\ \textbf{1 } & & 2 & & 3 & & 4 & & 5 & & 6 & & 7 & & 8 & & 9 & & 10 & & 11 & & 12 & & \cdots\\ & \textbf{2 } & & 6 & & 12 & & 20 & & 30 & & 42 & & 56 & & 72 & & 90 & & 110 & & 132 & & \cdots & \\ & & \textbf{3 } & & 2 & & 15 & & 6 & & 35 & & 12 & & 63 & & 20 & & 99 & & 30 & & \cdots & & \\ & & & \textbf{6 } & & 30 & & 10 & & 210 & & 420 & & 84 & & 1260 & & 1980 & & 330 & & \cdots & & & \\ & & & & \textbf{5 } & & 3 & & 21 & & 2 & & 5 & & 15 & & 77 & & 6 & & \cdots & & & & \\ & & & & & \textbf{15 } & & 7 & & 42 & & 10 & & 3 & & 1155 & & 462 & & \cdots & & & & & \\ & & & & & & \textbf{105 } & & 6 & & 105 & & 30 & & 385 & & 10 & & \cdots & & & & & & \\ & & & & & & & \textbf{70 } & & 70 & & 14 & & 462 & & 154 & & \cdots & & & & & & & \\ & & & & & & & & \textbf{1 } & & 5 & & 33 & & 3 & & \cdots & & & & & & & & \\ & & & & & & & & & \textbf{5 } & & 165 & & 11 & & \cdots & & & & & & & & & \\ & & & & & & & & & & \textbf{33 } & & 15 & & \cdots & & & & & & & & & & \\ & & & & & & & & & & & \textbf{55 } & & \cdots & & & & & & & & & & & \\ & & & & & & & & & & & & { \boldsymbol\cdots } & & & & & & & & & & & & \\ \end{array}\ ] ] a fundamental problem raised by the @xmath3-model concerns the shape and properties of the matrix grown from the first generation @xmath22 . its north - west corner is shown in figure [ figuretnns ] .
the @xmath12-tomographies of @xmath5 can be grown individually , by starting in the first row with the sequence @xmath23 .
of particular interest are questions related to the geometrical figures propagated from the cell @xmath24 . for any prime @xmath12 and integer @xmath25 ,
we denote by @xmath26 the collection of connected cells that starts from @xmath27 and contains only powers of @xmath12 larger than two .
we call these figures _ @xmath3-solitons _ and present two of them in figure [ figsolitons ] . unlike the set of cells containing powers of @xmath12 less than two , which forms a continuous texture all over the infinite matrix @xmath5 , the solitons are larger and larger with the power @xmath28 , but finite .
[ conjecturesolitons ] two distinct solitons @xmath29 , @xmath30 , neither overlap nor touch each other .
the characterization problem of the evolution of shape and size of the solitons for different @xmath12 and growing @xmath28 is the analog of the limiting shape problems studied in @xcite , @xcite , @xcite , @xcite , @xcite .
the solitons @xmath31 verify conjecture [ conjecturesolitons ] , as follows from their complete description presented in section [ sectionsoliton2 ] . for odd primes , @xmath26 are more complex , with an aspect of certain fringed parts of sierpinski triangles . for a fixed prime @xmath12 ,
the series of solitons @xmath26 , as @xmath28 increases , offers an intriguing spectacle and their growth appears to be proportional , as in the analog case of the abelian sandpiles @xcite , @xcite , @xcite , @xcite .
another zone of interest is the first column or the _ west edge _ of matrix @xmath4 .
we denote this sequence by @xmath32 .
the western edge can be viewed as the projection of @xmath2 through the entire @xmath3-process .
notice that the value of the @xmath33-th citadel on @xmath32 is influenced by the values of all first @xmath33 citadels of @xmath2 and by neither of the others .
an example is the western frontier of the triangle in figure [ figuretnns ] : @xmath34 various evidence obtained by computer verifications suggest that no square of any prime divides a term of this sequence . in other words , no soliton grows as large as to touch the western edge @xmath35 .
corollary [ corolarry2 ] shows that this is true for solitons @xmath31 , @xmath25 . at the exponents level , this is the counterpart of the gilbreath s conjecture ( * ? ? ?
* a10 ) , @xcite , which refers to the similar process that starts with the sequence of primes as @xmath2 and grows the future generations with children born by taking the absolute difference of their parents .
the gilbreath s conjecture says that the west edge of the triangle composed of these rows of successive gaps of gaps , contains only ones .
another example extending this widespread property @xcite is the matrix that starts with the sequence of sophie germain primes is a sophie germain prime if @xmath12 and @xmath36 are primes at the same time . ] and is generated in the same way , listing successively the gaps from the previous row of gaps .
for this matrix , john w. layman ( * ? ? ?
* a080209 ) observed and conjectured that the left edge consists only of @xmath13s in our multiplicative setting , we conjecture that the maximal power of any prime that appears in the decomposition of the numbers situated on the left edge of @xmath37 is one .
the left edge of the infinite triangle generated by the iterated application of the @xmath3-rule to the set of positive integers contain only square free numbers .
the object of sections [ sectionpp][sectionwestpp ] is to compare and analyze the similarities between @xmath5 and an analogue matrix that has no solitons .
the @xmath12-tomographies of this new matrix are generated by sequence @xmath38 and in theorem [ theoremperiodicity ] we show that these tomographies are eventually periodic for all @xmath12 . furthermore , we observe the closeness of the citadels on the western side of the two matrices and in theorem [ theoremwpp2lap ] we characterize the structure of the sinuous series of extreme values of the western edge of the matrix with no solitons .
theorem [ theoremtnnsdoi ] gives a complete characterization of the @xmath39-tomography of matrix @xmath5 .
in particular , it shows that there are no fours on the western edge of @xmath5 , solving a problem raised by n. j. a. sloane ( * ? ? ? *
a222313 ) .
our analysis in sections [ sectionpp][sectionwestpp ] allows us to formulate the precise conjecture [ conjecture3 ] regarding another problem raised by sloane ( * ? ? ?
* a222313 ) , ( * ? ? ?
* question 3 ) .
starting with a sequence of integers @xmath40 , we consider the matrix whose first row is @xmath41 and the following ones are generated with the @xmath3-rule .
we denote this matrix by @xmath42 , where @xmath43 , @xmath44 and @xmath45 if the initial sequence is a finite ordered set @xmath46 , we obtain the numerical triangle @xmath47 whose first row is @xmath43 , @xmath48 , and following ones are generated recursively by @xmath49 we say that @xmath50 , for @xmath51 , is the _ child _ of its _ parents _ @xmath52 and @xmath53 and in pictures we usually place the child in the middle , below its parents .
the @xmath54-th row of the matrix is called the @xmath54-th _ generation _ and we denote it by @xmath55 we denote the _ west - side _ of the triangle by @xmath56 the evolution at the exponents level is presented into the following tables : @xmath57 given an infinite matrix @xmath58 or a bounded triangle @xmath59 , we denote their @xmath12-_tomography _ ( also called the @xmath12__-slice _ _ or the @xmath12__-section _ _ ) by @xmath60 thus , the superposition of all @xmath12-slices recovers the full matrix : @xmath61 where the product over all primes @xmath12 is taken component - wise . for any positive integer @xmath11 ,
we denote by @xmath62 the largest square free number that divides @xmath11 , and by @xmath63 the sequence of these numbers : @xmath64 we denote by @xmath65 $ ] the ring of meromorphic series of variable @xmath66 and coefficients in the field with two elements @xmath67 and by @xmath65^*\subset { \mathbb f}_2[[x]]$ ] the collection of series that are sums of monomials @xmath68 with @xmath69 , only . as usual ,
the number of distinct prime factors of @xmath11 is denoted by @xmath70 and the notation for the multiplicative order of @xmath9 modulo @xmath12 is @xmath71 .
the real action on @xmath72 is on the exponents level and , to understand its result , we need to formalize it .
let @xmath73 denote the collection of nonempty finite words over the infinite alphabet @xmath74 .
we introduce the following sequence of words in @xmath73 , defined recursively : @xmath75 where `` @xmath76 '' denotes the concatenation of integers .
note that @xmath77 is the concatenation of @xmath78 integers . since @xmath77 is an initial sub - word of @xmath79 , for all @xmath80
, there exists a limit sequence @xmath81 , whose first @xmath82 terms coincides with the letters of @xmath77 , for @xmath80 .
we write : @xmath83 . similarly , starting with @xmath13 instead of @xmath84
, we define the sequence of words @xmath85 and obtain the limit sequence @xmath86 , whose first @xmath82 terms coincides with the letters of @xmath87 , for @xmath80 .
the first terms of @xmath88 and @xmath89 are : @xmath90 for a given sequence @xmath91 , we denote by @xmath92 _ the sequence of absolute differences _ between consecutive terms : @xmath93 and by @xmath94 _ the bubbled sequence _ ,
defined by @xmath95 we use the same notations for the similar operations applied on words , where the action is on the the sequences of their letters .
for example : @xmath96 and @xmath97 . for @xmath20 .
notice that the larger solitons are further and further away from the western edge of @xmath37 .
+ , scaledwidth=78.0% ] the next lemma shows how @xmath88 and @xmath89 relates to one another through these operations .
[ lemmaw01 ] the following properties hold true : 1 .
@xmath98 , with element - wise addition ; 2 .
@xmath99 , with term - wise application of the valuation @xmath100 ; 3 .
@xmath101 ; 4 .
@xmath102 .
\(1 ) the equality @xmath98 follows directly from the definitions .
\(2 ) we prove the equality by induction .
the initial step : @xmath103 . now
, suppose that @xmath104 for all @xmath105
. then @xmath106 and @xmath107 , for all @xmath108 , by the definition of the valuation . using the definition of @xmath88
, this means that @xmath104 , for @xmath109 , as needed .
\(3 ) @xmath101 follows from ( 1 ) and ( 2 ) .
\(4 ) the equality @xmath110 also follows directly from ( 1 ) and the definition of @xmath111 .
so it remains to prove that @xmath112 .
we proceed by induction .
the initial step : applying @xmath111 to the finite sequence @xmath113 ( which are the letters of @xmath114 , the beginning of @xmath89 ) , we get @xmath115 , the beginning of @xmath116 or , with the notation on words , @xmath117 .
the induction step : suppose that @xmath118 .
then @xmath119 since the last and the first letters of @xmath120 are equal to @xmath13 .
this completes the proof of the lemma .
thus , by lemma [ lemmaw01 ] we see that the sequences of gaps between consecutive terms of @xmath88 and @xmath89 both coincide with the bubbled sequence @xmath121 with the above notations , we see that the sequence of exponents of @xmath39 on the first row of @xmath72 coincides with @xmath88
. then , by lemma [ lemmaw01 ] , it follows that the subsequent sequences of exponents of @xmath39 on the following rows of @xmath72 are : @xmath122 in general , the @xmath33th row of @xmath72 is @xmath123 now we can describe the structure of the matrix of the exponents @xmath124 , which corresponds explicitly to the explicit description of the @xmath39-tomography of @xmath5 .
its initial cut - off triangle @xmath125 , composed of @xmath126 rows , is shown in figure [ figtomo2 ] .
we see that , geometrically , it is part of an infinite sierpinski triangle .
notice that the horizontal rows are grouped naturally in _ slices _ containing sequences of pairs of triangles .
the couple of triangles in each pair is colored with the same color and the change of colors from a pair to another corresponds to the change of numbers in the the sequence of slices @xmath127 in which @xmath124 is partitioned are larger and larger in size .
the slice @xmath128 is just the first row and it is exceptional .
the next slice , @xmath129 , is the second row .
then , for any @xmath130 , the slice @xmath131 groups @xmath132 rows , those from the @xmath133th till the @xmath134th .
in any slice , the largest triangles formed by cells of the same color are the top rows of the pascal arithmetic triangle modulo @xmath39 , with the odd entries replaced by a certain positive integer .
such a triangle depends on two parameters : the height @xmath135 and the weight @xmath136 , which is the value of the non - zero entries .
we denote it by @xmath137 ( see the left triangle in figure [ figurep2ht ] for such an example ) .
triangle @xmath137 is generated as pascal s classic triangle , by starting from the top with a symbolic variable @xmath136 , which satisfies the rule @xmath138 .
the same result is obtained if the top is placed somewhere in a row of zeros ( see the matrix from the right - side of figure [ figurep2ht ] ) .
@xmath139 [ figurep2710 ] @xmath140 [ figurep25 t ] [ lemmap2 t ] let @xmath135 be a positive integer and let @xmath136 be a formal variable .
let @xmath141 be a sequence of zeros , except one term @xmath142 and suppose that @xmath143 .
then , the matrix with rows @xmath144 contains triangle @xmath137
. it suffices to note that condition @xmath145 ensures that the object that develops from @xmath142 is not influenced by external obstacles and the operation of taking the absolute value of the difference acts on @xmath146 exactly as the operation that grows a pascal triangle with entries in @xmath67 .
we summarize the complete description of @xmath124 ( respectively @xmath72 ) in the next theorem .
[ theoremtnnsdoi ] ( 1 ) slice @xmath128 of the matrix @xmath124 is sequence @xmath88 .
the next rows of @xmath124 are grouped in slices @xmath131 , such that , for any @xmath69 , slice @xmath131 is formed by rows from the @xmath133th till the @xmath134th .
( 2 ) the single row of @xmath129 is sequence @xmath147 .
( 3 ) for any @xmath69 , the collection of non - zero elements in slice @xmath131 is the union of triangles @xmath148 and the sequence of their weights ( from left to right ) coincides with @xmath116 .
the top vertices of these triangles are on the first row of the slice and their bases are adjacent and partition the bottom row .
\(1 ) follows by the definitions .
( 2 ) is proved in lemma [ lemmaw01 ] .
( 3 ) the proof is by induction .
the initial step , @xmath149 , coincides with ( 2 ) .
suppose now that the stated description is valid for slice @xmath131 and let as look on @xmath150 .
we begin with the first row of @xmath150 . here , the first @xmath151 cells are zeros , because , by the induction hypothesis , on the previous row , the first @xmath152 cells where the adjacent bases of two triangles @xmath153 .
the next element , the @xmath152th , is @xmath154 , since on the previous slice , the weight of the second triangle was @xmath13 and the weight of the third triangle is @xmath39 . continuing in the same way
, we see that the next non - zero cell on the first row of slice @xmath150 is the @xmath155th and its value is equal with @xmath156 . in this way
we see that the non - zero cells on the first row of slice @xmath150 are those obtained as absolute differences of the parent cells that are vertices of neighbor triangles @xmath148 with different weights from the previous slice .
these are the @xmath157th , the @xmath158th , the @xmath159th , and so on . moreover , by the induction hypothesis , the values of integers occupying these cells are the integers in the sequence @xmath116 .
now , using lemma [ lemmap2 t ] , we find that from each of these cells , grows a triangle @xmath160 .
moreover , the weights of these triangles coincides with the integers on the non - zero cells in the first row of @xmath150 .
also , the size of the slice assures that the bases of these triangles are adjacent .
this concludes the proof of the induction step and of the theorem . in particular , theorem [ theoremtnnsdoi ] describes the western edge of the matrix @xmath72 .
[ corolarry1 ] the @xmath39-valuation of the elements of the west sequence @xmath35 are : @xmath161 also , theorem [ theoremtnnsdoi ] answers to a question of n. j. a. sloane , who asks whether there is a proof that @xmath162 can not appear on the western edge of the matrix @xmath5 ( * ? ? ?
* a222313 ) .
[ corolarry2 ] there is no @xmath162 on @xmath163 .
the powers of all primes in the decomposition of the terms in the sequence @xmath165 defined in are equal to one .
this allows us to employ operations in the the ring of meromorphic series @xmath65 $ ] to understand the structure of the @xmath12-tomographies of @xmath164 .
thus the initial generation of any @xmath12-tomography is sequence @xmath15 defined by .
the superposition ( component - wise multiplication ) of the @xmath12-tomographies for all @xmath12 gives the full description of matrix @xmath164 . for any prime @xmath12 , we look at matrix @xmath166 .
always , the first row is filled with zeros except the cells in the arithmetic progression @xmath167 , @xmath69 , which are equal to @xmath13 .
again , the prime @xmath20 is exceptional .
the second row of @xmath168 has all cells equal to @xmath13 and from the third row on , matrix @xmath168 is filled with zeros only .
we show that if @xmath12 is odd , the rows can be grouped in periodic slices .
the number of rows in such a slice is a _ period _ of @xmath166 and we denote by @xmath169 the length of the smallest period . if @xmath20 , the periodic slices contain just one row , which repeats from the third on .
if @xmath12 is odd , the first row of the first periodic slice is always the second row of @xmath166 .
one can check the small periods for some primes : @xmath170 , @xmath171 , @xmath172 , @xmath173 , @xmath174 , @xmath175 . as @xmath12 increases ,
the size of @xmath169 becomes large : @xmath176 , @xmath177 , @xmath178 , @xmath179 .
this fact produces the general aspect of randomness of @xmath166 , for @xmath180 .
also , in some areas this phenomenon is more pronounced than in others ( see figure [ figtomo13 - 19 ] ) .
any sequence @xmath181 can be identified uniquely with a series in @xmath65^*$ ] .
we use this identifications for the rows of the matrix @xmath182 and write @xmath183 the operation of passing from one generation to the next by applying the @xmath184-rule transfers on the side of the series to multiplication by @xmath185 .
this may produce a series in @xmath65\setminus{\mathbb f}_2[[x]]^{*}$ ] and we need to bring it back by dropping the meromorphic part and the constant term through the @xmath186 operation : @xmath187\,,\quad \text{for $ s(x)=\sum c_kx^k\in{\mathbb f}_2[[x]]$}\ , .
\end{split}\ ] ] then , to pass from the @xmath54th generation to the @xmath188th , we have to multiply repeatedly @xmath33 times by @xmath189 , so the general correspondence is @xmath190 where @xmath191 is the absolute value of the differences ( which , in this case , coincides with addition in @xmath67 ) , taken component wise . next
we show that this association is well defined .
[ propositioncommutes ] the above association between the rows of the matrix @xmath166 and the series in @xmath65^*$ ] and the operation of passing from one generation to the next is well defined , and diagram is commutative .
well defining is due to the correspondence between the absence of columns to the left of the first column of @xmath166 , so there is no influence from the left when @xmath191 is applied , and from ignoring of the meromorphic and constant terms of the series using the dropping function @xmath186 .
it remains to prove that diagram is commutative by induction . the initial step : suppose @xmath192 .
then , on the one hand , we have : @xmath193 and on the other hand @xmath194 the outcomes of and are identical , so the initial step is completed .
the induction step follows by using the associative property of the composition of functions @xmath186 and multiplication by @xmath189 and the fact that @xmath195 .
this completes the proof of the proposition .
let us see the periodicity of the matrix @xmath166 in two particular cases .
the first row of @xmath196 contains only zeros , except the cells with ones in the columns with ranks in the arithmetic progression @xmath197 .
the series that corresponds to the second row is @xmath198 then the series that corresponds to the @xmath199th row is @xmath200 comparing and , we see that the @xmath39nd and the @xmath199th rows coincide . therefore , the matrix @xmath196 is eventually periodic and @xmath201 is the length of a period .
the series that corresponds to the second row of @xmath202 is @xmath203 we take advantage of the fact that @xmath204 is a special number and all the binomial coefficients @xmath205 , @xmath206 , are odd . then @xmath207 dropping the meromorphic and the constant term , we find that @xmath208 which , compared with shows that the @xmath39nd row coincides with the @xmath209th .
thus the matrix @xmath202 is eventually periodic and @xmath204 is the length of a period .
one can check that there is no shorter period . for this
, it suffices to calculate the terms from the first column of the matrix @xmath202 .
the first @xmath209 of them are : @xmath210 for a general @xmath12 , we can also take advantage of the fact that there are integers @xmath212 for which all the binomial coefficients @xmath213 , @xmath214 are odd . such integers do exist , as follows from the next simple lemma .
[ lemmabinomodd ] for any integer @xmath215 , we have @xmath216 then , by lemma [ lemmabinomodd ] and the definition of the binomial coefficients , we see that @xmath217 another requirement for @xmath212 is to be divisible by @xmath12 .
a minimal value of @xmath11 for which @xmath218 is divisible by @xmath12 is @xmath219 .
( we denote by @xmath71 the smallest integer @xmath220 for which @xmath221 . )
now let @xmath222 for some integer @xmath223 .
the series associated to the second row of @xmath166 is @xmath224 to get the series @xmath225 corresponding to the @xmath226th row , we have to multiply @xmath227 by @xmath228 .
first , let us see that @xmath229 then @xmath230 here the meromorphic and constant terms occur only in the first sum .
dropping them , we arrive at @xmath231 in conclusion , we have proved the following theorem .
[ theoremperiodicity ] for any prime @xmath232 , the rows of the matrix @xmath166 are eventually periodic .
the pre - period contains only the first row of the matrix and the length of the smallest period is a divisor of @xmath233 .
we remark that @xmath233 is not always the size of the smallest period .
for example , if @xmath234 , @xmath235 and @xmath236 , but the length of the smallest period is @xmath237 . also ,
if @xmath238 , @xmath239 and @xmath240 , but the length of the smallest period is @xmath241 . as well , if @xmath242 , @xmath243 and @xmath244 . in this case
, again , the length of the smallest period is shorter , @xmath245 .
there are two classes of primes : the first one , for which the length of the period of @xmath166 is maximal ( that is , @xmath233 ) and the second one , for which the length of the period is strictly smaller than @xmath233 .
we do not know if either one or both of these classes contain infinitely many primes .
the reason for the shorter periods in these cases are the arithmetic properties that produce favorable patterns in the series of binomial coefficients .
thus , writing the binomial coefficients @xmath246 , as concatenated letters of a word and the repeated letters as powers , for @xmath247 , they are : @xmath248 and for @xmath249 , they are : @xmath250 a related pattern appears if @xmath242 . in this case @xmath251 and the word defined by the binomial coefficients is @xmath252 where @xmath253 and @xmath254 .
calculations using power series from @xmath65 $ ] allows us to quickly find a particular element of matrix @xmath164 . in particular
, we can find the `` extreme values '' of @xmath256 , @xmath257 .
they emerge on the @xmath33th row of @xmath164 in places where number @xmath33 , when written in base two , has either few or many ones , compared with the rank of rows in its neighborhood .
one can notice this property in the augmented oscillations of both graphs in figure [ figlop ] .
in the following , we present the concrete structure of the most pronounced extremes , the values of @xmath256 , with @xmath33 around powers of two . we have to consider only the influence of odd primes , since @xmath20 is involved only on the first two rows of @xmath164 .
let @xmath259 .
by we know that @xmath260 , for @xmath261 . to find the series associated to the @xmath262th row of the @xmath12-topography of @xmath164 , we have to multiply : @xmath263 then , @xmath12 divides @xmath258 if and only if the coefficient of @xmath66 in series is odd .
we see that primes @xmath264 are not involved and @xmath265 is impossible . for
any small primes @xmath266 , denote by @xmath267 the largest integer for which there exist integers @xmath268 , such that @xmath269 & \ \ , \vdots\\[-0.35em ] 1&=\lambda p+s_{\lambda}-m.\\ \end{split}\ ] ] note that @xmath270 exists and @xmath271 .
then , monomial @xmath66 appears in the @xmath65 $ ] series if and only if @xmath270 is odd .
for example , if @xmath272 , by a simple investigation we find that @xmath273 is even and @xmath274 is odd for @xmath275 , so @xmath276 . in the same way ,
if @xmath277 , we see that @xmath278 is even only for @xmath279 , so @xmath280 .
let @xmath282 with @xmath25 .
in @xmath65 $ ] we have @xmath283 finding the series that corresponds to the @xmath284th row of @xmath164 implicates the calculation : @xmath285 again , we have to look for terms whose power of @xmath66 is equal to one . in series , the terms corresponding to primes
@xmath264 do not contribute to the coefficient of @xmath66 . also , @xmath12 can not be equal to @xmath33 , because @xmath33 is even . if @xmath286 is a mersenne prime , then @xmath265 is equal with this prime
. then @xmath287 , so @xmath12 divides @xmath281 . for the remaining primes @xmath288 , let @xmath289 be the the maximal number of equalities
@xmath290 & \ \ , \vdots\\[-0.35em ] 1&=\mu p+t_{\mu}-m , \end{split}\ ] ] where @xmath291 are even numbers that belong to @xmath292 .
notice that @xmath293
. then monomial @xmath66 effectively appears in series if and only if @xmath294 .
therefore @xmath295 if and only if @xmath296 is odd .
examples : if @xmath297 , we find that @xmath298 ; @xmath299 ; and @xmath300 , so @xmath301 . if @xmath302 , @xmath265 is a mersenne prime .
for the smaller primes , we find that @xmath303 ; @xmath304 ; @xmath305 ; and @xmath306 .
this implies that @xmath307 .
let @xmath309 .
the smaller numbers on the west side of @xmath164 appear on the rows of rank @xmath310 .
this is due to the fact that @xmath311 in @xmath65 $ ] , that is , the binomial @xmath312 has fewest possible terms .
then , the series that correspond to the @xmath313th row sums the terms of positive powers of @xmath66 from the following @xmath314 for a given @xmath232 , on the right - hand side of may appear a single monomial @xmath66 , and this happens whenever there exists an integer @xmath223 , such that @xmath315 .
this implies that the only prime divisors @xmath12 of @xmath308 are those for which if @xmath316 .
for example , @xmath317 ; @xmath318 ; @xmath319 and @xmath320 ( because @xmath321 ) ; @xmath322 ( because @xmath323 ) .
other terms of sequence @xmath255 may be calculated in the same way .
a few more examples are listed in table [ tablewest2lag ] .
@xmath324 we collect the results from sections [ subsection1]-[subsection3 ] into the next theorem .
[ theoremwpp2lap ] let @xmath25 and let @xmath325 and @xmath326 be the integers defined by and .
then @xmath327 [ paradox1 ] explain why @xmath281 is larger than @xmath258 , even if in the definition of @xmath296 , in equalities , an extra parity restriction on numbers @xmath328 is imposed ( condition that is absent for the existence of numbers @xmath329 in ) .
removing duplicates and ordering @xmath255 , we obtain sequence @xmath330 this is related and has terms close to those of the analogues sequence ( * ? ? ?
* a222313 ) , ( * ? ? ?
* question 3 ) obtained by starting with the initial generation @xmath22 instead of @xmath63 .
a complete discussion based on the previous analysis might give a complete argument for the certainty of the ranks of terms in list .
by theorem [ theoremperiodicity ] it follows that sequence @xmath331 , the west edge of the matrix @xmath166 , is also periodic , for any odd prime and the pre - period contains only the first term of the sequence .
we do not know whether there is a prime @xmath12 for which the length of the period of @xmath331 is strictly smaller than the length of the period of @xmath166 .
if there is such a prime , then it should be larger than @xmath332 . comparing the general aspect of the @xmath12-tomographies of @xmath5 and @xmath164
, one can observe both similarities and significant differences .
thus , on the one hand , although there are more and more irregularities in @xmath166 as @xmath12 increases , it is still eventually periodic . on the other hand
, a big noise grows under the cells with larger and larger powers of @xmath12 , if the initial generation is @xmath22 .
the most noticeable difference is if @xmath333 , since @xmath168 has only zero - cells from the third row on , while @xmath124 sprouts the triangles in figure [ figtomo2 ] . for small powers of @xmath334 and @xmath335 ,
the results are shown side by side in figures [ figtomo3 ] and [ figtomo5 ] .
although the noise is transmitted till the west edge , it does not cover it completely .
the beginning of sequences @xmath35 and @xmath255 are : @xmath336 and @xmath337 they are equal in @xmath338 places , at terms of indices @xmath339 .
as far as we can check , this semblance remains valid , suggesting a general behavior . even in places where they differ , the terms are very close , both in size and in arithmetic structure .
compare figures [ figlon ] and [ figlop ] to see more similarities of sequences @xmath340 and @xmath255 .
more precisely , the closeness between the two sequences can be measured by the surplus number of prime factors between @xmath341 or @xmath342 and their greatest common divisor , @xmath343 , @xmath257 . for this , the appropriate counting functions are @xmath344 in table [ tablesurplus ] we counted the number of integers @xmath33 for which the surplus occurs .
notice that if @xmath345 , the largest surplus is @xmath346 .
this is small when compared with the maximum values of @xmath347 and @xmath348 in this range , which are equal to @xmath349 and @xmath350 .
@xmath351 equality between @xmath341 and @xmath352 for @xmath345 occurs @xmath353 times
. we also mention that even in a larger range , integers @xmath33 for which @xmath354 tend to appear in clusters , often grouping a varying number of consecutive numbers . an intricate pattern of the sets of primes that divide the larger values of @xmath355 around @xmath309 develops as @xmath28 increases .
let us see a typical example , the case @xmath356 . to emphasize the presence or absence and the position of prime divisors in the list all primes @xmath357 , we have listed them all , but in two distinguished ways .
thus @xmath355 is the product of primes written in normal font , while the primes that do not divide @xmath355 are shown in red color ( in the electronic form ) smaller font .
thus , we have : @xmath358 w_{{\mathbb{n}^*}}(256):\ & 2 , { \mbox{{\color{red}\footnotesize 3 } } } , 5 , 7 , 11 , { \mbox{{\color{red}\footnotesize 13 } } } , 17 , 19 , 23 , { \mbox{{\color{red}\footnotesize 29 , 31 , 37 , 41 } } } , 43 , 47 , { \mbox{{\color{red}\footnotesize 53 , 59 , 61 } } } , 67 , 71 , 73 , 79 , 83,\\ & { \mbox{{\color{red}\footnotesize 89 , 97 , 101 , 103 , 107 , 109 , 113 , 127 } } } , 131 , 137 , 139 , 149 , 151 , 157 , 163 , 167,\\ & 173 , 179 , 181 , 191 , 193 , 197 , 199 , 211 , 223 , 227 , 229 , 233 , 239 , 241 , 251\ , , \end{split}\ ] ] so @xmath359 and @xmath360 .
n. j. a. sloane ( * ? ? ?
* a222313 ) , ( * ? ? ?
* question 3 ) orders increasingly and eliminates duplicates from the terms of @xmath35 and obtains sequence and @xmath361 from the list of terms of size smaller than @xmath362 . ] @xmath363 he asks if the first part of the list contains all numbers @xmath364 that appear in this sequence .
the numbers from are obtained from the first @xmath365 terms of @xmath35 . examining the terms
, we observed a general formula for a numbers that make a big jump in the beginning , during the process of ordering .
[ conjecture3 ] for any integer @xmath366 , we have : @xmath367 where @xmath368 is the largest square that divides @xmath310 . for small ranks , if @xmath369 , conjecture [ conjecture3 ] verifies , since @xmath310 is prime , for @xmath370 , or a product of two distinct primes , for @xmath371 , and @xmath372 in these cases . for the remaining values
, we have : @xmath373 ; @xmath374 ; @xmath375 .
verifying the decomposition of a few hundred more numbers of the form @xmath310 and assuming , confer the above discussion , that the smallest local minimums of @xmath35 are attained at these ranks , we should expect a positive answer to sloane s question .
[ paradox2 ] explain the peculiarity : why , given that the towers of the @xmath12-tomographies are higher when starting with the initial generation @xmath22 instead of @xmath63 , more non - zero cells appear on the western edge in the second case . for example , counting only terms less than @xmath376 , we find @xmath377 terms in @xmath378 and @xmath379 terms in @xmath380 .
all calculations and images presented in this work were made using the free open - source mathematical software system @xcite .
a. granville , _ arithmetic properties of binomial coefficients . i. binomial coefficients modulo prime powers _ , organic mathematics ( burnaby , bc , 1995 ) , 253276 .
. proc . * 20 * , amer .
providence , ri , 1997 .
m. prunescu , _ sign - reductions , @xmath12-adic valuations , binomial coefficients modulo @xmath384 and triangular symmetries _ , preprint 2012 .
https://www.academia.edu/3136109/sign-reductions_p-adic_valuations_binomial_coefficients_modulo_p_k_and_triangular_symmetries[electronic version ] t. sadhu , d. dhar , _ pattern formation in growing sandpiles with multiple sources or sinks _ , j. stat .
* 138 * ( 2010 ) , no . 4 - 5 , 815837 .
w. stein et al . , _ sage mathematics software ( version 6.8 ) _ , the sage development team , 2015 . | we present an evolutionary self - governing model based on the numerical atomic rule @xmath0 , for @xmath1 positive integers . starting with a sequence of numbers , the initial generation @xmath2 ,
a new sequence is obtained by applying the @xmath3-rule to any neighbor terms .
likewise , applying repeatedly the same procedure to the newest generation , an entire matrix @xmath4 is generated .
most often , this matrix , which is the recorder of the whole process , shows a fractal aspect and has intriguing properties . if @xmath2 is the sequence of positive integers , in the associated matrix remarkable are the distinguished geometrical figures called the @xmath3-solitons and the sinuous evolution of the size of numbers on the western edge .
we observe that @xmath5 is close to the analogue free of solitons matrix generated from an initial generation in which each natural number is replaced by its largest divisor that is a product of distinct primes .
we describe the shape and the properties of this new matrix .
n. j. a. sloane raised a few interesting problems regarding the western edge of the matrix @xmath5 .
we solve one of them and present arguments for a precise conjecture on another . |
let us shortly remind the main constructions of works @xcite , @xcite , @xcite .
let there be given the universal covering of an oriented three - dimensional manifold , considered as a simplicial complex , in which every simplex is mapped into the three - dimensional euclidean space ( so that the intersections of their images in @xmath0 are possible ) . to each edge
is assigned thus a euclidean length , and each tetrahedron has a sign `` plus '' or `` minus '' depending on orientation of its image in @xmath1 ; the defect angle ( minus algebraic sum of dihedral angles ) at each edge is equal to zero modulo @xmath2 .
then we consider infinitesimal translations of vertices of the complex which lead to infinitesimal changes of edge lengths , while these latter lead , in their turn , to the infinitesimal changes of defect angles .
thus , the following sequence of vector spaces arises : @xmath3 here @xmath4 is the lie algebra of motions of three - dimensional euclidean space ; @xmath5 is the space of column vectors of vertex coordinates differentials ; @xmath6 is the space of column vectors of edge length differentials ; @xmath7 is the space of column vectors of differentials of defect angles .
the superscript @xmath8 means matrix transposition .
matrices @xmath9 , @xmath10 and @xmath11 consist of partial derivatives .
one can see @xcite that the sequence ( [ eq : acycl1 ] ) is an algebraic complex ( superposition of two next matrices is equal to zero ) . besides , it is possible to show @xcite that for manifolds with finite fundamental group this complex is acyclic . for lens spaces we consider two sets of representations of fundamental group ( recall that @xmath12 ) : @xmath13 the realization of representations @xmath14 was described in detail in @xcite ( there
they were denoted @xmath15 ) .
as to representations @xmath16 , we will describe them below when we consider the structure of spaces entering in complex ( [ eq : acycl1 ] ) .
) , width=207,height=302 ] in figure [ fig : complex ] we show the simplicial complex ( bipyramid ) corresponding to the lens space @xmath17 .
let index @xmath18 ( in figure ) vary from @xmath19 to @xmath20 .
then gluing together @xmath21 bipyramids numbered by @xmath18 by identical vertices , edges and @xmath22-simplices we obtain an image of simplicial complex for @xmath23 in @xmath0 .
every element of group @xmath24 can be represented in the group of motions of three - dimensional euclidean space as a rotation through angle @xmath25 around some axis which we denote @xmath26 .
the integer parameter @xmath27 numbers the characters of these representations .
thus , in the right cartesian coordinate system @xmath28 the coordinates of vertices of simplicial complex look like ( cf . @xcite ) : @xmath29 where @xmath30 , @xmath31 , @xmath32 , @xmath33 are some parameters that fix , to within the motions of @xmath0 , the location of vertices and hence all geometrical values of the complex such as edge lengths , volumes of tetrahedra , dihedral angles , etc .
for instance , a simple calculation shows that the oriented volumes of tetrahedra are determined in this case by the formula : @xmath34 let us write out now more explicitly some of vector spaces entering in ( [ eq : acycl1 ] ) ( due to the symmetry of the complex , it is enough for us to deal only with the first three of them ) : @xmath35 where @xmath36 and @xmath37 are infinitesimal translations and rotations in the described above coordinate system @xmath28 .
every element of group @xmath24 acts on spaces @xmath5 and @xmath6 by cyclic shifts as follows : it sends subspaces @xmath38 and @xmath39 to @xmath40 and @xmath41 respectively .
index @xmath42 changes from @xmath19 to @xmath20 .
so , these actions completely determine representations @xmath16
. subspaces @xmath38 and @xmath39 look like ( all indexes are taken modulo @xmath21 ) : @xmath43 in general , matrices @xmath9 , @xmath10 and @xmath11 specifying the mappings of vector spaces in complex ( [ eq : acycl1 ] ) are not block - diagonal . in order to make them such we perform unitary change - of - basis transformations in all spaces entering in the complex .
our goal is to split ( [ eq : acycl1 ] ) into the direct sum of subcomplexes each of which would correspond to the character of some representation @xmath16 ( see ( [ eq : gj ] ) ) .
the unitarity of transformations is necessary to preserve the symmetric form ( [ eq : acycl1 ] ) of the complex .
let @xmath44 denote the primitive root of unity of degree @xmath21 .
then * the change - of - basis matrix in space @xmath4 is @xmath45 * the change - of - basis matrix in space @xmath5 is @xmath46 where @xmath47 * the change - of - basis matrix in space @xmath6 is @xmath48 where @xmath49 is the identical matrix of size @xmath50 .
now , in an obvious way , @xmath51 and each of matrices @xmath9 , @xmath10 and @xmath11 becomes block - diagonal ( the rearrangement of rows or / and columns can be required ) .
thus , the acyclic complex ( [ eq : acycl1 ] ) splits into the following direct sum of @xmath21 acyclic subcomplexes : [ eq : acycl2 ] @xmath52 the superscript @xmath53 means operation of hermitian conjugation .
we define the torsions of complexes ( [ eq : acycl2 ] ) by formulas : @xmath54 \bigl|(\det { } _ { { \mathcal d}_j}{\vert } b_j { \vert}_{{\overline { \mathcal c}}_j } ) \bigr| ^2 \,(\det { } _ { { \mathcal c}_j}{\vert } a_j ) ^ { - 1 } , & j = 2 , \ldots , p - 2 .
\end{cases}\ ] ] here @xmath55 is a maximal subset of edges such that the restriction of matrix @xmath56 onto its corresponding subspace of @xmath6 ( we denote it as @xmath57 ) is nondegenerate .
@xmath58 is the complement of @xmath55 .
sets @xmath59 and @xmath60 consist of vertices and are defined similarly . on the base of torsions ( [ eq : torsions ] )
, one can construct @xcite invariant values not changing under the pachner moves @xmath61 and @xmath62 . in our case , these invariants take the form : @xmath63 here @xmath64 are usual euclidean edge lengths , @xmath65 are volumes of tetrahedra given by ( [ eq : volumes ] ) .
we use definition @xmath66 to emphasize that , in contrast with work @xcite , here we deal with nontrivial representations @xmath14 .
the calculations implied by formula ( [ eq : invs1 ] ) are as follows . using acyclicity of complexes
, we choose subsets @xmath55 and @xmath59 .
then we calculate the corresponding minors of matrices @xmath56 , @xmath67 and @xmath68 .
in particular , we use the known expressions for partial derivatives @xmath69 ( see @xcite for details ) . after this
, we find the torsions according to ( [ eq : torsions ] ) and substitute them into ( [ eq : invs1 ] ) .
the computer calculations which have been carried out for first several @xmath21 s allow us to suppose that the general formula for invariants of lens spaces ( [ eq : invs1 ] ) looks as follows : [ eq : invs2 ] @xmath70 where we have denoted @xmath71
let us note once again that the value @xmath72 determined in ( [ eq : deltam ] ) is nothing but the module of @xmath18th component of the reidemeister torsion of @xmath17 .
it is of course very intriguing to calculate invariants of such kind for other 3- and 4-manifolds . in the future we plan to investigate similar invariants based on @xmath73-solution of the pentagon equation @xcite , @xcite .
after that , we plan to move on from lens spaces to manifolds with noncommutative or / and infinite fundamental group . i. g. korepanov , e. v. martyushev , _ distinguishing three - dimensional lens spaces @xmath74 and @xmath75 by means of classical pentagon equation _ , j. nonlin .
phys . , 2002 , vol .
9 , 8698 , ` arxiv : math.gt/0210343 ` e. v. martuyshev , _ euclidean simplices and invariants of three - manifolds : a modification of the invariant for lens spaces _ , to appear in proceedings of chelyabinsk scientific center , ` arxiv : math.at/0212018 ` i. g. korepanov , _ euclidean 4-simplices and invariants of four - dimensional manifolds : ii .
an algebraic complex and moves @xmath76 _ , theor . math
. phys . , 2002 , vol .
133 , 13381347 , ` arxiv : math.gt/0211166 ` | this article is a continuation of work on construction and calculation various of modifications of invariant based on the use euclidean metric values attributed to elements of manifold triangulation .
we again address the well investigated lens spaces as a standard tool for checking the nontriviality of topological invariants . |
at present we have three scattering theories : 1 .
theory of spherical harmonics ( sht ) used for elastic scattering ( see , for example @xcite ) ; 2 .
more general theory of van hove correlation functions , which we call standard scattering theory ( sst ) ( see , for example @xcite ) ; 3 . the fundamental scattering theory ( fst ) ( see , for example @xcite ) , which is used mainly for justification of sst .
the first one is inconsistent , the second one is not a theory but only a set of rules for calculation of cross sections , and the third one contains an error , so it does not justify sst .
the error is hidden in definition of scattering probability .
it relates initial wave packet of the incident neutron to outgoing plane wave of the scattered neutron .
such a definition completely violates unitarity .
the unitarity is conserved when both initial and final particles are plane waves or wave packet , but in both cases fst gives not cross sections but only probabilities .
we applied the fundamental theory of plane waves scattering ( fpwst ) ( without generally used artificial finite volume @xmath0 ) to neutron scattering in monatomic gas , and found that the result is ambiguous , which is very easily checked .
the probability amplitude @xmath1 of the neutron scattering on an atom is ( see section 4 ) @xmath2 where @xmath3 are neutron s and @xmath4 are atom s initial and final momenta , @xmath5 is ratio of the neutron s @xmath6 to the atomic @xmath7 masses .
the amplitude ( [ intr ] ) is the product of scattering amplitude @xmath8 and two @xmath9-functions related to the momentum and energy conservation laws .
it is simplified after integration over @xmath10 , which eliminates the first @xmath9-function , and is reduced to @xmath11 where @xmath12 , @xmath13 , and @xmath14 .
expressions ( [ intr ] ) and ( [ intr1 ] ) are very natural , however the two ways of probability calculation : one directly in laboratory reference frame ( lf ) , and another one via center of mass ( cm ) reference frame give two different results .
[ [ in - lf ] ] in lf + + + + + we directly integrate over @xmath15 and get @xmath16 , where @xmath17 is neutron scattering angle , and @xmath18 .
with @xmath19 the probability @xmath20 of scattering in lf is @xmath21 .
[ [ in - the - standard - approach ] ] in the standard approach + + + + + + + + + + + + + + + + + + + + + + + + ( look , for instance @xcite ) we transform ( [ intr1 ] ) to cm , i.e. change variable : @xmath22 , integrate over @xmath23 , find @xmath24 of neutron scattering into the cm scattering angle @xmath25 , and obtain probability @xmath26 .
transformation of this probability back to lf gives @xmath27 .
[ [ the - ambiguity - is ] ] the ambiguity is : + + + + + + + + + + + + + + + + + @xmath28 .
now we present these calculations in details , but first we want to stress , that the direct calculation , which is the main new point , does not include change of reference frames , while all the calculations with change of frames lead to the standard result .
so it is not profitable to seek for errors in change of reference frames . since @xmath29
, we directly integrate @xmath9-function in ( [ intr1 ] ) over @xmath15 and get @xmath30 where @xmath31 is the total momentum of the neutron and atom , @xmath32 and @xmath33 is a unit vector in the scattering direction of @xmath34 .
substitution of ( [ intr2 ] ) into ( [ intr1 ] ) gives @xmath35 and the probability of scattering into @xmath36 is @xmath37 after backward transformation ( [ intr2 ] ) one restores @xmath9-function of energy conservation and obtains @xmath38 @xmath39 where @xmath40 , @xmath41 was replaced with @xmath42 , and we used relation @xmath43 we see , that the scattering probability ( it is obtained directly in the lf , without change to other reference frames ) depends on 3 variables @xmath44 , @xmath45 and @xmath46 contrary to the wide spread belief that it depends only on two variables @xmath44 and @xmath45 . the argument of the @xmath9-function can be represented as @xmath47 where @xmath31 is the total momentum of the cm , and @xmath48 , later denoted by @xmath49 , is the relative speed of the neutron and atom .
after change of variables @xmath50 we obtain @xmath51 integration of the @xmath9-function over @xmath52 gives @xmath53d^3k_c= \frac{qd\omega_c}{2(1+\mu)^2 } , \label{intr7b}\ ] ] and ( [ intr1 ] ) is reduced to @xmath54 the scattering probability from an atom with momentum @xmath55 is @xmath56 the backward transformation ( [ intr7b ] ) restores @xmath9-function @xmath57 and with account of ( [ intr1a ] ) we get @xmath58 , leads to the standard cross section @xmath59 ) over @xmath15 we obtain @xmath60 both ( [ intr11 m ] ) , ( [ intr6 ] ) and ( [ intr12 ] ) , ( [ intr5 ] ) are different .
the fpwst , if we accept it notwithstanding of the ambiguity ( it will be later eliminated ) , gives probabilities and not cross sections .
to see how to match them to experiment , we in the next section discuss what do experimentalists really measure , and how to get a cross section from probability .
this discussion inevitably leads to a wave packet description of free neutron .
in the third section we show what are the drawbacks of all the three theories pointed above , in fourth section we derive scattering amplitude ( [ intr ] ) and demonstrate how to get the standard cross section for neutron scattering in monatomic gas . in 5-th section some wave packets and their properties
are discussed . in particular , we consider scattering of a wave packet from a fixed center and show that in linear theory probability of scattering does not depend on impact parameter , so to get a cross section from probability we need a nonlinear interaction . in 6-th section some ways to resolve contradictions and ambiguities are proposed , in section 7 some experiments to investigate wave packet properties are discussed , and in conclusion the results of the paper are summarized .
almost all experiments ( exceptions are reflectometry and diffractometry ) are interpreted in terms of scattering cross sections . here
we analyze what is really measured , how cross section is extracted and how it is theoretically defined .
this analysis leads to conclusion that to get a cross section from probability we have to introduce a parameter @xmath61 with dimension of area , characterizing the size of the neutron wave function . for definition of the scattering cross section we can look at an experiment schematically shown in fig .
if the detector registers @xmath62 scattered neutrons per unit time , then the total probability @xmath63 for a single neutron to be scattered in the sample into a given direction is @xmath64 where @xmath65 is the neutron flux density , @xmath66 is the area of the sample immersed into the neutron flux , and @xmath67 is the total number of neutrons incident on the sample per unit time .
experimentalists divide this value by dimensional parameter @xmath68 , where @xmath69 is atomic number density in the sample , and @xmath70 is the sample width . as a result one obtains the cross section @xmath71 where @xmath72 is the volume of the sample illuminated by the incident neutron flux .
the expression ( [ ratio1 ] ) is commonly accepted , but gives no insight into interaction of a single neutron with a single atom .
it looks more reasonable from total probability @xmath63 ( [ prob ] ) of scattering of a single neutron in the whole sample to find the scattering probability @xmath73 of a single neutron on a single atom : @xmath74 where @xmath75 is the number of atoms met by a single neutron on its way during the flight through the sample . to find the number of atoms on the neutron s way we have to introduce a front area @xmath61 of the incident particle wave function , and suppose that scattering takes place only , if the scattering center crosses this area .
so , let the neutron wave function to have area @xmath61 , then @xmath76 . from ( [ prob1 ] )
we immediately find the scattering cross section of a single neutron per single atom : @xmath77 which coincides with ( [ ratio1 ] ) .
the left hand side is the cross section we must calculate , the right hand side is the experimentally defined cross section . to compare theory with experiment
we must be able to calculate @xmath78 and @xmath61 . according to all the textbooks
the scattering cross section is defined as a ratio of the count rate @xmath62 of scattered particles to the flux density , @xmath65 , of the incident particles : @xmath79 if , for instance , we have a small sample , the above ratio gives the cross section of the whole sample , and if we divide this ratio by the total number of atoms in the sample , @xmath80 , illuminated by the incident flux , we obtain the result ( [ ratio1 ] ) , which defines the cross section per one atom
. we call it a phenomenological definition , because it says nothing about interaction of a single neutron with a single atom . theoretically , if you want to find a number of scattered particles for a single target atom and a given incident flux @xmath65 , you must first find , how a single particle is scattered , and how this scattering depends on impact parameter .
after that we should integrate over all particles in the incident flux , and average over all possible positions of the scatterer .
this procedure gives the number @xmath62 of scattered particles for the given @xmath65 .
the ratio @xmath81 is an average cross section @xmath82 , and it can be compared with the phenomenological one .
the parameter @xmath61 includes also dimension of a single nucleus .
we can even suppose that it is related only to nucleus , and neutrons are point particles , propagating like rays of a wave .
in that case @xmath61 is equal to the size @xmath83 of the nucleus , and the total cross section can never be larger than @xmath83 , because the total probability @xmath73 can never be larger than unity .
this contradicts to the well known facts , that some capture cross sections can be many orders of magnitude larger than @xmath83 . to avoid this contradiction
, we have to assume that @xmath61 is related to neutron and is considerably larger than @xmath83 , so we can neglect contribution of @xmath83 into @xmath61 .
introduction of the finite front area means that the particle wave function is not a plane wave , but a wave packet .
this wave packet can not be spreading , because , if it were , the transmission of a sample would decrease , when the sample is shifted from source to detector , and no one , so far , had ever observed such a phenomenon .
one of the possible candidates for the nonspreading wave packet is the singular de broglie wave packet ( dbwp ) @xcite @xmath84 where @xmath85/2 $ ] , @xmath46 determines the packet width , and @xmath86 is the wave packet velocity , which in our units @xmath87 coincides with wave vector @xmath88 . the front area of ( [ d1 ] )
can be estimated as @xmath89 .
some reasonings related to solution of ultra cold neutrons anomaly @xcite , leads to @xmath90 which means that for thermal energies @xmath91 . because of that the short range interaction becomes a long range one , which is manifested in such effects as total reflection and diffraction in crystals . for description of these processes
one does not use cross sections , so the parameter @xmath61 can be even omitted , and neutrons can be represented by plane waves .
at present we have three theories and every one has its own drawbacks . sht .
: : it is used only for elastic scattering and two particles scattering in cm reference frame .
it is logically inconsistent and contradicts to canonical quantum mechanics ( cqm ) , because it does not describe free particles after scattering .
the spherical waves are not solutions of the free schrdinger equation .
to make sht self consistent one have to find asymptotical form of spherical waves , which is a superposition of free states or free plane waves .
however after this step no cross section appears in theory , but only dimensionless probabilities .
: : this theory , which uses van hove correlation functions , is not a theory , but only a set of rules how to calculate a cross section .
this `` theory '' uses free states after scattering , however it never calculates scattering amplitudes , operates with probability per unit time , which does not depend on time , introduces superfluous space cell of arbitrary large size @xmath92 , a flux density for a single particle and has other drawbacks discussed below .
: : its main goal is to justify sst .
it uses asymptotical wave function after scattering , which is a superposition of free states .
more over it calculates scattering probabilities .
however these probabilities violate unitarity , and transition from probabilities to cross sections contains some connivance , which is not valid .
now we shall consider them in details in this theory the wave function of a scattered neutron is represented ( see , for example , @xcite ) as @xmath93 where plane @xmath94 and spherical @xmath95 waves describe incident and scattered particles , respectively .
the scattering amplitude is @xmath96 where @xmath97 are scattering matrix elements of spherical harmonics and @xmath98 are legeandre polynomials .
the simplest process is elastic s - wave scattering from a fixed center with wave function , @xmath99 in which @xmath100const .
the approach with spherical waves is borrowed from classical theories of sound and classical light scattering , but it is not a quantum scattering theory , because the spherical wave does not describe a free particle .
it satisfies the equation @xmath101\frac{\exp(ikr)}{r}=-4\pi\delta({\mbox{\boldmath $ r$ } } ) , \label{q2a}\ ] ] with the right hand side containing the dirac @xmath9-function .
it is not identical zero in the whole space .
therefore ( [ q2a ] ) is not a free equation .
usually one says that we need wave function outside the point @xmath102 , and there the spherical wave satisfies the free schrdinger equation .
however even the wave function of the bound state does satisfy free schrdinger equation outside the potential .
nevertheless we do not consider the tails of the bound wave functions as free particles .
of course , in the case of bound states the wave function tails decay exponentially at infinity , and this fact justifies why these tails are not considered as free particles .
we shall show below , that the asymptotic @xmath103 of spherical waves also contains exponentially decaying waves . therefore to get real asymptotic of spherical waves at infinity we need to exclude exponentially decaying part from it too ( compare ( [ 2a ] ) and ( [ 2ad ] ) below ) .
to make sht really quantum theory we have to find asymptotical form of spherical waves , when particles are sufficiently separated after scattering .
this asymptotical wave function should be a superposition @xmath104 where @xmath105 is wave vector of the length @xmath106 pointing into the direction @xmath107 in the element @xmath108 of the solid angle @xmath109 .
then the probability of scattering into @xmath108 is @xmath110 the asymptotic wave function ( [ fqq2 ] ) can be found in two ways : stationary and nonstationary ones .
the spherical function has 2-dimensional fourier expansion @xmath111 where we fix the direction from the scatterer to the observation point as @xmath112-axis , and integrate over all components @xmath113 parallel to @xmath114 plane with @xmath112-component of the momentum being equal to @xmath115 .
the range of integration over @xmath113 ( [ 2a ] ) is infinite , and , in particular , it includes those @xmath113 , for which @xmath116 . at these @xmath113 the component @xmath117 is imaginary , and @xmath118 is an exponentially decaying function .
if the distance to the observation point is large enough ( later we discuss what does it mean `` enough '' ) , we can neglect exponentially decaying terms , and restrict integration to @xmath119 : @xmath120 in this integral we can substitute @xmath121 where @xmath122 , @xmath117 is a variable , and we introduced the step function @xmath123 , which is unity or zero , when inequality in its argument is satisfied or not , respectively .
substitution of ( [ sph8 ] ) into ( [ 2ad ] ) gives @xmath124 the terms which we neglected are of the order @xmath103 because at the observation point @xmath112 ( @xmath125 ) we have @xmath126 where we replaced @xmath112 by the distance @xmath127 between scatterer and observation point .
thus the asymptotical form of the wave function ( [ fq2 ] ) after scattering is just ( [ fqq2 ] ) with scattering probability amplitude @xmath128 and scattering probability @xmath129 where @xmath130 is the neutron wave length .
we see that ( [ fq2 ] ) is reduced to ( [ fqq2 ] ) , when we neglect the terms of the order @xmath131 . since the decision to neglect or not to neglect this term is at will of the physicist ,
then the distance @xmath127 from the center is not an asymptotical one , being even of light years size , if he uses the spherical wave . on the other side the distances of the order 1 are asymptotical ones , if @xmath131 is neglected . to find nonstationary asymptotic of the wave function ( [ fq2 ] ) we include in it the time dependent factor @xmath133 , where @xmath134 , and use 3-dimensional fourier expansion for
the spherical wave @xmath135 one adds and subtracts @xmath136 in the exponent , and represents the field ( [ a ] ) as a superposition of plane waves @xmath137 with amplitudes @xmath138t)}{\omega_p-\omega_k - i\epsilon } , \label{ps8}\ ] ] which depend on time @xmath139 .
there is a relation @xcite @xmath140t)}{\omega_p-\omega_k - i\epsilon } = i\int\limits_{-\infty}^{t}\exp(i[\omega_p-\omega_k]t')dt ' , \label{mt1}\ ] ] which in the limit @xmath132 gives the law of energy conservation : @xmath141 t ' ) dt'=2\pi i\delta(\omega_p-\omega_k)=4\pi i\delta(p^2-k^2 ) .
\label{mt2}\ ] ] in this limit ( [ mt3 ] ) is @xmath142 and we get dimensionless scattering probability amplitude ( [ sph7 ] ) and the total scattering probability @xmath143 , which coincide with ( [ sph7a ] ) .
the main result of all these considerations is : the correct asymptotical wave function gives not a cross section , but only dimensionless probability . with asymptotical wave function ( [ mt4 ] )
it is possible to define phenomenological semi integral cross section @xmath144 as a ratio of normal flux @xmath145 through an infinite plane @xmath146 with normal @xmath33 to the incident flux density @xmath147 @xmath148 . \label{mmt4}\ ] ] the wave function ( [ mt4 ] ) in general case , when scattering is nonisotropic ( @xmath149 ) , can be represented as @xmath150 where @xmath151 .
substitution of this function into ( [ mmt4 ] ) gives @xmath152 where @xmath153 denotes solid angle @xmath154 around normal vector @xmath33 .
we can define differential cross section as @xmath155 this definition is not unique , because some arbitrary function @xmath156 can be added to it , for which @xmath157 and we do not want to accept it , because it is a phenomenological one .
sst ( see , for example @xcite ) is more general than sht , because it is applied to both elastic and inelastic processes .
it lists the rules , to be justified by fst , one must use to derive scattering cross section .
these rules are : 1 .
define the cross section as a ratio @xmath158 where @xmath159 is the flux density of incident neutrons , and @xmath160 is probability of scattering per unit time : 2 .
define the probability of scattering per unit time according to the fermi golden rule @xmath161 where @xmath162 is a matrix element of the neutron - scatterer interaction potential @xmath163 between initial @xmath164 , @xmath165 , and final @xmath166 , @xmath167 , neutron and scatterer states , respectively , and @xmath168 is the density of the neutron final states : @xmath169 the delta - function factor corresponds to the energy conservation law .
it contains initial and final neutron @xmath170 and scatterer @xmath171 energies .
the last factor is the phase space density of the neutron final states , which includes momentum element @xmath172 and some volume @xmath0 with an arbitrary large size @xmath92 .
3 . define neutron states before and after scattering as ( they are really free states ) @xmath173 4 .
the ratio ( [ nqb4 ] ) gives the cross section @xmath174 taking into account that @xmath175 , @xmath176 , one obtains double differential scattering cross section @xmath177 to compare with experimentally measured value one averages ( [ g2a ] ) over initial states and sums over final states of the scatterer , getting @xmath178 @xmath179 where @xmath180 is the probability of scatterer to be initially in the state @xmath181 . in the case , when @xmath182 where @xmath9-function depends on neutron and scatterer positions @xmath183 , @xmath184 respectively , the cross section ( [ gg3a ] ) becomes @xmath185 where @xmath12 is momentum transferred to the scatterer .
one of contradictions here is that the probability of scattering per unit time ( [ wpg1a ] ) does not depend on time , so integral of @xmath160 over time is senseless , whereas according definition it should be less than unity .
two other dubious points are : introduction of the superfluous volume @xmath0 and of flux density for a * single * particle .
the sst has also another drawback one never calculates scattering amplitude in it .
however this amplitude can be important . with it
one would be able to find coherent amplitude , which is averaged over initial states @xmath181 of the scatterer .
this amplitude gives coherent scattering cross section , which some times could carry an additional information about scattering process . in fst
( see , for example @xcite ) , which was developed mainly to justify sst , the scattering process is divided into three stages : infinite past , where incident particle is free ; present , where the neutron interacts with a scatterer ; and infinite future , where scattered neutron is again a free particle . to work within the hilbert space of normalized states , the incident free neutron is represented by a wave packet @xmath186 with fourier expansion : @xmath187 where parameter @xmath46 characterizes dimension of the wave packet , @xmath88 is momentum of the packet , @xmath188 is a state corresponding to plane wave , @xmath189 , in coordinate space , and @xmath190 are numerical fourier coefficients .
the dynamics of the packet @xmath186 is determined by free hamiltonian @xmath191 : @xmath192 after scattering the wave function becomes @xmath193 where @xmath194 is scattering matrix . and
dimensionless probability of scattering is defined as @xmath195 this last formula is absolutely wrong .
it means that scattering transforms the wave packet into a plane wave .
such a process is impossible because it completely violates unitarity .
the state @xmath186 is normalized and @xmath188 is not normalizable . from unitarity of the @xmath66-matrix
it follows that norm of the wave function is preserved .
so the final wave function @xmath196 after scattering is ether a wave packet or a superposition of wave packets @xmath197 with some new @xmath198 if the parameter @xmath46 depends on @xmath147 .
then @xmath199 defines the amplitude of transition probability of the wave packet state @xmath200 with momentum @xmath88 into the wave packet state @xmath201 with momentum @xmath202 .
in fact , in the books @xcite and others only scattering of plane waves is considered , and the initial wave packet defines only spectrum of plane waves in the incident beam .
however plane waves scattering gives only probability , and averaging over initial spectrum of plane waves does not change the situation . in fst one also introduces the finite volume @xmath0 with an arbitrary large @xmath92 for justification of sst .
below we show that it is possible to avoid this artificial step .
we use fundamental scattering theory with plane waves ( fpwst ) and calculate scattering probability of neutron in monatomic gas .
this probability , as is shown in introduction , is ambiguous , however , on one side , there is a way to find the standard expression for cross section , and on the other side , knowledge of the ambiguity helps to find the way to fight it .
the authors of @xcite and other books , make one special step to get cross section from probability , however this step contains a connivance that there is no scattering , if the scatterer does not cross the neutron wave packet . in linear scattering theory it is false , as we shall see in section 5 .
nevertheless we shall accept this point , because otherwise it is absolutely impossible to get a cross section from probability .
we shall accept such a connivance as an implicit introduction of nonlinearity into neutron - atom interaction .
here we apply the principles of fst to plane wave neutron - atom scattering and show that the result is ambiguous .
we also show how one can obtain the known result of sst .
when we consider neutron scattering from a monatomic gas , we must treat the neutron and atom of the gas in the same way , it means that we need the same schrdinger equation for both particles @xmath203\psi({\mbox{\boldmath $ r$}}_n,{\mbox{\boldmath $ r$}}_a , t)=0 , \label{2aa2}\ ] ] where potential @xmath204 is @xmath205 @xmath183 , @xmath206 , @xmath6 , @xmath7 are coordinates and masses of the neutron and atom respectively , @xmath5 , and the time @xmath139 contains factor @xmath207 .
the green function of equation ( [ 2aa2 ] ) with excluded interaction is @xmath208 where @xmath209 , @xmath210 .
it can be easily checked by substitution that @xmath211 satisfies the equation @xmath212g({\mbox{\boldmath $ r$}}_n-{\mbox{\boldmath $ r$}}'_n,{\mbox{\boldmath $ r$}}_a-{\mbox{\boldmath $ r$}}'_a , t - t')=-\delta({\mbox{\boldmath $ r$}}_n-{\mbox{\boldmath $ r$}}_n')\delta({\mbox{\boldmath $ r$}}_a-{\mbox{\boldmath $ r$}}_a')\delta(t - t ' ) .
\label{grenf}\ ] ] the scattered part of the wave function is @xmath213 @xmath214 @xmath215 where @xmath216 , @xmath217 describe incident plane waves of the neutron and atom respectively with their energies @xmath218 , and @xmath219 .
the wave function ( [ a1 ] ) after integration over @xmath220 and @xmath42 can be found in the final form @xmath221-iet\right ) , \label{aa11}\ ] ] where @xmath222 is relative speed of particles , @xmath223 is their total momentum , and @xmath224 is their total energy .
this wave function has a very clear physical interpretation , however it is not an asymptotical one . to get the asymptotical wave function we represented ( [ a1 ] ) as a superposition of plane waves related to final states of the neutron , @xmath225 , and atom , @xmath226 : @xmath227 where @xmath228 with the relation ( [ mt1 ] ) we find in the limit @xmath132 that the probability amplitude for the particle to leave in the state @xmath42 , and for the atom to leave in the state @xmath220 is : @xmath229 which coincides with ( [ intr ] )
now we can show that with this amplitude it is possible to get the standard cross section .
for that we transform to cm reference frame and obtain the probability scattering amplitude ( [ intr9 ] ) : @xmath230 from which with the help of front area @xmath61 and ( [ intr10 ] ) we obtain the cross section in the cm frame @xmath231 for neutron scattering on an atom with the initial momentum @xmath55 .
transformation to the laboratory frame according to ( [ intr7b ] ) gives @xmath232 the number of collisions per unit time in a gas sample is @xmath233 where @xmath234 is relative velocity .
since the interaction time with the whole sample of width @xmath70 is @xmath235 , then the total probability of a single neutron scattering in the sample is @xmath236 where @xmath237 is maxwellian distribution of @xmath55 : @xmath238 if we accept @xmath239 ^ 2}{q^2},\ ] ] and define cross section as @xmath240 , then we obtain the standard scattering cross section in a monatomic gas @xmath241 it is necessary to point out that definition ( [ area ] ) of the parameter @xmath61 means that the wave packet contracts with increase of the relative energy , and because of that the wave optics at low energies transforms to geometrical one at high energies .
[ [ the - total - cross - section - for - an - atom - at - rest . ] ] the total cross section for an atom at rest .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + if instead of maxwellian distribution we substitute @xmath242 , we immediately obtain the total cross section @xmath243 which also coincides with the standard one .
all the standard results were obtained with the help of cm reference frame . if we repeat the same with scattering probability ( [ intr6 ] ) obtained directly in the laboratory reference frame we find a divergency when averaging with maxwellian distribution and an additional factor @xmath244 in ( [ bcmm1 ] ) for scattering from an atom at rest .
introduction of the parameter @xmath61 means that instead of plane waves we deal with wave packets . we can not deduce @xmath61 from a theory , however we can study its properties . we also need to learn how to describe scattering in terms of wave packets .
here we consider elastic scattering of a wave packet from a fixed center , and find at first sight unexpected though finally well understandable result that probability of scattering does not depend on position of the scattering center .
however first of all we need to discuss the types of the wave packets . in general all the wave packets can be represented by their fourier expansion @xmath245 , \label{q4q4}\ ] ] where parameter @xmath46 determines the width of the packet like in ( [ d1 ] ) , @xmath246 and @xmath247 are functions of invariant variables @xmath248 , @xmath249 and @xmath250 .
the primary wave packet describes a free incident particle .
its fourier expansion contains plane waves @xmath251 , which satisfy the equation @xmath252\exp(i{\mbox{\boldmath $ p$}}{\mbox{\boldmath $ r$}})=0 . \label{e1}\ ] ] all the packets are representable in the form ( [ q4q4 ] ) , and they differ by the fourier coefficients @xmath246 and dispersion @xmath247 .
we consider three types of the wave packets and discuss which one is the most appropriate for description of particles .
the most popular in the literature is the gaussian wave packet @xmath253 ^ 2}{2[1+its^2]}\right ) .
\label{g1}\ ] ] this packet is normalized to unity , satisfies the free schrdinger equation , but spreads in time .
because of this spreading its form in space does not completely coincides with ( [ q4q4 ] ) .
its fourier components are @xmath254 where @xmath46 is the width in momentum space .
the spectrum of wave vectors @xmath255 is spherically symmetrical with respect to the central point @xmath256 and decays away from it according to gaussian law .
the cross area of this packet moving in @xmath112 direction can be defined as @xmath257 where @xmath258 .
it is known that there are no nonspreading normalizable wave packets , which satisfy the free schrdinger equation .
however nonnormalizable wave packets do exist . as an example we can demonstrate nonsingular de broglie wave packet @xcite @xmath259 in which @xmath260 and @xmath261 in units @xmath262 .
the packet ( [ q4q2 ] ) is a spherical bessel function @xmath263 , which center is moving with the speed @xmath86 .
this packet satisfy the free schrdinger equation .
its fourier components are @xmath264 and spectrum of @xmath255 is a sphere of radius @xmath46 in momentum space with center at the point @xmath256 . since it is not normalizable , its front area is infinite , and such a wave packet is not a good candidate for description of free neutrons .
the singular de broglie wave packet @xcite @xmath265 is normalizable one with normalization constant @xmath266 defined by @xmath267 the parameter @xmath46 is the width of the packet in momentum space and reciprocal width in coordinate space , @xmath268 is the packet speed , and @xmath269 .
we see that @xmath270 is less than kinetic energy by the term @xmath271 , which can be thought of as bound energy of the packet .
the singular de broglie wave packet satisfies inhomogeneous schrdinger equation @xmath272\psi_{db}({\mbox{\boldmath $ r$}},{\mbox{\boldmath $ v$}},t , s)= -2\pi ce^{i(v^2+s^2)t/2}\delta({\mbox{\boldmath $ r$}}-{\mbox{\boldmath $ v$}}t ) , \label{ab1}\ ] ] with right hand side being zero everywhere except one point .
the fourier coefficients of the singular de broglie wave packet are @xmath273 and @xmath274/2=[p^2-({\mbox{\boldmath $ k$}}-{\mbox{\boldmath $ p$}})^2-s^2]/2 . \label{d1cc}\ ] ] the spectrum of wave vectors @xmath255 is spherically symmetrical with respect to the central point @xmath256 and decays away from it according to lorenzian law with width @xmath46 .
the fourier coefficients ( [ d1d ] ) and frequency ( [ d1cc ] ) become identical to those of spherical wave @xmath275 after substitution @xmath276 and @xmath277 , i.e. @xmath278 .
the front area of the singular de broglie wave packet , moving in @xmath112 direction can be defined as @xmath279 where @xmath258 .
after change of variables @xmath280 we get @xmath281 the singular de broglie wave packet descends from the spherical wave . indeed ,
let s consider the spherical wave with energy @xmath282 : @xmath283 this wave satisfies inhomogeneous schrdinger equation @xmath272\psi(r , t , q)= -2\pi \exp(-iq^2t/2)\delta({\mbox{\boldmath $ r$ } } ) .
\label{ab2a}\ ] ] the right hand side describes the center radiating the spherical wave .
if we change to reference frame moving with the speed @xmath261 then we must perform the following transformation of the function @xmath284 : @xmath285 the transformed function is the spherical wave around moving center .
it satisfies the equation @xmath286\psi= -2\pi \exp(i[k^2-q^2]t/2)\delta({\mbox{\boldmath $ r$}}-{\mbox{\boldmath $ k$}}t ) .
\label{ab2b}\ ] ] if the energy of the wave ( [ q5q ] ) is negative : @xmath287 , i.e. the wave ( [ q5q ] ) describes a bound state around the center , then ( [ q4q ] ) becomes @xmath288 with normalization constant @xmath289 expression ( [ q4q1 ] ) becomes identical to ( [ exc3 ] ) .
thus the singular de broglie wave packet is the spherical hankel function of imaginary argument moving with the speed @xmath86 .
the nonsingular de broglie wave packet is obtained by transformation to the moving reference frame of the nonsingular spherical wave @xmath290 which satisfies the homogeneous schrdinger equation .
this way we can construct a lot of nonsingular wave packets corresponding to different angular momenta @xmath291 . we see that the proof of validity of sst is not perfect because of unacceptable definition of scattering probability , according to which a wave packet after scattering transforms to plane waves , though according to unitarity it should remain a wave packet .
now we look at a wave packet elastic scattering from a fixed center .
we take the wave packet not as a preparation of a particle in some state , but as an intimate property of the particle , which means that after scattering the particle is the same packet as before it .
the wave packet ( [ q4q4 ] ) relates to a free particle . in the presence of a potential @xmath292
the plane wave components @xmath251 should be replaced by wave functions @xmath293 , which are solutions of the equation @xmath294\psi_p({\mbox{\boldmath $ r$}})=0 \label{e2x}\ ] ] containing @xmath251 as the incident wave .
substitution of @xmath295 into ( [ q4q4 ] ) transforms it to @xmath296 .
\label{q4q4x}\ ] ] now we have to find asymptotical form of ( [ q4q4x ] ) . for that we substitute asymptotical form of @xmath295 . for incident wave @xmath251 asymptotical wave function after scattering on a fixed center with an impact parameter @xmath297 is a superposition of plane waves : @xmath298 ) , \label{e2y}\ ] ] where @xmath299 is the probability amplitude of a plane wave with wave vector @xmath255 to be transformed to the plane wave with wave vector @xmath300 pointing into direction @xmath107 in the element of solid angle @xmath108 .
this amplitude for isotropic scattering is @xmath301 .
dependence on @xmath302 is an irritating moment , however , since the spectrum of wave packets has a sharp peak at @xmath303 , we can approximate @xmath299 by @xmath304 , having in mind that corrections to this value is of the order @xmath305 , where @xmath46 is the packet width in the momentum space .
this correction is small , when @xmath46 is small , i.e. area @xmath61 of the packet is large .
the vector @xmath300 in ( [ e2y ] ) is of length @xmath302 , but it is turned by angle @xmath109 from @xmath255 .
substitution of ( [ e2y ] ) into ( [ q4q4 ] ) for @xmath251 transforms ( [ q4q4 ] ) to the form @xmath306-i\omega({\mbox{\boldmath $ p$}},{\mbox{\boldmath $ k$}},s)t ] .
\label{q4qq}\ ] ] since @xmath246 , @xmath307 and @xmath247 are invariant with respect to rotation , we can replace them with @xmath308 , @xmath309 and @xmath310 .
after that we can transform integration variable @xmath311 , and drop the index @xmath109 of @xmath255 . as a result we transform ( [ q4qq ] ) to the form @xmath312-i\omega({\mbox{\boldmath $ p$}},{\mbox{\boldmath $ k$}}_\omega , s)t ] , \label{q4qq1}\ ] ] which can be represented as @xmath313 where @xmath314 denotes the wave packet of the the same form as that of the incident particle .
we see that the packet as a whole is scattered with probability @xmath315 , which , surprisingly , has no dependence on impact parameter @xmath297 as in the case of plane waves .
it shows that scattering of wave packets is almost the same as that of plane waves .
the independence of scattering amplitude on position of scatterer is well understandable in linear wave mechanics .
indeed , the wave packet is a superposition of plane waves , which exist in the whole space .
they cancel each other outside the packet , but they exist , and because they are scattered independently of each other , the whole packet s scattering does not depend on position of the scatterer . to get a cross section
we need an additional hypothesis which restricts scattering to those cases , when the wave packet overlaps the target position .
this hypothesis is outside of the wave mechanics and in the books @xcite it is accepted implicitly .
we can say that without this hypothesis the wave mechanics is incomplete theory , i.e. it is insufficient to describe scattering of particles .
introduction of this hypothesis is equivalent to inclusion of nonlinearity into quantum mechanics .
when we write @xmath316 , we implicitly accept nonlinearity .
we can not deduce @xmath61 , but we can explore its properties .
so we can ask ourselves how large is @xmath61 and what is its dependence on @xmath317 . at first sight @xmath61
should not be large .
indeed , if cross section of scattering from a fixed center , or from a heavy nucleus ( @xmath318 ) , according to ( [ bcm ] ) is @xmath319 , then , to get usual cross section @xmath320 , we must put @xmath321 , where @xmath322 is wave length of the neutron motion with respect to the center . however , if @xmath61 is small , then the total reflection of thermal neutrons becomes impossible .
indeed , the wave packet ( [ q4q4 ] ) contains plain waves with different wave vectors @xmath255 .
probability of wave packet reflection is determined by reflected wave function @xmath323 , \label{q4q4q}\ ] ] where @xmath324 the width of the spectrum of @xmath255 is proportional to @xmath46 .
if @xmath325 is small , then @xmath46 is large . if @xmath46 is large , the probability to find @xmath326 when @xmath327 is also large , so @xmath328 . to get large reflection we need small @xmath46 , so for @xmath327 probability to find plane wave components with @xmath329 is small .
in that case @xmath330 .
in particular , if @xmath331 , which is the case for ultracold neutrons ( ucn ) , @xmath332 , which means @xcite that @xmath333 ( [ area1 ] ) . however , if @xmath46 is small , than @xmath325 is large .
then to get @xmath334 to be equal to the measured cross section of the order 1 barn , we need to take @xmath8 of the order of @xmath335 cm . with such small @xmath8
we obtain optical potential @xmath336 too low to see ucn storage , and , more over , again we can not see the total reflection of thermal neutrons .
so we have a new contradiction : @xmath61 have to be large , but it can not be large .
later we shall see how to resolve this contradiction .
in any case the value of @xmath61 can be explored experimentally in two types of experiments : in measurement of transmission of thin films when the grazing angle of incident neutrons is below critical one @xcite , and deviation of reflection angle from the specular direction , when the grazing angle of incident neutrons is a little bit above the critical one @xcite .
both of these effects are proportional to @xmath46 , as it follows from ( [ q4q4q ] ) .
we have seen in section 4.2 that to get the standard cross section for neutron - monatomic gas scattering we need to suppose that @xmath61 is proportional to @xmath337 , where @xmath317 is the energy of relative neutron - atom motion .
if the area @xmath61 were constant then the cross section calculated via cm reference frame would depend on gas temperature @xmath338 proportionally @xcite @xmath339 .
so , to define whether @xmath61 does depend on energy or not , it is necessary to measure temperature dependence of neutron transmission through monatomic gas .
there is one publication @xcite on this matter .
however it was not the temperature dependence , which was measured there , but energy dependence of neutrons transmission through noble gases .
the temperature dependence was deduced , because transmission depends on dimensionless parameter @xmath340 .
so , it was desirable to measure directly the temperature dependence .
the experiment was done with cold neutrons @xcite and it had shown , that cross section grows with temperature proportionally to @xmath341 , which supports @xmath342 dependence .
this results contradicts to the experiment on subcritical transmission of mirrors , when grazing angle of incident neutrons is below critical one @xcite .
however , we believe that the last experiment should be repeated to seek more carefully for transmitted neutrons .
the sht theory , which is used for description of elastic scattering and two body scattering in cm reference frame contradicts canonical quantum mechanics , because spherical waves do not correspond to free particles after scattering . +
this contradiction is resolved , when we find asymptotical form of spherical waves in accordance with fst .
2 . sst obtains cross sections by introduction of scattering probability per unit time ( though it does not depend on time ) , and flux density for a single particle .
+ it can be avoided , if instead of sst we use fst .
3 . fst pretends to prove validity of sst , however this prove is inconsistent , because it violates unitarity .
this violation is related to transformation of initial wave packet into final plane waves .
+ we resolve this contradiction replacing wave packet in the initial state by a plane wave .
however , then we obtain only probabilities and not cross sections . +
this contradiction is resolved by introduction of wave packet and its cross area @xmath61 .
5 . to improve fst
we consider both initial and final neutrons described by the same wave packet .
however we find that wave packets scatter nearly the same as plane waves .
the scattering probability does not depend on impact parameter , so introduction of wave packets does not help to get cross sections from probabilities .
+ resolution of this contradiction requires introduction of nonlinearity into interaction .
we do not have such a theory so we include @xmath61 by hands , and suppose that such inclusion implicitly involves nonlinearity .
when we consider the value of @xmath61 we find that it should be large but can not be large .
+ we resolve this contradiction simultaneously with the following ambiguity .
when we apply fst with initial and final plane waves we find that calculation of neutron scattering from a moving atom gives different results , when we find scattering probability directly in lf reference frame , or via intermediate transformation to cm .
+ below we show how this ambiguity can be resolved , and this resolution simultaneously resolves the contradiction between ( [ area ] ) and ( [ area1 ] ) .
in the whole we must claim that there are no now a selfconsistent scattering theory .
however we are to proceed with that we have at our hands . the ambiguity arises because of the continuous spectrum of angular distribution @xmath343 no ambiguity would appear , if the angular distribution were discrete .
so we shall eliminate the ambiguity , if we replace the integral in ( [ wpg39n ] ) with the integral sum @xmath344 where discrete amplitudes @xmath345 are introduced .
in such a representation the probability of scattering into the angle @xmath346 is @xmath347 , and probability of scattering into the angular interval covering @xmath348 elements @xmath349 around some direction @xmath350 , is @xmath351 where @xmath352 and we supposed that @xmath353 are almost constant in the interval @xmath349 .
transformation of the integral to the sum can be made with an arbitrary choice of @xmath354 around every @xmath346 , which is also an ambiguity . to resolve it
we can make a step in style of quantization , i.e. we can require all the amplitude elements @xmath355 to be equal , which means that we introduce a quantum of the area @xmath356 . in the cm reference frame , where @xmath357 is constant , such a requirement means an introduction of quantum of the angular interval or uncertainty @xmath358 .
below we discuss what value this quantum can be of .
when we transform from one reference frame to another , we see a deformation of both @xmath359 and @xmath358 , however the amplitude elements @xmath360 and the number of such elements in the whole integral @xmath361 remain the same .
therefore , if we take some amount of amplitude elements in the center of mass reference frame , square every element , and after that transform to the laboratory reference frame , then we obtain the same amount of the squared elements in the laboratory frame , and they will be confined in some angular interval . if we transform from center of mass to laboratory frame without squaring the amplitude elements , and square them only after the transformation , we obtain the same number of squared elements in the same angular interval as before . which means that our result is completely invariant under galilean transformation .
therefore we have the full right to use center of mass reference frame without a danger to get an ambiguous result after changing of the reference frame . according to ( [ fbcm ] ) we have @xmath362
therefore the scattering amplitude element is @xmath363 probability of scattering into some interval of solid angle @xmath364 is @xmath365 the differential scattering cross section becomes @xmath366 and we immediately obtain that to get @xmath367 , we have to require @xmath368 it means that the angular quantum @xmath358 correlates with the wave packet cross area @xmath61 !
this correlation is very important because it resolves the contradiction of two results : ( [ area ] ) and ( [ area1 ] ) .
since , without doubt , the interval @xmath358 is small , we must have @xmath61 to be large .
so the contradiction between ( [ area ] ) and ( [ area1 ] ) is resolved in favor to ( [ area1 ] ) .
simple consideration of a scattering process shows a contradiction inherent in the sht .
on one side we use plane waves as eigenstates of a particle , and on the other side describe scattered particles with spherical waves , which are not even solutions of the free schrdinger equation .
rigorous approach , which resolves this contradiction , creates another one . with this approach
we can calculate only dimensionless probabilities of scattering , while for interpretation of experiments we need cross sections with dimensions of area . to resolve this second contradiction and to get cross sections instead of dimensionless probabilities
, we have to introduce some front area @xmath61 of the wave function for the incident particle .
thus we arrive at wave packets , and must describe scattering process with wave packets instead of plane waves .
however , while doing that we arrive at the next contradiction : the wave packets scatter similarly to plane waves , and scattering does not depend on whether the scatterer crosses the wave packet or not .
this contradiction is a result of linearity of the quantum mechanics . to resolve this contradiction
we must introduce a nonlinearity at least into interaction term to assure that scattering takes place only , when the scatterer is inside the front area of the wave packet .
however , even with this hypothesis in mind we have another contradiction : the wave packet have to be large , but can not be large .
if we abandon wave packet and calculate scattering probability with plane waves , we find that the probability of scattering is an ambiguous value , which depends on a way of calculation . to resolve this ambiguity we are to quantize the process of scattering .
this quantization leads to replacement of angular integrals with discrete integral sums .
it is amazing that resolution of the ambiguity simultaneously resolves the contradiction related to the value of area @xmath61 pointed out in ( [ area ] ) and ( [ area1 ] ) .
the last result proves that some truth is contained indeed in our investigations . as for sst
it is not a theory , because it is not justified by fst , but only a set of rules how to calculate cross sections .
it can be used by everyone , who is satisfied with these rules , notwithstanding that they are not well justified .
we hope that our research will help to understand better the merits of sst , of quantum mechanics as a whole , and will give an impetus for serch of a more consistent scattering theory .
the author is very grateful to j - f .
bloch , prof . of institute
francais politechnigue de grenoble , for his interest to the work and assistance , to dr .
e.p.shabalin of jinr ( dubna ) for his support , to f.gareev and v.furman for their encouragement , to v.nikolenko for his help , to d.koltun , prof of rochester university , and to dr .
a.shabad from lebedev physics institute of rf academy of science for their useful discussions .
despite my considerable efforts i have been unable to elicit any referee reports on your article and at this stage i do not think i am going to receive any reviews . in the circumstances and to save you losing further time i believe it would be in your interests if you submitted your article to another journal .
this paper makes the startling claim that there are fundamental flaws in the standard quantum scattering theory that was developed by moller , jauch , and others and is described in refs 4 and 5 and many others texts ( for example , _ scattering theory of waves and particles _ by roger newton ) .
claims of this sort must be taken seriously .
but they must also be supported by unusually good evidence , and , in my view , such evidence is definitely not present here .
therefore , it is my opinion that this paper should not be published in anything like its present form .
i shall not discuss all of the points raised in the paper , but three will , i hope , suffice . consider first the claim in connection with eqs .
( 60 ) and ( 61 ) that standard scattering theory violates unitarity .
the function @xmath369 is the momentum - space wave function of the incoming state @xmath370
. the function @xmath371 , with @xmath196 as defined in eq .
( 60 ) , is the momentum - space wave function of the corresponding outgoing state @xmath372 .
if @xmath370 is normalized , then so is @xmath373 , since @xmath66 is unitary .
equation ( 61 ) is simply the elementary proposition that the absolute value squared of the momentum - space wave function gives the probability density that a measurement of momentum will give an answer in the neighorhood of @xmath374 .
it in no way implies that the scattering process has produced a transition from the normalized @xmath370 to the improper state @xmath375 .
rather , the scattering process has produced the normalized state @xmath373 with a well - defined probability for its momentum to be found in the neighborhood of any chosen point @xmath376 .
the discussion of eq .
( 61 ) here and elsewhere seems to show a failure to understand the most basic principles of quantum mechanics .
consider next the observation that the wave function of eq .
( 28 ) is not a solution of the free schrdinger equation everywhere .
this observation is , of course , correct , but is in no way in conflict with standard scattering theory .
the form ( 28 ) is only valid in the asymptotic region , far outside the target potential .
thus it is only expected to satisfy the free schrdinger equation in this region .
its behavior near @xmath377 is completely irrelevant .
thus the presence of the delta function @xmath378 in eq .
( 29 ) is , in no way , the flaw that the author seems to believe it to be .
finally , the claim on page 29 that the scattering probability is independent of the position of the target violates both common sense and observed fact . any rigorous definition of the scattering cross section requires an integration over impact paramenters , and
this integration converges only if the scattering probabiliy goes to zero as the impact parameter gets large .
this is true in the standard scattering theory of the standard references .
if it is false in the author s version of scattering theory , this only proves that his version is undeniably unsatisfactory .
no , it is an amplitude of momentum state component @xmath380 in the composition of the initial state . if you mean the same but call it a momentum wave function - then it is only semantic difference between us .
`` equation ( 61 ) is simply the elementary proposition that the absolute value squared of the momentum - space wave function gives the probability density that a measurement of momentum will give an answer in the neighborhood of @xmath374 . ''
i agree that it produced the normalized state , however my question : what the scattering process means : transition from improper state @xmath382 i.e. @xmath383 to improper state @xmath384 or initial state has a well definite momentum @xmath385 ? then what is the wave function with momentum @xmath374 after scattering ?
look , please , @xmath66-matrix transforms plane waves into superposition of plane waves .
why do nt you comment my eq .
( 62 ) and text below it ?
if particles are described by normalized wave functions the scattered wave function should have momentum density not @xmath386 , but @xmath387 , where @xmath388 momentum of scattered particle with normalized outgoing wave function @xmath389 .
this is my argument .
why do you ignore it ?
\2 ) you write : `` the form ( 28 ) is only valid in the asymptotic region , far outside the target potential .
thus it is only expected to satisfy the free schrdinger equation in this region .
its behavior near @xmath377 is completely irrelevant . ''
my arguments : even bound state wave function satisfies free schrodinger equation outside the potential well .
you do not consider it as scattering wave function , because it has exponentially decaying asymptotic , not @xmath103 .
however the function @xmath103 also contains exponentially decaying waves .
therefore to be self consistent you must neglect exponentially decaying waves there too .
then you obtain real asymptotic the superposition of plane waves .
why do nt you discuss eq - s .
( 32 ) , ( 33 ) ?
for clarity i added few words after eq .
( 29 ) .
`` any rigorous definition of the scattering cross section requires an integration over impact parameters , this integration converges only if the scattering probability goes to zero as the impact parameter gets large . ''
i agree that such result is unsatisfactory . for me myself
it was a tragedy , when i discovered it .
and i submitted this paper in order to show : friends , look , what happens with standard approach .
( it is standard approach , not my version . )
you reject result , and did nt show where mathematics is wrong in my consideration of the wave packet scattering .
take into account , please , that i am the first , who tried to do it , and why do nt you discuss my arguments after eq .
if something is not clear , give me to know , please .
i ll try to improve .
dear referee , i appeal not to a student , who had just accumulated his knowledge , but to a researcher , who dares to doubt .
i understand that you have no time to read the paper carefully , but i ask you then to take a responsibility before killing a paper .
i am ready to discuss all my points openly .
i hope that you will be one of those , who will be interested enough to spare more time for discussion . [ [ reply - of - non - anonymous - referee - on - november-1-and - my - reply - on - november-02 ] ] reply of non anonymous referee on november 1 and my reply on november 02 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ the manuscript ah10019 , `` contradictions of quantum scattering theory , '' by vladimir k. ignatovich has been rejected by physical review a , and the author has requested reconsideration .
the decision was based on the position of the referee that the evidence presented in the manuscript is not sufficient to show that conventional scattering theory has fundamental flaws , as claimed by the author .
the referee singled out three examples of points made in the manuscript that were not convincing in this regard .
the author has attempted to show that the comments made by the referee are not valid .
the first point concerns eq .
( 61 ) and the statements made in the manuscript : `` this last formula is absolutely wrong .
it means that scattering transforms the wave packet into a plane wave . ''
the referee does not accept this statement , because the formula does not mean that the scattered wave function is a plane wave .
the referee correctly points out that eq .
( 61 ) represents the probability density that the scattered particle will have a particular momentum and that it does not imply that the final state is a plane wave .
this is the conventional and correct quantum mechanical interpretation of such an equation .
the subsequent discussion in the manuscript does not support the author s claim , as suggested in his resubmission letter . * my reply .
before to find momentum density of scattered particle , you should define the wave function of the particle .
your disagreement with my arguments shows how many unclear notions quantum mechanics contains . *
the second point concerns eq .
( 29 ) . in the manuscript
it is stated that `` the spherical wave does not describe a free particle . '' in fact , outside of the region containing the origin , the spherical wave does describe a free particle , as pointed out by the referee , and the claim by the author is not valid .
the author goes on to suggest that his claim supported by the fact that bound - state wave functions outside of the potential decay exponentially , however this analogy is not convincing .
in fact , the solution in eq .
( 29 ) is a free solution outside of the potential for either a scattering state or a bound state .
the only difference is that @xmath147 is imaginary in the case of a bound state . *
why do nt you comment my equations ( 32 ) and ( 33 ) ?
i tell just about imaginary component of @xmath147 .
the waves , which have imaginary component in some direction must be excluded from asymptotic .
* the third point made by the referee is that the conclusion by the author that the scattering of a wave packet is independent of the impact parameter is problematic .
the author claims that this is a problem with conventional scattering theory , however it appears more likely that the author s interpretation of scattering theory is at fault .
the source of the problem is indicated in the statement in the manuscript : `` the independence of the scattering amplitude on position of scatterer is well understandable in linear wave mechanics . indeed , the wave packet is a superposition of plane waves , which exist in the whole space .
they cancel each other outside the packet , but they exist , and because they are scattered independently of each other , the whole packets scattering does not depend on position of the scatterer . ''
the last statement is not correct and overlooks the fact that if the initial state is a wave packet , then the plane wave components will scatter individually in such a way that their coherent superposition will destructively interfere to give a small scattering probability if the impact parameter of the initial wave packet is large .
evidently , the calculation in the manuscript makes approximations that lose this information .
however , it is not appropriate for the referee to correct the manuscript line by line in order to show how to fix the problem .
in addition to the valid criticisms of the referee , it should be noted that the description of scattering in terms of wave packets is not a new result .
for example , the subject is discussed in the article `` scattering of wave packets , '' by eyvind h. wichmann , american journal of physics , 33 , 20 ( 1965 ) .
the abstract of that article notes that : `` the aim of this paper is mainly pedagogical , and the final results obtained are old and well established . ''
* my reply .
i thank you very much for the reference .
i rushed into library , grasped the journal , buried myself in it . and was greatly disappointed .
there are no wave packet at all , except words
. may be it is a special pedagogical trick of am.j.phys . ?
* * dear dr .
mohr , it is very difficult to discuss the paper point by point with a man , who does not want to listen arguments .
i was surprised that you did not understand my simplest arguments about exponential decay of some part of spherical wave asymptotic ( eq .
( 32 ) and ( 33 ) ) .
do you understand that the @xmath390 becomes imaginary , when @xmath116 ?
if you do not understand that , how can you understand my mathematics about wave packet scattering ? !
* [ [ my - last - reply - and - appellation - to - editor - in - chief - of - aps - on - november-2 ] ] my last reply and appellation to editor - in - chief of aps on november 2 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ dear editor , below ( here above ) i wrote my reply to dr . mohr .
i consider his report as incompetent one .
forward , please , my letter with my reply and my entire file to the editor - in - chief of the american physical society .
dear , editor - in - chief , you already have one of my appellations on your table .
it is on the paper about experiment on measurement of temperature dependence of neutron - helium scattering , submitted to physical review letters .
referees of that paper complained that they are unable to see my full paper on contradictions .
i tried hard to publish this paper in physical review a. look , please , how difficult it is , and how incompetent are referee reports .
both referee even do not understand that the square root @xmath391 becomes imaginary , when @xmath392 !
i am ready to defend all my results openly
! give me a chance , please , to do it .
i also ask your permission to put all my discussions with referees at the end of my paper .
i suppose it is the only way to punish irresponsible referees !
yours sincerely , vladimir ignatovich .
your paper was submitted to physical review a on 03 august 2005 .
this journal rejected your manuscript on 09september 05 , on the basis of an incontrovertibly negative referee report .
you appealed the rejection to the editorial board ; the appeal was adjudicated by board member dr .
peter mohr who denied your appeal .
now you apparently ask that the rejection of your appeasl by the board be further appealed to the editor - in - chief of the american physical society .
however , the closing sentence(s ) of the board adjudication state that `` under the revised editorial policies of the physical review ( copy enclosed ) , this completes the scientific review of your paper . '' while you may request that the case be reviewed by the editor - in - chief of the aps ( see enclosed editorial policies ) .
this request should be addressed to the editor , who will forward the entire file to the editor - in - chief of the aps.suvch an appeal must be based on the fairness of tehprocedures followed , and must not be a request for another scientific review .
the question to be answered in this review is : did the paper receive a fair hearing ?
the decision of the editor - in - chief concludes the consideration of the manuscript by the american physical society .
dear dr.bernd crasemann , thank you for reply .
may i ask you , are you just that editor to whom i am to ask to forward my appeal to editor - in chief . if yes , please , forward this letter and all my file to him .
may i ask you ? what does it mean `` fair hearing ? '' if i claim that 2x2=4 , and referee reports that he learned from text books that 2x2=5 , and dr .
mohr repeats that he believes the same , is it fair hearing .
may i ask all your editorial board to answer a simplest question : does the integral in my formula ( 32 ) contain exponentially decaying waves or not ?
dr mohr replies `` not '' , and i do nt believe he is a scientist .
you should check his diploma . [
[ reply - of - editor - in - chief - martin - blume - of-13-dec.-05-sent - by - postal - mail ] ] reply of editor - in - chief martin blume of 13 dec .
05 sent by postal mail ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ i have reviewed the file concerning this manuscript which was submitted to physical review a. the scientific review of your paper is the responsibility of the editor of physical review a , and resulted in the decision to reject your paper .
the editor - in - chief must assure that the procedures of our journals have been followed responsibly and fairly in arriving at this decision . the insulting language that you have used in your correspondence
does nothing to advance the cause of your manuscript , nor has it any place in scientific discourse .
we will not respond to any further correspondence on this matter .
on considering all aspects of this file i have concluded that our procedures have in fact been appropriately followed and that your paper received a fair review .
accordingly , i must uphold the decision of the editors .
landau and e.m .
lifshits quantum mechanics m. : nauka 1989 .
w.marshall and s.w.lovesey , _ theory of thermal neutron scattering .
_ clarendon press , oxford , 1971 .
s.w.lovesey , _ theory of neutron scattering from condensed matter . _ clarendon press , oxford , 1984 , v. 1 , ch . 1 .
m.r.goldberger , k.w.watson , _ collision theory _
( john wiley & sons inc . ,
london , sydney , toronto , 1964 ) .
j.r.taylor , _ scattering theory . the quantum theory of nonrelativistic collisions .
_ ( john wiley & sons inc .
, n.y . , london , sydney , toronto , 1972 ) . l. de broglie , _ non - linear wave mechanics : a causal interpretation .
_ elsevier , amsterdam , 1960 .
ignatovich , m. utsuro , `` tentative solution of ucn problem , '' _ phys .
lett . _ * a225 * , pp .
195 - 202 , 1997 .
m. utsuro , v. ignatovich , p. geltenbort , th .
brenner , j. butterworth , m. hino , k. okumura , m. sugimoto , `` an experimental search of subcritical transmission of very cold neutrons ( vcn ) described by the de broglie wave - packet , '' _ proceedings of vii international seminar on interaction of neutrons with nuclei : neutron spectroscopy , nuclear structure , related topics , dubna , may 25 - 28 , 1999 , _ * e3 - 98 - 212 * , pp . 110 - 125 , jinr , dubna , 1999 . v.k .
ignatovich , neutron reflection from condensed matter , the goos - haenchen effect and coherence .
a. 322 ( 2004 ) 36 - 46 .
ignatovich , apocrypha of standard scattering theory ( sst ) and quantum mechanics of the de broglie wave packet .
jinr e4 - 2001 - 42 , dubna , 2001 ; proceedings of isinn-9 , p. 116 - 127 .
dubna , 2001 ; proceedings of the xxxvi winter school in pinpi on physics of atomic nuclei and elementary particles , st.petersburg , 2002 , pp .
347 - 380 .
r. genin , h. beil , c. signarbieux a. o. , _ determination , des sections efficases dabsorption et de diffusion des gaz rares pour les neutrons thermique _ le journal de physique et le radium * 24 * ( 1963 ) 21 .
temperature dependence of neutron transmission by he-4 . to be published . | the standard scattering theory ( sst ) in non relativistic quantum mechanics is analyzed .
contradictions of sst are revealed .
justification of sst in textbooks with the help of fundamental scattering theory ( fst ) is shown to be unconvincing .
a theory of wave packets scattering on a fixed center is presented , and its similarity to plane wave scattering is demonstrated .
the neutron scattering in a monatomic gas is investigated , and ambiguity of the cross section is found . the ways to resolve contradictions and ambiguity is suggested , and experiments to explore properties of wave packets are discussed . |
a complete description of the momentum and spin degrees of freedom of quarks and antiquarks in the nucleon requires , at leading twist , the definition of three sets of parton distributions .
two of them , the momentum distribution @xmath2 and the helicity distribution @xmath3 , have been intensively investigated in the last few years ( for recent reviews see @xcite ) while the so called transversity distribution , @xmath4 , has come to the attention of theorists and experimentalists more recently in the analysis of drell yan spin asymmetries @xcite .
the main reason why it has passed unnoticed for such a longtime is related to its chiral - odd character . if quark masses are neglected in the qcd lagrangian , no interaction at lowest order of perturbation theory , can flip chirality . as a consequence ,
transversity is strongly suppressed ( by powers of @xmath5 ) in deep inelastic lepton - nucleon scattering ( dis ) and in general in any hard process that involves only one parton distribution . in hadron - hadron collisions
the chirality of the partons that annihilate is uncorrelated and the previous restrictions do not apply .
recently there has been a number of proposals to measure the transversity parton distributions ( see @xcite for a review ) .
lepton pair production in doubly polarized drell - yan processes are among the proposed scenarios and such experiments are included in the research program at rhic and the hera-@xmath1 project @xcite . the feasibility of measurements of double transverse asymmetries ( @xmath6 ) at polarized @xmath7 colliders ( rhic ) and fixed - target experiments ( hera-@xmath1 ) has been recently studied @xcite .
the expected maximal value for @xmath6 at the kinematic range covered by rhic is around 1 - 2 % , which would be difficult to measure with the present acceptances . for hera-@xmath1 , @xmath6 is expected to reach values as large as 5 % .
along with the experimental prospects , many efforts have also been made on the theoretical side @xcite .
in particular jaffe and ji calculated @xmath8 within the bag model of the nucleon @xcite , pointing out the relativistic character of the differences between transverse and longitudinal polarization properties .
the calculation of leading order ( lo ) @xcite and next - to - leading order ( nlo ) @xcite anomalous dimensions allowed to address quantitatively the differences between @xmath9 and @xmath8 due to perturbative qcd ( pqcd ) evolution in different models @xcite . in particular , scopetta and vento have shown that pqcd differences are sizeable only for low values of @xmath10 ( @xmath11 ) @xcite .
on the other hand the boundary conditions of the evolution equations , that are provided by the low energy input at the hadronic scale @xmath12 ( @xmath13 gev@xmath14 ) , can yield an additional , non perturbative , difference . as it was stressed by jaffe and ji @xcite , at this scale
the equality @xmath15 is a typical outcome of non - relativistic models of the nucleon , in which motion and spin observables are uncorrelated . in other words ,
any departure from the previous identity is a signature of relativity in the employed hadronic model .
a complete theoretical study of @xmath8 and @xmath9 has to account for both : the relativistic effects which distinguish @xmath8 from @xmath9 at the non - perturbative scale , and the pqcd evolution which differs for the two structure functions .
aim of the present work is the quantitative study of the relativistic effects in @xmath8 and @xmath9 due to the correlations of spin and parton motion in the hadronic systems . for this purpose
we use light - front dynamics formalism in which the interplay between motion and spin is made explicit through the melosh rotations .
in particular , we make use of the light - front covariant ( lfc ) quark model of ref .
@xcite to compute the leading twist contribution to the matrix elements at the hadronic scale @xmath16 .
the non - perturbative input is then evolved , at nlo , up to a higher @xmath0 scale . we shall show that relativistic corrections introduced at @xmath16 clearly survive evolution
. measurements of observables involving transverse and longitudinal asymmetries might put in evidence the corrections to the naive non - relativistic spin picture adopted in low - energy models of the nucleon .
the paper is organized as follows : in section [ lfc ] the framework to calculate @xmath8 and @xmath9 within the lfc quark model is described and a specific observable , related to double spin asymmetry experiments , is defined .
the definition is such that the view on relativistic spin effects is optimized .
results in the kinematic range covered by rhic and hera-@xmath1 are presented in section [ res ] , where the feasibility of the detection of these effects is also discussed .
finally , conclusions are drawn in section [ concl ] .
the helicity distribution @xmath17 of a parton with flavour @xmath18 is defined as the probability of having a ( longitudinally ) polarized parton @xmath18 with spin parallel to the longitudinal polarization of the parent nucleon minus the probability of finding the parton polarized in the opposite direction .
a similar definition applies to @xmath19 for transverse polarizations @xcite . in the quark model
a simple approach can be developed to connect parton and momentum densities at the hadronic scale @xmath12 where valence degrees of freedom dominate the matrix elements at leading twist @xcite .
in particular , @xmath20 @xcite and @xmath21 @xcite can be related to the longitudinal and transverse quark momentum densities respectively .
namely : @xmath22 where @xmath23 is the three - momentum of the struck quark , @xmath24 is the nucleon mass , @xmath25 and @xmath26 is the density of ( valence ) quarks with momentum @xmath23 and longitudinal polarization aligned ( antialigned ) with the longitudinal polarization of the parent nucleon . a similar notation ( @xmath27 ) is used for transverse polarization .
the explicit expressions for the densities are : @xmath28 where @xmath29 is the flavour projector . in any non - relativistic description of the wave function
the densities ( [ naupdef ] ) and ( [ nafwdef ] ) are equal and hence @xmath30 as discussed by jaffe and ji @xcite .
however the degeneracy is removed by relativistic covariance requirements , such as those which are implemented in the covariant quark models based on light - front dynamics @xcite .
light - front dynamics with a fixed number of particles has been widely discussed in the literature ( see @xcite for reviews ) , so that it is enough to highlight briefly the main features of the angular momentum in this formalism . in relativistic quantum theories
, the angular momentum operator is obtained from the generators of the poincar group . in the light front form of the dynamics problems
arise when considering addition of angular momenta , because usual composition rules are not satisfied . nevertheless , it is possible to restore them by means of a unitary transformation : the melosh rotation @xcite . from the physical viewpoint
these transformations relate the angular momentum eigenstates in a subsystem rest frame ( a quark , for example ) to the centre of mass frame . for spin 1/2 particles the melosh rotations ( mr ) that link light - front to
canonical spin states are represented by : @xmath31 = \frac{(m+\omega+k_z ) - i \vec{\sigma } \cdot ( \hat{z } \times \vec{k}_\perp ) } { ( ( m+\omega+k_z)^2 + \vec{k}_\perp^{\,2})^{1/2 } } \;\ ; , \label{mr}\ ] ] where @xmath23 is the three - momentum of the particle and @xmath32 its relativistic energy .
the fact that @xmath33 \rightarrow 1 $ ] in the limit @xmath34 reveals the relativistic origin of the mr . a new spin - flavor basis
is then defined ( the pauli - melosh basis ) in which the ordinary ( canonical ) pauli spinors for each individual particle are replaced by the melosh - rotated spinors : @xmath35 \chi_i^c = \frac{(m_i+\omega_i+k_{i_z } ) - i \vec{\sigma}^{(i ) } \cdot ( \hat{z } \times \vec{k}_{i_\perp } ) } { ( ( m_i+\omega_i+k_{i_z})^2 + \vec{k}_{i_\perp}^{\,2})^{1/2 } } \chi_i^c \ ; , \label{melosh2}\ ] ] where @xmath36 are the usual ( canonical ) pauli spinors for the particle @xmath37 .
the pauli - melosh basis constitutes a minimal relativistic spin basis @xcite that , despite of its non - manifest covariance , is compatible with poincar invariance requirements . in our calculation
we shall refer to the wave function @xmath38 of ref .
@xcite , where the relativistic mass equation : @xmath39 has been solved .
the free relativistic mass operator @xmath40 is supplemented with an hypercentral phenomenological interaction @xmath41 where the hyper - radius @xmath42 is a function of what , in a non relativistic treatment , is the vector position of the particles @xmath43 and the center of mass position @xmath44 .
@xmath45 , @xmath46 and @xmath47 are constants fixed to reproduce the basic features of the low - energy baryonic spectrum pattern in the @xmath48 channels . since this interaction is invariant under rotations and
does not depend on the total light - front momentum , all the commutation relations that guarantee the covariance requirement ( see @xcite ) are correctly satisfied .
the advantage of the hypercentral interactions , well known in the non - relativistic limit @xcite , is that they allow a straightforward solution of the schrdinger equation .
the resulting spin - isospin wave function is @xmath49 symmetric .
it should be emphasized that the degeneracy between @xmath8 and @xmath9 in the non - relativistic limit is independent of the presence of @xmath49breaking terms in the nucleon wave function . from eq .
( [ melosh2 ] ) it is evident the correlation between spin and motion in the pauli - melosh spin basis , so that the momentum densities ( [ naupdef ] ) and ( [ nafwdef ] ) are no longer equal , independently of its su(6)-symmetric character .
differences between @xmath50 and @xmath51 are now proportional to the internal transverse momentum @xmath52 , as expected : @xmath53 \nonumber \\ & & \times n(\vec{k } ) \ ; \delta \left ( \frac{x}{1-x } - \frac{k^+ } { m/\sqrt{2}}\right ) \ ; , \label{diff}\end{aligned}\ ] ] where @xmath54 is the unpolarized flavorless quark momentum density normalized to the number of particles as obtained by solving the mass equation ( [ rme ] ) .
equation ( [ diff ] ) is one of the central results of the present work .
it shows that differences between @xmath21 and @xmath20 have a clear relativistic origin depending on the kinematical structure of the mr .
the @xmath10-dependence of the quantity @xmath55 is embodied in the wave function term @xmath54 . within a fully non - relativistic approach
@xmath55 would vanish and the actual ( and identical ) values of @xmath21 and @xmath20 would be largely suppressed for @xmath56 due to the lack of high momentum components in @xmath54 for typical non - relativistic models @xcite .
in fact the relativistic effects show up , in the present approach , in a twofold way : i ) mr introduce the non - vanishing difference @xmath55 of eq.([diff ] ) ; ii ) high momentum components in @xmath54 which are mostly originated by the relativistic kinetic energy in the mass equation ( [ rme ] ) and that reinforce the mr contribution , as discussed in ref.@xcite for @xmath20 . concerning the scale dependence
, it should be stressed that eq.([diff ] ) is valid at a low - energy ( hadronic ) scale @xmath12 , where the non - perturbative models used to calculate the twist - two matrix elements can be applied . at higher ( experimental ) scales @xmath57
perturbative qcd evolution yields an additional contribution to the differences between longitudinal and transverse polarization observables ( cfr .
section [ res ] ) .
previous attempts @xcite to study the effects of mr in @xmath8 disregarded this scale dependence and did not include perturbative contributions .
let us now discuss a specific combination of longitudinally and transversely polarized parton distributions , namely the ratio @xmath58 where @xmath59 is the charge of a quark of flavour @xmath18 .
the arguments @xmath60 and @xmath61 are related , for drell - yan processes , to the center of mass energy @xmath62 , the invariant mass of the produced lepton pair @xmath0 , and the rapidity @xmath63 : @xmath64 @xmath65 does not dependent on the unpolarized parton distribution and involves only the ratio between @xmath9 and @xmath8 . to this respect
it is less model dependent than the single terms and is quite transparent for studying mr effects . the flavor combination of transverse parton distributions in eq .
( [ rdef ] ) appears in the lo double transverse asymmetry of drell - yan processes : @xmath66 where the arrows denote the transverse polarization of the beam and target and @xmath67 ( @xmath68 ) is the polar ( azimuthal ) angle of the detected lepton .
the corresponding lo asymmetry with longitudinally polarized hadrons , @xmath69 , is proportional to the combination of helicity distributions @xmath9 considered in eq .
( [ rdef ] ) : @xmath70 then , @xmath71 is just the ratio between double transverse and double longitudinal asymmetries at lo : @xmath72 the evolution procedure we are going to use is a nlo procedure and therefore the parton distribution and the corresponding cross sections should be consistently evaluated at nlo . as a consequence the asymmetries do not assume the simple forms ( [ attdef ] ) and ( [ alldef ] ) .
nevertheless it is still feasible to establish a rather close connection between the combinations that enter eq .
( [ rdef ] ) and the nlo asymmetries .
indeed , the differential cross section for longitudinally polarized process is written , at nlo , as @xcite : @xmath73 where @xmath74 is the polarization of gluons and the coefficient @xmath75 .
the nlo order coefficients are explicitly given in @xcite .
the nlo order introduces on the one hand the polarization of the gluons due to the subprocess @xmath76 and on the other hand the coefficients @xmath77 that contain terms which are not proportional to @xmath78 and therefore break the relationship ( [ xdefs ] ) we have used in our calculation .
a similar modification appears for the unpolarized dy cross section .
nevertheless , an estimation of the asymmetry @xmath69 by using available parameterization of polarized parton distributions indicate that the nlo asymmetry is dominated by the @xmath79 of the @xmath80 subprocess in the central rapidity region .
this term contains just the combination of polarized parton distributions that we are interested in and which enters in the ratio ( [ rdef ] ) .
alternatively , it would also be possible to extract this combination of helicity parton distributions from other experiments . in the transversely nlo polarized cross sections , gluons are absent since there is no equivalent transversity for gluons , and
hence @xcite : @xmath81 though the cross section is proportional to the combination of transversity distributions that we have used in ( [ rdef ] ) , the nlo order coefficient @xmath82 may spoil the relationship ( [ xdefs ] ) again .
however , it was shown in ref.@xcite , that the dominant term in this coefficient ( and in the equivalent one for the unpolarized cross section ) is proportional to @xmath78 , so that it is still possible a rather direct extraction of the combination @xmath83 from the analysis of the transverse asymmetry at nlo .
to reach the experimental regime we have evolved the calculated parton distributions from the hadronic scale up to a scale @xmath84 by using pqcd at nlo , in the @xmath85 scheme .
the sensitivity to the factorization scheme in this approach has been tested in previous calculations @xcite and found to be generally small .
the value of @xmath12 ( the hadronic scale ) is fixed @xcite by evolving backwards the parametrized experimental fits for unpolarized parton distributions @xcite , to the scale where the valence quarks carry the whole momentum of the nucleon .
we found @xmath86 gev@xmath14 and the reliability of the pqcd evolution procedure at such low scale was studied in detail in @xcite .
let us emphasize that the value @xmath87 gev@xmath14 is found by using an evolution code where the transcendental equation @xmath88 = 0 \label{trasceq}\ ] ] has been solved to get the nlo coupling constant .
the use of the full nlo expression ( [ trasceq ] ) is mandatory when evolving from / to a low energy scale @xcite and the approximate solution @xmath89 \;\ ; , \label{appeq}\ ] ] which is valid for large values of @xmath90 and often used in nlo codes ( e.g. ref.@xcite ) , would introduce quite spurious effects .
in particular , the evolution to energy scales as low as the hadronic point @xmath12 may yield a negative gluon contribution at @xmath12 , related to a large extent to the spurious effects introduced by eq.([appeq ] ) . in the present case , for example , evolving back the experimental fit @xcite to the hadronic scale @xmath87 gev@xmath14 , would give a fraction of momentum carried by the gluons of @xmath91 , which , although non - zero , is relatively small .
obviously , the situation is much worse when eq .
( [ appeq ] ) is used instead of eq .
( [ trasceq ] ) .
in addition , the evolution code to be used for hadronic model calculations must guarantee complete symmetry for the forward and backward paths @xmath92 and vice versa , as implied by genuine perturbative qcd expansion at nlo .
additional approximations associated with taylor expansion valid for large @xmath93 , must be avoided , as discussed in @xcite . in fig .
1 we show the curves for @xmath94 and @xmath95 at the hadronic scale @xmath12 ( fig .
1(a ) ) and at the partonic scale @xmath96 gev@xmath14 ( fig .
a remarkable difference between @xmath97 and @xmath98 appears at large @xmath10 , reaching a peak at @xmath99 .
quantitatively they are bigger that those obtained within bag models @xcite .
it is clear that the probability of transverse polarization is larger than the longitudinal one when relativistic effects are considered .
it is possible to appreciate better the mr effects by simply setting @xmath33 \rightarrow 1 $ ] . in this case
the remaining relativistic ingredient is the high momentum components generated by the relativized schrdinger equation ( [ rme ] ) .
the results are also shown in fig .
1(a ) and we get @xmath100 as expected .
after evolution ( fig .
1(b ) ) we see that when the mr are neglected , @xmath9 and @xmath8 differ mainly at low @xmath10 ( @xmath11 ) because of pqcd evolution @xcite , while the inclusion of the correlations between spin and motion produce large effects also in the medium and large @xmath10 region . in order to study the effects of mr in the combination of polarized parton distributions we are interested in ,
let us first consider the ratios @xmath101 and @xmath102 shown in fig .
2(a ) and fig .
2(b ) respectively as a function of the center of mass rapidity @xmath103 in a kinematic region ( @xmath104 gev , @xmath96 gev@xmath14 ) accessible by rhic experiments .
these values are of the order of a few percent , compatible with those obtained in the literature ( see for instance @xcite , where antiquarks are considered also at the hadronic scale ) . again it is possible to single out the contribution introduced by mr switching them off . in general , mr reduce both @xmath105 and @xmath106 but this reduction is far more significant for @xmath106 . in fig .
3 we show the ratio @xmath107 as a function of @xmath103 ( in the same kinematical region ) when considering melosh rotations and when they are neglected .
figure 3 represents our reference point to assess the importance of relativistic spin effects : they enhance the transverse ratio @xmath108 with respect to the longitudinal term @xmath109 .
namely , if melosh rotations are taken into account @xmath110 in the considered kinematic regime . on the contrary , if the spin - motion correlations are neglected ( non - relativistic limit ) , then in the same region @xmath111 . experiments may decide between these two alternatives , therefore probing the relevance of covariance effects in the dynamical description of the nucleon spin . before examining the possibility of measuring these effects
let us discuss some issues concerning the previous results : the coupling of @xmath9 to the polarized gluons in the evolution at nlo and the dependence of @xmath71 on the spatial nucleon wave function . in our valence model of the nucleon ,
the contribution to @xmath112 and @xmath113 comes from evolution only . on qualitative grounds one would expect @xmath114 ( i.e. @xmath115 ) , since @xmath116 at lo vanishes while @xmath117 receives contributions also from the lowest order in @xmath118 .
however , a careful analysis shows that this is not necessarily the case , since the final value of @xmath112 is not only determined by the order of @xmath118 but also from the behavior of the anomalous dimensions .
more specifically , in order to understand why one gets @xmath119 , let us consider the dglap equations for the longitudinally polarized parton distributions @xcite to disentangle the coupling of @xmath117 to quarks , antiquarks and gluons : @xmath120 @xmath121 where @xmath122 and @xmath123 corresponds to non - singlet combinations of @xmath20 and @xmath124 . for the sake of simplicity we will give results for @xmath125 at @xmath96 gev@xmath14 (
conclusions are not changed for other @xmath0 points we have checked ) at lo . in fig .
4 we show the contribution to the derivative of @xmath126 due to the coupling to @xmath74 and to the terms proportional to the quarks and antiquarks at lo .
the remarkable feature is the opposite sign of the coupling to the gluons and to the quarks+antiquarks at lo that makes the final derivative be not too large for @xmath127 .
this partial cancellation between different contributions produces a rather slow increase of @xmath128 with @xmath0 at order @xmath129 and the final value of @xmath128 at nlo are comparable to that of @xmath130 .
this happens for @xmath127 , while for smaller @xmath10 the growing of @xmath126 dominates over that of @xmath131 , due to some extent to the large size of the coupling to gluons , which is absent in the evolution of @xmath131 .
it should be reminded that @xmath71 vanishes at leading order since the sea quarks are entirely generated through evolution .
this fact prevent us from doing a quantitative comparison between the lo and nlo for @xmath71 .
however , scopetta and vento @xcite analyzed the lo and nlo evolution for other flavour combinations of @xmath8 and @xmath9 and it was found that these differences between lo and nlo were not too large and went in the same direction for @xmath9 and @xmath8 .
at this point one can argue that these results for @xmath65 could be strongly influenced by the presence of high momentum components in the spatial part of the wave function , carried over by the presence of a relativistic kinetic energy in the mass operator .
these components have been proved to be essential to describe electromagnetic transitions and elastic form factors @xcite and also the high @xmath10 region in dis @xcite .
we have therefore repeated the calculation of the ratio @xmath65 starting from a nucleon wave function obtained by making use of a a non - relativistic kinetic energy in the mass operator ( cfr .
( [ rme ] ) ) .
the interaction has been kept of the same form ( [ potential ] ) , while the values of the parameters ( @xmath45 , @xmath46 , @xmath47 ) have been reset according to ref @xcite . the corresponding result for @xmath65
is shown also in fig .
3 . we can see that the ratio obtained from a fully non - relativistic model , lies reasonably close to the curve obtained employing a relativistic kinetic energy but neglecting melosh rotations .
hence , we can safely conclude that @xmath65 is rather insensitive to the details of the mass operator , but very sensitive to the relativistic spin corrections derived from a light - front dynamics .
this is not so evident for the values of @xmath105 or @xmath106 which are roughly doubled by the same change in the mass operator ( see fig .
2 ) . this fact gives support to the suitability of the chosen observable .
moreover , due to this insensitivity to the details of the spatial wave function , we do not expect a dramatic change of our conclusions for other more sophisticated potential models @xcite . in order to study the feasibility of the experimental detection of the differences for @xmath71 we have to estimate the expected errors for @xmath71 under the conditions of rhic and hera@xmath1 .
we have calculated the statistical errors according to the expression @xcite : @xmath132 where @xmath133 ( @xmath134 ) is the degree of polarization of the beam ( target ) , @xmath135 is the expected luminosity and @xmath136 is the efficiency in the detection of events .
the unpolarized cross section @xmath137 is integrated over bins of @xmath138 and @xmath103 . for rhic we have taken @xcite @xmath139 and a luminosity of 240 pb@xmath140 .
the calculated errors for the figure 3 in the central rapidity region largely exceeds the separation between the two main curves , even by assuming a 100 % efficiency and integrating over an interval @xmath141 and @xmath142 around the central values .
therefore , these effects will be unlikely to be observed at rhic . for hera@xmath1
the expected degrees of polarization are @xcite @xmath143 and @xmath144 with a projected luminosity of @xmath145 pb@xmath140
. the center of mass energy will be @xmath146 gev , corresponding to a @xmath147 gev .
we have checked that , though the error bars are far smaller than those obtained for rhic , relativistic spin effects can not be disentangled for @xmath71 in the region @xmath148 gev , @xmath149 . in order to maximize the rates
we will integrate over the whole range of @xmath103 : @xmath150 the results for this ratio for the kinematics of hera@xmath1 are shown in figure 5 , where we can observe the same pattern of differences as the one seen in the y - dependent ratio ( fig .
the relative insensitivity to the details of the chosen potential is also evident in this representation . in the error bars shown in fig .
5 we have taken into account the limited acceptance of the detectors , as explained in @xcite , which roughly gives @xmath151 in the considered region .
we have also assumed that @xmath152 . a measurement in the region @xmath153 gev can not single out which is the right spin - flavor basis , while experiments in the low mass region ( @xmath154 gev ) could be more selective , though some overlap between the error bars still persists .
for rhic the acceptance corrections ( @xmath155 ) are too large to reveal relativistic differences .
we have shown that drell - yan processes can probe the relativistic effects embodied in a light - front dynamical description of the low energy spin structure of the nucleon .
for this purpose , the ratio @xmath71 turns out to be particularly suitable since it is largely independent of the structure of the mass operator and , therefore , of the choice of the phenomenological interaction .
this insensitivity to the details of the spatial degrees of freedom is not so evident for each asymmetry @xmath105 and @xmath106 separately . while predictions for @xmath71 are robust , the given values for @xmath105 and @xmath106 should be considered just at a semiquantitative level .
in fact , we have seen that relativistic spin effects would reduce @xmath105 and therefore the chance of measure it with respect to the maximal scenario presented in @xcite , but no definite quantitative conclusions about the feasibility of measuring transversity should be drawn from fig .
2 . by estimating the expected statistical errors for rhic and hera we have concluded that for hera@xmath1 it would be possible to determine the right spin - flavour basis for the nucleon wave function in the low mass region while this kind of measurement would be far more difficult at rhic . in the present work
we have limited ourselves to give predictions for d - y related observables in some kinematic regions accessible to rhic , but the model can be straightforwardly applied to other experimental conditions and to other kind of processes as well . in particular , the observed enhancement of the transverse polarization with respect to the longitudinal one may have some impact in the extraction of the twist-3 contribution to the spin structure function @xmath156 from dis experiments @xcite ( see also @xcite ) .
finally let us emphasize that we have used a valence quark model to evaluate the parton distributions at the hadronic scale , neglecting sea quark ( and antiquark ) distributions at that scale .
one could argue that the consideration of a non - vanishing @xmath157 and @xmath158 would mask the effects due to melosh rotations . though a quantitative study of the change in the initial conditions would require more elaborate models ( beyond the valence picture @xcite ) there are , however , some qualitative reasons to support the stability of the results presented in figs . 3 and 5 . a non - vanishing @xmath157 ( and @xmath158 )
has two main consequences on the overall scheme : firstly it slightly raises the initial scale @xmath16 and on the other hand it changes the small @xmath10 behavior of the parton distributions .
the ( small ) increment of the initial scale would produce a little shift in the @xmath0 scale in fig .
5 , but not large enough to give rise to a confusion between the two main curves .
with respect to the change of the small @xmath10 behavior , it should be noticed that in the central rapidity regions we are exploring ranges of @xmath10 around @xmath159 and also for the y - integrated observables this central region is dominant .
furthermore , since we deal with ratios we expect the influence of these changes to be minimized . 0.5 cm
we are grateful to s. scopetta for useful suggestions and for providing us with an updated version of the pqcd evolution code .
we thank v. vento for his encouragement and help during the present study and for a critical reading of the manuscript .
we also thank o. martin for useful correspondence about the calculation of the statistical errors and m. miyama for sending us his evolution code for @xmath8 @xcite .
c. bourrely and j. soffer , nucl .
b423 * , 329 ( 1994 ) ; nucl .
* b445 * , 341 ( 1995 ) ; p. pobylitsa and v. polyakov , phys . lett . *
b389 * , 350 ( 1996 ) ; b.i .
ioffe and a. khodjamirian , phys .
d * 51 * , 3373 ( 1995 ) ; r. kirschner , l. mankiewicz , a. schfer and l. szymanonski , z. phys . * c74 * , 501 ( 1997 ) ; r. jakob , p.j . mulders and j. rodrigues ,
a626 * , 937 ( 1997 ) .
keister and w.n .
polyzou , adv . in nucl
. phys . * 20 * , 225 ( 1991 ) ; f. coester , prog .
. phys . * 29 * , 1 ( 1992 ) ; j. carbonell , b. desplanques , v.a .
karmanov and j .- f .
mathiot , phys . rep .
* 300 * , 215 ( 1998 )
. * figure 1 * helicity and transversity distributions for the @xmath160 quark ( a ) at the hadronic scale @xmath86 gev@xmath14 and ( b ) after evolution up to @xmath96 gev@xmath14 . in fig .
a ) the solid line corresponds to @xmath161 , the dashed line to @xmath162 and the dotted line is the result when melosh rotation is not considered ( @xmath163 ) . in fig .
( b ) the solid and dashed lines represent @xmath8 and @xmath9 respectively .
the dotted and dash - dotted lines correspond to @xmath8 and @xmath9 when melosh rotation is neglected . * figure 2 .
* @xmath109 ( a ) and @xmath108 ( b ) ratios as defined in eqs .
( [ rldefinition],[rtdefinition ] ) as a function of the center of mass rapidity @xmath103 for @xmath96 gev@xmath14 and a center of mass energy @xmath164 gev ( solid line ) .
the dashed line corresponds to the case when the melosh rotation is switched off .
the dotted line is the result obtained in the non - relativistic model discussed in the text .
* figure 3 .
* ratio between transverse and longitudinal parton distributions ( eq . ( [ rdef ] ) ) as a function of the center of mass rapidity @xmath103 for @xmath96 gev@xmath14 and a center of mass energy @xmath164 gev .
notation as in fig .
* figure 4 .
* contributions to @xmath165 at lo coming from the terms proportional to the gluons ( solid line ) and to the quarks + antiquarks ( dashed line ) .
all the results correspond to @xmath166 gev@xmath14 .
* figure 5 .
* ratio between transverse and longitudinal parton distributions , eq .
( [ ryintdef ] ) , as a function of the invariant mass of the produced lepton pair ( @xmath0 ) at a center of mass energy corresponding to hera@xmath1 ( @xmath167 ) gev .
notation as in fig .
error bars have been calculated at lo and include acceptance corrections .
error bars in the lower curve have been slightly shifted to appreciate the overlap . | significant differences between transverse and longitudinal polarized parton distributions are found at low energies within a light - front covariant quark model of the nucleon .
these differences are due to relativistic spin effects introduced by the melosh rotations and survive evolution to higher @xmath0 scales . a specific observable related to double - spin asymmetries in lepton pair production in polarized hadron - hadron collisions
is defined .
the possibility of assessing the relevance of these relativistic spin effects in future experiments at rhic and hera@xmath1 is discussed . |
in a recent stimulating article with important implications for the use of finite basis sets , staroverov , scuseria , and davidson @xcite presented an exchange - only optimized effective potential ( xoep ) scheme that yields , for given finite gaussian orbital basis sets , ground state energies that surprisingly equal exactly the ground state hartree - fock ( hf ) energies for these basis sets .
moreover , their xoep scheme not only yields one unique but an infinite number of exchange potentials and each of the latter leads to the corresponding ground state hf energy if used as the exchange potential in the corresponding exchange - only ks hamiltonian operator . on the other hand
, it is known that in a complete basis set limit , which corresponds to a complete real space representation of all quantities , the xoep method is identical @xcite to the exact exchange - only kohn - sham method and yields ground state energies that always lie above @xcite the corresponding ground state hf energy .
staroverov , scuseria , and davidson then state : `` our conclusions may appear paradoxical . for any finite basis set ,
no matter how large , there exist infinitely many xoep s that deliver exactly the ground - state hf energy in that basis , however close it may be to the hf limit .
nonetheless , in the complete basis set limit , the xoep is unique and e(xoep ) is above e(hf ) '' .
( here e(xoep ) and e(hf ) denote the xoep and hf total energies , respectively , that are denoted @xmath0 and @xmath1 in this work . )
furthermore they state : `` the non - uniqueness of oeps in a finite basis set raises doubt about their usefulness in practical applications '' we here first show , by different means including a constrained - search one , that the above statement of staroverov , scuseria , and davidson , that it is always possible to construct optimized effective potentials that deliver exactly the ground state hf energy , holds if and only if the products of the orbital basis functions , or at least the products of the corresponding occupied and unoccupied hf orbitals from a given orbital basis set , form a linearly independent set .
otherwise , the xoep scheme for finite orbital basis sets , in general , does not deliver exactly the ground state hf energy .
secondly , we show that the xoep approach of staroverov , scuseria , and davidson , does not really represent an exchange - only ks method and does not yield physically meaningfull ks exchange potentials , even if the products of orbital basis functions are linearly independent . in order to get physically meaningfull ks exchange potentials via xoep schemes ,
the latter have to be set up in a way that they represent ks methods , otherwise they are indeed of little usefulness in practical applications .
however , if xoep schemes are set up properly then they are of great usefulness in practice as demonstrated , e.g. , by numerically stable plane - wave xoep procedures for solids @xcite .
we start by briefly reconsidering the xoep approach of staroverov , scuseria , and davidson @xcite .
the relevant hamiltonian operators are the hf hamiltonian operator @xmath2 and the exchange - only ks hamiltonian operator @xmath3 atomic units are used throughout . in eqs .
( [ hf - hamiltonian ] ) and ( [ ks - hamiltonian ] ) , @xmath4 denotes the external potential , usually the electrostatic potential of the nuclei , @xmath5 is the hartree potential , i.e. , the coulomb potential of the electron density , @xmath6 is the local multiplicative ks exchange potential , @xmath7 the effective ks potential , and @xmath8 the nonlocal exchange operator with the kernel @xmath9 in a real space representation . here
@xmath10 designates the first - order density matrix . in the hf - hamiltonian operator of eq.([hf - hamiltonian ] ) the first - order density matrix occuring in the nonlocal exchange operator of eq .
( [ hf - kernel ] ) equals the hf first order density matrix @xmath11 and the nonlocal exchange operator subsequently equals the hf exchange operator . for simplicity
we consider closed shell systems with non - degenerate ground states . in this case
orbitals , first - order density matrices , and basis functions can all be chosen to be real - valued .
next we introduce an orbital basis set @xmath12 of dimension @xmath13 .
the representations of the hf- and exchange - only ks - hamiltonian operators in this basis set are @xmath14 and @xmath15 respectively .
the matrices @xmath16 , @xmath17 , @xmath18 , @xmath19 and @xmath20 are defined by the corresponding matrix elements @xmath21 , @xmath22 , @xmath23 , @xmath24 , and @xmath25 , respectively , and by @xmath26 . because the orbital basis functions are real - valued all matrices are symmetric
now we expand the ks exchange potential in an auxiliary basis set @xmath27 of dimension @xmath28 , i.e. , @xmath29 the auxiliary basis set , of course , shall be chosen such that its basis functions are linearly independent .
the crucial question arising now is how many and what types of matrices @xmath20 representing the ks density - functional exchange potential can be constructed for a given auxiliary basis set @xmath27 .
this question was answered in ref .
firstly we consider the case when the @xmath30 different products @xmath31 of orbital basis functions are linearly independent . in this case , if @xmath32 and the auxiliary basis functions span the same space as the products of the orbital basis functions , then a symmetric matrix @xmath20 can be constructed in a unique way by determining appropriate expansion coefficients @xmath33 for the exchange potential .
the reason is that the determination of the @xmath32 expansion coefficients @xmath33 for the construction of the @xmath32 different matrix elements of the symmetric matrix @xmath20 leads to a linear system of equations @xmath34 with @xmath35 for the coefficients @xmath33 of dimension @xmath32 that is nonsingular and thus has a unique solution @xcite . in eq .
( [ equationforxcoefficients ] ) , @xmath36 is a @xmath37 matrix that contains the overlap matrix elements @xmath38 .
the first index of @xmath36 , i.e. , @xmath39 , is a superindex refering to products of orbital basis functions while the second index , @xmath40 , refers to auxiliary basis functions .
the vector @xmath41 collects the expansion coefficients of eq .
( [ vx - aux ] ) for the exchange potential and the right hand side @xmath42 , a vector with superindices @xmath39 , contains the @xmath43 independent elements of an arbitrarily chosen matrix @xmath20 . if we chose @xmath20 to be equal to the matrix representation of an arbitrary nonlocal operator with respect to the orbital basis set then eqs . ( [ vx - aux ] ) and ( [ equationforxcoefficients ] )
define a local potential with the same matrix representation .
this demonstrates that a distinction of local multiplicative and nonlocal operators is not clearly possible for orbital basis sets with linearly independent products of orbital basis functions .
if @xmath44 and the space spanned by the auxiliary functions contains the space spanned by the product of orbital functions then @xcite an infinite number of sets of coefficients @xmath33 lead to any given symmetric matrix @xmath20 .
the real space ks exchange potentials @xmath45 corresponding according to eq .
( [ vx - aux ] ) to these sets of coefficients @xmath33 are all different but all represent local multiplicative potentials .
next we construct ks hamiltonian operators ( [ ks - hamiltonian ] ) by adding these different ks exchange potentials to always the same external and hartree potential .
the resulting effective ks potentials in real space , i.e. , the @xmath46 are all different . nevertheless the resulting basis set representations @xmath47 of the corresponding ks hamiltonian operators are all identical because the basis set representations @xmath48 of the different exchange potentials @xmath6 , by construction , are all identical . as a consequence the ks orbitals resulting from diagonalizing the ks hamiltonian matrix @xmath47 and subsequently also the resulting ground state electron densities are identical in all cases .
we thus have a situation where different local multiplicative ks potentials @xmath46 lead to the same ground state electron density .
this seems to constitute a violation of the hohenberg - kohn theorem .
indeed it was shown in ref .
[ ] and discussed in ref .
[ ] that the hohenberg - kohn theorem does not hold for finite orbital basis sets in its original formulation , i.e. , that different local potentials , e.g. , local potentials obtained by different linear combinations of auxiliary basis funtions , must lead to different ks determinants and thus different ks electron densities . we will come back to this point later on . finally , if @xmath49 then not all symmetric matrices @xmath20 can be constructed from a local ks exchange potential given by an expansion ( [ vx - aux ] ) . in their xoep approach staroverov , scuseria , and davidson @xcite can expand the ks exchange potential in @xmath32 auxiliary basis functions and determine the coefficients such that the resulting matrix @xmath48 exactly equals the hf exchange matrix @xmath19 .
if additionally the ks hartree potential is set equal to the hf one then the resulting hf and ks hamiltonian operators are identical . subsequently also the hf and ks orbitals , the ground state electron densities , and the ground state energies are identical . because the hf and the ks electron densities turn out to be identical , the coulomb potential of this density can equally well be considered as a hf or a ks hartree potential .
it follows immediately that the ks exchange potential constructed in this way is the xoep exchange potential : the hf total energy is the lowest total energy any slater determinant can yield .
thus if a local multiplicative ks potential leads to this total energy it is clearly the optimized effective potential defined as the potential that yields the lowest total energy achievable by any local multiplicative ks potential .
the xoep ground state energy resulting from this construction equals the corresponding hf energy .
moreover by enlarging the number of auxiliary basis functions , resulting in @xmath44 , not only one optimized exchange potential leading to the hf energy but infinitely many can be constructed .
staroverov , scuseria , and davidson obtained the hf energy in their xoep scheme even if the number of auxiliary functions only equaled the product @xmath50 of occupied and virtual orbitals @xcite . in this case a similarity transformation of the hf and the ks hamiltonian matrices and their constituents was carried out in order to obtain representations of all matrices with respect to the hf orbitals .
then it is sufficient to chose the expansion coefficients of the ks exchange potential such that only the occupied - virtual block of the ks exchange matrix equals that of the hf exchange matrix .
the resulting ks hamiltonian matrix then may differ from the hf hamiltonian matrix in the occupied - occupied and the virtual - virtual block but this merely leads to unitary transformations of the occupied and virtual orbitals among themselves and thus does not change the ground state energy or the electron density .
indeed it is straightforward to show that the occupied - virtual block of the exchange matrix equals that of the hf exchange matrix if the products of occupied and unoccupied orbitals are linearly independent . to that end
we consider the xoep equation determining the xoep exchange potential @xcite @xmath51 in eq .
( [ exxb ] ) @xmath52 and @xmath53 denote occupied and unoccupied ks orbitals , respectively , with eigenvalues @xmath54 and @xmath55 . both sides of eq .
( [ exxb ] ) are a linear combination of products @xmath56 of occupied and unoccupied ks orbitals with coefficients @xmath57 and @xmath58 , respectively . however , if the products @xmath56 are linearly independent then the two linear combinations can only be identical if the coefficients multiplying the products are all identical .
this , however , requires that @xmath59 , i.e. , that the occupied - virtual block of the ks exchange matrix equals that of the corresponding exchange matrix of a nonlocal exchange operator of the form of the hf exchange operator .
replacement of the ks exchange matrix by the matrix of the nonlocal exchange operator thus again leads only to a unitary transformation of the occupied and virtual orbitals among themselves .
therefore the corresponding xoep determinant can also be interpreted as hf determinant .
next we consider the crucial point what happens if the products of orbital basis functions @xmath31 are linearly dependent .
then the rows of the matrix @xmath36 of eq.([equationforxcoefficients ] ) are linearly dependent , thus the rank of the matrix @xmath36 is lower than @xmath60 , and as consequence eq.([equationforxcoefficients ] ) , in general , has no solution .
for an alternative argument , observe that for linear dependent products of orbital basis functions @xmath31 , there exists at least one linear combination of such products that equals zero @xmath61 in eq .
( [ linear - dependence ] ) the @xmath62 denote the coefficients of that linear combination .
the corresponding sum of matrix elements of @xmath48 also equals zero , i.e. , @xmath63 for any choice of expansion coefficients @xmath64 in eq .
( [ vx - aux ] ) because the product of any local function and thus of any ks exchange potential @xmath6 with the sum ( [ linear - dependence ] ) equals zero .
the products @xmath65 for two different arguments @xmath66 and @xmath67 , on the other hand , are always linearly independent because the orbital basis set @xmath12 has to be linearly independent
. therefore the linear combination @xmath68 can not be identical to zero for all values of the arguments @xmath66 and @xmath67 .
then , however , also the integral of this linear combination with @xmath69 , i.e. , with the kernel of the nonlocal hf exchange operator , in general , is not equal to zero , i.e. , in general @xmath70 comparison of eqs .
( [ zero - matrix - elements ] ) and ( [ zero - matrix - elementshf ] ) shows that , in general , the exchange matrices @xmath20 and @xmath19 are different no matter how the expansion coefficients @xmath33 of the ks exchange potential , eq .
( [ vx - aux ] ) , are chosen .
this demonstrates that , in general , neither the xoep scheme of staroverov , scuseria , and davidson @xcite nor any other leads to an xoep hamiltonian operator that equals the hf hamiltonian operator when the orbital basis products are linearly dependent .
if we consider the version of staroverov , scuseria , and davidson s xoep scheme that refers only to the occupied - virtual block of the xoep and hf exchange matrices then by completely analogous arguments it follows that this scheme only works if the products of occupied and unoccupied hf orbitals are linearly independent .
however , in general , if the products of occupied and unoccupied hf orbitals are linearly dependent then it is not possible to obtain the hf ground state energy via an xoep scheme . before we discuss the question how products of basis functions can become linearly dependent for given orbital basis sets we elucidate the situation from a constrained - search @xcite point of view
we start with a constrained - search proof that the xoep ground state energy , @xmath0 , must equal the hf ground state energy , @xmath1 , in their common finite orbital basis , when there is no linear dependence in the products of orbital basis functions . to accomplish this we appeal to the work of harriman @xcite .
he showed that only one first - order density matrix may yield any density generated by a given finite orbital basis whose basis products form a linearly independent set .
this means that since an idempotent first - order density matrix uniquely fixes a corresponding single determinant , it follows that only one single determinant , constructed from a given finite orbital basis whose products are linearly independent , may yield a density that is constructed from this same basis .
consequently , with use of a common finite orbital basis set , the xoep single determinant must equal the hf single determinant if there exists an effective ks potential @xmath71 in eq ( [ ks - hamiltonian ] ) such that the corresponding ks ground state density is the same as the hartree - fock density . that this @xmath71 exists for the situation when the basis products are linearly independent , as discussed above , follows from ref .
[ ] and was shown in practice by staroverov , scuseria , and davidson @xcite . what happens when the products are not linearly independent ?
due to the idempotency property of the first - order density matrix for a single determinant , a density generated from a given finite orbital basis could still generate a unique determinant if the basis products are linearly dependent , provided that this linear dependency is mild enough @xcite , i.e. , if the products of occupied and unoccupied orbitals remain linearly independent .
however , if the linear dependency of the basis product pairs is not sufficiently mild , then the situation changes dramatically in that more than one single determinant will yield the same density from a given finite basis set @xcite . in this case
we do not have equality @xmath72 .
instead , we have inequality @xmath73 , which arises from the following contradiction .
assume that the xoep determinant @xmath74 equals the hf determinant @xmath75 through respective optimizations in their common finite orbital basis set .
then it follows that their densities must be the same .
but , from a constrained - search analysis @xcite , the xoep determinant @xmath74 would yield this hf density and minimize , within this common basis , just the expectation value @xmath76 of the kinetic energy , while the hf determinant @xmath75 yields this hf density and minimizes , within the common basis , the expectation value @xmath77 of the kinetic energy plus the electron - electron repulsion energy .
here @xmath78 denotes the many - electron kinetic energy operator , @xmath79 the corresponding electron - electron repulsion operator , and @xmath80 slater determinants that yield the hf density .
( equivalently , the xoep determinant would yield the hf density and minimize @xmath81 while the hf determinant yields this hf density and of course minimizes @xmath82 . here
@xmath83 denotes the many - electron hamiltonian operator . ) . because the slater determinants @xmath74 and @xmath75 minimize different expectation values , i.e. , @xmath84 , respectively , they are different , in general , and the inequality @xmath85 applies for this common finite orbital basis case .
however , there is only one possible determinant @xmath80 that yields the hf density from a given finite basis when the basis products are linearly independent or the extent of linear dependency is weak . in this case
both minimizations yield this one slater determinant simply because both minimization only run over one slater determinant .
thus there is no contradiction and the finite basis set conclusion of staroverov , scuseria , and davidson follows in that the equality @xmath72 applies .
hence we are now able to provide the resolution of the xoep paradox @xcite stated by staroverov , scuseria , and davidson : for a finite basis set case , no matter how large the basis , @xmath0 equals @xmath1 provided that the basis products form a linearly independent set or the extent of linear dependence is sufficiently weak .
however , in going from any starting finite basis set to the complete basis set limit , @xmath0 may become greater than @xmath1 somewhere along the way because as more and more basis orbitals are added to the finite basis set , the onset of sufficient linear dependency eventually occurs ( see appendix ) .
we have provided an explanation for what might very well seem counterintuitive to the reader without knowledge of the analysis provided here .
as one keeps adding more and more orbital basis functions , both @xmath1 and @xmath0 decrease and they continually remain equal to each other . past a certain critical point in the addition of orbital basis functions , however , @xmath1 and @xmath0 may start to differ from each other and @xmath1 keeps decreasing while the behavior of @xmath0 depends on the chosen orbital basis set and it might actually be that @xmath0 rises ! the latter behavior for example occurs if the exact hf orbitals as they correspond to a real space representation are themselves chosen as the basis set . if the basis set is restricted to the occupied hf orbitals , @xmath1 and @xmath0 are of course equal .
if unoccupied hf orbitals are added to the basis set , @xmath1 remains unchanged at first .
in contrast , beyond a certain point @xmath0 raises .
the cause , of course , is the appearance of sufficient linear dependence at the critical point .
[ ] does analyse certain linear dependency situations but the authors do not discuss the energy consequences for finite basis sets . ) next we consider how products of orbitals basis functions become linearly dependent .
as example we consider a plane wave basis set corresponding to a unit cell defined by the three linearly independent lattice vectors @xmath86 , @xmath87 , and @xmath88 . the plane waves representing the orbital basis set @xmath89 then are given by @xmath90 with @xmath91 and @xmath92 in eq .
( [ g ] ) , @xmath93 , @xmath94 , @xmath95 denote three reciprocal lattice vectors defined by the conditions @xmath96 for @xmath97 .
by @xmath98 the space of all integer numbers is denoted , @xmath99 denotes the cutoff that determines the size of the plane wave basis set , and @xmath100 stands for the crystal volume .
we have assumed before that basis functions are real - valued .
this is not the case for plane waves .
however , we can always obtain a real valued basis set by linear combining all pairs of plane waves with wave vectors @xmath101 and @xmath102 to real - valued basis functions .
this real valued basis set and the original complex - valued plane wave basis set are related by a unitary transformation that does not change any of the arguments of this paper .
all arguments therefore are also valid for the complex - valued plane wave basis sets considered here and below .
the number @xmath60 of basis functions roughly equals @xmath103 .
the exact value of @xmath60 depends on whether reciprocal lattice vectors @xmath101 that lie in the immediate vicinity of the surface of the sphere with radius @xmath99 have lengths that are slightly larger or slightly smaller than @xmath99 .
the relation @xmath104 shows that the products of plane waves of the orbital basis set are again plane waves of the same type with reciprocal lattice vectors @xmath105 that obey the relation @xmath106 . due to the latter relation
the number of different products @xmath107 is about 8 times as large as the number @xmath13 of orbital basis functions , i.e. , equals about @xmath108 .
if @xmath109 then @xmath110 . in this case
the number of _ different _ products of orbital basis functions is smaller than the number of products of orbital functions .
thus some products of orbital functions are equal and thus linearly dependent . for realistic systems
the number of plane wave basis functions is much larger than 15 . in a plane wave framework
therefore xoep and hf methods , in general , lead to different ground state energies with @xmath73 .
results from plane wave xoep and hf calculations for silicon discussed below illustrate this point .
in this section we show that the xoep approach of staroverov , scuseria , and davidson @xcite does not really correspond to an exact exchange ks method and does not yield a ks exchange potential , irrespective of whether or not the products of basis functions of the chosen orbital basis set are linearly independent . to this end
we consider the xoep or exact exchange ( exx ) equation written in a form that slightly differs from that of eq.([exxb ] ) @xmath111 the response function @xmath112 in eq .
( [ exx ] ) is given by @xmath113 eq .
( [ exx ] ) can be derived in completely different ways .
firstly , following refs . [ ] and [ ] , one can consider the expression of the hf total energy and search for those orbitals that minimize this energy under the constraint that the orbitals are eigenstates of a schrdinger equation with an hamiltonian operator of the form @xmath114 the search for these orbitals is tantamount to searching the optimal effective potential @xmath115 , therefore the name optimized effective potential method .
the optimized effective potential @xmath115 can always be expressed as @xmath116 with the hartree potential given as the coulomb potential of the electron density generated by the orbitals . as shown in refs .
[ ] and [ ] the optimized effective potential @xmath115 is obtained if the exchange potential potential @xmath117 of eq.([oep - potential ] ) obeys the xoep or exx equation ( [ exx ] ) .
alternatively the xoep or exx equation ( [ exx ] ) can be derived within an exact exchange - only ks framework . the hamiltonian operator @xmath118 of the exact exchange - only ks equation is given by eq .
( [ ks - hamiltonian ] ) with the effective ks potential @xmath119 the ks exchange potential in eq .
( [ ks - potential ] ) is defined as the functional derivative of the exchange energy @xmath120 with respect to the electron density @xmath121 , i.e , as @xmath122 following ref .
[ ] we now exploit that according to the hohenberg - kohn theorem there exists a one - to - one mapping between effective potentials @xmath123 and resulting electron densities @xmath121 .
therefore all quantities that are functionals of the electron density , here in particular the exchange energy , can be simultaneously considered as functionals of the effective potential @xmath71 .
taking the functional derivative @xmath124 of the exchange energy with respect to the effective potential @xmath71 in two different ways with the help of the chain rule yields @xmath125 the functional derivative @xmath126 equals the response function ( [ xs ] ) and the right hand side of eq.([chain - rule ] ) equals the right hand side of the xoep or exx equation ( [ exx ] ) .
furthermore the response function @xmath112 is symmetric in its arguments for real valued orbitals .
therefore eq .
( [ chain - rule ] ) is identical to the oep or exx equation ( [ exx ] ) .
this shows that the exchange potentials arising in the xoep and the exact exchange - only ks schemes and subsequently the xoep and the exact exchange - only ks schemes itself are identical . the xoep or exx equation can be derived in various ways within a ks framework @xcite . a crucial point , however , is that all derivations within a ks framework rely on real space representations in the sense that functional derivatives are taken within real space because the ks exchange potential is defined in real space as the functional derivative @xmath127 .
thus the above conclusion that the xoep and the exact exchange - only ks schemes are equivalent holds only in real space , i.e. , if all quantities are respresented in real space .
calculations , however , are usually carried out in basis sets and we will show next that in this case an xoep and an exact exchange - only ks scheme , in general , are not equivalent .
the xoep or exx equation ( [ exx ] ) turns into the matrix equation @xmath128 with matrix and vector elements elements @xmath129 @xmath130 and @xmath131 if an auxiliary basis set @xmath132 is introduced to represent the response function , the exchange potential , and the right hand side of the exx equation ( [ exx ] ) . for simplicity
we assume at this point that the auxiliary basis set is an orthonormal basis set .
this is actually the case for plane wave basis sets but not for gaussian basis sets .
however , without changing the following arguments we can assume that we have orthonormalized any auxiliary gaussian basis set .
as long as the orbitals are represented in real space there is an infinite number of them and the summations over unoccupied orbitals in the response function ( [ xs ] ) and the right hand side of the xoep or exx equation ( [ exx ] ) remains infinite and complete . for simplicity
we assume that the considered electron system is either periodic and thus exhibits periodic boundary conditions or , in case of a finite system , is enclosed in a large but finite box with an infinite external potential outside the box .
then the number of orbitals is infinite but countable .
as long as all orbitals are taken into account in the summation over unoccupied orbitals , the basis set representation of the exchange potential resulting from the basis set xoep or exx equation ( [ exx - basis ] ) becomes the more accurate the larger the auxiliary basis set and converges against the real space representation of the exchange potential and can be interpreted both as exact exchange - only ks or xoep exchange potential .
this changes dramatically if the orbitals are represented in a finite orbital basis set .
then , provided a reasonable orbital basis set is chosen , the occupied and the energetically low unoccupied orbitals are well represented .
most of the energetically higher unoccupied orbitals , however , are not represented at all simply because a finite orbital basis set can not give rise to an infinite number of unoccupied orbitals .
moreover , the energetically higher orbitals arising in a finite orbital basis set are quite poor representations of true unoccupied orbitals .
let us now concentrate on the representation of the response function .
the integrals @xmath133 occuring in the matrix elements ( [ xs - elements ] ) of the response function contain the three functions @xmath52 , @xmath53 , and @xmath134 .
the occupied orbitals @xmath52 have few nodes and thus are relatively smooth functions .
the energetically low lying unoccupied orbitals still are relatively smooth , the higher ones however , with an increasing number of nodes and with increasing kinetic energy become more and more rapidly oscillating . for smooth auxiliary basis functions @xmath134 the integrals latexmath:[$\langle \phi_a high unoccupied orbital @xmath53 because the product of the smooth functions @xmath134 and @xmath52 again is a smooth function and the integral of this smooth product with a rapidly oscillatory unoccupied orbital @xmath53 is zero due to the fact that any integral of a smooth with a rapidly oscillating function vanishes .
this means that for matrix elements @xmath136 of the response function with two sufficiently smooth functions @xmath134 and @xmath137 the summation over unoccupied orbitals in eq .
( [ xs - elements ] ) can be restricted to unoccupied orbitals @xmath53 below a certain energy depending on the smoothness of the involved auxiliary basis functions @xmath134 and @xmath138 . for sufficiently smooth functions @xmath134 and @xmath138 the contributing unoccupied orbitals @xmath53
thus are well represented in a finite orbital basis set
. therefore the matrix elements @xmath136 of the response functions are correct for indices @xmath40 and @xmath139 referring to sufficiently smooth auxiliary basis functions . for a more rapidly oscillating auxiliary basis function
nonvanishing matrix elements @xmath133 with energetically high unoccupied orbitals @xmath53 occur .
the energetically high unoccupied orbitals @xmath53 , however , are poorly described in the finite orbital basis set and moreover there are too few of them .
therefore the matrix elements @xmath136 of the response functions turn out to be wrong if at least one index refers to a more rapidly oscillating auxiliary basis functions . indeed ,
if an auxilliary function @xmath140 oscillates much more rapidly than the energetically highest unoccupied orbitals @xmath53 obtained for a given orbital basis set then all matrix elements @xmath141 and thus all corresponding elements @xmath136 of the response function is erroneously zero . for a given auxiliary basis set , according to the above argument
, a representation of the response function is correct only if the orbital basis set is balanced to the auxiliary basis set in the sense that it describes well unoccupied orbitals up to a sufficiently high energy .
otherwise an incorrect representation of the response function is obtained .
the matrix representation of the response function like the response function itself is negative semidefinite .
this is easily seen if a matrix element of the type @xmath142 for an arbitrary function @xmath143 is considered .
such a matrix element is obtained by summing up the contributions occuring in the summation over occupied and unoccupied orbitals in eq .
( [ xs ] ) . each single contribution and
thus also the complete sum is nonpositive .
therefore an insufficient orbital basis set leading to too few energetically high unoccupied orbitals results in eigenvalues of the response matrix that have a too small magnitude .
solutions of the matrix equation ( [ exx - basis ] ) are given by the product of the inverse of the response matrix with the right hand side of the equation , i.e. , by @xmath144 .
if @xmath145 contains eigenvalues that are too small then the corresponding eigenvectors contribute with a too large magnitude to the solution of equation ( [ exx - basis ] ) .
the eigenvectors with too small eigenvalues correspond to rapidly oscillatory functions .
therefore the resulting exchange potential exhibits rapidly oscillatory features .
this is exactly what is observed in the xoep scheme of staroverov , scuseria , and davidson @xcite . if the response matrix even contains eigenvectors with eigenvalues that are erroneously zero then an infinite number of solutions arise of the matrix equation ( [ exx - basis ] ) corresponding to an infinite number of exchange potentials , which yield , within the finite basis set , the same ks orbitals . therefore if the auxilliary and the orbital basis sets are chosen unbalanced , e.g. , if one chooses a too small orbital basis set for a given auxiliary basis set or a too large auxiliary basis set for a given orbital basis set , then the resulting response matrix @xmath145 is corrupted and no longer represents a proper representation of the response function in real space . in this case
the xoep scheme no longer represents an exact exchange ks scheme and the resulting exchange potential is unphysical and no longer represents the ks exchange potential .
however , even in this case the xoep scheme still is a proper optimized potential scheme in the sense that it yields a linear combination of auxiliary basis functions that results in the lowest total energy for this orbital basis set that can be obtained if the exchange potential shall be a linear combination of the auxiliary basis functions . while the resulting exchange potential is unphysical and does not resemble the ks exchange potential it obeys the above requirement of the xoep scheme .
the reason is that the arguments used for the xoep derivation of the real space exx or xoep equation can also be used if orbital and auxiliary basis sets are introduced whereas no analogue to the dft derivation exists anymore in this case .
we now illustrate the arguments of the previous two sections by specific examples .
these examples also demonstrate that an auxiliary basis set that consists of all products of occupied and unoccupied orbitals is not balanced to the corresponding orbital basis set in the sense that a correct representation of the response function and a proper ks exchange potential can not be obtained for such an auxiliary basis set .
firstly a system of electrons in a box with periodic boundary conditions and an external potential equal to a constant is considered .
the box shall be defined by corresponding unit cell vectors @xmath146 with @xmath147 .
if the box , i.e. , the unit cell vectors , become infinitely large then the system turns into an homogeneous electron gas .
the ks eigenstates @xmath148 of such a system are determined by symmetry and are simple plane waves @xmath149 as they are given in eq .
( [ plane - waves ] ) .
all plane waves with @xmath101 vectors of a length smaller than some given constant g@xmath150 , i.e. , with @xmath151 , shall represent occupied ks orbitals , all plane waves with @xmath152 represent unoccupied ks orbitals .
the maximal length @xmath153 of the vectors @xmath101 of the occupied orbitals determines the fermi level . for the orbital basis set as well as for the auxiliary basis
set we chose plane waves , @xmath149 and @xmath154 , respectively , again as given in eq.([plane - waves ] ) .
thus , for the considered system , arises the special case that each orbital basis function @xmath149 represents a ks orbital @xmath148 . obviously , the cutoff @xmath155 of the orbital basis set has to be chosen equal to or larger than @xmath156 .
the matrix representation @xmath145 of the response function in the considered case is diagonal with diagonal elements @xmath157 the auxiliary basis set shall be characterized by the cutoff radius @xmath158 , i.e. , the auxiliary basis set shall consist of all plane waves @xmath154 with @xmath159 .
note that the auxiliary function with @xmath160 that equals a constant function has to be excluded from the auxiliary basis set because the xoep or exx equation in agreement with the basic formalism determines the exchange potential only up to an additive constant .
a constant function would be an eigenfunction of the reponse function with zero eigenvalue .
now three cases can be distinguished : ( i ) if @xmath161 then the corresponding matrix elements @xmath162 of the reponse function are obtained with their correct value in a basis set calculation with an orbital basis set characterized by the cutoff radius @xmath155 because all unoccupied orbitals @xmath163 occuring in the summation in eq .
[ xsgg ] can be represented by the orbital basis set .
( ii ) if @xmath164 then for the matrix elements @xmath162 with @xmath165 incorrect values are obtained because some of the unoccupied orbitals @xmath163 occuring for these matrix elements in the sum in eq .
[ xsgg ] can not be represented in the orbital basis set and therefore are not taken into account . because all terms in the sum in eq .
[ xsgg ] have the same sign the magnitudes of the resulting matrix elements @xmath162 are too small .
( iii ) if @xmath166 then the resulting @xmath167 not only contains elements with a too small magnitude but additionally all matrix elements @xmath168 with @xmath169 are erroneously zero because all of the unoccupied orbitals @xmath163 occuring in the summation in eq .
[ xsgg ] can not be represented in the orbital basis set and therefore are not taken into account .
if the auxiliary basis set is chosen to be the space spanned by all products of occupied and unoccupied orbitals then it consists of all plane waves @xmath154 with @xmath170 , i.e. , @xmath171 .
thus the auxiliary basis set is chosen according to the above cases ( ii ) . therefore some of the resulting matrix elements @xmath168 of the reponse function are incorrect .
this demonstrates that an auxiliary basis set given by all products of occupied and unoccupied orbitals is not balanced with the corresponding orbital basis set .
the considered system is special in that the right hand side of the xoep or exx matrix equation is zero due to the translational symmetry .
therefore also the resulting exchange potential is zero or more precisely equals an arbitrary constant .
if the auxiliary basis set is chosen according to the above cases ( i ) and ( ii ) then a basis set calculation yields the correct exchange potential , i.e. , zero or a constant .
if the auxiliary basis set contains functions according to the above case ( iii ) , however , then the xoep or exx matrix equation erroneously has an infinite number of solutions that equal a constant plus an arbitrary contribution of auxiliary basis functions with @xmath169 .
the reason why the correct exchange potential is obtained for an auxiliary basis set chosen according to the above case ( ii ) despite the fact that in this case the response function is already corrupted is that for the special system considered here the right hand side of the xoep or exx matrix equation is zero .
therefore any values for the diagonal elements @xmath162 that differ from zero lead to the correct result .
however , in general the right hand side of xoep or exx matrix equation is not equal to zero and then a response matrix with eigenvalues with erroneously too small magnitudes leads to a wrong exchange potential that exhibits too large contributions from those linear combinations of auxiliary basis functions that correspond to the too small eigenvalues of the response matrix .
this is demonstrated in the following example .
we consider plane wave xoep calculations for bulk silicon carried out with the method of ref .
the integrable singularity occuring in hf and xoep exchange energies in plane wave treatments of solids is taken into account according to ref .
the lattice constant was set to the experimental value of 5.4307 .
the set of used * k*-points was chosen as a uniform @xmath172 mesh covering the first brillouin zone . in all calculations , all unoccupied orbitals resulting for a given orbital basis set were taken into account for the construction of the response function and the right hand side of the xoep equation .
exx pseudopotentials @xcite with angular momenta @xmath173 and cutoff radii , in atomic units , of @xmath174 , @xmath175 , and @xmath176 were employed .
the pseudopotential with @xmath177 was chosen as local pseudopotential .
[ fig.chicut5 ] and [ fig.chicut10 ] display xoep exchange potentials along the silcion - silicon bond axis , i.e. , the unit cell s diagonal , for auxiliary basis set cutoffs @xmath178 of 5.0 and 10.0 a.u .
( @xmath158 of 3.2 and 4.5 a.u . ) , and for various different orbital basis set cutoffs @xmath179 .
note that in figures and tables instead of the cutoffs @xmath155 and @xmath158 that refer to the length of the reciprocal lattice vectors of the plane waves the corresponding energy cutoffs @xmath180 and @xmath181 are displayed .
[ fig.chicut5 ] shows that the combination of an auxiliary basis set with cutoff @xmath182 ( @xmath183 ) with a orbital basis set with cutoff @xmath184 ( @xmath185 ) leads to a highly oscillating unphysical exchange potential .
the cutoff of the auxiliary basis set in the considered case is about twice as large as the cutoff of the orbital basis .
this means that the space spanned by the auxiliary basis is the same as that of all products of occupied and unoccupied orbitals . in this case
the matrix representing the response function is corrupted and the resulting exchange potential turns out to be unphysical .
with increasing cutoff @xmath179 of the orbital basis set the xoep exchange potentials converge towards the physical ks exchange potential , more precisely towards the representation of the physical ks exchange potential in an auxiliary basis set with cutoff @xmath182 ( @xmath183 ) .
if the cutoff @xmath179 of the orbital basis set is about 1.5 times as large as the cutoff of the auxiliary basis set @xmath178 , i.e. , equals 7.5 ( @xmath186 ) , then the exchange potential is converged .
a further increase of @xmath179 to @xmath187 ( @xmath188 ) leads to an exchange potential that is indistinguishable from that for @xmath189 ( @xmath186 ) on the scale of fig .
[ fig.chicut5 ] . fig .
[ fig.chicut10 ] gives an analogous picture for a cutoff of the auxiliary basis set of @xmath190 ( @xmath191 ) .
again , if the space spanned by the auxiliary basis set equals that of the product of occupied and unoccupied orbitals , curve for @xmath192 ( @xmath193 ) , an highly oscillating unphysical exchange potential is obtained . if @xmath194 then the exchange potential is converged towards the representation of the physical ks exchange potential in an auxiliary basis set with cutoff @xmath195 ( @xmath191 ) .
this demonstrates the point that the xoep scheme only represents a ks scheme if the orbital basis set is balanced to the auxiliary basis set . in the case of a plane wave basis
set this requires the energy cutoff @xmath179 of the orbital basis set to be about 1.5 times larger than the energy cutoff @xmath178 of the auxiliary basis set .
table [ tab.energies ] lists for a number of orbitals basis set cutoffs @xmath179 exchange and ground state energies for series of auxiliary basis set cutoffs @xmath178 . table [ tab.energies ] shows that the ground state energies for a given @xmath179 always decrease with increasing @xmath178 even if the values of @xmath178 is that large that the resulting exchange potential is unphysical .
this demonstrates that the xoep scheme remains well - defined even if unbalanced basis sets are used . in this case , however , the xoep scheme no longer represents a ks method and the resulting exchange potential is unphysical and does not represent the ks exchange potential .
table [ tab.energies ] also lists the differences of the xoep and hf ground state energies and shows that the xoep energy does not converge to the hf energy . in the combinations @xmath196 , @xmath197 , and
@xmath198 the space spanned by the auxiliary basis set roughly equals that of the product of occupied and unoccupied orbitals .
the ground state xoep energies in these cases is de facto the lowest that can be achieved by the xoep method for the given orbital basis set .
the fact that this energy is higher than the hf total energy shows that the xoep energy does not reach the hf ground state energy if the products of occupied and unoccupied orbitals become linearly dependent as it is usually the case in plane wave calculations and as it is the case in the presented calculations .
we have given arguments leading to the conclusion that exchange - only optimized potential ( xoep ) methods , with finite basis sets , can not in general yield the hartree - fock ( hf ) ground state energy , but a ground state energy that is higher .
this holds true even if the exchange potential that is optimized in xoep schemes is expanded in an arbitrarily large auxiliary basis set .
the hf ground state energy can only be obtained via an xoep scheme in the special case that all products of occupied and unoccupied orbitals emerging for the orbital basis set are linearly independent from each other . in this case , however , exchange potentials leading to the hf ground state energy exhibit unphysical oscillations and do not represent kohn - sham ( ks ) exchange potentials . these findings solve the seemingly paradoxical results of staroverov , scuseria and davidson @xcite that certain finite basis set xoep calculations lead to the hf ground state energy despite the fact that it was shown @xcite that within a real space representation ( complete basis set ) the xoep ground state energy is always higher than the hf energy .
a key point is that the orbital products of a complete basis are linearly dependent . moreover , whether or not the products of occupied and unoccupied orbitals are linearly independent , we have shown that basis set xoep methods only represent exchange - only ( exx ) ks methods , i.e. , proper density - functional methods , if the orbital basis set and the auxiliary basis set representing the exchange potential are balanced to each other , i.e. , if the orbital basis set is comprehensive enough for a given auxiliary basis set
otherwise xoep schemes do not represent exx ks methods .
we have found that auxiliary basis sets that consist of all products of occupied and unoccupied orbitals are not balanced to the corresponding orbital basis set .
the xoep method , even in cases of unbalanced orbital and auxiliary basis sets , works properly in the sense that it determines among all exchange potentials that can be represented by the auxiliary basis set the one that yields the lowest ground state energy .
however , in these cases the resulting exchange potential is unphysical and does not represent a ks exchange potential .
therefore the xoep method is of little practical use in those cases for which it does not represent a exx ks method .
remember that , at present , the main reason to carry out xoep methods in most cases is to obtain a qualitatively correct ks one - particle spectrum , either for the purposes of interpretation or as input for other approaches like time - dependent density - functional methods . however , the unphysical oscillations of the exchange - potential of xoep schemes with unbalanced basis sets affect the unoccupied orbitals and eigenvalues .
another reason to carry out xoep methods that represent exx ks methods is that the latter may be combined with new , possibly orbital - dependent , correlation functionals to arrive at a new generation of density - functional methods .
also in this case it is important that the xoep methods represents proper ks methods .
a balancing of auxiliary and orbital basis sets is straightforward for plane wave basis sets . in this case
xoep schemes are proper exx ks methods if the energy cutoff for the orbital basis set set is about 1.5 times as large as that of the auxiliary basis set .
this as well as other results of this work were illustrated with plane wave calculations for bulk silicon .
for gaussian basis sets on the other hand , a proper generally applicable and reasonably simple balancing scheme of orbital and auxiliary basis sets is so far not available despite much efforts @xcite . therefore effective exact exchange - only methods like the kli @xcite , the localized hartree - fock @xcite , the equivalent common energy denominator approximation method @xcite , or the closely related very recent method of ref .
, are in use as numerically stable alternatives that yield results very cose to those of full exx ks methods . .
xoep exchange and ground state energies @xmath199 and @xmath200 , respectively , and the difference between hf and xoep ground state energy for various combinations of and orbital auxiliary basis sets , characterized by energy cutoffs @xmath179 and @xmath178 , respectively .
( @xmath13 and @xmath201 denote the corresponding number of basis functions . )
all quantities are given in a.u .
[ tab : energies ] [ cols="<,>,>,^,^,^",options="header " , ] [ tab.energies ] and different orbital basis set cutoffs @xmath179 .
the upper and lower panels differ in the energy scale . the curve for @xmath202 is only displayed in the upper panel.,width=415 ] and different orbital basis set cutoffs @xmath179 .
the upper and lower panels differ in the energy scale .
the curve for @xmath203 is only displayed in the upper panel.,width=415 ] 40 v. k. staroverov , g. e. scuseria , and e. r. davidson , j. chem .
phys . * 124 * , 141103 ( 2006 ) .
v. sahni , j. gruenbaum , and j. p. perdew , phys .
b * 26 * , 4371 ( 1982 ) .
s. ivanov and m. levy , j. chem . phys . * 119 * , 7087 ( 2003 ) .
m. stdele , j. a. majewski , p. vogl , and a. grling , phys .
79 * , 2089 ( 1997 ) .
m. stdele , m. moukara , j. a. majewski , p. vogl , and a. grling , phys .
b * 59 * , 10031 ( 1999 ) .
r. j. magyar , a. fleszar , and e.k.u .
gross , phys .
b * 69 * , 045111 ( 2004 ) .
a. qteish , a.i .
al - sharif , m. fuchs , m. scheffler , s. boeck , and j. neugebauer , comp .
* 169 * , 28 ( 2005 ) .
p. rinke , a. qteish , j. neugebauer , c. freysoldt , and m. scheffler , new j. phys .
* 7 * , 126 ( 2005 ) .
a. grling and m. ernzerhof , phys .
a * 51 * , 4501 ( 1995 ) .
r. t. sharp and g. k. horton , phys .
rev . * 90 * , 317 ( 1953 ) .
j. d. talman and w. f. shadwick , phys .
a * 14 * , 36 ( 1976 ) .
w. kohn and l. j. sham , phys . rev . *
140 * , a1133 ( 1965 ) .
m. levy , proc .
usa * 76 * , 6062 ( 1979 ) .
j. e. harriman , phys .
a * 27 * , 632 ( 1983 ) ; * 34 * , 29 ( 1986 ) . m. levy and j. a. goldstein , phys .
b * 35 * , 7887 ( 1987 ) . see the constrained search discussion in ref [ ] , which includes appropriate references on the approach in this context .
h. ou - yang and m. levy , phys .
lett . * 65 * , 1036 ( 1990 ) .
m. levy and j. p. perdew , phys .
* a3 * 2 , 2010 ( 1985 ) .
m. levy and a. grling , phys .
a * 53 * , 3140 ( 1996 ) .
a. grling and m. levy , int .
j. quantum chem .
* 29 * , 93 ( 1995 ) .
a. grling , j. chem . phys . * 123 * , 062203 ( 2005 ) ; and references therein . p. carrier , s. rohra , and a. grling , submitted .
e. engel , a. hck , and s. varga , phys .
b * 63 * , 125121 ( 2001 ) .
e. engel , a. hck , r. n. schmid , and r. m. dreizler phys .
b * 64 * , 125111 ( 2001 ) .
a. grling , phys .
lett . * 83 * , 5459 ( 1999 ) .
s. ivanov , s. hirata , and r. j. bartlett , phys .
lett . * 83 * , 5455 ( 1999 ) .
s. hamel , m. e. casida , and d. r. salahub , j. chem .
phys . * 114 * , 7342 ( 2001 ) . l. veseth , j. chem . phys .
* 114 * , 8789 ( 2001 ) .
s. hirata , s. ivanov , i. grabowski , r. bartlett , k. burke , and j. d. talman , j. chem . phys . * 115 * , 1635 ( 2001 ) .
w. yang and q. wu , phys .
* 89 * , 143002 ( 2002 ) .
j. b. krieger , y. li , and g. j. iafrate , phys .
a * 46 * , 5453 ( 1992 ) .
f. della sala and a. grling , j. chem . phys . * 115 * , 5718 ( 2001 ) .
o. v. gritsenko and e. j. baerends , phys .
a * 64 * , 042506 ( 2001 ) .
v. k. staroverov , g. e. scuseria , and e. r. davidson , j. chem .
phys . * 125 * , 081104 ( 2006 ) .
let @xmath204 be a complete set of functions of a complex valued variable @xmath205 such that any arbitrary square integrable function can be written as a linear combination of the functions in the complete set .
we show that the set @xmath206 is linearly dependent . using our complete sets , an arbitrary function @xmath207 of two complex valued variables @xmath205 and @xmath208
may be expanded in terms of @xmath204 and @xmath209 @xmath210 set @xmath211 to get : @xmath212 now choose a function @xmath213 and a @xmath214 out of the set @xmath204 such that ( i ) @xmath215 and ( ii ) at least one @xmath216 when @xmath217 and @xmath218 . since @xmath219 is just a function of @xmath205 , we may expand it in term of the @xmath204 : @xmath220 solving for @xmath213 , @xmath221 and equating eq .
( [ eq:1 ] ) with eq .
( [ eq:2 ] ) , we get @xmath222 or by setting @xmath223 , @xmath224 or @xmath225 eq . ( [ eq:4 ] ) is a linear combination of a subset of @xmath206 broken up into disjoint components and equated to zero .
if a subset of a set is linearly dependent , then the set must also be linearly dependent .
we show such a case by contradiction : according to eq .
( [ eq:4 ] ) , for the subset @xmath226 ( appearing in the equation ) to be linearly independent , three conditions must be met : but according to our condition on @xmath213 there is at least one @xmath235 with @xmath234 and @xmath217 , which is a contradiction to number three of our linear independence criteria
. therefore @xmath226 must be linearly dependent by contradiction , and therefore @xmath206 for all @xmath236 is linearly dependent because @xmath226 is linearly dependent . | arguments showing that exchange - only optimized effective potential ( xoep ) methods , with finite basis sets , can not in general yield the hartree - fock ( hf ) ground state energy , but a higher one , are given . while the orbital products of a complete basis are linearly dependent
, the hf ground state energy can only be obtained via a basis set xoep scheme in the special case that all products of occupied and unoccupied orbitals emerging from the employed orbital basis set are linearly independent from each other . in this case , however , exchange potentials leading to the hf ground state energy exhibit unphysical oscillations and do not represent a kohn - sham ( ks ) exchange potential .
these findings solve the seemingly paradoxical results of staroverov , scuseria and davidson that certain finite basis set xoep calculations lead to the hf ground state energy despite the fact that within a real space ( or complete basis ) representation the xoep ground state energy is always higher than the hf energy .
moreover , whether or not the occupied and unoccupied orbital products are linearly independent , it is shown that basis set xoep methods only represent exact exchange - only ( exx ) ks methods , i.e. , proper density - functional methods , if the orbital basis set and the auxiliary basis set representing the exchange potential are balanced to each other , i.e. , if the orbital basis is comprehensive enough for a given auxiliary basis .
otherwise xoep methods do not represent exx ks methods and yield unphysical exchange potentials .
the question whether a xoep method properly represents a ks method with an exchange potential that is a functional derivative of the exchange energy is related to the problem of the definition of local multiplicative operators in finite basis representations and to the fact that the hohenberg kohn theorem does not apply in finite basis representations .
plane wave calculations for bulk silicon illustrate the findings of this work . |
let @xmath1 be a graph with @xmath2 and @xmath3 being respectively the sets of all vertices and edges of @xmath4 . throughout of the present paper
, we always assume that graphs satisfy the standard conditions , that is , they are finite , undirected , simple , connected and regular .
assume that @xmath4 is a @xmath5-regular graph .
it is well known that , as an analogue of the prime number theorem , we have so - called the prime geodesic theorem for @xmath4 .
we here briefly review it ( more precisely , see e.g. , @xcite ) . for
a positive integer @xmath6 , let @xmath7 be the number of all reduced cycles @xmath0 in @xmath4 with @xmath8 where @xmath9 is the length of @xmath0 and @xmath10 the number of all equivalence classes @xmath11 $ ] of the prime reduced cycles @xmath12 in @xmath4 with @xmath13 . here
, the cycle @xmath0 is called reduced if @xmath14 has no backtrack and is prime if it can not be expressed as @xmath15 for any cycle @xmath16 and @xmath17 .
it is easy to see that @xmath18 and hence @xmath19 by the mbius inversion formula .
let @xmath20 be the edge adjacency matrix of @xmath4 , which is a square matrix of size @xmath21 , and @xmath22 the set of all eigenvalues of @xmath20 with multiplicities .
then , it holds that @xmath23 and hence @xmath24 if @xmath25 ( and @xmath26 otherwise ) . here , @xmath27 is the greatest common divisor of all lengths of prime reduced cycles in @xmath4 .
actually , we can obtain from by using the result obtained by kotani and sunada @xcite , which asserts that the eigenvalues of @xmath20 with the largest absolute value are given by @xmath28 for @xmath29 . as the prime number theorem is obtained via the riemann zeta function , these formulas are also related to the zeta function @xmath30 ,
called the ihara zeta function @xcite , of @xmath4 defined by the following euler product ; @xmath31}\bigl(1-u^{l(p)}\bigr)^{-1 } \quad ( |u|<q^{-1}).\ ] ] here , in the product , @xmath11 $ ] runs over all equivalence classes of the prime reduced cycles in @xmath4 .
in fact , since we have from the definition @xmath32 one obtains by the following determinant expression of @xmath30 with respect to @xmath20 ; @xmath33 here , for a positive integer @xmath34 , @xmath35 is the identity matrix of size @xmath34 .
remark that we will encounter another type of the determinant expression of @xmath30 ( see [ subsec : ihara_zeta ] ) .
the aim of this paper is to establish an explicit prime geodesic theorem , which are _ not _ an asymptotic formula , for the discrete tori . here , for @xmath36 , the discrete torus @xmath37 of dimension @xmath38
is defined by the cayley graph of the group @xmath39 associated with the generating set @xmath40 with @xmath41 .
this is a @xmath42-regular graph having @xmath43 vertices and @xmath44 edges .
because of the simplicity of the structure of the graph , their harmonic analysis are well studied .
in particular , very recently , there are various results on the _ complexities _ of the discrete tori or their degenerated ones by establishing the theory of the heat kernel on the graphs ( @xcite ) .
to state the result , we need a generalization of the jacobi polynomial : for @xmath45 , @xmath46 and @xmath47 , define @xmath48 \alpha_1 + 1,\ldots,\alpha_d+1 \end{array } ; \,\frac{1-x}{2},\ldots,\frac{1-x}{2 } \right),\ ] ] where @xmath49 is the pochhammer symbol with @xmath50 being the gamma function , @xmath51 and , for @xmath52 , @xmath53 is the lauricella multivariable hypergeometric function of type @xmath0 . remark that @xmath54 is equal to the gauss hypergeometric function @xmath55 and hence @xmath56 coincides with the classical jacobi polynomial .
when @xmath57 , @xmath58 equals the appell hypergeometric function @xmath59 .
it is worth commenting that , if @xmath60 , then the above generalized jacobi polynomial @xmath61 can be expressed by the generalized hypergeometric function @xmath62 ( more precisely , see in example [ ex : d=2 ] ) .
however , for general @xmath63 , we can not confirm such degeneracies .
write @xmath64 for short .
the following is our main result .
[ thm : main ] for @xmath65 , it holds that @xmath66 where , for @xmath67 , @xmath68 and , for @xmath69 , @xmath70 here , @xmath71 is the kronecker delta and @xmath72 with @xmath73 denotes the multinomial coefficient .
we remark that though we can not calculate the right hand side of explicitly ( not numerically ) in general , we can definitely the one of because it is a finite sum of polynomials whose coefficients are given concretely .
we also remark that above result is easily extended to general discrete tori corresponding to the groups @xmath74 for @xmath75 . in this paper , for simplicity , we only consider such diagonal cases .
we give a proof of theorem [ thm : main ] in section [ sec : spectral_zeta ] .
it is achieved by calculating the spectral zeta function of @xmath76 in two ways : one is done by using the ihara zeta function and the other is by the theory of the heat kernel obtained in @xcite . in section [ sec : normalized_dt ] , we investigate a special case , that is , the normalized discrete torus @xmath77 .
we give some numerical examples and observations obtained from the examples .
we see that , in this case , our formula seems to give a graph analogue of a refinement of the prime geodesic theorem for compact riemannian manifolds of negative curvature , which counts closed geodesics lying in a fixed homology class .
let @xmath1 be a @xmath5-regular graph . define a spectral zeta function @xmath78 of @xmath4 by the following finite sum .
@xmath79 here , @xmath80 is the combinatoric laplacian on @xmath4 , that is , @xmath81 with @xmath82 being the ( vertex ) adjacency matrix of @xmath4 .
we sometimes understand that @xmath80 is a linear operator on the @xmath83-vector space @xmath84 , endowed with the inner product @xmath85 , @xmath86 , acting by @xmath87 here we denote @xmath88 by the edge which connects vertices @xmath89 and @xmath90 . the idea for obtaining the main result is to calculate the spectral zeta function @xmath91 of @xmath76 in two ways ; one is via the ihara zeta function and the other is via the heat kernel on @xmath76 .
first , it is useful to give here the explicit descriptions of the eigenvalues and the corresponding eigenfunctions of the combinatoric laplacian @xmath92 of @xmath76 .
let @xmath93 notice that @xmath94 , @xmath95 are the sets of all vertices of @xmath76 and its dual @xmath96 ( see @xcite ) , respectively . moreover , for @xmath97 and @xmath98 , put @xmath99 .
[ lem : specdata ] for @xmath100 , let @xmath101 then , we have @xmath102 and see that @xmath103 is an orthonormal eigenfunction corresponding to @xmath104 , that is , @xmath105 forms an orthonormal basis of @xmath106 .
see @xcite .
let @xmath1 be a @xmath5-regular graph .
it is easy to see that @xmath107 we now show that , using the determinant expression @xmath108 of the ihara zeta function @xmath109 with respect to @xmath82 , the spectral zeta function @xmath110 can be written by the logarithmic derivative of @xmath109 .
let @xmath1 be a @xmath5-regular graph .
then , it holds that @xmath111 where @xmath112 substituting @xmath113 into with the relation @xmath114 , we have @xmath115 now the desired formula is derived from and with making a change of variable @xmath116 that is , , together with the relation @xmath117 . from this lemma , because @xmath37 is a @xmath42-regular graph ( i.e. , @xmath118 ) , one immediately obtains the following proposition .
we have @xmath119 where @xmath120 we notice from that @xmath121
we will encounter this quadratic transformation in [ subsec : hg ] . because @xmath122 , what we have to do next is to expand @xmath123 in a series in the variable @xmath124 .
we next start from the fact that the spectral zeta function @xmath110 can be expressed as the laplace transform of the theta function @xmath125 of @xmath4 defined by @xmath126 actually , noticing that @xmath127 $ ] , we have @xmath128 from a general theory , we know that @xmath125 is essentially given by the heat kernel @xmath129 on the graph @xmath4 , which is the unique solution of the heat equation @xmath130 \displaystyle{\lim_{t\to 0}}f(x , t)=\delta_{o}(x ) . \end{array } \right.\ ] ] here , @xmath131 is a fixed base point of @xmath4 and @xmath132 is the kronecker delta , that is , @xmath133 if @xmath134 and @xmath135 otherwise . the following is well known .
let @xmath1 be a @xmath5-regular graph with and @xmath131 .
let @xmath136 be an orthonormal eigenfunction of @xmath80 with respect to @xmath137 .
then , we have @xmath138 see , e.g. , @xcite .
now , let us calculate @xmath139 and @xmath140 .
we notice that , in the case of studying @xmath37 , we always take a base point @xmath141 . from lemma [ lem : specdata ] and
, we have @xmath142 and hence @xmath143 it is shown in @xcite ( see also @xcite ) that the heat kernel @xmath144 on @xmath145 with @xmath146 is given by @xmath147 where @xmath148 is the @xmath149-bessel function ( or the modified bessel function of the first kind ) having the expansion @xmath150 hence , by the uniqueness of the heat kernel , we see that the heat kernel @xmath151 on @xmath152 can be written as @xmath153 for @xmath154 . moreover , periodizing this as a function on @xmath155 , we have @xmath156 this shows from that @xmath157 here , one obtains the last equation above as follows : let @xmath158 with @xmath159 being the symmetric group of degree @xmath38 .
we see that @xmath20 naturally acts on @xmath152 and can take the set of all partitions of length less than or equal to @xmath38 as a set of all representatives of the @xmath20-orbits on @xmath152 . for a partition @xmath160 of the length @xmath161 , we have @xmath162 where @xmath163 is the @xmath20-orbit of @xmath160 .
then , for @xmath67 , we have @xmath164 and , because @xmath165 is @xmath20-invariant ( notice that @xmath166 ) , obtain the desired equation .
now , from , it holds that @xmath167 where , for @xmath168 , we put @xmath169 here , we have used the relation . we will study some properties of these functions in
[ subsec : hg ] .
let us give a proof of the main result .
[ proof of theorem [ thm : main ] ] from , we have @xmath170 therefore , based on the equation , one can obtain a formula for @xmath171 by expanding the right hand side of in a series in the variable @xmath124 .
the first term on the right hand side of can be easily expanded as follows : @xmath172 moreover , since we have from together with , which we will prove in [ subsec : hg ] , @xmath173 with @xmath174 , the second term of can be written as @xmath175 here , in the first equality , we have used the equation @xmath176 and , for @xmath177 , @xmath178 which are easily seen from the definition .
let us write the coefficient of @xmath179 of the rightmost hand side of the above formula as , say , @xmath180 .
noticing that @xmath181 and @xmath182 because @xmath183 for all @xmath184 , we have @xmath185 this shows that the second term on the right hand side of is expanded as follows : @xmath186 combining and , and noticing that @xmath187 , we obtain the desired formula .
this ends the proof .
we here prove the following proposition , which contains a key formula for our results .
let @xmath189 .
@xmath190 we have @xmath191 z_1 + 1,\ldots , z_d+1 \end{array } ; \,\frac{4}{x^2},\ldots,\frac{4}{x^2 } \right)\\ \label{for : f_d2 } & = \frac{1}{x^{|z|+1}}\sum^{\infty}_{n=0}a^{(d)}_z(n)\frac{(\frac{|z|}{2}+\frac{1}{2})_n(\frac{|z|}{2}+1)_n}{(h+1)_n}\frac{(\frac{4}{x^2})^n}{n!},\end{aligned}\ ] ] where @xmath174 and @xmath192 @xmath193 we have @xmath194 where @xmath195 is the generalization of the jacobi polynomial defined by . moreover ,
if @xmath196 , then it can be written as @xmath197 put @xmath198 .
using , we have @xmath199 here , we have employed the identities @xmath200 and @xmath201 for @xmath202 .
hence we obtain .
the formula is easily obtained from .
we next concentrate on @xmath203 . from
, it holds that @xmath204 the generalized binomial theorem yields @xmath205 and hence @xmath206 therefore we obtain .
the equation follows in the same manner as .
[ ex : d=1 ] when @xmath207 , we have respectively from and @xmath208 h+1 \end{array } ; \,\frac{4}{x^2 } \right ) , \nonumber\\ \label{for : g1 } g^{(1)}_h(u ) & = u^{h+1}\sum^{\infty}_{n=0}p^{(h,0)}_{1,k}(-1)(-u^2)^{k}=\frac{u^{h+1}}{1+u^2}.\end{aligned}\ ] ] here , in the last equality in , we have used the well - known formula @xmath209 we remark that is also obtained from the pfaff transformation @xmath210 a - b+1 \end{array } ; \,\frac{4x}{(1+x)^2 } \right ) = ( 1+x)^a { } _ 2f_{1}\left ( \begin{array}{c } a , b\\[3pt ] a - b+1 \end{array } ; \,x \right)\ ] ] for the gauss hypergeometric functions @xmath211 . from this
, we can say that is a kind of generalization of the pfaff transformation for @xmath188 with special parameters .
[ ex : d=2 ] we next consider the case @xmath57 . let @xmath212 .
it is easy to see that @xmath213 hence , we have respectively from and @xmath214 |z|+1,z_1 + 1,z_2 + 1 \end{array } ; \,\frac{16}{x^2 } \right),\\ g^{(2)}_z(u ) & = c_zu^{|z|+1}\sum^{\infty}_{k=0 } p^{(z,0)}_{2,k}\bigl(\frac{1}{3}\bigr)(-3u^2)^{k}.\end{aligned}\ ] ] here , @xmath215 is the generalized hypergeometric function defined by @xmath216 moreover , from and , we have for @xmath217 @xmath218 |\alpha|+1,\alpha_1 + 1,\alpha_2 + 1 \end{array } ; \,2(1-x ) \right).\ ] ]
in this section , we consider a special case , that is , a normalized discrete torus @xmath220 for @xmath221 . in this case , the result obtained in the previous section becomes more simple form as below .
here , we put @xmath222 .
[ thm : main2 ] for @xmath65 , it holds that @xmath223 where , for a partition @xmath224 of length @xmath225 with @xmath226 being the multiplicity of @xmath227 in @xmath228 , @xmath229 consider the case @xmath207 .
let us check the trivial result @xmath230 from our formula .
first , we have @xmath231 if @xmath232 ( the empty partition ) and @xmath233 otherwise .
moreover , from , it holds that @xmath234 hence , from , one can actually obtain the desired formula .
we next consider the case @xmath57 .
it holds that @xmath235 here , @xmath236 if @xmath6 is even and @xmath135 otherwise .
notice that , for @xmath237 with @xmath238 , @xmath239 if @xmath240 and @xmath233 otherwise . using
, we have @xmath241 1,1 \end{array } ; \,\frac{4}{3 } \right).\ ] ] moreover , for @xmath242 and @xmath237 with @xmath238 , letting @xmath243 , we have @xmath244 mh+1,mh+1,1 \end{array } ; \,\frac{4}{3 } \right),\\ x^{(2)}_{m , h}\bigl(n;(\mu_1,\mu_2)\bigr ) & = \frac{4n(-3)^{k}}{n+mh}\binom{mh}{m\mu_1,m\mu_2}\frac{(mh+1)_k}{k ! } { } _ 4f_{3}\left ( \begin{array}{c } -k , k+mh,\frac{mh}{2}+\frac{1}{2},\frac{mh}{2}+1\\[3pt ] m\mu_1 + 1,m\mu_2 + 1,mh+1 \end{array } ; \,\frac{4}{3 } \right).\end{aligned}\ ] ] for example , let us consider the case @xmath245 and @xmath246 . since @xmath247 with @xmath248 1,1 \end{array } ; \,\frac{4}{3 } \right ) = 2 + 2(-27)\bigl(-\frac{11}{27}\bigr ) = 24 , \\
x^{(2)}_{3,2}\bigl(6;(2)\bigr ) & = { } _ 4f_{3}\left ( \begin{array}{c } 0,6,\frac{7}{2},4\\[3pt ] 7,7,1 \end{array } ; \,\frac{4}{3 } \right ) = 1,\\ x^{(2)}_{3,2}\bigl(6;(1,1)\bigr ) & = \binom{6}{3 } { } _ 4f_{3}\left ( \begin{array}{c } 0,6,\frac{7}{2},4\\[3pt ] 4,4,7 \end{array } ; \,\frac{4}{3 } \right ) = 20,\end{aligned}\ ] ] we have @xmath249 for the other values of @xmath250 with @xmath251 , see the table below .
.the values of @xmath252 [ cols="^,^,^,^,^,^",options="header " , ] from the table above , we first expect the following .
[ conj : xinteger ] it holds that @xmath253 .
if conjecture [ conj : xinteger ] is true , then it seems that @xmath254 counts something special type of cycles in @xmath255 . to see this ,
let us subdivide @xmath256 into small pieces by the following manner .
fix @xmath257 .
we notice that there is an one - to - one correspondence between a cycle @xmath0 in @xmath255 starting from and ending to @xmath258 of length @xmath6 and a path @xmath259 in @xmath152 starting from the origin @xmath260 and ending to @xmath261 for some @xmath262 of length @xmath6 .
let us call a path @xmath259 in @xmath152 _ reduced modulo @xmath34 _ if the corresponding cycle @xmath0 in @xmath255 is reduced .
let @xmath263 be the set of all reduced cycles in @xmath255 starting from and ending to @xmath258 of length @xmath6 and , for @xmath264 , @xmath265 the set of all reduced paths modulo @xmath34 in @xmath152 starting from the origin @xmath260 and ending to @xmath261 of length @xmath6 .
it is clear that @xmath266 where , for @xmath267 of length @xmath225 , @xmath268 is defined in and @xmath269 is the number of all reduced paths modulo @xmath34 in @xmath152 starting from the origin @xmath260 and ending to @xmath270 of length @xmath6 .
notice that @xmath271 represents the number of choices of the starting points .
now , it is natural from to expect the following .
[ conj : xn ] it holds that @xmath272 .
it is clear that conjecture [ conj : xinteger ] follows from conjecture [ conj : xn ] because @xmath273 .
the following figures support that conjecture [ conj : xn ] is true for the case @xmath57 , @xmath245 and @xmath246 , that is , @xmath274 , @xmath275 and @xmath276 . here , the black dots in the figures represent the lattices points . in particular , the big black dots denote points of the form of @xmath277 for some @xmath278 . by the same manner , we have already checked that the equation @xmath272 holds for @xmath251 . ; it is the number of all reduced paths @xmath259 modulo @xmath279 in @xmath280 starting from @xmath281 and ending to @xmath282 of length @xmath283.,width=529 ] ; it is the number of all reduced paths @xmath259 modulo @xmath279 in @xmath280 starting from @xmath281 and ending to @xmath282 of length @xmath283.,width=188 ] ; it is the number of all reduced paths @xmath259 modulo @xmath279 in @xmath280 starting from @xmath281 and ending to @xmath282 of length @xmath283.,width=566 ] let @xmath284 be a compact riemannian manifold of negative curvature .
it is known that there exist countably infinitely many closed geodesics in @xmath284 and that a closed geodesic in @xmath284 corresponds to a unique non - trivial conjugacy class @xmath285 of @xmath286 where @xmath287 is the fundamental group of @xmath284 .
let us write the closed geodesic corresponding to @xmath285 as @xmath288 .
let @xmath289 be the number of all closed geodesics in @xmath284 of length @xmath290 .
then , we have the following prime geodesic theorem for @xmath284 ( @xcite ) ; @xmath291 where @xmath292 is the topological entropy of the geodesic flow over @xmath284 .
this is an analogue of the classical prime number theorem .
moreover , we have also an analogue of the dirichlet theorem on arithmetic progressions , which counts the closed geodesics lying in a fixed homology class : let @xmath293 be the first homology group of @xmath284 and @xmath294 $ ] be the natural projection . here , @xmath295 $ ] is the commutant subgroup of @xmath287 . for
a fixed @xmath296 , let @xmath297 be the number of all closed geodesics @xmath298 in @xmath284 of length @xmath290 satisfying @xmath299 .
then , it is shown in @xcite that there exists a constant @xmath300 , not depending on @xmath301 , such that @xmath302 here , @xmath303 is the first betti number of @xmath284 , that is , the rank of @xmath293 .
now , we may regard the claim in conjecture [ conj : xn ] as a graph analogue of . note that @xmath304 where @xmath305 is the real torus of dimension @xmath38 .
the author would like to thank professor hiroyuki ochiai for carefully reading the manuscript and giving many helpful comments .
999999 t. adachi and t. sunada , homology of closed geodesics in a negatively curved manifold , _ j. diff .
_ , * 26 * ( 1987 ) , 8199 .
g. chinta , j. jorgenson and a. karlsson , zeta functions , heat kernels , and spectral asymptotics on degenerating families of discrete tori , _ nagoya math .
* 198 * ( 2010 ) , 121172 .
g. chinta , j. jorgenson and a. karlsson , complexity and heights of tori , dynamical systems and group actions , 8998 , _ contemp .
_ , * 567 * , amer .
soc . , providence , ri , 2012 .
g. chinta , j. jorgenson and a. karlsson , heat kernels on regular graphs and generalized ihara zeta function formulas , _ monatsh .
_ , * 178 * ( 2015 ) , 171190 . f. chung , spectral graph theory .
cbms regional conference series in mathematics , 92 . published for the conference board of the mathematical sciences , washington , dc ; by the american mathematical society , providence , ri , 1997 .
y. ihara , on discrete subgroups of the two by two projective linear group over @xmath306-adic fields , _ j. math .
japan _ , * 18 * ( 1966 ) , 219235 . a. karlsson , applications of heat kernels on abelian groups : @xmath307 , quadratic reciprocity , bessel integrals .
number theory , analysis and geometry . in memory of serge lang , springer verlag , 307320 2012 .
a. karlsson and m. neuhauser , heat kernels , theta identities , and zeta functions on cyclic groups , topological and asymptotic aspects of group theory , 177189 , contemp .
, 394 , amer .
soc . , providence , ri , 2006 .
m. kotani and t. sunada , zeta functions of finite graphs , _ j. math . sci
, * 7 * ( 2000 ) , 725 .
s. p. lalley , closed geodesics in homology classes on surfaces of variable negative curvature , _ duke math . j. _ , * 58 * ( 1989 ) , 795821 .
j. louis , asymptotics for the number of spanning trees in circulant graphs and degenerating @xmath38-dimensional discrete tori , _ ann .
_ , * 19 * ( 2015 ) , 513543 .
j. louis , a formula for the number of spanning trees in circulant graphs with nonfixed generators and discrete tori , _ bull .
_ , * 92 * ( 2015 ) , 365373 .
g. margulis , applications of ergodic theory to the investigation of manifolds of negative curvatures , _ funkt .
i ego pril .
_ , * 3 * ( 1969 ) , 8990 .
r. phillips and p. sarnak , geodesics in homology classes , _ duke math . j. _ , * 55 * ( 1987 ) , 287297 .
a. terras , zeta functions of graphs .
a stroll through the garden , cambridge studies in advanced mathematics , 128 .
cambridge university press , cambridge , 2011 . | the discrete tori are graph analogues of the real tori , which are defined by the cayley graphs of a finite product of finite cyclic groups . in this paper , using the theory of the heat kernel on the discrete tori established by chinta , jorgenson and karlsson , we derive an explicit prime geodesic theorem for the discrete tori , which is not an asymptotic formula . to describe the formula , we need generalizations of the classical jacobi polynomials , which are defined by the lauricella multivariable hypergeometric function of type @xmath0 . _ primary _ 11m36 _ secondary _ 05c30 , 33c65 discrete tori , prime geodesic theorem , heat kernels on graphs , lauricella hypergeometric functions , jacobi polynomials . |
thanks to the reasonably large amount of data we are obtaining with the swift satellite [ 4 ] we are slowly improving our understanding of the mechanism generating the afterglow .
bright bursts as in the case of grb080319b [ 5 ] shed light on the emission mechanisms of the likely multi - component jet and on its time evolution and , however , basic uncertainties remain regarding the mechanism of emission and the characteristics of the central object . on the other hand
the nature itself of the phenomenon , high variability , tell us that the clues of the event are indeed related to the observed variability so that understanding it on various temporal scale may guide toward the comprehension of the phenomenon and to the construction of a realistic model . to this end
we decided to approach the problem in a systematic way using different time scales .
in addition to the temporal analysis of the prompt emission on the time scale of the spikes width and the temporal decay of the afterglow , pasotti et al .
( in preparation ) , we started the systematic analysis of the flares on temporal scale of the flare width [ larger than tens of seconds ] , chincarini et al .
( in preparation ) , and of the analysis of the prompt emission , afterglow and flares of bright grbs on temporal scales from less than 0.1 s to tens of seconds , margutti et al . ,
( in preparation ) .
these analyses are based on the data reduction of the bat ( guidorzi et al . , ( in preparation ) ) and xrt data carried out by the italian swift group , pasotti et al . ,
( in preparation ) . in brief
while the bat data have been reduced using the standard time resolution and binning , the xrt data have been both reduced using a standard @xmath3 along the light curve and a binning that optimizes the number of photons per bin for each grb .
spectra , on the other hand , are based on a minimum of 2000 photons have been measured as a function of time and the light curve fluxes have been derived from counts using a conversion factor that changes with time according to the measured spectra .
the spectra have been measured both using a constant mean @xmath4 and leaving @xmath4 as a free parameter . at the time of writing the archives
were up dated to april 2008 and the present sample covers the period april 2005-april 2008 .
both in the flare sample and in the flare definition we have a certain degree of subjectivity . the first decision of the analyst
is that of fitting the underlying light curve .
it is assumed , see [ 1 ] and [ 2 ] and references therein , that the flares are superimposed on an otherwise rather regular afterglow light curve . in most cases the eye is unambiguously guided in deciding the proper functional shape while in others it remains , due to lack of contiguous observations , open to slightly different interpretation .
this leads to uncertainties .
the @xmath5 value not always allows to distinguishing between different models , however the parameters of the flares generally remain , even using different profiles , within the uncertainties of the analysis . whether a rather faint flare is or not present is again a somewhat subjective decision that depends on the observer
this can be checked statistically afterwards but in this sample we limited the selection to flares that were reasonably easily detected by the eye ; the mini - flare sample is not yet at hand and in any case to the mini flare statistics a high resolution temporal analysis [ 6 ] may be preferable .
the sample is rather sharply defined by the distribution of flare fluence , figure 1 , and flare to underlying light curve continuum ratio ( in preparation ) .
finally two related questions may create a sample bias as far as early flares are concerned : a ) in some cases the early flares , often blended , may be fitted by the fitting functions without subtracting the underlying continuum . in this case
the decay of the blended flares mimics the underlying continuum . a good example , but not the only one , is grb060714 , figure 2 and krimm et al . ,
2007 and b ) early flares may simply be the extension of the prompt emission spikes at lower energies especially when we are dealing [ but often we do not have the redshift ] with high redshift objects .
in [ 1 ] we used for the rising part a power law or exponential profile while for the decaying part we used a power law [ the fitting of the total observed light curve using simultaneously a broken power law and various gaussian functions was used only for the measure of width and intensity peak since it was shown that such parameters were not sensitive , within errors , to the fitting function selected ] .
the inconvenient of this function is the cusp at the peak that likely does not corresponds to reality , furthermore the derivations of fluence lack in part of homogeneity . similar inconvenient ( cusps ) are presented by other profile that we used as the norris [ 8 ] and kobayashi@xcite that were used to fit the prompt emission spikes .
kocevski , ryde and liang@xcite proposed for the prompt emission spikes a function that fits quite satisfactory the profile of the prompt emission , does not have any cusp and through which the definition of the flare parameters is quite robust .
norris 2005 defined another function that also does not have a cusp , fits properly the prompt emission and allows the robust estimate of the spikes parameters .
this works quite well also for the flares and after experimenting with the various functions proposed in the literature , we decided to choose the latter for simplicity and to ease the comparison with the parameters derived by norris for the prompt emission spikes .
this function , as shown in the figure 3 , works very nicely also for blended flares .
the present sample consists of 56 grbs and for 20 of these we have redshift . of these 36 grbs
present a single flare , 15 two flares and in 5 cases we were able to measure 3 flares .
the total number of flares of the sample is therefore 81 .
heavily blended multi - component flares are discussed separately to check whether the flare parameters are indeed the same as for single flares .
the flare width is now defined as the distance between the rising and decaying profile at the folding of the peak intensity [ width of the profile at i = 0.37 @xmath6 .
a rather symmetric profile can be fitted reasonably well also by a gaussian profile and in this case the width as defined above corresponds to @xmath7 and practically coincides with the gaussian width measured at 0.37 ipeak [ within about 1@xmath8 .
this allows easy comparisons with previous values measured by chincarini et al . since the two definition of width
are related by the relation : @xmath9 . in figure 4
we plot the ratio @xmath10 as a function of @xmath11 , as it can be seen the ratio remains rather constant as a function of time with a mean value : @xmath12 .
the large error is due to the fact , as it can be seen by the large number of points in the figure , that we plotted - to make it easier to give the global view - the values derived in all the 5 bands : 0.3 - 10 kev , 0.3 - 1 kev , 1 - 2 kev , 2 - 3 kev and 3 - 10 kev . in some of the energy bands
used the counts are smaller and the uncertainties larger . note
that according to the above relation this would correspond using a gaussian fit to @xmath13 that is in excellent agreement with the value measured in [ 1 ] : @xmath14 .
we computed all the parameters derived by norris et al .
2005 for the prompt emission spikes to verify whether or not we have similar correlations .
we do ( the verious correlations will be given elsewhere ) .
the spikes observed during the prompt emission have a large spread in symmetry for each width .
flares observed during the afterglow seem to have a rather constant mean asymmetry @xmath15 and rather large spread , figure 4 .
one of the main reasons why we derived all the light curves in 5 energy channels is to estimate the correlation of the width of the flares with energy .
these data will be complemented ( in progress ) , when available , with measures of the same flares detected simultaneously by the bat instrument . indeed due to the much broader energy coverage in this case
the result will be more robust . in figure 5
we show the result of the analysis obtained using the xrt data alone from which it is very clearly shown that the median width of the soft energy channel is larger than the median width of the harder energy channel . and the maximum @xmath16 .
errors are in the range @xmath17 and @xmath18 . ] for a detailed discussion of this grb . on the bottom
right the evolution of the hardness ratio miming the light curve . ] , as a function of the width for 76 flares in the channel 0.3 - 10 kev . the mean value is 0.37 with a standard deviation of 0.58 . ] versus the fluence ( log ) of the prompt emission .
a similar plot , as stated in the text , is obtained for the redshift sample where we can transform the fluence in energy [ see cospar paper ] .
the dashed line has a slope of -1.5 and due very likely to a sensitivity selection bias .
right : the mean luminosity has been computed spreading the flare energy over the flare t90 . as in lazzati et al .
2008 this would correspond to a rectangular profile .
note however that this approach does not account for the band correction due to the different redshifts and that it largely overestimates the decaying part of the profile so that it should underestimate the slope of the `` global flare light curve '' . ]
of the 20 grbs with measured redshift 11 have a single flare , 7 two flares and 2 three measurable early flares for a total of 31 flares .
we do not detect any correlation of the distribution with redshift [ this means we do not have a distance bias ] except that at larger z the dispersion in energy is larger with more grbs with high energies [ likely a cosmological volume effect ] .
the flare energetic will further and robustly constrain the models and related efficiency . if we plot the energy observed in the prompt emission [ 15 - 150 kev ] versus the energy observed in flares we have an ( almost ) expected result .
grbs with fainter prompt emission have also a fainter total emission in flares .
this result is however somewhat weak since we have only two faint prompt emission flares of which on is a short grb [ grb070124 ] and the second had a controversial redshift estimate [ grb060512 for which we assumed z = 0.4428 @xcite ] . in the large prompt emission energy region of the plot
we have also a tendency of a decreasing percentage of the flare emission respect to the energy emitted in the prompt emission .
it is however premature to conclude that faint grbs may have a more efficient flare emission since selection effects are not yet fully understood . in figure 6
we plot the prompt emission fluence versus the ratio xrt / bat fluence to evidence a possible bias due to a ) xrt sensitivity limit and b ) the sample of faint flares is statistically not significant and it may be improper to mix , to this end , short and long grbs .
lazzati et al . ,
[ 12 ] , estimate the average flare light curve of the subsample of swift afterglows that show flaring activity and for which redshift has been measured . while of course this is one of the things that need to be done , we have many doubts on the validity [ completeness ] of the samples we have at hand at present and on the procedures .
for instance it is true that for grb050904 we observe a flare light curve with slope of about -1 but it is also true that this slope is not observed , for instance , in grb071117a and other flares .
we will further discuss this point at the cospar meeting proceedings and present part of the analysis we are completing on this matter .
the mean flare luminosity of the grbs of the present sample and the related t90 is given in figure 6 . | the flare activity that is observed in grbs soon after the prompt emission with the xrt ( 0.3 - 10 kev ) instrument on board of the swift satellite is leading to important clues in relation to the physical characteristics of the mechanism generating the emission of energy in gamma ray bursts .
we will briefly refer to the results obtained with the recent analysis [ 1 ] and [ 2 ] and discuss the preliminary results we obtained with a new larger sample of grbs [ limited to early flares ] based on fitting of the flares using the norris 2005 profile .
we find , in agreement with previous results , that xrt flares follow the main characteristics observed in [ 3 ] for the prompt emission spikes .
the estimate of the flare energy for the subsample with redshift is rather robust and an attempt is made , using the redshisft sample , to estimate how the energy emitted in flares depends on time .
we used a @xmath0 , @xmath1 , @xmath2 cosmology .
address = inaf , osservatorio astronomico di brera , via e. bianchi 46 , i-23807 , merate ( lc ) , italy , altaddress = universit degli studi di milano - bicocca , dipartimento di fisica , piazza della scienza 3 , i-20126 milano , italy address = inaf , osservatorio astronomico di brera , via e. bianchi 46 , i-23807 , merate ( lc ) , italy address = inaf , osservatorio astronomico di brera , via e. bianchi 46 , i-23807 , merate ( lc ) , italy , altaddress = universit degli studi di milano - bicocca , dipartimento di fisica , piazza della scienza 3 , i-20126 milano , italy address = inaf , osservatorio astronomico di brera , via e. bianchi 46 , i-23807 , merate ( lc ) , italy , altaddress = universit degli studi di milano - bicocca , dipartimento di fisica , piazza della scienza 3 , i-20126 milano , italy address = inaf , osservatorio astronomico di brera , via e. bianchi 46 , i-23807 , merate ( lc ) , italy address = universit di `` roma la sapienza '' , icra , icranet , italy |
the kuiper belt is the vast swarm of small bodies that inhabit the outer solar system beyond the orbit of neptune .
these kuiper belt objects ( kbos ) that inhabit this belt are relics of the solar system s primordial planetesimal disk
they are bits of debris that failed to coalesce into other large planets .
the kuiper belt is also of great interest since it preserves a record of the outer solar system s early dynamical history .
this is reflected in the kbos curious distribution of orbits , which suggest that there was considerable readjustment of the solar system s early architecture . the possibility that the orbits of the giant planets may have shifted significantly ( that is , after the solar nebula gas had already dissipated )
was first demonstrated by the accretion simulations of @xcite ; they showed that as the growing giant planets gravitationally scatter the residual planetesimal debris , they can exchange angular momentum with the debris disk in a way that causes the planets orbits to drift .
@xcite later showed that an episode of outwards migration by neptune by at least @xmath10 au could also account for pluto s peculiar orbit , which resides at neptune s 3:2 resonance with an eccentricity of @xmath11 . in this scenario ,
pluto s large eccentricity is a consequence of it having been captured by neptune s advancing 3:2 resonance , which pumped pluto s @xmath12 up as it shepherded the small planet outwards .
further support for this planet
migration scenario is provided by the subsequent discovery of numerous other kbos also inhabiting neptune s 3:2 resonance with eccentricities similar to model predictions @xcite , as well as by more modern nbody simulations of the orbital evolution of giant planets while they are still embedded in a massive planetesimal disk ( @xcite ) .
the purpose of the present work is to use higher resolution simulations to update this conventional model of neptune s migration into the kuiper belt .
this model s strengths , as well as its weaknesses , will be assessed quantitatively by rigorously comparing the simulations endstates to current observations of the belt . in the following , we execute two simulations that track the orbital evolution of the four migrating giant planets plus @xmath13 massless test particles ( the latter representing the kbos ) over the age of the solar system . in one simulation
the initial state of the kuiper belt is dynamically cold ( _ i.e. _ , the particles have initial eccentricities and inclinations of @xmath12 and @xmath14 ) , while the second simulation is of a kuiper belt that is initially stirred up a modest amount ( _ i.e. _ , @xmath12 and @xmath15 ) .
we then use a monte carlo method to assign sizes ( and hence magnitudes ) to the simulated kbos ; this allows us to account for the telescopic biases that tends to select those kbos that inhabit orbits that are more favorable for discovery over those kbos in less favorable orbits .
then , by comparing the resulting model kuiper belts with the current observational data , we rigorously test the planet
migration scenario as well as obtain a more realistic assessment of the abundance of kbos .
this analysis will also provide the relative abundance of the belt s various subpopulations
the resonant kbos , the main belt objects , the scattered disk , plus the centaurs and neptune s trojans . the paper is organized as follows .
section [ model ] describes the so
called ` standard model ' that is considered here , as well as the numerical methods to be employed . results from two simulations of the kuiper belt are reported in sections [ cold ] and [ hot ] , while section [ i ] examines the kuiper belt inclination problem .
section [ census ] details the monte carlo model that is then used in sections [ edge][extendedsd ] to assess the relative abundance of the belt s various subpopulations , with a final tally of the belt s total population given in section [ calibration ] .
section [ unmodeled ] comments on some important unmodeled effects , and section [ discussion ] summarizes the results .
the mercury6 nbody integrator @xcite is used to track the orbital evolution of the four giant planets plus numerous massless particles . in our simulations , planet migration
is implemented by applying an external torque to each planet s orbit so that its semimajor axis @xmath16 varies as @xmath17 where @xmath18 is planet @xmath19 s final semimajor axis , @xmath20 is the planet s net radial displacement , and @xmath21 is the e fold timescale for planet migration ; this form of planet migration was first used in @xcite . to implement this in mercury6 ,
the integrator is modified so that each planet s velocity @xmath22 is incremented by the small velocity kick @xmath23 with each timestep @xmath24 .
this additional velocity kick is directed along the planet s velocity vector , and results in a torque @xmath25 being applied to each planet . since @xmath26 where @xmath27 the planet s angular momentum , these velocity kicks cause the planet s orbit to vary at the rate @xmath28 , which then recovers eqn .
( [ a_j(t ) ] ) when integrated . the simulations reported below adopt the current planets masses and orbits as initial conditions , except that their initial semimajor axes are displaced by an amount @xmath29 so that the migration torque ultimately delivers these planets into orbits similar to their present ones .
the free parameters that describe this migration are the planets radial displacements @xmath20 and the migration timescale @xmath21 . at present , there is only one strong constraint on the @xmath20 s , namely , that neptune s orbit must expand by @xmath30 au if resonance trapping is to account for the kbos having eccentricities of @xmath31 at neptune s 3:2 resonance ( see appendix [ appendix_bi ] ) .
another constraint , on the magnitude of jupiter s inward migration , can be obtained from the orbital distribution of asteroids in the outer asteroid belt .
@xcite show that the severe depletion of the outer asteroid belt can be explained if jupiter migrated inward by at least 0.2 au , and franklin et al ( 2004 ) show that the orbits of the hilda asteroids at jupiter s 3:2 resonance are consistent with jupiter having migrated inwards by about 0.45 au .
the remaining @xmath20 s for saturn and uranus are less well constrained , but stability considerations require them to be neither too large nor too small . with this in mind , our simulations adopt the following values for the @xmath20 s : @xmath32 au for jupiter , @xmath33 au for saturn , @xmath34 au for uranus , and @xmath35 au for neptune .
all of the simulations reported here also employ a planet migration timescale of @xmath36 years .
this value is supported by the self consistent nbody simulations by @xcite of the giant planets migration while they are embedded in a planetesimal disk .
those simulations show that a planetesimal disk having a mass @xmath37 m@xmath6 spread over @xmath38 au will cause neptune s orbit to expand @xmath39 au over a characteristic timescale of @xmath40 years ( see also @xcite ) .
we note that the orbital evolution adopted here is constructed so that the migrating planets eccentricities are always comparable to their present values , and that the migration proceeds along nearly circular orbits . but this particular choice for the planets eccentricities is merely a simplifying assumption since we do not know the @xmath12-evolution of the giant planets during the migration epoch .
for instance , it is possible that dynamical friction with the particle disk would have conspired to keep the planets eccentricities low , but there may also have been other transient protoplanets roaming about the outer solar system , and their perturbations would tend to pump up the planets eccentricities . given the uncertainty in the relative rates of these effects , we adopt the simplest possible model , one that assumes that the giant planets eccentricities were always comparable to their present one .
however alternate migration schemes are possible ; for instance , @xcite consider a scenario where the giant planets pump up their eccentricities as they pass through mutual resonances .
but it is uncertain as to whether this possible history would have altered the bulk properties of the kuiper belt , and it is not considered here . in order to enforce migration in nearly circular orbits , our simulations have jupiter migrating outwards @xmath41 au , whereas other self consistent simulations show that jupiter usually migrates inwards a small amount @xcite .
note that this choice avoids having jupiter approach the 5:2 resonance with saturn , which tends to excite the planets eccentricities above current levels .
but that eccentricity excitation might then have been damped back to current levels by dynamical friction with the particle disk , but that is a phenomena that goes unmodeled in our massless particle disk .
we simply avoid this event altogether by instead having jupiter migrate outwards a modest amount .
but this not a concern here since our interest is in the kuiper belt , whose endstate is not likely to preserve any memory of whether jupiter migrated slightly inwards or outwards .
the remaining @xmath20 s are similarly chosen to avoid all major resonances , and the simulated planets semimajor axes @xmath42 are also shown in fig . [ migrate ] . in the simulations described below
, the model kuiper belt is initially composed of @xmath13 massless particles having semimajor axes randomly distributed over @xmath43 au with a surface number density that varies as @xmath44 .
the inner edge of this particle disk is 1.4 au inwards of neptune s initial semimajor axis , and the disk extends well beyond the outer reaches of the observed main belt . all simulations described here use a timestep of @xmath45 years , which is sufficiently short to accurately evolve particles in eccentric orbits down to perihelia as low as @xmath46 au without suffering the perihelion instability described in @xcite .
of course , particles can still achieve orbits having perihelia lower than @xmath47 , but their orbits will not be calculated correctly in our simulations .
however this is of little consequence since these planet crossing bodies have very short dynamical lifetimes and are quickly removed from the system anyway .
accretion models have shown that the observed kbo population must have formed in an environment that was initially dynamically cold , that is , the known kbos must have formed from seeds that were in nearly circular and coplanar orbits with initial @xmath12 s and @xmath48 s that were @xmath49 @xcite . in anticipation of this , many investigations of the dynamical history of the kuiper belt have adopted initial kbo orbits that are dynamically cold ( e.g. , @xcite ) .
figure [ cold belt ] shows the results of a simulation of neptune s smooth migration into a dynamically cold swarm of massless kuiper belt objects having initial @xmath12 s that are rayleigh distributed about a mean value @xmath50 , and initial inclinations similarly distributed with a mean @xmath51 .
this system is evolved for @xmath52 years . as is well known from previous studies , neptune s smooth migration is very efficient at inserting particles into the planet s mean motion resonances , principally the 2:1 , 5:3 , and the 3:2 .
however , it is also well recognized that the endstate of this idealized model differs from the observed kbo orbits ( the red dots in fig .
[ cold belt ] ) in several ways .
for example , one prominent discrepancy is that the 2:1 resonance is densely populated with simulated particles while only sparsely populated by observed kbos .
however the discrepancy that is most important to this discussion lies in the @xmath53 au zone between the 7:4 and the 2:1 resonances , which is the outer half of the main belt that is conventionally defined as the region between the 3:2 and 2:1 resonances .
although these simulated particles managed to slip through the advancing 2:1 resonance , they still reside in orbits that are only modestly disturbed with @xmath54 and @xmath55 , whereas the observed kbos inhabit orbits that are considerably more excited .
thus neptune s smooth migration into a dynamically cold kuiper belt is unable to account for the belt s stirred up state .
evidently , some other event has also disturbed the kuiper belt , and this stirring event may have taken place prior to or after the onset of neptune s migration
. however section [ hot ] provides reason to believe that this stirring event occurred before the onset of neptune s migration into the kuiper belt .
to examine the effects of neptune s migration and its resonance sweeping of a previously stirred up kuiper belt population , we repeat the numerical integrations with @xmath13 simulated kbos , but with initial @xmath12 s rayleigh distributed about a mean value of @xmath56 and initial @xmath57 s distributed similarly about a mean value of @xmath51 . however this time
the simulation is evolved for the age of the solar system , 4.5 gyrs , with fig .
[ hot belt ] showing the resulting kuiper belt endstate .
first , we note that in this case we find an outer solar system that is far more depleted in transient particles like centaurs ( which are scattered particles having semimajor axes interior to neptune ) and scattered disk objects [ which are particles that were lofted into eccentric neptune crossing orbits due to a close
encounter with neptune @xcite ] ; those bodies usually reside in orbits having perihelia @xmath58 between the @xmath59 and @xmath60 au curves seen in figs .
[ cold belt][hot belt ] .
this difference is primarily due to the simulation s longer integration time .
another prominent difference with the ` cold belt ' simulation is that neptune s weaker higher order resonances , such as the 3:1 and 5:2 , are considerably more efficient at capturing particles when neptune migrates into a hot disk , a phenomenon that was first noted by @xcite .
this result was rather surprising , because low - order resonance capture theory theory predicts a generally lower capture probability for particles having higher eccentricities @xcite .
however , a careful examination of the theory of adiabatic resonance capture ( e.g. , @xcite ) shows that there are two reasons for this result .
( 1 ) the higher order resonances have capture probabilities that drop off more slowly with eccentricity than first order resonances : although the 1st order resonance capture probability varies as @xmath61 , the 2nd order resonance capture probability falls off more slowly as @xmath62 while the 3rd order resonance capture probability varies as @xmath63 .
( 2 ) the threshold migration speeds for adiabatic resonance capture are also lower for the higher order resonances , and they also depend more strongly upon the initial eccentricities . for capture at a @xmath64 resonance , the requirement for adiabatic resonance capture
is that neptune s migrate rate , which is @xmath65 au / yr in these simulations , be sufficiently slow , namely , that @xmath66 where @xmath67 and @xmath68 are neptune s mass and orbital period , and @xmath69 is a function of laplace coefficients .
for example , the 5:2 resonance has @xmath70 , so the migration speed threshold that permits adiabatic resonance sweeping is @xmath71 au / yr among particles having @xmath72 , while the threshold is reduced to @xmath73 au / yr among particles having @xmath74 .
it is clear then that a dynamically cold particle swarm has no chance of adiabatic capture at neptune s high order mean motion resonances , while particles that are stirred up to @xmath72 are at least near the threshold for adiabatic resonance capture . and
as @xcite point out , the fact that seven eccentric kbos are known to librate at neptune s 5:2 resonance also lends support to the pre - stirred kuiper belt scenario .
another advantage of this stirred up kuiper belt scenario is that it recovers eccentricities that are observed to be as large as @xmath75 in the main belt that lies between the 3:2 and 2:1 resonances at @xmath76 au ( the red dots in fig .
[ hot belt ] ) .
this is a feature that the cold belt scenario ( fig .
[ cold belt ] ) does not account for .
of course , fig .
[ hot belt ] also shows that the simulated main belt is densely populated by low
eccentricity particles having @xmath54 at @xmath77 au , whereas the observed kuiper belt is only sparely populated here . but section [ edge ] will show that there are a variety of possible explanations for this discrepancy such a change in the kbo size distribution , or perhaps an outer edge in the primordial kuiper belt .
figure [ hot belt ] also shows that trapping at the distant high order resonances like the 5:2 and 3:1 is quite effective at promoting bodies into eccentric orbits having @xmath78 au and perihelia @xmath79 au .
this domain is usually regarded as the scattered disk .
this result then suggests the possibility that some of the observed kbos in the @xmath79 au zone may actually be resonantly trapped bodies that are masquerading as members of the scattered disk . of course
, particles scattered by neptune also tend to spend a large fraction of their time near resonances due to the resonance sticking phenomenon ( e.g. , @xcite ) .
therefore , the discrimination between scattered and resonantly trapped particles must be done carefully . towards this end , we examine the orbital histories of all surviving particles in the shaded zone in fig .
[ sd ] that have @xmath80 and @xmath9 au .
the usual definition of being ` in ' a @xmath64 mean motion resonance is that the particle s resonance angle @xmath81 , eqn .
( [ phi_jk ] ) , librates about some fixed value with some modest amplitude @xmath82 that is usually @xmath83 , while the resonance angle for a scattered body that is temporarily ` stuck ' in a resonance will have a @xmath81 that circulates over @xmath84 . however , we find that @xmath82 is not the best discriminant for identifying trapped and scattered particles because a small but significant fraction of particles do get trapped at a resonance with a @xmath81 that is either circulating or else librating with a very large amplitude . for some trapped particles ,
this distinction is unclear due to this simulation s infrequent time sampling that occurs every @xmath85 years .
rather , a more reliable discriminant between trapped and scattered particles is based on brouwer s integral @xmath86 , eqn .
( [ b ] ) .
this integral is conserved by resonantly trapped particles but is not conserved by scattered particles that are temporarily exhibiting the ` resonance sticking ' phenomenon .
of the 134 particles that inhabit the gray zone in fig .
[ sd ] , only 12 , or about @xmath87 of these particles , are truly scattered particles whose orbits @xmath88 evolve stochastically about the @xmath79 au zone ; these scattered particles are indicated by the crosses in fig .
the remaining particles are resonantly trapped particles , most s and @xmath57 s of some resonantly trapped particles will still oscillate with constant @xmath89 in a manner that preserves their jacobi integral ; this evolution usually occurs after migration has ceased , and these particular motions do _ not _ preserve @xmath86 . ]
of which preserved their @xmath86 integral to within @xmath90 .
the orbits of all particles having perihelia @xmath91 au have also been inspected , and those resonances inhabited by trapped particles having libration amplitudes @xmath92 are indicated by the vertical dashes in the figure .
we find that particles get trapped at a number of exotic resonances like the 11:6 , 13:7 , 13:6 , 9:4 , 12:5 , 8:3 , and the 11:4 .
inspection of the inclinations shown in fig .
[ hot belt ] suggests that the smooth migration model does not produce sufficient numbers of bodies in high inclination orbits .
this has been recognized in previous studies ( @xcite ) , but has not been quantified .
however one should not directly compare the simulation s
@xmath57 s to the observed kbo inclinations , since the latter is heavily biased by telescopic selection effects .
note that most telescopic surveys of the kuiper belt observe near the ecliptic , which favors the discovery of lower@xmath57 kbos that spend a larger fraction of their time at lower latitudes @xcite . to mitigate this selection effect
, one should instead consider the ecliptic inclination distribution , which is the inclination distribution of objects having latitudes @xmath93 very near the ecliptic ( e.g. , @xcite ) .
the ecliptic inclination distribution for multi opposition kbos having perihelia @xmath94 au and latitudes @xmath95 ( red curve ) is shown in fig .
[ idist ] , as well as the simulated ecliptic inclination distribution for particles from fig .
[ hot belt ] that are selected similarly ( black curve ) .
the agreement seen in fig .
[ idist ] among bodies having inclinations of @xmath96 shows that the smooth migration model can readily recover the kuiper belt s lower inclination members . of course
, this agreement is partly due to the particles initial inclinations being distributed around @xmath97 . but fig .
[ idist ] also shows this model to be quite deficient in producing sufficient numbers of the high@xmath57 bodies having @xmath98 .
similar results are also obtained among bodies orbiting at neptune s 3:2 resonance .
this is a serious discrepancy , since @xcite has shown that there are two inclination populations in the kuiper belt : a minor population of low@xmath57 having characteristic inclinations of @xmath99 , and a high@xmath57 population having @xmath100 containing about three
quarters of all kbos .
note that these high@xmath57 bodies are very underrepresented in fig .
[ idist ] due to telescopic selection effects .
of course , neptune scattered particles routinely achieve high inclinations of @xmath98 ; for instance , many of the high@xmath57 particles seen in fig .
[ cold belt ] were scattered by neptune .
could the scattered disk be a source of the high@xmath57 kbos that are found elsewhere in the belt ?
recent nbody simulations by @xcite show that a small fraction of these neptune scattered particles can evolve from very eccentric , neptune crossing orbits into less eccentric orbits in the main belt . in gomes simulations , this occurs principally at secular and mean motion resonances that drive large oscillations in a scattered particle s eccentricity .
when a scattered particle visits a resonance , it can have its @xmath12 temporarily lowered and its @xmath58 raised . if this occurs during the planet
migration epoch , this process becomes irreversible and can strand scattered particles in the main belt with their high inclinations .
such bodies are identified by gomes as ` evaders ' since they are neptune crossing bodies that ultimately manage to evade neptune when deposited in the main belt .
note , however , that the efficiency of this process is quite low , affecting only @xmath101 of all scattered particles in the simulation that is evolved over the age of the solar system by @xcite .
however all of the high@xmath57 particles seen in our simulation ( fig .
[ hot belt ] ) achieved their inclinations while temporarily or permanently trapped in neptune s advancing resonances .
there were no neptune scattered evaders having @xmath102 that survived in our simulations . despite the evader mechanism s inefficiency , a model
can still be constructed that yields a kbo inclination distribution that is quite similar to the observed one .
for instance , this can be achieved by making the number density of small bodies initially orbiting interior to @xmath103 au about 60 times higher than the density of bodies initially orbiting beyond 27 au .
as @xcite show , neptune s migration through this densely populated inner disk creates the kuiper belt s high@xmath57 evaders , while the sparse outer disk provides the belt s low@xmath57 component .
although this scheme yields an @xmath57distribution that does indeed agree with the observations , that success is achieved via a special configuration of the initial particle disk .
however , @xcite avoid this problem of special initial conditions by assuming that the initial planetesimal disk simply ended at @xmath104 au .
this is the ` push out ' model , which argues that most of the kuiper belt is a consequence of neptune s advancing 2:1 resonance , which can drag bodies outwards to litter the main belt with low@xmath57 kbos .
the belt s high@xmath57 component is then presumed to be due to the evader mechanism .
while the push out model remains quite intriguing , it would be interesting to see this scenario subjected to greater scrutiny to see whether it can indeed reproduce the kuiper belt s curious mix of high and low inclination kbos in a self consistent manner .
figure [ abundance ] plots the relative abundance , over time , of the simulated belt s various dynamical classes among particles having perihelia @xmath105 au .
these curves are normalized such that the final abundance of the main belt ( mb ) , where @xmath106 au , is unity .
note that this model predicts a 2:1 resonance that is 1.4 times more abundant than the main belt , and 2.5 times more abundant than the 3:2 , while the observations ( fig .
[ hot belt ] ) show a 2:1 that is only sparsely populated .
of course , when comparing the simulated population to the observed population , one must first deal with the observational selection effect that strongly favors the discovery of larger and/or nearer kbos .
however it is shown below that the effects of this bias can be accounted for by using a monte carlo method that assigns random sizes to the simulated population .
this then allows one to make a fair comparison of the relative abundances of the simulated and observed populations .
begin by letting @xmath107 the number of bodies in the simulated population having radii exceeding a radius @xmath108 .
also let @xmath109 be a random number that is uniformly distributed between zero and one , and interpret this number as the probability of selecting a body with a radius that is smaller than @xmath108 .
this is also equal to the probability of _ not _ selecting a body of radius @xmath110 , so @xmath111 where @xmath112 is the total number of bodies in this population .
since most small body populations have a cumulative size distribution @xmath113 that varies as a power law , adopt @xmath114 where @xmath115 is the radius of the smallest member of the swarm .
then @xmath116 , but @xmath117 can be replaced with @xmath109 since these random numbers have the same distribution , so @xmath118 equation ( [ rrandom ] ) is then used to generate random sizes for the simulated population of fig .
[ hot belt ] that have apparent r band magnitudes of @xmath119\end{aligned}\ ] ] where @xmath120 is the particle s heliocentric distance , @xmath121 au , @xmath122 is the sun s apparent @xmath108 magnitude , @xmath123 km is adopted here , and the observation is presumed to occur at solar opposition .
all of our calculations will also adopt the usual albedo of @xmath124 so that our findings can be readily compared to past results obtained by others .
however if an alternate albedo @xmath8 is desired , simply revise all kbo radii reported here by a factor of @xmath125 , and all masses by a factor of @xmath126 . finally , note that a power law size distribution results in a cumulative luminosity function that varies as @xmath127 , where @xmath128 is the sky plane number density of kbos brighter than apparent magnitude @xmath129 and the logarithmic slope is @xmath130 @xcite .
hubble space telescope observations reveal that the bright end of the kuiper belt s luminosity function has a steep logarithmic slope of @xmath131 for bodies having magnitudes @xmath132 , while the faint end ( @xmath133 ) of the luminosity function has a shallow logarithmic slope of @xmath134 @xcite ; the steeper slope of the bright end of the luminosity function was also confirmed recently by @xcite .
this luminosity function can be interpreted as evidence that the kbo size distribution is actually two power laws that break even at a magnitude @xmath135 , which corresponds to a body of radius @xmath136 km orbiting at a characteristic distance of @xmath137 au assuming it has an albedo of @xmath124 .
however our application will concentrate only on those kbos that have known orbits , and @xmath138 of those bodies have magnitudes @xmath139 .
consequently , this study will be sensitive only to the larger end of the kbo size spectrum , and such bodies will be characterized here via a single power
law size distribution having @xmath140 and @xmath141 . although @xmath142 simulated particles in fig .
[ hot belt ] manage to survive over the age of the solar system , the monte carlo model assigns far too few of them with sizes that would be detected by any telescopic survey of the kuiper belt . to boost the statistics of the detectable portion of the simulated population , each survivor is replicated @xmath13 times such that each particle s orbit elements @xmath143 are preserved while its mean anomaly @xmath144 is randomly distributed over @xmath145 .
it should be noted that this step effectively assumes that the particles longitudes are uniformly distributed over @xmath146 , which is not quite correct since neptune s resonant perturbations tend to arrange the particles longitudes in a non uniform manner @xcite .
nonetheless , this is not a major concern since the observed kbos were discovered along lines of sight that are roughly distributed uniformly in ecliptic longitude , which effectively washes out neptune s azimuthal arrangement of the belt ; see appendix [ symmetry ] for a more detailed examination and justification of this assumption .
lastly , this monte carlo model is then tested by verifying that the randomly generated population does indeed exhibit the expected luminosity function that varies as @xmath147 .
further comparison of the monte carlo model of the kuiper belt to any observations must be done carefully . note that the brighter kbos tend to be discovered in shallow , wide
angle surveys that observe a large area @xmath148 on the sky , while the fainter kbos tend to be discovered in deeper surveys that observe smaller areas @xmath148 .
consequently , the observed abundances of the various kbo subclasses ( e.g. , the main belt , the scattered disk , _ etc .
_ ) are proportional to all of these surveys total area @xmath149 , which itself is some function of the limiting magnitude @xmath150 .
however appendix [ appendix_ratios ] shows that this dependence upon @xmath149 can be factored out by constructing ratios of the belt s various subclasses .
that appendix also shows that the ratios of the observed abundance of any two dynamical classes of kbos is approximately equal to the ratio of the intrinsic abundances of the much larger unseen populations . ] .
thus by plotting _ ratios _ of the simulated populations to the observed kbo populations , we can compare the model to the observations in a manner that is insensitive to survey details like their individual sky
coverage @xmath148 .
figure [ ratios ] shows the apparent abundances of the 2:1 and the 3:2 populations relative to the main belt ( mb ) as a function of their @xmath108band magnitudes @xmath150 .
the upper light gray curve is the simulated ratio , which predicts an apparent 2:1 abundance that is about @xmath151 of the mb , while the dark gray curve is the observed ratio .
taking the ratio of these two curves reveals that their discrepancy at magnitudes @xmath152 ( which refers to about 90% of the observed sub
populations ) is a factor of @xmath153the observed 2:1 resonance is markedly underabundant relative to the observed main belt population .
there are @xmath154 monte carlo particles in the simulated main belt having radii @xmath3 km , and @xmath155 particles in the 2:1 . if we let @xmath156 represent the inferred ratio of 2:1 to main belt objects , then @xmath157 , which is comparable ( albeit lower by a factor of @xmath158 ) to the ratio that @xcite infer from telescopic surveys of the kuiper belt .
the observed 3:2/mb ratio plotted in figure [ ratios ] also shows that this resonant population is underabundant relative to model predictions by a factor of @xmath159 among bright objects with @xmath160 , and by a factor of @xmath161 at fainter magnitudes .
close inspection of the observations suggests that there indeed is a deficiency of fainter kbos in the 3:2 , and that this curve is not due to some overabundance of main belt kbos having magnitudes of @xmath162 .
it should be noted that the results given in fig .
[ ratios ] are not particularly sensitive to the detailed location of the main belt s outer edge . for instance , if we assumed the belt s primordial edge where instead at @xmath163 au ( e.g. , @xcite ; see also section [ edge ] ) , this reduces the 2:1 and mb populations both by about 40% while leaving the 3:2 population unchanged .
consequently , the 2:1/mb ratios of fig . [ ratios ] are largely unchanged for both the simulated and observed populations , while the 3:2/mb ratios increase by a factor of @xmath164 .
however the _ discrepancy _ between the simulated and observed populations is still the same factors of @xmath15960 .
although there are several possible interpretations of the discrepancies seen fig .
[ ratios ] , the most plausible explanation is that other unmodeled processes are responsible for ( _ i. _ ) reducing the trapping efficiencies of the 2:1 and 3:2 resonances by factors of @xmath15960 , or ( _ ii . _ ) causing trapped particles to diffuse out of the resonances and into nearby regions of phase space that are quite unstable ( cf .
fig . 1 of @xcite )
, resulting in their ejection from the kuiper belt .
such unmodeled processes include the collisions and gravitational scatterings that occurred with ever greater vigor during earlier times when the belt was more crowded .
the scattering of these planetesimals by neptune was of course responsible for driving that planet s migration , so the occasional scattering of a large and/or close planetesimal will cause that planet s orbit and hence its resonances to shudder some . likewise ,
scattering events among the kbos themselves would also cause their semimajor axes to diffuse some , as would collisions .
this means that scatterings and collisions will have driven a random walk in the resonant particles semimajor axes , as well as a random walk in the location of the resonances themselves .
it is possible then that these unmodeled effects can drive particles out of resonances and reduce the resonant population by the large factors indicated by fig .
[ ratios ] , a scenario that is also explored in simulations by @xcite and @xcite .
the magnitude dependence of the observed 3:2/mb ratio shown in fig .
[ ratios ] is also quite curious .
the fact that the observed ratio varies with apparent magnitude @xmath150 , while the simulated ratio remains constant at magnitudes fainter than @xmath165 , suggests that the @xmath166 power
law that was universally applied throughout the entire belt is overly simplistic .
one way for the model to achieve better agreement with the observations is to assume that larger , brighter bodies are more abundant in the 3:2 , and that smaller , fainter bodies are less abundant there than they are in the mb , which requires a shallower size distribution .
the dashed curve in fig . [ ratios ] illustrates this possibility , which shows the simulated 3:2/mb ratio assuming that the 3:2 bodies have a shallow @xmath167 size distribution with @xmath168 km ( note that reducing @xmath115 has the effect of reducing the total number of visible objects ) while the mb bodies have the usual distribution with @xmath169 and @xmath123 km .
the shallow size distribution that is inferred here for the 3:2 population is also consistent with the logarithmic slope of @xmath170 that @xcite recently reported for the luminosity function of their ` resonant ' population that is dominated by 3:2 kbos ; the size distribution inferred from that work is @xmath171 .
kbo sizes can also vary with inclination [ @xcite , but see also footnote 1 of @xcite ] .
in particular , @xcite report that the bright end of the luminosity function for high inclination ( @xmath172 ) kbos have a logarithmic slope of @xmath173 and a size distribution @xmath174 that is much shallower than the low@xmath57 kbos having @xmath175 and a @xmath176 . in our monte carlo model
there are only @xmath177 bodies larger than @xmath178 km , so their numerical abundance relative to the main belt is @xmath179 , which again is comparable ( but again lower by a factor of @xmath158 ) to the ratio reported in @xcite .
the largest monte carlo body at the 3:2 has a radius @xmath180 km , which is comparable to the size of the largest multi opposition kbo there . ] with @xmath181 km , assuming @xmath124 . this range of power
law indices that is inferred for the kuiper belt , @xmath182 , is comparable to the values of @xmath183 that are observed at various sites throughout the asteroid belt .
the near earth objects have a fairly shallow size distribution with @xmath184 @xcite , while the asteroid families exhibit steeper size distributions .
for instance , fig . 1 .
of @xcite shows size distributions for several prominent asteroid families having values of @xmath185 .
note also that nonfamily asteroids have @xmath186 @xcite , which is slightly steeper than the canonical @xmath187 value that results from a collisional cascade @xcite . since the various asteroid subclasses exhibit such a wide variation in their size distributions over a relatively narrow range of semimajor axes of @xmath188 au ,
perhaps it should be of no surprise that the spatially much wider kuiper belt might also exhibit some variety in @xmath183 .
inspection of figure [ hot belt ] shows a prominent absence of observed kbos having modest eccentricities of @xmath72 near and beyond neptune s 2:1 resonance .
the prevailing interpretation of this observed feature is that there is a boundary near @xmath189 au that marks the outer edge of the solar system s primordial kuiper belt @xcite .
the dots in fig . [ erosion ] shows the belt s surface density @xmath190 that is inferred from these observations , which peaks at @xmath191 au .
we have nonetheless allowed our simulated kuiper belt to extend out to @xmath192 au in order to use the dearth of observed distant kbos to place quantitative upper limits on the abundance of hypothetical kbos that might live beyond 50 au .
the nbody / monte carlo of the previous section can be used to predict how many kbos should have been observed in the @xmath9 au zone ( which we identify here as the outer belt , or ob ) assuming ( _ i. _ ) that the primordial kuiper belt extends smoothly out to @xmath192 au , and ( _ ii . _ ) that all kbos everywhere have the same size distribution with the usual parameters @xmath169 and @xmath123 km .
this simulation s ratio of outer belt / main belt objects , @xmath193 , is plotted versus magnitude @xmath150 in fig .
[ outer_belt ] .
this is the ratio of @xmath194 the number of bodies in the simulated outer belt ( whose members have semimajor axes @xmath195 au and eccentricities @xmath196 ) , to @xmath197 the number of bodies in the simulated main belt ( where @xmath198 au ) in the magnitude interval @xmath199 where @xmath200 . according to the figure , the expected ob / mb ratio is @xmath201 . at present
there are @xmath202 kbos in the mb that have been observed for 2 or more oppositions , and the dimmest member of this group of kbos has an apparent magnitude @xmath203 .
the nbody / monte carlo model thus predicts that there should also be @xmath204 objects brighter than @xmath205 orbiting in the ob beyond @xmath206 au .
this prediction is in marked contrast with the observations which show that there are no known multi
opposition objects orbiting in the ob with magnitudes brighter than @xmath205 , _ i.e._,@xmath207 .
this prediction that the outer belt would have an observed abundance that is @xmath208 of the main belt differs considerably from that of @xcite who estimated that the observed outer belt population should only be @xmath209 of the total observed population .
however this much lower estimate was obtained by assuming that the current belt s surface number density resembles its primordial @xmath190 which likely varied as @xmath210 or so .
this very common assumption causes the inner part of the belt to be more concentrated than its outer part .
however a real kuiper belt has been dynamically eroded over the eons by the giant planets gravitational perturbations .
this dynamical erosion is illustrated in fig .
[ erosion ] , which shows the simulated belt s primordial surface density ( gray curve ) and its final surface density ( dashed curve ) ; similar erosion is also seen in the long term integrations of @xcite .
figure [ erosion ] shows that @xmath190 for an eroded belt is a sharply increasing function of @xmath120 for @xmath211 au , which implies that the inner observable portion of the belt is very underdense relative to the more distant @xmath212 au zone .
this dynamical erosion accounts for the discrepancies between our outer belt predictions and that by @xcite .
recall that figs .
[ hot belt ] and [ ratios ] show that the 2:1 and 3:2 populations are very depleted relative to model predictions , and that the zone beyond neptune s 2:1 resonance is either empty or inhabited by bodies too small and faint to be seen . to account for these depletions , the solid curve in fig .
[ erosion ] also shows a revised surface density curve that is obtained from the simulated belt that is truncated at @xmath163 au ( about 3 au inwards of neptune s 2:1 resonance ) , and with the negligible contribution from the 3:2 populations also being ignored .
this result is a curve that agrees quite well with the belt s observed surface density variations . despite this good agreement in the radial distributions of the simulated and observed kuiper belts , fig .
[ hot belt ] shows that this apparent edge at @xmath163 au is still rather fuzzy since there are four multi - opposition kbos of low eccentricity ( @xmath213 ) orbiting in the main belt at @xmath214 au .
close inspection of fig .
[ hot belt ] shows that a hard edge at @xmath163 au also could not account for the kbos having @xmath72 in the @xmath214 au , unless the advancing 2:1 resonance also dragged some bodies out of the @xmath215 au zone and deposited them here , reminiscent of the scenario suggested by @xcite . the remainder of this section places upper limits on the size and abundance of any unseen kbos that might lurk beyond @xmath206 au .
of course there are multiple interpretations of the dearth of observed multiple
opposition bodies orbiting beyond @xmath206 with modest eccentricities of @xmath72 , _
i.e. _ , that @xmath207 .
one interpretation of this upper limit is that assumption ( _ i. _ ) is incorrect that the primordial kuiper belt s density did not extend smoothly beyond neptune s 2:1 , but that it instead was reduced by a factor @xmath216 ( relative to the smooth model s density ) in the @xmath9 au zone . in this case
, the ob / mb ratio becomes @xmath217 , which implies that the primordial density of the ob was smaller than the mb by a factor @xmath218 .
alternatively , assumption ( _ ii . _ ) could be incorrect , namely , that the kbo size distribution was not uniform everywhere .
for instance , the absence of any multi
opposition bodies in the ob having magnitudes brighter than @xmath203 could simply mean that bodies beyond @xmath219 au are dimmer than @xmath205 and thus have radii smaller than @xmath220 km ( see eqn .
note that @xcite obtained a similar limit , but that they came to regard this scenario as unlikely .
it is also possible that the outer belt s size distribution is steeper , _
i.e. _ , has a larger @xmath183 , than the main belt s size distribution .
an increase in @xmath183 decreases the abundance of bright bodies , as is illustrated by the curve in fig .
[ outer_belt ] which gives the the @xmath221 ratio for an ob having a @xmath222 size distribution while bodies in the mb have the usual @xmath169 distribution .
this particular @xmath183 is also the minimum value that is consistent with the observed upper limit of @xmath223 ; outer belts with with a smaller @xmath183 would contain at least 1 kbo brighter than @xmath205 in the @xmath9 au zone for every 264 kbos detected in the main belt , while an ob having a larger @xmath183 would be undetected .
this particular model is near the threshold of detection , and its largest member has a radius of @xmath224 km . @xcite also considered this scenario , but they concluded that the absence of distant kbos requires a steeper @xmath225 size distribution .
the origin of this discrepancy is unclear .
it is also interesting to note that the low
inclination kbos have a logarithmic slope of @xmath175 along the bright end of their luminosity function @xcite , which implies a steep size distribution of @xmath176 .
such bodies , if they inhabit the outer belt beyond @xmath206 au with the same abundances as adopted by our model , could conceivably have avoided detection to date due to their steep size distribution . in other words , a main belt
whose low@xmath57 population extends beyond @xmath206 au while its high@xmath57 population terminates at @xmath206 au could be quite consistent with their non detection . from these considerations
it may be concluded that the observed absence of multi opposition kbos in the @xmath9 au zone having modest eccentricities @xmath72 implies : ( _ i. _ ) that this part of the primordial kuiper belt was underdense by a factor of @xmath218 relative to the @xmath226 au zone , or that ( _ ii . _ ) these distant kbos have radii @xmath227 km , or that ( _ iii . _ ) their size distribution has a power
law index @xmath228 , or perhaps ( _ iv .
_ ) some combination of the above effects .
it is generally accepted that centaurs are those bodies that have diffused inwards from the kuiper belt into orbits that cross the giant planets ( e.g. , @xcite ) .
presently there are 27 known centaurs observed for more than one opposition ; these are the red dots in fig .
[ centaur_fig ] having @xmath229 au . since planet
crossers are quickly ejected or accreted , centaurs have short dynamical lifetimes of only @xmath230 years @xcite .
consequently , the density of these ` escapees from the kuiper belt ' ( e.g. , @xcite ) is very tenuous inside of @xmath231 au ( see fig . [ erosion ] ) .
indeed , only @xmath232 centaurs are detected during the final @xmath233 billion years of our simulation that was only sparsely time sampled once every @xmath85 years , so the instantaneous number of centaurs is @xmath234 at the end of the simulation .
there are also @xmath235 bodies in the main belt , so the centaur / main belt ratio is provisionally estimated at @xmath236 . the open circles on fig . [ centaur_fig ]
show the orbital elements of these seven centaurs at time @xmath237 years , which is at a time when planet migration has only recently ceased .
thus the open dots indicate the locations where neptune has parked these proto centaurs in the kuiper belt .
note that all seven centaurs originate from sites in / near neptune s mean motion resonances , namely , the 3:2 , 5:3 , 13:7 , 2:1 , and the 5:2 .
their subsequent motions at times @xmath238 years are shown as black dots ( again , poorly time sampled ) , which show that the eccentricities of nearly all proto
centaurs initially wander up and down with constant @xmath89 until they have a close encounter with neptune , scatter off that planet , make a brief apparition in the @xmath229 au centaur zone , and then are quickly removed from the system .
these seven bodies have initial semimajor axes of @xmath239 au at time @xmath240 , so centaurs can also be regarded as samples that have been drawn from a wide swath of the outer solar nebula .
[ centaur_fig ] also shows that the simulated centaurs are concentrated just inside of neptune s orbit ; their mean heliocentric distance is @xmath241 au , and their mean inclination is @xmath242 .
note also that three of the seven centaurs emerged from the 3:2 and 2:1 resonances , which section [ census ] showed to be heavily depleted relative to the model s predictions .
consequently , the centaur / main belt ratio reported above should instead be interpreted as an upper limit , e.g. , @xmath243 .
it is shown later in section [ calibration ] that our model predicts that there are @xmath244 main belt kbos having radii @xmath3 km , so this model also predicts that there are @xmath245 similarly
sized centaurs .
although the centaur upper limit reported here is comparable to the population that @xcite infer from the centaur luminosity function , there is still a prominent disconnect in the heliocentric distances of the simulated and observed populations ; our simulated centaurs all reside at @xmath246 au , while the three centaurs that @xcite used to construct the centaur luminosity were detected at heliocentric distances of @xmath247 au
. one possible interpretation of this excess of centaurs at @xmath248 au is that centaurs may be breaking up and spawning new centaurs ( e.g. , @xcite ) as they wander among the giant planets .
finally , we note that deep , wide angle surveys of the kuiper belt , such as the legacy survey that is currently being implemented at the canada france hawaii telescope , may soon reveal the existence of the more distant centaurs that is anticipated by this model to reside at distances of @xmath249 au .
figure [ hot belt ] also shows that @xmath250 particles managed to survive the length of the simulation at neptune s 1:1 resonance .
these simulated particles are of course neptune s trojans , of which two are presently known : 2001 qr@xmath251 @xcite and 2004 up@xmath252 @xcite . for this simulation
the trojan / mb ratio is @xmath253 , where @xmath235 is the number of survivors that persist in the main belt .
the spatial coordinates of the two observed and five simulated trojans are shown in figure [ trojans ] , which indicates that these particles can roam about with longitudes of @xmath254 from neptune s triangular lagrange points with semimajor axes of @xmath255 au from neptune s .
the extent of these trojan sites are similar to that seen in integrations by @xcite and @xcite .
note that no special effort was made to start any of the simulated particles at neptune s lagrange points .
rather , all particles were distributed randomly about a disk according to a smooth surface density law , with the inner edge of the disk being well inside of neptune s initial tadpole region . in our simulation
the five survivors had initial semimajor axes of @xmath256 au from neptune s initial @xmath89 , and there were a total of 68 particles initially in neptune s trojan source region ( _ i.e. _ , @xmath257 au and @xmath258 ) , so the surviving trojan fraction is about @xmath259 .
this survival fraction is comparable to that obtained by @xcite in a similar simulation . that work also showed that as planets migrate , several secondary resonances sweep across the 1:1 , which result in a heavy loss of neptune s trojans during the migration epoch .
neptune s trojans are of interest since they might place constraints on some models of the early evolution of the outer solar system .
for example , @xcite postulate that neptune originally formed in the vicinity of jupiter and saturn and was tossed outwards after scattering off the larger planets .
but the existence of 2001 qr@xmath251 and 2004 up@xmath252 might cast doubt on this scenario since trojans seem unlikely to persevere at neptune s langrange during such a scattering event .
however it has since been shown that a recently scattered neptune can still acquire its trojans later as that planet s orbit is circularized by a dense kuiper belt ( levison 2005 , personal communication ) .
it is also conceivable that neptune may have captured its trojan from a heliocentric orbit _ after _
neptune s orbit has settled down .
although @xcite shows that the direct capture of trojans from heliocentric space is rare and results in only transient trojans , @xcite show that mutual collisions may have inserted small bodies into stable orbits at neptune s lagrange points after its orbit has circularized .
figure [ oss ] shows the orbits of those scattered particles that have been tossed into very wide orbits about the sun .
most of the simulated scattered particles have perihelia between @xmath79 , as do most of the observed scattered kbos .
however there are two exceptions to this rule , namely , 2000 cr@xmath260 and 2003 vb@xmath261 ( also known as sedna ) which have respective perihelia of @xmath262 @xcite and @xmath263 au @xcite .
@xcite classify those scattered kbos having perihelia higher than @xmath264 au as members of a so called extended scattered disk .
sedna s large radius of @xmath265 km makes this object a particular curiosity since its discovery circumstance suggests that there may be a few hundred other unseen senda
sized objects @xcite .
since sedna has a mass of @xmath266 m@xmath6 , the implied mass that might be hidden in the extended scattered disk is a few tenths of an earth mass
. thus sedna by itself may represent an enormous reservoir of unseen mass that is comparable to the ` conventional ' kuiper belt ( see section [ calibration ] ) .
the extended scattered disk is also of dynamical interest since , as @xcite note , most dynamical models of the kuiper belt ( including this one ) generally produce scattered objects in lower perihelia orbits having @xmath267 au .
@xcite review a number of scenarios that might explain how a kbo might get promoted from a nearly circular orbit into a wide , eccentric orbit having @xmath268 au ; these include : ( _ i. _ ) chaotic diffusion of scattered bodies , ( _ ii . _ ) gravitational scattering by long gone massive protoplanets , ( _ iii . _ ) scattering by an undiscovered distant planet , and ( _ iv .
_ ) scattering by a single star that passes to within @xmath269 au of the sun .
however all of these scenarios are problematic .
for instance , billion
year integrations of @xmath270 particles in chaotic neptune scattered orbits fail to diffuse into orbits having perihelia as high as that of 2000 cr@xmath260 @xcite .
@xcite also cast doubt on scenarios ( _ ii iii .
_ ) by showing that @xmath271 of any distant population of protoplanets would have persisted over the age of the solar system , and that some fraction of these large bodies should already have been discovered by one of the various wide angle kuiper belt surveys . scenario ( _ iv .
_ ) is also in doubt since simulations of a close encounter with a single star generally produce disturbances in an outer kuiper belt that is quite unlike that seen in the observed belt ( e.g. , @xcite ) .
however @xcite have shown that repeated encounters by more distant stars can produce sedna like orbits .
this may have occurred early while the sun was still a member of the open cluster from which it presumably formed . in this scenario ,
the giant planets scatter small bodies into wide orbits of @xmath272 to 1000 au , which the nearby cluster stars then perturb into sedna like orbits having higher perihelia .
scattering by a passing star was recently re
examined by @xcite , and their simulations also support this scenario .
although our simulations did not produce any sedna like objects in orbits that are well decoupled from the giant planets , we did find a single scattered object in a 2000 cr@xmath260like orbit in the extended scattered disk with a semimajor axis of @xmath273 au and a perihelion of @xmath274 au ( the large black dot in figure [ oss ] ) .
the orbital history of this scattered particle is shown in fig .
[ p2847 ] , which shows that as the particle inhabited neptune s 16:3 resonance during times @xmath275 gyrs , some process raised this scattered particle s perihelion up and into the extended scattered disk on a billion year timescale .
this kind of behavior was first reported in @xcite and @xcite , whose simulations also show that some scattered particles can achieve high perihelia orbits while in or near mean motion resonances .
however we have not identified any particular resonance as being responsible for raising the perihelia of our one cr@xmath260candidate shown in fig .
[ p2847 ] .
for instance , a kozai resonance is not implicated since the argument of perihelion @xmath276 does not librate .
the possibility of other pluto like ` super resonances ' ( e.g. , @xcite ) was also examined ; this is the libration of a resonance angle of the form @xmath277 , where @xmath278 is the particle s longitude of the ascending node and the @xmath279 subscript refers to neptune s orbit elements .
angles having @xmath280 were examined , and although the angle @xmath281 angle did in fact librate for about 1 gyrs , that occurred well after the time when the particle s @xmath58 was raised .
a resonance involving interactions with multiple planets is also unlikely since the particle s tisserand parameter @xmath282 ( which is simply its jacobi integral sans the interaction energy due to neptune ) was well preserved during these times .
although the particular mechanism that drove this particle into the extended scattered disk in not understood , this particle does demonstrate that it is indeed possible for scattered particles to diffuse into the extended scattered disk via planetary perturbations alone , with other external agents ( like a stellar encounters ) being absent .
this transport from the scattered disk to the extended scattered disk via mean motion resonances is also evident in the simulations of @xcite and @xcite .
however this transport has an extremely low flux since only one of the @xmath283 particles initially in the scattered disk did manage to enter the extended scattered disk and persist over the age of the solar system . external perturbations from passing stars @xcite may indeed be more effective at producing members of the extended scattered disk .
note also the simulated particles represented by crosses in fig .
[ oss ] . even though their perihelia of @xmath284 au might suggest that they also inhabit the domain of the extended scattered disk , they are in fact resonant particles that were trapped at the 3:1 , 7:2 , and 4:1 resonances during the migration epoch .
most of these particles have libration amplitudes less than @xmath285 .
these particles also had initial semimajor axes of @xmath286 au , which is noteworthy since , if any resonant kbos are ever discovered in these orbits , they could be interpreted as evidence that the outer edge of the solar system lies beyond @xmath286 au .
however that interpretation would still be ambiguous , since neptune scattered evaders , which originated from smaller semimajor axes , can also settle into these same resonances see fig .
18 of @xcite for an example .
figure [ sd_ratio ] shows the apparent abundance of so called scattered disk objects relative to the main belt , as predicted by the nbody / monte carlo model .
there are @xmath287 monte carlo bodies in orbits having @xmath288 au and perihelia @xmath289 au , while @xmath290 monte carlo particles survive in the main belt , so the model predicts an intrinsic sd / mb ratio of @xmath291 .
the apparent ratio of these two populations is @xmath292 , also shown in fig .
[ sd_ratio ] .
note that the intrinsic sd / mb ratio inferred here is about one fourth that reported in @xcite .
the ecliptic luminosity function of @xcite is shown in fig.[lum_fn_fit ] , and its bright end varies as @xmath293 deg@xmath294 where @xmath140 and @xmath295 , for magnitudes @xmath296 .
this luminosity function gives the number density of kbos near the ecliptic that are brighter than magnitude @xmath150 .
since this curve scales with the total number of kbos , it can be used to calibrate the simulation to determine the total number of objects in the kuiper belt .
sections [ census][edge ] show that the observed 3:2 and 2:1 populations are severely depleted relative to model predictions , and that bodies in an outer belt beyond the 2:1 are either absent or too faint to be seen . to account for these depletions ,
the truncated kuiper belt similar to that of section [ edge ] is adopted ; this belt is formed by discarding any bodies orbiting beyond the 2:1 , as well as all bodies orbiting within @xmath297 au of neptune s 3:2 and 2:1 resonances .
there are @xmath298 nbody particles in this truncated kuiper belt , and they are replicated @xmath299 times with sizes and magnitudes assigned to them according to the monte carlo method of section [ census ] , with @xmath169 and @xmath123 km .
the simulation s median inclination is low ( e.g. , section [ i ] ) , only @xmath300 , which is much lower than the median inclination @xmath301 that is inferred from the debiased kbo inclination distribution reported by @xcite . due to these low inclinations ,
the simulation s ecliptic luminosity function would thus be artificially overdense by a factor @xmath302 , so it is revised downwards by this factor to compensate .
the simulation s @xmath303 is then multiplied by a factor @xmath304 to fit it to the bright end of the observed luminosity function ; this accounts for the different populations in the simulated and observed kuiper belts , and results in the curve shown in fig .
[ lum_fn_fit ] .
the size distribution adopted here is valid down to a radius of about @xmath305 km ( see section [ census ] ) , so the inferred number of kbos larger than @xmath306 is @xmath307 . to estimate the total number of kbos larger than the fiducial radius of @xmath308 km , note that the faint end of the observed luminosity function has a logarithmic slope of @xmath309 @xcite , which implies a power
law index of @xmath310 for bodies having radii @xmath311 .
the total number of bodies larger than @xmath312 is thus @xmath313 .
the total mass of these bodies is obtained from their cumulative size distribution , which for the large bodies with @xmath314 can be written @xmath315 .
the differential size distribution is then @xmath316 , and if @xmath317 mass of a body having a radius @xmath108 , the total mass of bodies having radii in the interval @xmath318 is @xmath319 n_{break}m_{break}\end{aligned}\ ] ] where @xmath320 gm , which is the mass of a body of radius @xmath305 km assuming it has a density @xmath7 and albedo @xmath8 .
the total mass of kbos larger than @xmath306 with semimajor axes inside of neptune s 2:1 is thus @xmath321 m@xmath6 for a @xmath169 size distribution that extends to radii as large as @xmath322 km . to get the total mass of bodies at the fiducial size @xmath323 , add to the above the mass of bodies in the size interval @xmath324 , which is roughly @xmath325 .
the total mass of bodies larger than @xmath308 km is then @xmath326 m@xmath6 .
note that the @xmath327 m@xmath6 prefactor is a consequence of adopting the oft
employed halley albedo of @xmath124 .
however recent observations indicate kbos have an average albedo of @xmath328 @xcite , which in turn lowers the kuiper belt mass to @xmath329 m@xmath6 assuming they have a unit density .
this population estimate is comparable to , but a bit higher than , previous estimates that rely on far simpler models of the kuiper belt .
for instance , @xcite report a main belt population of @xmath330 objects of mass @xmath331 m@xmath6 among bodies having radii @xmath3 km .
they also estimate the belt s total population to be @xmath332 the main belt population , so a total population of @xmath333 bodies larger than @xmath312 having mass @xmath334 m@xmath6 is inferred .
a similar estimate is also inferred from the hst survey by @xcite ; according to their fig .
8 , the sky
plane number density of kbos larger than @xmath312 is @xmath335 deg@xmath294 .
since the kuiper belt subtends a total solid angle of @xmath336 deg@xmath337 @xcite , the total number of kbos larger than @xmath312 is @xmath338 having a total mass of @xmath339 m@xmath6 .
sections [ census][extendedsd ] show that the simulated belt s various dynamical classes have abundances of @xmath340 , @xmath341 , @xmath243 , @xmath342 , and @xmath343 relative to the main belt , so the main belt fraction is @xmath344 and thus there are are @xmath345 main belt kbos having radii @xmath3 km .
the numerical abundance of the @xmath346 dynamical class is @xmath347 , and its mass is @xmath348 where @xmath349 m@xmath6 assuming @xmath350 gm/@xmath351 and @xmath124 ; these abundances and masses are listed in table [ pop_table ] .
the exception is the 3:2 mass estimate which adopts the @xmath167 power law size distribution described in section [ census ] ; if this subgroup really does have such a flat size distribution , then eqn . ( [ mtotal ] ) must be used to calculate its mass , @xmath352 , and @xmath353 replaced by @xmath312 , @xmath354 , and @xmath355 . ] .
note also that the preceding ratios assume that the main belt terminates just inwards of the 2:1 resonance at @xmath356 au .
if , however , one wishes to adopt an outer edge at @xmath163 au , then section [ census ] shows that this reduces the mb population by @xmath208 , so that the ratios @xmath120 quoted above should then be raised by a factor of 1.7 .
the exception to this rule are the bodies at the 2:1 resonance their abundance relative to the mb is unchanged .
however the _ total _ number of kbos reported here is still insensitive to the detailed location of the main belt s outer edge , since that number is obtained by fitting the simulated kbos luminosity function to the observed @xmath303 , which is quite insensitive to the detailed location of the main belt s outer edge .
it should also be noted that this study employed an initial @xmath357 disk surface density , but our findings are readily adapted for an alternate surface density law .
for instance , if the canonical @xmath358 law were instead desired , then this shallower power law would result fewer objects trapped in the 3:2 resonance relative to the main belt population .
since the 3:2 objects are drawn from the @xmath359 au part of the disk , while the main belt objects formed at @xmath360 au , this revised surface density law would reduce the 3:2/mb ratio reported here by a factor @xmath361 , which is a @xmath362 change in relative abundance .
of course , the 2:1/mb ratio would remain unchanged since their source populations are the same .
also , we conservatively interpret the abundance of neptune s trojans reported table [ pop_table ] as an upper limit on their real abundance .
it was argued in section [ census ] that other unmodeled processes , possibly the scattering of planetesimals by neptune or amongst themselves , reduced the trapping efficiency of the 3:2 and 2:1 resonances by factors of @xmath363 .
thus it is possible that the same unmodeled phenomena might also have destabilized orbits at neptune s 1:1 resonance , so the actual number of trojan survivors may be smaller than that reported in section [ trojans ] .
finally , upper limits on the abundance of kbos inhabiting a hypothetical outer belt are reported for the @xmath195 au zone assuming these bodies have the shallowest possible size distribution , namely @xmath222 down to @xmath123 km ( see section [ edge ] ) . in this case
, there are at most @xmath364 bodies in the ob having radii of @xmath365 km and a total mass of @xmath366 m@xmath6 assuming a density of @xmath350 gm/@xmath351 and an albedo @xmath124 .
it should be noted that the model used here only accounts for the belt s dynamical erosion that is a consequence of neptune s gravitational perturbations
it does not account for the collisional erosion of the kuiper belt that is often invoked to account for the belt s depleted appearance ( e.g. , @xcite ) .
in particular , models of kbo accretion , as well as the self consistent nbody simulations of neptune s migration , all suggest that the kuiper belt s primordial mass was of order @xmath367 m@xmath6 ( @xcite ) , which is at least @xmath368 times more than the current mass . however the model used here , which only accounts for the dynamical erosion , results in a depletion by a factor of about 3 in the @xmath369 au zone of fig .
[ hot belt ] .
this suggests that collisional erosion , which is not modeled here , may have been responsible for reducing the belt s mass by an additional factor .
this is distinct from the push out model which need not rely on any collisional depletion of the kuiper belt @xcite .
] of @xmath269 . nonetheless , the abundance and mass estimates obtained here should still be reliable provided the belt s collisional erosion was relatively uniform across the observable @xmath370 au zone . if , however , collisional erosion was more vigorous in some parts of the belt , and less so in other parts , then the estimates obtained above will only be accurate in an order of magnitude sense .
a comparable problem also occurs with the model s inclinations .
section [ calibration ] shows that the simulated kuiper belt is too thin by a factor of @xmath371 .
this is compensated for by reducing the simulation s luminosity function @xmath303 by the factor @xmath372 , which is equivalent to increasing each particles inclination by this factor .
again , this crude treatment should still yield a reliable estimate of the kbo population provided the factor @xmath372 is uniform everywhere and independent of semimajor axis @xmath89 . if , however , @xmath372 is _ not _ independent of @xmath89 , then this will result in errors in the relative abundances of the belt s various subpopulations reported in table [ pop_table ] .
we also note that the relative abundances of the belt s various subpopulations are determined by a model that invokes a smooth outward migration by neptune by @xmath373 au , with the results reported in fig .
[ abundance ] .
that figure shows that the smooth migration scenario predicts a combined 3:2 + 2:1 population that is comparable to the main belt population .
this is because smooth migration is very efficient at trapping particles at neptune s resonances , and this results in densely populated resonances .
however a detailed comparison of the model predictions to the observed abundances indicates that the resonant kbo population is really only about @xmath374 of the main belt population ( see section [ census ] and table [ pop_table ] ) .
the seemingly low abundance of resonant kbos is likely due to unmodeled effects that may have occurred during the migration epoch , possibly due to the mutual scattering that might occur among bodies trapped at resonance , or perhaps due to the gravitational scattering of large planetesimals by neptune ( e.g. , @xcite ) .
there is also evidence indicating that a wide swath of the early kuiper belt was stirred up prior to the onset of neptune s migration .
recall that simulations of neptune s outward migration into a dynamically cold kuiper belt is unable to account for the eccentricities of @xmath72 observed among main belt kbos ( section [ cold belt ] ) .
this suggests that the belt was stirred up , either prior to or after the onset of neptune s migration
. however section [ hot belt ] shows that this stirring event likely occurred prior to migration : migration into a stirred up kuiper belt facilitates trapping at a multitude of weak , high order mean motion resonances which , as @xcite point out , is consistent with the detection of seven kbos now known to librate at neptune s 5:2 resonance . it is then natural to ask what mechanism might be responsible for stirring up a broad swath of the kuiper belt , particularly since accretion models tell us that kbos must have formed in a dynamically cold environment , _
i.e. _ , the particles initial @xmath12 s and @xmath57 s were @xmath49 ( @xcite ) . note that this disturbance was probably not due to gravitational stirring by a number of long gone protoplanets since , as @xcite point out , a sizable fraction of such bodies would still persist in the kuiper belt and would likely have been discovered by now .
note that this stirring mechanism must also have had a large reach since it must afflict kbos across the entire main belt , at least out to neptune s 2:1 resonance .
one mechanism that comes to mind is secular resonance sweeping , which is the only mechanism known to us that might stir eccentricities in the belt up to @xmath72 across its entire width @xcite .
however this @xmath12excitation is coherent in the sense that neighboring particles will have similar eccentricities .
so it is unclear whether secular resonance sweeping of the main belt , which would then followed by sweeping mean
motion resonances due to neptune s migration , will result in the range of eccentricities that is seen in fig .
[ hot belt ] .
secular resonance sweeping is a consequence of the dispersal of the solar nebula gas ; the removal of that gas alters the giant planets precession rates which in turn shifts the location of secular resonances @xcite .
the magnitude of the disturbance caused by a sweeping secular resonance depends on the timescale over which the nebula is depleted ; a longer depletion timescales @xmath375 results in larger eccentricity pumping .
the simulations of nebula dispersal by @xcite show that a disturbance of @xmath72 across much of the kuiper belt requires a nebula depletion timescale of @xmath376 years ( but see also @xcite ) .
one of the goals of this study is to determine how the adoption of a particular kuiper belt model might affect an assessment of the belt s total population and mass .
note that some models of the belt assume that the kbos are distributed according to a primordial surface density distribution that might vary with distance at @xmath377 or so ( e.g. , @xcite ) , while the kbos in another model are essentially equidistant @xcite .
however fig .
[ erosion ] shows that a realistic kuiper belt has been eroded from the inside out by the giant planets gravitational perturbations , which suggests that the earlier models might not apply .
however it turns out that an estimate of the total kbo population does _ not _ depend strongly upon a particular model s radial variation .
as section [ calibration ] shows , all three models yield population estimates that are within a factor of @xmath158 of each other .
this is because the observable kbos really do inhabit a relatively narrow belt centered on @xmath191 au having a radial half width that is only @xmath378 au ( see fig .
[ erosion ] ) , so the the assumption of equidistant kbos ( e.g. , @xcite ) appears to be good enough .
it should also be noted that the magnitude interval over which a model kuiper belt can be compared to the observed belt is given by the brightness of those kbos having reliable orbits , and this sample is presently dominated by bodies having a relatively limited magnitude range of only @xmath379 .
further testing of this model , as well as the development of alternative models of the belt , would be greatly facilitated if they could be compared to a larger sample of multi opposition kbos having reliable orbits and also exhibiting a broader range of apparent magnitudes and sizes .
this larger kbo sample would be very useful in many ways .
for example , it could be used to test the possibility that the various kuiper belt subpopulations do exhibit variations in their size distributions ( e.g. , section [ census ] ) .
this larger sample might also permit a better understanding of certain rare and unusual kbos , such as those that inhabit the extended scattered disk ( section [ extendedsd ] ) .
a deeper understanding of the kuiper belt , and what the belt tells us about the early evolution of the outer solar system , would be facilitated by deeper kbo surveys over larger portions of the sky in a systematic way that leads to efficient kbo recoveries and reliable orbit determinations .
* accretion models have shown that kuiper belt objects must have formed in a dynamically cold environment where the initial kbo seeds had nearly circular and coplanar orbits with eccentricities and inclinations @xmath380 @xcite .
simulations of neptune s outwards migration into a dynamically cold kuiper belt , described in section [ cold ] , show that the survivors in the main belt still maintain low eccentricities and inclinations .
however this conflicts with the main belt s observed @xmath12 s and @xmath48 s of @xmath292 .
this discrepancy suggests that some other process has also stirred up the kuiper belt .
this stirring event could have occurred prior to or after the onset of planet migration . *
the existence of several kbos librating at neptune s 5:2 resonance suggest that this stirring event occurred prior to the onset of planet migration .
simulations by @xcite have shown that if neptune migrates into a stirred up kuiper belt having eccentricities of @xmath72 , then trapping at neptune s higher order resonances , such as the 5:2 , becomes more efficient .
this result is confirmed by a higher resolution study of this phenomena described in section [ hot ] , which reveals that additional trapping also occurs at a number of exotic mean motion resonances like the 11:6 , 13:7 , 13:6 , 9:4 , 7:3 , 12:5 , 8:3 , 3:1 , 7:2 , and the 4:1 ; such resonances are not populated when neptune migrates into a dynamically cold disk .
not surprisingly , neptune s migration into a previously stirred up kuiper belt also accounts for the eccentricities of @xmath72 observed in the main belt .
* however the planet migration scenario investigated here does not account for the observed kbos having inclinations above @xmath100 ( section [ i ] ) , which is the main deficiency of this model .
this is a serious discrepancy since half of all kbos have inclinations @xmath381 according to the debiased inclination distribution reported by @xcite . *
neptune s migration into a stirred up kuiper belt traps particles in eccentric orbits at a number of resonances beyond @xmath206 au , the most prominent of these being the 5:2 and the 3:1 .
many of these distant particles that are trapped at semimajor axes @xmath9 au also have perihelia @xmath79 au , which is the domain conventionally known as the scattered disk . however section [ hot ] shows that only about 10% of the simulated particles that inhabit the
so called scattered disk or the extended scattered disk ( such as the gray zone in fig .
[ sd ] where @xmath195 au and @xmath80 ) are truly scattered particles .
the vast majority of these particles never had a close encounter with neptune ; rather , they were placed in these wide , eccentric orbits by neptune s sweeping mean motion resonances .
note that the origin of these bodies as being due to resonant trapping is very distinct from the scattering scenario originally suggested by @xcite .
* of the @xmath13 particles simulated here , only one managed to persist over the age of the solar system in the extended scattered disk , which is loosely defined as scattered orbits having perihelia @xmath268 au .
this particle s orbit is qualitatively similar to 2000 cr@xmath260 which has a perihelion of @xmath382 au .
however our simulations did not produce any extreme members of the extended scattered disk that are similar to 2003 vb@xmath261 ( sedna ) , which has a perihelion of @xmath383 au . * the output of the nbody model is coupled to a monte carlo model that assigns radii @xmath108 to the simulated particles according to a power law type cumulative size distribution that varies as @xmath166 .
magnitudes are computed for the simulation s particles , which then allows us to directly compare the simulated belt to the observed belt in a manner that accounts for telescopic selection effects .
section [ census ] compares the observed abundance of 2:1 objects to known main belt objects , and it is shown that the observed 2:1 population is underdense by a factor of 20 relative to model predictions . similarly , the observed 3:2 population is also depleted relative to model expectations .
another curious feature of the 3:2 is its lower than expected abundance ( relative to the main belt kbos ) of fainter kbos having magnitudes @xmath384 .
section [ census ] shows that this dearth of fainter kbos at the 3:2 can be interpreted as a dearth of small bodies , which implies that the 3:2 population has a @xmath385 size distribution that is substantially shallower than the canonical @xmath169 power law that holds for the larger members of the main belt .
* the simulated centaurs are quite sparse owing to their short dynamical lifetimes ; only seven centaurs were detected in the simulation during its final 2 gyrs .
interestingly , all seven originated at / near neptune s mean motion resonances in the kuiper belt ( section [ centaurs ] ) .
the model puts an upper limit of @xmath386 centaurs having radii larger than @xmath178 km , assuming they have a material density @xmath350 gm/@xmath351 and an albedo @xmath124 .
it should be noted that all of the simulated centaurs inhabit heliocentric distances of @xmath246 au , while the three centaurs reported in @xcite were detected at @xmath387 au .
if the simulated centaurs are representative of reality , then this discrepancy in their heliocentric distances may indicate that centaurs can break up and spawn additional centaurs ( e.g. , @xcite ) after evolving inwards from the kuiper belt . *
this model also estimates that there are at most @xmath388 trojans larger than @xmath178 km having a total mass of @xmath389 m@xmath6 orbiting at neptune s triangular lagrange points , assuming the usual @xmath350 gm/@xmath351 and @xmath124 . *
the absence of any distant kbos having low eccentricities at @xmath9 au places tight upper limits on the abundance of any kbos that might inhabit a hypothetical outer belt .
several upper limits are inferred from this null result : ( _ i. _ ) the primordial density of outer belt objects beyond 50 au is smaller than the primordial main belt density by a factor @xmath218 , ( _ ii .
_ ) these distant kbos are fainter than @xmath390 and thus have radii smaller than @xmath220 km , ( _ iii . _ ) the cumulative size distribution of outer belt objects is steep , having a power
law index of @xmath228 , or ( _ iv .
_ ) some combination of the above . * the luminosity function of the nbody / monte carlo model is fitted to the kbo s observed luminosity function , which then yields an estimate of the belt s total population of @xmath4 kbos larger than @xmath391 km having a total mass of @xmath392 m@xmath6 .
the population and mass of the belt s various subclasses ( e.g. , centaurs , neptune trojans , 3:2 and 2:1 populations , the main belt , and the scattered disk ) are also assessed in section [ calibration ] and listed in table [ pop_table ] . * acknowledgements * the authors thank matt holman for his comments on this paper , rodney gomes for discussions of the neptune evader mechanism , and hal levison for his review this paper .
the derivation of brouwer s @xmath86 integral appearing in appendix [ appendix_bi ] was suggested to us several years ago by bill ward .
support for this research was provided to jmh by a discovery grant from the natural sciences and engineering research council of canada ( nserc ) .
several of the simulations reported here were executed on the institute for computational astrophysics ( ica ) pluto computer cluster , which is funded by a grant from the canada foundation for innovation ( cfi ) .
additional simulations were also performed on the mckenzie computer cluster at the canadian institute for theoretical astrophysics ( cita ) at the university of toronto ; those machines are funded by the cfi as well as the ontario innovation trust ( oit ) .
an early generation of these simulations was also performed while jmh was in residence at the lunar and planetary institute ( lpi ) , and that portion of this research was supported by the national aeronautics and space administration ( nasa ) via an origins of solar systems grant nag5 - 10946 .
rm acknowledges research support from nasa s programs in planetary geology & geophysics and in origins of solar systems .
consider the orbital evolution of a particle trapped by a migrating planet s @xmath64 mean motion resonance .
the planet s disturbing function contains the the resonant term @xmath393 where @xmath394 is the planet / particle semimajor axis ratio for a particle having an eccentricity @xmath12 and a resonance angle having the form @xmath395 where @xmath396 and @xmath397 are the particle and planet s mean longitude while @xmath398 is the particle s longitude of periapse @xcite .
the relevant lagrange planetary equations are where @xmath400 is the particle s mean longitude at epoch , @xmath401 is the particle s mean motion , @xmath402 is time , and coplanar orbits are assumed . according to the averaging principal , the non resonant terms in the planet s disturbing function are of high frequency and average out during the particle s libration period , and the planet s disturbing function is simply @xmath403 .
the derivatives of @xmath108 are @xmath404 with these , the planetary equations ( [ planetary_eqns ] ) can be combined into a single differential equation , @xmath405 it will be convenient to replace @xmath12 with the variable @xmath406 ; since @xmath407 , the above equation can then be recast as @xmath408 where @xmath409 .
this differential equation is now easily integrated and yields @xmath410 where the integration constant @xmath86 can also be expressed as @xmath411 ( _ c.f . _ , @xcite ) .
the earliest derivation of this integral of the motion known to us is given in @xcite .
brouwer originally used this integral to consider the motion of a particle being perturbed by a planet in a static orbit .
however @xcite recognized that this integral is preserved even when the planet is migrating , which means that eqn .
( [ b ] ) can be used to predict the particle s eccentricity as its orbit is expanded by the planet s resonance : @xmath412 if , prior to capture , the particle is in a circular orbit with an initial semimajor axis @xmath413 , then @xmath414 ^ 2 $ ] and @xmath415 and the particle s eccentricity grows as the particle s orbit expands .
an approximate form of this expression is also derived in @xcite . thus if neptune s orbital expansion is indeed responsible for the kbos seen at neptune s 3:2 resonance at @xmath416 au having eccentricities as high as @xmath417 ( see fig .
[ cold belt ] ) , eqn .
( [ e(a)2 ] ) indicates they must first have been orbiting at @xmath418 au at the time of capture , and that neptune was initially orbiting at @xmath419 au and hence migrated a distance @xmath420 au .
the following will show that the ratio of the observed abundance of any two dynamical classes of kbos is approximately equal to the ratio of the absolute abundances of the much larger unseen populations .
begin by letting @xmath421 be the number of observed kbos that inhabit some dynamical class @xmath422 that are brighter than apparent magnitude @xmath129 , where @xmath422 might represent , say , the 3:2 population .
also assume that all kbos have the same heliocentric distance @xmath120 , which is a common assumption ( e.g. , @xcite ) that simplifies this analysis considerably , and is actually not a bad assumption since fig .
[ erosion ] shows that most of the observed multi
opposition kbos considered here do indeed inhabit a rather narrow belt , with more that half of these kbos found at heliocentric distances within 3 au of @xmath423 au .
we will also assume that the kuiper belt has azimuthal symmetry , which is justified further in appendix [ symmetry ] .
we shall also assume that all the kbo astronomers are observing this belt largely along the ecliptic , which also is not a bad assumption , since @xmath424 of the multi - opposition kbos studied here have latitudes of @xmath425 , which is the typical inclination of the belt s low@xmath57 component @xcite .
this indicates that the astronomers who discovered most of the kbos in this sample ( represented by the red dots in fig . [ hot belt ] ) were observing largely along the belt s midplane .
so , although these assumptions are not rigorously correct in detail , they are good enough to allow us to assess qualitatively the relative abundances of the various kbo populations .
let @xmath426 represent the sky
plane number density of class@xmath422 kbos ( _ i.e. _ , the cumulative luminosity function of class@xmath422 kbos ) , while @xmath427 be the _ differential _ luminosity function for class @xmath422
. then number of observed kbos brighter than magnitude @xmath129 is @xcite @xmath428 where @xmath429 is the total solid angle that has been surveyed by all kbo astronomers to a limiting magnitude @xmath129 , and @xmath430 is the efficiency at which a kbo of class @xmath422 and magnitude @xmath129 and is detected .
thus the number of observed class @xmath422 kbos of having magnitudes in the interval @xmath431 is @xmath432 if we were to compare class @xmath422 kbos to , say , class @xmath433 kbos , then the ratio of their observed abundances is @xmath434 where it is assumed that the kbo detection efficiency is insensitive to dynamical class , _
i.e. _ , @xmath435 . note that this ratio depends only on the kbos differential luminosity functions @xmath436 , and that it is insensitive to their discovery details such as the total solid angle @xmath429 that the kbo astronomers have surveyed to depth @xmath129 .
the cumulative luminosity function for class @xmath422 can be written as @xmath437 where @xmath438 is the kuiper belt s total solid angle , and @xmath439 is the cumulative magnitude distribution for class @xmath422 , _ ie , _ the total number of kbos in class @xmath422 that are brighter than magnitude @xmath129 .
thus @xmath440 since a body s magnitude @xmath129 is a function of its radius @xmath108 via eqn .
( [ mr ] ) , for which @xmath441 is the differential size distribution for class @xmath422 which has a total of @xmath442 bodies having radii in the interval @xmath443 ( see eqn .
[ n(r ) ] ) . and
since @xmath444 ( see eqn .
[ mr ] ) , the differential luminosity function can be written @xmath445 thus if any two kbo classes @xmath422 and @xmath433 have the same size distribution @xmath183 , then the ratio of their observed abundances is simply @xmath446 in other words , the ratio of the observed abundance of any two classes of kbos is approximately equal to the ratio of their intrinsic abundances , provided all bodies have the same size distribution @xmath183 .
but if populations @xmath422 and @xmath433 have different size distributions , then @xmath447 will vary with @xmath129 ( see section [ census ] ) . the monte carlo model of section [ census
] replicates each nbody survivor seen in fig .
[ hot belt ] @xmath13 times by randomizing the particles mean anomalies , which results in a model kuiper belt having azimuthal symmetry .
thus the model implicitly assumes that the kuiper belt has this same symmetry .
however one might question this assumption , since it is well known that a planet s resonant perturbations can rearrange the longitudes of a small body population .
for instance , particles at neptune s 3:2 tend to approach perihelion at longitudes that are @xmath448 away from the planet @xcite .
this is illustrated in the upper left corner of fig .
[ longitudes ] , which shows the ecliptic coordinates of all simulated particles that are in or very near neptune s 3:2 resonance .
this figure shows that neptune tends to arrange the 3:2 bodies preferentially away from the sun
neptune line , causing regions that lead / trail neptune by @xmath449 to be more densely populated , termed ` sweet spots ' in @xcite .
figure [ longitudes ] shows that resonant shepherding also occurs at neptune s 2:1 , again similar to that seen in @xcite .
red dots indicate the positions of the multi
opposition kbos considered here , which shows that kbos in or near the 3:2 resonance tend to be discovered at longitudes that are roughly @xmath449 away from neptune , as expected .
it is thus possible that neptune s rearrangement of the belt might skew our estimate of the resonant populations .
if , for example , kbo astronomers were preferentially surveying the belt along longitudes that are @xmath449 away from neptune , they would detect a rather high sky
plane number density of kbos in the 3:2 .
if we then assumed that this 3:2 number density were uniform about the entire ecliptic , we would overestimate the total 3:2 population .
similarly , if astronomers systematically observed the belt towards neptune or @xmath450 away , we would underestimate the 3:2 s average sky plane density and undercount its total population .
but if astronomers surveyed all longitudes with equal frequency , then these competing effects due to neptune pushing kbos away from certain longitudes and towards other longitudes should wash out , resulting in an estimate of the 3:2 population that is approximately reliable . the upper right part of fig .
[ longitudes ] shows the ecliptic coordinates of the observed multi opposition main belt kbos .
since the main belt population is likely azimuthally symmetric , these red dots should be a good indicator of where astronomers are looking for kbos .
it is quite clear from this figure that these astronomers lines of sight are not distributed uniformly about the ecliptic .
for instance , these kbo hunters tend to avoid the galactic plane , which passes through the ecliptic along the dashed line .
the mb figure also shows that astronomers have avoided a narrow portion of the 3:2 s ` sweet spot ' that leads neptune by about @xmath451 .
but it is also quite clear that these same astronomers are not systematically staring at the 3:2 s sweet spots , nor are they systematically avoiding them .
consequently , our simple visual inspection of this kbo sample indicates that our estimate of the 3:2 population will not be significantly baised towards under or overcounting the 3:2 population , and that the 3:2 abundance reported in section [ calibration ] is indeed representative .
a similar conclusion is also drawn for the much sparser 2:1 population . of course , this treatment succeeds only if the monte carlo method of section [ census ] does not alter the particles radial distribution as they have their mean anomalies randomized by the replication process .
however it is straighforward to show that this is indeed the case .
for instance , randomizing the mean anomalies of the 3:2 population seen fig .
[ longitudes ] does not change their radial distribution in any significant way . , .
, tabachnik , s. , rafikov , r. , lupton , r. h. , quinn , t. , hammergren , m. , eyer , l. , chu , j. , armstrong , j. c. , fan , x. , finlator , k. , geballe , t. r. , gunn , j. e. , hennessy , g. s. , knapp , g. r. , leggett , s. k. , munn , j. a. , pier , j. r. , rockosi , c. m. , schneider , d. p. , strauss , m. a. , yanny , b. , brinkmann , j. , csabai , i. , hindsley , r. b. , kent , s. , lamb , d. q. , margon , b. , mckay , t. a. , smith , j. a. , waddel , p. , york , d. g. , & the sdss collaboration .
2001 , , 122 , 2749 centaurs & @xmath452 & @xmath453 & @xmath454 + trojans & @xmath455 & @xmath456 & @xmath457 + 3:2 & @xmath458 & @xmath459 & @xmath460 + main belt & @xmath106 & @xmath461 & 0.059 + 2:1 & @xmath462 & @xmath463 & @xmath464 + scattered disk & @xmath465 & @xmath466 & 0.011 | nbody simulations are used to examine the consequences of neptune s outward migration into the kuiper belt , with the simulated endstates being compared rigorously and quantitatively to the observations .
these simulations confirm the findings of @xcite , who showed that neptune s migration into a previously stirred up kuiper belt can account for the kuiper belt objects ( kbos ) known to librate at neptune s 5:2 resonance .
we also find that capture is possible at many other weak , high order mean motion resonances , such as the 11:6 , 13:7 , 13:6 , 9:4 , 7:3 , 12:5 , 8:3 , 3:1 , 7:2 , and the 4:1 .
the more distant of these resonances , such as the 9:4 , 7:3 , 5:2 , and the 3:1 , can also capture particles in stable , eccentric orbits beyond 50 au , in the region of phase space conventionally known as the scattered disk .
indeed , @xmath0 of the simulated particles that persist over the age of the solar system in the so called scattered disk zone never had a close encounter with neptune , but instead were promoted into these eccentric orbits by neptune s resonances during the migration epoch .
this indicates that the observed scattered disk might not be so scattered .
this model also produced only a handful of centaurs , all of which originated at neptune s mean motion resonances in the kuiper belt . however a noteworthy deficiency of the migration model considered here is that it does not account for the observed abundance of main belt kbos having inclinations higher than @xmath1 .
in order to rigorously compare the model endstate with the observed kuiper belt in a manner that accounts for telescopic selection effects , monte carlo methods are used to assign sizes and magnitudes to the simulated particles that survive over the age of the solar system .
if the model considered here is indeed representative of the outer solar system s early history , then the following conclusions are obtained : ( _ i. _ ) the observed 3:2 and 2:1 resonant populations are both depleted by a factor of @xmath2 relative to model expectations ; this depletion is likely due to unmodeled effects , possibly perturbations by other large planetesimals , ( _ ii .
_ ) the size distribution of those kbos inhabiting the 3:2 resonance is significantly shallower than the main belt s size distribution , ( _ iii . _ ) the total number of kbos having radii @xmath3 km and orbiting interior to neptune s 2:1 resonance is @xmath4 ; these bodies have a total mass of @xmath5 m@xmath6 assuming they have a material density @xmath7 and an albedo @xmath8 .
we also report estimates of the abundances and masses of the belt s various subpopulations ( e.g. , the resonant kbos , the main belt , and the so called scattered disk ) , and also provide upper limits on the abundance of centaurs and neptune s trojans , as well as upper limits on the sizes and abundances of hypothetical kbos that might inhabit the @xmath9 au zone . |
mounting evidence has lent strong support for the origin of long - duration gamma - ray bursts ( grbs ) in the death of massive stars ( see woosley & bloom 2006 for a review ; but also see fynbo et al.2006a , gal - yam et al .
2006 , & della valle et al .
2006 for recent discoveries ) .
specifically , a direct association between grbs and type ib / c supernovae ( sne ) is established for grbs 980425 , 030329 , and 031203 ( patat et al .
2001 ; hjorth et al . 2003 ; stanek et al . 2003 ; malesani et al .
these sne of type ic show no hydrogen or strong siii transitions in their spectra , and have never been seen in elliptical galaxies ( filippenko 1997 ) . while it is conceivable that all long - duration grbs occur contemporaneously with a type ic sn
, their low event rate indicates that only a small fraction ( @xmath4% ) of type ic sne are accompanied by a grb phase ( van putten 2004 ; podsiadlowski et al .
2004 ; soderberg et al .
2006 ) . for stars to lose their hydrogen envelope prior to the core collapse and form a grb , current theoretical models have homed in on progenitors with initial stellar mass of @xmath5 , fast rotation speed , and low metallicity @xmath6 ( woosley & heger 2007 ) .
these grbs are therefore expected to originate preferentially in metal - poor environment and/or underluminous galaxies .
observations of five nearby ( @xmath7 ) host galaxies indeed show that the host galaxies tend to be underluminous and that two of these galaxies also have an integrated hii - region abundance below what is expected from a nominal luminosity - metallicity relation of field galaxies ( e.g. prochaska et al .
2004 ; modjaz et al .
2006 ) . at the same time , chemical abundance studies of high - redshift ( @xmath8 ) hosts have been carried out based on absorption lines identified in early - time afterglow spectra ( e.g. fynbo et al . 2006b ; savaglio 2006 ; prochaska et al .
2007 ) , where the emission lines of an hii region become too faint to be detected .
abundance measurements of the cold interstellar medium ( ism ) based on absorption - line studies , despite representing only a small volume along a pencil beam from the afterglow , are less subject to uncertainties in the ionization fraction or temperature of the gas , and therefore offer a more accurate assessment of the metallicity .
comprehensive studies of afterglow absorption - line spectra have also allowed us to resolve gas at @xmath9 pc from the progenitor stars , providing direct constraints on the ism properties of the progenitor environment ( e.g. prochaska , chen , & bloom 2006 ; chen et al . 2007 ; vreeswijk et al .
2007 ) . however , our understanding of the progenitor environment is far from clear .
for example , roughly 50% of grb afterglows are located behind a large column of neutral gas with @xmath10 ( jakobsson et al . 2006 ) but no trace of molecular gas to very low limits @xmath11{\lesssim}10^{-6}$ ] ( tumlinson et al .
2007 ) . using a sample of 16 afterglow spectra ,
prochaska et al . (
2007 ) showed that grb hosts have sub - solar chemical abundances and an @xmath1-element enhancement relative to fe in the ism , but also that a non - negligible fraction exhibit @xmath12% solar abundances .
in addition , these authors also showed that even with a large neutral gas column , there can be little dust extinction along the sightlines in the host ism toward the afterglows ( see also starling et al .
the lack of h@xmath13 can , in principle , be explained by a high destruction rate due to the enhanced photo - ionization radiation field in the star - forming cloud that hosts the grb progenitor and by a low formation rate inferred from the observed low metallicity .
but at high enough gas density ( @xmath14 @xmath15 ) , self - shielding is expected to be effective and will promote h@xmath13 survival ( e.g. tumlinson et al .
2007 ) .
super star clusters provide an interesting local analogue for understanding the formation of grbs for two main reasons .
first , wolf - rayet stars , a popular candidate for grb progenitors ( e.g.woosely & bloom 2006 ) , are often found in young star clusters ( see crowther 2006 for references ) .
second , star clusters contain a large number of massive stars in a small volume and are expected to have on average a higher rate of stellar collision and mass transfer between individual stars for nurturing the formation of a grb .
various authors have proposed that dense stellar environments facilitate stellar coalescence for forming blackholes that power the grbs ( e.g.hansen & murali 1998 ; efremov 2001 ) .
in addition , grbs have been considered as a possible source for powering giant hi supperbubble shells found in nearby galaxies ( efremov , elmegreen , & hodge 1998 ; loeb & perna 1998 ) a direct comparison between absorption - line properties in the ism surrounding super star clusters and those identified in grb host galaxies may therefore provide important calibrations for interpreting afterglow absorption - line observations . to gain further insights for interpreting the afterglow absorption - line spectra of high - redshift grbs ,
we have searched for ultraviolet echelle spectra of nearby super star clusters in the hubble space telescope ( hst ) data archive .
ngc1705 - 1 , a super star cluster sitting at the center of a blue compact dwarf galaxy ngc1705 at redshift @xmath16 , is the only star cluster with high quality echelle spectra from the space telescope imaging spectrograph ( stis ) .
the galaxy is a well known starburst galaxy that exhibits evidence of galactic scale gas outflows .
it has a total luminosity of @xmath17 and an hii region abundance of @xmath18 35% solar ( e.g. lee & skillman 2004 ) .
a comparison between the on - going star formation rate and @xmath19-band luminosity indicates a formation time scale that is much shorter than normal irregular galaxies and comparable to what is known for grb host galaxies at high redshift ( e.g. christensen , hjorth , & gorosabel 2004 ) .
the spectra of the super star cluster were originally taken for studying the physical conditions of the galactic winds driven by starburst outflows ( see e.g. vzquez et al .
here we re - visit this data set and present new column density measurements of various low - ionization species . we compare the ism properties of the dense star cluster with those of high - redshift grb host galaxies . this paper is organized as the following . in section 2 , we review known ism properties integrated over the entire galaxy ngc1705 and ism properties that are local to the super star cluster ngc1705 - 1 . in section 3 , we present our absorption - line measurements of the cold gas along the line of sight toward ngc1705 - 1 . in section 4 ,
we compare the ism properties of ngc1705 - 1 with measurements for high - redshift grb hosts .
we describe the value of collecting echelle spectra of nearby supper star clusters using future space uv facilities such as the cosmic origin spectrograph , for understanding the grb progenitor environment at high redshift .
ngc1705 is a well - known starburst galaxy at @xmath20 that has been studied extensively over a broad spectral window , from ultraviolet through radio frequencies ( e.g. lamb et al .
1985 ; calzetti et al . 1994 ; meurer et al . 1998 ; lee & skillman 2004 ) .
this blue irregular galaxy hosts a super star cluster , ngc1705 - 1 ( figure 1 ) , that is among the brightest star clusters in the local universe and contributes roughly 40% of the total uv light of the galaxy ( meurer et al .
spatially resolved strong emission lines indicate that the warm ism in the hii region can be charaterized with an electron temperature of @xmath21 k ( lee & skillman 2004 ) .
one of the most notable features in the ism of ngc1705 is the multiple super bubble shells , indicating gas outflow driven by the last episode of starburst roughly @xmath22 years ago that also formed most of the stars ( @xmath23 ) seen today in ngc1705 - 1 ( meurer et al .
the galaxy also exhibits very little intrinsic reddening ( calzetti et al .
1994 ) and shows no sign of molecular gas to very sensitive levels ( greve et al . 1996
; hoopes et al . 2004 ) .
moderate - to - high resolution uv spectra of ngc1705 - 1 from different space facilities are published in the literature . while most of the absorption lines identified in the spectra are of ism origin ( york et al .
1990 ) , detailed stellar population analysis based on a few stellar photospheric lines have placed strong constraints on the age , metallicity , and imf of the cluster ( vzquez et al .
the ism absorption lines exhibit two strong components with respective blue - shifted velocity of @xmath24 and @xmath25 ( heckman et al .
2001 ; vzquez et al .
the low velocity component is believed to trace the cold , neutral ism , while the high - velocity component is interpreted to originate in a gas outflow .
abundance measurements based on low ions such as n@xmath26 , s@xmath3 and fe@xmath3 support the scenario that the cold neutral ism has been enriched primarily by type ii sne that occurred in the star cluster roughly @xmath22 yr ago ( sahu & blades 1997 ; sahu 1998 ) .
in addition , the observed kinematic profile and the absorption strength of the ovi doublet are consistent with the expectations of superwinds undergoing radiative cooling ( heckman et al .
2001 ) .
a summary of known properties of galaxy ngc1705 and the super star cluster ngc1705 - 1 is presented in table 1 .
we note that ism absorption - line measurements in the table are based primarily on previous fuse observations ( e.g. heckman et al .
2001 ; hoopes et al .
we have retrieved all available stis echelle spectra of ngc1705 - 1 in the hst data archive .
the observations were carried out using a @xmath27 slit and three different grating set - ups for a total exposure time of 28,569 s ( pid=8297 ) .
individual echelle spectra were reduced , extracted , and calibrated using standard pipeline techniques . to form a final spectrum for absorption line studies , we calculated a weighted average of individual spectra and the corresponding 1-@xmath28 error array per grating set - up with the weighting factor determined by the inverse variance .
we normalized the mean spectrum with a best - fit , low - order polynomial continuum .
the combined and normalized spectra have a spectral resolution of @xmath29 , covering wavelength ranges of @xmath30 with the e140 m grating at signal - to - noise ratio @xmath31 and @xmath32 with the e230 m grating at @xmath33 per resolution element .
absorption lines of heavy ions identified in the ism of ngc1705 exhibit complex absorption profiles . in figure 2
, we present in the three left panels low ionization species , and in the right column high - ionization species .
the zero relative velocity ( @xmath34 ) corresponds to the systematic redshift of ngc1705 @xmath20 , measured in hi 21 cm ( meurer et al .
1992 ) and in stellar photospheric transitions ( vzquez et al .
the spectra include absorption features from the milky way and a foreground high - velocity cloud ( sahu & blades 1997 ) which occasionally blend with features from ngc1705 .
these ` contaminating ' features have been identified and are indicated with dotted lines in the panels . in addition , stellar photospheric features are apparent in transitions such as civ and siiv , as a broad , shallow component extending from @xmath35 through @xmath36 . the low- and high - ionization species in the ism of ngc1705 exhibit distinct kinematic signatures .
the low - ion resonance lines , ( which excludes excited c@xmath3 transition , cii * 1335 ) , show two narrow components at @xmath37 and @xmath38 , while the high - ionization elements are broad and show velocity centroids systematically blue - shifted to @xmath39 .
these high ions share the same kinematic feature as the ovi absorption doublet discussed in heckman et al .
( 2001 ) .
we measure the column density of each ion using the apparent optical depth method ( savage & sembach 1991 ) over a velocity interval of @xmath40 to @xmath41 for low - ionization transitions and @xmath42 to @xmath43 for high - ionization transitions .
one exception is for the sii 1259 transition , where the line is contaminated by the siii 1260 transition from the foreground high - velocity cloud .
we determine @xmath44 using the vpfit software package , including multiple components .
the measurements are summarized in table 2 .
the hi transition of ngc1705 is present in the stis spectrum .
it is saturated and blended with absorption from the milky way ( figure 3 ) .
we note , however , that the @xmath45 value of ngc1705 is well constrained by the red damping wing .
we have performed a voigt profile analysis and found @xmath46 for ngc1705 and @xmath47 for the milky way .
this is consistent with the finding of sahu ( 1998 ) and heckman et al .
( 2001 ) based on hst ghrs and fuse observations , respectively .
the best - fit model profile combining absorption from the milky way and ngc1705 is presented as the red curve in figure 3 .
the total hi gas column density observed along the sightline toward ngc1705 - 1 is only about 10% of what is observed in hi 21 cm emission ( meurer et al .
1992 ) , indicating that the super cluster is located near the front edge of the galaxy and the sightline only probes a fraction of the ( presumably outer ) gaseous disk .
based on the column density measurements presented in table 2 , we derive an iron abundance of @xmath48=-1.1 \pm 0.2 $ ] and @xmath49=-0.6 \pm 0.1 $ ] with respect to the solar values from asplund , grevesse , & sauval ( 2005 ) .
the sulfer abundance in the cold neutral medium is consistent with the oxygen abundance determined for the hii region , @xmath50=-0.5\pm 0.05 $ ] ( table 1)$ ] determined by heckman et al .
( 2001 ) based on oi absorption lines observed in fuse spectra .
given the observed column density , however , we expect the line to be saturated .
therefore , the reported value should be treated as a lower limit . ] .
the large ratio of @xmath51=+0.5 \pm 0.2 $ ] indicates a an @xmath1-element enhancement , which is qualitatively consistent with a chemical enrichment history dominated by type ii supernovae and little / no reddening observed for the ism of ngc1705 ( calzetti et al.1994 ) . although dust depletion can be substantial even when reddening is negligible ( e.g. savaglio 2006 ; prochaska et al .
2007 ) , the significant underabundance of nitrogen , @xmath52=-2.1 \pm 0.1 $ ] , from heckman et al . ( 2001 ) indicates that more evolved , intermediate - mass stars have not formed to produce abundant nitrogen and is consistent with a chemical enrichment history dominated by type ii supernovae ( e.g. henry et al .
2000 ) .
in contrast with the grb host environment observed at high redshift , we do not observe excited ions ( e.g. strong cii * 1335 transition ) for the neutral gas in the ism toward ngc1705 - 1 .
instead , the observed cii * 1335 transition is blueshifted by @xmath53 from the neutral species ( see discussion in the following section ) .
the lack of excited c@xmath3 in the neutral gas implies a warm neutral medium and/or low dust content ( wolfe , prochaska , & gawiser 2003 ) . as noted earlier , high - ionization transitions such as civ 1548 , 1550 , siiii 1206 , siiv 1393 , 1402 , and alii 1854 , 1862
are observed to be blue - shifted at @xmath25 in the stis echelle spectra ( figure 2 ) .
these coincide in velocity space with the ovi absorption doublet reported by heckman et al .
in particular , we note the presence of strong cii * at the same velocity offset , where little feii absorption has been detected .
this is similar to what is observed in the ism toward hd192639 ( sonnentrucker et al .
2002 ) , whose origin is interpreted to be in shocks produced by expanding supperbubble shells or stellar winds .
the broad line widths seen in unsaturated lines such as cii * 1335 and aliii 1854 are suggestive of a collisional ionization origin .
adopting the observed siii and siv transitions that share the same kinematic features with the ovi absorber ( heckman et al .
2001 ) , we find @xmath54 and constrain the gas temperature at @xmath55 k under a collisional ionization equilibrium . further constraints for the gas temperature can be obtained by comparing the line width of cii * 1335 and aliii 1854 .
we measure a doppler parameter of @xmath56 for c@xmath3 and @xmath57 for al@xmath58 from the stis spectra . assuming that these ions originate at the same location , we further infer a gas temperature of @xmath59 k and a turbulent velocity of @xmath60 . under a collisional excitation scenario
, we can also constrain the density of ionized gas based on the relative strengths between resonance absorption and absorption from excited states .
given that the cii 1334 transition is fully saturated , we obtain the contraint using the relative abundance between excited si@xmath3 and ground - state si@xmath3 .
the high - velocity component at @xmath61 is moderately resolved from the warm and cold gas components .
we perform a multiple - component voigt profie analysis and measure @xmath62 and @xmath63 for the high - velocity component . the lack of siii * yields @xmath64 , leading to an electron density @xmath65 @xmath15 for @xmath66 k.
recent observations of grb host environments have yielded two main results . first , the ism of long - duration grbs appear to be essentially devoid of molecular gas , despite exhibiting strong neutral - hydrogen absorption columns .
second , these grbs occur preferentially in metal - poor , low - metallicity , low - luminosity , actively star forming galaxies .
together these suggest that global properties of the hosts are particularly well - suited for the creation of the massive star progenitors thought to give rise to grbs .
if true , the study of local analogues of the sorts of environments amenable to grb progenitor birth should help in understanding the nature of distant grb birthsites .
ngc1705 is a well known starburst galaxy in the nearby universe , with special interest because of the clear presence of substantial o@xmath67 in a starburst - driven outflow .
we have re - visited and presented new measurements of the ism properties along the line of sight toward the super star cluster , ngc1705 - 1 , using archived by unanalyzed stis echelle spectra .
the study is motivated by the hypothesis that massive young star clusters offer a birthsite for the progenitor stars of high - redshift grbs , because of the dense stellar environment ( e.g. hansen & murali 1998 ; efremov 2001 ) .
the available stis uv echelle spectra of the cluster show numerous ionic transitions with a complex velocity structure , and allow us to use the same set of metal absorption lines for a direct comparison between known properties of the ism near the progenitors of high - redshift grbs and the ism properties around ngc1705 - 1 . here
we discuss the insights we gain from the uv echelle spectra of nearby super star clusters for interpreting the afterglow absorption - line studies of distant grb hosts .
the complex velocity structure displayed in the absorption - line profiles of different ionization species indicate a multi - phase ism near the super star cluster ngc1705 - 1 .
based on fuse observations of the starburst , heckman et al .
( 2001 ) presented a comprehensive study of the thermal state of the ism .
these authors identified ( 1 ) a neutral medium as demonstrated by hi , oi , siii , and feii absorption , ( 2 ) a warm photoionized phase as demonstrated by the h@xmath1 emission nebula ( meurer et al .
1992 ; marlowe et al .
1995 ) and nii and siii absorption , and ( 3 ) coronal gas as demonstrated by the ovi absorption doublet .
these results are well supported by our analysis of the stis echelle observations .
given the small velocity offset of the neutral gas ( @xmath68 ) with respect to the nebular lines , heckman et al .
propose that the neutral gaseous clouds represent neutral ism that has been engulfed by the expanding superbubble . under this scenario ,
the distance of the neutral gas must be @xmath69 pc , the observed size of the superbubble , and the gas density is required to be @xmath70 for effective self - shieding from the ionizing radiation of the star cluster .
together these constrain the cloud to be @xmath71 pc for producing the observed hi column density .
given the large observed column density of o@xmath72 at @xmath73 , heckman et al
. also concluded that the coronal gas can not originate from behind the shocks induced by the expanding superbubble , but rather blowout gas ( super winds ) that are undergoing radiative cooling between photo - ionized fragments of the shells . in this picture , the ism around ngc1705 - 1 has been heated by both the expansion of super bubbles and the uv radiation of the central star cluster .
the stis echelle observations allow us to constrain the gas temperature at @xmath74 k and a bulk motion of of @xmath75 .
the primary objectives of the present study are ( 1 ) to explore an alternative scenario in which young super star clusters provide a likely birthsite for long - duration grbs , and ( 2 ) to examine whether available uv echelle spectra of a nearby super star clusters offer additional insights for interpreting afterglow absorption - line studies of high - redshift bursts . our current understanding of the host ism around high - redshift bursts can be summarized as the following . first , prochaska , chen , & bloom ( 2006 ) argued that the large neutral gas column observed along an afterglow sightline is located at @xmath76 pc from the grb progenitor star , in order for mg@xmath26 to survive the intense uv radiation field .
the associated heavy ions such as c@xmath3 , si@xmath3 , and fe@xmath3 in the neutral gas clouds would be subsequently excited by the afterglow uv photons for producing the observed absorption features from their excited states ( confirmed by vreeswijk et al.2007 for grb060418 ) .
the neutral gas clouds in the ism model of ngc1705 - 1 from heckman et al .
( 2001 ) offers a natural candidate for the presence of dense neutral clouds in the vicinity of a grb progenitor at high redshift that was centered on ngc1705 - 1 , in contrast to the @xmath77 aperture available for the fuse observations .
the agreement between our respective @xmath45 and @xmath78 measurements , therefore , confirms that the neutral gas close to the super star cluster dominates the absorption column observed in fuse . ] .
note that the observed @xmath45 along the sightline toward ngc1705 - 1 is more similar to strong lyman limit systems of @xmath79 observed in the hosts of grbs021004 , 050908 , and 060526 ( jakobsson et al .
higher column density absorbers would imply either a smaller cloud size or higher neutral gas density .
nonetheless , should a grb occur in ngc1705 - 1 , the neutral gas at @xmath80 pc would still remain neutral under the afterglow radiation field , but would be excited to produce the commonly observed transitions from excited ions .
in addition , the ism of ngc1705 exhibits little / no trace of molecular gas , neither co nor h@xmath13 , despite the recent starburst episode @xmath81 myr ago and on - going star formation @xmath82 yr@xmath83 .
ngc1705 resembles the small magellanic cloud both in size and in the total metal content , but has on average a substantially smaller molecular fraction : @xmath84 for ngc1705 - 1 ( table 1 ) versus @xmath85 near star - forming regions in smc ( tumlinson et al.2002 ) .
the super star cluster ngc1705 - 1 near the center of ngc1705 is among the brightest star clusters in the local universe ( meurer et al .
1995 ) , containing massive stars of @xmath86 and @xmath71 myr old .
the lack of molecules in the ism around ngc1705 - 1 is reminiscent of what is observed in the high column density clouds surrounding grb progenitors ( hatsukade et al . 2007 ; tumlinson et al .
the low molecular fraction and dust extinction for moderately enriched ism and high gas density appear to be remarkably similar , implying that the lack of molecules and dust may be due to the enhanced uv radiation from massive stars in the star - forming region that hosts the grb progenitor ( e.g. hoopes et al .
2004 ; tumlinson et al . 2007 ) .
we further note that similar to what is observed in the afterglow spectra of high - redshift grbs , low - ionization species in ngc1705 - 1 are more confined in a narrow veolicty interval whereas high - ionization transitions exhibit larger velocity spread .
the apparent offset in the velocity profiles of low - ionization ( such as feii ) and high - ionization transitions ( such as civ and siiv ) observed toward ngc1705 - 1 ( figure 2 ) is now understood as due to starburst driven outflows , based on extensive imaging and spectroscopy studies .
it suggests that important clues for possible starburst nature in grb host galaxies may already be learned based on afterlow spectra alone . the is particularly valuable because the host galaxies of distant grbs are faint and challenging to be found in emission .
of the six grb host studied in chen et al .
( 2007 ) , we identify velocity offsets between siii and civ / siiv over the range from @xmath87 to @xmath88 , along the sightlines toward grbs021004 , 050730 , 050908 , and 060418 . finally , we present in table 3 a direct comparison of different ism properties between ngc1705 - 1 and grb hosts .
the table shows that the two sources occupy a comparable parameter space .
in particular , absorption - line studies of grb host galaxies show that the host ism of high - redshift grbs exhibits clear @xmath1-element enhancement with little dust extinction / depletion along the sightlines toward the afterglows .
the similarity in the ism properties shows that this scenario for long - duration grbs to originate in massive star clusters is at least viable .
the relatively low nitrogen abundance also signifies the young age of the progenitor star clusters , because nitrogen is produced primarily in intermediate - mass agb stars ( e.g. prochaska et al .
ngc1705 - 1 is unique among known super star clusters locally , because of its young age and massive stellar content .
conclusive evidence requires a large uv spectroscopic sample of nearby super star clusters .
these data are necessary for a complete statistical analysis of their ism properties to be compared with afterglow absorption - line studies . for example , a comparison of the @xmath45 distribution and the size distribution of neutral clouds in the vicinity of young star clusters may offer important support , or otherwise , for the star cluster origin of long - duration grbs . we expect that a direct comparison with known statistical properties of high - redshift grb hosts will allow us to identify the key physical parameters that facilitate the formation of grbs .
it is worthwhile to note also the direct evidence of inflow gas around ngc1705 , as evidenced by the redshifted components seen in civ and siiv
. a larger sample may be useful for studying gas acretion around galaxies .
future spectroscopic observations of nearby super star clusters using the cosmic origin spectrograph to be installed on the hst are necessary .
we thank j. gallagher for a coffee discussion that sparked this analysis and d. welty for his valuable input on hd192639 .
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 .
jxp , and jsb acknowledge support from nasa grant nng05gf55 g .
w.c . acknowledges partial support from an nsf grant ast-0607510 . )
corresponds to @xmath20 .
contaminating features are dotted out .
stellar photospheric features are present in transitions such as civ and siiv , as a broad , shallow component extending from @xmath35 through @xmath36 . while strong low - ionization transitions such as cii 1334 and siii 1260 are saturated , weaker transitions such as ni , sii 1259 , and feii 1608 clearly exhibit two narrow components in the neutral ism at @xmath89 and @xmath90 .
high ionization transitions such as aliii , civ , and siiv exhibit a dominant component that is systematically shifted to @xmath39 , coincident with the kinematic signatures of the ovi absorption doublet discussed in heckman et al .
we also note a weak contribution from highly ionized gas at @xmath91 , presumably from photospheric absorption . ]
a voigt profile analysis that includes two hi components at @xmath20 ( the position of the vertical dashed line ) and @xmath92 ( the emission feature ) yield @xmath93 for ngc1705 and @xmath47 for the milky way .
the best - fit model is represented by the red curve with the yellow band marking the 1-@xmath28 uncertainties in the ngc 1705 profile .
the error is dominated by systematic uncertainty due to continuum placement which is described by the nearly horizontal dashed line .
although the absorption feature is blended with the hi absorption from the milky way , the total neutral hydrogen absorption column density @xmath45 is well constrained by the red damping wind .
our @xmath45 measurements agree very well with previous measurements by sahu ( 1998 ) and heckman et al .
( 2001 ) based on hst ghrs and fuse observations , respectively .
we note that the absorption features at @xmath94 are due to siiii 1206 transitions at @xmath95 . ]
p3incl distance ( mpc ) & @xmath96 & tosi et al .
( 2001 ) @xmath97 & @xmath98 & marlowe et al .
( 1999 ) star formation rate ( @xmath99 yr@xmath83 ) & 0.06 & lee & skillman ( 2004 ) formation time scale ( @xmath100 gyr ) & @xmath101 & lee & skillman ( 2004 ) stellar mass ( @xmath102 ) & @xmath103 & smith & gallagher ( 2001 ) effective radius ( pc ) & @xmath104 & smith & gallagher ( 2001 ) age ( myr ) & @xmath105 & vzquez et al .
( 2004 ) hii region @xmath106 & @xmath107 & lee & skillman ( 2004 ) hii region @xmath108 & @xmath109 & lee & skillman ( 2004 ) @xmath110 ( @xmath111 k ) & 1.12 & lee & skillman ( 2004 ) @xmath112 & @xmath113 mag & calzetti et al .
( 1994 ) @xmath114 ( ) & @xmath115 & vzquez et al .
( 2004 ) @xmath116 ( @xmath117 ) & @xmath118 & hoopes et al .
( 2004 ) @xmath45 ( @xmath117 ) ( 21 cm ) & @xmath119 & meurer et al .
( 1998 ) @xmath120 ( absorption ) & @xmath121 & heckman et al .
( 2001 ) @xmath122 & @xmath123 & heckman et al .
( 2001 ) @xmath124 & @xmath125 & heckman et al .
( 2001 ) @xmath126 & @xmath127 & heckman et al .
( 2001 ) @xmath128 & @xmath129 & heckman et al .
( 2001 ) @xmath114 ( ) & @xmath130 & vzquez et al .
( 2004 ) @xmath131 & @xmath132 & heckman et al .
( 2001 ) @xmath133 & @xmath134 & heckman et al .
( 2001 ) lccccccccc + & 0 & 0.00 & 1277.2450 & @xmath135&@xmath136$]&@xmath137&@xmath138&@xmath139 + & 0 & 0.00 & 1560.3092 & @xmath140&@xmath141$]&@xmath142&@xmath143 & + & 0 & 0.00 & 1656.9283 & @xmath144&@xmath141$]&@xmath145&@xmath146 & + + & 1/2 & 0.00 & 1334.5323 & @xmath147&@xmath141$]&@xmath148&@xmath149&@xmath150 + & 3/2 & 63.40 & 1335.7077 & @xmath151&@xmath141$]&@xmath152&@xmath153&@xmath154 + + & & 0.00 & 1548.1950 & @xmath155&@xmath156$]&@xmath157&@xmath158&@xmath159 + & & 0.00 & 1550.7700 & @xmath160&@xmath156$]&@xmath161&@xmath162 & + + & 2 & 0.00 & 1302.1685 & @xmath163&@xmath164$]&@xmath165&@xmath166&@xmath167 + & 1 & 158.26 & 1304.8576 & @xmath168&@xmath141$]&@xmath169&@xmath170&@xmath170 + + & & 0.00 & 2852.9642 & @xmath171&@xmath141$]&@xmath172&@xmath173&@xmath174 + + & & 0.00 & 2796.3520 & @xmath175&@xmath176$]&@xmath177&@xmath178&@xmath179 + & & 0.00 & 2803.5310 & @xmath180&@xmath141$]&@xmath181&@xmath182 & + + & & 0.00 & 1670.7874 & @xmath183&@xmath141$]&@xmath184&@xmath185&@xmath185 + + & & 0.00 & 1854.7164 & @xmath186&@xmath141$]&@xmath187&@xmath188&@xmath189 + & & 0.00 & 1862.7895 & @xmath190&@xmath141$]&@xmath191&@xmath192 & + + & 1/2 & 0.00 & 1260.4221 & @xmath193&@xmath176$]&@xmath194&@xmath195&@xmath196 + & 1/2 & 0.00 & 1304.3702 & @xmath197&@xmath141$]&@xmath198&@xmath199 & + & 1/2 & 0.00 & 1526.7066 & @xmath200&@xmath141$]&@xmath201&@xmath202 & + & 3/2 & 287.24 & 1264.7377 & @xmath203&@xmath141$]&@xmath204&@xmath205&@xmath206 + + & & 0.00 & 1259.519 & @xmath207 & ... & ... & @xmath208 & @xmath209 + + & & 0.00 & 2056.2539 & @xmath210&@xmath141$]&@xmath211&@xmath212&@xmath213 + + & 9/2 & 0.00 & 1608.4511 & @xmath214&@xmath141$]&@xmath215&@xmath216&@xmath217 + & 9/2 & 0.00 & 1611.2005 & @xmath218&@xmath141$]&@xmath219&@xmath220 & + & 9/2 & 0.00 & 2260.7805 & @xmath221&@xmath141$]&@xmath222&@xmath223 & + & 9/2 & 0.00 & 2344.2140 & @xmath224&@xmath141$]&@xmath225&@xmath226 & + & 9/2 & 0.00 & 2374.4612 & @xmath227&@xmath141$]&@xmath228&@xmath229 & + & 9/2 & 0.00 & 2382.7650 & @xmath230&@xmath141$]&@xmath231&@xmath232 & + + & & 0.00 & 1370.1310 & @xmath233&@xmath141$]&@xmath234&@xmath235&@xmath236 + + & & 0.00 & 2026.1360 & @xmath237&@xmath141$]&@xmath238&@xmath239&@xmath240 + p1.5incc @xmath120 & @xmath241 $ ] & @xmath242 $ ] @xmath243 $ ] & @xmath244 & @xmath245 $ ] @xmath246 $ ] & @xmath247 & @xmath248 $ ] @xmath249 $ ] & @xmath250 & @xmath251 $ ] @xmath252 & @xmath253 & @xmath254 @xmath255 & @xmath256 & @xmath257 $ ] @xmath258 ( pc ) & @xmath259 & 1001000 | recent observations suggest that global properties of the host galaxies for long - duration @xmath0-ray bursts ( grbs ) are particularly well - suited for creating the massive star progenitors of these grbs .
motivated by the hypothesis that massive young star clusters located in metal - poor , low - metallicity galaxies are a natural birthsite for grb progenitors , we present a comparison study of the ism properties along the sightline toward the super star cluster , ngc1705 - 1 , and those in distant grb hosts . using the same set of metal transitions in the uv and applying known ism structures in ngc1705
, we find that ngc1705 - 1 resembles distant grb host galaxies in its high neutral gas column , low molecular gas fraction , low metallicity , @xmath1-element enhancement , and low dust depletion .
the lack of molecular gas is due to the enhanced uv radiation field in the starburst environment , consistent with the expectations for grb progenitors .
in addition , the known presence of dense neutral gas clouds at @xmath2 pc from ngc1705 - 1 provides a natural reservoir of c@xmath3 , si@xmath3 , and fe@xmath3 ions that may be subsequently excited by the afterglow uv radiation field to produce excited lines commonly seen in grb host ism .
we further argue that the apparent offset in the velocity profiles of low- and high - ionization transitions from absorption - line observations alone already offers important clues for related starburst episodes in grb host galaxies .
our study shows that a statistical comparison between the ism around star clusters and high - redshift grb progenitors is important for identifying the key physical parameters that facilitate the formation of grbs . |
low - mass and low - surface brightness ( lsb ) galaxies , including in this definition both dwarf ellipticals and irregulars , have since long surged to a central role in the general debate about galaxy formation and evolution .
these objects , typically characterized by an absolute blue magnitude fainter than @xmath10 and a mean surface brightness @xmath11 mag arcsec@xmath12 @xcite , display in fact a quite wide range of morphological and spectrophotometric properties ( e.g. * ? ? ?
* ) such as to rise the question whether they assume an individual role as a sort of pristine building blocks to form more massive systems ( as in the cdm theoretical scheme , e.g. * ? ? ?
* ) , or they rather originate as coarse `` by - products '' of the formation of `` standard '' ellipticals and spirals @xcite .
the problem of the genetic signature of low - mass systems is not a negligible one as , although in a more silent way , dwarf galaxies display in many cases continuous star formation activity , which makes them possible efficient `` engines '' to enrich the interstellar medium ( e.g. * ? ? ?
actually , a major problem in this framework resides in the apparent dichotomy between active ongoing star formation , widely observed in lsb galaxies , and the exceedingly low metallicity of their present - day stellar populations @xcite .
the most striking case in this context is surely that of izw 18 , as recently discussed by @xcite .
again , while for some dwarf ellipticals we might be dealing with a somewhat quiescent evolutionary scenario , reminiscent of old globular cluster history , in other situations a `` tuned '' star formation could have proceeded allover the entire galaxy life ( e.g. * ? ? ?
* ; * ? ? ?
* ) still leaving , nowadays , an important fraction of fresh ( primordial ? ) gas ( e.g. * ? ? ?
* ; * ? ? ?
as far as galaxy spatial distribution is concerned , the ubiquitous presence of lsb irregulars in the different cosmic environments @xcite is clearly at odds with the morphology - density relation @xcite , which has been proven to hold for `` standard '' galaxies for a wide range of densities and scales @xcite .
a systematic observational work to map the lsb galaxy distribution in selected zones of the sky has been carried out by different teams , leading to complete surveys and morphological catalogues of some loose groups of galaxies and nearby clusters @xcite .
these morphological surveys serve as a basic reference to any further spectrophotometric analysis of individual galaxies @xcite .
however , efficient detection and accurate observation of lsb galaxies still remain a formidable task , even for hst observations ( e.g. * ? ? ? * ) especially when moving to distances beyond the virgo and fornax clusters , the two best studied aggregates so far ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
a complementary view may be gained through the study of smaller groups .
in fact , the wide range of densities which characterizes these aggregates makes them the natural link between galaxy clusters and the field .
detailed mapping of the distribution and nature of low - mass galaxies in small groups will fill the `` void '' between fully relaxed systems ( like actually the case of ngc 5044 group , e.g. * ? ? ? * ) to very loose and possibly unbound aggregates @xcite . besides probing the properties of dwarf galaxies in different environments , these have the advantage that even a moderate - sized sample would be fairly representative of the whole group population . at the same time ,
depth effects would be minimized allowing a better analysis of distance - dependent quantities . in this work
we want to further follow up a long - range study on the low - mass lsb galaxy population of the ngc 5044 group .
in particular , after assessing group morphology and membership in previous contributions @xcite , we would like to tackle here the problem of galaxy evolutionary properties , according to a multiwavelength observational approach that includes integrated and surface - brightness multicolour photometry , further complemented in most cases with a spectroscopic input . our analysis will rely on original stellar population synthesis models , to which refer galaxy data in order to lead to a convenient and physically self - consistent picture of the galaxy evolutionary status for the ngc 5044 group . in this framework
, we will arrange our discussion by presenting the observed database in sec . 2 .
its morphological characterization will be carried out in sec . 3 by fitting galaxy surface brightness with a @xcite profile .
these results in most cases refine our previous discussion in @xcite leading also to a homogeneous morphological classification of the full galaxy population in the ngc 5044 group . together with up - to - date redshift measurements , the revised morphological parameters proved to be a useful additional piece of information to constrain group membership and eventually assess the overall physical properties of the galaxy group as a whole . as a central step in this direction , in sec .
4 we study galaxy distribution in different magnitude and colour domains . starting from apparent magnitudes and morphological types , a match with the @xcite template galaxy models also allows us to lead to a realistic estimate of galaxy stellar mass and a reference birthrate figure for the galaxy group as a whole . the spectroscopic properties for our sample are finally discussed in some detail in sec . 5 , where lick narrow - band indices are derived for most of the observed galaxies .
the diagnostic diagrams compare @xmath9-element features , like magnesium and calcium , versus iron and balmer line stregth , also including the contribution from carbon molecules like ch ( g band ) and c@xmath13 .
we will discuss our results and summarize our main conclusions in sec .
our galaxy sample consists of a total of 79 mostly dwarf and lsb galaxies , mainly selected from the @xcite catalogs , and representative of the faint galaxy population surrounding the standard elliptical ngc 5044 , in the range @xmath14 mag , can be estimated , according to the mean spectroscopic redshift of all group members in @xcite . for @xmath15
kms@xmath6 mpc@xmath6 this leads to a value of @xmath16 mag , that we will adopt throughout this paper . ] regardless of morphology .
johnson @xmath17 photometry and @xmath18 imaging in the gunn system @xcite , as well as mid - resolution ( @xmath19 - 1000 ) spectroscopy for a subsample of 24 objects have been collected during different observing runs from 1996 to 2000 .
specifically , 40 objects have been observed during two observing campaigns ( on the nights of apr 16 - 17 , 1999 and apr 29-may 1 , 2000 ) at the eso 3.6 m telescope of la silla ( chile ) , while previous extended observations at the 2.15 m telescope of the observatory , in san juan ( argentina ) were carried out along three observing runs ( on the nights of may 10 - 13 1996 , apr 8 - 10 1997 , and mar 19 - 21 1999 ) providing supplementary data for 57 galaxies , 18 of which in common with the eso sample . ]
eso observations have been carried out with efosc2 , equipped with a loral 2k ccd in a @xmath20 binned mode providing a platescale of 0.32 arcsec pixel@xmath6 .
the @xmath21 field of view allowed in most cases to image more than one galaxy within each frame .
a total of 24 fields were imaged in the @xmath18 bands ( see fig . 1 in * ? ? ?
* for the exact location map ) , including 33 known low - luminosity galaxies , plus the bright sbi - ii ngc5054 , ngc 5044 itself , and 6 newly discovered lsb galaxies ( see * ? ? ?
atmospheric conditions were photometric all the way , with sub - arcsec seeing during the first run , and a poorer performance ( fwhm @xmath22 arcsec ) in the 2000 run . in case of multiple exposures , a final image of the field
was obtained by stacking the individual frames after all processing steps were completed .
standard stars from the lists of @xcite and @xcite were also observed during each run for calibration purposes .
johnson @xmath17 imaging at the casleo telescope has been carried out , on the other hand , during two runs in 1996 and 1999 .
the telescope was equipped with a tektronix 1024@xmath231024 ccd , providing a 9@xmath24 circular field of view in direct - imaging mode , with a platescale of 0.83 arcsec pixel@xmath6 .
standard stars for magnitude calibration were selected from the list of @xcite .
seeing conditions were typically poorer than eso observations , ranging between 2.0 and 2.8 arcsec .
partial results from the 1996 data have been presented in @xcite .
image processing of both eso and casleo data was done using iraf , complemented with a few of our own fortran routines .
each bias - corrected frame was twilight flat - fielded , while residual fringe patterns on eso @xmath25 and @xmath26 images were corrected , mostly within a final @xmath27% accuracy level , by subtraction of the appropriate templates ( see * ? ? ? * for more details on the procedure ) .
cosmic rays were cleaned up using the iraf task cosmicrays , and the sky background , fitted with a tilted plane , was finally subtracted to obtain the final images of each field .
the photometric reduction procedure yielded a nominal figure ( @xmath28 mag ) for the internal magnitude uncertainty of bright galaxies , being the total photometric error dominated throughout by the external uncertainty in the zero - point calibration .
overall , for a @xmath29 galaxy we estimate a full magnitude uncertainty of the order of @xmath30 mag , a figure that may raise to @xmath31 mag for the faintest ( @xmath32 ) objects ( @xmath33 mag for casleo data ) .
a check of internal self - consistency of the @xmath17 photometry has been made possible by the combined observations of galaxies n49 ( @xmath34 ) and n83a ( @xmath35 ) , in common to the 1996 and 1999 runs .
for the first bright object , the coincidence was excellent , as both @xmath36 and @xmath37 isophotal magnitudes and surface brightness have been reproduced within @xmath38 mag , and @xmath39 mag arcsec@xmath12 , respectively .
much poorer is the match of n83a , instead , a very faint dsph galaxy . in the latter case , in fact , repeated @xmath37 photometry resulted in a nearly negligible offset of 0.06 mag , that raised however to 0.35 mag for the @xmath36 band .
a similar figure has been found for the surface - brightness measurements , with a difference of 0.002 and 0.45 mag arcsec@xmath12 in @xmath37 and @xmath36 , respectively .
this warns of the fact that photometric parameters of very faint galaxies may actually show considerable errors . in order to ease a more homogenous discussion of the whole galaxy sample
one would like to convert casleo @xmath17 photometry into eso gunn system .
to this aim we took advantage of the 18 galaxies in common between the two subsamples , as summarized in table [ t1 ] . for these galaxies , integrated magnitudes within one @xmath40-effective radius , as identified by the srsic fitting profile ( see table [ a2 ] ) , have been computed matching casleo @xmath17 and eso @xmath41 photometry .
the linear fit to these data is reported in table [ t2 ] .
one has to remark , however , that the strong clustering of the data in the colour domain ( see fig . [ f1 ] ) , poorly constrains the relations of transformation .
alternatively , more standard sets of empirical conversion relationships can be found in the literature , mainly relying on the observation of stellar grids .
this is the case , for instance , of the works by @xcite and @xcite , also reported in table [ t2 ] .
a further set of fully theoretical conversion equations has been considered in the table , relying on the @xcite template galaxy models .
the latter calibration may in principle be more suitable for our aims , as it better accounts for the slightly different locus of stars and galaxies in the colour domain , given the `` broader '' spectral energy distribution ( sed ) of the latter . | with this third paper of a series we present johnson - gunn @xmath0 multicolour photometry for 79 objects , including a significant fraction of the faintest galaxies around ngc 5044 , assessing group membership on the basis of apparent morphology ( through accurate srsic profile fitting ) and low - resolution ( @xmath1 ) optical spectroscopy to estimate redshift for 21 objects .
early- and late - type systems are found to clearly separate in the srsic parameter space , with the well - known luminosity vs. shape relation being mostly traced by different morphological types spanning different ranges in the shape parameter @xmath2 .
a significantly blue colour is confirmed for magellanic irregulars ( sm / im s ) , while a drift toward bluer integrated colours is also an issue for de s .
both features point to a moderate but pervasive star - formation activity even among nominally `` quiescent '' stellar systems .
together , de s and i m s provide the bulk of the galaxy luminosity function , around @xmath3 , while the s0 and dsph components dominate , respectively , the bright and faint - end tails of the distribution .
this special mix places the ngc 5044 group just `` midway '' between the high - density cosmic aggregation scale typical of galaxy clusters , and the low - density environment of looser galaxy clumps like our local group .
the bright mass of the 136 member galaxies with available photometry and morphological classification , as inferred from appropriate m / l model fitting , amounts to a total of @xmath4 m@xmath5 .
this is one seventh of the total dynamical mass of the group , according to its x - ray emission .
current sfr within the group turns to be about 23 m@xmath5 yr@xmath6 , a figure that may however be slightly increased facing the evident activity among dwarf ellipticals , as shown by enhanced h@xmath7 emission in their spectra .
lick narrow - band indices have been computed for 17 galaxies , probing all the relevant atomic and molecular features in the 4300 - 5800 wavelength range .
dwarf ellipticals are found to share a sub - solar metallicity ( @xmath8 \lesssim -0.5 $ ] ) with a clear decoupling between iron and @xmath9 elements , as already established also for high - mass systems . both de s and ds0 s
are consistent with an old age , about one hubble time , although a possible bias , toward higher values of age , may be induced by the gas emission affecting the h@xmath7 strength . [ firstpage ] galaxies : clusters : ngc 5044 group galaxies : dwarf galaxies : elliptical and lenticular , cd galaxies : photometry galaxies : fundamental parameters |
a _ dehn surgery _ on a knot @xmath0 in a @xmath1-manifold @xmath2 is an operation to create a new @xmath1-manifold from @xmath2 and @xmath0 as follows : remove an open tubular neighborhood of @xmath0 , and glue a solid torus back . by gluing a solid torus back as it was
, the surgery gives the original manifold again .
so such a surgery is called _ trivial _ , and we will ignore it in general . on knots in the @xmath1-sphere @xmath3 , it is an interesting problem to determine and classify all non - trivial dehn surgeries which produce @xmath1-manifolds with cyclic or finite fundamental groups , which we call _ cyclic surgeries _
/ _ finite surgeries _ respectively .
as part of the hyperbolic dehn surgery theorem , thurston @xcite established that there are finitely many cyclic and finite surgeries .
in fact , culler , gordon , luecke and shalen @xcite ( respectively , boyer and zhang @xcite ) proved there are at most three cyclic ( resp .
, five finite ) surgeries .
furthermore , it is conjectured that knots admitting cyclic ( resp . , finite ) surgeries are doubly primitive ( resp .
, primitive / seifert fibered ) knots as introduced by berge @xcite ( resp . ,
dean @xcite ) .
see ( * ? ? ?
* problem 1.77 ) for more information .
cyclic and finite surgeries have been studied extensively for some classes of knots .
for example , it was shown by delman and roberts in @xcite that no hyperbolic alternating knot admits a cyclic or finite surgery .
one of the other well - known classes of knots , containing non - alternating ones , is the montesinos knots .
montesinos knot _ is defined as a knot admitting a diagram obtained by putting rational tangles together in a circle .
see figure [ fig : montesinosknot ] for instance .
in particular , a montesinos knot @xmath0 is called a _ @xmath4-pretzel knot _ if the rational tangles in @xmath0 are of the form @xmath5 . in this paper ,
based on studies by delman @xcite and mattman @xcite , we give a complete classification of cyclic / finite surgeries on montesinos knots as follows .
( 100,80 ) ( 0,0 ) [ mainthm ] let @xmath0 be a hyperbolic montesinos knot .
if @xmath0 admits a non - trivial cyclic surgery , then @xmath0 must be equivalent to the @xmath6-pretzel knot and the surgery slope is @xmath7 or @xmath8 . if @xmath0 admits a non - trivial acyclic finite surgery , then @xmath0 must be equivalent to either the @xmath6-pretzel knot and the surgery slope is @xmath9 , or the @xmath10-pretzel knot and the surgery slope is @xmath11 or @xmath12 . as a direct corollary , together with the result by wu @xcite , we have the following . [ cor ]
let @xmath0 be a hyperbolic arborescent knot .
if @xmath0 admits a non - trivial cyclic surgery , then @xmath0 must be equivalent to the @xmath6-pretzel knot and the surgery slope is @xmath7 or @xmath8 . if @xmath0 admits a non - trivial acyclic finite surgery , then @xmath0 must be equivalent to either the @xmath6-pretzel knot and the surgery slope is @xmath9 , or the @xmath10-pretzel knot and the surgery slope is @xmath11 or @xmath12 . recently , using khovanov homology , it was shown in ( * ? ? ?
* theorem 7.5 ) that the @xmath13-pretzel knot does not admit finite surgeries for @xmath14 . very recently , futer , ishikawa , kabaya , mattman , and shimokawa @xcite obtained , independently , a complete classification of finite surgeries on @xmath15-pretzel knots with odd positive integers @xmath16 and @xmath17 .
it is already known which montesinos knots are non - hyperbolic . if a montesinos knot is equivalent to one consisting of at most two rational tangles , then it actually is a two - bridge knot .
menasco @xcite showed that the non - hyperbolic two - bridge knots are the @xmath18-torus knots .
the only other non - trivial non - hyperbolic montesinos knots are the @xmath19- and @xmath20-pretzel knots , which are actually the @xmath21- and @xmath22-torus knots , respectively .
this was originally shown by oertel ( * ? ? ?
* corollary 5 ) as well as in an unpublished monograph @xcite by bonahon and siebenmann .
the cyclic and finite surgeries of torus knots have been completely classified by moser @xcite . to prove theorem [ mainthm ] ,
we will prepare two propositions , propositions [ prop1 ] and [ prop2 ] , which will be shown in sections [ sec2 ] and [ sec3 ] respectively .
then , in the last section , theorem [ mainthm ] will be proved from these propositions together with a result of mattman @xcite .
the authors would like to thank mikami hirasawa for useful conversations about the fiberedness of pretzel knots as in the proof of claim [ clm : fibered ] .
they also wish to thank hiroshi matsuda for his useful comments about the rank and the euler characteristic of heegaard floer homology as in the proof of claim [ clm : lsurgery ] , and also thank thomas mattman for his useful comments about cyclic and finite surgeries on arborescent knots as in corollary [ cor ] .
they also thank the referee for careful reading of our draft and useful suggestions to improve the proof of claim 2 .
in this section , we prove the following proposition .
[ prop1 ] let @xmath0 be a hyperbolic montesinos knot admitting a non - trivial cyclic or finite surgery .
then @xmath0 is equivalent to a @xmath23-pretzel knot , where @xmath24 is a non - zero integer and @xmath25 are odd positive integers with @xmath26 .
furthermore all non - zero coefficients of the alexander polynomial for @xmath0 are @xmath27 .
suppose that a hyperbolic montesinos knot @xmath0 admits a non - trivial cyclic or finite surgery .
then delman showed in @xcite that @xmath0 must be equivalent to either a @xmath28-pretzel knot , a @xmath23-pretzel knot or a @xmath29-pretzel knot with an integer @xmath24 , integers @xmath30 and odd positive integers @xmath25 ( @xmath26 ) . also see ( * ? ? ?
* section 2 , section 3 ) .
actually delman showed that any montesinos knot except for those listed above admits an essential lamination in its exterior which survives all non - trivial dehn surgeries . _
essential laminations _ were introduced by gabai and oertel in @xcite and , actually , they showed that if a 3-manifold admits an essential lamination , then its universal cover must be the 3-space @xmath31 . in particular its fundamental group is not cyclic or finite .
see @xcite for the precise definition . by virtue of delman
s result , in order to prove proposition [ prop1 ] , it suffices to show that the first and the third types of pretzel knots described above can not have cyclic or finite surgeries .
note here that a @xmath15-pretzel knot ( the case @xmath32 in the first ) is equivalent to a @xmath33-pretzel knot ( the case @xmath34 in the second ) .
also a @xmath35-pretzel knot ( the case @xmath36 in the third ) is equivalent to a @xmath37-pretzel knot ( the case @xmath38 in the second ) .
thus , excluding over laps , we are assuming @xmath39 . among the classes of knots described above , the first one was already studied by mattman in @xcite .
he actually showed in ( * ? ? ? * theorem 1.1 and 1.2 ) that any @xmath28-pretzel knot with @xmath40 and odd positive integers @xmath25 ( @xmath26 ) has neither cyclic surgeries nor finite surgeries .
thus , in the following , we focus on the third class above .
we here use the following strong result by ni , ( * ? ? ?
* corollary 1.3 ) , established by using the heegaard floer homology theory : if a knot in @xmath3 admits a cyclic or finite surgery , then it must be a fibered knot .
actually he showed that a knot @xmath0 in @xmath3 must be fibered if @xmath0 admits a surgery yielding an l - space . here a rational homology sphere @xmath41 is called an _
l - space _ if the rank of @xmath42 is equal to @xmath43 .
in fact , any @xmath1-manifold with a cyclic or finite fundamental group is an l - space , as is shown in ( * ? ? ?
* proposition 2.3 ) .
now , the next claim , together with the result by ni , imply the first conclusion of proposition [ prop1 ] . [ clm : fibered ] let @xmath44 and @xmath45 be odd positive integers @xmath46 . the @xmath29-pretzel knot is not fibered .
we just apply the algorithm given in ( * ? ? ?
* theorem 6.7 ) .
here we include only an outline , assuming that the reader is rather familiar with ( * ? ? ?
* theorem 6.7 ) .
please see @xcite for details .
let @xmath0 be a @xmath47-pretzel knot with an integer @xmath48 and odd positive integers @xmath25 ( @xmath49 ) .
we start to apply the algorithm in ( * ? ? ?
* theorem 6.7 ) with @xmath50 , @xmath51 , @xmath52 , @xmath53 , @xmath54 . after a cyclic permutation ,
the surface @xmath55 obtained by applying seifert s algorithm is of type ii in ( * ? ? ?
* type ii.6.5 ) with @xmath56 , @xmath57 , @xmath58 , @xmath59 , @xmath60 .
( see ( * ? ? ?
* figure 6.3 ) . )
we now see case 2 in ( * ? ? ?
* theorem 6.7 ) . here
we note that the associated oriented pretzel link @xmath61 ( defined in ( * ? ? ?
* type ii.6.5 ) ) is of type @xmath62 . since @xmath63 and @xmath61 is of type @xmath64 if @xmath48 , we check case 2b in ( * ? ? ?
* theorem 6.7 ) .
then we see that @xmath0 is fibered if and only if @xmath61 is fibered . for @xmath61
, we check case 1 in ( * ? ? ?
* theorem 6.7 ) , and verify that @xmath61 is not fibered since no @xmath65 is @xmath66 and @xmath61 is not equivalent to a pretzel link of type @xmath67 with an integer @xmath24 .
therefore we conclude that @xmath0 is not fibered .
the second conclusion of proposition [ prop1 ] follows from results of ozsvth and szab , also achieved by using the heegaard floer homology theory .
we first prepare the following claim , which is implicitly used in ( * ? ? ?
* proof of corollary 1.3 ) .
[ clm : lsurgery ] if @xmath68-dehn surgery on a non - trivial knot @xmath0 in @xmath3 yields an l - space for some coprime integers @xmath69 with @xmath70 , then @xmath71-dehn surgery on @xmath0 also yields an l - space . given coprime integers @xmath69 and a knot @xmath0 in @xmath3 , let @xmath72 denote the 3-manifold obtained from @xmath3 by @xmath68-surgery on @xmath0 .
we recall the following general formula given in ( * ? ? ?
* proposition 9.5 ) : @xmath73 this holds for any pair of coprime integers @xmath69 . for simplicity ,
let @xmath74 denote @xmath75 and @xmath41 denote @xmath76 .
then we have latexmath:[\[\label{rkhf } \mathrm{rk } \widehat{hf } ( s^3_{\alpha / \beta } ( k ) ) = now , for some coprime integers @xmath69 with @xmath70 , we assume that @xmath72 is an l - space , i.e. , by definition , @xmath78 it then suffices to show that @xmath79 is an l - space , i.e. , @xmath80 . on the other hand , in general , we see that @xmath81 for any integer @xmath71 as follows . in the proof of ( * ? ? ?
* proposition 5.1 ) , it is claimed that @xmath82 also see ( * ? ? ?
* section 2 ) . by definition ,
the euler characteristic ( the left - hand side ) is the alternating sum of the dimensions of @xmath83 .
hence , it is not greater than the total rank of @xmath83 , i.e. , @xmath84 from this equation , in order to obtain @xmath80 , it suffices to show that @xmath85 .
actually , we have from equation ; @xmath86 note here that we have @xmath87 as follows .
it is seen that @xmath88 from equation and the assumption that @xmath89 .
thus , together with @xmath90 by definition , we have @xmath87 .
if @xmath91 , then @xmath92 holds . since @xmath87 , together with equation , we obtain that @xmath85 as desired . if @xmath93 , then we have @xmath94 from the assumption that @xmath95 .
thus , together with @xmath87 and equation , we obtain that @xmath96 together with equation , this implies that @xmath85 as desired .
then , in ( * ? ? ?
* corollary 1.3 ) , ozsvth and szab proved that if a knot @xmath0 in @xmath3 admits an integral dehn surgery yielding an l - space , then the alexander polynomial @xmath97 has the form @xmath98 for some increasing sequence of positive integers @xmath99 .
this means that all non - zero coefficients of @xmath100 are @xmath27 . in the above proof ,
claim [ clm : lsurgery ] is actually necessary for the following reason . by the cyclic surgery theorem established in @xcite , all cyclic surgeries on hyperbolic knots in @xmath3 are shown to be integral surgeries .
however , the finite surgery theorem of @xcite shows that finite surgeries on hyperbolic knots in @xmath3 are half - integral or integral . in other words , at present , we can not rule out the possibility of a half - integral surgery and it is currently only a conjecture that such finite surgeries are integral : see ( * ? ? ?
* problem 1.77 a(6 ) ) for more information .
in this section , we prove the following proposition , which will be shown by direct calculations of the alexander polynomials .
[ prop2 ] let @xmath0 be a pretzel knot of type @xmath23 , where @xmath24 is an integer and @xmath25 are odd positive integers with @xmath49 . if every non - zero coefficient of the alexander polynomial of @xmath0 is @xmath27 , then @xmath101 and @xmath102 .
recall that the alexander polynomial @xmath103 for a link @xmath104 satisfies the following skein relation ( see @xcite for example ) : @xmath105 where @xmath106 , @xmath107 , and @xmath108 possess diagrams @xmath109 , @xmath110 and @xmath111 which differ only in a small neighborhood as shown in figure [ fig : skein ] .
[ rem : nabla ] let @xmath112 be a positive integer , and @xmath113 the alexander polynomial of a @xmath114-torus link . set @xmath115 .
then we have @xmath116 .
see @xcite for example .
we divide our proof of proposition [ prop2 ] into three claims .
we denote by @xmath117 a pretzel link of type @xmath118 , and by @xmath119_j$ ] the coefficient of @xmath120 in a polynomial @xmath121 .
[ clm:-2n ] let @xmath24 be an integer with @xmath122 .
let @xmath16 and @xmath17 be odd integers with @xmath123 .
let @xmath0 be a pretzel knot of type @xmath124 .
then we have @xmath125_1 = \left\ { \begin{array}{ll } -4 & \text{if } ~n=1 , \\
-3 & \text{if } ~n \geq 2 , \\
\end{array } \right .
\end{aligned}\ ] ] where @xmath100 is normalized so that @xmath126 and @xmath127_0 > 0 $ ] .
let @xmath128 with @xmath129 and @xmath123 . by applying the skein formula at crossings in the @xmath130-twists repeatedly , we can obtain a resolving tree such that each leaf node corresponds to either @xmath131 or @xmath132 .
notice that @xmath133 is equivalent to a connected sum of a @xmath134-torus knot and a @xmath135-torus knot .
then we have @xmath136 next we calculate the alexander polynomial of @xmath137 by the same argument as above .
by applying the skein formula at crossings in the @xmath16-twists repeatedly , we can obtain a resolving tree such that each leaf node corresponds to either @xmath138 or @xmath139 . note that @xmath140 is equivalent to a @xmath141-torus knot and that @xmath142 is equivalent to a @xmath143-torus link .
then we have @xmath144 hence we have @xmath145 to calculate easily , we consider the polynomial obtained by substituting @xmath146 in the alexander polynomial , namely , @xmath147 . by remark [ rem : nabla ]
, we have @xmath148 here the symbol @xmath149 means that both sides are equal up to multiplications by units of the laurent polynomial ring @xmath150 $ ] .
here we recall that @xmath129 and @xmath123 .
then we have @xmath151_1 & = \left\ { \begin{array}{ll } 2 & \text{if } ~n=1 , \\ 3 & \text{if } ~n \geq 2,\\ \end{array } \right . \\
[ f_{2n-1}f_{p-2}f_{q-1}]_1 & = 3,\\ [ f_{2n-1}f_{p-1}f_{q-2}]_1 & = 3.\end{aligned}\ ] ] therefore we have @xmath152_1 = \left\ { \begin{array}{ll } -2 + 3 + 3 = 4 & \text{if } ~n=1 , \\ -3 + 3 + 3 = 3 & \text{if } ~n \geq 2,\\ \end{array } \right . \end{aligned}\ ] ] that is , @xmath125_1 = \left\ { \begin{array}{ll } -4 & \text{if } ~n=1 , \\
-3 & \text{if } ~n \geq 2.\\ \end{array } \right .
\end{aligned}\ ] ] [ clm:+2n ] let @xmath24 be an integer with @xmath153 .
let @xmath16 and @xmath17 be odd integers with @xmath123 .
let @xmath0 be a pretzel knot of type @xmath154 .
then we have @xmath125_3 = 2 , \end{aligned}\ ] ] where @xmath100 is normalized so that @xmath126 and @xmath127_0 > 0 $ ] .
the proof is similar to that of claim [ clm:-2n ] .
let @xmath155 with @xmath156 and @xmath123 . by applying the skein formula at crossings in the @xmath157-twists repeatedly , we can obtain a resolving tree such that each leaf node corresponds to @xmath158 or @xmath159 .
then we have @xmath160 by applying the same argument as above at crossings in the @xmath16-twists , we have @xmath161 notice that @xmath162 is equivalent to a @xmath163-torus link . by applying the skein formula
, we have @xmath164 .
hence we have @xmath165 and then we have @xmath166 here we recall that @xmath156 and @xmath123 .
then we have @xmath167_3 & = [ f_{2n-2 } f_{p-1 } f_{q-1}]_2\\ & = 6,\\ [ t f_{2n-1 } f_{p-2}f_{q-1}]_3 & = [ f_{2n-1 } f_{p-2 } f_{q-1}]_2\\ & = \left\ { \begin{array}{ll } 5 & \text{if } ~p=3,~ q \geq 3,\\ 6 & \text{if } ~5 \leq
p \leq q , \end{array } \right.\\ [ t f_{2n-1 } f_{p-1 } f_{q-2}]_3 & = [ f_{2n-1 } f_{p-1 } f_{q-2}]_2\\ & = \left\ { \begin{array}{ll } 5 & \text{if } ~p=3,~q=3 , \\ 6 & \text{if } ~p \geq 3,~q \geq 5 , \end{array } \right.\\ [ ( 1+t ) f_{2n-1 } f_{p-1 } f_{q-1}]_3 & = [ f_{2n-1 } f_{p-1 } f_{q-1}]_3 + [ f_{2n-1 } f_{p-1 } f_{q-1}]_2\\ & = \left\ { \begin{array}{ll } 8 + 6 = 14 & \text{if } ~p=3,~q=3 , \\ 9 + 6 = 15 & \text{if } ~p=3,~q \geq 5,\\ 10 + 6 = 16 & \text{if } ~5 \leq p \leq q. \end{array } \right.\end{aligned}\ ] ] therefore we have @xmath152_3&= \left\ { \begin{array}{ll } -6 -5 -5 + 14 = -2 & \text{if } ~p=3,~q=3 , \\
-6 -5 -6 + 15 = -2 & \text{if } ~p=3,~q \geq 5,\\ -6 -6 -6 + 16 = -2 & \text{if } ~5 \leq p \leq q , \end{array } \right.\end{aligned}\ ] ] that is , @xmath127_3=2 $ ] . here
we note that @xmath168 is equivalent to @xmath169 .
[ clm:-2pq ] let @xmath16 and @xmath17 be odd integers with @xmath170 .
let @xmath0 be a pretzel knot of type @xmath15 .
then we have @xmath125_4 = -2 , \end{aligned}\ ] ] where @xmath100 is normalized so that @xmath126 and @xmath127_0 > 0 $ ] .
let @xmath171 with @xmath172 . by applying the skein formula at a crossing in the @xmath173-twists
, we have @xmath174 then we have @xmath175 here we recall that @xmath172
. then we have @xmath176_4 = [ f_{p-1}f_{q-1}]_3 = 4 $ ] and @xmath177_4 = 1 + 1 = 2 $ ] .
therefore we have @xmath178_4 = -4 + 2 = -2 $ ] , that is , @xmath127_4
= -2 $ ] .
this completes the proof of proposition [ prop2 ] .
by propositions [ prop1 ] and [ prop2 ] , if a hyperbolic montesinos knot @xmath0 admits a non - trivial cyclic or finite surgery , then @xmath0 is equivalent to a @xmath179-pretzel knot , where @xmath17 is an odd positive integer with @xmath180 .
this @xmath0 is actually equivalent to a @xmath181-pretzel knot .
then mattman showed in ( * ? ? ? * theorem 1.1 and 1.2 ) that , among such pretzel knots , only the @xmath182- and @xmath183- can have cyclic / finite surgeries , and the surgery slopes are as described in theorem [ mainthm ] .
this completes the proof of theorem [ mainthm ] . | we give a complete classification of the dehn surgeries on montesinos knots which yield manifolds with cyclic or finite fundamental groups . |
the fermilab experiment 687 collaboration has collected a large sample of high - energy photoproduction events , recorded with the e687 spectrometer @xcite@xcite during the 1990/91 fixed - target runs at the wideband photon beamline at fermilab .
although the experiment is focussed on charm physics , a very large sample of diffractively photoproduced light - meson events was also recorded .
this paper reports on a study@xcite of the diffractive photoproduction of the @xmath4 final state and the observation of a narrow dip in the mass spectrum at 1.9 gev / c@xmath0 .
preliminary results on diffractive photoproduction of @xmath2 final states are also shown .
in e687 , a forward multiparticle spectrometer is used to measure the interactions of high - energy photons on a 4-cm - thick be target .
it is a large - aperture , fixed - target spectrometer with excellent vertexing , particle identification , and reconstruction capabilities for photons and @xmath5 s .
the photon beam is derived from the bremsstrahlung of secondary electrons of @xmath6 gev endpoint energy , which were produced by the 800 gev/@xmath7 tevatron proton beam .
the charged particles emerging from the target are tracked by a system of twelve planes of silicon microstrip detectors arranged in three views .
these provide high - resolution separation of primary ( production ) and secondary ( charm decay or interaction ) vertices .
the momentum of a charged particle is determined by measuring its deflections in two analysis magnets of opposite polarity with five stations of multiwire proportional chambers .
three multicell threshold erenkov counters are used to discriminate between pions , kaons , and protons .
photons and neutral pions are reconstructed by electromagnetic ( em ) calorimetry .
hadron calorimetry and muon detectors provide triggering and additional particle identification .
invariant mass after applying a cut on the total energy deposited in the calorimeters with respect to the total energy in the spectrometer.,width=288,height=288 ] pions are produced in photon interactions in the be target .
the data acquisition trigger requires a minimum energy deposition in the hadron calorimeters located behind the electromagnetic calorimeters and at least three charged tracks outside the pair region .
the microstrip system and the forward spectrometer measure the six - pion final state ( in this paper,6@xmath8 refers to the @xmath9 state ) with a mass resolution @xmath10 at a total invariant mass of about 2 gev / c@xmath0 .
it is required that a single six - prong vertex be reconstructed in the target region by the microstrip detector , with a good confidence level .
such a requirement rejects background due to secondary interactions in the target .
exclusive final states are selected by also requiring that the same number of tracks be reconstructed in the magnetic spectrometer .
the six microstrip tracks and the six spectrometer tracks are required to be linked , with no ambiguity in associating the microstrip and spectrometer tracks .
events with particles identified by the erenkov system as definite electrons , kaons , or protons , or as kaon / proton ambiguous are eliminated and at least four out of six particles have to be positively identified as @xmath11 .
particle identification is tested by assuming that one or two out of the six tracks is a @xmath12 , by computing all two - track invariant mass combinations , and verifying that there is no evidence of a peak at the @xmath13 or at the @xmath14 mass .
we eliminated final states with @xmath5 s by rejecting events with visible energy in the electromagnetic calorimeters that was not associated with the charged tracks .
a cut in this variable ( @xmath15 ) is applied on the calorimetric neutral energy normalized to the six - pion energy measured in the spectrometer .
the distribution of the six - pion invariant mass after these cuts is shown in fig . 1 .
the plot shows a structure at 1.9 gev / c@xmath0 . in the following
, only the 6@xmath8 mass region around this structure will be analyzed . for diffractive reactions at our energies ,
the square of the four - momentum transfer @xmath16 can be approximated by the square of the total transverse momentum @xmath17 of the diffractively produced hadronic final state . using this definition , the @xmath17 distribution of diffractive events , fig .
2 , is well described by two exponentials : a coherent contribution with a slope b@xmath18(gev
/ c)@xmath19 consistent with the be form factor @xcite and an incoherent contribution with a slope b@xmath20(gev
/ c)@xmath19 .
taking only events with @xmath21gev@xmath22 , we evaluated a contamination of about @xmath23 from nondiffractive events .
this incoherent contribution shows no structure in the [email protected] gev / c@xmath0 mass range , fig .
the diffractive mass distribution was obtained by subtracting this contribution , parametrized by a polynomial fit , and dividing the yield by the detection efficiency .
the detection efficiency was computed by modeling diffractive photoproduction of a mass m , using the experimentally found slope b@xmath25 , and simulating the decay m@xmath26 according to phase space @xcite .
there is no threshold or discontinuity for the efficiency , particularly in the region of the dip structure . at @xmath27 , the ( self - normalized relative ) efficiency _ a _ varied as _ da / a / dm_@xmath28 = 0.15/@xmath29 .
the efficiency and the efficiency - corrected distribution of the six - pion invariant mass for diffractive events , in the mass range 1.4@xmath30 , are shown in fig .
there no evidence , albeit with large combinatorial backgrounds , for resonance substructure , e.g. , @xmath31 , in the 6@xmath8 data below @xmath32 , either at the mass region of the dip or in nearby sidebands .
similarly , the efficiency or acceptance exhibited no threshold , edge , or discontinuity over the entire mass region observed , when the six pion state was simulated as a sequence of decays of intermediate two - body resonances , for example @xmath33 , even under extreme assumptions of full longitudinal or transverse polarizations for the initial state .
the presence of a dip at 1.9 gev / c@xmath0 was verified by several systematics checks .
diffractive photoproduction of pairs or the associated production of 0bar plus a charm baryon at low t ( where the decay products of the other charm meson or charm baryon are missed ) followed by the 6 @xmath8 decay of the d@xmath34 or 0bar are estimated by monte carlo simulation to be negligible contributions .
it was also checked that demanding more stringent cuts ( i.e. , requiring that all six particles be identified as @xmath11 , applying a sharper cut on the calorimeter neutral energy , or subtracting the incoherent contribution bin by bin ) increases the statistical errors without significantly affecting the behavior shown in fig .
[ fig : third ] .
a three - parameter polynomial fit was performed , solid line in fig.4 , to explore the hypothesis that any structure in this distribution is a statistical fluctuation .
the normalized residual distribution , evaluated for each 10-mev / c@xmath0 bin , is good in the full invariant mass range , with the exception of the interval centered at 1.9 gev / c@xmath0 , the region of the claimed dip , where a poor @xmath35 confidence level interval , is obtained , making it highly unlikely that the observed dip is a statistical fluctuation .
incoherently adding a breit - wigner to the fit does not improve the fit quality much .
invariant mass in the [email protected] gev / c@xmath0 mass range : coherent plus incoherent contribution .
dotted distribution : incoherent contribution .
, width=288,height=288 ]
invariant mass for diffractive events after subtracting incoherent contribution .
solid line : second - order polynomial fit . dashed line : polynomial fit with incoherently - added breit - wigner .
upper dot line : relative detection efficiency ( arbitrary normalization ) .
, width=288,height=288 ] because of the narrow width , the e687 spectrometer mass resolution , @xmath36 at 2 gev / c@xmath0 , was unfolded by applying the method described in ref .
@xcite : the experimentally observed data distribution @xmath37 and the unfolded mass distribution @xmath38 are related by @xmath39 , where @xmath40 is the second derivative with respect to m@xmath28 for the observed distribution .
this relationship results from applying a fourier transform and approximating the resolution function by @xmath41 , which is marginally different from a gaussian .
a fit similar to the following one is used to obtain @xmath40 from the unfold ed data .
the data after unfolding are shown in fig .
[ fig : fifth ] . the dip structure at 1.9
gev / c@xmath0 has been characterized by a two - component fit , adding coherently a relativistic breit - wigner resonance to a diffractive continuum contribution .
the continuum probability distribution @xmath42 has been modeled after a jacob - slansky diffractive parameterization@xcite , plus a constant term @xmath43 @xmath44 the jacob - slansky amplitude @xmath45 is assumed to be the purely real ( @xmath46 ) square root of the probability function @xmath47 . for the fit , a relative phase factor @xmath48 , independent of mass , and a normalizing factor @xmath49 multiplied a relativistic breit - wigner resonance term , giving the overall amplitude @xmath50 fit results are shown in table [ tab : unfolded ] and in fig .
[ fig : fifth ] for a fitted mass range from 1.65 to 2.15 gev / c@xmath0 , symmetric with respect to the dip .
quantities shown are the mass and width of the resonance , the amplitude ratio @xmath51 between the breit - wigner function and the jacob - slansky continuum , the relative phase and the @xmath52/dof .
fit values show consistent evidence for a narrow resonance at m@xmath53 = 1.911 @xmath54 0.004 @xmath54 0.001 gev / c@xmath0 with a width @xmath55 = 29 @xmath54 11 @xmath54 4 mev / c@xmath0 , where the errors quoted are statistical and systematic , respectively .
the fit values shown in fig . [ fig : fifth ] and represented by the parameters of table 1 are stable with acceptable @xmath56 over varying mass ranges from 1.65 to 2.3 gev / c@xmath0 .
we quote as systematic error the sample variance of the fit values due to our choice of fit mass range .
the quality of the fit deteriorates somewhat as the upper limit of the fit for this simple model is extended from 2.1 to @xmath57 .
however , the only fit parameter that is affected is the width , which varies from @xmath58 to @xmath59 .
@xmath60 + @xmath55 ( mev / c@xmath0 ) & @xmath61 + @xmath51 & @xmath62 + @xmath14 ( deg . ) & @xmath63 + @xmath52/dof & 1.1 + invariant mass for diffractive events .
the mass resolution has been unfolded .
fit parameters are listed in table 1 .
, width=288,height=288 ]
about one million of 2(@xmath64 ) events have been recorded by e687 in the 1990/91 fixed target runs at the fermi national accelerator laboratory.their analysis is in progress and in the following only a preliminary presentation is given .
as in the @xmath65 study , exclusive final states are selected based on the total number of charged tracks seen in the spectrometer and final states with @xmath5 s are rejected by requiring no visible energy in the electromagnetic calorimeters . the particle identification has been done applying the same cuts described in section 3 .
the t - distribution shows a @xmath17 slope of about 60 ( gev / c)@xmath19 , in good agrement with the @xmath65 sample .
taking only events with @xmath66gev@xmath22 , our 2(@xmath64 ) sample remains with a contamination of about @xmath67 from nondiffractive events .
the acceptance corrected 2(@xmath64 ) mass spectrum is shown in fig .
[ fig : sixth](top).although we expect , to first order , that the bump in the mass spectrum below 2.0 gev / c@xmath0 is dominated by the @xmath68(1450 ) vector meson , our fit is not good ( solid line ) even when we add 3 p - wave interfering breit - wigner : the @xmath52 is 467.0 for 274 degrees of freedom.as the acceptance is almost flat for m@xmath69gev@xmath22 , the deviations observed should be of physical origin .
the residuals ( data - fitted values ) shown in fig .
[ fig : sixth](bottom ) clearly show structures with a non negligible statistical significance .
their study is in progress . for the time being we can say that the ratio of the residuals to the total yield in 4@xmath8 is of the same order of the @xmath70 mixing effect in 2@xmath8@xcite , shown in fig .
[ fig : seventh ] . ) mass ( top ) and residual ( data - fitted values ) distributions ( bottom).the solid line is the 3 p - wave breit - wigner fit . ,
width=288,height=480 ] spectrum in the range 0.5 - 1.0 gev / c@xmath0 , around the @xmath71 pole mass ( top ) ; b- residual ( fit - data ) once the mixing amplitude is fixed to 0 ; c- residuals ( fit - data ) when the interfering amplitude is included .
, width=288,height=480 ]
the diffractive photoproduction of @xmath4 has been studied by e687 .
evidence has been found for a narrow structure near @xmath72 = 1.9 gev/@xmath73 .
if this dip is characterized as the destructive interference of a resonance with the continuum background , then the parameters of this resonance would be m@xmath53 = 1.911 @xmath54 0.004 @xmath54 0.001 gev / c@xmath0 , with @xmath55 = 29 @xmath54 11 @xmath54 4 mev/[email protected] a resonance could be assigned the photon quantum numbers ( j@xmath74 ) and g=+1 , i=1 due to the final state multiplicity .
9 p. l. frabetti _ et al . _
( e687 coll . ) , hep - ex/0106029 .
frabetti _ et al . _
( e687 coll . ) , nucl .
instrum .
meth . a * 320 * , 519 ( 1992 ) .
frabetti _ et al . _
( e687 coll . ) , nucl .
instrum .
meth . a * 329 * , 62 ( 1993 ) .
de jager , h. de vries , and c. de vries , atom . data and nucl .
data tables * 36 * ( 1987 ) 495 .
f. james , genbod n - body monte - carlo event generator , cern computer centre program library .
e. sjntoft , nucl .
* 163 * ( 1979 ) 519 .
m. jacob and r. slansky , phys .
b * 37 * ( 1971 ) 408 , and phys . rev .
d * 5 * ( 1972 ) 1847 .
frabetti _ et al . _
( e687 coll . ) , eps - hep97,jerusalem,19 - 26 august ( 1993 ) . | a narrow dip structure has been observed at 1.9 gev / c@xmath0 in a study of diffractive photoproduction of the @xmath1 final state performed by the fermilab experiment e687 .
preliminary results on new structures in @xmath2 final states are also given . 0bar@xmath3 |
in recent years , the research of the density dependent symmetry energy is still one of the hot topics in nuclear physics and astrophysics communities .
this is simply because the symmetry energy governs many nuclear and astrophysical phenomena , such as the cooling of neutron stars @xcite , the mass - radius relations of neutron stars @xcite , the study of gravitational waves @xcite , properties of nuclei involved in r - process nucleosynthesis @xcite , and observables in heavy - ion collisions at intermediate energies @xcite .
while significant progress has been made in constraining the symmetry energy around saturation density @xcite , it is rather uncertain at supra - saturation densities .
fortunately , more related experiments are currently underway or being planed at several advanced radioactive ion beam facilities such as csr / china @xcite , frib / usa @xcite , gsi / germany @xcite , riken / japan @xcite , or koria / korea @xcite .
it thus provides more opportunities and enhances our confidences to study the symmetry energy although there are still many uncertain factors , such as the effects of pion potential @xcite , the isospin dependence of in - medium nuclear strong interactions @xcite , the short - range tensor force @xcite and the @xmath5 dynamics @xcite .
it is noted that the sensitivities of the @xmath1 ratio to the symmetry energy shown in ref .
@xcite is quite different from that shown in ref .
the discrepancies observed among the results of different models may be because the latter uses a momentum - dependent nucleon symmetry potential while the previous two studies use a momentum - independent nucleon symmetry potential .
the momentum - dependent nucleon symmetry potential could increase the sensitivity of dense neutron to proton ratio ( thus the @xmath1 ratio ) to the symmetry energy @xcite . while in this study , we use a modified symmetry potential which including the effects of the short - range correlations of neutron and proton @xcite
. the short - range correlations of neutron and proton reduce the kinetic symmetry energy , the strength of the modified symmetry potential is thus increased @xcite .
the effects of such modified symmetry potential on the @xmath1 ratio may be also enlarged . in the following studies
, we confirmed the above deduction . in refs .
@xcite it is argued that due to the difference of @xmath0 potential and nucleon potential , threshold conditions of pion and @xmath0 productions need to be modified .
relating to the question of non - conservation of energy on @xmath0 or @xmath4 productions , in this study , we also investigate how the difference of @xmath0 and nucleon potentials affects pion production .
if the effect of @xmath0 potential on @xmath1 ratio can be neglected , one can use nucleon potential to replace @xmath0 potential .
our study shows that one can replace the @xmath0 potential by the nucleon isoscalar potential if a soft symmetry energy is employed .
_ _ in this study , we use the semi - classical transport model ibuu , in which a momentum - independent and isospin - dependent mean - field potential is adopted @xcite .
the symmetry potential reads @xmath6(\frac{\rho}{\rho_0})^{\gamma}\nonumber\\ & & \times[\pm2\frac{\rho_n-\rho_p}{\rho}+(\gamma-1)(\frac{\rho_n-\rho_p}{\rho})^2],\end{aligned}\ ] ] where @xmath7 = 31.6 mev is the symmetry energy at saturation density , @xmath8 = -6.71 mev is the kinetic symmetry energy at saturation density , and @xmath9 denotes the degree of stiffness of the symmetry potential .
the isoscalar potential for @xmath0 resonance has been studied by many people using various many - body approaches and interactions @xcite .
considering that the mean field potential of @xmath0 resonance mainly depends on its interactions with surrounding nucleons , one can make an assumption that the isoscalar potential of @xmath0 resonance is the same as that for nucleon .
the isovector potential for @xmath0 resonance can be an weighted average of that for neutrons and protons according to the charge state of @xmath0 resonance in the processes @xmath10 @xcite .
it reads @xmath11 for comparison , we also adopt the potential of @xmath0 resonance to be equal to the isoscalar potential of nucleon as the other choice , i.e. , @xmath12 by phenomenology @xmath0 potential has a depth of about -30 mev at @xmath13 while nucleon potential is approximately -50 mev deep @xcite .
therefore , one can also use the form @xcite @xmath14 shown in fig .
[ deltapot ] is our used density - dependent potentials of @xmath0 resonance with soft ( @xmath15 ) and stiff ( @xmath16 ) symmetry potentials , respectively .
one can see that the density - dependent @xmath0 potential may also depend on the symmetry potential used .
ratio with different forms of @xmath0 potential and nucleon symmetry potential ( @xmath17 ) in @xmath2 reaction at beam energies of 200 and 400 mev / a , respectively.,scaledwidth=50.0% ] ratio with different forms of @xmath0 and nucleon symmetry potential ( @xmath17 ) with or without pion potentials in @xmath2 reaction at a beam energy of 400 mev / a.,scaledwidth=50.0% ] , but for the beam energy of 200 mev / a.,scaledwidth=50.0% ] resonance potential on @xmath18 and @xmath19 productions with soft symmetry potential ( @xmath15 ) in central @xmath2 collision at beam energies of 200 and 400mev / a , respectively.,scaledwidth=50.0% ] , but with stiff symmetry potential ( @xmath16).,scaledwidth=50.0% ] resonance potential on @xmath1 ratio with soft symmetry potential ( @xmath15 ) in @xmath2 reaction at beam energies of 200 and 400mev / a , respectively.,scaledwidth=50.0% ] , but with stiff symmetry potential ( @xmath16).,scaledwidth=50.0% ] resonance potential to the @xmath1 ratio in different reaction periods in @xmath2 reaction at a beam energy of 400mev / a.,scaledwidth=50.0% ] before studying the effect of @xmath0 resonance potential on charged pion production , it is instructive to see if @xmath0 resonance potential affects pre - equilibrium free @xmath3 ratio .
[ npt ] shows the effect of @xmath0 resonance potential on pre - equilibrium free @xmath3 ratio .
it is seen that whether for soft or stiff symmetry potentials , the form of @xmath0 resonance potential does not affect pre - equilibrium free @xmath3 ratio at incident beam energies of 200 and 400 mev / a .
only a small fraction of nucleons go through @xmath0 resonances in the whole reaction process , one thus does not see any effect of @xmath0 potential on pre - equilibrium free @xmath3 ratio @xcite .
now we move to the study of pion production . shown in fig .
[ e400piek ] is kinetic energy dependence of @xmath1 ratio with different forms of @xmath0 resonance potential and different symmetry potentials ( @xmath20 ) and with or without pion potentials @xcite in the central collision of @xmath2 at a beam energy of 400mev / a .
[ e200piek ] is same as fig .
[ e400piek ] , but for @xmath21 = 200 mev / a . unlike the pion potential used in ref .
@xcite , our pion potential includes the isoscalar part and the isovector part and both of them are momentum dependent @xcite .
it is seen that the value of @xmath1 ratio with @xmath0 resonance potential including isoscalar and iso - vector potentials is almost equal to that with only isoscalar @xmath0 resonance potential within error . around the coulomb peak @xcite
, there are small differences with different @xmath0 resonance potentials , especially for different isoscalar potentials . because differences of @xmath0 resonance potentials with different charge states are in fact very small ( as shown in fig .
[ deltapot ] ) , one sees negligible effect of @xmath0 isovector potential on the charged @xmath1 ratio .
the general relative large discrepancy of the result corresponding @xmath22 potential is due to its larger difference from the other two as shown in fig .
[ deltapot ] .
the shallower potential well of @xmath22 potential causes @xmath0 resonance has shorter time to exist in nuclear matter , thus less affected by the coulomb potential .
therefore , around the coulomb peak @xcite , the value of the @xmath1 ratio is not so large .
it is also instructive to see if @xmath0 potential affects total charged pion production . shown in fig .
[ g03-npi ] and fig .
[ g15-npi ] are total charged pion productions at different incident beam energies with different @xmath0 potentials and nucleon symmetry potentials .
again , one sees negligible effect of @xmath0 potential on the charged @xmath4 production .
[ gamma03 ] and fig .
[ gamma15 ] show the effect of @xmath0 resonance potential on @xmath1 ratio with soft and stiff symmetry potentials ( @xmath15 , @xmath16 ) at beam energies of 200 and 400mev / a , respectively .
one sees negligible effect of @xmath0 potential on the charged @xmath4 production at higher incident beam energy @xmath21 = 400 mev / a whereas one may see observable effect ( about 5% ) of @xmath0 potential on the @xmath1 ratio at lower incident beam energy @xmath21 = 200 mev / a , especially for the stiff symmetry potential .
the stiff symmetry potential causes larger difference among different charged ( i.e. , different isospin - dependent ) @xmath0 potentials .
it is not surprise to see larger value of @xmath1 ratio with the isovector @xmath0 potential .
since @xmath0 resonance lies in different reaction periods , it is thus interesting to see in which reaction period the @xmath0 potential shows sensitivity to the @xmath1 ratio . in order to know in which reaction period
the @xmath1 ratio shows sensitivity to the @xmath0 potential , similar with previous studies @xcite , in the whole reaction process @xmath23 fm / c we use @xmath24 as the standard calculation , which gives a value of @xmath25 , i.e. , @xmath26 to see the relative sensitivity in different reaction periods ( i.e. , @xmath27 ) , we turn on the @xmath0 potential ( e.g , @xmath28 ) in one reaction period but keep it null in the residual reaction period .
we thus get the other value of the @xmath1 ratio @xmath29 , i.e. , @xmath30 by comparing these new computational results with the standard calculation @xmath25 , one can obtain the relative sensitivity in a certain reaction period , i.e. , @xmath31 .
shown in fig .
[ piratio ] is the decomposition of the sensitivity of the @xmath0 resonance potential to the @xmath1 ratio in different reaction periods .
it is seen that the effect of @xmath0 resonance potential on the @xmath1 ratio in different reaction periods are not more than 2% , thus the total effect of @xmath0 resonance potential on the @xmath1 ratio is roughly 5% . as expected , the maximal effect of @xmath0 resonance potential on the @xmath1 ratio is in the maximal compression period of the @xmath2 reaction .
in summary , in the framework of the isospin - dependent boltzmann - uehling - uhlenbeck(ibuu ) transport model , we studied the effects of @xmath0 resonance potential on the free n / p and @xmath1 ratios in @xmath2 reaction at beam energies of 200 and 400 mev / a .
it is shown that the effect of @xmath0 potential on the ratio of pre - equilibrium free @xmath3 is negligible .
the effect of @xmath0 potential on the ratio of @xmath1 may be observable only at lower incident beam energies . relating to the question of non - conservation of energy on @xmath0 or @xmath4 productions in heavy - ion collisions , one can replace @xmath0 potential by nucleon isoscalar potential especially a soft symmetry energy is employed .
this work is supported by the national natural science foundation of china ( grant nos .
11375239 , 11375062 , 11175219 ) , the 973 program of china ( grant no.2013cb834405 ) , the knowledge innovation project ( grant no .
kjcx2-ew - n01 ) of chinese academy of sciences , the project sponsored by srf for rocs , sem and the national key basic research program of china ( no .
2013cb834400 ) .
99 j. m. lattimer , c. j. pethick , m. prakash , p. haensel , phys .
lett . * 66 * , 2701 ( 1991 ) .
m. prakash , t. l. ainsworth , j. m. lattimer , phys . rev . lett . * 61 * , 2518 ( 1988 )
. f. j. fattoyev , w. g. newton , b.a .
j. a * 50 * , 45 ( 2014 ) .
n. nikolov , n. schunck , w. nazarewicz , m. bender , and j. pei , phys .
c * 83 * , 034305 ( 2011 ) .
b. a. li , c. m. ko , z. z. ren , phys .
lett . * 78 * , 1644 ( 1997 ) .
m. b. tsang et al . , phys .
lett . * 102 * , 122701,(2009 ) . m. b. tsang , z. chajecki , d. coupland , p. danielewicz , f. famiano , r. hodges , m. kilburn , f. lu , w. g. lynch , j. winkelbauer , m. youngs , and yingxun zhang , prog .
. phys . * 66 * , 400 ( 2011 ) .
y. x. zhang , z. li , k. zhao , h. liu , and m. b. tsang , nuclear science and techniques * 24 * , 050503 ( 2013 ) .
k. s. jeong and s. h. lee , nuclear science and techniques * 24 * , 050506 ( 2013 ) .
z. g. xiao et al . , j. phys .
g * 36 * , 064040 ( 2009 )
. symmetry energy project ( sep ) , http://groups.nscl.msu.edu/hira/sep.htm .
roy lemmon for the asy - eos collaboration , talk given at the second international symposium on nuclear symmetry energy ( nusym11 ) , june 17 - 20 , 2011 , northampton , massachusetts , http://www.smith.edu / nusym11/. riken samurai collaboration , http://rarfaxp.riken.go.jp/ribf-tac05/10samurai.pdf .
byungsik hong , talk given at the second international symposium on nuclear symmetry energy ( nusym11 ) , june 17 - 20 , 2011 , northampton , massachusetts , http://www.smith.edu/nusym11 .
w. m. guo , g. c. yong , and w. zuo , phys .
c * 91 * , 054616 ( 2015 ) .
jun hong , p. danielewicz , phys .
c * 90 * , 024605 ( 2014 ) .
g. c. yong , w. zuo , x. c. zhang , phys .
b , * 705 * , 240 ( 2011 ) . v.
r. pandharipande and v. k. garde , phys .
b , * 39 * , 608 ( 1972 ) . c. xu , b. a. li , http://arxiv.org/abs/1104.2075 , arxiv:1104.2075 ( 2011 ) .
t. song , c. m. ko , http://arxiv.org/abs/1403.7363 , arxiv:1403.7363 ( 2014 ) .
g. ferini , m. colonna , t. gaitanos , m. di toro , nucl .
a * 762 * , 147 ( 2005 ) .
m. d. cozma , http://arxiv.org/abs/1409.3110 , arxiv:1409.3110 ( 2014 ) . or hen , b. a. li , w. j. guo , l. b. weinstein , and e. piasetzky , phy .
c * 91 * , 025803 ( 2015 ) .
e. oset and a. plalanques - mestre , nucl .
phys . a * 359 * , 289 ( 1981 ) .
m. baldo and l. s. ferreira , phys .
c * 41 * , 2298 ( 1990 ) . g. e. brown and w. weise , phys .
* 22 * , 279 ( 1975 ) .
t. ericson andw.weise , pions and nuclei , oxford univ .
press , 1988 , isbn 0 - 19 - 852008 - 5 .
b. a. li , nucl .
a * 708 * 365 ( 2002 ) .
oliver buss , `` photon- and pion - induced reactions in a transport approach '' , phd thesis , http://www.uni-giessen.de/cms/fbz/fb07/fachgebiete/physik/einrichtungen/theorie/theorie1/publications/dissertation/buss_diss , ( 2008 ) . b. a. li , phys .
c * 92 * , 034603 ( 2015 ) .
g. c. yong , b. a. li , l. w. chen , and w. zuo , phys .
c * 73 * , 034603 ( 2006 )
. h. l. liu , g. c. yong , d. h. wen , phys .
c * 91 * , 044609 ( 2015 ) .
h. l. liu , g. c. yong , d. h. wen , phys .
c * 91 * , 024604 ( 2015 ) . | based on the isospin - dependent boltzmann - uehling - uhlenbeck(ibuu ) transport model , effects of @xmath0 resonance potential on the free n / p and @xmath1 ratios in the central collision of @xmath2 at beam energies of 200 and 400 mev / a are studied .
it is found that the effect of @xmath0 potential on the ratio of pre - equilibrium free @xmath3 is invisible .
the effect of @xmath0 isovector potential on the kinetic energy integrating ratio of @xmath1 may be observable only at lower incident beam energies and with stiffer symmetry energy .
the strength of the @xmath0 isoscalar potential affects the height of the @xmath1 ratio around the coulomb peak but does not affect the kinetic energy integrating ratio of @xmath1 . in heavy - ion collisions at intermediate energies , relating to the question of non - conservation of energy on @xmath0 or @xmath4 productions , one can replace the @xmath0 potential by nucleon isoscalar potential especially a soft symmetry energy is employed . |
the understanding of the gamma - ray burst phenomenon has been greatly advanced in recent years . while the prompt gamma - ray emission is generally interpreted as due to internal dissipation processes within the relativistic ejecta due to internal shocks near or far above the photosphere ( rees & mszros 1994 ; kobayashi et al .
1997 ; daigne & mochkovitch 1998 ; rees & mszros 2005 ; peer et al . 2006 ) or internal magnetic energy dissipation processes ( e.g. usov 1992 ; thompson 1994 ; drenkhahn & spruit 2004 ; giannios & spruit 2006 ; zhang & yan 2011 ) , the afterglow is usually interpreted as arising from the external shock formed as the fireball is decelerated by the ambient medium ( mszros & rees 1997 ; sari et al .
1998 ) . in the pre-_swift _ era ,
afterglow observations were mostly made in the optical bands .
the data were well explained by the external shock model ( e.g. , mszros & rees 1997 ; sari et al . 1998 ; panaitescu et al .
1998 ; panaitescu & kumar 2001 ; huang et al .
2000 ; see zhang & mszaros 2004 for review ) . however , simultaneous observations with the xrt and uvot on board _ swift _ ( gehrels et al .
2004 ) as well as ground - based optical telescopes in the early afterglow phase have opened a new window to understand grb afterglow ( mszro 2006 ; zhang 2007 ; liang 2010 ) .
in particular , early x - ray afterglow observations revealed erratic flares and early plateaus that are difficult to interpret within the standard theoretical framework ( zhang et al .
2006 , nousek et al .
the flares are believed to be produced by late central engine activity ( burrows et al . 2005 ; fan & wei 2005 ; zhang et al .
2006 ; dai et al .
2006 ; proga & zhang . 2006 ; perna et al .
2006 ) , and the shallow decay segment likely signals a long - lasting wind powered by the grb central engine after the prompt gamma - ray phase ( dai & lu 1998a , b ; zhang et al .
these features indicate that the grb central engine does not die out quickly .
the observed afterglow emission is a superposition of the traditional external shock afterglow and an afterglow related to the late central engine activity ( zhang 2011 ) .
the mix of various emission components makes afterglow lightcurves more diverse ( e.g. , liang et al . 2008 ) .
one may decompose the lightcurves using two different approaches : one through theoretical modeling and the other through empirical fitting .
theoretical modeling prevailing in the pre-_swift _ era ( e.g. , panaitescu et al . 1998 ; panaitescu & kumar 2001 ; huang et al . 2000
; wu et al . 2005 ) is found increasingly difficult in the _ swift _ era , because of the large amount of data , and more importantly , the chromatic behavior that defeats the simplest external shock model .
we therefore take the more empirical approach to perform the analysis ( e.g. , liang & zhang 2006 ; panaitescu & vestrant 2008 , 2011 ; kann et al . 2010 ; 2011 ) .
our results are presented in a series of papers . in the first paper of this series ( li et al .
2012 ; paper i ) , we presented the general features of a `` synthetic `` optical lightcurve based on our decomposition analysis .
the ' ' synthetic '' optical lightcurve shows eight possible components with distinct physical origins ( see figure 2 of paper i ) : these include prompt optical flares ; an early optical flare of the reverse shock origin ; an early shallow decay segment ; the standard afterglow component ( an onset hump followed by a normal decay segment ) ; the post jet break phase ; optical flares ; a re - brightening hump ; and a late supernova ( sn ) bump .
the optical flares and the shallow decay segment may signal late activity of the grb central engine , which have been discussed in detail in paper i. this paper is dedicated to the onset bump ( component iii ) and the late re - brightening bump ( component vi ) in the synthetic lightcurve .
the reason to discuss them together is because physically they share a similar origin of jet - medium interaction , and they are potentially good probes for the fireball and the ambient density profile ( e.g. , rykoff et al.2009 ; oates et al .
2009 ; liang et al . 2010
; l et al . 2012 ; ghirlanda et al .
2012 ; yi et al .
2012 ) .
the fireball model predicts a clear early onset hump in the afterglow lightcurve as the fireball is decelerated by the ambient medium ( rees & mszros 1992 ; mszros & rees 2003 ; sari & piran 1999 for the thin shell case , and kobayashi et al .
1999 ; kobayashi & zhang 2007 for the thick shell case ) .
this prediction was examined by some authors with early optical and x - ray afterglow observations ( e.g. , molinari et al . 2007 ; xue et al .
2009 ; melandri et al .
2010 ; liang et al .
the epoch of deceleration depends on the energy of the fireball , density of the medium , and more sensitively the initial lorentz factor of the ejecta , @xmath7 .
the rising slopes of the onset bumps are determined by both the medium density profile ( the @xmath9 value by assuming @xmath10 ) and the electron spectral index , @xmath11 .
therefore , the afterglow onset bumps would be an ideal probe to study the properties of the fireball and the circumburst medium .
the re - brightening feature is likely related to another emission component , and hence , may probe the jet structure .
we present the data and lightcurve fitting results in 2 and compare the properties of the onset and late re - brightening bumps in 3 .
we measure @xmath7 , @xmath11 , and @xmath9 with the onset bumps in 4 .
an analysis for the physical origin of the re - brightening bumps is presented in 5 .
a summary and discussion of our results are presented in 6 . throughout
, a convention @xmath12 is used for the physical parameters in cgs units , and the slope of a lightcurve and the power - law index of a spectrum is defined with a convention @xmath13 .
all the grbs that have optical afterglow detections by november 2011 ( from feb .
28 1997 to nov . 2011 ) are included in our analysis .
a sample of 225 optical lightcurves are compiled from published papers , or from gcn circulars if no published paper is available .
well - sampled lightcurves are available for 146 grbs .
the sample has been presented in table 1 of paper i , in which grb name , redshift , optical spectral information , observational time interval , as well as @xmath14-ray spectral parameters are summarized ( li et al .
we collected the optical spectral indices @xmath15 is adopted for those grbs whose @xmath15 is not available . ] and the host galaxy extinction @xmath16 for each burst from the same literature in order to reduce the uncertainties introduced by different authors .
galactic extinction correction is made by using a reddening map presented by schlegel et al .
( 1998 ) . since the @xmath16 values are available only for some grbs and the @xmath16 is derived from the spectral fits using different extinction curves , we do not make correction for the grb host galaxy extinction .
the @xmath9-correction in magnitude is calculated by @xmath17 .
the late epoch data ( @xmath18 seconds after the grb triggers ) are heavily contaminated with the flux from the host galaxy .
we fit the host galaxy flux with the late time data and subtract it from the entire light curve .
the isotropic gamma - ray energy ( @xmath19 ) is derived in the rest frame @xmath20 kev energy band using the spectral parameters .
we fit the lightcurves with a model of multiple power - law components . the basic component of our model is either a power - law function , @xmath21 or a smooth broken power - law function , @xmath22^{1/s}$ ] , where @xmath23 , @xmath24 , @xmath25 are the temporal slopes , @xmath26 is the peak time of a bump , and @xmath27 measures the sharpness of a peak of the lightcurve component . in some cases ,
we adopt a tripple broken - power - law model .
we developed an idl code to make best fits with a subroutine called mpfitcraigm / idl / fitting.html . ] .
for the details of our lightcurve fitting , see paper i. as shown in paper i , eight components are decomposed from the observed optical lightcurves .
this paper focuses on the early afterglow onset and late re - brightening afterglow components .
the reason to discuss them together is because they may be both related to the similar physical origin , i.e. , multiple fireball components are decelerated by the circum - burst medium with the same or different jet axes so that the line - of - sight may view them with on - axis or off - axis geometries .
an afterglow onset feature is characterized by a smooth bump with a peak less than 1 hour post trigger , which is followed by a normal ( decay index between -0.75 and -2 ) power - law decay component .
a rebrightening bump is similar to an onset bump but peaks at a much later time .
the supernova bump is a special late rebrightening peaking at around 1 - 2 weeks after a grb trigger , which usually shows a red color .
we exclude the supernovae bumps in this analysis by restricting the peak time of the bumps to be earlier than 7 days after a grb trigger . throughout
, we mark the parameters of the early onset bumps and the late re - brightening bumps with the superscripts `` on '' and `` re '' , respectively .
we classify the optical lightcurves with the detections of an onset bump and/or a late re - brightening bump into five groups as described below .
* group i : this group includes bursts with an onset bump followed by a normal decay segment without a late re - brightening bump , as shown in figure [ onset ] .
these are a textbook version of lightcurves as predicted by the standard fireball model .
erratic flares are superimposed on some of the lightcurves , but they may be produced by late internal shocks as presented in paper i. we have 27 cases in the sample .
* group ii : this group includes bursts with an initial afterglow onset bump followed by a late re - brightening hump .
we have 11 cases in the sample .
they are shown in figure [ onset_rb ] . * group iii : this group includes bursts with an initial shallow decay segment followed by a late re - brightening hump .
we have 8 cases in the sample ( figure [ shallow_rb ] ) .
* group iv : this group includes bursts with an initial normal decay followed by a late re - brightening bump .
we have 7 cases in the sample ( figure [ normal_rb ] ) .
* group v : this group includes bursts with a single bump that peaks at @xmath28 seconds post the grb trigger .
we have 3 cases ( grbs 060614 , 070306 and 100418 ) in the sample ( figure [ late_bump ] ) . since we do not detect a decay component before the bump
, one can not decide with confidence whether it is an onset or re - brightening bump .
we therefore do not include them in the statistical analysis of the two types of bumps .
altogether we have 38 early afterglow onset bumps and 26 late re - brightening bumps for our statistical analysis .
the parameters of the onset and re - brightening bumps in our sample derived from our fits , include the peak flux ( @xmath29 ) , slopes ( @xmath24 and @xmath25 ) , peak time ( @xmath26 ) , the full - width - at - half - maximum ( fwhm , @xmath30 ) , the rising and decaying timescales measured at fwhm ( @xmath31 and @xmath32 ) , the ratio ( @xmath33 ) of @xmath34 , and the ratio ( @xmath35 ) of @xmath36 , are reported in tables 1 and 2 .
we derive the isotropic energy release by integrating the emission over the @xmath37 band ( with the central frequency and the full - width - half - maximum as @xmath38 and @xmath39 , respectively ) , in units of @xmath40 ergs ( @xmath41 ) for the afterglow onset and re - brightening bumps in the time interval [ @xmath42 .
they are also listed in tables 1 and 2 .
the redshift ( @xmath43 ) and the isotropic gamma - ray energy in units of @xmath44 ergs ( @xmath45 ) of these grbs are also reported in tables 1 and 2 .
figure [ onset_rb_dis ] compares the distributions of @xmath24 , @xmath25 , @xmath26 , the r - band luminosity at @xmath26 ( @xmath46 ) and the ratio * ( @xmath33 ) * of rising - to - decaying timescales measured at fwhm .
it is found that both @xmath47 and @xmath48 distributions are well consistent with each other , falling in the range of @xmath49 , with a typical value 1.5 .
both @xmath50 and @xmath51 are also consistent .
they are narrowly distributed in the range [ -1.8 , -0.6 ] with a typical value @xmath52 , except for three outliers whose @xmath51 are steeper than 2 .
the distribution of the ratio * @xmath53 * is clustered around @xmath54 .
the distribution of @xmath55 spans the range 30 to 3000 seconds , while @xmath56 is distributed in a much wider range from several hundreds of seconds to days .
the @xmath57 distribution shows a bimodal feature that peaks at @xmath58 erg s@xmath59 and @xmath40 erg s@xmath59 , respectively , but this feature is not statistically significant in the current sample ) and a corresponding chance probability @xmath60 .
a bimodal feature is statistically acceptable if @xmath61 .
we get @xmath62 and @xmath63 , which means that the bimodal feature is still not statistically acceptable . ] .
a sharp cutoff at the left side of the first peak is likely due to an instrumental detection limit , which selects against faint optical afterglow onset bumps .
the log - normal function fit to the @xmath64 distribution yields @xmath65 ( @xmath66 ) .
figure [ onset_rb_corr ] shows @xmath30 , @xmath67 , and @xmath68 as a function of @xmath26 for both the onset and re - brightening bumps .
we find that @xmath69 with a spearman correlation coefficient @xmath70 and @xmath71 for the onset bumps .
the rebrightening bumps share the same @xmath72 relation , with a slope of @xmath73 .
the peak luminosity is anti - correlated with the peak time for both the onset and re - brightening bumps , but the power - law indices of the correlations are different .
the best fits give @xmath74 $ ] ( with @xmath75 , @xmath71 ) and @xmath76 $ ] ( with @xmath77 , @xmath71 ) .
the @xmath78 is tentatively anti - correlated with @xmath55 , i.e. , @xmath79 $ ] with @xmath80 and @xmath81 .
a correlation analysis between @xmath82 and @xmath56 gives @xmath83 and @xmath84 , confidently indicating that there is no correlation between the two quantities .
both onset and re - brightening bumps are detected for 11 grbs .
the luminosity @xmath64 is usually smaller than @xmath85 , but the energy @xmath82 is comparable to @xmath78 or even larger than @xmath78 , as shown in figure [ lr_lb_corr ] .
we examine the pair correlations for @xmath86 and @xmath87 , and obtain @xmath88 with @xmath89 for @xmath86 and @xmath90 with @xmath91 for @xmath87 .
these indicate that @xmath85 is not corrected with @xmath64 , and there is only a weak correlation between @xmath92 and @xmath82 .
figure [ eiso_onset_rb_corr ] shows @xmath93 and @xmath68 as a function of @xmath5 for both the onset and re - brightening bumps .
it is found that both @xmath85 and @xmath78 are correlated with @xmath5 .
the best fit gives @xmath94 ( with @xmath95 , @xmath96 ) and @xmath97 ( with @xmath98 and @xmath71 ) .
however , we do not find similar correlations for the late re - brightening bumps in the sample .
both x - ray and optical afterglows were observed for most grbs in our sample .
the x - ray afterglow lightcurves observed with _
swift_/xrt are also shown in figures 1 - 5 .
we check whether an onset or re - brightening bump also exist in the x - ray lightcurve at the time of optical onset / rebrightening bump . for the onset bumps ,
the x - ray lightcurves are usually characterized by a steep decay segment or flares .
no associated onset bumps are identified ( see figure 1 ) . for the re - brightening bumps , on the other hand , the associations seem common . a simultaneous bump in the x - ray band
is detected in about half of the optical re - brightening bumps ( figure 2 ) .
these results suggest that the optical lightcurves are less contaminated by the prompt emission tail and late central engine activity emission , and therefore a good probe for the external shock afterglow ( e.g. , rykoff et al.2009 ; oates et al .
2009 ; liang et al . 2010 ; l et al . 2012 ; ghirlanda et al . 2012 ; yi et al .
within the framework of the synchrotron external shock model , a clear onset bump is expected in the early afterglow lightcurve as the fireball is decelerated by the surrounding medium ( e.g. , sari & piran 1999 ; kobayashi & zhang 2007 ) .
the tight @xmath99 correlation and the values of the rising and decay slopes are all consistent with the model that invokes early deceleration of the grb fireball ( shallower than 1.5 ) would disfavor such a possibility . for
a more stratified medium , the transition time to a steeper phase can be longer , which may accommodate some of the observations .
however , such a model predicts a gradual steepening feature in the lightcurve , which is not observed from the data . ] . in this section
we take the deceleration model to interpret the onset bumps , and use the data to probe the properties of fireball and circumburst medium density profile . within the deceleration model ,
the rising and decaying slopes of an afterglow onset bump depend on the circumburst medium density profile parameter @xmath9 and the radiation spectral regime ( e.g. , sari et al .
1999 ; xue et al .
as shown in table 1 of paper i , the spectral index ( @xmath100 ) for most optical afterglows are between 0.5 - 0.7 , indicating that the optical emission is usually in the spectral regime @xmath101 and @xmath102 , where @xmath103 and @xmath104 are the typical and cooling frequencies of synchrotron radiation . in this spectral regime , we have @xmath105 , @xmath106 and @xmath107 for the constant density ism model , and @xmath108 and @xmath109 for the wind model . with the @xmath11 values derived from @xmath100
, we find that the predicted rising slopes of both the ism and wind medium are significantly inconsistent with the observed @xmath24 values of most grbs in our sample . the observed @xmath25 values , on the other hand , generally agree with the prediction of the ism model
. therefore , we consider a circumburst medium density profile @xmath110 where @xmath111 is the transition radius at which the medium turns into a constant density ism , @xmath112 @xmath113 . if the condition @xmath114 is satisfied , where @xmath115 is the deceleration radius , the thin - shell external shock model gives a rising slope is satisfied , in principle one would see a transition from a slow - rising phase to a fast - rising phase .
this is not observed , which suggests that usually @xmath114 is satisfied . ]
@xmath116 since the decaying phase after the peak usually lasts a much longer time , its decay slope is essentially defined by the phase of @xmath117 , so that the long - term decay slope is defined by the constant - density ism scaling cm in some grbs ( e.g. dai & wu 2003 ) . in this case the decay slope @xmath109 .
] , i.e. @xmath118 with the observed @xmath24 and @xmath25 one can derive the @xmath9 and @xmath11 values for each burst .
we find that 32 out of the 38 bursts in our sample can be explained with this simple model .
six grbs can not be explained with this model .
the early bumps in grbs 090102 , 110205 , and 110906 have a very rapid rising with an index 5.2 , 4.0 , and [email protected] , respectively .
this may be due to the contamination of the reversed shock emission ( e.g. , zhang et al .
2003 , zheng et al .
the decay slope of the bumps in grbs 030418 , 071010b and 081126 , on the other hand , are too shallow ( -0.55@xmath119 0.04 , [email protected] , -0.39@xmath119 0.01 , respectively ) to be interpreted in the standard fireball model , and requires significant energy injection .
we therefore only include 32 grbs in our following analysis .
figure [ pkonset ] shows the distributions of @xmath11 and @xmath9 .
the @xmath11 values range from 2 to 3.5 , with a typical value of 2.5 .
both analytical and numerical studies show that the spectrum of electrons accelerated by ultra - relativistic shocks are a power - law with an index in this range ( kirk et al .
2000 , achterberg et al .
2001 , bednarz & ostrowski 1998 , lemonine & pelletier 2003 ) .
the derived @xmath11 values from our analysis are generally consistent with this , but they are distributed in a wider range ( see also panaitescu & kumar 2001 ; yost et al .
2003 ; shen et al .
2006 ; curran et al .
2010 ) .
the @xmath9 values vary from 0 to 1.5 among grbs , and its typical value is 1 . most of them are narrowly distributed in the range @xmath120 .
this is an intermediate regime between ism and stellar wind .
it may be formed via episodic energy ejection of matter by the progenitor star some time before the grb . the exact mechanism to shape such a density profile is subject to further theoretical study .
liang et al . (
2010 ) derived the intial lorentz factor @xmath7 and @xmath115 assuming that the early optical afterglow onset bumps are due to the deceleration of a grb fireball in an ism medium , and discovered a tight correlation between @xmath7 and @xmath19 .
we revisit this relation in view of the modified density profile as discussed above .
we define a dimensionless parameter @xmath121 where @xmath122 is the deceleration radius of the grb fireball in the ism model , which can be calculated by @xmath123^{\frac{1}{4}},\end{aligned}\ ] ] where @xmath124 is the ambient density and @xmath125 is the ratio between the isotropic gamma - ray energy and the isotropic blastwave kinetic energy , @xmath126 is the mass of proton , @xmath127 speed of light . in the moderate density profile
@xmath128 as we derived above , the deceleration radius should be @xmath129^{\frac{1}{3-k}}.\end{aligned}\ ] ] the initial lorentz factor can be then estimated as @xmath130^{\frac{1}{\rm 8 - 2k } } , & \xi \gg 1 \\
1.4\times\left[\frac{{\rm 3}e_{\rm\gamma , iso}(1+z)^{3}}{32\pi\eta n_0m_{\rm p}c^{\rm 5}t_{\rm p}^{\rm 3}}\right]^{\frac{1}{\rm 8 } } , & \xi \ll 1 , \end{array } \right.\end{aligned}\ ] ] where @xmath131 is a dimensionless parameter as a function of @xmath9 ( figure [ gk ] ) .
we take @xmath112 @xmath113 , @xmath132 and @xmath133 to calculate @xmath134 and @xmath135 .
the distributions of @xmath134 and @xmath135 are shown in figure [ gamma_0 ] .
it is found that @xmath136 is distributed in the range [ 50 , 500 ] , and @xmath137 is around 10@xmath138 cm .
the relation between @xmath139 and @xmath140 is shown in figure [ gamma_0-eiso ] .
the best fit yields @xmath141 or @xmath142 .
the slope of this relation is consistent with that reported in liang et al .
( 2010 ) , but the coefficient is smaller since here we adopt the density profile as discussed above and replace the constant factor 2 by 1.4 in our calculation of the initial lorentz factor .
as summarized in 2 , late rebrightenning bumps are observed in 26 grbs .
several models may produce a rebrightening bump , including a medium density jump ( dai & lu 2002 ; dai & wu 2003 ; lazzati et al . 2002 ) , a refreshed shock ( zhang & mszros 2002a ; bjrnsson et al . 2004 ) ; a structured jet ( nakar et al . 2003 ; berger et al .
2003 ; huang et al . 2004 ; liu et al . 2006 ; jia et al .
2012 ) , and emission from a long - lived reverse shock due to certain stratified ejecta profile ( uhm et al . 2012 ) .
however , with a density bump it is difficult to produce very sharp rebrightenings ( nakar & granot 2007 ) . the refreshed shock scenario requires a large total energy budget to be at least comparable to the existing energy in the blastwave , even though a reverse shock can be bright enough to power a rapid optical rebrightening feature ( zhang & mszros 2002a ) . the long - lasting reverse shock model ( uhm et al .
2012 ) needs to have the forward shock suppressed in order to have a significant rebrightening feature from the reverse shock . in the following we focus on the structured jet model .
three scenarios involving structured jets may be employed to explain the late re - brightening bumps .
first , the outflow may be `` patchy '' , and have at least two components with different jet axes .
the on - axis one gives rise to the prompt emission and early afterglow , while the more energetic off - axis one enters the field of view at a later time and gives the re - brightening feature ( e.g. granot et al . 2002 ) .
this model interprets the rebrightening bump as the full emergence of the off - axis jet in the field of view ( huang et al .
2004 ; panaitescu & vestrand 2008 ; guidorzi et al .
2009 , margutti et al .
this requires a uniform jet with a sharp edge , with the line of sight initially outside the jet cone .
the peak of the bump corresponds to the epoch when the @xmath143 cone enters the line of sight , i.e. , @xmath144 , where @xmath145 and @xmath146 are the viewing angle and the jet opening angle , respectively .
this model predicts that the rising index of the lightcurve is steep , i.e. @xmath147 ( panaitescu & vestrand 2008 ) , and the decaying should correspond to the post jet break phase , which should be steeper than 1.5 ( e.g. , liang et al .
as shown in figure [ onset_rb_dis ] , the @xmath24 values are @xmath148 and the @xmath51 values are shallower than 1.5 the late re - brightening bumps in our sample .
all these suggest that the off - axis two - component jet model is not favored by the data of the majority grbs that show a rebrightening bump , even though with proper contrived parameters , some grbs can be still interpreted by this model .
the second model invokes a structured jet with smoothly varying energy per solid angle , usually as a power law or gaussian distribution ( e.g. mszros et al .
1998 ; zhang & mszros 2002b ; rossi et al .
for the gaussian model ( e.g. zhang et al . 2004 ) , an off - axis observer at a large viewing angle would see a smooth rebrightening feature ( kumar & granot 2003 ) .
however , this model predicts too shallow a hump to interpret the observed rebrightening bumps in our sample .
the third model invokes a two - component jet that has a fast , narrow jet being encompassed by a slow , wider jet , with the fast one powering the prompt emission and early afterglow onset and the slow one powering the rebrightening bump ( e.g. racusin et al .
this is likely the best model to interpret the rebrightening bumps .
since the rebrightening is interpreted as deceleration of the second slow jet , the rising and decaying slopes and the @xmath149 ratio should be similar to those of the onset bumps .
this is indeed the case , as shown shown in figure [ onset_rb_dis ] . on the other hand ,
the properties of the re - brightening humps are not correlated with the prompt gamma - ray properties .
this suggests that the slow component responsible for the re - brightening bump is likely an independent component that is not related to the prompt gamma - ray emission .
one may estimate the initial lorentz factor of the slow component if it is decelerated by the same medium as the fast component .
the emission in the re - brightening component should be in the spectral regime @xmath150 .
the luminosity of the radiation from external shocks at @xmath151 can be given by ( e.g. , sari et al .
1998 ) , @xmath152 where @xmath153 is the kinetic energy of the slow component . assuming that the microphysical parameters in the two components are the same
, one can estimate the ratio of @xmath154 for the two components @xmath155 therefore , the ratio of the initial lorentz factors of the two jet components can be estimated as @xmath156 taking typical values from our analysis results , i.e. , @xmath157 , @xmath158 erg s@xmath59 , @xmath159 erg s@xmath59 , @xmath160s and @xmath161s , one can get @xmath162 and @xmath163 for @xmath164 .
therefore , the typical kinetic energy of the slow component is about one order of magnitude larger than the fast component but with a much smaller initial lorentz factor .
we have presented a detailed analysis of the optical afterglow onset and late re - brightening bumps with a sample of 146 well - sampled optical lightcurves .
we made empirical fits to the lightcurves to identify these bumps and studied the statistical properties of various parameters .
we summarize our results in the following . * a smooth onset bump is observed in 38 grbs . among them
27 show a clear onset feature followed by a normal decay segment without detection of a late re - brightening bump , while 11 others are followed by a late re - brightening bump .
sixteen grbs in our sample show a power - law decay segment followed by a late re - brightening hump .
no x - ray onset bump was observed to coincide the optical onset bumps in our sample .
however , an associated x - ray re - brightening bump was detected for half of the optical re - brightening bumps . *
the distributions of @xmath24 and @xmath25 of both the onset and re - brighten bumps are consistent with each other .
the @xmath24 distribution is in the range of [ 0.3 , 0.4 ] with a typical value of @xmath0 , while @xmath25 is narrowly distributed in the range [ -1.8 , -0.6]with a typical value about @xmath165 .
the distribution of the ratio @xmath166 is clustered around @xmath167 .
the @xmath55 distribution ranges from 30 to 3000 seconds , while @xmath56 is distributed in a much wider range from several hundreds of seconds to days . *
both the onset and re - brighten bumps share the same width - peak luminosity relation , i.e. , @xmath168 , indicating that a bump peaking earlier tends to be narrower .
the peak luminosity is anti - correlated with the peak time for both the onset and re - brightening bumps , but the power - law indices of the correlations are different , i.e. , @xmath169 for the onset bumps and @xmath170 for the re - brightening bumps .
no correlation between the peak luminosities of the onset and re - brighten bumps is found . *
it is found that @xmath85 is proportional to @xmath5 , i.e. , @xmath171 , and the isotropic energy release during in the onset bump is also correlated with @xmath5 .
these results indicate that the afterglow onset bumps is likely due to the deceleration of the grb fireballs . *
we take the onset bumps as a probe for the properties of the grb fireball and the circumburst medium density profile .
we find that the electron spectral index @xmath11 is distributed in the range from 2 to 3.5 among bursts , with a typical value of 2.5 .
the medium density profile is characterized by @xmath172 , with the @xmath9 values narrowly distributed in the range [ 0.75 , 1.25 ] , with a typical value @xmath173 .
this profile is intermediate between a constant density ism and a stellar wind .
the physical origin of this profile is subject to further study . with this medium density profile , we re -
derive the initial lorentz factor of the fireball and confirm the @xmath174 correlation discovered in our previous work ( liang et al .
* the peak time and peak luminosity of the rebrightening bumps are not correlated with the properties of the prompt gamma - ray emission and those of the onset bumps .
their rising and decaying slopes are not consistent with the prediction of off - axis two - component jet models .
rather , it seems to be consistent with an on - axis two - component jet model , with the rebrightening bump signaling the slow component . within this interpretation ,
the typical kinetic energy of the slow component is about one order of magnitude larger that the fast component that is responsible for the prompt emission and onset afterglow component , for typical values of the observed peak luminosities of the two components .
its initial lorentz factor is typically only a few tens , being much smaller than that of the fast component .
our results indicate that the optical afterglow is a critical probe for the grb external shock . inspecting the x - ray lightcurves in figures 1 - 5
, one can find that the early x - ray afterglow lightcurves are usually dominated by flares and the tail emission of prompt gamma - rays .
the optical afterglow lightcurves are much less contaminated by these emission components .
the smooth onset bumps are a clean probe of the key parameters of the fireball and circumburst medium density profile .
we acknowledge the use of the public data from the swift data archive .
we appreciate helpful suggestions by the referee .
this work is supported by the 973 " program of china ( 2009cb824800 ) , the national natural science foundation of china ( grants no .
11025313 , 11203008 , 11078008 , 11063001 , 11163001,11033002 ) , special foundation for distinguished expert program of guangxi , the guangxi natural science foundation ( 2013gxnsffa019001 ; 2010gxnsfa013112 , 2011gxnsfb018063 and 2010gxnsfc013011 ) , the special funding for national outstanding young scientist ( contract no .
2011 - 135 ) .
bz acknowledges support from nsf ( ast-0908362 ) .
achterberg , a. , gallant , y. a. , kirk , j. g. , & guthmann , a. w. 2001 , , 328 , 393 antonelli , l. a. , testa , v. , romano , p. , et al .
2006 , , 456 , 509 bednarz , j. , & ostrowski , m. 1998 , physical review letters , 80 , 3911 berger , e. , kulkarni , s. r. , pooley , g. , et al .
2003 , , 426 , 154 bjrnsson , g. , gudmundsson , e. h. , & jhannesson , g. 2004 , , 615 , l77 burrows , d. n. , romano , p. , falcone , a. , et al .
2005 , science , 309 , 1833 cenko , s. b. , kasliwal , m. , harrison , f. a. , et al .
2006 , , 652 , 490 cenko , s. b. , kelemen , j. , harrison , f. a. , et al .
2009 , , 693 , 1484 cenko , s. b. , frail , d. a. , harrison , f. a. , et al .
2011 , , 732 , 29 chester , m. m. , wang , x. y. , cummings , j. r. , et al .
2008 , american institute of physics conference series , 1000 , 421 covino , s. , davanzo , p. , klotz , a. , et al .
2008 , , 388 , 347 cucchiara , a. , cenko , s. b. , bloom , j. s. , et al .
2011 , , 743 , 154 curran , p. a. , evans , p .
a. , de pasquale , m. , page , m. j. , & van der horst , a. j. 2010 , , 716 , l135 dai , z. g. , & lu , t. 2002 , , 565 , l87 dai , z. g. , & wu , x. f. 2003 , , 591 , l21 dai , z. g. , wang , x. y. , wu , x. f. , & zhang , b. 2006 , science , 311 , 1127 daigne , f. , & mochkovitch , r. 1998 , , 296 , 275 drenkhahn , g. , & spruit , h. c. 2004 , astronomical society of the pacific conference series , 312 , 357 fan , y. z. , & wei , d. m. 2005 , , 364 , l42 ferrero , p. , kann , d. a. , klose , s. , et al . 2008 , american institute of physics conference series , 1000 , 257 ferrero , p. , klose , s. , kann , d. a. , et al .
2009 , , 497 , 729 fynbo , j. p. u. , jakobsson , p. , prochaska , j. x. , et al .
2009 , , 185 , 526 gehrels , n. , chincarini , g. , giommi , p. , et al .
2004 , , 611 , 1005 gendre , b. , klotz , a. , palazzi , e. , et al .
2010 , , 405 , 2372 giannios , d. , & spruit , h. c. 2006 , , 450 , 887 golenetskii , s. , aptekar , r. , mazets , e. , et al .
2006 , grb coordinates network , 5837 , 1 golenetskii , s. , aptekar , r. , mazets , e. , et al .
2007 , grb coordinates network , 6879 , 1 gorbovskoy , e. s. , lipunova , g. v. , lipunov , v. m. , et al .
2012 , , 421 , 1874 granot , j. , panaitescu , a. , kumar , p. , & woosley , s. e. 2002 , , 570 , l61 granot , j. , nakar , e. , & piran , t. 2003 , , 426 , 138 granot , j. 2005 , , 631 , 1022 greiner , j. , krhler , t. , fynbo , j. p. u. , et al .
2009 , , 693 , 1610 grupe , d. , gronwall , c. , wang , x .- y . , et al .
2007 , , 662 , 443 guidorzi , c. , clemens , c. , kobayashi , s. , et al . 2009 , , 499 , 439 guidorzi , c. , kobayashi , s. , perley , d. a. , et al
. 2011 , , 417 , 2124 he , h .-
n . , wu , x .- f . ,
toma , k. , wang , x .- y .
, & mszros , p. 2011
, , 733 , 22 huang , y. f. , gou , l. j. , dai , z. g. , & lu , t. 2000 , , 543 , 90 huang , y. f. , wu , x. f. , dai , z. g. , ma , h. t. , & lu , t. 2004 , , 605 , 300 huang , k. y. , wang , s. y. , & urata , y. 2009 , american institute of physics conference series , 1133 , 212 jia , l .- w .
, wu , x .- f . ,
l , h .- j . , hou , s .- j . , & liang , e .- w . 2012 , research in astronomy and astrophysics , 12 , 411 jin , z. p. , xu , d. , covino , s. , et al .
2009 , , 400 , 1829 kann , d. a. , klose , s. , & zeh , a. 2006 , , 641 , 993 kann , d. a. , klose , s. , zhang , b. , et al .
2010 , , 720 , 1513 kann , d. a. , klose , s. , zhang , b. , et al .
2011 , , 734 , 96 kirk , j. g. , guthmann , a. w. , gallant , y. a. , & achterberg , a. 2000 , , 542 , 235 klotz , a. , gendre , b. , stratta , g. , et al .
2008 , , 483 , 847 klotz , a. , gendre , b. , atteia , j. l. , et al .
2009 , , 697 , l18 kobayashi , s. , & zhang , b. 2007 , , 655 , 973 kobayashi , s. , piran , t. , & sari , r. 1997 , , 490 , 92 krhler , t. , greiner , j. , afonso , p. , et al .
2009 , , 508 , 593 krhler , t. , greiner , j. , mcbreen , s. , et al .
2009 , , 697 , 758 krimm , h. a. , yamaoka , k. , sugita , s. , et al .
2009 , , 704 , 1405 kuin , n. p. m. , landsman , w. , page , m. j. , et al .
2009 , , 395 , l21 kumar , p. & granot , j. 2003 , , 591 , 1075 laas - bourez , m. , klotz , a. , coward , d. , et al .
2010 , grb coordinates network , 11382 , 1 lazzati , d. , rossi , e. , covino , s. , ghisellini , g. , & malesani , d. 2002 , , 396 , l5 lazzati , d. , covino , s. , di serego alighieri , s. , et al .
2003 , , 410 , 823 li , w. , & filippenko , a. v. 2008 , grb coordinates network , 7475 , 1 li , w. , filippenko , a. v. , chornock , r. , & jha , s. 2003 , , 586 , l9 li , l. , liang , e .- w . , tang , q .- w . , et al .
2012 , , 758 , 27 ( paper i ) liang , e. , & zhang , b. 2006 , , 638 , l67 liang , e. , racusin , j. l. , zhang , b. , zhang , b .- b . , & burrows , d. n. 2008 , american institute of physics conference series , 1000 , 204 liang , e .- w . ,
yi , s .- x . , zhang , j. , et al .
2010 , , 725 , 2209 liang , e. , & zhang , b. 2006 , , 638 , l67 littlejohns , o. m. , willingale , r. , obrien , p. t. , et al .
2012 , , 421 , 2692 liu , x. w. , wu , x. f. , & lu , t. 2008 , , 487 , 503 l , j. , zou , y .-
c.,lei , w .- h . , et al . 2012 , , 751 , 49 malesani , d. , fynbo , j. p. u. , vreeswijk , p. m. , & villforth , c. 2008
, grb coordinates network , 7436 , 1 mao , j. , malesani , d. , davanzo , p. , et al .
2012 , , 538 , a1 margutti , r. , guidorzi , c. , chincarini , g. , et al .
2010 , , 406 , 2149 melandri , a. , guidorzi , c. , kobayashi , s. , et al .
2009 , , 395 , 1941 melandri , a. , kobayashi , s. , mundell , c. g. , et al . 2010 , , 723 , 1331 mszros , p. , razzaque , s. , & wang , x. y. 2006 , particles and nuclei , 842 , 1007 mszros , p. , & rees , m. j. 1997 , , 476 , 232 mszros , p. , & rees , m. j. 2003 , , 591 , l91 mszros , p. , rees , m. j. , & wijers , r. a. m. j. 1998 , , 499 , 301 molinari , e. , vergani , s. d. , malesani , d. , et al .
2007 , , 469 , l13 nakar , e. , & granot , j. 2007 , , 380 , 1744 nardini , m. , ghisellini , g. , ghirlanda , g. , et al .
2006 , , 451 , 821 nardini , m. , greiner , j. , krhler , t. , et al .
2011 , , 531 , a39 oates , s. r. , page , m .
j. , schady , p. , et al .
2009 , , 395 , 490 page , k. l. , willingale , r. , bissaldi , e. , et al .
2009 , , 400 , 134 panaitescu , a. , & kumar , p. 2001
, , 554 , 667 panaitescu , a. , & kumar , p. 2001
, , 560 , l49 panaitescu , a. , & vestrand , w. t. 2008 , , 387 , 497 panaitescu , a. , & vestrand , w. t. 2011 , , 414 , 3537 panaitescu , a. , mszros , p. , & rees , m. j. 1998 , , 503 , 314 pandey , s. b. , swenson , c. a. , perley , d. a. , et al .
2010 , , 714 , 799 peer , a. , mszros , p. , & rees , m. j. 2006 , , 642 , 995 pelassa , v. , & ohno , m. 2010 , arxiv:1002.2863 perley , d. a. , li , w. , chornock , r. , et al .
2008 , , 688 , 470 perley , d. a. , bloom , j. s. , klein , c. r. , et al .
2010 , , 406 , 2473 perna , r. , armitage , p. j. , & zhang , b. 2006 , , 636 , l29 prochaska , j. x. , chen , h .- w .
, bloom , j. s. , et al .
2007 , , 168 , 231 proga , d. , & zhang , b. 2006 , , 370 , l61 racusin , j. l. , karpov , s. v. , sokolowski , m. , et al .
2008 , , 455 , 183 rana , v. , cenko , b. , harrison , f. , fox , d. , & kelemen , j. 2009 , american astronomical society meeting abstracts # 213 , 213 , # 610.03 rees , m. j. , & mszros , p. 1994
, , 430 , l93 rees , m. j. , & mszros , p. 2005
, , 628 , 847 rees , m. j. , & meszaros , p. 1992
, , 258 , 41p resmi , l. , ishwara - chandra , c. h. , castro - tirado , a. j. , et al .
2005 , , 440 , 477 robertson , b. e. , & ellis , r. s. 2012 , , 744 , 95 rossi , e. , lazzati , d. , & rees , m. j. 2002 , , 332 , 945 rossi , a. , schulze , s. , klose , s. , et al .
2011 , , 529 , a142 ruiz - velasco , a. e. , swan , h. , troja , e. , et al .
2007 , , 669 , 1 rykoff , e. s. , smith , d. a. , price , p. a. , et al .
2004 , , 601 , 1013 rykoff , e. s.,aharonian , f. , akerlof , c. w. , et al .
2009 , , 702 , 489 sakamoto , t. , barthelmy , s. d. , baumgartner , w. h. , et al .
2010 , grb coordinates network , 11169 , 1 sari , r. , & piran , t. 1999 , , 520 , 641 sari , r. , piran , t. , & narayan , r. 1998 , , 497 , l17 schady , p. , de pasquale , m. , page , m. j. , et al .
2008 , american institute of physics conference series , 1000 , 200 schlegel , d. j. , finkbeiner , d. p. , & davis , m. 1998 , , 500 , 525 shen , r. , kumar , p. , & robinson , e. l. 2006 , , 371 , 1441 thne , c. c. , kann , d. a. , jhannesson , g. , et al .
2010 , , 523 , a70 thompson , c. 1994 , , 270 , 480 uehara , t. , uemura , m. , arai , a. , et al .
2011 , , 526 , a92 uhm , z. l. , zhang , b. , hascoet , r. , et al . 2012 , arxiv:1208.2347 usov , v. v. 1992 , , 357
, 472 wren , j. , vestrand , w. t. , wozniak , p. r. , davis , h. , & norman , b. 2009 , grb coordinates network , 9778 , 1 wu , x. f. , dai , z. g. , huang , y. f. , & lu , t. 2005 , , 357 , 1197 xue , r .-
r . , fan , y .- z . , & wei , d .-
2009 , , 498 , 671 yost , s. a. , harrison , f .
a. , sari , r. , & frail , d. a. 2003 , , 597 , 459 yost , s. a. , alatalo , k. , rykoff , e. s. , et al .
2006 , , 636 , 959 yuan , f. , rykoff , e. s. , schaefer , b. e. , et al . 2008 , american institute of physics conference series , 1065 , 103 yuan , f. , schady , p. , racusin , j. l. , et al .
2010 , , 711 , 870 zafar , t. , watson , d. , fynbo , j. p. u. , et al .
2011 , , 532 , a143 zhang , b. , & mszros , p. 2002a , , 566 , 712 zhang , b. , & mszros , p. 2002b , , 571 , 876 zhang , b. , & mszros , p. 2004
, international journal of modern physics a , 19 , 2385 zhang , b. , & yan , h. 2011 , , 726 , 90 zhang , b. , kobayashi , s. , & mszros , p. 2003
, , 595 , 950 zhang , b. , dai , x. , lloyd - ronning , n. m. , & mszros , p. 2004
, , 601 , l119 zhang , b. , fan , y. z. , dyks , j. , et al . 2006 , , 642 , 354 zhang , b. , zhang , b .- b . , liang , e .- w .
, et al . 2007 , , 655 , l25 zhang , b. 2007 , , 7 , 1 zhang , b. 2011 , comptes rendus physique , 12 , 206 lcccccccccccccccccc 030418(v)@xmath175&[email protected]&[email protected]&[email protected]&1190@xmath119109&6015&926&5089&0.18&0.71 & ... & ... & ... & ... & ... & ... + 050502a(v)@xmath176&[email protected]&0.87&[email protected]&58@xmath1195&96&34&62&0.55&0.63&@xmath177& ... &162.5&[email protected] & ... & ... + 050820a(r)@xmath178&[email protected]&[email protected]&[email protected]&477@xmath1196&896&258&637&0.41&0.53&@xmath179&15924@xmath1191244&301.6&[email protected]&226&1.99 + 060110(v)@xmath180&[email protected]&[email protected]&[email protected]&50&132&30&101&0.30&0.58& ... &320@xmath11960& ... & ... &190&0.51 + 060418(h)@xmath181&[email protected]&[email protected]&[email protected]&170@xmath1195&252&85&167&0.51&0.50&@xmath182&4859@xmath1191056&200.3&[email protected]&252&1.27 + 060605(r)@xmath183&[email protected]&[email protected]&[email protected]&590@xmath11945&1216&336&880&0.38&0.75&@xmath184&283@xmath11945&362.3&[email protected]&133&0.62 + 060607a(h)@xmath185&[email protected]&[email protected]&[email protected]&179@xmath1193&220&65&155&0.42&0.35&@xmath186&2342@xmath119149&202.0&[email protected]&310&1.34 + 060904b(r)@xmath187&[email protected]&[email protected]&[email protected]&400&616&195&421&0.46&0.45&@xmath188&77@xmath11910&8.6&[email protected]&108&0.88 + 060906(r)@xmath189&[email protected]&[email protected]&[email protected]&1149@xmath119355&2177&788&1389&0.57&0.76&@xmath190&1727@xmath119139&27.2&[email protected]&120&0.98 + 061007(r)@xmath183&[email protected]&2.00&-1.67&77@xmath1191&130&41&90&0.45&0.50&@xmath191&42104@xmath1194190&1277.2&[email protected]&473&2.37 + 061121(v)@xmath192&[email protected]&[email protected]&[email protected]&208@xmath11925&394&114&280&0.41&0.54&@xmath193&28468@xmath1193272&830.2&[email protected]&210&1.15 + 070318(v)@xmath194&[email protected]&[email protected]&[email protected]&507@xmath11946&947&347&600&0.58&0.77&@xmath195&135@xmath11933&28.7&[email protected]&84&0.53 + 070411(r)@xmath196&[email protected]&0.61&[email protected]&739@xmath11910&1196&473&723&0.65&0.74&@xmath197&1000@xmath119200&79.8&[email protected]&129&0.87 + 070419a(r)@xmath198&[email protected]&[email protected]&[email protected]&765@xmath11930&1212&437&775&0.56&0.59&@xmath199&19@xmath1192&1.8&[email protected]&62&0.43 + 070420(r)@xmath200&[email protected]&[email protected]&[email protected]&202@xmath11922&424&107&317&0.34&0.51& ... &3100@xmath119500& ... & ... &157&1.45 + 071010a(r)@xmath201&[email protected]&1.50&[email protected]&586@xmath11966&996&289&707&0.41&0.48&@xmath202&13@xmath1192&8.3&[email protected]&72&0.45 + 071010b(r)@xmath203&[email protected]&[email protected]&[email protected]&287@xmath119145&1296&215&1081&0.20&0.91&@xmath204&174@xmath11990&6.4&[email protected] & ... & ...
+ 071025(j)@xmath205&[email protected]&[email protected]&[email protected]&548@xmath11929&1090&309&781&0.40&0.55&@xmath206&1500@xmath119300& ... & ...
&193&0.95 + 071031(r)@xmath207&[email protected]&[email protected]&[email protected]&1213@xmath1192&3312&853&2459&0.35&0.75&@xmath208&390@xmath11960&40.9&[email protected]&89&0.71 + 071112c(r)@xmath209&[email protected]&[email protected]&[email protected]&178@xmath11913&396&91&306&0.30&0.47&@xmath210& ... &2.1&[email protected] & ... & ... + 080310(r)@xmath211&[email protected]&1.50&[email protected]&184@xmath11912&286&85&200&0.43&0.45&@xmath212&590@xmath119100&9.9&[email protected]&219&0.77 + 080319a(r)@xmath213&[email protected]&1.80&[email protected]&238@xmath11917&784&127&657&0.19&0.46& ... &800@xmath119100 & ... & ... & ... & ... + 080330(r)@xmath214&[email protected]&0.33&[email protected]&578@xmath11925&1207&395&811&0.49&0.91&@xmath215&41@xmath1196&10.7&[email protected]&74&0.33 + 080603a(r)@xmath216&[email protected]&[email protected]&[email protected]&1044@xmath119167&6296&1130&5167&0.22&0.70&@xmath217& ... & ... & ... &95&1.08 + 080710(r)@xmath209&[email protected]&[email protected]&[email protected]&1934@xmath11946&4212&1174&3038&0.39&0.58&@xmath218&80@xmath11940&23.8&[email protected]&57&1.03 + 080810(r)@xmath219&[email protected]&[email protected]&[email protected]&117@xmath1192&308&78&230&0.34&0.65&@xmath220&3000@xmath1192000&836.7&[email protected]&337&0.89 + 081008(r)@xmath221&[email protected]&[email protected]&[email protected]&163@xmath1192&290&59&230&0.26&0.33&@xmath222& ...
&16.9&[email protected]&250&1.12 + 081109a(h)@xmath223&[email protected]&[email protected]&[email protected]&559@xmath119127&1348&344&1005&0.34&0.98&@xmath202&530@xmath11980& ... & ...
&68&0.68 + 081126(r)@xmath224&[email protected]&[email protected]&[email protected]&159@xmath1192&1450&127&1323&0.10&0.63& ... &900@xmath119200 & ... & ... & ... & ... + 081203a(u)@xmath225&[email protected]&[email protected]&[email protected]&295@xmath1192&766&205&561&0.36&0.55&@xmath226&1700@xmath119400&904.0&[email protected]&226&1.57 + 090102(r)@xmath227&[email protected]&5.22&[email protected]&50@xmath1191&45&11&34&0.31&0.20&@xmath228&1400@xmath119500&26.6&[email protected]&61&0.84 + 090313(r)@xmath229&[email protected]&[email protected]&[email protected]&1315@xmath119109&2126&690&1436&0.48&0.53&@xmath230&460@xmath11950&445.7&[email protected]&107&1.02 + 090510(r)@xmath231&[email protected]&[email protected]&[email protected]&1579@xmath119650&5046&853&4194&0.20&0.90&@xmath232&42@xmath1194&0.3&[email protected]&61&0.84 + 090812(r)@xmath233&[email protected]&[email protected]&[email protected]&71@xmath1198&104&35&69&0.51&0.49&@xmath234&4586@xmath119597&27.6&[email protected]&399&0.98 + 100901a(r)@xmath235&[email protected]&[email protected]&[email protected]&1260@xmath11976&2349&577&1772&0.33&0.43&@xmath236&245&22.2&[email protected]&87&1.21 + 100906a(v)@xmath235&[email protected]&[email protected]&[email protected]&101@xmath1194&245&55&189&0.29&0.43&@xmath237& ... &124.2&[email protected] & ... & ... + 110205a(r)@xmath238&[email protected]&4.00&[email protected]&948@xmath1193&971&246&725&0.34&0.24 & ... & ... & ... & ... & ... & ... + 110213a(v)@xmath238&[email protected]&[email protected]&[email protected]&293@xmath11912&641&153&488&0.31&0.48&@xmath239& ... & ... & ... &165&0.97 + lccccccccccccccccc 021211(r)@xmath175&[email protected]&1.80&[email protected]&0.50&3.88&0.31&3.56&0.09&0.50&@xmath240&111@xmath11910&1.4&[email protected] + 021004(r)@xmath176&[email protected]&2.50&[email protected]&[email protected]&228.51&62.01&166.49&0.37&0.34&@xmath241&500@xmath119115&373.5&[email protected] + 030329(r)@xmath178&[email protected]&[email protected]&[email protected]&[email protected]&265.42&80.05&185.37&0.43&0.39&@xmath242&155@xmath11914&43.1&[email protected] + 050502a(v)@xmath180&[email protected]&0.80&[email protected]&[email protected]&4.40&1.65&2.74&0.60&0.65&@xmath243& ... &46.3&[email protected] + 050721(r)@xmath181&[email protected]&0.50&-1.06&[email protected]&214.89&67.97&146.92&0.46&0.81& ... &460@xmath11990&0 & ... + 050820a(r)@xmath183&[email protected]&[email protected]&[email protected]&[email protected]&39.87&12.09&27.78&0.44&0.65&@xmath244&15924@xmath1191244&287.2&[email protected] + 060526(r)@xmath185&[email protected]&1.60&-2.28&[email protected]&104.72&45.73&58.99&0.78&0.42&@xmath245&606@xmath119303&183.7&[email protected] + 060729(u)@xmath187&[email protected]&[email protected]&[email protected]&[email protected]&67.93&18.61&49.32&0.38&0.65&@xmath246&65@xmath1195&167.8&[email protected] + 060904b(r)@xmath189&[email protected]&8.00&[email protected]&[email protected]&3.46&0.60&2.86&0.21&0.27&@xmath247&77@xmath11910&12.3&[email protected] + 060906(r)@xmath192&[email protected]&[email protected]&-1.20&[email protected]&13.32&3.28&10.04&0.33&0.31&@xmath248&1727@xmath119139&151.3&[email protected] + 060927(v)@xmath194&@xmath249&1.40&-1.20&0.48&0.79&0.24&0.55&0.44&0.49&@xmath250&5815@xmath119862&183.8&1095.82 + 070318(v)@xmath196&[email protected]&[email protected]&-1.60&[email protected]&73.03&29.10&43.93&0.66&0.59&@xmath251&135@xmath11933&38.9&[email protected] + 071003(r)@xmath198&[email protected]&1.00&-1.10&20.00&38.31&11.75&26.56&0.44&0.60&@xmath252&1800@xmath119600&306.1&[email protected] + 080310(r)@xmath200&[email protected]&0.51&-1.07&[email protected]&3.48&1.21&2.28&0.53&0.79&@xmath253&590@xmath119100&119.0&[email protected] + 080319c(v)@xmath201&[email protected]&[email protected]&[email protected]&[email protected]&0.56&0.11&0.45&0.25&0.34&@xmath254&5206@xmath1191041&16.5&[email protected] + 080330(r)@xmath203&[email protected]&1.20&[email protected]&[email protected]&3.15&0.96&2.19&0.44&0.54&@xmath255&41@xmath1196&15.8&[email protected] + 080413a(r)@xmath205&[email protected]&1.00&[email protected]&0.15&0.26&0.09&0.17&0.49&0.59&@xmath256&1855@xmath119397&105.6&[email protected] + 080710(r)@xmath207&[email protected]&[email protected]&[email protected]&[email protected]&13.96&5.90&8.06&0.73&0.90&@xmath257&80@xmath11940&26.4&[email protected] + 080913(j)@xmath209&[email protected]&2.30&-0.95&[email protected]&173.07&38.53&134.54&0.29&0.43&@xmath258&711@xmath11989&51.3&[email protected] + 081029(r)@xmath213&[email protected]&1.77&[email protected]&[email protected]&14.57&5.71&8.86&0.64&0.40&@xmath259& ... &462.7&[email protected] + 090102(r)@xmath214&[email protected]&0.80&[email protected]&[email protected]&13.19&3.73&9.46&0.39&0.67&@xmath260&1400@xmath119500&8.2&[email protected] + 090902b(r)@xmath216&[email protected]&1.20&-0.81&[email protected]&360.45&82.29&278.15&0.30&0.56&@xmath261&4586@xmath119597&77.5&[email protected] + 100219a(r)@xmath219&[email protected]&0.91&-2.24&[email protected]&22.71&10.93&11.79&0.93&0.59&@xmath262& ... &34.6&[email protected]
+ 100901a(r)@xmath263&[email protected]&[email protected]&-1.50&[email protected]&31.58&9.57&22.01&0.43&0.37&@xmath264&245&371.2&[email protected] + 101024a(r)@xmath221&[email protected]&0.30&[email protected]&[email protected]&5.78&1.84&3.94&0.47&0.92 & ... & ... & ... & ... + 110213a(v)@xmath223&[email protected]&[email protected]&[email protected]&[email protected]&7.72&3.12&4.60&0.68&0.53&@xmath265& ... &685.7&[email protected] and deceleration radius @xmath135 for the medium density profile derived from the afterglow onset bumps in our sample.,title="fig : " ] and deceleration radius @xmath135 for the medium density profile derived from the afterglow onset bumps in our sample.,title="fig : " ] | we continue our systematic statistical study of various components in gamma - ray burst ( grb ) optical lightcurves .
we decompose the early onset bump and the late re - brightening bump with empirical fits and analyze their statistical properties . among the 146 grbs that have well - sampled optical lightcurves , the onset and re - brightening bumps
are observed in 38 and 26 grbs , respectively .
it is found that the typical rising and decaying slopes for both the onset and re - brightening bumps are @xmath0 and @xmath1 , respectively .
no early onset bumps in the x - ray band are detected to be associated with the optical onset bumps , while an x - ray re - brightening bump is detected for half of the re - brightening optical bumps .
the peak luminosity is anti - correlated with the peak time , @xmath2 for the onset bumps and @xmath3 for the re - brightening bumps .
both @xmath4 and the isotropic energy release of the onset bumps are correlated with @xmath5 , whereas no similar correlation is found for the re - brightening bumps .
these results suggest that the afterglow onset bumps are likely due to the deceleration of the grb fireballs .
taking the onset bumps as probes for the properties of the fireballs and their ambient medium , we find that the typical power - law index of the relativistic electrons is 2.5 and the medium density profile behaves as @xmath6 within the framework of the synchrotron external shock models . with the medium density profile obtained from our analysis , we also confirm the correlation between initial lorentz factor ( @xmath7 ) and @xmath8 in our previous work .
the jet component that produces the re - brightening bump seems to be on - axis and independent of the prompt emission jet component .
its typical kinetic energy budget would be about one order of magnitude larger than the prompt emission component , but with a lower @xmath7 , typically several tens . |
_ dehn surgery _ is the process of removing a tubular neighborhood of a ( tame ) knot in the @xmath0-sphere @xmath2 and gluing a solid torus back by a homeomorphism of the boundaries to get a new @xmath0-manifold .
this was first used in 1910 by dehn for easy and effective constructions of homology spheres @xcite .
in fact , a fundamental theorem due to lickorish and wallace says that every connected closed oriented @xmath0-manifold can be obtained by dehn surgery along a * link * from @xmath2 @xcite .
this connection between knot theory in @xmath2 and @xmath0-manifold topology in general lead to many interesting results in both areas .
one of the most striking applications of dehn surgery to knot theory is the _ knot complement theorem _ for knots in @xmath2 . in 1908
tietze asked if a knot in @xmath2 is determined by the homeomorphism type of its complement @xcite . while this is not true for links in @xmath2 @xcite and for knots in general @xmath0-manifolds @xcite
, the original question remained a long time open , until it was spectacularly answered in the affirmative by gordon and luecke in 1989 @xcite .
gordon and luecke proved a slightly more general statement about dehn surgery , namely that if a non - trivial dehn surgery along a single knot @xmath3 in @xmath2 gives back @xmath2 ( a so - called _ cosmetic _ dehn surgery ) , then the knot @xmath3 has to be equivalent to the unknot . from this
it follows very easily that every knot in @xmath2 is determined by its exterior and , by work of edwards @xcite , also by its complement .
the connection between these two different results is as follows .
given the meridian on a knot exterior , it is easy to reconstruct the knot . first one can glue a @xmath4-disk @xmath5 to the meridian in a unique way
. then the boundary of the resulting object is a @xmath4-sphere @xmath6 to which one can glue a @xmath0-ball @xmath7 in a unique way .
therefore there is a unique way to glue a solid torus to the boundary of a knot exterior , such that the meridian of this solid torus maps to the meridian of the knot .
the knot is then recovered as the spine of this newly glued - in solid torus . with this discussion
the knot complement problem reduces to the question , in how many different ways one can glue a solid torus to the boundary of a knot exterior to get @xmath2 back , which is nothing but the existence question of cosmetic dehn surgeries . here
we want to study similar questions for transverse knots in contact @xmath0-manifolds .
a ( coorientable ) _ contact structure _ on a smooth @xmath0-manifold @xmath8 is a _ completely non - integrable _ @xmath4-plane field @xmath9 , meaning that there exists a globally defined @xmath10-form @xmath11 on @xmath8 , such that @xmath12 is a volume form of @xmath8 .
the so - called standard contact structure @xmath13 on @xmath14 is given in cartesian coordinates as @xmath15 .
a smooth knot @xmath16 in a contact manifold @xmath17 is called _ transverse _ , if it is always transverse to the contact structure .
it is easy to see that every topological knot can be @xmath18-close approximated by a transverse knot . for this and other background on contact
topology we refer the reader to the standard textbook @xcite . by using that every transverse knot looks locally the same , martinet could make dehn surgery along transverse knots compatible with the contact structures ( * ? ? ?
* theorem 4.1.2 ) .
together with the theorem of lickorish and wallace it follows that every closed oriented @xmath0-manifold carries a contact structure . in fact , every contact structure on any closed @xmath0-manifold can be obtained by contact surgery along a transverse link from @xmath19 @xcite .
here we study cosmetic contact dehn surgeries along transverse knots in @xmath19 , i.e. contact dehn surgeries yielding again @xmath19 . from a result of etnyre and ghrist ( * ? ? ? * theorem 2.2 ) it follows , that cosmetic contact surgeries do exist along a transverse unknot with self - linking number @xmath20 .
in contrast to this , is our main result which is proven in section [ section : proofdehnsurgery ] .
[ transverse contact surgery theorem][thm : transversesurgeryalg ] let @xmath16 be a transverse knot in @xmath19 , which is not isotopic to a transverse unknot with @xmath20 .
then there exists no non - trivial contact dehn surgery along @xmath16 in a sufficiently small tubular neighborhood of @xmath16 yielding @xmath19 back .
in section [ section : prelim ] we provide the necessary background about transverse knots in contact @xmath0-manifolds and contact dehn surgery along transverse knots and make these statements more precise . as an easy application of this result we prove in section [ section : proofknotexterior ] the following generalization of the theorem of gordon and luecke .
[ transverse knot exterior theorem ] [ cor : transversalg ] two transverse knots in @xmath19 are isotopic if and only if their sufficiently big exteriors are contactomorphic .
we conclude in section [ section : open ] with a discussion about the size of the standard tubular neighborhoods of transverse knots , with explicit counterexamples to corollary [ cor : transversalg ] in the case of transverse links in @xmath19 and transverse knots in general contact @xmath0-manifolds and state a few related open problems .
the results presented here are part of my thesis @xcite , which was partially supported by the _ dfg - graduiertenkolleg 1269 globale strukturen in geometrie und analysis _ at the universitt zu kln .
i would like to thank my advisor hansjrg geiges and my colleagues sebastian durst and christian evers .
in this section we recall the necessary background about transverse knots and contact dehn surgery along them and give the precise statements of the theorem [ thm : transversesurgeryalg ] and corollary [ cor : transversalg ] .
a knot @xmath16 in a contact @xmath0-manifold @xmath17 with cooriented contact structure is called * transverse * , if the contact planes @xmath21 are everywhere transverse to the tangent space of the knot .
transverse knots admits a preferred orientation by the requirement that they are positive transverse to the contact structure . here
all transverse knots are understood to be oriented like this , but all results holds similar also for negative transverse knots .
an important fact about transverse knots is that they locally all look the same ( * ? ? ?
* example 2.5.16 ) .
[ neighborhoods of transverse knots ] let @xmath16 be a transverse knot in @xmath17 . then there exists a * standard tubular neighborhood * @xmath22 of @xmath16 in @xmath8 such that @xmath23 is contactomorphic to @xmath24 where @xmath25 is an angular coordinate on @xmath26 and @xmath27 are polar coordinates on the disk @xmath28 with radius @xmath29 and the contactomorphism sends @xmath16 to @xmath30 . the size @xmath29 of such neighborhoods @xmath22 for transverse knots @xmath16 is not arbitrary .
the maximal value of this size is an invariant of the transverse knot @xmath16 , as studied for example in @xcite .
let @xmath16 be a transverse knot in @xmath17 .
then one calls @xmath31 the * complement * of @xmath16 and @xmath32 an * exterior * of @xmath16 . here
i want to consider transverse knots up to coarse equivalence .
let @xmath33 and @xmath34 be two transverse knots in @xmath17
. then @xmath33 is * ( coarse ) equivalent * to @xmath34 , if there exists a contactomorphism @xmath35 of @xmath17 @xmath36 that maps @xmath33 to @xmath34 .
the ( coarse ) equivalence is in general a weaker condition than the equivalence given by transverse isotopy ( for example in overtwisted contact structures on @xmath2 @xcite ) .
but one can show that in @xmath37 this two concepts are the same ( see ( * ? ? ?
* section 4.3 ) ) .
if two transverse knots are coarse equivalent , their complements and exteriors are contactomorphic . the transverse knot complement or exterior problem asks if the reverse also holds . in section [ section : proofknotexterior ]
we prove corollary [ cor : transversalg ] , which can be made more precise as follows .
[ transverse knot exterior theorem ] [ cor : transverse ] let @xmath33 and @xmath34 be two transverse knots in @xmath37 with contactomorphic sufficiently big exteriors ( meaning that the size of the tubular neighborhoods is sufficiently small ) .
then @xmath33 and @xmath34 are equivalent .
next , i recall the definition of contact dehn surgery along transverse knots . roughly speaking one cuts out the neighborhood of a transverse knot and glues a contact solid torus back in a different way to obtain a new contact @xmath0-manifold .
more precisely : let @xmath16 be a transverse knot in @xmath17 . take a standard tubular neighborhood @xmath22 and a non - trivial simple closed curve @xmath38 on @xmath39 and a diffeomorphism @xmath40 , such that @xmath41 then define @xmath42 where the contact structure @xmath43 on @xmath44 is chosen , such that @xmath43 and @xmath21 fits together to a global new contact structure @xmath45 on @xmath46 .
one says that @xmath47 is obtained out of @xmath8 by * contact dehn surgery * along @xmath16 with * slope * @xmath38 .
it is easy to show that @xmath46 is again a @xmath0-manifold independent of the choice of @xmath40 .
also it is a standard fact that every nullhomologous knot @xmath16 ( for example if @xmath8 is @xmath2 ) has a * seifert surface * , i.e. a compact oriented surface @xmath48 with oriented boundary @xmath16 .
then one can obtain a special parallel copy of @xmath16 on @xmath49 , called the * surface longitude * , by pushing @xmath16 in a seifert surface @xmath48 .
( this is independent of the choice of the seifert surface . )
then one can write any non - trivial simple closed curve @xmath38 uniquely as @xmath50 with @xmath51 the meridian of @xmath22 and @xmath52 and @xmath53 coprime integers .
therefore , it is often easier to think of @xmath38 instead of a simple closed curve as a rational number @xmath54 , where @xmath55 means @xmath56 and @xmath57 .
this rational number is called * surgery coefficient*. also it is easy to show that there exists always a contact structure @xmath43 on @xmath58 that fits together with the old contact structure @xmath21 ( see for example the proof of theorem 4.1.2 in @xcite ) , but this contact structure is in general not unique .
therefore , the resulting contact structure @xmath45 on the new manifold is also not unique and depends moreover highly on the size of the tubular neighborhood @xmath22 .
here we are interested in so - called * cosmetic contact dehn surgeries * along transverse knots , i.e. situations where @xmath47 is again contactomorphic to @xmath17 . that cosmetic contact dehn surgeries do exist follows from the following theorem due to etnyre and ghrist ( * ? ? ?
* theorem 2.2 ) .
[ existence of cosmetic contact surgeries][thm : etnyre - ghrist ] there exists a transverse unknot @xmath16 with self - linking number @xmath20 in @xmath19 , such that for every surgery coefficient @xmath59 there exists a contact dehn surgery along @xmath16 yielding a tight contact manifold .
since dehn surgery along an unknot with surgery coefficient of the form @xmath60 , for @xmath61 , gives back @xmath2 and since on @xmath2 the standard contact structure @xmath13 is the only tight contact structure , it follows that there exist cosmetic contact surgeries from @xmath19 to itself for every surgery coefficient of the form @xmath60 .
observe also that transverse unknots in tight contact @xmath0-manifolds are completely classified by there self - linking numbers ( * ? ? ?
* theorem 5.1.1 ) .
moreover , the possible values for the self - linking numbers are all negative odd integers . in particular
, there is only one transverse unknot with self - linking number @xmath20 .
our main result , theorem [ thm : transversesurgeryalg ] , is in contrast to the result of etnyre and ghrist and can be stated precisely as follows .
[ transverse contact surgery
theorem][thm : transverse ] let @xmath16 be a transverse knot in @xmath19 .
if @xmath62 , constructed using a sufficiently small tubular neighborhood @xmath22 , is contactomorphic to @xmath19 for some @xmath63 , then @xmath16 is equivalent to the transverse unknot with self - linking number @xmath20 .
notice also , that the contact surgeries in theorem [ thm : etnyre - ghrist ] are done with respect to a sufficiently big tubular neighborhood of the knot @xmath16 to ensure tightness of the resulting manifold as explained in ( * ? ? ?
* remark 2.1 ) . for small tubular neighborhoods
this theorem is not true . this
i will explain in more detail in section [ section : open ] .
in @xcite a similar theorem is proven for * legendrian * knots , i.e. smooth knots always tangent to the contact structure .
the proof here follows the same ideas .
first , it follows from the theorem of gordon and luecke @xcite that @xmath16 is topological equivalent to the unknot and an easy computation in homology shows that the surgery coefficient has to be of the form @xmath64 , for @xmath65 .
therefore by eliashberg s classification of transverse unknots ( * ? ? ?
* theorem 5.1.1 ) , @xmath16 is equivalent to the unique transverse unknot @xmath66 with self - linking number @xmath67 . the idea of the proof is now to consider , for every transverse unknot @xmath66 ( other than @xmath68 ) and every surgery coefficient of the form @xmath64 , a legendrian knot @xmath69 in the exterior @xmath70 of @xmath66 .
this legendrian knot can also be seen as a legendrian knot in the new surgered contact manifold @xmath71 . in this situation
the formula for computing the new * thurston bennequin invariant * @xmath72 of the legendrian knot @xmath69 in the new contact manifold from @xcite simplifies to @xmath73 where @xmath74 is the old thurston
bennequin number of @xmath69 in @xmath19 and @xmath75 denotes the linking - number between @xmath69 and @xmath16 in @xmath19 .
although the new contact structure @xmath76 is not unique , the new thurston
bennequin invariant @xmath72 is . a theorem by eliashberg @xcite states that a contact manifold is * tight ( not overtwisted ) * , if and only if every nullhomologous legendrian knot fulfills the * bennequin inequality * @xmath77 , where @xmath78 denotes the genus of the knot . the strategy is then to find a legendrian knot that violates the bennequin inequality in the new surgered manifold , which then can not be contactomorphic to @xmath19 .
we want to present these situations in diagrams , called front projections .
observe that @xmath79 is contactomorphic to @xmath80 .
therefore , one can present knots in @xmath19 in their * front projection * @xmath81 .
if one parametrizes a knot as @xmath82 then the contact condition tells that the knot is legendrian if and only if @xmath83 . and similar such a knot is transverse if and only if @xmath84 .
this conditions give restrictions on the behavior of these knots in crossings and vertical tangencies in front projections ( * ? ? ?
* chapter 3 ) .
for example in figure [ fig : classificationoftransverseunknots ] the transverse unknots @xmath66 are pictured in their front - projections and in figure [ fig : crossings ] one can see the 4 possibilities of intersections of a legendrian knot @xmath69 and a transverse knot @xmath16 in the front - projection . from the conditions before it follows immediately that the intersection of type @xmath85 and @xmath86 is uniquely determined , but in type @xmath87 and @xmath88 in general both is possible .
but here it is enough to know the behavior in type @xmath85 and @xmath86 ( see also @xcite and @xcite ) .
consider now the front projection in figure [ fig : nsmallerzero ] of a transverse unknot @xmath16 of type @xmath66 and a legendrian unknot @xmath69 in the exterior of @xmath16 .
if one does a @xmath89-surgery along @xmath66 then the knot @xmath69 is ( topologically ) again an unknot in some new contact @xmath2 ( this can be seen by doing a * rolfsen twist * along @xmath66 , i.e. twisting @xmath90-times along the seifert disk of @xmath66 ) .
if the resulting manifold is again @xmath37 then the bennequin inequality holds , i.e. @xmath91 . on the other hand we compute @xmath92 with the above formula .
so it follows that for @xmath93 the resulting contact @xmath0-sphere is not the standard one .
for @xmath94 one considers the diagram from figure [ fig : nbiggerzero ] .
observe that this does not work for the transverse unknot with self - linking number @xmath1 , but for all other transverse unknots .
the linking number is @xmath95 and therefore the thurston
bennequin number of @xmath69 stays the same , i.e. @xmath96 . to show that the resulting contact @xmath0-sphere is always overtwisted we first determine the new knot type @xmath97 of @xmath69 in the new @xmath2 .
therefore , observe that topologically the link @xmath98 is the whitehead link , which can also be pictured like in figure [ fig : newlink](i ) . by doing a rolfsen twist along @xmath16
we get the new knot @xmath99 in the new @xmath2 pictured in figure [ fig : newlink](ii ) . of course here the bennequin inequality can not help because the minimal bound in the bennequin inequality is @xmath1 .
therefore , one has to use a finer bound for the thurston
bennequin number .
one such bound is the so - called * kauffman bound * @xcite , which says that for a legendrian knot @xmath69 in @xmath37 the following inequality holds @xmath100 where @xmath101 is the kauffman polynomial of @xmath69 . from the work of tanaka @xcite and yokota
@xcite it follows that in the case of a reduced alternating knot diagram @xmath102 of @xmath99 ( as is the case here ) this inequality transforms to @xmath103 where @xmath104 is the self - linking number of the knot diagram @xmath102 and @xmath105 is the number of regions in the knot diagram @xmath102 . by coloring the complement of an alternating knot diagram according to the rule in figure [ fig : regions](i ) one gets the regions as the colored areas .
( moreover , in @xcite and @xcite it is shown that this bound is sharp , but this is not important for the argument here . ) in figure [ fig : regions](ii ) one can count @xmath106 and @xmath107 .
therefore , every legendrian realization of @xmath99 in @xmath37 has @xmath108 .
but earlier i computed @xmath96 in the new surgered contact @xmath0-sphere .
therefore this new contact @xmath0-sphere has to be overtwisted .
the proof of the transverse knot exterior theorem [ cor : transverse ] works now very similar as for topological or transverse knots .
compare @xcite .
pick a contactomorphism between the exteriors @xmath110 and then consider the following commutative diagram : + + ( 0,40)*+(s^3,_st)=``a'';(8,40)*+=``a1 '' ; ( 24,40)*+(s^3_t_1 ( _ 1),_t_1 ( _ 1))=``b '' ; ( 40,40)*+:==``b1 '' ; ( 52,40)*+(s^1d^2,_s)=``c '' ; ( 75,40)*++=``d '' ; ( 100,40)*+(s^3,_st)=``e'';(114,40)*+/_~=``e1'';(53,35)*+_0=``f '' ; ( 99,35)*+_1=``g'';(75,20)*+=``k1'';(53,5)*+_0=``l '' ; ( 92,5)*+r_2:=h(_1)=``m'';(24,0)*+(s^3_t_2 ( r_2),_t_2 ( r_2))=``n'';(40,0)*+:==``n1 '' ; ( 52,0)*+(s^1d^2,_s)=``o '' ; ( 75,0)*++=``p '' ; ( 100,0)*+(s^3,_st)=``q'';(114,0)*+/_~=``q1'';@>_f `` b'';``n'';@->@/_1pc/ _ `` c'';``o'';@->@/^1pc/^h `` e'';``q'';@|->__1 `` f'';``g'';@|->^h_1 `` l'';``m '' ; + + here the contact solid torus @xmath111 is chosen to be equal to @xmath112 . because @xmath113 and @xmath114 on the two factors are contactomorphisms and because they both send the characteristic foliations of the boundaries to each other , these two maps glue together to a contactomorphism @xmath35 of the whole contact manifolds ( * ? ? ?
* section 2.5.4 ) . from the transverse contact dehn surgery theorem [ thm : transverse ]
it follows that @xmath115 is equal to @xmath116 , or @xmath34 is equivalent to the transverse unknot @xmath68 with self - linking number @xmath20 . if @xmath117 then this is a trivial contact dehn surgery , and so the contactomorphism @xmath35 maps @xmath33 to @xmath34 . in the other case
the same argument with @xmath33 and @xmath34 reversed shows that @xmath33 is also equivalent to the transverse unknot @xmath68 . @xmath109
maybe it is not directly clear where in the proofs it is important that the standard neighborhoods of the transverse knots are sufficiently small . to see why it is ,
observe that the proof of the transverse dehn surgery theorem shows a bit more .
it is not only shown that if there exists a cosmetic contact dehn surgery along a transverse knot in @xmath37 , then this knot has to be the transverse unknot @xmath68 , but also the surgery coefficient has to be of the form @xmath64 for @xmath94 ( so negative integers are excluded ) . at a first glance it looks as if this is a contradiction to the result of etnyre ghrist ( theorem [ thm : etnyre - ghrist ] ) about the existence of tight transverse knots .
but the surgeries in this theorem are done with respect to a sufficiently big tubular neighborhood of the transverse knot @xmath16 to ensure tightness of the resulting manifold as explained in ( * ? ? ?
* remark 2.1 ) .
by contrast , in the proof of the transverse dehn surgery theorem one has to choose sufficiently small neighborhoods of the transverse knot , in fact so small that one has enough space in the exterior for the legendrian knot @xmath69 .
both results together show that the legendrian knot @xmath69 from figure [ fig : nsmallerzero ] can not realized in a sufficiently small exterior of @xmath16 .
what is the maximal size of the tubular neighborhoods of @xmath16 , such that one can realize the legendrian knot @xmath69 from figure [ fig : nsmallerzero ] in its complement ? how is this size related to other notions of maximal neighborhoods of transverse knots ( for example in
@xcite ) ? observe , that if theorem [ thm : transverse ] holds for arbitrary neighborhoods then the transverse knot exterior problem is also true for arbitrary neighborhoods , since in its proof the size of the neighborhood is not used .
as already mentioned in the introduction from the work of edwards @xcite it follows that for topological knots two knot complements are homeomorphic if and only if the knot exteriors are homeomorphic . for transverse knots in contact manifold
this is not clear ( see ( * ? ? ?
* section 6.4 ) for more discussion on this problem ) .
in particular , the knot complement problem for transverse knots in @xmath19 remains open .
an obvious question is if the transverse knot exterior theorem is also true for links .
it will turn out that this is not the case . there
exist ( infinitly many ) non - equivalent transverse @xmath4-component links in @xmath19 with contactomorphic exteriors .
the idea for constructing such examples is the same as for topological links @xcite .
consider the so - called whitehead link from figure [ fig : newlink](i ) .
chose @xmath16 in @xmath19 to be a transverse unknot with @xmath20 with standard tubular neighborhood from theorem [ thm : etnyre - ghrist ] of etnyre - ghrist . so that contact @xmath89-dehn surgery along @xmath16 with respect to this standard tubular neighborhood yields again @xmath19 .
next chose a transverse realization of @xmath69 in the exterior of @xmath16 ( with respect to the chosen tubular neighborhood of @xmath16 ) . a contact @xmath89-dehn surgery along @xmath16 yields @xmath19 back but the transverse knot @xmath69 changes to a transverse knot with underlying topological knot type @xmath99 from figure [ fig : newlink](ii ) which is clearly non - equivalent to @xmath69 .
moreover from the proof of theorem [ thm : etnyre - ghrist ] it follows directly that the new glued - in solid torus is the standard tubular neighborhood of a new transverse knot @xmath119 .
again , it is unclear how the above example depends on the size of the tubular neighborhoods of the knots .
the neighborhood of @xmath16 has to be big ( such that the cosmetic contact dehn surgery of etnyre and ghrist works ) and the neighborhood of @xmath69 has to be so small , that it does not intersect the neighborhood of @xmath16 .
it remains open if such examples exist for arbitrary size of standard tubular neighborhoods of the transverse links .
another obvious question is , if transverse knots in general contact manifolds are also determined by the contactomorphism type of their exteriors . again , as for topological knots , this is in general not true .
consider a standard neighborhood @xmath22 of a transverse knot @xmath16 with size @xmath121 .
the characteristic foliation on the boundary @xmath122 is given by the vector field @xmath123 if one denotes the longitude @xmath124 of @xmath122 by @xmath125 , then the characteristic foliation of @xmath122 is given by the linear curves @xmath126 .
( -14,12)*+ = `` a0'';(-3,12)*+l(9,7)=``a5'';(8,12)*+==``a6'';(20,12)*+l(9,-2)=``a'';(32,12)*+==``a1 '' ; ( 40,12)*+v_1=``c '' ; ( 60,12)*++=``d '' ; ( 80,12)*+v_2=``e'';(95,12)*+/_~=``e1'';(40,6)*+_1=``f '' ; ( 80,6)*+-2_2 + 9_2,=``g'';(40,0)*+_1=``h '' ; ( 80,0)*+-_2 + 4_2.=``i'';@|- > `` f'';``g '' ; @|- > `` h'';``i '' ; + + the characteristic foliation given by @xmath129 of @xmath130 is mapped to the characteristic foliation given by @xmath131 of @xmath132 .
therefore , the contact structures on the two solid tori glue together to a global contact structure on the lens space @xmath133 . if one chooses the tubular neighborhoods of @xmath134 as @xmath135 , then by construction the corresponding exteriors are the same .
but it is easy to see that the knots @xmath33 and @xmath34 are not even topologically equivalent .
( a homeomorphism of the lens space mapping one knot to the other , restricts to the complementary solid torus to a map not isotopic to a composition of dehn twist , which is impossible .
see @xcite or ( * ? ? ?
* section 1.3 ) for details . )
j. baldwin and j. etnyre , admissible transverse surgery does not preserve tightness , _ math .
ann . _ * 37 * ( 2013 ) , 441468 .
j. conway , transverse surgery on knots in contact 3-manifolds , ` arxiv:1409.7077 ` ( 2015 ) .
m. dehn , ber die topologie des dreidimensionalen raumes , _ math .
* 69 * ( 1910 ) , 137168 .
f. ding and h. geiges , a legendrian surgery presentation of contact @xmath0-manifolds , _ math .
cambridge philos .
soc . _ * 136 * ( 2004 ) , 583598 .
s. durst and m. kegel , computing rotation and self - linking numbers in contact surgery diagrams , _ acta math .
_ * 150 * ( 2016 ) , 524540 .
j. edwards , concentricity in @xmath0-manifolds , _ trans .
* 113 * ( 1964 ) , 406423 .
y. eliashberg , legendrian and transverse knots in tight contact @xmath0-manifolds , in : _ topological methods in modern mathematics _
( stony brook , 1991 ) , publish or perish , houston ( 1993 ) , 171193 .
y. eliashberg and m. fraser , topologically trivial legendrian knots , _ j. symplectic geom .
_ * 7 * ( 2009 ) , 77127 .
j. etnyre and r. ghrist , tight contact structures via dynamics , _ proc .
_ * 127 * ( 1999 ) , 36973706 .
j. etnyre and k. honda , cabling and transverse simplicity , _ ann . of math .
( 2 ) _ * 162 * ( 2005 ) , 13051333 .
e. ferrand , on legendrian knots and polynomial invariants , _ proc .
soc . _ * 130 * ( 2002 ) , 11691176 .
d. gay , symplectic @xmath138-dimensional @xmath4-handles and contact surgery along transverse knots , ph.d . thesis , uc berkeley ( 1999 ) .
d. gay , symplectic @xmath4-handles and transverse links , _ trans .
* 354 * ( 2002 ) , 10271047 .
h. geiges , _ an introduction to contact topology _ , cambridge stud . adv .
* 109 * ( cambridge university press , 2008 ) .
c. gordon and j. luecke , knots are determined by their complements , _
j. amer .
* 2 * ( 1989 ) , 371415 .
m. kegel , the legendrian knot complement problem , ` arxiv:1604.05196 ` ( 2016 ) .
m. kegel , legendrian knots in surgery diagrams and the knot complement problem , doktorarbeit ( ph.d . thesis ) , universitt zu kln ( 2017 ) .
r. lickorish , a representation of orientable combinatorial @xmath0-manifolds , _ ann . of math .
( 2 ) _ * 76 * ( 1962 ) , 531540 .
y. mathieu , closed @xmath0-manifolds unchanged by dehn surgery , _
j. knot theory ramifications _ * 1 * ( 1992 ) , 279296
. y. rong , some knots not determined by their complements , in : `` quantum topology '' , ser .
knots everything * 3 * , ( world sci .
publishing , river edge , nj , 1993 ) , 339353 .
l. rudolph , a congruence between link polynomials , _ math .
soc . _ * 107 * ( 1990 ) , 319327 .
t. tanaka , maximal thurston
bennequin numbers of alternating links , _ topology appl .
_ * 153 * ( 2006 ) , 24762483 .
h. tietze , ber die topologischen invarianten mehrdimensionaler mannigfaltigkeiten , _ monatshefte fr mathematik und physik _ * 19 * ( 1908 ) , 1118 .
t. vogel , non - loose unknots , overtwisted discs , and the contact mapping class group of @xmath2 , ` arxiv:1612.06557 ` ( 2016 ) .
a. wallace , modifications and cobounding manifolds , _ canad .
_ * 12 * ( 1960 ) , 503528 .
j. whitehead , on doubled knots , _ j. london math .
* 12 * ( 1937 ) , 6371 .
y. yokota , the kauffman polynomial of alternating links , _ topology appl . _
* 65 * ( 1995 ) , 229236 . | we study cosmetic contact surgeries along transverse knots in the standard contact @xmath0-sphere , i.e. contact surgeries that yield again the standard contact @xmath0-sphere .
the main result is that we can exclude non - trivial cosmetic contact surgeries ( in sufficiently small tubular neighborhoods ) along all transverse knots not isotopic to the transverse unknot with self - linking number @xmath1 . as a corollary it follows that every transverse knot in the standard contact @xmath0-sphere is determined by the contactomorphism type of any sufficiently big exterior .
moreover , we give counterexamples to this for transverse links in the standard contact @xmath0-sphere and for transverse knots in general contact manifolds . |
we are grateful to a. e. dorokhov for many discussions and the collaboration in the appendix calculation .
this research was partially supported by the conselho nacional de desenvolvimento cientfico e tecnolgico ( cnpq ) , by coordenao de aperfeioamento do pessoal de ensino superior ( capes ) , by grant 2013/22079 - 8 of fundao de amparo pesquisa do estado de so paulo ( fapesp ) , and by fundao de amparo pesquisa do estado do rio grande do sul ( fapergs ) .
let us find the small @xmath93 behavior of @xmath8 by using the schwinger representation ( [ eq1l ] ) .
we also make the following approximation for the denominators of the integrand@xmath113 then , first we need to find the laplace transformation of the factor @xmath114@xmath115 ^{\gamma/2}}\\ & = \frac{\mu^{1/2}}{p^{2}+\mu^{2}}\frac{1}{\gamma_{\gamma/2}}\int d\alpha\alpha^{\gamma/2 - 1 } \times \\ & \,\,\,\,\,\ , \exp\left [ -\alpha\left ( 1+\overline{b}\ln\left ( \frac{p^{2}+\mu^{2}}{\mu^{2}}\right ) \right ) \right ] \\ & = \frac{\mu^{1/2}}{p^{2}+\mu^{2}}\frac{1}{\gamma_{\gamma/2}}\int d\alpha\alpha^{\gamma/2 - 1}e^{-\alpha}\left ( \frac{\mu^{2}}{p^{2}+\mu^{2}}\right ) ^{\alpha\overline{b}}\\ & = \frac{\mu^{1/2}}{\gamma_{\gamma/2}}\int d\alpha\alpha^{\gamma /2 - 1}e^{-\alpha}\left ( \mu^{2}\right ) ^{\alpha\overline{b } } \times \\ & \,\,\,\,\,\ , \frac{1}{\gamma\left ( 1+\alpha\overline{b}\right ) } \int d\beta e^{-\beta\left ( p^{2}+\mu^{2}\right ) } \beta^{\alpha\overline{b}}\\ & = \frac{\mu^{1/2}}{\overline{b}^{\gamma/2}}\int d\beta e^{-\beta p^{2}}e^{-\beta\mu^{2 } } \times \\ & \,\,\,\,\,\,\ , \int_{0}^{\infty}d\alpha e^{-\alpha/\overline{b}}\frac{\alpha^{\gamma/2 - 1}\left ( \mu^{2}\beta\right ) ^{\alpha}}{\gamma_{\gamma/2}\gamma\left ( 1+\alpha\right ) } \equiv\int d\beta e^{-\beta p^{2}}g_{1/2}\left ( \beta\right ) , \end{aligned}\ ] ] where @xmath116 and@xmath117 similarly we obtain the laplace transform of the factor @xmath118 as@xmath119 next we substitute these expressions in ( [ eq1l ] ) keeping in mind that we are interested in the @xmath83 behavior
. then one has for the integrand@xmath120 and for the pion da@xmath121 where @xmath122 in the asymptotic regime @xmath83 , we can take @xmath123 at @xmath124 and obtain @xmath125 this is the result given above in equation ( [ phito0as ] ) . note
, that this last result is proportional to @xmath126 with the argument @xmath127 of log substituted by @xmath128 . if @xmath129 then the convergence to this asymptotic behavior is rather slow .
f. halzen , g.i .
krein , a.a .
natale , phys .
d * 47 * ( 1993 ) 295 ; e.g.s .
luna , phys .
b * 641 * ( 2006 ) 171 ; a.a .
natale , pos qcd - tnt * 09 * ( 2009 ) 031 ; e.g.s .
luna , a.l .
dos santos , a.a .
natale , phys .
b * 698 * ( 2011 ) 52 ; e.g.s .
luna and p.c .
beggio , nucl .
phys . a * 929 * ( 2014 ) 230 . | the pion distribution amplitude ( da ) can be related to the fundamental qcd green s functions as a function of the quark self - energy and the quark - pion vertex , which in turn are associated with the pion wave function through the bethe - salpeter equation .
considering the extreme hard asymptotic behavior in momentum space allowed for a pseudoscalar wave function , which is limited by its normalization condition , we compute the pion da and its second moment . from the resulting amplitude , representing the field theoretical upper limit on the da behavior , we calculate the photon - pion transition form factor @xmath0 . the resulting upper limit on the pion transition form factor
is compared with existing data published by cleo , babar and belle collaborations .
a few years ago new data were published @xcite for the @xmath1 process , where one of the photons is far off mass shell ( large @xmath2 ) and the other one is near mass shell ( @xmath3 ) .
these measurements of the photon - pion transition form factor @xmath0 , taken in single - tagged two - photon @xmath4 reaction , was performed in a wide range of momentum transfer squared ( @xmath5 gev@xmath6 ) . at sufficiently high @xmath2
it is expected that the standard factorization approach can be applied @xcite ( for a review , see @xcite ) .
the amplitude for this process at high virtuality has the form @xmath7 this equation is obtained assuming factorization of the pion distribution amplitude ( da ) @xmath8 and the hard scattering amplitude @xmath9 given by @xcite @xmath10 .
\label{eqb}\ ] ] however the babar collaboration data @xcite seems to be in contradiction with this approach since , in accordance with perturbative qcd ( pqcd ) , @xmath11 should be limited to the value @xmath12 gev @xcite , hereafter called bl limit . at the same time , the belle collaboration data @xcite presented in the same range of transferred momenta show that the pion transition form factor may not increase as fast as shown by the babar results . as a consequence of these experiments
there were many theoretical papers speculating why the data should ( or not ) obey the bl limit @xcite . the first attempt to explain the babar result
can be found in ref . @xcite . among these
there were proposals claiming that the pion distribution amplitude should be modified @xcite , leading to a broader or flatter distribution in the place of the asymptotic form @xmath13 @xcite .
a flat da would be consistent with the babar data , although a field theoretical support for such possibility is still missing .
some papers claim that other transition form factors of heavier mesons are compatible among themselves and with the saturation required by factorization theorems obtained from pqcd @xcite .
however , for heavier mesons than the pion the da may be more peaked away from the end points @xcite .
a common statement in all papers is the need for more data to settle this problem .
meanwhile , the pion transition form factor is the most sensitive physical quantity to observe a non - perturbative contribution to the da .
other quantities , for instance , like the pion form factor , may already contain a hard scattering amplitude at leading order with a soft behavior , due to the effect of extra coupling constants or gluon propagators @xcite .
this means that they do not lead to such a simple integral over a da as the one shown by eq.([eqa ] ) . as claimed in ref.@xcite
, we may assume that at present there is no definite conclusion on which is the asymptotic form of the pion da , and it is possible that in the future a combined analysis of data of the processes involving pions will shed light on the pion distribution amplitude @xcite .
notwithstanding , considering the possibility that a flatter pion da seems to be favored by the babar data @xcite , we can establish a field theoretical limit on how flat this da can be , and , consequently , compare this limit to the experimental data . in order to do this
we will study the da dependence on the non - perturbative dynamics of the theory , and ultimately on the asymptotic behavior of the pion - quark vertex and the quark self - energy .
like in the nambu - jona - lasinio four - fermion approach , in qcd or any asymptotically free non - abelian gauge theory , the fermion masses are dynamically generated along with bound state goldstone bosons ( the pions ) .
the dynamical quark mass ( @xmath14 ) , giving by the schwinger - dyson equation is exactly identical to the pseudoscalar bethe - salpeter equation ( bse ) at zero momentum transfer ( @xmath15 ) , as demonstrated by delbourgo and scadron @xcite @xmath16 which is a consequence of the fact that they are related through the ward - takahashi identity .
the homogeneous bse can be , in general , written as @xmath17 where the amplitude depends on the quarks total ( @xmath18 ) and relative ( @xmath19 ) momenta , with @xmath20 , @xmath21 , and @xmath22 , where @xmath23 is the momentum fraction parameter . in eq.([eq02 ] ) @xmath24 is the fully amputated quark - antiquark scattering kernel , @xmath25 are the dressed quark propagators , and the homogeneous bse is valid on - shell , i.e. @xmath26 in the pion case .
note that we suppressed all indices ( color , etc ... ) in eq.([eq02 ] ) .
the bse , eq.([eq02 ] ) , is an integral equation that can be transformed into a second order differential equation .
the two solutions of the differential equation can be found , for example , in ref.@xcite and are characterized by one soft asymptotic solution @xmath27 and by the extreme hard high energy asymptotic behavior of a bound state wave function @xmath28 ^{-\gamma}\ , , \label{eq1a}\ ] ] where @xmath29 , @xmath30 is the casimir eigenvalue for quarks in the fundamental representation ( @xmath31 is the number of colors , and @xmath32 is the number of quark flavors ) , and @xmath33 .
the asymptotic expression shown in eq.([eq1a ] ) was determined in the appendix of ref.@xcite and it satisfies the callan - symanzik equation .
this last solution is constrained by the bse normalization condition @xcite , which imply @xmath34 , or @xmath35 @xcite , otherwise it is not consistent with a possible bound state solution in a @xmath36 non - abelian gauge theory .
we will take @xmath37 as will be explained later .
this solution is one alternative to the soft one ( @xmath38 ) @xcite which leads to the standard da @xmath39 .
nowadays it is known that we may have solutions with a momentum behavior varying between eq.([eq03 ] ) and eq.([eq1a ] ) depending on the theory dynamics @xcite .
note that the bethe - salpeter equation ( bse ) can be transformed into a second order differential equation .
this equation has two possible solutions , one that asymptotically behaves as @xmath40 as in eq.(5 ) and the other one as @xmath41^{-\gamma}$ ] ( eq.(6 ) ) .
however this result comes out from the homogeneous bse .
the non - homogeneous bse also includes a normalization condition , as discussed in refs.[27,30,31 ] , that is obeyed by eq.(5 ) but when applied to eq.(6 ) implies @xmath42 .
this constraint appears because the wave function is very hard " , i.e. decreases very slowly with the momentum and can not be normalized ( square integrable ) if @xmath43 .
this condition on @xmath44 gives the bound @xmath42 .
this limit on @xmath44 was obtained by mandelstam in ref.[30 ] , in qcd for the first time in ref.[27 ] and recently , in a different context , in ref.[31 ] .
if @xmath45 only the solution of eq.(5 ) exists , because it would be the only one obeying the bse normalization condition .
this also means that if @xmath46 qcd may have a chiral broken phase whose self - energy is given by eq.(6 ) .
nowadays it is known that the chiral phase diagram for a non - abelian theory may change considerably as we change the number of flavors .
for instance , if the theory contains contributions of higher order operators it may have its quark self - energy or bound state solution varying between eq.(5 ) and ( 6 ) as discussed in ref.[33 ] .
we are just saying that if @xmath42 eq.(6 ) is a possible bound state solution , and the hardest one that we may have .
it has been argued that eq.([eq1a ] ) may be a realistic wave function in a scenario where the chiral symmetry breaking is associated to confinement and the gluons have a dynamically generated mass @xcite .
this solution also appears when using an improved renormalization group approach in qcd , associated to a finite quark condensate @xcite , and it minimizes the vacuum energy as long as @xmath35 @xcite .
moreover , this specific solution is the only one consistent with regge - pole like solutions @xcite .
the important fact is that this is the hardest ( in momentum space ) asymptotic behavior allowed for a bound state solution in a non - abelian gauge theory , and it is exactly for this reason that the constraint on @xmath44 arises from the bse normalization condition .
no matter this solution is realized in nature or not , it will lead to the flattest pion da , any other flatter distribution than this one can not be a realistic bse wave function , and would not be consistent with a composite pion .
a totally flat da can only be related to a fundamental pion . a realistic da , in principle ,
should be related to a solution of the bse and should obey a normalization condition peculiar to a well behaved wave function .
the infrared behavior of the gap equation ( or bse ) is approximately constant at small momenta , @xmath47 , where @xmath48 , of order of a few hundred mev , is the characteristic scale of dynamical quark mass generation . in order to compute the pion da
we will perform an integral over the wave function in the full range of momenta ( i.e. up to @xmath49 , this is why we will consider @xmath50 ) . to obtain the extreme field theoretical limit on the pion da , we shall also work with a simple interpolating expression that roughly reflects the full behavior of the hardest " quark self - energy ( or bse solution ) discussed in the previous paragraphs , namely @xcite @xmath51 ^{-\gamma}\ , .
\label{eq2}\ ] ] note that the @xmath48 factor introduced into the logarithm denominator leads to the right infrared ( ir ) behavior ( @xmath52 ) .
furthermore , this is just one possible ansatz for the full behavior of the self - energy and other possible interpolations between the ir and ultraviolet ( uv ) behaviors are possible , but as long as @xmath14 shows the logarithmic uv behavior our final result will not change .
the coupling constant @xmath53 is calculated at the chiral symmetry breaking scale @xmath48 , and given by @xmath54}\ , , \label{eq6}\ ] ] which is an infrared finite coupling determined in qcd where gluons have an effective dynamical mass @xmath55 @xcite and is consistent with the models of ref.@xcite .
@xmath56 is the qcd characteristic scale . within this approach ,
the pion distribution amplitude at leading twist , as a function of the quark self - energy and the pion - quark vertex , is given by @xcite @xmath57 \ , , \label{eq1}\ ] ] where the @xmath58-variable plays the role of the quark transverse momentum squared , @xmath59 and @xmath60 are the longitudinal projections of the quark momentum on the light cone directions ( @xmath61 ) , @xmath62 is the dynamical quark mass given by ( [ eq2]),@xmath63 and the function @xmath64 is the momentum dependent part of the quark - pion vertex , which can be approximated by @xmath65 , where @xmath66 and @xmath67 are the quark and anti - quark momenta .
the pion da at leading twist is normalized as @xmath68 it is also useful to write down the expression for the pion da in the form found in @xcite ( the so called schwinger representation ) _ ( x ) & = & _ 0^e^xlp^2 , [ eq1l]where @xmath69 and @xmath70 are the laplace transformations of @xmath71 correspondingly . for example,@xmath72 for the model calculations we take the following parameters : @xmath73 mev , @xmath74 mev and @xmath75 mev @xcite . to describe fig .
( 1 ) we used @xmath76 mev and for @xmath50 a more appropriate value would be @xmath77 mev @xcite .
however it should be noticed that the result is more dependent on the ratio @xmath78 than on the proper @xmath56 value .
we also emphasize that the largest origin of uncertainty in our result is the assumption of eq .
( 7 ) for the self - energy in the full range of momenta . in the intermediate and infrared region of momenta eq.(7 ) may give a poor description of the self - energy , although the good point is that the flat da behavior is totally credited to the hard asymptotic self - energy behavior .
within the model considered we take the expression for the pion decay constant in the so - called pagels - stokar form @xcite @xmath79 where @xmath80 . with the given set of parameters
we obtain @xmath81 mev . actually , it is possible to obtain this @xmath82 value with different values for @xmath48 and @xmath55 , just changing the formula that interpolates between the ir and uv regimes , although these values should stay around a few hundreds of mev .
the pion da obtained with above parameters and eqs.([eq2 ] ) and ( [ eq1 ] ) is shown in fig.([fig1 ] ) ; for comparison we also draw the asymptotic wave function @xmath39 .
the da turns out to be quite flat , and we have not observed any significant variation as we change @xmath55 and @xmath48 as long as we do not modify the @xmath82 value in eq.([eq1 ] ) and maintain the logarithmic uv behavior , _ which is at the origin of the flat da behavior_. the da flatness is totally dependent on the logarithmic behavior of the self - energy .
the asymptotic behavior as @xmath83 for the model considered here is given by @xmath84 where @xmath85 .
this behavior is actually the expected one for a pion da with a vertex function @xmath86 similar to ours , where @xmath86 goes to zero in the limit @xmath87 @xcite . in the appendix
we determine the asymptotic behavior shown in eq.([phito0as ] ) . ) and ( [ eq2 ] ) , and using the parameters @xmath88 mev and @xmath73 mev ( solid curve ) .
the perturbative - qcd asymptotic pion da , @xmath13 , is shown by the dashed curve for comparison . ]
our pion da numerical result can be reasonably reproduced by using the normalized form @xmath89 where @xmath90 .
however , it is worth noting that the calculations performed in this work have been carried out using numerical values of the pion da obtained from eq .
( [ eq1 ] ) .
the leading asymptotic behavior of the form factor is expressed through the pion da ( [ eq3a ] ) as @xcite @xmath91 if we were considering a totally flat da this integral would diverge . however , as emphasized by radyushkin @xcite , the finite size @xmath92 of the pion should provide a cut - off for the @xmath93 integral .
therefore the @xmath94 in the denominator of eq.([asf ] ) will be changed as xq^2xq^2 + m^2 .
[ eqxx ] in principle the factor @xmath95 should be related to the dynamical quark mass .
it was also proposed by radyushkin that @xmath95 could be treated as an effective gluon mass .
indeed the meson radius may have a deep connection with the effective gluon mass as discussed in @xcite .
therefore , no matter we have one case or another , the transition form factor will be giving by f_^ ( 0;q^2,0 ) = f__0 ^ 1dx .
[ fasmod]@xmath95 being a dynamical mass should have a momentum dependence showing the decrease of the mass with the momentum .
however when @xmath94 is small we can safely substitute @xmath96 by @xmath95 in eq.([fasmod ] ) , and for large @xmath94 the momentum behavior of @xmath96 is negligible .
the result for the transition form factor given by eq.([fasmod ] ) is shown in fig.([fig2 ] ) . in fig.([fig2 ] ) we are just assuming @xmath97 mev no matter this is a quark or gluon dynamical mass .
we are neglecting the effects of the qcd evolution in eq.(19 ) and , consequently , in fig.(2 ) , because , as verified by radyushkin @xcite , these effects are small for such a flat da . computed with the extreme da calculated in this work .
the continuous red curve is the limit obtained with eq.([fasmod ] ) .
the horizontal hatched area correspond to the asymptotic bl limit .
the experimental data are taken from cleo ( @xmath98 ) , belle ( @xmath99 ) and babar ( @xmath100 ) collaborations . ]
the result of fig.([fig2 ] ) for the photon - pion transition form factor can be compared to the parameterization fit of babar @xcite and belle @xcite collaborations , @xmath101 where @xmath102 gev and @xmath103 for babar , and @xmath104 gev and @xmath105 for belle . with our numerical pion da
we can also compute the second moment of the da , which is given by @xmath106 where @xmath107 .
our result can be compared to the lattice value @xmath108 @xcite , where the first error is statistical and the second is systematic .
the @xmath109 lattice value has been computed at the scale @xmath110 gev@xmath6 , while our da calculation is obtained at the scale of the chiral symmetry breaking . since the da evolution with the momentum can only diminishes the ir value of the integral in eq.([eq3a ] ) we can consider our result as an upper limit on the value of the second moment of the da , i.e. @xmath111 which is the highest value for all the parameters that we have considered .
radyushkin @xcite and polyakov @xcite have proposed a flat behavior for the pion da in the light of babar data .
this behavior , if it indeed happens in nature , should be justified field theoretically .
we have verified that a flat behavior for the pion da is totally dependent on the asymptotic uv logarithmic behavior of the self - energy . when the pion da is calculated with eq.([eq1 ] ) using softer quark self - energies the result turns out to be more peaked at @xmath112 .
the effect of the self - energy behavior at small and intermediate momenta are erased by the normalization condition eq.([eq1b ] ) .
the high energy limit of the pion transition form factor with a flat da must have a natural cut - off ( @xmath95 ) as proposed by radyushkin @xcite and polyakov @xcite , and this one was introduced by us into eq.([fasmod ] ) .
we have assumed a cut - off of the order of the dynamical masses of the theory , while in refs.@xcite these values are quite high and difficult to associate to some physical scale .
this is why our calculation can only be considered as an upper limit on the pion transition form factor . if we had assumed a larger @xmath95 value we would have a better agreement with the babar data
of course , radiative corrections may also bring down the red curve in fig.([fig2 ] ) implying a better adjust of the data .
hence our result provides a field theoretical limit on the pion da and transition form factor based on the hardest asymptotic quark self - energy allowed in qcd .
we also obtained a limit on the second moment of the pion da compatible with the one of recent qcd lattice simulation . |
polarized transport in spin - dependent nanostructures is a subject of intense study in the emerging field of spintronics,@xcite due to its relevance to the development of novel spin - based devices.@xcite in addition , transport through qds provides information about fundamental physical phenomena in spin - dependent and strongly correlated systems , such as the kondo effect,@xcite the coulomb- and spin - blockade effects,@xcite spin valve effect and tunnelling magnetoresistance ( tmr),@xcite etc .
novel spin filters and pumps have also been proposed using qds coupled to normal metal leads.@xcite a system of particular interest in this context comprises a quantum dot or a metallic grain coupled to ferromagnetic leads .
the ferromagnetism of the leads introduces spin - dependent tunneling rates between the leads and the central region .
this results in a nonzero net spin in the central region for asymmetric magnetization geometries .
this effect is called spin accumulation or spin imbalance.@xcite it has been shown that spin accumulation affects several transport properties , such as magnetoresistance,@xcite ( negative ) differential resistance@xcite and the zero - bias anomaly.@xcite in addition it provides a way to generate and control the current spin polarization via gates or bias voltages,@xcite which is one of the main goals within spintronics .
systems composed of a nonmagnetic lead and a ferromagnetic lead with a quantum dot or a quantum wire as spacer have been analyzed recently .
it was pointed out that if the spacer is a dot and the ferromagnetic lead is half - metallic , a novel mesoscopic current - diode effect arises.@xcite spin - current rectification was also predicted in an asymmetric system composed of a ferromagnetic ( fe or ni ) and nonmagnetic ( au or pd ) contacts coupled to each other via a molecular wire.@xcite additionally , it was pointed out that a nm - qd - fm system can operate as a spin - filter and as a spin - diode.@xcite in ref .
[ ] the authors use the bias voltage to change the resonance position of the dot level with respect to the spin - split density of states of the ferromagnetic lead .
this gives rise to spin - dependent currents .
here we study spin - resolved currents in a single - level quantum dot attached to a nonmagnetic lead ( left lead " ) and to a ferromagnetic lead ( right lead " ) , fig .
1 . as we shall show , the magnetic asymmetry between the left and right terminals results in a rectification of the current polarization for a particular bias range for which the single electron channel @xmath2 is on - resonance , and the double - occupancy channel @xmath3 is off - resonance .
more precisely , when the nonmagnetic lead operates as an emitter and the ferromagnetic lead as a collector , defined as the positive bias ( @xmath4 ) , the current is unpolarized .
in contrast , when the ferromagnetic lead is the emitter and the nonmagnetic lead is the collector ( negative bias ) a spin - polarized current arises .
importantly , this rectification occurs only in this particular bias range , as we shall demonstrate both analytically and numerically .
this is attributed to an interplay between nonequilibrium spin accumulation and coulomb interaction within the dot . for high enough bias voltages ,
the current polarization is essentially symmetric with respect to the bias , and no rectification is found . in the main body of the text
we employ the master - equation approach of glazman and matveev@xcite to describe the spin - dependent transport through the nm - qd - fm junction in the sequential tunneling regime ( @xmath5,@xcite where @xmath6 is a characteristic tunneling rate ) . an alternative description in terms of nonequilibrium green functions
is also presented in the appendix , that corroborates our results obtained via master equation .
the nm - qd - fm system we study is schematically illustrated in fig .
an external bias voltage @xmath7 drives the system away from equilibrium thus imposing a chemical potential imbalance between the left ( l ) and the right ( r ) leads : @xmath8 , where @xmath9 is the chemical potential of the lead @xmath10 and @xmath11 is the absolute value of the electron charge ( @xmath12 ) .
the system hamiltonian is @xmath13 where @xmath14 is the free - electron energy with wave vector @xmath15 and spin @xmath16 in lead @xmath17 ( @xmath18 ) , @xmath2 is the spin - degenerate dot level , @xmath19 is the on - site coulomb repulsion and the operators @xmath20 ( @xmath21 ) and @xmath22 ( @xmath23 ) destroy ( create ) an electron with spin @xmath16 in the lead @xmath17 and in the dot , respectively .
the matrix element @xmath24 gives the lead - dot coupling .
we do not consider any spin - flip processes . to calculate the current we use rate equations,@xcite which yield @xmath25\nonumber\\ & & + e[\widetilde{\g}_{01\s}^\eta(n_{\overline{\s}}-n_{\uparrow \downarrow } ) - \widetilde{\g}_{10\s}^\eta n_{\uparrow \downarrow}],\end{aligned}\ ] ] where we have assumed @xmath26 .
the parameter @xmath27 corresponds to the rate of adding one electron to the dot coming from lead @xmath17 , and @xmath28 is the rate of moving one electron from the dot to lead @xmath17 .
in addition , @xmath29 and @xmath30 give the rates of moving one electron with spin @xmath16 to and from the dot , respectively , when it is already occupied by one electron with opposite spin .
following ref .
[ ] we define @xmath31 and @xmath32 ( @xmath33 ) as the dot single and double average occupancies , respectively .
the tunneling rates are @xmath34 where @xmath35 + 1\}$ ] and @xmath36 + 1\}$ ] .
the rates @xmath37 and @xmath38 are related to the spin - resolved density of states of lead @xmath17 via @xmath39 . here
we assume @xmath40 and @xmath41 .
this reflects the fact that the density of states of the left lead is spin - degenerate while the right one is spin - split . assuming a constant density of states and a constant tunneling parameter @xmath42
, we have @xmath43 . with this assumption the terms with @xmath44 in eq .
( [ isetageneral ] ) cancel out,@xcite and one simply finds @xmath45.\end{aligned}\ ] ] to calculate the current via eq .
( [ iseta ] ) we need to find @xmath46 from@xcite @xmath47-\g_{10\s } [ n_\s - n_{\uparrow \downarrow}]\nonumber\\&&\phantom{xxx}+\wid{\g}_{01 \s } [ n_{\bar{\s}}-n_{\uparrow \downarrow}]-\wid{\g}_{10\s } n_{\uparrow \downarrow},\end{aligned}\ ] ] where @xmath48 when @xmath49 , eq .
( [ eqns1general ] ) becomes @xmath50-\g_{10\s } n_\s + \widetilde{\g}_{01 \s } n_{\bar{\s}},\ ] ] where the terms with @xmath44 cancel out .
_ stationary regime_. in this regime ( @xmath51 ) eq .
( [ eqns1 ] ) reduces to @xmath52 which can be solved for each spin component , thus resulting in @xmath53 where @xmath54 . using eq .
( [ nsfinal ] ) into eq .
( [ iseta ] ) we obtain @xmath55 from eq .
( [ ifinal ] ) we can readily evaluate the current polarization @xmath56 .
next ( sec .
iii ) we provide some simple analytical results valid when double - occupancy is energetically forbidden .
numerical results are presented in sec .
as we shall see , the most interesting behavior takes place when the channel @xmath2 is completely within the conduction window and @xmath3 is far above the fermi energy of the emitter . with this channel configuration
we approximate @xmath57 , and @xmath58 , @xmath59 for @xmath4 and @xmath60 , @xmath61 for @xmath62 . using this into the occupation and current equations , eqs .
( [ nsfinal ] ) and ( [ ifinal ] ) , we find analytical expressions for the first plateau that appears in the current and its polarization for both positive and negative bias voltage .
equation ( [ nsfinal ] ) then becomes @xmath63 where @xmath64 , @xmath65 for @xmath4 and @xmath66 , @xmath67 for @xmath62 .
the current of the left lead then becomes @xmath68 where @xmath66 and the @xmath69 sign corresponds to @xmath4 , while @xmath64 and the @xmath70 sign to @xmath62 .
the right side current is simply given by @xmath71 for a spin - conserving stationary regime .
equation ( [ closedi ] ) gives the ( bias - independent ) current in the regime addressed here .
for the particular case of spin - independent tunneling rates , i.e. , @xmath72 and @xmath73 , we obtain @xmath74 in accordance with results already known in the literature.@xcite using eq .
( [ closedi ] ) into the definition @xmath75 , we obtain the current polarization plateau @xmath76 where @xmath64 for @xmath4 and @xmath66 for @xmath62 .
we model the tunneling rates by @xmath77 and @xmath78 , where @xmath79 $ ] is the spin polarization degree of the ferromagnetic right lead@xcite and @xmath80 the lead - dot coupling . within this model eq .
( [ somepol ] ) gives @xmath81 thus , when only the level @xmath2 is within the conduction window , the current becomes unpolarized for positive bias , while spin - polarized for negative bias .
therefore the nm - qd - fm junction functions as a _ current - polarization diode_.
we assume that the dot level depends on the bias voltage according to @xmath82 , where @xmath83 accounts for asymmetric voltage drops along the left and right tunnel barriers.@xcite @xmath84 can be controlled via gate voltages . for the numerics
we take @xmath85 mev , @xmath86 , @xmath87 , @xmath88 @xmath89ev , @xmath90 mev and @xmath91ev.@xcite in secs .
b , c and d we assume a symmetric potential drop across the system with @xmath92 . in sec .
e we briefly discuss the asymmetric case with @xmath93 .
figure [ fig2 ] shows the current polarization as a function of the external bias @xmath94 .
we observe that for positive bias the current polarization decreases for increasing bias , reaching zero around @xmath954 mev .
conversely , for the negative biases we obtain a maximum polarization @xmath96 around @xmath97 mev , confirming the analytical result found in sec .
( [ polresults ] ) .
the voltage range for this behavior scales with the parameter @xmath19 . for high enough bias voltages ( @xmath98 7 mev ) the polarization reaches the same nonzero plateaus for both positive and negative voltages .
both the suppression ( @xmath4 ) and the enhancement ( @xmath62 ) of the current polarization are due to the interplay of coulomb interaction and spin accumulation in the quantum dot .
quite interestingly this interplay affects @xmath1 differently with the bias sign , namely , for direct bias it suppresses @xmath1 while for reverse bias it enhances @xmath1.@xcite the suppression of @xmath1 for positive bias results in the zero polarization seen for all @xmath96 values except @xmath99 . in the half - metallic case ( @xmath99 ) ,
there is only spin up current flowing in the system ( @xmath100 , @xmath101 ) , so the polarization becomes simply @xmath102 . on the other hand , for negative bias , the maximum polarization plateau changes as @xmath96 varies .
in particular , @xmath1 attains a plateau equal to the polarization degree of the ferromagnetic lead , according to eq .
( [ polresults ] ) .
to gain a more detailed understanding of the spin - diode effect we investigate next the spin accumulation @xmath103 and the spin - resolved @xmath104 curves as a function of the bias .
figure [ fig3 ] shows the spin accumulation @xmath105 as a function of the bias voltage , for distinct polarization parameters @xmath96 .
for all the @xmath96 values considered here we note that @xmath106 for positive bias and @xmath107 for negative bias .
this spin - imbalance can be understood in terms of the tunneling rates @xmath37 between dot and leads . due to the ferromagnetism of the right lead ,
the rates @xmath108 and @xmath109 become asymmetric .
for example , for @xmath110 the rates are @xmath111 @xmath89ev , @xmath112 @xmath89ev and @xmath113 @xmath89ev . for positive bias ,
@xmath108 becomes the ingoing tunneling rate for electrons with spin @xmath16 and @xmath109 the outgoing tunneling rate .
due to the inequality @xmath114 , the spin up electrons can tunnel out the dot faster than they come into it .
on the other hand , since @xmath115 , the spin down electrons leave the dot slower than they come into it .
so on average the spin down electrons spend more time in the dot than the spin up ones for @xmath4 , thus @xmath116 .
a similar reasoning applies to the other @xmath96 values , except for @xmath117 for which there is no accumulation . for negative bias ,
@xmath118 and @xmath119 are the outgoing tunneling rates while @xmath120 and @xmath121 become the ingoing tunneling rates . as a consequence of this interchange ,
the spin accumulation inverts its sign @xmath122 . for small @xmath96 values
the spin accumulation is essentially an odd function of the bias , fig .
[ fig3 ] . when @xmath96 increases , though , the imbalance becomes stronger for positive bias .
in particular for @xmath99 , @xmath123 reaches @xmath124 in the positive bias range corresponding to single occupancy ( @xmath125 ) , and a constant plateau for all negative bias .
this happens because no spin - down states are available in the right lead for @xmath99 , so a spin - down electron that enters the dot , coming from the left side ( @xmath4 ) , can not leave the dot to the right side .
hence a spin - up electron can not hop into the dot when @xmath125 , so the accumulation becomes completely spin - down polarized for positive bias . for high enough bias voltages an additional electron with opposite spin can jump into the dot ( for both positive and negative bias ) , thus resulting in a suppression ( in modulus ) of @xmath123 . in figure [ fig4 ]
we show the spin resolved currents @xmath126 and @xmath127 as a function of the bias voltage for differing polarization parameters @xmath96 .
we observe that for positive bias the spin up and spin down currents coincide in the plateaus indicated by arrows for any @xmath96 value .
this results in the zero current - polarization seen in fig .
[ fig2 ] . in the second plateau ,
though , @xmath126 attains higher values compared to @xmath127 , which enhances @xmath1 .
the strong suppression of @xmath126 in the first plateau ( @xmath4 ) is attributed to the spin imbalance @xmath106 observed for the corresponding bias range ( see fig .
[ fig3 ] ) .
more specifically , since the dot is predominantly spin - down occupied for positive bias , the spin - up electrons tend to be more blocked than the spin - down ones , thus reducing further @xmath126 and interestingly locking it on top of @xmath127 . in contrast , for negative bias we have the population inversion @xmath107 .
this gives a stronger suppression of @xmath127 as compared to @xmath126 , which enhances the difference between @xmath126 and @xmath127 , and consequently @xmath1 .
when the channel @xmath3 reaches resonance ( @xmath128 mev ) both the @xmath126 and @xmath127 plateaus attain values somewhat closer to each other , thus reducing the current polarization ( see fig .
[ fig2 ] ) . in particular for @xmath99 the @xmath127 is zero for any bias voltage since there are no spin - down states available in the right lead .
the @xmath126 increases slightly ( for positive bias ) while the dot is becoming populated .
when the population is high enough the coulomb interaction plays a role and the spin up current goes down to zero.@xcite this gives rise to a negative differential conductance at the beginning of the first plateau for @xmath4 [ see fig .
[ fig4 ] with @xmath99].@xcite for negative bias ( and @xmath99 ) @xmath126 attains one plateau instead of two steps as for the other @xmath96 values .
this is expected because the spin - down electrons do not participate in the transport in this case , so no coulomb interaction effect arises .
note that for @xmath99 the system can operate as a mesoscopic current diode.@xcite here we consider the effects of an asymmetric bias drop , i.e. , @xmath130 . as fig .
[ fig5 ] shows , the asymmetry in the bias drop gives rise to quantitative , but not qualitative changes .
for @xmath131 the current polarization @xmath1 goes to zero much slower with the bias than it does for @xmath92 .
this is so because the resonance condition @xmath132 ( @xmath4 ) , which is necessary to have @xmath133 [ see sec.([analytical ] ) ] , happens for higher bias when @xmath83 decreases . for negative bias the resonance @xmath134
is reached faster ( i.e. , at lower biases as compared with the @xmath92 case ) for decreasing @xmath83 .
this , in turn , translates into a steeper enhancement of @xmath1 which then attains a plateau at @xmath135 [ see eq .
( [ polresults ] ) ] .
in addition , for @xmath131 the zero current - polarization plateau ( @xmath4 ) enlarges while the maximum plateau ( @xmath62 ) shrinks compared to the respective @xmath92 widths . for @xmath136
the resonance @xmath132 ( @xmath4 ) takes place faster with the bias when compared to the @xmath131 and @xmath92 cases .
this results in the steeper suppression of @xmath1 and the shrinkage of the zero current - polarization bias range . for negative biases
the resonance condition @xmath134 for @xmath136 is more slowly attained with the bias as compared to the @xmath131 and @xmath92 cases .
consequently , the polarization @xmath1 reaches the plateau at @xmath135 for higher bias voltages ( in modulus ) .
we propose a nm - qd - fm system which operates as a diode for the current polarization .
more specifically , when double - occupancy is forbidden in the system , i.e. , the channel @xmath3 is far above the chemical potential of the emitter lead , the system carries an unpolarized current for positive bias and a spin - polarized current for negative bias .
this effect is a result of the interplay between spin accumulation in the dot and the coulomb interaction .
interestingly , for positive biases the spin - resolved currents @xmath126 and @xmath127 lock onto the same plateau for a particular bias range , thus resulting in @xmath133 .
the authors acknowledge g. platero , k. flensberg , t. novotn , and j. p. morten for helpful comments .
fms and jce acknowledge the kind hospitality at the centre for advanced study ( oslo ) during the revision stage of this work .
jce acknowledges support from cnpq and fapesp .
in the main text we have formulated the problem via master equation formalism . here
we show that eqs .
( [ nupdwsolved1 ] ) and ( [ closedi ] ) , from which our main result eq .
( [ polresults ] ) directly follows , can be obtained via the keldysh formalism .
we start with the well known equation for the stationary current@xcite @xmath137f_\eta(\e)+g_{\s\s}^<(\e)\}\}\nonumber,\ ] ] where @xmath138 , @xmath139 and @xmath140 are the retarded , advanced and lesser green functions , respectively . to calculate these
we apply the equation of motion technique and use the hartree - fock approximation to factorize high - order correlation functions in the resulting chain of equations.@xcite the retarded green function becomes @xmath141 where @xmath142 is the non - interacting tunneling self - energy given in the wide band approximation by @xmath143 , and @xmath144 is the dot green function without coupling to leads , @xmath145 where @xmath146 is the dot occupation number , with @xmath147 .
this occupation can be calculated self - consistently via @xmath148 where the correlation function @xmath149 is given by the keldysh equation @xmath150 the advanced green function @xmath151 is given by @xmath152^*$ ] , while @xmath153 $ ] . in order to consider the same channel configuration adopted in sec .
iii , we assume a large @xmath19 ( @xmath154 ) so the retarded green function becomes @xmath155 and the lesser green function reads @xmath156 substituting eq .
( [ glesserlargeu ] ) into eq .
( [ nintglesser ] ) we have @xmath157 now assuming that the dot level @xmath2 is completely on resonance within the conduction window between @xmath158 and @xmath159 for positive or negative bias@xcite we can integrate eq .
( [ nsintegral ] ) in order to obtain @xmath160 where @xmath64 for @xmath4 or @xmath66 for @xmath62 .
solving eq .
( [ nsafterint ] ) for each spin component we find exactly eq .
( [ nupdwsolved1 ] ) . to obtain eq .
( [ closedi ] ) from the green functions , we substitute eqs .
( [ gdotlargeu ] ) and ( [ glesserlargeu ] ) into the current formula ( [ currentgf ] ) , which gives @xmath161 solving eq . ( [ closedinegf ] ) with the same assumptions adopted
previously@xcite we find @xmath162 where @xmath69 and @xmath70 signs correspond to @xmath4 and @xmath62 , respectively .
( [ islafterint ] ) can also be written as eq .
( [ closedi ] ) . for an overview
see e.g. _ semiconductor spintronics and quantum computation _ , eds .
d. d. awschalom , d. loss , and n. samarth , springer , berlin , 2002 ; i. zutic , j. fabian , and s. das sarma , rev .
76 * , 323 ( 2004 ) .
f. m. souza , j. c. egues , and a. p. jauho , _ proceedings of the 26th international conference on the physics of semiconductors _
( icps 26 ) , edinburgh 2002 , edited by j. h. davies and a. r. long .
see cond - mat/0209263 .
typical values of @xmath80 and @xmath19 can be found in , for example , a. kogan _
et al . _ ,
phys . rev .
lett . * 93 * , 166602 ( 2004 ) ; f. simmel _
* 83 * , 804 ( 1999 ) ; d. goldhaber - gordon _ et al .
_ , nature * 391 * , 156 ( 1998 ) . in the context of nonequilibrium green function ( see appendix ) we note that by increasing the coupling parameter @xmath80 , the width of the dot level increases as well , thus resulting in more smooth curves in figs .
[ fig2]-[fig5 ] . for a system with two ferromagnetic leads coupled to a quantum dot ( fm - qd - fm ) , it is possible to have an enhancement of @xmath1 for negative biases and a suppression of @xmath1 for positive biases when the lead magnetizations are aligned antiparallel .
however , the polarization does not vanish for any particular bias range and consequently no rectification effect is observed - f. m. souza , phd thesis , university of so paulo , brazil ( 2004 ) . | we report a study of spin dependent transport in a system composed of a quantum dot coupled to a normal metal lead and a ferromagnetic lead ( nm - qd - fm ) .
we use the master equation approach to calculate the spin - resolved currents in the presence of an external bias and an intra - dot coulomb interaction .
we find that for a range of positive external biases ( current flow from the normal metal to the ferromagnet ) the current polarization @xmath0 is suppressed to zero , while for the corresponding negative biases ( current flow from the ferromagnet to the normal metal ) @xmath1 attains a relative maximum value .
the system thus operates as a rectifier for spin current polarization .
this effect follows from an interplay between coulomb interaction and nonequilibrium spin accumulation in the dot . in the parameter range
considered , we also show that the above results can be obtained via nonequilibrium green functions within a hartree - fock type approximation . |
Subsets and Splits